linux/net/dsa/port.c

2076 lines
50 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Handling of a single switch port
*
* Copyright (c) 2017 Savoir-faire Linux Inc.
* Vivien Didelot <vivien.didelot@savoirfairelinux.com>
*/
#include <linux/if_bridge.h>
#include <linux/netdevice.h>
#include <linux/notifier.h>
#include <linux/of_mdio.h>
#include <linux/of_net.h>
#include "dsa.h"
#include "port.h"
#include "switch.h"
#include "tag_8021q.h"
#include "user.h"
/**
* dsa_port_notify - Notify the switching fabric of changes to a port
* @dp: port on which change occurred
* @e: event, must be of type DSA_NOTIFIER_*
* @v: event-specific value.
*
* Notify all switches in the DSA tree that this port's switch belongs to,
* including this switch itself, of an event. Allows the other switches to
* reconfigure themselves for cross-chip operations. Can also be used to
* reconfigure ports without net_devices (CPU ports, DSA links) whenever
* a user port's state changes.
*/
static int dsa_port_notify(const struct dsa_port *dp, unsigned long e, void *v)
{
return dsa_tree_notify(dp->ds->dst, e, v);
}
static void dsa_port_notify_bridge_fdb_flush(const struct dsa_port *dp, u16 vid)
2021-08-08 14:35:25 +00:00
{
struct net_device *brport_dev = dsa_port_to_bridge_port(dp);
struct switchdev_notifier_fdb_info info = {
.vid = vid,
2021-08-08 14:35:25 +00:00
};
/* When the port becomes standalone it has already left the bridge.
* Don't notify the bridge in that case.
*/
if (!brport_dev)
return;
call_switchdev_notifiers(SWITCHDEV_FDB_FLUSH_TO_BRIDGE,
brport_dev, &info.info, NULL);
}
net: dsa: centralize fast ageing when address learning is turned off Currently DSA leaves it down to device drivers to fast age the FDB on a port when address learning is disabled on it. There are 2 reasons for doing that in the first place: - when address learning is disabled by user space, through IFLA_BRPORT_LEARNING or the brport_attr_learning sysfs, what user space typically wants to achieve is to operate in a mode with no dynamic FDB entry on that port. But if the port is already up, some addresses might have been already learned on it, and it seems silly to wait for 5 minutes for them to expire until something useful can be done. - when a port leaves a bridge and becomes standalone, DSA turns off address learning on it. This also has the nice side effect of flushing the dynamically learned bridge FDB entries on it, which is a good idea because standalone ports should not have bridge FDB entries on them. We let drivers manage fast ageing under this condition because if DSA were to do it, it would need to track each port's learning state, and act upon the transition, which it currently doesn't. But there are 2 reasons why doing it is better after all: - drivers might get it wrong and not do it (see b53_port_set_learning) - we would like to flush the dynamic entries from the software bridge too, and letting drivers do that would be another pain point So track the port learning state and trigger a fast age process automatically within DSA. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-08 14:35:23 +00:00
static void dsa_port_fast_age(const struct dsa_port *dp)
{
struct dsa_switch *ds = dp->ds;
if (!ds->ops->port_fast_age)
return;
ds->ops->port_fast_age(ds, dp->index);
2021-08-08 14:35:25 +00:00
/* flush all VLANs */
dsa_port_notify_bridge_fdb_flush(dp, 0);
}
static int dsa_port_vlan_fast_age(const struct dsa_port *dp, u16 vid)
{
struct dsa_switch *ds = dp->ds;
int err;
if (!ds->ops->port_vlan_fast_age)
return -EOPNOTSUPP;
err = ds->ops->port_vlan_fast_age(ds, dp->index, vid);
if (!err)
dsa_port_notify_bridge_fdb_flush(dp, vid);
return err;
}
static int dsa_port_msti_fast_age(const struct dsa_port *dp, u16 msti)
{
DECLARE_BITMAP(vids, VLAN_N_VID) = { 0 };
int err, vid;
err = br_mst_get_info(dsa_port_bridge_dev_get(dp), msti, vids);
if (err)
return err;
for_each_set_bit(vid, vids, VLAN_N_VID) {
err = dsa_port_vlan_fast_age(dp, vid);
if (err)
return err;
}
return 0;
net: dsa: centralize fast ageing when address learning is turned off Currently DSA leaves it down to device drivers to fast age the FDB on a port when address learning is disabled on it. There are 2 reasons for doing that in the first place: - when address learning is disabled by user space, through IFLA_BRPORT_LEARNING or the brport_attr_learning sysfs, what user space typically wants to achieve is to operate in a mode with no dynamic FDB entry on that port. But if the port is already up, some addresses might have been already learned on it, and it seems silly to wait for 5 minutes for them to expire until something useful can be done. - when a port leaves a bridge and becomes standalone, DSA turns off address learning on it. This also has the nice side effect of flushing the dynamically learned bridge FDB entries on it, which is a good idea because standalone ports should not have bridge FDB entries on them. We let drivers manage fast ageing under this condition because if DSA were to do it, it would need to track each port's learning state, and act upon the transition, which it currently doesn't. But there are 2 reasons why doing it is better after all: - drivers might get it wrong and not do it (see b53_port_set_learning) - we would like to flush the dynamic entries from the software bridge too, and letting drivers do that would be another pain point So track the port learning state and trigger a fast age process automatically within DSA. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-08 14:35:23 +00:00
}
static bool dsa_port_can_configure_learning(struct dsa_port *dp)
{
struct switchdev_brport_flags flags = {
.mask = BR_LEARNING,
};
struct dsa_switch *ds = dp->ds;
int err;
if (!ds->ops->port_bridge_flags || !ds->ops->port_pre_bridge_flags)
return false;
err = ds->ops->port_pre_bridge_flags(ds, dp->index, flags, NULL);
return !err;
}
bool dsa_port_supports_hwtstamp(struct dsa_port *dp)
{
struct dsa_switch *ds = dp->ds;
struct ifreq ifr = {};
int err;
if (!ds->ops->port_hwtstamp_get || !ds->ops->port_hwtstamp_set)
return false;
/* "See through" shim implementations of the "get" method.
* Since we can't cook up a complete ioctl request structure, this will
* fail in copy_to_user() with -EFAULT, which hopefully is enough to
* detect a valid implementation.
*/
err = ds->ops->port_hwtstamp_get(ds, dp->index, &ifr);
return err != -EOPNOTSUPP;
}
int dsa_port_set_state(struct dsa_port *dp, u8 state, bool do_fast_age)
{
struct dsa_switch *ds = dp->ds;
int port = dp->index;
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
if (!ds->ops->port_stp_state_set)
return -EOPNOTSUPP;
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
ds->ops->port_stp_state_set(ds, port, state);
if (!dsa_port_can_configure_learning(dp) ||
(do_fast_age && dp->learning)) {
/* Fast age FDB entries or flush appropriate forwarding database
* for the given port, if we are moving it from Learning or
* Forwarding state, to Disabled or Blocking or Listening state.
* Ports that were standalone before the STP state change don't
* need to fast age the FDB, since address learning is off in
* standalone mode.
*/
if ((dp->stp_state == BR_STATE_LEARNING ||
dp->stp_state == BR_STATE_FORWARDING) &&
(state == BR_STATE_DISABLED ||
state == BR_STATE_BLOCKING ||
state == BR_STATE_LISTENING))
net: dsa: centralize fast ageing when address learning is turned off Currently DSA leaves it down to device drivers to fast age the FDB on a port when address learning is disabled on it. There are 2 reasons for doing that in the first place: - when address learning is disabled by user space, through IFLA_BRPORT_LEARNING or the brport_attr_learning sysfs, what user space typically wants to achieve is to operate in a mode with no dynamic FDB entry on that port. But if the port is already up, some addresses might have been already learned on it, and it seems silly to wait for 5 minutes for them to expire until something useful can be done. - when a port leaves a bridge and becomes standalone, DSA turns off address learning on it. This also has the nice side effect of flushing the dynamically learned bridge FDB entries on it, which is a good idea because standalone ports should not have bridge FDB entries on them. We let drivers manage fast ageing under this condition because if DSA were to do it, it would need to track each port's learning state, and act upon the transition, which it currently doesn't. But there are 2 reasons why doing it is better after all: - drivers might get it wrong and not do it (see b53_port_set_learning) - we would like to flush the dynamic entries from the software bridge too, and letting drivers do that would be another pain point So track the port learning state and trigger a fast age process automatically within DSA. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-08 14:35:23 +00:00
dsa_port_fast_age(dp);
}
dp->stp_state = state;
return 0;
}
static void dsa_port_set_state_now(struct dsa_port *dp, u8 state,
bool do_fast_age)
{
struct dsa_switch *ds = dp->ds;
int err;
err = dsa_port_set_state(dp, state, do_fast_age);
if (err && err != -EOPNOTSUPP) {
dev_err(ds->dev, "port %d failed to set STP state %u: %pe\n",
dp->index, state, ERR_PTR(err));
}
}
int dsa_port_set_mst_state(struct dsa_port *dp,
const struct switchdev_mst_state *state,
struct netlink_ext_ack *extack)
{
struct dsa_switch *ds = dp->ds;
u8 prev_state;
int err;
if (!ds->ops->port_mst_state_set)
return -EOPNOTSUPP;
err = br_mst_get_state(dsa_port_to_bridge_port(dp), state->msti,
&prev_state);
if (err)
return err;
err = ds->ops->port_mst_state_set(ds, dp->index, state);
if (err)
return err;
if (!(dp->learning &&
(prev_state == BR_STATE_LEARNING ||
prev_state == BR_STATE_FORWARDING) &&
(state->state == BR_STATE_DISABLED ||
state->state == BR_STATE_BLOCKING ||
state->state == BR_STATE_LISTENING)))
return 0;
err = dsa_port_msti_fast_age(dp, state->msti);
if (err)
NL_SET_ERR_MSG_MOD(extack,
"Unable to flush associated VLANs");
return 0;
}
int dsa_port_enable_rt(struct dsa_port *dp, struct phy_device *phy)
{
struct dsa_switch *ds = dp->ds;
int port = dp->index;
int err;
if (ds->ops->port_enable) {
err = ds->ops->port_enable(ds, port, phy);
if (err)
return err;
}
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
if (!dp->bridge)
dsa_port_set_state_now(dp, BR_STATE_FORWARDING, false);
if (dp->pl)
phylink_start(dp->pl);
return 0;
}
int dsa_port_enable(struct dsa_port *dp, struct phy_device *phy)
{
int err;
rtnl_lock();
err = dsa_port_enable_rt(dp, phy);
rtnl_unlock();
return err;
}
void dsa_port_disable_rt(struct dsa_port *dp)
{
struct dsa_switch *ds = dp->ds;
int port = dp->index;
if (dp->pl)
phylink_stop(dp->pl);
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
if (!dp->bridge)
dsa_port_set_state_now(dp, BR_STATE_DISABLED, false);
if (ds->ops->port_disable)
ds->ops->port_disable(ds, port);
}
void dsa_port_disable(struct dsa_port *dp)
{
rtnl_lock();
dsa_port_disable_rt(dp);
rtnl_unlock();
}
static void dsa_port_reset_vlan_filtering(struct dsa_port *dp,
struct dsa_bridge bridge)
{
struct netlink_ext_ack extack = {0};
bool change_vlan_filtering = false;
struct dsa_switch *ds = dp->ds;
struct dsa_port *other_dp;
bool vlan_filtering;
int err;
if (ds->needs_standalone_vlan_filtering &&
!br_vlan_enabled(bridge.dev)) {
change_vlan_filtering = true;
vlan_filtering = true;
} else if (!ds->needs_standalone_vlan_filtering &&
br_vlan_enabled(bridge.dev)) {
change_vlan_filtering = true;
vlan_filtering = false;
}
/* If the bridge was vlan_filtering, the bridge core doesn't trigger an
* event for changing vlan_filtering setting upon user ports leaving
* it. That is a good thing, because that lets us handle it and also
* handle the case where the switch's vlan_filtering setting is global
* (not per port). When that happens, the correct moment to trigger the
* vlan_filtering callback is only when the last port leaves the last
* VLAN-aware bridge.
*/
if (change_vlan_filtering && ds->vlan_filtering_is_global) {
dsa_switch_for_each_port(other_dp, ds) {
struct net_device *br = dsa_port_bridge_dev_get(other_dp);
if (br && br_vlan_enabled(br)) {
change_vlan_filtering = false;
break;
}
}
}
if (!change_vlan_filtering)
return;
err = dsa_port_vlan_filtering(dp, vlan_filtering, &extack);
if (extack._msg) {
dev_err(ds->dev, "port %d: %s\n", dp->index,
extack._msg);
}
if (err && err != -EOPNOTSUPP) {
dev_err(ds->dev,
"port %d failed to reset VLAN filtering to %d: %pe\n",
dp->index, vlan_filtering, ERR_PTR(err));
}
}
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
static int dsa_port_inherit_brport_flags(struct dsa_port *dp,
struct netlink_ext_ack *extack)
net: dsa: configure better brport flags when ports leave the bridge For a DSA switch port operating in standalone mode, address learning doesn't make much sense since that is a bridge function. In fact, address learning even breaks setups such as this one: +---------------------------------------------+ | | | +-------------------+ | | | br0 | send receive | | +--------+-+--------+ +--------+ +--------+ | | | | | | | | | | | | | swp0 | | swp1 | | swp2 | | swp3 | | | | | | | | | | | | +-+--------+-+--------+-+--------+-+--------+-+ | ^ | ^ | | | | | +-----------+ | | | +--------------------------------+ because if the switch has a single FDB (can offload a single bridge) then source address learning on swp3 can "steal" the source MAC address of swp2 from br0's FDB, because learning frames coming from swp2 will be done twice: first on the swp1 ingress port, second on the swp3 ingress port. So the hardware FDB will become out of sync with the software bridge, and when swp2 tries to send one more packet towards swp1, the ASIC will attempt to short-circuit the forwarding path and send it directly to swp3 (since that's the last port it learned that address on), which it obviously can't, because swp3 operates in standalone mode. So DSA drivers operating in standalone mode should still configure a list of bridge port flags even when they are standalone. Currently DSA attempts to call dsa_port_bridge_flags with 0, which disables egress flooding of unknown unicast and multicast, something which doesn't make much sense. For the switches that implement .port_egress_floods - b53 and mv88e6xxx, it probably doesn't matter too much either, since they can possibly inject traffic from the CPU into a standalone port, regardless of MAC DA, even if egress flooding is turned off for that port, but certainly not all DSA switches can do that - sja1105, for example, can't. So it makes sense to use a better common default there, such as "flood everything". It should also be noted that what DSA calls "dsa_port_bridge_flags()" is a degenerate name for just calling .port_egress_floods(), since nothing else is implemented - not learning, in particular. But disabling address learning, something that this driver is also coding up for, will be supported by individual drivers once .port_egress_floods is replaced with a more generic .port_bridge_flags. Previous attempts to code up this logic have been in the common bridge layer, but as pointed out by Ido Schimmel, there are corner cases that are missed when doing that: https://patchwork.kernel.org/project/netdevbpf/patch/20210209151936.97382-5-olteanv@gmail.com/ So, at least for now, let's leave DSA in charge of setting port flags before and after the bridge join and leave. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:54 +00:00
{
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
const unsigned long mask = BR_LEARNING | BR_FLOOD | BR_MCAST_FLOOD |
BR_BCAST_FLOOD | BR_PORT_LOCKED;
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
struct net_device *brport_dev = dsa_port_to_bridge_port(dp);
int flag, err;
net: dsa: configure better brport flags when ports leave the bridge For a DSA switch port operating in standalone mode, address learning doesn't make much sense since that is a bridge function. In fact, address learning even breaks setups such as this one: +---------------------------------------------+ | | | +-------------------+ | | | br0 | send receive | | +--------+-+--------+ +--------+ +--------+ | | | | | | | | | | | | | swp0 | | swp1 | | swp2 | | swp3 | | | | | | | | | | | | +-+--------+-+--------+-+--------+-+--------+-+ | ^ | ^ | | | | | +-----------+ | | | +--------------------------------+ because if the switch has a single FDB (can offload a single bridge) then source address learning on swp3 can "steal" the source MAC address of swp2 from br0's FDB, because learning frames coming from swp2 will be done twice: first on the swp1 ingress port, second on the swp3 ingress port. So the hardware FDB will become out of sync with the software bridge, and when swp2 tries to send one more packet towards swp1, the ASIC will attempt to short-circuit the forwarding path and send it directly to swp3 (since that's the last port it learned that address on), which it obviously can't, because swp3 operates in standalone mode. So DSA drivers operating in standalone mode should still configure a list of bridge port flags even when they are standalone. Currently DSA attempts to call dsa_port_bridge_flags with 0, which disables egress flooding of unknown unicast and multicast, something which doesn't make much sense. For the switches that implement .port_egress_floods - b53 and mv88e6xxx, it probably doesn't matter too much either, since they can possibly inject traffic from the CPU into a standalone port, regardless of MAC DA, even if egress flooding is turned off for that port, but certainly not all DSA switches can do that - sja1105, for example, can't. So it makes sense to use a better common default there, such as "flood everything". It should also be noted that what DSA calls "dsa_port_bridge_flags()" is a degenerate name for just calling .port_egress_floods(), since nothing else is implemented - not learning, in particular. But disabling address learning, something that this driver is also coding up for, will be supported by individual drivers once .port_egress_floods is replaced with a more generic .port_bridge_flags. Previous attempts to code up this logic have been in the common bridge layer, but as pointed out by Ido Schimmel, there are corner cases that are missed when doing that: https://patchwork.kernel.org/project/netdevbpf/patch/20210209151936.97382-5-olteanv@gmail.com/ So, at least for now, let's leave DSA in charge of setting port flags before and after the bridge join and leave. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:54 +00:00
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
for_each_set_bit(flag, &mask, 32) {
struct switchdev_brport_flags flags = {0};
net: dsa: configure better brport flags when ports leave the bridge For a DSA switch port operating in standalone mode, address learning doesn't make much sense since that is a bridge function. In fact, address learning even breaks setups such as this one: +---------------------------------------------+ | | | +-------------------+ | | | br0 | send receive | | +--------+-+--------+ +--------+ +--------+ | | | | | | | | | | | | | swp0 | | swp1 | | swp2 | | swp3 | | | | | | | | | | | | +-+--------+-+--------+-+--------+-+--------+-+ | ^ | ^ | | | | | +-----------+ | | | +--------------------------------+ because if the switch has a single FDB (can offload a single bridge) then source address learning on swp3 can "steal" the source MAC address of swp2 from br0's FDB, because learning frames coming from swp2 will be done twice: first on the swp1 ingress port, second on the swp3 ingress port. So the hardware FDB will become out of sync with the software bridge, and when swp2 tries to send one more packet towards swp1, the ASIC will attempt to short-circuit the forwarding path and send it directly to swp3 (since that's the last port it learned that address on), which it obviously can't, because swp3 operates in standalone mode. So DSA drivers operating in standalone mode should still configure a list of bridge port flags even when they are standalone. Currently DSA attempts to call dsa_port_bridge_flags with 0, which disables egress flooding of unknown unicast and multicast, something which doesn't make much sense. For the switches that implement .port_egress_floods - b53 and mv88e6xxx, it probably doesn't matter too much either, since they can possibly inject traffic from the CPU into a standalone port, regardless of MAC DA, even if egress flooding is turned off for that port, but certainly not all DSA switches can do that - sja1105, for example, can't. So it makes sense to use a better common default there, such as "flood everything". It should also be noted that what DSA calls "dsa_port_bridge_flags()" is a degenerate name for just calling .port_egress_floods(), since nothing else is implemented - not learning, in particular. But disabling address learning, something that this driver is also coding up for, will be supported by individual drivers once .port_egress_floods is replaced with a more generic .port_bridge_flags. Previous attempts to code up this logic have been in the common bridge layer, but as pointed out by Ido Schimmel, there are corner cases that are missed when doing that: https://patchwork.kernel.org/project/netdevbpf/patch/20210209151936.97382-5-olteanv@gmail.com/ So, at least for now, let's leave DSA in charge of setting port flags before and after the bridge join and leave. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:54 +00:00
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
flags.mask = BIT(flag);
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
if (br_port_flag_is_set(brport_dev, BIT(flag)))
flags.val = BIT(flag);
net: dsa: configure better brport flags when ports leave the bridge For a DSA switch port operating in standalone mode, address learning doesn't make much sense since that is a bridge function. In fact, address learning even breaks setups such as this one: +---------------------------------------------+ | | | +-------------------+ | | | br0 | send receive | | +--------+-+--------+ +--------+ +--------+ | | | | | | | | | | | | | swp0 | | swp1 | | swp2 | | swp3 | | | | | | | | | | | | +-+--------+-+--------+-+--------+-+--------+-+ | ^ | ^ | | | | | +-----------+ | | | +--------------------------------+ because if the switch has a single FDB (can offload a single bridge) then source address learning on swp3 can "steal" the source MAC address of swp2 from br0's FDB, because learning frames coming from swp2 will be done twice: first on the swp1 ingress port, second on the swp3 ingress port. So the hardware FDB will become out of sync with the software bridge, and when swp2 tries to send one more packet towards swp1, the ASIC will attempt to short-circuit the forwarding path and send it directly to swp3 (since that's the last port it learned that address on), which it obviously can't, because swp3 operates in standalone mode. So DSA drivers operating in standalone mode should still configure a list of bridge port flags even when they are standalone. Currently DSA attempts to call dsa_port_bridge_flags with 0, which disables egress flooding of unknown unicast and multicast, something which doesn't make much sense. For the switches that implement .port_egress_floods - b53 and mv88e6xxx, it probably doesn't matter too much either, since they can possibly inject traffic from the CPU into a standalone port, regardless of MAC DA, even if egress flooding is turned off for that port, but certainly not all DSA switches can do that - sja1105, for example, can't. So it makes sense to use a better common default there, such as "flood everything". It should also be noted that what DSA calls "dsa_port_bridge_flags()" is a degenerate name for just calling .port_egress_floods(), since nothing else is implemented - not learning, in particular. But disabling address learning, something that this driver is also coding up for, will be supported by individual drivers once .port_egress_floods is replaced with a more generic .port_bridge_flags. Previous attempts to code up this logic have been in the common bridge layer, but as pointed out by Ido Schimmel, there are corner cases that are missed when doing that: https://patchwork.kernel.org/project/netdevbpf/patch/20210209151936.97382-5-olteanv@gmail.com/ So, at least for now, let's leave DSA in charge of setting port flags before and after the bridge join and leave. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:54 +00:00
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
err = dsa_port_bridge_flags(dp, flags, extack);
if (err && err != -EOPNOTSUPP)
return err;
}
return 0;
}
static void dsa_port_clear_brport_flags(struct dsa_port *dp)
{
const unsigned long val = BR_FLOOD | BR_MCAST_FLOOD | BR_BCAST_FLOOD;
const unsigned long mask = BR_LEARNING | BR_FLOOD | BR_MCAST_FLOOD |
BR_BCAST_FLOOD | BR_PORT_LOCKED;
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
int flag, err;
for_each_set_bit(flag, &mask, 32) {
struct switchdev_brport_flags flags = {0};
flags.mask = BIT(flag);
flags.val = val & BIT(flag);
err = dsa_port_bridge_flags(dp, flags, NULL);
if (err && err != -EOPNOTSUPP)
dev_err(dp->ds->dev,
"failed to clear bridge port flag %lu: %pe\n",
flags.val, ERR_PTR(err));
net: dsa: configure better brport flags when ports leave the bridge For a DSA switch port operating in standalone mode, address learning doesn't make much sense since that is a bridge function. In fact, address learning even breaks setups such as this one: +---------------------------------------------+ | | | +-------------------+ | | | br0 | send receive | | +--------+-+--------+ +--------+ +--------+ | | | | | | | | | | | | | swp0 | | swp1 | | swp2 | | swp3 | | | | | | | | | | | | +-+--------+-+--------+-+--------+-+--------+-+ | ^ | ^ | | | | | +-----------+ | | | +--------------------------------+ because if the switch has a single FDB (can offload a single bridge) then source address learning on swp3 can "steal" the source MAC address of swp2 from br0's FDB, because learning frames coming from swp2 will be done twice: first on the swp1 ingress port, second on the swp3 ingress port. So the hardware FDB will become out of sync with the software bridge, and when swp2 tries to send one more packet towards swp1, the ASIC will attempt to short-circuit the forwarding path and send it directly to swp3 (since that's the last port it learned that address on), which it obviously can't, because swp3 operates in standalone mode. So DSA drivers operating in standalone mode should still configure a list of bridge port flags even when they are standalone. Currently DSA attempts to call dsa_port_bridge_flags with 0, which disables egress flooding of unknown unicast and multicast, something which doesn't make much sense. For the switches that implement .port_egress_floods - b53 and mv88e6xxx, it probably doesn't matter too much either, since they can possibly inject traffic from the CPU into a standalone port, regardless of MAC DA, even if egress flooding is turned off for that port, but certainly not all DSA switches can do that - sja1105, for example, can't. So it makes sense to use a better common default there, such as "flood everything". It should also be noted that what DSA calls "dsa_port_bridge_flags()" is a degenerate name for just calling .port_egress_floods(), since nothing else is implemented - not learning, in particular. But disabling address learning, something that this driver is also coding up for, will be supported by individual drivers once .port_egress_floods is replaced with a more generic .port_bridge_flags. Previous attempts to code up this logic have been in the common bridge layer, but as pointed out by Ido Schimmel, there are corner cases that are missed when doing that: https://patchwork.kernel.org/project/netdevbpf/patch/20210209151936.97382-5-olteanv@gmail.com/ So, at least for now, let's leave DSA in charge of setting port flags before and after the bridge join and leave. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:54 +00:00
}
}
net: bridge: move the switchdev object replay helpers to "push" mode Starting with commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"), DSA has introduced some bridge helpers that replay switchdev events (FDB/MDB/VLAN additions and deletions) that can be lost by the switchdev drivers in a variety of circumstances: - an IP multicast group was host-joined on the bridge itself before any switchdev port joined the bridge, leading to the host MDB entries missing in the hardware database. - during the bridge creation process, the MAC address of the bridge was added to the FDB as an entry pointing towards the bridge device itself, but with no switchdev ports being part of the bridge yet, this local FDB entry would remain unknown to the switchdev hardware database. - a VLAN/FDB/MDB was added to a bridge port that is a LAG interface, before any switchdev port joined that LAG, leading to the hardware database missing those entries. - a switchdev port left a LAG that is a bridge port, while the LAG remained part of the bridge, and all FDB/MDB/VLAN entries remained installed in the hardware database of the switchdev port. Also, since commit 0d2cfbd41c4a ("net: bridge: ignore switchdev events for LAG ports which didn't request replay"), DSA introduced a method, based on a const void *ctx, to ensure that two switchdev ports under the same LAG that is a bridge port do not see the same MDB/VLAN entry being replayed twice by the bridge, once for every bridge port that joins the LAG. With so many ordering corner cases being possible, it seems unreasonable to expect a switchdev driver writer to get it right from the first try. Therefore, now that DSA has experimented with the bridge replay helpers for a little bit, we can move the code to the bridge driver where it is more readily available to all switchdev drivers. To convert the switchdev object replay helpers from "pull mode" (where the driver asks for them) to a "push mode" (where the bridge offers them automatically), the biggest problem is that the bridge needs to be aware when a switchdev port joins and leaves, even when the switchdev is only indirectly a bridge port (for example when the bridge port is a LAG upper of the switchdev). Luckily, we already have a hook for that, in the form of the newly introduced switchdev_bridge_port_offload() and switchdev_bridge_port_unoffload() calls. These offer a natural place for hooking the object addition and deletion replays. Extend the above 2 functions with: - pointers to the switchdev atomic notifier (for FDB replays) and the blocking notifier (for MDB and VLAN replays). - the "const void *ctx" argument required for drivers to be able to disambiguate between which port is targeted, when multiple ports are lowers of the same LAG that is a bridge port. Most of the drivers pass NULL to this argument, except the ones that support LAG offload and have the proper context check already in place in the switchdev blocking notifier handler. Also unexport the replay helpers, since nobody except the bridge calls them directly now. Note that: (a) we abuse the terminology slightly, because FDB entries are not "switchdev objects", but we count them as objects nonetheless. With no direct way to prove it, I think they are not modeled as switchdev objects because those can only be installed by the bridge to the hardware (as opposed to FDB entries which can be propagated in the other direction too). This is merely an abuse of terms, FDB entries are replayed too, despite not being objects. (b) the bridge does not attempt to sync port attributes to newly joined ports, just the countable stuff (the objects). The reason for this is simple: no universal and symmetric way to sync and unsync them is known. For example, VLAN filtering: what to do on unsync, disable or leave it enabled? Similarly, STP state, ageing timer, etc etc. What a switchdev port does when it becomes standalone again is not really up to the bridge's competence, and the driver should deal with it. On the other hand, replaying deletions of switchdev objects can be seen a matter of cleanup and therefore be treated by the bridge, hence this patch. We make the replay helpers opt-in for drivers, because they might not bring immediate benefits for them: - nbp_vlan_init() is called _after_ netdev_master_upper_dev_link(), so br_vlan_replay() should not do anything for the new drivers on which we call it. The existing drivers where there was even a slight possibility for there to exist a VLAN on a bridge port before they join it are already guarded against this: mlxsw and prestera deny joining LAG interfaces that are members of a bridge. - br_fdb_replay() should now notify of local FDB entries, but I patched all drivers except DSA to ignore these new entries in commit 2c4eca3ef716 ("net: bridge: switchdev: include local flag in FDB notifications"). Driver authors can lift this restriction as they wish, and when they do, they can also opt into the FDB replay functionality. - br_mdb_replay() should fix a real issue which is described in commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"). However most drivers do not offload the SWITCHDEV_OBJ_ID_HOST_MDB to see this issue: only cpsw and am65_cpsw offload this switchdev object, and I don't completely understand the way in which they offload this switchdev object anyway. So I'll leave it up to these drivers' respective maintainers to opt into br_mdb_replay(). So most of the drivers pass NULL notifier blocks for the replay helpers, except: - dpaa2-switch which was already acked/regression-tested with the helpers enabled (and there isn't much of a downside in having them) - ocelot which already had replay logic in "pull" mode - DSA which already had replay logic in "pull" mode An important observation is that the drivers which don't currently request bridge event replays don't even have the switchdev_bridge_port_{offload,unoffload} calls placed in proper places right now. This was done to avoid unnecessary rework for drivers which might never even add support for this. For driver writers who wish to add replay support, this can be used as a tentative placement guide: https://patchwork.kernel.org/project/netdevbpf/patch/20210720134655.892334-11-vladimir.oltean@nxp.com/ Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:03 +00:00
static int dsa_port_switchdev_sync_attrs(struct dsa_port *dp,
struct netlink_ext_ack *extack)
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
{
struct net_device *brport_dev = dsa_port_to_bridge_port(dp);
struct net_device *br = dsa_port_bridge_dev_get(dp);
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
int err;
err = dsa_port_inherit_brport_flags(dp, extack);
if (err)
return err;
err = dsa_port_set_state(dp, br_port_get_stp_state(brport_dev), false);
if (err && err != -EOPNOTSUPP)
return err;
err = dsa_port_vlan_filtering(dp, br_vlan_enabled(br), extack);
if (err && err != -EOPNOTSUPP)
return err;
err = dsa_port_ageing_time(dp, br_get_ageing_time(br));
if (err && err != -EOPNOTSUPP)
return err;
return 0;
}
static void dsa_port_switchdev_unsync_attrs(struct dsa_port *dp,
struct dsa_bridge bridge)
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
{
/* Configure the port for standalone mode (no address learning,
* flood everything).
* The bridge only emits SWITCHDEV_ATTR_ID_PORT_BRIDGE_FLAGS events
* when the user requests it through netlink or sysfs, but not
* automatically at port join or leave, so we need to handle resetting
* the brport flags ourselves. But we even prefer it that way, because
* otherwise, some setups might never get the notification they need,
* for example, when a port leaves a LAG that offloads the bridge,
* it becomes standalone, but as far as the bridge is concerned, no
* port ever left.
*/
dsa_port_clear_brport_flags(dp);
/* Port left the bridge, put in BR_STATE_DISABLED by the bridge layer,
* so allow it to be in BR_STATE_FORWARDING to be kept functional
*/
dsa_port_set_state_now(dp, BR_STATE_FORWARDING, true);
dsa_port_reset_vlan_filtering(dp, bridge);
/* Ageing time may be global to the switch chip, so don't change it
* here because we have no good reason (or value) to change it to.
*/
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
}
static int dsa_port_bridge_create(struct dsa_port *dp,
struct net_device *br,
struct netlink_ext_ack *extack)
{
struct dsa_switch *ds = dp->ds;
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
struct dsa_bridge *bridge;
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
bridge = dsa_tree_bridge_find(ds->dst, br);
if (bridge) {
refcount_inc(&bridge->refcount);
dp->bridge = bridge;
return 0;
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
}
bridge = kzalloc(sizeof(*bridge), GFP_KERNEL);
if (!bridge)
return -ENOMEM;
refcount_set(&bridge->refcount, 1);
bridge->dev = br;
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
bridge->num = dsa_bridge_num_get(br, ds->max_num_bridges);
if (ds->max_num_bridges && !bridge->num) {
NL_SET_ERR_MSG_MOD(extack,
"Range of offloadable bridges exceeded");
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
kfree(bridge);
return -EOPNOTSUPP;
net: dsa: add support for bridge TX forwarding offload For a DSA switch, to offload the forwarding process of a bridge device means to send the packets coming from the software bridge as data plane packets. This is contrary to everything that DSA has done so far, because the current taggers only know to send control packets (ones that target a specific destination port), whereas data plane packets are supposed to be forwarded according to the FDB lookup, much like packets ingressing on any regular ingress port. If the FDB lookup process returns multiple destination ports (flooding, multicast), then replication is also handled by the switch hardware - the bridge only sends a single packet and avoids the skb_clone(). DSA keeps for each bridge port a zero-based index (the number of the bridge). Multiple ports performing TX forwarding offload to the same bridge have the same dp->bridge_num value, and ports not offloading the TX data plane of a bridge have dp->bridge_num = -1. The tagger can check if the packet that is being transmitted on has skb->offload_fwd_mark = true or not. If it does, it can be sure that the packet belongs to the data plane of a bridge, further information about which can be obtained based on dp->bridge_dev and dp->bridge_num. It can then compose a DSA tag for injecting a data plane packet into that bridge number. For the switch driver side, we offer two new dsa_switch_ops methods, called .port_bridge_fwd_offload_{add,del}, which are modeled after .port_bridge_{join,leave}. These methods are provided in case the driver needs to configure the hardware to treat packets coming from that bridge software interface as data plane packets. The switchdev <-> bridge interaction happens during the netdev_master_upper_dev_link() call, so to switch drivers, the effect is that the .port_bridge_fwd_offload_add() method is called immediately after .port_bridge_join(). If the bridge number exceeds the number of bridges for which the switch driver can offload the TX data plane (and this includes the case where the driver can offload none), DSA falls back to simply returning tx_fwd_offload = false in the switchdev_bridge_port_offload() call. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-22 15:55:40 +00:00
}
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
dp->bridge = bridge;
return 0;
}
static void dsa_port_bridge_destroy(struct dsa_port *dp,
const struct net_device *br)
{
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
struct dsa_bridge *bridge = dp->bridge;
dp->bridge = NULL;
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
if (!refcount_dec_and_test(&bridge->refcount))
return;
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
if (bridge->num)
dsa_bridge_num_put(br, bridge->num);
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
kfree(bridge);
net: dsa: add support for bridge TX forwarding offload For a DSA switch, to offload the forwarding process of a bridge device means to send the packets coming from the software bridge as data plane packets. This is contrary to everything that DSA has done so far, because the current taggers only know to send control packets (ones that target a specific destination port), whereas data plane packets are supposed to be forwarded according to the FDB lookup, much like packets ingressing on any regular ingress port. If the FDB lookup process returns multiple destination ports (flooding, multicast), then replication is also handled by the switch hardware - the bridge only sends a single packet and avoids the skb_clone(). DSA keeps for each bridge port a zero-based index (the number of the bridge). Multiple ports performing TX forwarding offload to the same bridge have the same dp->bridge_num value, and ports not offloading the TX data plane of a bridge have dp->bridge_num = -1. The tagger can check if the packet that is being transmitted on has skb->offload_fwd_mark = true or not. If it does, it can be sure that the packet belongs to the data plane of a bridge, further information about which can be obtained based on dp->bridge_dev and dp->bridge_num. It can then compose a DSA tag for injecting a data plane packet into that bridge number. For the switch driver side, we offer two new dsa_switch_ops methods, called .port_bridge_fwd_offload_{add,del}, which are modeled after .port_bridge_{join,leave}. These methods are provided in case the driver needs to configure the hardware to treat packets coming from that bridge software interface as data plane packets. The switchdev <-> bridge interaction happens during the netdev_master_upper_dev_link() call, so to switch drivers, the effect is that the .port_bridge_fwd_offload_add() method is called immediately after .port_bridge_join(). If the bridge number exceeds the number of bridges for which the switch driver can offload the TX data plane (and this includes the case where the driver can offload none), DSA falls back to simply returning tx_fwd_offload = false in the switchdev_bridge_port_offload() call. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-22 15:55:40 +00:00
}
static bool dsa_port_supports_mst(struct dsa_port *dp)
{
struct dsa_switch *ds = dp->ds;
return ds->ops->vlan_msti_set &&
ds->ops->port_mst_state_set &&
ds->ops->port_vlan_fast_age &&
dsa_port_can_configure_learning(dp);
}
int dsa_port_bridge_join(struct dsa_port *dp, struct net_device *br,
struct netlink_ext_ack *extack)
{
struct dsa_notifier_bridge_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.extack = extack,
};
struct net_device *dev = dp->user;
net: bridge: switchdev: let drivers inform which bridge ports are offloaded On reception of an skb, the bridge checks if it was marked as 'already forwarded in hardware' (checks if skb->offload_fwd_mark == 1), and if it is, it assigns the source hardware domain of that skb based on the hardware domain of the ingress port. Then during forwarding, it enforces that the egress port must have a different hardware domain than the ingress one (this is done in nbp_switchdev_allowed_egress). Non-switchdev drivers don't report any physical switch id (neither through devlink nor .ndo_get_port_parent_id), therefore the bridge assigns them a hardware domain of 0, and packets coming from them will always have skb->offload_fwd_mark = 0. So there aren't any restrictions. Problems appear due to the fact that DSA would like to perform software fallback for bonding and team interfaces that the physical switch cannot offload. +-- br0 ---+ / / | \ / / | \ / | | bond0 / | | / \ swp0 swp1 swp2 swp3 swp4 There, it is desirable that the presence of swp3 and swp4 under a non-offloaded LAG does not preclude us from doing hardware bridging beteen swp0, swp1 and swp2. The bandwidth of the CPU is often times high enough that software bridging between {swp0,swp1,swp2} and bond0 is not impractical. But this creates an impossible paradox given the current way in which port hardware domains are assigned. When the driver receives a packet from swp0 (say, due to flooding), it must set skb->offload_fwd_mark to something. - If we set it to 0, then the bridge will forward it towards swp1, swp2 and bond0. But the switch has already forwarded it towards swp1 and swp2 (not to bond0, remember, that isn't offloaded, so as far as the switch is concerned, ports swp3 and swp4 are not looking up the FDB, and the entire bond0 is a destination that is strictly behind the CPU). But we don't want duplicated traffic towards swp1 and swp2, so it's not ok to set skb->offload_fwd_mark = 0. - If we set it to 1, then the bridge will not forward the skb towards the ports with the same switchdev mark, i.e. not to swp1, swp2 and bond0. Towards swp1 and swp2 that's ok, but towards bond0? It should have forwarded the skb there. So the real issue is that bond0 will be assigned the same hardware domain as {swp0,swp1,swp2}, because the function that assigns hardware domains to bridge ports, nbp_switchdev_add(), recurses through bond0's lower interfaces until it finds something that implements devlink (calls dev_get_port_parent_id with bool recurse = true). This is a problem because the fact that bond0 can be offloaded by swp3 and swp4 in our example is merely an assumption. A solution is to give the bridge explicit hints as to what hardware domain it should use for each port. Currently, the bridging offload is very 'silent': a driver registers a netdevice notifier, which is put on the netns's notifier chain, and which sniffs around for NETDEV_CHANGEUPPER events where the upper is a bridge, and the lower is an interface it knows about (one registered by this driver, normally). Then, from within that notifier, it does a bunch of stuff behind the bridge's back, without the bridge necessarily knowing that there's somebody offloading that port. It looks like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v call_netdevice_notifiers | v dsa_slave_netdevice_event | v oh, hey! it's for me! | v .port_bridge_join What we do to solve the conundrum is to be less silent, and change the switchdev drivers to present themselves to the bridge. Something like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | hardware domain for v | this port, and zero dsa_slave_netdevice_event | if I got nothing. | | v | oh, hey! it's for me! | | | v | .port_bridge_join | | | +------------------------+ switchdev_bridge_port_offload(swp0, swp0) Then stacked interfaces (like bond0 on top of swp3/swp4) would be treated differently in DSA, depending on whether we can or cannot offload them. The offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | switchdev mark for v | bond0. dsa_slave_netdevice_event | Coincidentally (or not), | | bond0 and swp0, swp1, swp2 v | all have the same switchdev hmm, it's not quite for me, | mark now, since the ASIC but my driver has already | is able to forward towards called .port_lag_join | all these ports in hw. for it, because I have | a port with dp->lag_dev == bond0. | | | v | .port_bridge_join | for swp3 and swp4 | | | +------------------------+ switchdev_bridge_port_offload(bond0, swp3) switchdev_bridge_port_offload(bond0, swp4) And the non-offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge waiting: call_netdevice_notifiers ^ huh, switchdev_bridge_port_offload | | wasn't called, okay, I'll use a v | hwdom of zero for this one. dsa_slave_netdevice_event : Then packets received on swp0 will | : not be software-forwarded towards v : swp1, but they will towards bond0. it's not for me, but bond0 is an upper of swp3 and swp4, but their dp->lag_dev is NULL because they couldn't offload it. Basically we can draw the conclusion that the lowers of a bridge port can come and go, so depending on the configuration of lowers for a bridge port, it can dynamically toggle between offloaded and unoffloaded. Therefore, we need an equivalent switchdev_bridge_port_unoffload too. This patch changes the way any switchdev driver interacts with the bridge. From now on, everybody needs to call switchdev_bridge_port_offload and switchdev_bridge_port_unoffload, otherwise the bridge will treat the port as non-offloaded and allow software flooding to other ports from the same ASIC. Note that these functions lay the ground for a more complex handshake between switchdev drivers and the bridge in the future. For drivers that will request a replay of the switchdev objects when they offload and unoffload a bridge port (DSA, dpaa2-switch, ocelot), we place the call to switchdev_bridge_port_unoffload() strategically inside the NETDEV_PRECHANGEUPPER notifier's code path, and not inside NETDEV_CHANGEUPPER. This is because the switchdev object replay helpers need the netdev adjacency lists to be valid, and that is only true in NETDEV_PRECHANGEUPPER. Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch: regression Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Tested-by: Horatiu Vultur <horatiu.vultur@microchip.com> # ocelot-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:01 +00:00
struct net_device *brport_dev;
int err;
if (br_mst_enabled(br) && !dsa_port_supports_mst(dp))
return -EOPNOTSUPP;
net: dsa: enable flooding for bridge ports Switches work by learning the MAC address for each attached station by monitoring traffic from each station. When a station sends a packet, the switch records which port the MAC address is connected to. With IPv4 networking, before communication commences with a neighbour, an ARP packet is broadcasted to all stations asking for the MAC address corresponding with the IPv4. The desired station responds with an ARP reply, and the ARP reply causes the switch to learn which port the station is connected to. With IPv6 networking, the situation is rather different. Rather than broadcasting ARP packets, a "neighbour solicitation" is multicasted rather than broadcasted. This multicast needs to reach the intended station in order for the neighbour to be discovered. Once a neighbour has been discovered, and entered into the sending stations neighbour cache, communication can restart at a point later without sending a new neighbour solicitation, even if the entry in the neighbour cache is marked as stale. This can be after the MAC address has expired from the forwarding cache of the DSA switch - when that occurs, there is a long pause in communication. Our DSA implementation for mv88e6xxx switches disables flooding of multicast and unicast frames for bridged ports. As per the above description, this is fine for IPv4 networking, since the broadcasted ARP queries will be sent to and received by all stations on the same network. However, this breaks IPv6 very badly - blocking neighbour solicitations and later causing connections to stall. The defaults that the Linux bridge code expect from bridges are for unknown unicast and unknown multicast frames to be flooded to all ports on the bridge, which is at odds to the defaults adopted by our DSA implementation for mv88e6xxx switches. This commit enables by default flooding of both unknown unicast and unknown multicast frames whenever a port is added to a bridge, and disables the flooding when a port leaves the bridge. This means that mv88e6xxx DSA switches now behave as per the bridge(8) man page, and IPv6 works flawlessly through such a switch. Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Reviewed-by: Vivien Didelot <vivien.didelot@gmail.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-20 23:35:06 +00:00
/* Here the interface is already bridged. Reflect the current
* configuration so that drivers can program their chips accordingly.
*/
err = dsa_port_bridge_create(dp, br, extack);
if (err)
return err;
net: bridge: switchdev: let drivers inform which bridge ports are offloaded On reception of an skb, the bridge checks if it was marked as 'already forwarded in hardware' (checks if skb->offload_fwd_mark == 1), and if it is, it assigns the source hardware domain of that skb based on the hardware domain of the ingress port. Then during forwarding, it enforces that the egress port must have a different hardware domain than the ingress one (this is done in nbp_switchdev_allowed_egress). Non-switchdev drivers don't report any physical switch id (neither through devlink nor .ndo_get_port_parent_id), therefore the bridge assigns them a hardware domain of 0, and packets coming from them will always have skb->offload_fwd_mark = 0. So there aren't any restrictions. Problems appear due to the fact that DSA would like to perform software fallback for bonding and team interfaces that the physical switch cannot offload. +-- br0 ---+ / / | \ / / | \ / | | bond0 / | | / \ swp0 swp1 swp2 swp3 swp4 There, it is desirable that the presence of swp3 and swp4 under a non-offloaded LAG does not preclude us from doing hardware bridging beteen swp0, swp1 and swp2. The bandwidth of the CPU is often times high enough that software bridging between {swp0,swp1,swp2} and bond0 is not impractical. But this creates an impossible paradox given the current way in which port hardware domains are assigned. When the driver receives a packet from swp0 (say, due to flooding), it must set skb->offload_fwd_mark to something. - If we set it to 0, then the bridge will forward it towards swp1, swp2 and bond0. But the switch has already forwarded it towards swp1 and swp2 (not to bond0, remember, that isn't offloaded, so as far as the switch is concerned, ports swp3 and swp4 are not looking up the FDB, and the entire bond0 is a destination that is strictly behind the CPU). But we don't want duplicated traffic towards swp1 and swp2, so it's not ok to set skb->offload_fwd_mark = 0. - If we set it to 1, then the bridge will not forward the skb towards the ports with the same switchdev mark, i.e. not to swp1, swp2 and bond0. Towards swp1 and swp2 that's ok, but towards bond0? It should have forwarded the skb there. So the real issue is that bond0 will be assigned the same hardware domain as {swp0,swp1,swp2}, because the function that assigns hardware domains to bridge ports, nbp_switchdev_add(), recurses through bond0's lower interfaces until it finds something that implements devlink (calls dev_get_port_parent_id with bool recurse = true). This is a problem because the fact that bond0 can be offloaded by swp3 and swp4 in our example is merely an assumption. A solution is to give the bridge explicit hints as to what hardware domain it should use for each port. Currently, the bridging offload is very 'silent': a driver registers a netdevice notifier, which is put on the netns's notifier chain, and which sniffs around for NETDEV_CHANGEUPPER events where the upper is a bridge, and the lower is an interface it knows about (one registered by this driver, normally). Then, from within that notifier, it does a bunch of stuff behind the bridge's back, without the bridge necessarily knowing that there's somebody offloading that port. It looks like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v call_netdevice_notifiers | v dsa_slave_netdevice_event | v oh, hey! it's for me! | v .port_bridge_join What we do to solve the conundrum is to be less silent, and change the switchdev drivers to present themselves to the bridge. Something like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | hardware domain for v | this port, and zero dsa_slave_netdevice_event | if I got nothing. | | v | oh, hey! it's for me! | | | v | .port_bridge_join | | | +------------------------+ switchdev_bridge_port_offload(swp0, swp0) Then stacked interfaces (like bond0 on top of swp3/swp4) would be treated differently in DSA, depending on whether we can or cannot offload them. The offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | switchdev mark for v | bond0. dsa_slave_netdevice_event | Coincidentally (or not), | | bond0 and swp0, swp1, swp2 v | all have the same switchdev hmm, it's not quite for me, | mark now, since the ASIC but my driver has already | is able to forward towards called .port_lag_join | all these ports in hw. for it, because I have | a port with dp->lag_dev == bond0. | | | v | .port_bridge_join | for swp3 and swp4 | | | +------------------------+ switchdev_bridge_port_offload(bond0, swp3) switchdev_bridge_port_offload(bond0, swp4) And the non-offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge waiting: call_netdevice_notifiers ^ huh, switchdev_bridge_port_offload | | wasn't called, okay, I'll use a v | hwdom of zero for this one. dsa_slave_netdevice_event : Then packets received on swp0 will | : not be software-forwarded towards v : swp1, but they will towards bond0. it's not for me, but bond0 is an upper of swp3 and swp4, but their dp->lag_dev is NULL because they couldn't offload it. Basically we can draw the conclusion that the lowers of a bridge port can come and go, so depending on the configuration of lowers for a bridge port, it can dynamically toggle between offloaded and unoffloaded. Therefore, we need an equivalent switchdev_bridge_port_unoffload too. This patch changes the way any switchdev driver interacts with the bridge. From now on, everybody needs to call switchdev_bridge_port_offload and switchdev_bridge_port_unoffload, otherwise the bridge will treat the port as non-offloaded and allow software flooding to other ports from the same ASIC. Note that these functions lay the ground for a more complex handshake between switchdev drivers and the bridge in the future. For drivers that will request a replay of the switchdev objects when they offload and unoffload a bridge port (DSA, dpaa2-switch, ocelot), we place the call to switchdev_bridge_port_unoffload() strategically inside the NETDEV_PRECHANGEUPPER notifier's code path, and not inside NETDEV_CHANGEUPPER. This is because the switchdev object replay helpers need the netdev adjacency lists to be valid, and that is only true in NETDEV_PRECHANGEUPPER. Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch: regression Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Tested-by: Horatiu Vultur <horatiu.vultur@microchip.com> # ocelot-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:01 +00:00
brport_dev = dsa_port_to_bridge_port(dp);
net: dsa: keep the bridge_dev and bridge_num as part of the same structure The main desire behind this is to provide coherent bridge information to the fast path without locking. For example, right now we set dp->bridge_dev and dp->bridge_num from separate code paths, it is theoretically possible for a packet transmission to read these two port properties consecutively and find a bridge number which does not correspond with the bridge device. Another desire is to start passing more complex bridge information to dsa_switch_ops functions. For example, with FDB isolation, it is expected that drivers will need to be passed the bridge which requested an FDB/MDB entry to be offloaded, and along with that bridge_dev, the associated bridge_num should be passed too, in case the driver might want to implement an isolation scheme based on that number. We already pass the {bridge_dev, bridge_num} pair to the TX forwarding offload switch API, however we'd like to remove that and squash it into the basic bridge join/leave API. So that means we need to pass this pair to the bridge join/leave API. During dsa_port_bridge_leave, first we unset dp->bridge_dev, then we call the driver's .port_bridge_leave with what used to be our dp->bridge_dev, but provided as an argument. When bridge_dev and bridge_num get folded into a single structure, we need to preserve this behavior in dsa_port_bridge_leave: we need a copy of what used to be in dp->bridge. Switch drivers check bridge membership by comparing dp->bridge_dev with the provided bridge_dev, but now, if we provide the struct dsa_bridge as a pointer, they cannot keep comparing dp->bridge to the provided pointer, since this only points to an on-stack copy. To make this obvious and prevent driver writers from forgetting and doing stupid things, in this new API, the struct dsa_bridge is provided as a full structure (not very large, contains an int and a pointer) instead of a pointer. An explicit comparison function needs to be used to determine bridge membership: dsa_port_offloads_bridge(). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-12-06 16:57:56 +00:00
info.bridge = *dp->bridge;
net: dsa: permit cross-chip bridging between all trees in the system One way of utilizing DSA is by cascading switches which do not all have compatible taggers. Consider the following real-life topology: +---------------------------------------------------------------+ | LS1028A | | +------------------------------+ | | | DSA master for Felix | | | |(internal ENETC port 2: eno2))| | | +------------+------------------------------+-------------+ | | | Felix embedded L2 switch | | | | | | | | +--------------+ +--------------+ +--------------+ | | | | |DSA master for| |DSA master for| |DSA master for| | | | | | SJA1105 1 | | SJA1105 2 | | SJA1105 3 | | | | | |(Felix port 1)| |(Felix port 2)| |(Felix port 3)| | | +--+-+--------------+---+--------------+---+--------------+--+--+ +-----------------------+ +-----------------------+ +-----------------------+ | SJA1105 switch 1 | | SJA1105 switch 2 | | SJA1105 switch 3 | +-----+-----+-----+-----+ +-----+-----+-----+-----+ +-----+-----+-----+-----+ |sw1p0|sw1p1|sw1p2|sw1p3| |sw2p0|sw2p1|sw2p2|sw2p3| |sw3p0|sw3p1|sw3p2|sw3p3| +-----+-----+-----+-----+ +-----+-----+-----+-----+ +-----+-----+-----+-----+ The above can be described in the device tree as follows (obviously not complete): mscc_felix { dsa,member = <0 0>; ports { port@4 { ethernet = <&enetc_port2>; }; }; }; sja1105_switch1 { dsa,member = <1 1>; ports { port@4 { ethernet = <&mscc_felix_port1>; }; }; }; sja1105_switch2 { dsa,member = <2 2>; ports { port@4 { ethernet = <&mscc_felix_port2>; }; }; }; sja1105_switch3 { dsa,member = <3 3>; ports { port@4 { ethernet = <&mscc_felix_port3>; }; }; }; Basically we instantiate one DSA switch tree for every hardware switch in the system, but we still give them globally unique switch IDs (will come back to that later). Having 3 disjoint switch trees makes the tagger drivers "just work", because net devices are registered for the 3 Felix DSA master ports, and they are also DSA slave ports to the ENETC port. So packets received on the ENETC port are stripped of their stacked DSA tags one by one. Currently, hardware bridging between ports on the same sja1105 chip is possible, but switching between sja1105 ports on different chips is handled by the software bridge. This is fine, but we can do better. In fact, the dsa_8021q tag used by sja1105 is compatible with cascading. In other words, a sja1105 switch can correctly parse and route a packet containing a dsa_8021q tag. So if we could enable hardware bridging on the Felix DSA master ports, cross-chip bridging could be completely offloaded. Such as system would be used as follows: ip link add dev br0 type bridge && ip link set dev br0 up for port in sw0p0 sw0p1 sw0p2 sw0p3 \ sw1p0 sw1p1 sw1p2 sw1p3 \ sw2p0 sw2p1 sw2p2 sw2p3; do ip link set dev $port master br0 done The above makes switching between ports on the same row be performed in hardware, and between ports on different rows in software. Now assume the Felix switch ports are called swp0, swp1, swp2. By running the following extra commands: ip link add dev br1 type bridge && ip link set dev br1 up for port in swp0 swp1 swp2; do ip link set dev $port master br1 done the CPU no longer sees packets which traverse sja1105 switch boundaries and can be forwarded directly by Felix. The br1 bridge would not be used for any sort of traffic termination. For this to work, we need to give drivers an opportunity to listen for bridging events on DSA trees other than their own, and pass that other tree index as argument. I have made the assumption, for the moment, that the other existing DSA notifiers don't need to be broadcast to other trees. That assumption might turn out to be incorrect. But in the meantime, introduce a dsa_broadcast function, similar in purpose to dsa_port_notify, which is used only by the bridging notifiers. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2020-05-10 16:37:41 +00:00
err = dsa_broadcast(DSA_NOTIFIER_BRIDGE_JOIN, &info);
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
if (err)
goto out_rollback;
/* Drivers which support bridge TX forwarding should set this */
dp->bridge->tx_fwd_offload = info.tx_fwd_offload;
net: dsa: add support for bridge TX forwarding offload For a DSA switch, to offload the forwarding process of a bridge device means to send the packets coming from the software bridge as data plane packets. This is contrary to everything that DSA has done so far, because the current taggers only know to send control packets (ones that target a specific destination port), whereas data plane packets are supposed to be forwarded according to the FDB lookup, much like packets ingressing on any regular ingress port. If the FDB lookup process returns multiple destination ports (flooding, multicast), then replication is also handled by the switch hardware - the bridge only sends a single packet and avoids the skb_clone(). DSA keeps for each bridge port a zero-based index (the number of the bridge). Multiple ports performing TX forwarding offload to the same bridge have the same dp->bridge_num value, and ports not offloading the TX data plane of a bridge have dp->bridge_num = -1. The tagger can check if the packet that is being transmitted on has skb->offload_fwd_mark = true or not. If it does, it can be sure that the packet belongs to the data plane of a bridge, further information about which can be obtained based on dp->bridge_dev and dp->bridge_num. It can then compose a DSA tag for injecting a data plane packet into that bridge number. For the switch driver side, we offer two new dsa_switch_ops methods, called .port_bridge_fwd_offload_{add,del}, which are modeled after .port_bridge_{join,leave}. These methods are provided in case the driver needs to configure the hardware to treat packets coming from that bridge software interface as data plane packets. The switchdev <-> bridge interaction happens during the netdev_master_upper_dev_link() call, so to switch drivers, the effect is that the .port_bridge_fwd_offload_add() method is called immediately after .port_bridge_join(). If the bridge number exceeds the number of bridges for which the switch driver can offload the TX data plane (and this includes the case where the driver can offload none), DSA falls back to simply returning tx_fwd_offload = false in the switchdev_bridge_port_offload() call. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-22 15:55:40 +00:00
net: bridge: move the switchdev object replay helpers to "push" mode Starting with commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"), DSA has introduced some bridge helpers that replay switchdev events (FDB/MDB/VLAN additions and deletions) that can be lost by the switchdev drivers in a variety of circumstances: - an IP multicast group was host-joined on the bridge itself before any switchdev port joined the bridge, leading to the host MDB entries missing in the hardware database. - during the bridge creation process, the MAC address of the bridge was added to the FDB as an entry pointing towards the bridge device itself, but with no switchdev ports being part of the bridge yet, this local FDB entry would remain unknown to the switchdev hardware database. - a VLAN/FDB/MDB was added to a bridge port that is a LAG interface, before any switchdev port joined that LAG, leading to the hardware database missing those entries. - a switchdev port left a LAG that is a bridge port, while the LAG remained part of the bridge, and all FDB/MDB/VLAN entries remained installed in the hardware database of the switchdev port. Also, since commit 0d2cfbd41c4a ("net: bridge: ignore switchdev events for LAG ports which didn't request replay"), DSA introduced a method, based on a const void *ctx, to ensure that two switchdev ports under the same LAG that is a bridge port do not see the same MDB/VLAN entry being replayed twice by the bridge, once for every bridge port that joins the LAG. With so many ordering corner cases being possible, it seems unreasonable to expect a switchdev driver writer to get it right from the first try. Therefore, now that DSA has experimented with the bridge replay helpers for a little bit, we can move the code to the bridge driver where it is more readily available to all switchdev drivers. To convert the switchdev object replay helpers from "pull mode" (where the driver asks for them) to a "push mode" (where the bridge offers them automatically), the biggest problem is that the bridge needs to be aware when a switchdev port joins and leaves, even when the switchdev is only indirectly a bridge port (for example when the bridge port is a LAG upper of the switchdev). Luckily, we already have a hook for that, in the form of the newly introduced switchdev_bridge_port_offload() and switchdev_bridge_port_unoffload() calls. These offer a natural place for hooking the object addition and deletion replays. Extend the above 2 functions with: - pointers to the switchdev atomic notifier (for FDB replays) and the blocking notifier (for MDB and VLAN replays). - the "const void *ctx" argument required for drivers to be able to disambiguate between which port is targeted, when multiple ports are lowers of the same LAG that is a bridge port. Most of the drivers pass NULL to this argument, except the ones that support LAG offload and have the proper context check already in place in the switchdev blocking notifier handler. Also unexport the replay helpers, since nobody except the bridge calls them directly now. Note that: (a) we abuse the terminology slightly, because FDB entries are not "switchdev objects", but we count them as objects nonetheless. With no direct way to prove it, I think they are not modeled as switchdev objects because those can only be installed by the bridge to the hardware (as opposed to FDB entries which can be propagated in the other direction too). This is merely an abuse of terms, FDB entries are replayed too, despite not being objects. (b) the bridge does not attempt to sync port attributes to newly joined ports, just the countable stuff (the objects). The reason for this is simple: no universal and symmetric way to sync and unsync them is known. For example, VLAN filtering: what to do on unsync, disable or leave it enabled? Similarly, STP state, ageing timer, etc etc. What a switchdev port does when it becomes standalone again is not really up to the bridge's competence, and the driver should deal with it. On the other hand, replaying deletions of switchdev objects can be seen a matter of cleanup and therefore be treated by the bridge, hence this patch. We make the replay helpers opt-in for drivers, because they might not bring immediate benefits for them: - nbp_vlan_init() is called _after_ netdev_master_upper_dev_link(), so br_vlan_replay() should not do anything for the new drivers on which we call it. The existing drivers where there was even a slight possibility for there to exist a VLAN on a bridge port before they join it are already guarded against this: mlxsw and prestera deny joining LAG interfaces that are members of a bridge. - br_fdb_replay() should now notify of local FDB entries, but I patched all drivers except DSA to ignore these new entries in commit 2c4eca3ef716 ("net: bridge: switchdev: include local flag in FDB notifications"). Driver authors can lift this restriction as they wish, and when they do, they can also opt into the FDB replay functionality. - br_mdb_replay() should fix a real issue which is described in commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"). However most drivers do not offload the SWITCHDEV_OBJ_ID_HOST_MDB to see this issue: only cpsw and am65_cpsw offload this switchdev object, and I don't completely understand the way in which they offload this switchdev object anyway. So I'll leave it up to these drivers' respective maintainers to opt into br_mdb_replay(). So most of the drivers pass NULL notifier blocks for the replay helpers, except: - dpaa2-switch which was already acked/regression-tested with the helpers enabled (and there isn't much of a downside in having them) - ocelot which already had replay logic in "pull" mode - DSA which already had replay logic in "pull" mode An important observation is that the drivers which don't currently request bridge event replays don't even have the switchdev_bridge_port_{offload,unoffload} calls placed in proper places right now. This was done to avoid unnecessary rework for drivers which might never even add support for this. For driver writers who wish to add replay support, this can be used as a tentative placement guide: https://patchwork.kernel.org/project/netdevbpf/patch/20210720134655.892334-11-vladimir.oltean@nxp.com/ Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:03 +00:00
err = switchdev_bridge_port_offload(brport_dev, dev, dp,
&dsa_user_switchdev_notifier,
&dsa_user_switchdev_blocking_notifier,
dp->bridge->tx_fwd_offload, extack);
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
if (err)
goto out_rollback_unbridge;
net: bridge: move the switchdev object replay helpers to "push" mode Starting with commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"), DSA has introduced some bridge helpers that replay switchdev events (FDB/MDB/VLAN additions and deletions) that can be lost by the switchdev drivers in a variety of circumstances: - an IP multicast group was host-joined on the bridge itself before any switchdev port joined the bridge, leading to the host MDB entries missing in the hardware database. - during the bridge creation process, the MAC address of the bridge was added to the FDB as an entry pointing towards the bridge device itself, but with no switchdev ports being part of the bridge yet, this local FDB entry would remain unknown to the switchdev hardware database. - a VLAN/FDB/MDB was added to a bridge port that is a LAG interface, before any switchdev port joined that LAG, leading to the hardware database missing those entries. - a switchdev port left a LAG that is a bridge port, while the LAG remained part of the bridge, and all FDB/MDB/VLAN entries remained installed in the hardware database of the switchdev port. Also, since commit 0d2cfbd41c4a ("net: bridge: ignore switchdev events for LAG ports which didn't request replay"), DSA introduced a method, based on a const void *ctx, to ensure that two switchdev ports under the same LAG that is a bridge port do not see the same MDB/VLAN entry being replayed twice by the bridge, once for every bridge port that joins the LAG. With so many ordering corner cases being possible, it seems unreasonable to expect a switchdev driver writer to get it right from the first try. Therefore, now that DSA has experimented with the bridge replay helpers for a little bit, we can move the code to the bridge driver where it is more readily available to all switchdev drivers. To convert the switchdev object replay helpers from "pull mode" (where the driver asks for them) to a "push mode" (where the bridge offers them automatically), the biggest problem is that the bridge needs to be aware when a switchdev port joins and leaves, even when the switchdev is only indirectly a bridge port (for example when the bridge port is a LAG upper of the switchdev). Luckily, we already have a hook for that, in the form of the newly introduced switchdev_bridge_port_offload() and switchdev_bridge_port_unoffload() calls. These offer a natural place for hooking the object addition and deletion replays. Extend the above 2 functions with: - pointers to the switchdev atomic notifier (for FDB replays) and the blocking notifier (for MDB and VLAN replays). - the "const void *ctx" argument required for drivers to be able to disambiguate between which port is targeted, when multiple ports are lowers of the same LAG that is a bridge port. Most of the drivers pass NULL to this argument, except the ones that support LAG offload and have the proper context check already in place in the switchdev blocking notifier handler. Also unexport the replay helpers, since nobody except the bridge calls them directly now. Note that: (a) we abuse the terminology slightly, because FDB entries are not "switchdev objects", but we count them as objects nonetheless. With no direct way to prove it, I think they are not modeled as switchdev objects because those can only be installed by the bridge to the hardware (as opposed to FDB entries which can be propagated in the other direction too). This is merely an abuse of terms, FDB entries are replayed too, despite not being objects. (b) the bridge does not attempt to sync port attributes to newly joined ports, just the countable stuff (the objects). The reason for this is simple: no universal and symmetric way to sync and unsync them is known. For example, VLAN filtering: what to do on unsync, disable or leave it enabled? Similarly, STP state, ageing timer, etc etc. What a switchdev port does when it becomes standalone again is not really up to the bridge's competence, and the driver should deal with it. On the other hand, replaying deletions of switchdev objects can be seen a matter of cleanup and therefore be treated by the bridge, hence this patch. We make the replay helpers opt-in for drivers, because they might not bring immediate benefits for them: - nbp_vlan_init() is called _after_ netdev_master_upper_dev_link(), so br_vlan_replay() should not do anything for the new drivers on which we call it. The existing drivers where there was even a slight possibility for there to exist a VLAN on a bridge port before they join it are already guarded against this: mlxsw and prestera deny joining LAG interfaces that are members of a bridge. - br_fdb_replay() should now notify of local FDB entries, but I patched all drivers except DSA to ignore these new entries in commit 2c4eca3ef716 ("net: bridge: switchdev: include local flag in FDB notifications"). Driver authors can lift this restriction as they wish, and when they do, they can also opt into the FDB replay functionality. - br_mdb_replay() should fix a real issue which is described in commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"). However most drivers do not offload the SWITCHDEV_OBJ_ID_HOST_MDB to see this issue: only cpsw and am65_cpsw offload this switchdev object, and I don't completely understand the way in which they offload this switchdev object anyway. So I'll leave it up to these drivers' respective maintainers to opt into br_mdb_replay(). So most of the drivers pass NULL notifier blocks for the replay helpers, except: - dpaa2-switch which was already acked/regression-tested with the helpers enabled (and there isn't much of a downside in having them) - ocelot which already had replay logic in "pull" mode - DSA which already had replay logic in "pull" mode An important observation is that the drivers which don't currently request bridge event replays don't even have the switchdev_bridge_port_{offload,unoffload} calls placed in proper places right now. This was done to avoid unnecessary rework for drivers which might never even add support for this. For driver writers who wish to add replay support, this can be used as a tentative placement guide: https://patchwork.kernel.org/project/netdevbpf/patch/20210720134655.892334-11-vladimir.oltean@nxp.com/ Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:03 +00:00
err = dsa_port_switchdev_sync_attrs(dp, extack);
net: bridge: switchdev: let drivers inform which bridge ports are offloaded On reception of an skb, the bridge checks if it was marked as 'already forwarded in hardware' (checks if skb->offload_fwd_mark == 1), and if it is, it assigns the source hardware domain of that skb based on the hardware domain of the ingress port. Then during forwarding, it enforces that the egress port must have a different hardware domain than the ingress one (this is done in nbp_switchdev_allowed_egress). Non-switchdev drivers don't report any physical switch id (neither through devlink nor .ndo_get_port_parent_id), therefore the bridge assigns them a hardware domain of 0, and packets coming from them will always have skb->offload_fwd_mark = 0. So there aren't any restrictions. Problems appear due to the fact that DSA would like to perform software fallback for bonding and team interfaces that the physical switch cannot offload. +-- br0 ---+ / / | \ / / | \ / | | bond0 / | | / \ swp0 swp1 swp2 swp3 swp4 There, it is desirable that the presence of swp3 and swp4 under a non-offloaded LAG does not preclude us from doing hardware bridging beteen swp0, swp1 and swp2. The bandwidth of the CPU is often times high enough that software bridging between {swp0,swp1,swp2} and bond0 is not impractical. But this creates an impossible paradox given the current way in which port hardware domains are assigned. When the driver receives a packet from swp0 (say, due to flooding), it must set skb->offload_fwd_mark to something. - If we set it to 0, then the bridge will forward it towards swp1, swp2 and bond0. But the switch has already forwarded it towards swp1 and swp2 (not to bond0, remember, that isn't offloaded, so as far as the switch is concerned, ports swp3 and swp4 are not looking up the FDB, and the entire bond0 is a destination that is strictly behind the CPU). But we don't want duplicated traffic towards swp1 and swp2, so it's not ok to set skb->offload_fwd_mark = 0. - If we set it to 1, then the bridge will not forward the skb towards the ports with the same switchdev mark, i.e. not to swp1, swp2 and bond0. Towards swp1 and swp2 that's ok, but towards bond0? It should have forwarded the skb there. So the real issue is that bond0 will be assigned the same hardware domain as {swp0,swp1,swp2}, because the function that assigns hardware domains to bridge ports, nbp_switchdev_add(), recurses through bond0's lower interfaces until it finds something that implements devlink (calls dev_get_port_parent_id with bool recurse = true). This is a problem because the fact that bond0 can be offloaded by swp3 and swp4 in our example is merely an assumption. A solution is to give the bridge explicit hints as to what hardware domain it should use for each port. Currently, the bridging offload is very 'silent': a driver registers a netdevice notifier, which is put on the netns's notifier chain, and which sniffs around for NETDEV_CHANGEUPPER events where the upper is a bridge, and the lower is an interface it knows about (one registered by this driver, normally). Then, from within that notifier, it does a bunch of stuff behind the bridge's back, without the bridge necessarily knowing that there's somebody offloading that port. It looks like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v call_netdevice_notifiers | v dsa_slave_netdevice_event | v oh, hey! it's for me! | v .port_bridge_join What we do to solve the conundrum is to be less silent, and change the switchdev drivers to present themselves to the bridge. Something like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | hardware domain for v | this port, and zero dsa_slave_netdevice_event | if I got nothing. | | v | oh, hey! it's for me! | | | v | .port_bridge_join | | | +------------------------+ switchdev_bridge_port_offload(swp0, swp0) Then stacked interfaces (like bond0 on top of swp3/swp4) would be treated differently in DSA, depending on whether we can or cannot offload them. The offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | switchdev mark for v | bond0. dsa_slave_netdevice_event | Coincidentally (or not), | | bond0 and swp0, swp1, swp2 v | all have the same switchdev hmm, it's not quite for me, | mark now, since the ASIC but my driver has already | is able to forward towards called .port_lag_join | all these ports in hw. for it, because I have | a port with dp->lag_dev == bond0. | | | v | .port_bridge_join | for swp3 and swp4 | | | +------------------------+ switchdev_bridge_port_offload(bond0, swp3) switchdev_bridge_port_offload(bond0, swp4) And the non-offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge waiting: call_netdevice_notifiers ^ huh, switchdev_bridge_port_offload | | wasn't called, okay, I'll use a v | hwdom of zero for this one. dsa_slave_netdevice_event : Then packets received on swp0 will | : not be software-forwarded towards v : swp1, but they will towards bond0. it's not for me, but bond0 is an upper of swp3 and swp4, but their dp->lag_dev is NULL because they couldn't offload it. Basically we can draw the conclusion that the lowers of a bridge port can come and go, so depending on the configuration of lowers for a bridge port, it can dynamically toggle between offloaded and unoffloaded. Therefore, we need an equivalent switchdev_bridge_port_unoffload too. This patch changes the way any switchdev driver interacts with the bridge. From now on, everybody needs to call switchdev_bridge_port_offload and switchdev_bridge_port_unoffload, otherwise the bridge will treat the port as non-offloaded and allow software flooding to other ports from the same ASIC. Note that these functions lay the ground for a more complex handshake between switchdev drivers and the bridge in the future. For drivers that will request a replay of the switchdev objects when they offload and unoffload a bridge port (DSA, dpaa2-switch, ocelot), we place the call to switchdev_bridge_port_unoffload() strategically inside the NETDEV_PRECHANGEUPPER notifier's code path, and not inside NETDEV_CHANGEUPPER. This is because the switchdev object replay helpers need the netdev adjacency lists to be valid, and that is only true in NETDEV_PRECHANGEUPPER. Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch: regression Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Tested-by: Horatiu Vultur <horatiu.vultur@microchip.com> # ocelot-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:01 +00:00
if (err)
goto out_rollback_unoffload;
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
return 0;
net: bridge: switchdev: let drivers inform which bridge ports are offloaded On reception of an skb, the bridge checks if it was marked as 'already forwarded in hardware' (checks if skb->offload_fwd_mark == 1), and if it is, it assigns the source hardware domain of that skb based on the hardware domain of the ingress port. Then during forwarding, it enforces that the egress port must have a different hardware domain than the ingress one (this is done in nbp_switchdev_allowed_egress). Non-switchdev drivers don't report any physical switch id (neither through devlink nor .ndo_get_port_parent_id), therefore the bridge assigns them a hardware domain of 0, and packets coming from them will always have skb->offload_fwd_mark = 0. So there aren't any restrictions. Problems appear due to the fact that DSA would like to perform software fallback for bonding and team interfaces that the physical switch cannot offload. +-- br0 ---+ / / | \ / / | \ / | | bond0 / | | / \ swp0 swp1 swp2 swp3 swp4 There, it is desirable that the presence of swp3 and swp4 under a non-offloaded LAG does not preclude us from doing hardware bridging beteen swp0, swp1 and swp2. The bandwidth of the CPU is often times high enough that software bridging between {swp0,swp1,swp2} and bond0 is not impractical. But this creates an impossible paradox given the current way in which port hardware domains are assigned. When the driver receives a packet from swp0 (say, due to flooding), it must set skb->offload_fwd_mark to something. - If we set it to 0, then the bridge will forward it towards swp1, swp2 and bond0. But the switch has already forwarded it towards swp1 and swp2 (not to bond0, remember, that isn't offloaded, so as far as the switch is concerned, ports swp3 and swp4 are not looking up the FDB, and the entire bond0 is a destination that is strictly behind the CPU). But we don't want duplicated traffic towards swp1 and swp2, so it's not ok to set skb->offload_fwd_mark = 0. - If we set it to 1, then the bridge will not forward the skb towards the ports with the same switchdev mark, i.e. not to swp1, swp2 and bond0. Towards swp1 and swp2 that's ok, but towards bond0? It should have forwarded the skb there. So the real issue is that bond0 will be assigned the same hardware domain as {swp0,swp1,swp2}, because the function that assigns hardware domains to bridge ports, nbp_switchdev_add(), recurses through bond0's lower interfaces until it finds something that implements devlink (calls dev_get_port_parent_id with bool recurse = true). This is a problem because the fact that bond0 can be offloaded by swp3 and swp4 in our example is merely an assumption. A solution is to give the bridge explicit hints as to what hardware domain it should use for each port. Currently, the bridging offload is very 'silent': a driver registers a netdevice notifier, which is put on the netns's notifier chain, and which sniffs around for NETDEV_CHANGEUPPER events where the upper is a bridge, and the lower is an interface it knows about (one registered by this driver, normally). Then, from within that notifier, it does a bunch of stuff behind the bridge's back, without the bridge necessarily knowing that there's somebody offloading that port. It looks like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v call_netdevice_notifiers | v dsa_slave_netdevice_event | v oh, hey! it's for me! | v .port_bridge_join What we do to solve the conundrum is to be less silent, and change the switchdev drivers to present themselves to the bridge. Something like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | hardware domain for v | this port, and zero dsa_slave_netdevice_event | if I got nothing. | | v | oh, hey! it's for me! | | | v | .port_bridge_join | | | +------------------------+ switchdev_bridge_port_offload(swp0, swp0) Then stacked interfaces (like bond0 on top of swp3/swp4) would be treated differently in DSA, depending on whether we can or cannot offload them. The offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | switchdev mark for v | bond0. dsa_slave_netdevice_event | Coincidentally (or not), | | bond0 and swp0, swp1, swp2 v | all have the same switchdev hmm, it's not quite for me, | mark now, since the ASIC but my driver has already | is able to forward towards called .port_lag_join | all these ports in hw. for it, because I have | a port with dp->lag_dev == bond0. | | | v | .port_bridge_join | for swp3 and swp4 | | | +------------------------+ switchdev_bridge_port_offload(bond0, swp3) switchdev_bridge_port_offload(bond0, swp4) And the non-offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge waiting: call_netdevice_notifiers ^ huh, switchdev_bridge_port_offload | | wasn't called, okay, I'll use a v | hwdom of zero for this one. dsa_slave_netdevice_event : Then packets received on swp0 will | : not be software-forwarded towards v : swp1, but they will towards bond0. it's not for me, but bond0 is an upper of swp3 and swp4, but their dp->lag_dev is NULL because they couldn't offload it. Basically we can draw the conclusion that the lowers of a bridge port can come and go, so depending on the configuration of lowers for a bridge port, it can dynamically toggle between offloaded and unoffloaded. Therefore, we need an equivalent switchdev_bridge_port_unoffload too. This patch changes the way any switchdev driver interacts with the bridge. From now on, everybody needs to call switchdev_bridge_port_offload and switchdev_bridge_port_unoffload, otherwise the bridge will treat the port as non-offloaded and allow software flooding to other ports from the same ASIC. Note that these functions lay the ground for a more complex handshake between switchdev drivers and the bridge in the future. For drivers that will request a replay of the switchdev objects when they offload and unoffload a bridge port (DSA, dpaa2-switch, ocelot), we place the call to switchdev_bridge_port_unoffload() strategically inside the NETDEV_PRECHANGEUPPER notifier's code path, and not inside NETDEV_CHANGEUPPER. This is because the switchdev object replay helpers need the netdev adjacency lists to be valid, and that is only true in NETDEV_PRECHANGEUPPER. Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch: regression Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Tested-by: Horatiu Vultur <horatiu.vultur@microchip.com> # ocelot-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:01 +00:00
out_rollback_unoffload:
net: bridge: move the switchdev object replay helpers to "push" mode Starting with commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"), DSA has introduced some bridge helpers that replay switchdev events (FDB/MDB/VLAN additions and deletions) that can be lost by the switchdev drivers in a variety of circumstances: - an IP multicast group was host-joined on the bridge itself before any switchdev port joined the bridge, leading to the host MDB entries missing in the hardware database. - during the bridge creation process, the MAC address of the bridge was added to the FDB as an entry pointing towards the bridge device itself, but with no switchdev ports being part of the bridge yet, this local FDB entry would remain unknown to the switchdev hardware database. - a VLAN/FDB/MDB was added to a bridge port that is a LAG interface, before any switchdev port joined that LAG, leading to the hardware database missing those entries. - a switchdev port left a LAG that is a bridge port, while the LAG remained part of the bridge, and all FDB/MDB/VLAN entries remained installed in the hardware database of the switchdev port. Also, since commit 0d2cfbd41c4a ("net: bridge: ignore switchdev events for LAG ports which didn't request replay"), DSA introduced a method, based on a const void *ctx, to ensure that two switchdev ports under the same LAG that is a bridge port do not see the same MDB/VLAN entry being replayed twice by the bridge, once for every bridge port that joins the LAG. With so many ordering corner cases being possible, it seems unreasonable to expect a switchdev driver writer to get it right from the first try. Therefore, now that DSA has experimented with the bridge replay helpers for a little bit, we can move the code to the bridge driver where it is more readily available to all switchdev drivers. To convert the switchdev object replay helpers from "pull mode" (where the driver asks for them) to a "push mode" (where the bridge offers them automatically), the biggest problem is that the bridge needs to be aware when a switchdev port joins and leaves, even when the switchdev is only indirectly a bridge port (for example when the bridge port is a LAG upper of the switchdev). Luckily, we already have a hook for that, in the form of the newly introduced switchdev_bridge_port_offload() and switchdev_bridge_port_unoffload() calls. These offer a natural place for hooking the object addition and deletion replays. Extend the above 2 functions with: - pointers to the switchdev atomic notifier (for FDB replays) and the blocking notifier (for MDB and VLAN replays). - the "const void *ctx" argument required for drivers to be able to disambiguate between which port is targeted, when multiple ports are lowers of the same LAG that is a bridge port. Most of the drivers pass NULL to this argument, except the ones that support LAG offload and have the proper context check already in place in the switchdev blocking notifier handler. Also unexport the replay helpers, since nobody except the bridge calls them directly now. Note that: (a) we abuse the terminology slightly, because FDB entries are not "switchdev objects", but we count them as objects nonetheless. With no direct way to prove it, I think they are not modeled as switchdev objects because those can only be installed by the bridge to the hardware (as opposed to FDB entries which can be propagated in the other direction too). This is merely an abuse of terms, FDB entries are replayed too, despite not being objects. (b) the bridge does not attempt to sync port attributes to newly joined ports, just the countable stuff (the objects). The reason for this is simple: no universal and symmetric way to sync and unsync them is known. For example, VLAN filtering: what to do on unsync, disable or leave it enabled? Similarly, STP state, ageing timer, etc etc. What a switchdev port does when it becomes standalone again is not really up to the bridge's competence, and the driver should deal with it. On the other hand, replaying deletions of switchdev objects can be seen a matter of cleanup and therefore be treated by the bridge, hence this patch. We make the replay helpers opt-in for drivers, because they might not bring immediate benefits for them: - nbp_vlan_init() is called _after_ netdev_master_upper_dev_link(), so br_vlan_replay() should not do anything for the new drivers on which we call it. The existing drivers where there was even a slight possibility for there to exist a VLAN on a bridge port before they join it are already guarded against this: mlxsw and prestera deny joining LAG interfaces that are members of a bridge. - br_fdb_replay() should now notify of local FDB entries, but I patched all drivers except DSA to ignore these new entries in commit 2c4eca3ef716 ("net: bridge: switchdev: include local flag in FDB notifications"). Driver authors can lift this restriction as they wish, and when they do, they can also opt into the FDB replay functionality. - br_mdb_replay() should fix a real issue which is described in commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"). However most drivers do not offload the SWITCHDEV_OBJ_ID_HOST_MDB to see this issue: only cpsw and am65_cpsw offload this switchdev object, and I don't completely understand the way in which they offload this switchdev object anyway. So I'll leave it up to these drivers' respective maintainers to opt into br_mdb_replay(). So most of the drivers pass NULL notifier blocks for the replay helpers, except: - dpaa2-switch which was already acked/regression-tested with the helpers enabled (and there isn't much of a downside in having them) - ocelot which already had replay logic in "pull" mode - DSA which already had replay logic in "pull" mode An important observation is that the drivers which don't currently request bridge event replays don't even have the switchdev_bridge_port_{offload,unoffload} calls placed in proper places right now. This was done to avoid unnecessary rework for drivers which might never even add support for this. For driver writers who wish to add replay support, this can be used as a tentative placement guide: https://patchwork.kernel.org/project/netdevbpf/patch/20210720134655.892334-11-vladimir.oltean@nxp.com/ Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:03 +00:00
switchdev_bridge_port_unoffload(brport_dev, dp,
&dsa_user_switchdev_notifier,
&dsa_user_switchdev_blocking_notifier);
net: dsa: flush switchdev workqueue on bridge join error path There is a race between switchdev_bridge_port_offload() and the dsa_port_switchdev_sync_attrs() call right below it. When switchdev_bridge_port_offload() finishes, FDB entries have been replayed by the bridge, but are scheduled for deferred execution later. However dsa_port_switchdev_sync_attrs -> dsa_port_can_apply_vlan_filtering() may impose restrictions on the vlan_filtering attribute and refuse offloading. When this happens, the delayed FDB entries will dereference dp->bridge, which is a NULL pointer because we have stopped the process of offloading this bridge. Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000 Workqueue: dsa_ordered dsa_slave_switchdev_event_work pc : dsa_port_bridge_host_fdb_del+0x64/0x100 lr : dsa_slave_switchdev_event_work+0x130/0x1bc Call trace: dsa_port_bridge_host_fdb_del+0x64/0x100 dsa_slave_switchdev_event_work+0x130/0x1bc process_one_work+0x294/0x670 worker_thread+0x80/0x460 ---[ end trace 0000000000000000 ]--- Error: dsa_core: Must first remove VLAN uppers having VIDs also present in bridge. Fix the bug by doing what we do on the normal bridge leave path as well, which is to wait until the deferred FDB entries complete executing, then exit. The placement of dsa_flush_workqueue() after switchdev_bridge_port_unoffload() guarantees that both the FDB additions and deletions on rollback are waited for. Fixes: d7d0d423dbaa ("net: dsa: flush switchdev workqueue when leaving the bridge") Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://lore.kernel.org/r/20220507134550.1849834-1-vladimir.oltean@nxp.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-05-07 13:45:50 +00:00
dsa_flush_workqueue();
net: dsa: inherit the actual bridge port flags at join time DSA currently assumes that the bridge port starts off with this constellation of bridge port flags: - learning on - unicast flooding on - multicast flooding on - broadcast flooding on just by virtue of code copy-pasta from the bridge layer (new_nbp). This was a simple enough strategy thus far, because the 'bridge join' moment always coincided with the 'bridge port creation' moment. But with sandwiched interfaces, such as: br0 | bond0 | swp0 it may happen that the user has had time to change the bridge port flags of bond0 before enslaving swp0 to it. In that case, swp0 will falsely assume that the bridge port flags are those determined by new_nbp, when in fact this can happen: ip link add br0 type bridge ip link add bond0 type bond ip link set bond0 master br0 ip link set bond0 type bridge_slave learning off ip link set swp0 master br0 Now swp0 has learning enabled, bond0 has learning disabled. Not nice. Fix this by "dumpster diving" through the actual bridge port flags with br_port_flag_is_set, at bridge join time. We use this opportunity to split dsa_port_change_brport_flags into two distinct functions called dsa_port_inherit_brport_flags and dsa_port_clear_brport_flags, now that the implementation for the two cases is no longer similar. This patch also creates two functions called dsa_port_switchdev_sync and dsa_port_switchdev_unsync which collect what we have so far, even if that's asymmetrical. More is going to be added in the next patch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-03-22 23:51:49 +00:00
out_rollback_unbridge:
dsa_broadcast(DSA_NOTIFIER_BRIDGE_LEAVE, &info);
out_rollback:
dsa_port_bridge_destroy(dp, br);
return err;
}
net: bridge: move the switchdev object replay helpers to "push" mode Starting with commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"), DSA has introduced some bridge helpers that replay switchdev events (FDB/MDB/VLAN additions and deletions) that can be lost by the switchdev drivers in a variety of circumstances: - an IP multicast group was host-joined on the bridge itself before any switchdev port joined the bridge, leading to the host MDB entries missing in the hardware database. - during the bridge creation process, the MAC address of the bridge was added to the FDB as an entry pointing towards the bridge device itself, but with no switchdev ports being part of the bridge yet, this local FDB entry would remain unknown to the switchdev hardware database. - a VLAN/FDB/MDB was added to a bridge port that is a LAG interface, before any switchdev port joined that LAG, leading to the hardware database missing those entries. - a switchdev port left a LAG that is a bridge port, while the LAG remained part of the bridge, and all FDB/MDB/VLAN entries remained installed in the hardware database of the switchdev port. Also, since commit 0d2cfbd41c4a ("net: bridge: ignore switchdev events for LAG ports which didn't request replay"), DSA introduced a method, based on a const void *ctx, to ensure that two switchdev ports under the same LAG that is a bridge port do not see the same MDB/VLAN entry being replayed twice by the bridge, once for every bridge port that joins the LAG. With so many ordering corner cases being possible, it seems unreasonable to expect a switchdev driver writer to get it right from the first try. Therefore, now that DSA has experimented with the bridge replay helpers for a little bit, we can move the code to the bridge driver where it is more readily available to all switchdev drivers. To convert the switchdev object replay helpers from "pull mode" (where the driver asks for them) to a "push mode" (where the bridge offers them automatically), the biggest problem is that the bridge needs to be aware when a switchdev port joins and leaves, even when the switchdev is only indirectly a bridge port (for example when the bridge port is a LAG upper of the switchdev). Luckily, we already have a hook for that, in the form of the newly introduced switchdev_bridge_port_offload() and switchdev_bridge_port_unoffload() calls. These offer a natural place for hooking the object addition and deletion replays. Extend the above 2 functions with: - pointers to the switchdev atomic notifier (for FDB replays) and the blocking notifier (for MDB and VLAN replays). - the "const void *ctx" argument required for drivers to be able to disambiguate between which port is targeted, when multiple ports are lowers of the same LAG that is a bridge port. Most of the drivers pass NULL to this argument, except the ones that support LAG offload and have the proper context check already in place in the switchdev blocking notifier handler. Also unexport the replay helpers, since nobody except the bridge calls them directly now. Note that: (a) we abuse the terminology slightly, because FDB entries are not "switchdev objects", but we count them as objects nonetheless. With no direct way to prove it, I think they are not modeled as switchdev objects because those can only be installed by the bridge to the hardware (as opposed to FDB entries which can be propagated in the other direction too). This is merely an abuse of terms, FDB entries are replayed too, despite not being objects. (b) the bridge does not attempt to sync port attributes to newly joined ports, just the countable stuff (the objects). The reason for this is simple: no universal and symmetric way to sync and unsync them is known. For example, VLAN filtering: what to do on unsync, disable or leave it enabled? Similarly, STP state, ageing timer, etc etc. What a switchdev port does when it becomes standalone again is not really up to the bridge's competence, and the driver should deal with it. On the other hand, replaying deletions of switchdev objects can be seen a matter of cleanup and therefore be treated by the bridge, hence this patch. We make the replay helpers opt-in for drivers, because they might not bring immediate benefits for them: - nbp_vlan_init() is called _after_ netdev_master_upper_dev_link(), so br_vlan_replay() should not do anything for the new drivers on which we call it. The existing drivers where there was even a slight possibility for there to exist a VLAN on a bridge port before they join it are already guarded against this: mlxsw and prestera deny joining LAG interfaces that are members of a bridge. - br_fdb_replay() should now notify of local FDB entries, but I patched all drivers except DSA to ignore these new entries in commit 2c4eca3ef716 ("net: bridge: switchdev: include local flag in FDB notifications"). Driver authors can lift this restriction as they wish, and when they do, they can also opt into the FDB replay functionality. - br_mdb_replay() should fix a real issue which is described in commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"). However most drivers do not offload the SWITCHDEV_OBJ_ID_HOST_MDB to see this issue: only cpsw and am65_cpsw offload this switchdev object, and I don't completely understand the way in which they offload this switchdev object anyway. So I'll leave it up to these drivers' respective maintainers to opt into br_mdb_replay(). So most of the drivers pass NULL notifier blocks for the replay helpers, except: - dpaa2-switch which was already acked/regression-tested with the helpers enabled (and there isn't much of a downside in having them) - ocelot which already had replay logic in "pull" mode - DSA which already had replay logic in "pull" mode An important observation is that the drivers which don't currently request bridge event replays don't even have the switchdev_bridge_port_{offload,unoffload} calls placed in proper places right now. This was done to avoid unnecessary rework for drivers which might never even add support for this. For driver writers who wish to add replay support, this can be used as a tentative placement guide: https://patchwork.kernel.org/project/netdevbpf/patch/20210720134655.892334-11-vladimir.oltean@nxp.com/ Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:03 +00:00
void dsa_port_pre_bridge_leave(struct dsa_port *dp, struct net_device *br)
{
net: bridge: switchdev: let drivers inform which bridge ports are offloaded On reception of an skb, the bridge checks if it was marked as 'already forwarded in hardware' (checks if skb->offload_fwd_mark == 1), and if it is, it assigns the source hardware domain of that skb based on the hardware domain of the ingress port. Then during forwarding, it enforces that the egress port must have a different hardware domain than the ingress one (this is done in nbp_switchdev_allowed_egress). Non-switchdev drivers don't report any physical switch id (neither through devlink nor .ndo_get_port_parent_id), therefore the bridge assigns them a hardware domain of 0, and packets coming from them will always have skb->offload_fwd_mark = 0. So there aren't any restrictions. Problems appear due to the fact that DSA would like to perform software fallback for bonding and team interfaces that the physical switch cannot offload. +-- br0 ---+ / / | \ / / | \ / | | bond0 / | | / \ swp0 swp1 swp2 swp3 swp4 There, it is desirable that the presence of swp3 and swp4 under a non-offloaded LAG does not preclude us from doing hardware bridging beteen swp0, swp1 and swp2. The bandwidth of the CPU is often times high enough that software bridging between {swp0,swp1,swp2} and bond0 is not impractical. But this creates an impossible paradox given the current way in which port hardware domains are assigned. When the driver receives a packet from swp0 (say, due to flooding), it must set skb->offload_fwd_mark to something. - If we set it to 0, then the bridge will forward it towards swp1, swp2 and bond0. But the switch has already forwarded it towards swp1 and swp2 (not to bond0, remember, that isn't offloaded, so as far as the switch is concerned, ports swp3 and swp4 are not looking up the FDB, and the entire bond0 is a destination that is strictly behind the CPU). But we don't want duplicated traffic towards swp1 and swp2, so it's not ok to set skb->offload_fwd_mark = 0. - If we set it to 1, then the bridge will not forward the skb towards the ports with the same switchdev mark, i.e. not to swp1, swp2 and bond0. Towards swp1 and swp2 that's ok, but towards bond0? It should have forwarded the skb there. So the real issue is that bond0 will be assigned the same hardware domain as {swp0,swp1,swp2}, because the function that assigns hardware domains to bridge ports, nbp_switchdev_add(), recurses through bond0's lower interfaces until it finds something that implements devlink (calls dev_get_port_parent_id with bool recurse = true). This is a problem because the fact that bond0 can be offloaded by swp3 and swp4 in our example is merely an assumption. A solution is to give the bridge explicit hints as to what hardware domain it should use for each port. Currently, the bridging offload is very 'silent': a driver registers a netdevice notifier, which is put on the netns's notifier chain, and which sniffs around for NETDEV_CHANGEUPPER events where the upper is a bridge, and the lower is an interface it knows about (one registered by this driver, normally). Then, from within that notifier, it does a bunch of stuff behind the bridge's back, without the bridge necessarily knowing that there's somebody offloading that port. It looks like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v call_netdevice_notifiers | v dsa_slave_netdevice_event | v oh, hey! it's for me! | v .port_bridge_join What we do to solve the conundrum is to be less silent, and change the switchdev drivers to present themselves to the bridge. Something like this: ip link set swp0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | hardware domain for v | this port, and zero dsa_slave_netdevice_event | if I got nothing. | | v | oh, hey! it's for me! | | | v | .port_bridge_join | | | +------------------------+ switchdev_bridge_port_offload(swp0, swp0) Then stacked interfaces (like bond0 on top of swp3/swp4) would be treated differently in DSA, depending on whether we can or cannot offload them. The offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge: Aye! I'll use this call_netdevice_notifiers ^ ppid as the | | switchdev mark for v | bond0. dsa_slave_netdevice_event | Coincidentally (or not), | | bond0 and swp0, swp1, swp2 v | all have the same switchdev hmm, it's not quite for me, | mark now, since the ASIC but my driver has already | is able to forward towards called .port_lag_join | all these ports in hw. for it, because I have | a port with dp->lag_dev == bond0. | | | v | .port_bridge_join | for swp3 and swp4 | | | +------------------------+ switchdev_bridge_port_offload(bond0, swp3) switchdev_bridge_port_offload(bond0, swp4) And the non-offload case: ip link set bond0 master br0 | v br_add_if() calls netdev_master_upper_dev_link() | v bridge waiting: call_netdevice_notifiers ^ huh, switchdev_bridge_port_offload | | wasn't called, okay, I'll use a v | hwdom of zero for this one. dsa_slave_netdevice_event : Then packets received on swp0 will | : not be software-forwarded towards v : swp1, but they will towards bond0. it's not for me, but bond0 is an upper of swp3 and swp4, but their dp->lag_dev is NULL because they couldn't offload it. Basically we can draw the conclusion that the lowers of a bridge port can come and go, so depending on the configuration of lowers for a bridge port, it can dynamically toggle between offloaded and unoffloaded. Therefore, we need an equivalent switchdev_bridge_port_unoffload too. This patch changes the way any switchdev driver interacts with the bridge. From now on, everybody needs to call switchdev_bridge_port_offload and switchdev_bridge_port_unoffload, otherwise the bridge will treat the port as non-offloaded and allow software flooding to other ports from the same ASIC. Note that these functions lay the ground for a more complex handshake between switchdev drivers and the bridge in the future. For drivers that will request a replay of the switchdev objects when they offload and unoffload a bridge port (DSA, dpaa2-switch, ocelot), we place the call to switchdev_bridge_port_unoffload() strategically inside the NETDEV_PRECHANGEUPPER notifier's code path, and not inside NETDEV_CHANGEUPPER. This is because the switchdev object replay helpers need the netdev adjacency lists to be valid, and that is only true in NETDEV_PRECHANGEUPPER. Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch: regression Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Tested-by: Horatiu Vultur <horatiu.vultur@microchip.com> # ocelot-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:01 +00:00
struct net_device *brport_dev = dsa_port_to_bridge_port(dp);
/* Don't try to unoffload something that is not offloaded */
if (!brport_dev)
return;
net: bridge: move the switchdev object replay helpers to "push" mode Starting with commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"), DSA has introduced some bridge helpers that replay switchdev events (FDB/MDB/VLAN additions and deletions) that can be lost by the switchdev drivers in a variety of circumstances: - an IP multicast group was host-joined on the bridge itself before any switchdev port joined the bridge, leading to the host MDB entries missing in the hardware database. - during the bridge creation process, the MAC address of the bridge was added to the FDB as an entry pointing towards the bridge device itself, but with no switchdev ports being part of the bridge yet, this local FDB entry would remain unknown to the switchdev hardware database. - a VLAN/FDB/MDB was added to a bridge port that is a LAG interface, before any switchdev port joined that LAG, leading to the hardware database missing those entries. - a switchdev port left a LAG that is a bridge port, while the LAG remained part of the bridge, and all FDB/MDB/VLAN entries remained installed in the hardware database of the switchdev port. Also, since commit 0d2cfbd41c4a ("net: bridge: ignore switchdev events for LAG ports which didn't request replay"), DSA introduced a method, based on a const void *ctx, to ensure that two switchdev ports under the same LAG that is a bridge port do not see the same MDB/VLAN entry being replayed twice by the bridge, once for every bridge port that joins the LAG. With so many ordering corner cases being possible, it seems unreasonable to expect a switchdev driver writer to get it right from the first try. Therefore, now that DSA has experimented with the bridge replay helpers for a little bit, we can move the code to the bridge driver where it is more readily available to all switchdev drivers. To convert the switchdev object replay helpers from "pull mode" (where the driver asks for them) to a "push mode" (where the bridge offers them automatically), the biggest problem is that the bridge needs to be aware when a switchdev port joins and leaves, even when the switchdev is only indirectly a bridge port (for example when the bridge port is a LAG upper of the switchdev). Luckily, we already have a hook for that, in the form of the newly introduced switchdev_bridge_port_offload() and switchdev_bridge_port_unoffload() calls. These offer a natural place for hooking the object addition and deletion replays. Extend the above 2 functions with: - pointers to the switchdev atomic notifier (for FDB replays) and the blocking notifier (for MDB and VLAN replays). - the "const void *ctx" argument required for drivers to be able to disambiguate between which port is targeted, when multiple ports are lowers of the same LAG that is a bridge port. Most of the drivers pass NULL to this argument, except the ones that support LAG offload and have the proper context check already in place in the switchdev blocking notifier handler. Also unexport the replay helpers, since nobody except the bridge calls them directly now. Note that: (a) we abuse the terminology slightly, because FDB entries are not "switchdev objects", but we count them as objects nonetheless. With no direct way to prove it, I think they are not modeled as switchdev objects because those can only be installed by the bridge to the hardware (as opposed to FDB entries which can be propagated in the other direction too). This is merely an abuse of terms, FDB entries are replayed too, despite not being objects. (b) the bridge does not attempt to sync port attributes to newly joined ports, just the countable stuff (the objects). The reason for this is simple: no universal and symmetric way to sync and unsync them is known. For example, VLAN filtering: what to do on unsync, disable or leave it enabled? Similarly, STP state, ageing timer, etc etc. What a switchdev port does when it becomes standalone again is not really up to the bridge's competence, and the driver should deal with it. On the other hand, replaying deletions of switchdev objects can be seen a matter of cleanup and therefore be treated by the bridge, hence this patch. We make the replay helpers opt-in for drivers, because they might not bring immediate benefits for them: - nbp_vlan_init() is called _after_ netdev_master_upper_dev_link(), so br_vlan_replay() should not do anything for the new drivers on which we call it. The existing drivers where there was even a slight possibility for there to exist a VLAN on a bridge port before they join it are already guarded against this: mlxsw and prestera deny joining LAG interfaces that are members of a bridge. - br_fdb_replay() should now notify of local FDB entries, but I patched all drivers except DSA to ignore these new entries in commit 2c4eca3ef716 ("net: bridge: switchdev: include local flag in FDB notifications"). Driver authors can lift this restriction as they wish, and when they do, they can also opt into the FDB replay functionality. - br_mdb_replay() should fix a real issue which is described in commit 4f2673b3a2b6 ("net: bridge: add helper to replay port and host-joined mdb entries"). However most drivers do not offload the SWITCHDEV_OBJ_ID_HOST_MDB to see this issue: only cpsw and am65_cpsw offload this switchdev object, and I don't completely understand the way in which they offload this switchdev object anyway. So I'll leave it up to these drivers' respective maintainers to opt into br_mdb_replay(). So most of the drivers pass NULL notifier blocks for the replay helpers, except: - dpaa2-switch which was already acked/regression-tested with the helpers enabled (and there isn't much of a downside in having them) - ocelot which already had replay logic in "pull" mode - DSA which already had replay logic in "pull" mode An important observation is that the drivers which don't currently request bridge event replays don't even have the switchdev_bridge_port_{offload,unoffload} calls placed in proper places right now. This was done to avoid unnecessary rework for drivers which might never even add support for this. For driver writers who wish to add replay support, this can be used as a tentative placement guide: https://patchwork.kernel.org/project/netdevbpf/patch/20210720134655.892334-11-vladimir.oltean@nxp.com/ Cc: Vadym Kochan <vkochan@marvell.com> Cc: Taras Chornyi <tchornyi@marvell.com> Cc: Ioana Ciornei <ioana.ciornei@nxp.com> Cc: Lars Povlsen <lars.povlsen@microchip.com> Cc: Steen Hegelund <Steen.Hegelund@microchip.com> Cc: UNGLinuxDriver@microchip.com Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Ioana Ciornei <ioana.ciornei@nxp.com> # dpaa2-switch Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-21 16:24:03 +00:00
switchdev_bridge_port_unoffload(brport_dev, dp,
&dsa_user_switchdev_notifier,
&dsa_user_switchdev_blocking_notifier);
net: dsa: flush switchdev workqueue when leaving the bridge DSA is preparing to offer switch drivers an API through which they can associate each FDB entry with a struct net_device *bridge_dev. This can be used to perform FDB isolation (the FDB lookup performed on the ingress of a standalone, or bridged port, should not find an FDB entry that is present in the FDB of another bridge). In preparation of that work, DSA needs to ensure that by the time we call the switch .port_fdb_add and .port_fdb_del methods, the dp->bridge_dev pointer is still valid, i.e. the port is still a bridge port. This is not guaranteed because the SWITCHDEV_FDB_{ADD,DEL}_TO_DEVICE API requires drivers that must have sleepable context to handle those events to schedule the deferred work themselves. DSA does this through the dsa_owq. It can happen that a port leaves a bridge, del_nbp() flushes the FDB on that port, SWITCHDEV_FDB_DEL_TO_DEVICE is notified in atomic context, DSA schedules its deferred work, but del_nbp() finishes unlinking the bridge as a master from the port before DSA's deferred work is run. Fundamentally, the port must not be unlinked from the bridge until all FDB deletion deferred work items have been flushed. The bridge must wait for the completion of these hardware accesses. An attempt has been made to address this issue centrally in switchdev by making SWITCHDEV_FDB_DEL_TO_DEVICE deferred (=> blocking) at the switchdev level, which would offer implicit synchronization with del_nbp: https://patchwork.kernel.org/project/netdevbpf/cover/20210820115746.3701811-1-vladimir.oltean@nxp.com/ but it seems that any attempt to modify switchdev's behavior and make the events blocking there would introduce undesirable side effects in other switchdev consumers. The most undesirable behavior seems to be that switchdev_deferred_process_work() takes the rtnl_mutex itself, which would be worse off than having the rtnl_mutex taken individually from drivers which is what we have now (except DSA which has removed that lock since commit 0faf890fc519 ("net: dsa: drop rtnl_lock from dsa_slave_switchdev_event_work")). So to offer the needed guarantee to DSA switch drivers, I have come up with a compromise solution that does not require switchdev rework: we already have a hook at the last moment in time when the bridge is still an upper of ours: the NETDEV_PRECHANGEUPPER handler. We can flush the dsa_owq manually from there, which makes all FDB deletions synchronous. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-10-26 09:25:55 +00:00
dsa_flush_workqueue();
}
void dsa_port_bridge_leave(struct dsa_port *dp, struct net_device *br)
{
struct dsa_notifier_bridge_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
};
int err;
net: dsa: fix panic when removing unoffloaded port from bridge If a bridged port is not offloaded to the hardware - either because the underlying driver does not implement the port_bridge_{join,leave} ops, or because the operation failed - then its dp->bridge pointer will be NULL when dsa_port_bridge_leave() is called. Avoid dereferncing NULL. This fixes the following splat when removing a port from a bridge: Unable to handle kernel access to user memory outside uaccess routines at virtual address 0000000000000000 Internal error: Oops: 96000004 [#1] PREEMPT_RT SMP CPU: 3 PID: 1119 Comm: brctl Tainted: G O 5.17.0-rc4-rt4 #1 Call trace: dsa_port_bridge_leave+0x8c/0x1e4 dsa_slave_changeupper+0x40/0x170 dsa_slave_netdevice_event+0x494/0x4d4 notifier_call_chain+0x80/0xe0 raw_notifier_call_chain+0x1c/0x24 call_netdevice_notifiers_info+0x5c/0xac __netdev_upper_dev_unlink+0xa4/0x200 netdev_upper_dev_unlink+0x38/0x60 del_nbp+0x1b0/0x300 br_del_if+0x38/0x114 add_del_if+0x60/0xa0 br_ioctl_stub+0x128/0x2dc br_ioctl_call+0x68/0xb0 dev_ifsioc+0x390/0x554 dev_ioctl+0x128/0x400 sock_do_ioctl+0xb4/0xf4 sock_ioctl+0x12c/0x4e0 __arm64_sys_ioctl+0xa8/0xf0 invoke_syscall+0x4c/0x110 el0_svc_common.constprop.0+0x48/0xf0 do_el0_svc+0x28/0x84 el0_svc+0x1c/0x50 el0t_64_sync_handler+0xa8/0xb0 el0t_64_sync+0x17c/0x180 Code: f9402f00 f0002261 f9401302 913cc021 (a9401404) ---[ end trace 0000000000000000 ]--- Fixes: d3eed0e57d5d ("net: dsa: keep the bridge_dev and bridge_num as part of the same structure") Signed-off-by: Alvin Šipraga <alsi@bang-olufsen.dk> Reviewed-by: Vladimir Oltean <olteanv@gmail.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Link: https://lore.kernel.org/r/20220221203539.310690-1-alvin@pqrs.dk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-21 20:35:38 +00:00
/* If the port could not be offloaded to begin with, then
* there is nothing to do.
*/
if (!dp->bridge)
return;
info.bridge = *dp->bridge;
/* Here the port is already unbridged. Reflect the current configuration
* so that drivers can program their chips accordingly.
*/
dsa_port_bridge_destroy(dp, br);
net: dsa: permit cross-chip bridging between all trees in the system One way of utilizing DSA is by cascading switches which do not all have compatible taggers. Consider the following real-life topology: +---------------------------------------------------------------+ | LS1028A | | +------------------------------+ | | | DSA master for Felix | | | |(internal ENETC port 2: eno2))| | | +------------+------------------------------+-------------+ | | | Felix embedded L2 switch | | | | | | | | +--------------+ +--------------+ +--------------+ | | | | |DSA master for| |DSA master for| |DSA master for| | | | | | SJA1105 1 | | SJA1105 2 | | SJA1105 3 | | | | | |(Felix port 1)| |(Felix port 2)| |(Felix port 3)| | | +--+-+--------------+---+--------------+---+--------------+--+--+ +-----------------------+ +-----------------------+ +-----------------------+ | SJA1105 switch 1 | | SJA1105 switch 2 | | SJA1105 switch 3 | +-----+-----+-----+-----+ +-----+-----+-----+-----+ +-----+-----+-----+-----+ |sw1p0|sw1p1|sw1p2|sw1p3| |sw2p0|sw2p1|sw2p2|sw2p3| |sw3p0|sw3p1|sw3p2|sw3p3| +-----+-----+-----+-----+ +-----+-----+-----+-----+ +-----+-----+-----+-----+ The above can be described in the device tree as follows (obviously not complete): mscc_felix { dsa,member = <0 0>; ports { port@4 { ethernet = <&enetc_port2>; }; }; }; sja1105_switch1 { dsa,member = <1 1>; ports { port@4 { ethernet = <&mscc_felix_port1>; }; }; }; sja1105_switch2 { dsa,member = <2 2>; ports { port@4 { ethernet = <&mscc_felix_port2>; }; }; }; sja1105_switch3 { dsa,member = <3 3>; ports { port@4 { ethernet = <&mscc_felix_port3>; }; }; }; Basically we instantiate one DSA switch tree for every hardware switch in the system, but we still give them globally unique switch IDs (will come back to that later). Having 3 disjoint switch trees makes the tagger drivers "just work", because net devices are registered for the 3 Felix DSA master ports, and they are also DSA slave ports to the ENETC port. So packets received on the ENETC port are stripped of their stacked DSA tags one by one. Currently, hardware bridging between ports on the same sja1105 chip is possible, but switching between sja1105 ports on different chips is handled by the software bridge. This is fine, but we can do better. In fact, the dsa_8021q tag used by sja1105 is compatible with cascading. In other words, a sja1105 switch can correctly parse and route a packet containing a dsa_8021q tag. So if we could enable hardware bridging on the Felix DSA master ports, cross-chip bridging could be completely offloaded. Such as system would be used as follows: ip link add dev br0 type bridge && ip link set dev br0 up for port in sw0p0 sw0p1 sw0p2 sw0p3 \ sw1p0 sw1p1 sw1p2 sw1p3 \ sw2p0 sw2p1 sw2p2 sw2p3; do ip link set dev $port master br0 done The above makes switching between ports on the same row be performed in hardware, and between ports on different rows in software. Now assume the Felix switch ports are called swp0, swp1, swp2. By running the following extra commands: ip link add dev br1 type bridge && ip link set dev br1 up for port in swp0 swp1 swp2; do ip link set dev $port master br1 done the CPU no longer sees packets which traverse sja1105 switch boundaries and can be forwarded directly by Felix. The br1 bridge would not be used for any sort of traffic termination. For this to work, we need to give drivers an opportunity to listen for bridging events on DSA trees other than their own, and pass that other tree index as argument. I have made the assumption, for the moment, that the other existing DSA notifiers don't need to be broadcast to other trees. That assumption might turn out to be incorrect. But in the meantime, introduce a dsa_broadcast function, similar in purpose to dsa_port_notify, which is used only by the bridging notifiers. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2020-05-10 16:37:41 +00:00
err = dsa_broadcast(DSA_NOTIFIER_BRIDGE_LEAVE, &info);
if (err)
dev_err(dp->ds->dev,
"port %d failed to notify DSA_NOTIFIER_BRIDGE_LEAVE: %pe\n",
dp->index, ERR_PTR(err));
dsa_port_switchdev_unsync_attrs(dp, info.bridge);
}
int dsa_port_lag_change(struct dsa_port *dp,
struct netdev_lag_lower_state_info *linfo)
{
struct dsa_notifier_lag_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
};
bool tx_enabled;
net: dsa: create a dsa_lag structure The main purpose of this change is to create a data structure for a LAG as seen by DSA. This is similar to what we have for bridging - we pass a copy of this structure by value to ->port_lag_join and ->port_lag_leave. For now we keep the lag_dev, id and a reference count in it. Future patches will add a list of FDB entries for the LAG (these also need to be refcounted to work properly). The LAG structure is created using dsa_port_lag_create() and destroyed using dsa_port_lag_destroy(), just like we have for bridging. Because now, the dsa_lag itself is refcounted, we can simplify dsa_lag_map() and dsa_lag_unmap(). These functions need to keep a LAG in the dst->lags array only as long as at least one port uses it. The refcounting logic inside those functions can be removed now - they are called only when we should perform the operation. dsa_lag_dev() is renamed to dsa_lag_by_id() and now returns the dsa_lag structure instead of the lag_dev net_device. dsa_lag_foreach_port() now takes the dsa_lag structure as argument. dst->lags holds an array of dsa_lag structures. dsa_lag_map() now also saves the dsa_lag->id value, so that linear walking of dst->lags in drivers using dsa_lag_id() is no longer necessary. They can just look at lag.id. dsa_port_lag_id_get() is a helper, similar to dsa_port_bridge_num_get(), which can be used by drivers to get the LAG ID assigned by DSA to a given port. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:49 +00:00
if (!dp->lag)
return 0;
/* On statically configured aggregates (e.g. loadbalance
* without LACP) ports will always be tx_enabled, even if the
* link is down. Thus we require both link_up and tx_enabled
* in order to include it in the tx set.
*/
tx_enabled = linfo->link_up && linfo->tx_enabled;
if (tx_enabled == dp->lag_tx_enabled)
return 0;
dp->lag_tx_enabled = tx_enabled;
return dsa_port_notify(dp, DSA_NOTIFIER_LAG_CHANGE, &info);
}
net: dsa: create a dsa_lag structure The main purpose of this change is to create a data structure for a LAG as seen by DSA. This is similar to what we have for bridging - we pass a copy of this structure by value to ->port_lag_join and ->port_lag_leave. For now we keep the lag_dev, id and a reference count in it. Future patches will add a list of FDB entries for the LAG (these also need to be refcounted to work properly). The LAG structure is created using dsa_port_lag_create() and destroyed using dsa_port_lag_destroy(), just like we have for bridging. Because now, the dsa_lag itself is refcounted, we can simplify dsa_lag_map() and dsa_lag_unmap(). These functions need to keep a LAG in the dst->lags array only as long as at least one port uses it. The refcounting logic inside those functions can be removed now - they are called only when we should perform the operation. dsa_lag_dev() is renamed to dsa_lag_by_id() and now returns the dsa_lag structure instead of the lag_dev net_device. dsa_lag_foreach_port() now takes the dsa_lag structure as argument. dst->lags holds an array of dsa_lag structures. dsa_lag_map() now also saves the dsa_lag->id value, so that linear walking of dst->lags in drivers using dsa_lag_id() is no longer necessary. They can just look at lag.id. dsa_port_lag_id_get() is a helper, similar to dsa_port_bridge_num_get(), which can be used by drivers to get the LAG ID assigned by DSA to a given port. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:49 +00:00
static int dsa_port_lag_create(struct dsa_port *dp,
struct net_device *lag_dev)
{
struct dsa_switch *ds = dp->ds;
struct dsa_lag *lag;
lag = dsa_tree_lag_find(ds->dst, lag_dev);
if (lag) {
refcount_inc(&lag->refcount);
dp->lag = lag;
return 0;
}
lag = kzalloc(sizeof(*lag), GFP_KERNEL);
if (!lag)
return -ENOMEM;
refcount_set(&lag->refcount, 1);
net: dsa: support FDB events on offloaded LAG interfaces This change introduces support for installing static FDB entries towards a bridge port that is a LAG of multiple DSA switch ports, as well as support for filtering towards the CPU local FDB entries emitted for LAG interfaces that are bridge ports. Conceptually, host addresses on LAG ports are identical to what we do for plain bridge ports. Whereas FDB entries _towards_ a LAG can't simply be replicated towards all member ports like we do for multicast, or VLAN. Instead we need new driver API. Hardware usually considers a LAG to be a "logical port", and sets the entire LAG as the forwarding destination. The physical egress port selection within the LAG is made by hashing policy, as usual. To represent the logical port corresponding to the LAG, we pass by value a copy of the dsa_lag structure to all switches in the tree that have at least one port in that LAG. To illustrate why a refcounted list of FDB entries is needed in struct dsa_lag, it is enough to say that: - a LAG may be a bridge port and may therefore receive FDB events even while it isn't yet offloaded by any DSA interface - DSA interfaces may be removed from a LAG while that is a bridge port; we don't want FDB entries lingering around, but we don't want to remove entries that are still in use, either For all the cases below to work, the idea is to always keep an FDB entry on a LAG with a reference count equal to the DSA member ports. So: - if a port joins a LAG, it requests the bridge to replay the FDB, and the FDB entries get created, or their refcount gets bumped by one - if a port leaves a LAG, the FDB replay deletes or decrements refcount by one - if an FDB is installed towards a LAG with ports already present, that entry is created (if it doesn't exist) and its refcount is bumped by the amount of ports already present in the LAG echo "Adding FDB entry to bond with existing ports" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond, then removing ports one by one" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link set swp1 nomaster ip link set swp2 nomaster ip link del br0 ip link del bond0 Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:53 +00:00
mutex_init(&lag->fdb_lock);
INIT_LIST_HEAD(&lag->fdbs);
net: dsa: create a dsa_lag structure The main purpose of this change is to create a data structure for a LAG as seen by DSA. This is similar to what we have for bridging - we pass a copy of this structure by value to ->port_lag_join and ->port_lag_leave. For now we keep the lag_dev, id and a reference count in it. Future patches will add a list of FDB entries for the LAG (these also need to be refcounted to work properly). The LAG structure is created using dsa_port_lag_create() and destroyed using dsa_port_lag_destroy(), just like we have for bridging. Because now, the dsa_lag itself is refcounted, we can simplify dsa_lag_map() and dsa_lag_unmap(). These functions need to keep a LAG in the dst->lags array only as long as at least one port uses it. The refcounting logic inside those functions can be removed now - they are called only when we should perform the operation. dsa_lag_dev() is renamed to dsa_lag_by_id() and now returns the dsa_lag structure instead of the lag_dev net_device. dsa_lag_foreach_port() now takes the dsa_lag structure as argument. dst->lags holds an array of dsa_lag structures. dsa_lag_map() now also saves the dsa_lag->id value, so that linear walking of dst->lags in drivers using dsa_lag_id() is no longer necessary. They can just look at lag.id. dsa_port_lag_id_get() is a helper, similar to dsa_port_bridge_num_get(), which can be used by drivers to get the LAG ID assigned by DSA to a given port. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:49 +00:00
lag->dev = lag_dev;
dsa_lag_map(ds->dst, lag);
dp->lag = lag;
return 0;
}
static void dsa_port_lag_destroy(struct dsa_port *dp)
{
struct dsa_lag *lag = dp->lag;
dp->lag = NULL;
dp->lag_tx_enabled = false;
if (!refcount_dec_and_test(&lag->refcount))
return;
net: dsa: support FDB events on offloaded LAG interfaces This change introduces support for installing static FDB entries towards a bridge port that is a LAG of multiple DSA switch ports, as well as support for filtering towards the CPU local FDB entries emitted for LAG interfaces that are bridge ports. Conceptually, host addresses on LAG ports are identical to what we do for plain bridge ports. Whereas FDB entries _towards_ a LAG can't simply be replicated towards all member ports like we do for multicast, or VLAN. Instead we need new driver API. Hardware usually considers a LAG to be a "logical port", and sets the entire LAG as the forwarding destination. The physical egress port selection within the LAG is made by hashing policy, as usual. To represent the logical port corresponding to the LAG, we pass by value a copy of the dsa_lag structure to all switches in the tree that have at least one port in that LAG. To illustrate why a refcounted list of FDB entries is needed in struct dsa_lag, it is enough to say that: - a LAG may be a bridge port and may therefore receive FDB events even while it isn't yet offloaded by any DSA interface - DSA interfaces may be removed from a LAG while that is a bridge port; we don't want FDB entries lingering around, but we don't want to remove entries that are still in use, either For all the cases below to work, the idea is to always keep an FDB entry on a LAG with a reference count equal to the DSA member ports. So: - if a port joins a LAG, it requests the bridge to replay the FDB, and the FDB entries get created, or their refcount gets bumped by one - if a port leaves a LAG, the FDB replay deletes or decrements refcount by one - if an FDB is installed towards a LAG with ports already present, that entry is created (if it doesn't exist) and its refcount is bumped by the amount of ports already present in the LAG echo "Adding FDB entry to bond with existing ports" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond, then removing ports one by one" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link set swp1 nomaster ip link set swp2 nomaster ip link del br0 ip link del bond0 Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:53 +00:00
WARN_ON(!list_empty(&lag->fdbs));
net: dsa: create a dsa_lag structure The main purpose of this change is to create a data structure for a LAG as seen by DSA. This is similar to what we have for bridging - we pass a copy of this structure by value to ->port_lag_join and ->port_lag_leave. For now we keep the lag_dev, id and a reference count in it. Future patches will add a list of FDB entries for the LAG (these also need to be refcounted to work properly). The LAG structure is created using dsa_port_lag_create() and destroyed using dsa_port_lag_destroy(), just like we have for bridging. Because now, the dsa_lag itself is refcounted, we can simplify dsa_lag_map() and dsa_lag_unmap(). These functions need to keep a LAG in the dst->lags array only as long as at least one port uses it. The refcounting logic inside those functions can be removed now - they are called only when we should perform the operation. dsa_lag_dev() is renamed to dsa_lag_by_id() and now returns the dsa_lag structure instead of the lag_dev net_device. dsa_lag_foreach_port() now takes the dsa_lag structure as argument. dst->lags holds an array of dsa_lag structures. dsa_lag_map() now also saves the dsa_lag->id value, so that linear walking of dst->lags in drivers using dsa_lag_id() is no longer necessary. They can just look at lag.id. dsa_port_lag_id_get() is a helper, similar to dsa_port_bridge_num_get(), which can be used by drivers to get the LAG ID assigned by DSA to a given port. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:49 +00:00
dsa_lag_unmap(dp->ds->dst, lag);
kfree(lag);
}
int dsa_port_lag_join(struct dsa_port *dp, struct net_device *lag_dev,
struct netdev_lag_upper_info *uinfo,
struct netlink_ext_ack *extack)
{
struct dsa_notifier_lag_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.info = uinfo,
.extack = extack,
};
struct net_device *bridge_dev;
int err;
net: dsa: create a dsa_lag structure The main purpose of this change is to create a data structure for a LAG as seen by DSA. This is similar to what we have for bridging - we pass a copy of this structure by value to ->port_lag_join and ->port_lag_leave. For now we keep the lag_dev, id and a reference count in it. Future patches will add a list of FDB entries for the LAG (these also need to be refcounted to work properly). The LAG structure is created using dsa_port_lag_create() and destroyed using dsa_port_lag_destroy(), just like we have for bridging. Because now, the dsa_lag itself is refcounted, we can simplify dsa_lag_map() and dsa_lag_unmap(). These functions need to keep a LAG in the dst->lags array only as long as at least one port uses it. The refcounting logic inside those functions can be removed now - they are called only when we should perform the operation. dsa_lag_dev() is renamed to dsa_lag_by_id() and now returns the dsa_lag structure instead of the lag_dev net_device. dsa_lag_foreach_port() now takes the dsa_lag structure as argument. dst->lags holds an array of dsa_lag structures. dsa_lag_map() now also saves the dsa_lag->id value, so that linear walking of dst->lags in drivers using dsa_lag_id() is no longer necessary. They can just look at lag.id. dsa_port_lag_id_get() is a helper, similar to dsa_port_bridge_num_get(), which can be used by drivers to get the LAG ID assigned by DSA to a given port. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:49 +00:00
err = dsa_port_lag_create(dp, lag_dev);
if (err)
goto err_lag_create;
net: dsa: create a dsa_lag structure The main purpose of this change is to create a data structure for a LAG as seen by DSA. This is similar to what we have for bridging - we pass a copy of this structure by value to ->port_lag_join and ->port_lag_leave. For now we keep the lag_dev, id and a reference count in it. Future patches will add a list of FDB entries for the LAG (these also need to be refcounted to work properly). The LAG structure is created using dsa_port_lag_create() and destroyed using dsa_port_lag_destroy(), just like we have for bridging. Because now, the dsa_lag itself is refcounted, we can simplify dsa_lag_map() and dsa_lag_unmap(). These functions need to keep a LAG in the dst->lags array only as long as at least one port uses it. The refcounting logic inside those functions can be removed now - they are called only when we should perform the operation. dsa_lag_dev() is renamed to dsa_lag_by_id() and now returns the dsa_lag structure instead of the lag_dev net_device. dsa_lag_foreach_port() now takes the dsa_lag structure as argument. dst->lags holds an array of dsa_lag structures. dsa_lag_map() now also saves the dsa_lag->id value, so that linear walking of dst->lags in drivers using dsa_lag_id() is no longer necessary. They can just look at lag.id. dsa_port_lag_id_get() is a helper, similar to dsa_port_bridge_num_get(), which can be used by drivers to get the LAG ID assigned by DSA to a given port. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:49 +00:00
info.lag = *dp->lag;
err = dsa_port_notify(dp, DSA_NOTIFIER_LAG_JOIN, &info);
if (err)
goto err_lag_join;
bridge_dev = netdev_master_upper_dev_get(lag_dev);
if (!bridge_dev || !netif_is_bridge_master(bridge_dev))
return 0;
err = dsa_port_bridge_join(dp, bridge_dev, extack);
if (err)
goto err_bridge_join;
return 0;
err_bridge_join:
dsa_port_notify(dp, DSA_NOTIFIER_LAG_LEAVE, &info);
err_lag_join:
net: dsa: create a dsa_lag structure The main purpose of this change is to create a data structure for a LAG as seen by DSA. This is similar to what we have for bridging - we pass a copy of this structure by value to ->port_lag_join and ->port_lag_leave. For now we keep the lag_dev, id and a reference count in it. Future patches will add a list of FDB entries for the LAG (these also need to be refcounted to work properly). The LAG structure is created using dsa_port_lag_create() and destroyed using dsa_port_lag_destroy(), just like we have for bridging. Because now, the dsa_lag itself is refcounted, we can simplify dsa_lag_map() and dsa_lag_unmap(). These functions need to keep a LAG in the dst->lags array only as long as at least one port uses it. The refcounting logic inside those functions can be removed now - they are called only when we should perform the operation. dsa_lag_dev() is renamed to dsa_lag_by_id() and now returns the dsa_lag structure instead of the lag_dev net_device. dsa_lag_foreach_port() now takes the dsa_lag structure as argument. dst->lags holds an array of dsa_lag structures. dsa_lag_map() now also saves the dsa_lag->id value, so that linear walking of dst->lags in drivers using dsa_lag_id() is no longer necessary. They can just look at lag.id. dsa_port_lag_id_get() is a helper, similar to dsa_port_bridge_num_get(), which can be used by drivers to get the LAG ID assigned by DSA to a given port. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:49 +00:00
dsa_port_lag_destroy(dp);
err_lag_create:
return err;
}
void dsa_port_pre_lag_leave(struct dsa_port *dp, struct net_device *lag_dev)
{
struct net_device *br = dsa_port_bridge_dev_get(dp);
if (br)
dsa_port_pre_bridge_leave(dp, br);
}
void dsa_port_lag_leave(struct dsa_port *dp, struct net_device *lag_dev)
{
struct net_device *br = dsa_port_bridge_dev_get(dp);
struct dsa_notifier_lag_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
};
int err;
net: dsa: create a dsa_lag structure The main purpose of this change is to create a data structure for a LAG as seen by DSA. This is similar to what we have for bridging - we pass a copy of this structure by value to ->port_lag_join and ->port_lag_leave. For now we keep the lag_dev, id and a reference count in it. Future patches will add a list of FDB entries for the LAG (these also need to be refcounted to work properly). The LAG structure is created using dsa_port_lag_create() and destroyed using dsa_port_lag_destroy(), just like we have for bridging. Because now, the dsa_lag itself is refcounted, we can simplify dsa_lag_map() and dsa_lag_unmap(). These functions need to keep a LAG in the dst->lags array only as long as at least one port uses it. The refcounting logic inside those functions can be removed now - they are called only when we should perform the operation. dsa_lag_dev() is renamed to dsa_lag_by_id() and now returns the dsa_lag structure instead of the lag_dev net_device. dsa_lag_foreach_port() now takes the dsa_lag structure as argument. dst->lags holds an array of dsa_lag structures. dsa_lag_map() now also saves the dsa_lag->id value, so that linear walking of dst->lags in drivers using dsa_lag_id() is no longer necessary. They can just look at lag.id. dsa_port_lag_id_get() is a helper, similar to dsa_port_bridge_num_get(), which can be used by drivers to get the LAG ID assigned by DSA to a given port. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:49 +00:00
if (!dp->lag)
return;
/* Port might have been part of a LAG that in turn was
* attached to a bridge.
*/
if (br)
dsa_port_bridge_leave(dp, br);
net: dsa: create a dsa_lag structure The main purpose of this change is to create a data structure for a LAG as seen by DSA. This is similar to what we have for bridging - we pass a copy of this structure by value to ->port_lag_join and ->port_lag_leave. For now we keep the lag_dev, id and a reference count in it. Future patches will add a list of FDB entries for the LAG (these also need to be refcounted to work properly). The LAG structure is created using dsa_port_lag_create() and destroyed using dsa_port_lag_destroy(), just like we have for bridging. Because now, the dsa_lag itself is refcounted, we can simplify dsa_lag_map() and dsa_lag_unmap(). These functions need to keep a LAG in the dst->lags array only as long as at least one port uses it. The refcounting logic inside those functions can be removed now - they are called only when we should perform the operation. dsa_lag_dev() is renamed to dsa_lag_by_id() and now returns the dsa_lag structure instead of the lag_dev net_device. dsa_lag_foreach_port() now takes the dsa_lag structure as argument. dst->lags holds an array of dsa_lag structures. dsa_lag_map() now also saves the dsa_lag->id value, so that linear walking of dst->lags in drivers using dsa_lag_id() is no longer necessary. They can just look at lag.id. dsa_port_lag_id_get() is a helper, similar to dsa_port_bridge_num_get(), which can be used by drivers to get the LAG ID assigned by DSA to a given port. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:49 +00:00
info.lag = *dp->lag;
dsa_port_lag_destroy(dp);
err = dsa_port_notify(dp, DSA_NOTIFIER_LAG_LEAVE, &info);
if (err)
dev_err(dp->ds->dev,
"port %d failed to notify DSA_NOTIFIER_LAG_LEAVE: %pe\n",
dp->index, ERR_PTR(err));
}
/* Must be called under rcu_read_lock() */
static bool dsa_port_can_apply_vlan_filtering(struct dsa_port *dp,
bool vlan_filtering,
struct netlink_ext_ack *extack)
{
struct dsa_switch *ds = dp->ds;
struct dsa_port *other_dp;
int err;
/* VLAN awareness was off, so the question is "can we turn it on".
* We may have had 8021q uppers, those need to go. Make sure we don't
* enter an inconsistent state: deny changing the VLAN awareness state
* as long as we have 8021q uppers.
*/
if (vlan_filtering && dsa_port_is_user(dp)) {
struct net_device *br = dsa_port_bridge_dev_get(dp);
struct net_device *upper_dev, *user = dp->user;
struct list_head *iter;
netdev_for_each_upper_dev_rcu(user, upper_dev, iter) {
struct bridge_vlan_info br_info;
u16 vid;
if (!is_vlan_dev(upper_dev))
continue;
vid = vlan_dev_vlan_id(upper_dev);
/* br_vlan_get_info() returns -EINVAL or -ENOENT if the
* device, respectively the VID is not found, returning
* 0 means success, which is a failure for us here.
*/
err = br_vlan_get_info(br, vid, &br_info);
if (err == 0) {
NL_SET_ERR_MSG_MOD(extack,
"Must first remove VLAN uppers having VIDs also present in bridge");
return false;
}
}
}
if (!ds->vlan_filtering_is_global)
return true;
/* For cases where enabling/disabling VLAN awareness is global to the
* switch, we need to handle the case where multiple bridges span
* different ports of the same switch device and one of them has a
* different setting than what is being requested.
*/
dsa_switch_for_each_port(other_dp, ds) {
struct net_device *other_br = dsa_port_bridge_dev_get(other_dp);
/* If it's the same bridge, it also has same
* vlan_filtering setting => no need to check
*/
if (!other_br || other_br == dsa_port_bridge_dev_get(dp))
continue;
if (br_vlan_enabled(other_br) != vlan_filtering) {
NL_SET_ERR_MSG_MOD(extack,
"VLAN filtering is a global setting");
return false;
}
}
return true;
}
int dsa_port_vlan_filtering(struct dsa_port *dp, bool vlan_filtering,
struct netlink_ext_ack *extack)
{
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
bool old_vlan_filtering = dsa_port_is_vlan_filtering(dp);
struct dsa_switch *ds = dp->ds;
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
bool apply;
int err;
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
if (!ds->ops->port_vlan_filtering)
return -EOPNOTSUPP;
/* We are called from dsa_user_switchdev_blocking_event(),
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
* which is not under rcu_read_lock(), unlike
* dsa_user_switchdev_event().
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
*/
rcu_read_lock();
apply = dsa_port_can_apply_vlan_filtering(dp, vlan_filtering, extack);
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
rcu_read_unlock();
if (!apply)
return -EINVAL;
if (dsa_port_is_vlan_filtering(dp) == vlan_filtering)
return 0;
err = ds->ops->port_vlan_filtering(ds, dp->index, vlan_filtering,
extack);
if (err)
return err;
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
if (ds->vlan_filtering_is_global) {
struct dsa_port *other_dp;
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
ds->vlan_filtering = vlan_filtering;
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
dsa_switch_for_each_user_port(other_dp, ds) {
struct net_device *user = other_dp->user;
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
/* We might be called in the unbind path, so not
* all user devices might still be registered.
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
*/
if (!user)
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
continue;
err = dsa_user_manage_vlan_filtering(user,
vlan_filtering);
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
if (err)
goto restore;
}
} else {
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
dp->vlan_filtering = vlan_filtering;
net: dsa: propagate switchdev vlan_filtering prepare phase to drivers A driver may refuse to enable VLAN filtering for any reason beyond what the DSA framework cares about, such as: - having tc-flower rules that rely on the switch being VLAN-aware - the particular switch does not support VLAN, even if the driver does (the DSA framework just checks for the presence of the .port_vlan_add and .port_vlan_del pointers) - simply not supporting this configuration to be toggled at runtime Currently, when a driver rejects a configuration it cannot support, it does this from the commit phase, which triggers various warnings in switchdev. So propagate the prepare phase to drivers, to give them the ability to refuse invalid configurations cleanly and avoid the warnings. Since we need to modify all function prototypes and check for the prepare phase from within the drivers, take that opportunity and move the existing driver restrictions within the prepare phase where that is possible and easy. Cc: Florian Fainelli <f.fainelli@gmail.com> Cc: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Cc: Hauke Mehrtens <hauke@hauke-m.de> Cc: Woojung Huh <woojung.huh@microchip.com> Cc: Microchip Linux Driver Support <UNGLinuxDriver@microchip.com> Cc: Sean Wang <sean.wang@mediatek.com> Cc: Landen Chao <Landen.Chao@mediatek.com> Cc: Andrew Lunn <andrew@lunn.ch> Cc: Vivien Didelot <vivien.didelot@gmail.com> Cc: Jonathan McDowell <noodles@earth.li> Cc: Linus Walleij <linus.walleij@linaro.org> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com> Cc: Claudiu Manoil <claudiu.manoil@nxp.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-10-02 22:06:46 +00:00
err = dsa_user_manage_vlan_filtering(dp->user,
vlan_filtering);
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
if (err)
goto restore;
}
return 0;
net: dsa: don't advertise 'rx-vlan-filter' when not needed There have been multiple independent reports about dsa_slave_vlan_rx_add_vid being called (and consequently calling the drivers' .port_vlan_add) when it isn't needed, and sometimes (not always) causing problems in the process. Case 1: mv88e6xxx_port_vlan_prepare is stubborn and only accepts VLANs on bridged ports. That is understandably so, because standalone mv88e6xxx ports are VLAN-unaware, and VTU entries are said to be a scarce resource. Otherwise said, the following fails lamentably on mv88e6xxx: ip link add br0 type bridge vlan_filtering 1 ip link set lan3 master br0 ip link add link lan10 name lan10.1 type vlan id 1 [485256.724147] mv88e6085 d0032004.mdio-mii:12: p10: hw VLAN 1 already used by port 3 in br0 RTNETLINK answers: Operation not supported This has become a worse issue since commit 9b236d2a69da ("net: dsa: Advertise the VLAN offload netdev ability only if switch supports it"). Up to that point, the driver was returning -EOPNOTSUPP and DSA was reconverting that error to 0, making the 8021q upper think all is ok (but obviously the error message was there even prior to this change). After that change the -EOPNOTSUPP is propagated to vlan_vid_add, and it is a hard error. Case 2: Ports that don't offload the Linux bridge (have a dp->bridge_dev = NULL because they don't implement .port_bridge_{join,leave}). Understandably, a standalone port should not offload VLANs either, it should remain VLAN unaware and any VLAN should be a software VLAN (as long as the hardware is not quirky, that is). In fact, dsa_slave_port_obj_add does do the right thing and rejects switchdev VLAN objects coming from the bridge when that bridge is not offloaded: case SWITCHDEV_OBJ_ID_PORT_VLAN: if (!dsa_port_offloads_bridge_port(dp, obj->orig_dev)) return -EOPNOTSUPP; err = dsa_slave_vlan_add(dev, obj, extack); But it seems that the bridge is able to trick us. The __vlan_vid_add from br_vlan.c has: /* Try switchdev op first. In case it is not supported, fallback to * 8021q add. */ err = br_switchdev_port_vlan_add(dev, v->vid, flags, extack); if (err == -EOPNOTSUPP) return vlan_vid_add(dev, br->vlan_proto, v->vid); So it says "no, no, you need this VLAN in your life!". And we, naive as we are, say "oh, this comes from the vlan_vid_add code path, it must be an 8021q upper, sure, I'll take that". And we end up with that bridge VLAN installed on our port anyway. But this time, it has the wrong flags: if the bridge was trying to install VLAN 1 as a pvid/untagged VLAN, failed via switchdev, retried via vlan_vid_add, we have this comment: /* This API only allows programming tagged, non-PVID VIDs */ So what we do makes absolutely no sense. Backtracing a bit, we see the common pattern. We allow the network stack to think that our standalone ports are VLAN-aware, but they aren't, for the vast majority of switches. The quirky ones should not dictate the norm. The dsa_slave_vlan_rx_add_vid and dsa_slave_vlan_rx_kill_vid methods exist for drivers that need the 'rx-vlan-filter: on' feature in ethtool -k, which can be due to any of the following reasons: 1. vlan_filtering_is_global = true, and some ports are under a VLAN-aware bridge while others are standalone, and the standalone ports would otherwise drop VLAN-tagged traffic. This is described in commit 061f6a505ac3 ("net: dsa: Add ndo_vlan_rx_{add, kill}_vid implementation"). 2. the ports that are under a VLAN-aware bridge should also set this feature, for 8021q uppers having a VID not claimed by the bridge. In this case, the driver will essentially not even know that the VID is coming from the 8021q layer and not the bridge. 3. Hellcreek. This driver needs it because in standalone mode, it uses unique VLANs per port to ensure separation. For separation of untagged traffic, it uses different PVIDs for each port, and for separation of VLAN-tagged traffic, it never accepts 8021q uppers with the same vid on two ports. If a driver does not fall under any of the above 3 categories, there is no reason why it should advertise the 'rx-vlan-filter' feature, therefore no reason why it should offload the VLANs added through vlan_vid_add. This commit fixes the problem by removing the 'rx-vlan-filter' feature from the slave devices when they operate in standalone mode, and when they offload a VLAN-unaware bridge. The way it works is that vlan_vid_add will now stop its processing here: vlan_add_rx_filter_info: if (!vlan_hw_filter_capable(dev, proto)) return 0; So the VLAN will still be saved in the interface's VLAN RX filtering list, but because it does not declare VLAN filtering in its features, the 8021q module will return zero without committing that VLAN to hardware. This gives the drivers what they want, since it keeps the 8021q VLANs away from the VLAN table until VLAN awareness is enabled (point at which the ports are no longer standalone, hence in the mv88e6xxx case, the check in mv88e6xxx_port_vlan_prepare passes). Since the issue predates the existence of the hellcreek driver, case 3 will be dealt with in a separate patch. The main change that this patch makes is to no longer set NETIF_F_HW_VLAN_CTAG_FILTER unconditionally, but toggle it dynamically (for most switches, never). The second part of the patch addresses an issue that the first part introduces: because the 'rx-vlan-filter' feature is now dynamically toggled, and our .ndo_vlan_rx_add_vid does not get called when 'rx-vlan-filter' is off, we need to avoid bugs such as the following by replaying the VLANs from 8021q uppers every time we enable VLAN filtering: ip link add link lan0 name lan0.100 type vlan id 100 ip addr add 192.168.100.1/24 dev lan0.100 ping 192.168.100.2 # should work ip link add br0 type bridge vlan_filtering 0 ip link set lan0 master br0 ping 192.168.100.2 # should still work ip link set br0 type bridge vlan_filtering 1 ping 192.168.100.2 # should still work but doesn't As reported by Florian, some drivers look at ds->vlan_filtering in their .port_vlan_add() implementation. So this patch also makes sure that ds->vlan_filtering is committed before calling the driver. This is the reason why it is first committed, then restored on the failure path. Reported-by: Tobias Waldekranz <tobias@waldekranz.com> Reported-by: Alvin Šipraga <alsi@bang-olufsen.dk> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-23 21:22:57 +00:00
restore:
ds->ops->port_vlan_filtering(ds, dp->index, old_vlan_filtering, NULL);
if (ds->vlan_filtering_is_global)
ds->vlan_filtering = old_vlan_filtering;
else
dp->vlan_filtering = old_vlan_filtering;
return err;
}
/* This enforces legacy behavior for switch drivers which assume they can't
* receive VLAN configuration when joining a bridge with vlan_filtering=0
*/
bool dsa_port_skip_vlan_configuration(struct dsa_port *dp)
{
struct net_device *br = dsa_port_bridge_dev_get(dp);
struct dsa_switch *ds = dp->ds;
if (!br)
return false;
return !ds->configure_vlan_while_not_filtering && !br_vlan_enabled(br);
}
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
int dsa_port_ageing_time(struct dsa_port *dp, clock_t ageing_clock)
{
unsigned long ageing_jiffies = clock_t_to_jiffies(ageing_clock);
unsigned int ageing_time = jiffies_to_msecs(ageing_jiffies);
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
struct dsa_notifier_ageing_time_info info;
int err;
info.ageing_time = ageing_time;
net: switchdev: remove the transaction structure from port attributes Since the introduction of the switchdev API, port attributes were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port attribute notifier structures, and converts drivers to not look at this member. In part, this patch contains a revert of my previous commit 2e554a7a5d8a ("net: dsa: propagate switchdev vlan_filtering prepare phase to drivers"). For the most part, the conversion was trivial except for: - Rocker's world implementation based on Broadcom OF-DPA had an odd implementation of ofdpa_port_attr_bridge_flags_set. The conversion was done mechanically, by pasting the implementation twice, then only keeping the code that would get executed during prepare phase on top, then only keeping the code that gets executed during the commit phase on bottom, then simplifying the resulting code until this was obtained. - DSA's offloading of STP state, bridge flags, VLAN filtering and multicast router could be converted right away. But the ageing time could not, so a shim was introduced and this was left for a further commit. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de> # hellcreek Reviewed-by: Linus Walleij <linus.walleij@linaro.org> # RTL8366RB Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:50 +00:00
err = dsa_port_notify(dp, DSA_NOTIFIER_AGEING_TIME, &info);
if (err)
return err;
dp->ageing_time = ageing_time;
return 0;
}
int dsa_port_mst_enable(struct dsa_port *dp, bool on,
struct netlink_ext_ack *extack)
{
if (on && !dsa_port_supports_mst(dp)) {
NL_SET_ERR_MSG_MOD(extack, "Hardware does not support MST");
return -EINVAL;
}
return 0;
}
int dsa_port_pre_bridge_flags(const struct dsa_port *dp,
net: dsa: act as passthrough for bridge port flags There are multiple ways in which a PORT_BRIDGE_FLAGS attribute can be expressed by the bridge through switchdev, and not all of them can be emulated by DSA mid-layer API at the same time. One possible configuration is when the bridge offloads the port flags using a mask that has a single bit set - therefore only one feature should change. However, DSA currently groups together unicast and multicast flooding in the .port_egress_floods method, which limits our options when we try to add support for turning off broadcast flooding: do we extend .port_egress_floods with a third parameter which b53 and mv88e6xxx will ignore? But that means that the DSA layer, which currently implements the PRE_BRIDGE_FLAGS attribute all by itself, will see that .port_egress_floods is implemented, and will report that all 3 types of flooding are supported - not necessarily true. Another configuration is when the user specifies more than one flag at the same time, in the same netlink message. If we were to create one individual function per offloadable bridge port flag, we would limit the expressiveness of the switch driver of refusing certain combinations of flag values. For example, a switch may not have an explicit knob for flooding of unknown multicast, just for flooding in general. In that case, the only correct thing to do is to allow changes to BR_FLOOD and BR_MCAST_FLOOD in tandem, and never allow mismatched values. But having a separate .port_set_unicast_flood and .port_set_multicast_flood would not allow the driver to possibly reject that. Also, DSA doesn't consider it necessary to inform the driver that a SWITCHDEV_ATTR_ID_BRIDGE_MROUTER attribute was offloaded, because it just calls .port_egress_floods for the CPU port. When we'll add support for the plain SWITCHDEV_ATTR_ID_PORT_MROUTER, that will become a real problem because the flood settings will need to be held statefully in the DSA middle layer, otherwise changing the mrouter port attribute will impact the flooding attribute. And that's _assuming_ that the underlying hardware doesn't have anything else to do when a multicast router attaches to a port than flood unknown traffic to it. If it does, there will need to be a dedicated .port_set_mrouter anyway. So we need to let the DSA drivers see the exact form that the bridge passes this switchdev attribute in, otherwise we are standing in the way. Therefore we also need to use this form of language when communicating to the driver that it needs to configure its initial (before bridge join) and final (after bridge leave) port flags. The b53 and mv88e6xxx drivers are converted to the passthrough API and their implementation of .port_egress_floods is split into two: a function that configures unicast flooding and another for multicast. The mv88e6xxx implementation is quite hairy, and it turns out that the implementations of unknown unicast flooding are actually the same for 6185 and for 6352: behind the confusing names actually lie two individual bits: NO_UNKNOWN_MC -> FLOOD_UC = 0x4 = BIT(2) NO_UNKNOWN_UC -> FLOOD_MC = 0x8 = BIT(3) so there was no reason to entangle them in the first place. Whereas the 6185 writes to MV88E6185_PORT_CTL0_FORWARD_UNKNOWN of PORT_CTL0, which has the exact same bit index. I have left the implementations separate though, for the only reason that the names are different enough to confuse me, since I am not able to double-check with a user manual. The multicast flooding setting for 6185 is in a different register than for 6352 though. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:56 +00:00
struct switchdev_brport_flags flags,
struct netlink_ext_ack *extack)
{
struct dsa_switch *ds = dp->ds;
net: dsa: act as passthrough for bridge port flags There are multiple ways in which a PORT_BRIDGE_FLAGS attribute can be expressed by the bridge through switchdev, and not all of them can be emulated by DSA mid-layer API at the same time. One possible configuration is when the bridge offloads the port flags using a mask that has a single bit set - therefore only one feature should change. However, DSA currently groups together unicast and multicast flooding in the .port_egress_floods method, which limits our options when we try to add support for turning off broadcast flooding: do we extend .port_egress_floods with a third parameter which b53 and mv88e6xxx will ignore? But that means that the DSA layer, which currently implements the PRE_BRIDGE_FLAGS attribute all by itself, will see that .port_egress_floods is implemented, and will report that all 3 types of flooding are supported - not necessarily true. Another configuration is when the user specifies more than one flag at the same time, in the same netlink message. If we were to create one individual function per offloadable bridge port flag, we would limit the expressiveness of the switch driver of refusing certain combinations of flag values. For example, a switch may not have an explicit knob for flooding of unknown multicast, just for flooding in general. In that case, the only correct thing to do is to allow changes to BR_FLOOD and BR_MCAST_FLOOD in tandem, and never allow mismatched values. But having a separate .port_set_unicast_flood and .port_set_multicast_flood would not allow the driver to possibly reject that. Also, DSA doesn't consider it necessary to inform the driver that a SWITCHDEV_ATTR_ID_BRIDGE_MROUTER attribute was offloaded, because it just calls .port_egress_floods for the CPU port. When we'll add support for the plain SWITCHDEV_ATTR_ID_PORT_MROUTER, that will become a real problem because the flood settings will need to be held statefully in the DSA middle layer, otherwise changing the mrouter port attribute will impact the flooding attribute. And that's _assuming_ that the underlying hardware doesn't have anything else to do when a multicast router attaches to a port than flood unknown traffic to it. If it does, there will need to be a dedicated .port_set_mrouter anyway. So we need to let the DSA drivers see the exact form that the bridge passes this switchdev attribute in, otherwise we are standing in the way. Therefore we also need to use this form of language when communicating to the driver that it needs to configure its initial (before bridge join) and final (after bridge leave) port flags. The b53 and mv88e6xxx drivers are converted to the passthrough API and their implementation of .port_egress_floods is split into two: a function that configures unicast flooding and another for multicast. The mv88e6xxx implementation is quite hairy, and it turns out that the implementations of unknown unicast flooding are actually the same for 6185 and for 6352: behind the confusing names actually lie two individual bits: NO_UNKNOWN_MC -> FLOOD_UC = 0x4 = BIT(2) NO_UNKNOWN_UC -> FLOOD_MC = 0x8 = BIT(3) so there was no reason to entangle them in the first place. Whereas the 6185 writes to MV88E6185_PORT_CTL0_FORWARD_UNKNOWN of PORT_CTL0, which has the exact same bit index. I have left the implementations separate though, for the only reason that the names are different enough to confuse me, since I am not able to double-check with a user manual. The multicast flooding setting for 6185 is in a different register than for 6352 though. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:56 +00:00
if (!ds->ops->port_pre_bridge_flags)
return -EINVAL;
net: dsa: act as passthrough for bridge port flags There are multiple ways in which a PORT_BRIDGE_FLAGS attribute can be expressed by the bridge through switchdev, and not all of them can be emulated by DSA mid-layer API at the same time. One possible configuration is when the bridge offloads the port flags using a mask that has a single bit set - therefore only one feature should change. However, DSA currently groups together unicast and multicast flooding in the .port_egress_floods method, which limits our options when we try to add support for turning off broadcast flooding: do we extend .port_egress_floods with a third parameter which b53 and mv88e6xxx will ignore? But that means that the DSA layer, which currently implements the PRE_BRIDGE_FLAGS attribute all by itself, will see that .port_egress_floods is implemented, and will report that all 3 types of flooding are supported - not necessarily true. Another configuration is when the user specifies more than one flag at the same time, in the same netlink message. If we were to create one individual function per offloadable bridge port flag, we would limit the expressiveness of the switch driver of refusing certain combinations of flag values. For example, a switch may not have an explicit knob for flooding of unknown multicast, just for flooding in general. In that case, the only correct thing to do is to allow changes to BR_FLOOD and BR_MCAST_FLOOD in tandem, and never allow mismatched values. But having a separate .port_set_unicast_flood and .port_set_multicast_flood would not allow the driver to possibly reject that. Also, DSA doesn't consider it necessary to inform the driver that a SWITCHDEV_ATTR_ID_BRIDGE_MROUTER attribute was offloaded, because it just calls .port_egress_floods for the CPU port. When we'll add support for the plain SWITCHDEV_ATTR_ID_PORT_MROUTER, that will become a real problem because the flood settings will need to be held statefully in the DSA middle layer, otherwise changing the mrouter port attribute will impact the flooding attribute. And that's _assuming_ that the underlying hardware doesn't have anything else to do when a multicast router attaches to a port than flood unknown traffic to it. If it does, there will need to be a dedicated .port_set_mrouter anyway. So we need to let the DSA drivers see the exact form that the bridge passes this switchdev attribute in, otherwise we are standing in the way. Therefore we also need to use this form of language when communicating to the driver that it needs to configure its initial (before bridge join) and final (after bridge leave) port flags. The b53 and mv88e6xxx drivers are converted to the passthrough API and their implementation of .port_egress_floods is split into two: a function that configures unicast flooding and another for multicast. The mv88e6xxx implementation is quite hairy, and it turns out that the implementations of unknown unicast flooding are actually the same for 6185 and for 6352: behind the confusing names actually lie two individual bits: NO_UNKNOWN_MC -> FLOOD_UC = 0x4 = BIT(2) NO_UNKNOWN_UC -> FLOOD_MC = 0x8 = BIT(3) so there was no reason to entangle them in the first place. Whereas the 6185 writes to MV88E6185_PORT_CTL0_FORWARD_UNKNOWN of PORT_CTL0, which has the exact same bit index. I have left the implementations separate though, for the only reason that the names are different enough to confuse me, since I am not able to double-check with a user manual. The multicast flooding setting for 6185 is in a different register than for 6352 though. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:56 +00:00
return ds->ops->port_pre_bridge_flags(ds, dp->index, flags, extack);
}
net: dsa: centralize fast ageing when address learning is turned off Currently DSA leaves it down to device drivers to fast age the FDB on a port when address learning is disabled on it. There are 2 reasons for doing that in the first place: - when address learning is disabled by user space, through IFLA_BRPORT_LEARNING or the brport_attr_learning sysfs, what user space typically wants to achieve is to operate in a mode with no dynamic FDB entry on that port. But if the port is already up, some addresses might have been already learned on it, and it seems silly to wait for 5 minutes for them to expire until something useful can be done. - when a port leaves a bridge and becomes standalone, DSA turns off address learning on it. This also has the nice side effect of flushing the dynamically learned bridge FDB entries on it, which is a good idea because standalone ports should not have bridge FDB entries on them. We let drivers manage fast ageing under this condition because if DSA were to do it, it would need to track each port's learning state, and act upon the transition, which it currently doesn't. But there are 2 reasons why doing it is better after all: - drivers might get it wrong and not do it (see b53_port_set_learning) - we would like to flush the dynamic entries from the software bridge too, and letting drivers do that would be another pain point So track the port learning state and trigger a fast age process automatically within DSA. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-08 14:35:23 +00:00
int dsa_port_bridge_flags(struct dsa_port *dp,
net: dsa: act as passthrough for bridge port flags There are multiple ways in which a PORT_BRIDGE_FLAGS attribute can be expressed by the bridge through switchdev, and not all of them can be emulated by DSA mid-layer API at the same time. One possible configuration is when the bridge offloads the port flags using a mask that has a single bit set - therefore only one feature should change. However, DSA currently groups together unicast and multicast flooding in the .port_egress_floods method, which limits our options when we try to add support for turning off broadcast flooding: do we extend .port_egress_floods with a third parameter which b53 and mv88e6xxx will ignore? But that means that the DSA layer, which currently implements the PRE_BRIDGE_FLAGS attribute all by itself, will see that .port_egress_floods is implemented, and will report that all 3 types of flooding are supported - not necessarily true. Another configuration is when the user specifies more than one flag at the same time, in the same netlink message. If we were to create one individual function per offloadable bridge port flag, we would limit the expressiveness of the switch driver of refusing certain combinations of flag values. For example, a switch may not have an explicit knob for flooding of unknown multicast, just for flooding in general. In that case, the only correct thing to do is to allow changes to BR_FLOOD and BR_MCAST_FLOOD in tandem, and never allow mismatched values. But having a separate .port_set_unicast_flood and .port_set_multicast_flood would not allow the driver to possibly reject that. Also, DSA doesn't consider it necessary to inform the driver that a SWITCHDEV_ATTR_ID_BRIDGE_MROUTER attribute was offloaded, because it just calls .port_egress_floods for the CPU port. When we'll add support for the plain SWITCHDEV_ATTR_ID_PORT_MROUTER, that will become a real problem because the flood settings will need to be held statefully in the DSA middle layer, otherwise changing the mrouter port attribute will impact the flooding attribute. And that's _assuming_ that the underlying hardware doesn't have anything else to do when a multicast router attaches to a port than flood unknown traffic to it. If it does, there will need to be a dedicated .port_set_mrouter anyway. So we need to let the DSA drivers see the exact form that the bridge passes this switchdev attribute in, otherwise we are standing in the way. Therefore we also need to use this form of language when communicating to the driver that it needs to configure its initial (before bridge join) and final (after bridge leave) port flags. The b53 and mv88e6xxx drivers are converted to the passthrough API and their implementation of .port_egress_floods is split into two: a function that configures unicast flooding and another for multicast. The mv88e6xxx implementation is quite hairy, and it turns out that the implementations of unknown unicast flooding are actually the same for 6185 and for 6352: behind the confusing names actually lie two individual bits: NO_UNKNOWN_MC -> FLOOD_UC = 0x4 = BIT(2) NO_UNKNOWN_UC -> FLOOD_MC = 0x8 = BIT(3) so there was no reason to entangle them in the first place. Whereas the 6185 writes to MV88E6185_PORT_CTL0_FORWARD_UNKNOWN of PORT_CTL0, which has the exact same bit index. I have left the implementations separate though, for the only reason that the names are different enough to confuse me, since I am not able to double-check with a user manual. The multicast flooding setting for 6185 is in a different register than for 6352 though. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:56 +00:00
struct switchdev_brport_flags flags,
struct netlink_ext_ack *extack)
{
struct dsa_switch *ds = dp->ds;
net: dsa: centralize fast ageing when address learning is turned off Currently DSA leaves it down to device drivers to fast age the FDB on a port when address learning is disabled on it. There are 2 reasons for doing that in the first place: - when address learning is disabled by user space, through IFLA_BRPORT_LEARNING or the brport_attr_learning sysfs, what user space typically wants to achieve is to operate in a mode with no dynamic FDB entry on that port. But if the port is already up, some addresses might have been already learned on it, and it seems silly to wait for 5 minutes for them to expire until something useful can be done. - when a port leaves a bridge and becomes standalone, DSA turns off address learning on it. This also has the nice side effect of flushing the dynamically learned bridge FDB entries on it, which is a good idea because standalone ports should not have bridge FDB entries on them. We let drivers manage fast ageing under this condition because if DSA were to do it, it would need to track each port's learning state, and act upon the transition, which it currently doesn't. But there are 2 reasons why doing it is better after all: - drivers might get it wrong and not do it (see b53_port_set_learning) - we would like to flush the dynamic entries from the software bridge too, and letting drivers do that would be another pain point So track the port learning state and trigger a fast age process automatically within DSA. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-08 14:35:23 +00:00
int err;
net: dsa: act as passthrough for bridge port flags There are multiple ways in which a PORT_BRIDGE_FLAGS attribute can be expressed by the bridge through switchdev, and not all of them can be emulated by DSA mid-layer API at the same time. One possible configuration is when the bridge offloads the port flags using a mask that has a single bit set - therefore only one feature should change. However, DSA currently groups together unicast and multicast flooding in the .port_egress_floods method, which limits our options when we try to add support for turning off broadcast flooding: do we extend .port_egress_floods with a third parameter which b53 and mv88e6xxx will ignore? But that means that the DSA layer, which currently implements the PRE_BRIDGE_FLAGS attribute all by itself, will see that .port_egress_floods is implemented, and will report that all 3 types of flooding are supported - not necessarily true. Another configuration is when the user specifies more than one flag at the same time, in the same netlink message. If we were to create one individual function per offloadable bridge port flag, we would limit the expressiveness of the switch driver of refusing certain combinations of flag values. For example, a switch may not have an explicit knob for flooding of unknown multicast, just for flooding in general. In that case, the only correct thing to do is to allow changes to BR_FLOOD and BR_MCAST_FLOOD in tandem, and never allow mismatched values. But having a separate .port_set_unicast_flood and .port_set_multicast_flood would not allow the driver to possibly reject that. Also, DSA doesn't consider it necessary to inform the driver that a SWITCHDEV_ATTR_ID_BRIDGE_MROUTER attribute was offloaded, because it just calls .port_egress_floods for the CPU port. When we'll add support for the plain SWITCHDEV_ATTR_ID_PORT_MROUTER, that will become a real problem because the flood settings will need to be held statefully in the DSA middle layer, otherwise changing the mrouter port attribute will impact the flooding attribute. And that's _assuming_ that the underlying hardware doesn't have anything else to do when a multicast router attaches to a port than flood unknown traffic to it. If it does, there will need to be a dedicated .port_set_mrouter anyway. So we need to let the DSA drivers see the exact form that the bridge passes this switchdev attribute in, otherwise we are standing in the way. Therefore we also need to use this form of language when communicating to the driver that it needs to configure its initial (before bridge join) and final (after bridge leave) port flags. The b53 and mv88e6xxx drivers are converted to the passthrough API and their implementation of .port_egress_floods is split into two: a function that configures unicast flooding and another for multicast. The mv88e6xxx implementation is quite hairy, and it turns out that the implementations of unknown unicast flooding are actually the same for 6185 and for 6352: behind the confusing names actually lie two individual bits: NO_UNKNOWN_MC -> FLOOD_UC = 0x4 = BIT(2) NO_UNKNOWN_UC -> FLOOD_MC = 0x8 = BIT(3) so there was no reason to entangle them in the first place. Whereas the 6185 writes to MV88E6185_PORT_CTL0_FORWARD_UNKNOWN of PORT_CTL0, which has the exact same bit index. I have left the implementations separate though, for the only reason that the names are different enough to confuse me, since I am not able to double-check with a user manual. The multicast flooding setting for 6185 is in a different register than for 6352 though. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-12 15:15:56 +00:00
if (!ds->ops->port_bridge_flags)
return -EOPNOTSUPP;
net: dsa: centralize fast ageing when address learning is turned off Currently DSA leaves it down to device drivers to fast age the FDB on a port when address learning is disabled on it. There are 2 reasons for doing that in the first place: - when address learning is disabled by user space, through IFLA_BRPORT_LEARNING or the brport_attr_learning sysfs, what user space typically wants to achieve is to operate in a mode with no dynamic FDB entry on that port. But if the port is already up, some addresses might have been already learned on it, and it seems silly to wait for 5 minutes for them to expire until something useful can be done. - when a port leaves a bridge and becomes standalone, DSA turns off address learning on it. This also has the nice side effect of flushing the dynamically learned bridge FDB entries on it, which is a good idea because standalone ports should not have bridge FDB entries on them. We let drivers manage fast ageing under this condition because if DSA were to do it, it would need to track each port's learning state, and act upon the transition, which it currently doesn't. But there are 2 reasons why doing it is better after all: - drivers might get it wrong and not do it (see b53_port_set_learning) - we would like to flush the dynamic entries from the software bridge too, and letting drivers do that would be another pain point So track the port learning state and trigger a fast age process automatically within DSA. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-08 14:35:23 +00:00
err = ds->ops->port_bridge_flags(ds, dp->index, flags, extack);
if (err)
return err;
if (flags.mask & BR_LEARNING) {
bool learning = flags.val & BR_LEARNING;
if (learning == dp->learning)
return 0;
net: dsa: avoid fast ageing twice when port leaves a bridge Drivers that support both the toggling of address learning and dynamic FDB flushing (mv88e6xxx, b53, sja1105) currently need to fast-age a port twice when it leaves a bridge: - once, when del_nbp() calls br_stp_disable_port() which puts the port in the BLOCKING state - twice, when dsa_port_switchdev_unsync_attrs() calls dsa_port_clear_brport_flags() which disables address learning The knee-jerk reaction might be to say "dsa_port_clear_brport_flags does not need to fast-age the port at all", but the thing is, we still need both code paths to flush the dynamic FDB entries in different situations. When a DSA switch port leaves a bonding/team interface that is (still) a bridge port, no del_nbp() will be called, so we rely on dsa_port_clear_brport_flags() function to restore proper standalone port functionality with address learning disabled. So the solution is just to avoid double the work when both code paths are called in series. Luckily, DSA already caches the STP port state, so we can skip flushing the dynamic FDB when we disable address learning and the STP state is one where no address learning takes place at all. Under that condition, not flushing the FDB is safe because there is supposed to not be any dynamic FDB entry at all (they were flushed during the transition towards that state, and none were learned in the meanwhile). Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-08 22:56:49 +00:00
if ((dp->learning && !learning) &&
(dp->stp_state == BR_STATE_LEARNING ||
dp->stp_state == BR_STATE_FORWARDING))
net: dsa: centralize fast ageing when address learning is turned off Currently DSA leaves it down to device drivers to fast age the FDB on a port when address learning is disabled on it. There are 2 reasons for doing that in the first place: - when address learning is disabled by user space, through IFLA_BRPORT_LEARNING or the brport_attr_learning sysfs, what user space typically wants to achieve is to operate in a mode with no dynamic FDB entry on that port. But if the port is already up, some addresses might have been already learned on it, and it seems silly to wait for 5 minutes for them to expire until something useful can be done. - when a port leaves a bridge and becomes standalone, DSA turns off address learning on it. This also has the nice side effect of flushing the dynamically learned bridge FDB entries on it, which is a good idea because standalone ports should not have bridge FDB entries on them. We let drivers manage fast ageing under this condition because if DSA were to do it, it would need to track each port's learning state, and act upon the transition, which it currently doesn't. But there are 2 reasons why doing it is better after all: - drivers might get it wrong and not do it (see b53_port_set_learning) - we would like to flush the dynamic entries from the software bridge too, and letting drivers do that would be another pain point So track the port learning state and trigger a fast age process automatically within DSA. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-08 14:35:23 +00:00
dsa_port_fast_age(dp);
dp->learning = learning;
}
return 0;
}
net: dsa: felix: manage host flooding using a specific driver callback At the time - commit 7569459a52c9 ("net: dsa: manage flooding on the CPU ports") - not introducing a dedicated switch callback for host flooding made sense, because for the only user, the felix driver, there was nothing different to do for the CPU port than set the flood flags on the CPU port just like on any other bridge port. There are 2 reasons why this approach is not good enough, however. (1) Other drivers, like sja1105, support configuring flooding as a function of {ingress port, egress port}, whereas the DSA ->port_bridge_flags() function only operates on an egress port. So with that driver we'd have useless host flooding from user ports which don't need it. (2) Even with the felix driver, support for multiple CPU ports makes it difficult to piggyback on ->port_bridge_flags(). The way in which the felix driver is going to support host-filtered addresses with multiple CPU ports is that it will direct these addresses towards both CPU ports (in a sort of multicast fashion), then restrict the forwarding to only one of the two using the forwarding masks. Consequently, flooding will also be enabled towards both CPU ports. However, ->port_bridge_flags() gets passed the index of a single CPU port, and that leaves the flood settings out of sync between the 2 CPU ports. This is to say, it's better to have a specific driver method for host flooding, which takes the user port as argument. This solves problem (1) by allowing the driver to do different things for different user ports, and problem (2) by abstracting the operation and letting the driver do whatever, rather than explicitly making the DSA core point to the CPU port it thinks needs to be touched. This new method also creates a problem, which is that cross-chip setups are not handled. However I don't have hardware right now where I can test what is the proper thing to do, and there isn't hardware compatible with multi-switch trees that supports host flooding. So it remains a problem to be tackled in the future. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-05-11 09:50:17 +00:00
void dsa_port_set_host_flood(struct dsa_port *dp, bool uc, bool mc)
{
struct dsa_switch *ds = dp->ds;
if (ds->ops->port_set_host_flood)
ds->ops->port_set_host_flood(ds, dp->index, uc, mc);
}
int dsa_port_vlan_msti(struct dsa_port *dp,
const struct switchdev_vlan_msti *msti)
{
struct dsa_switch *ds = dp->ds;
if (!ds->ops->vlan_msti_set)
return -EOPNOTSUPP;
return ds->ops->vlan_msti_set(ds, *dp->bridge, msti);
}
int dsa_port_mtu_change(struct dsa_port *dp, int new_mtu)
net: dsa: configure the MTU for switch ports It is useful be able to configure port policers on a switch to accept frames of various sizes: - Increase the MTU for better throughput from the default of 1500 if it is known that there is no 10/100 Mbps device in the network. - Decrease the MTU to limit the latency of high-priority frames under congestion, or work around various network segments that add extra headers to packets which can't be fragmented. For DSA slave ports, this is mostly a pass-through callback, called through the regular ndo ops and at probe time (to ensure consistency across all supported switches). The CPU port is called with an MTU equal to the largest configured MTU of the slave ports. The assumption is that the user might want to sustain a bidirectional conversation with a partner over any switch port. The DSA master is configured the same as the CPU port, plus the tagger overhead. Since the MTU is by definition L2 payload (sans Ethernet header), it is up to each individual driver to figure out if it needs to do anything special for its frame tags on the CPU port (it shouldn't except in special cases). So the MTU does not contain the tagger overhead on the CPU port. However the MTU of the DSA master, minus the tagger overhead, is used as a proxy for the MTU of the CPU port, which does not have a net device. This is to avoid uselessly calling the .change_mtu function on the CPU port when nothing should change. So it is safe to assume that the DSA master and the CPU port MTUs are apart by exactly the tagger's overhead in bytes. Some changes were made around dsa_master_set_mtu(), function which was now removed, for 2 reasons: - dev_set_mtu() already calls dev_validate_mtu(), so it's redundant to do the same thing in DSA - __dev_set_mtu() returns 0 if ops->ndo_change_mtu is an absent method That is to say, there's no need for this function in DSA, we can safely call dev_set_mtu() directly, take the rtnl lock when necessary, and just propagate whatever errors get reported (since the user probably wants to be informed). Some inspiration (mainly in the MTU DSA notifier) was taken from a vaguely similar patch from Murali and Florian, who are credited as co-developers down below. Co-developed-by: Murali Krishna Policharla <murali.policharla@broadcom.com> Signed-off-by: Murali Krishna Policharla <murali.policharla@broadcom.com> Co-developed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-03-27 19:55:42 +00:00
{
struct dsa_notifier_mtu_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
net: dsa: configure the MTU for switch ports It is useful be able to configure port policers on a switch to accept frames of various sizes: - Increase the MTU for better throughput from the default of 1500 if it is known that there is no 10/100 Mbps device in the network. - Decrease the MTU to limit the latency of high-priority frames under congestion, or work around various network segments that add extra headers to packets which can't be fragmented. For DSA slave ports, this is mostly a pass-through callback, called through the regular ndo ops and at probe time (to ensure consistency across all supported switches). The CPU port is called with an MTU equal to the largest configured MTU of the slave ports. The assumption is that the user might want to sustain a bidirectional conversation with a partner over any switch port. The DSA master is configured the same as the CPU port, plus the tagger overhead. Since the MTU is by definition L2 payload (sans Ethernet header), it is up to each individual driver to figure out if it needs to do anything special for its frame tags on the CPU port (it shouldn't except in special cases). So the MTU does not contain the tagger overhead on the CPU port. However the MTU of the DSA master, minus the tagger overhead, is used as a proxy for the MTU of the CPU port, which does not have a net device. This is to avoid uselessly calling the .change_mtu function on the CPU port when nothing should change. So it is safe to assume that the DSA master and the CPU port MTUs are apart by exactly the tagger's overhead in bytes. Some changes were made around dsa_master_set_mtu(), function which was now removed, for 2 reasons: - dev_set_mtu() already calls dev_validate_mtu(), so it's redundant to do the same thing in DSA - __dev_set_mtu() returns 0 if ops->ndo_change_mtu is an absent method That is to say, there's no need for this function in DSA, we can safely call dev_set_mtu() directly, take the rtnl lock when necessary, and just propagate whatever errors get reported (since the user probably wants to be informed). Some inspiration (mainly in the MTU DSA notifier) was taken from a vaguely similar patch from Murali and Florian, who are credited as co-developers down below. Co-developed-by: Murali Krishna Policharla <murali.policharla@broadcom.com> Signed-off-by: Murali Krishna Policharla <murali.policharla@broadcom.com> Co-developed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-03-27 19:55:42 +00:00
.mtu = new_mtu,
};
return dsa_port_notify(dp, DSA_NOTIFIER_MTU, &info);
}
int dsa_port_fdb_add(struct dsa_port *dp, const unsigned char *addr,
u16 vid)
{
struct dsa_notifier_fdb_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.addr = addr,
.vid = vid,
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
.db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
},
};
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
/* Refcounting takes bridge.num as a key, and should be global for all
* bridges in the absence of FDB isolation, and per bridge otherwise.
* Force the bridge.num to zero here in the absence of FDB isolation.
*/
if (!dp->ds->fdb_isolation)
info.db.bridge.num = 0;
return dsa_port_notify(dp, DSA_NOTIFIER_FDB_ADD, &info);
}
int dsa_port_fdb_del(struct dsa_port *dp, const unsigned char *addr,
u16 vid)
{
struct dsa_notifier_fdb_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.addr = addr,
.vid = vid,
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
.db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
},
};
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
if (!dp->ds->fdb_isolation)
info.db.bridge.num = 0;
return dsa_port_notify(dp, DSA_NOTIFIER_FDB_DEL, &info);
}
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
static int dsa_port_host_fdb_add(struct dsa_port *dp,
const unsigned char *addr, u16 vid,
struct dsa_db db)
{
struct dsa_notifier_fdb_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.addr = addr,
.vid = vid,
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
.db = db,
};
return dsa_port_notify(dp, DSA_NOTIFIER_HOST_FDB_ADD, &info);
}
int dsa_port_standalone_host_fdb_add(struct dsa_port *dp,
const unsigned char *addr, u16 vid)
{
struct dsa_db db = {
.type = DSA_DB_PORT,
.dp = dp,
};
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
return dsa_port_host_fdb_add(dp, addr, vid, db);
}
int dsa_port_bridge_host_fdb_add(struct dsa_port *dp,
const unsigned char *addr, u16 vid)
{
struct net_device *conduit = dsa_port_to_conduit(dp);
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
struct dsa_db db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
};
int err;
net: dsa: fix db type confusion in host fdb/mdb add/del We have the following code paths: Host FDB (unicast RX filtering): dsa_port_standalone_host_fdb_add() dsa_port_bridge_host_fdb_add() | | +--------------+ +------------+ | | v v dsa_port_host_fdb_add() dsa_port_standalone_host_fdb_del() dsa_port_bridge_host_fdb_del() | | +--------------+ +------------+ | | v v dsa_port_host_fdb_del() Host MDB (multicast RX filtering): dsa_port_standalone_host_mdb_add() dsa_port_bridge_host_mdb_add() | | +--------------+ +------------+ | | v v dsa_port_host_mdb_add() dsa_port_standalone_host_mdb_del() dsa_port_bridge_host_mdb_del() | | +--------------+ +------------+ | | v v dsa_port_host_mdb_del() The logic added by commit 5e8a1e03aa4d ("net: dsa: install secondary unicast and multicast addresses as host FDB/MDB") zeroes out db.bridge.num if the switch doesn't support ds->fdb_isolation (the majority doesn't). This is done for a reason explained in commit c26933639b54 ("net: dsa: request drivers to perform FDB isolation"). Taking a single code path as example - dsa_port_host_fdb_add() - the others are similar - the problem is that this function handles: - DSA_DB_PORT databases, when called from dsa_port_standalone_host_fdb_add() - DSA_DB_BRIDGE databases, when called from dsa_port_bridge_host_fdb_add() So, if dsa_port_host_fdb_add() were to make any change on the "bridge.num" attribute of the database, this would only be correct for a DSA_DB_BRIDGE, and a type confusion for a DSA_DB_PORT bridge. However, this bug is without consequences, for 2 reasons: - dsa_port_standalone_host_fdb_add() is only called from code which is (in)directly guarded by dsa_switch_supports_uc_filtering(ds), and that function only returns true if ds->fdb_isolation is set. So, the code only executed for DSA_DB_BRIDGE databases. - Even if the code was not dead for DSA_DB_PORT, we have the following memory layout: struct dsa_bridge { struct net_device *dev; unsigned int num; bool tx_fwd_offload; refcount_t refcount; }; struct dsa_db { enum dsa_db_type type; union { const struct dsa_port *dp; // DSA_DB_PORT struct dsa_lag lag; struct dsa_bridge bridge; // DSA_DB_BRIDGE }; }; So, the zeroization of dsa_db :: bridge :: num on a dsa_db structure of type DSA_DB_PORT would access memory which is unused, because we only use dsa_db :: dp for DSA_DB_PORT, and this is mapped at the same address with dsa_db :: dev for DSA_DB_BRIDGE, thanks to the union definition. It is correct to fix up dsa_db :: bridge :: num only from code paths that come from the bridge / switchdev, so move these there. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Simon Horman <simon.horman@corigine.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Link: https://lore.kernel.org/r/20230329133819.697642-1-vladimir.oltean@nxp.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2023-03-29 13:38:19 +00:00
if (!dp->ds->fdb_isolation)
db.bridge.num = 0;
/* Avoid a call to __dev_set_promiscuity() on the conduit, which
net: dsa: avoid call to __dev_set_promiscuity() while rtnl_mutex isn't held If the DSA master doesn't support IFF_UNICAST_FLT, then the following call path is possible: dsa_slave_switchdev_event_work -> dsa_port_host_fdb_add -> dev_uc_add -> __dev_set_rx_mode -> __dev_set_promiscuity Since the blamed commit, dsa_slave_switchdev_event_work() no longer holds rtnl_lock(), which triggers the ASSERT_RTNL() from __dev_set_promiscuity(). Taking rtnl_lock() around dev_uc_add() is impossible, because all the code paths that call dsa_flush_workqueue() do so from contexts where the rtnl_mutex is already held - so this would lead to an instant deadlock. dev_uc_add() in itself doesn't require the rtnl_mutex for protection. There is this comment in __dev_set_rx_mode() which assumes so: /* Unicast addresses changes may only happen under the rtnl, * therefore calling __dev_set_promiscuity here is safe. */ but it is from commit 4417da668c00 ("[NET]: dev: secondary unicast address support") dated June 2007, and in the meantime, commit f1f28aa3510d ("netdev: Add addr_list_lock to struct net_device."), dated July 2008, has added &dev->addr_list_lock to protect this instead of the global rtnl_mutex. Nonetheless, __dev_set_promiscuity() does assume rtnl_mutex protection, but it is the uncommon path of what we typically expect dev_uc_add() to do. So since only the uncommon path requires rtnl_lock(), just check ahead of time whether dev_uc_add() would result into a call to __dev_set_promiscuity(), and handle that condition separately. DSA already configures the master interface to be promiscuous if the tagger requires this. We can extend this to also cover the case where the master doesn't handle dev_uc_add() (doesn't support IFF_UNICAST_FLT), and on the premise that we'd end up making it promiscuous during operation anyway, either if a DSA slave has a non-inherited MAC address, or if the bridge notifies local FDB entries for its own MAC address, the address of a station learned on a foreign port, etc. Fixes: 0faf890fc519 ("net: dsa: drop rtnl_lock from dsa_slave_switchdev_event_work") Reported-by: Oleksij Rempel <o.rempel@pengutronix.de> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-18 12:13:02 +00:00
* requires rtnl_lock(), since we can't guarantee that is held here,
* and we can't take it either.
*/
if (conduit->priv_flags & IFF_UNICAST_FLT) {
err = dev_uc_add(conduit, addr);
net: dsa: avoid call to __dev_set_promiscuity() while rtnl_mutex isn't held If the DSA master doesn't support IFF_UNICAST_FLT, then the following call path is possible: dsa_slave_switchdev_event_work -> dsa_port_host_fdb_add -> dev_uc_add -> __dev_set_rx_mode -> __dev_set_promiscuity Since the blamed commit, dsa_slave_switchdev_event_work() no longer holds rtnl_lock(), which triggers the ASSERT_RTNL() from __dev_set_promiscuity(). Taking rtnl_lock() around dev_uc_add() is impossible, because all the code paths that call dsa_flush_workqueue() do so from contexts where the rtnl_mutex is already held - so this would lead to an instant deadlock. dev_uc_add() in itself doesn't require the rtnl_mutex for protection. There is this comment in __dev_set_rx_mode() which assumes so: /* Unicast addresses changes may only happen under the rtnl, * therefore calling __dev_set_promiscuity here is safe. */ but it is from commit 4417da668c00 ("[NET]: dev: secondary unicast address support") dated June 2007, and in the meantime, commit f1f28aa3510d ("netdev: Add addr_list_lock to struct net_device."), dated July 2008, has added &dev->addr_list_lock to protect this instead of the global rtnl_mutex. Nonetheless, __dev_set_promiscuity() does assume rtnl_mutex protection, but it is the uncommon path of what we typically expect dev_uc_add() to do. So since only the uncommon path requires rtnl_lock(), just check ahead of time whether dev_uc_add() would result into a call to __dev_set_promiscuity(), and handle that condition separately. DSA already configures the master interface to be promiscuous if the tagger requires this. We can extend this to also cover the case where the master doesn't handle dev_uc_add() (doesn't support IFF_UNICAST_FLT), and on the premise that we'd end up making it promiscuous during operation anyway, either if a DSA slave has a non-inherited MAC address, or if the bridge notifies local FDB entries for its own MAC address, the address of a station learned on a foreign port, etc. Fixes: 0faf890fc519 ("net: dsa: drop rtnl_lock from dsa_slave_switchdev_event_work") Reported-by: Oleksij Rempel <o.rempel@pengutronix.de> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-18 12:13:02 +00:00
if (err)
return err;
}
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
return dsa_port_host_fdb_add(dp, addr, vid, db);
}
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
static int dsa_port_host_fdb_del(struct dsa_port *dp,
const unsigned char *addr, u16 vid,
struct dsa_db db)
{
struct dsa_notifier_fdb_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.addr = addr,
.vid = vid,
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
.db = db,
};
return dsa_port_notify(dp, DSA_NOTIFIER_HOST_FDB_DEL, &info);
}
int dsa_port_standalone_host_fdb_del(struct dsa_port *dp,
const unsigned char *addr, u16 vid)
{
struct dsa_db db = {
.type = DSA_DB_PORT,
.dp = dp,
};
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
return dsa_port_host_fdb_del(dp, addr, vid, db);
}
int dsa_port_bridge_host_fdb_del(struct dsa_port *dp,
const unsigned char *addr, u16 vid)
{
struct net_device *conduit = dsa_port_to_conduit(dp);
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
struct dsa_db db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
};
int err;
net: dsa: fix db type confusion in host fdb/mdb add/del We have the following code paths: Host FDB (unicast RX filtering): dsa_port_standalone_host_fdb_add() dsa_port_bridge_host_fdb_add() | | +--------------+ +------------+ | | v v dsa_port_host_fdb_add() dsa_port_standalone_host_fdb_del() dsa_port_bridge_host_fdb_del() | | +--------------+ +------------+ | | v v dsa_port_host_fdb_del() Host MDB (multicast RX filtering): dsa_port_standalone_host_mdb_add() dsa_port_bridge_host_mdb_add() | | +--------------+ +------------+ | | v v dsa_port_host_mdb_add() dsa_port_standalone_host_mdb_del() dsa_port_bridge_host_mdb_del() | | +--------------+ +------------+ | | v v dsa_port_host_mdb_del() The logic added by commit 5e8a1e03aa4d ("net: dsa: install secondary unicast and multicast addresses as host FDB/MDB") zeroes out db.bridge.num if the switch doesn't support ds->fdb_isolation (the majority doesn't). This is done for a reason explained in commit c26933639b54 ("net: dsa: request drivers to perform FDB isolation"). Taking a single code path as example - dsa_port_host_fdb_add() - the others are similar - the problem is that this function handles: - DSA_DB_PORT databases, when called from dsa_port_standalone_host_fdb_add() - DSA_DB_BRIDGE databases, when called from dsa_port_bridge_host_fdb_add() So, if dsa_port_host_fdb_add() were to make any change on the "bridge.num" attribute of the database, this would only be correct for a DSA_DB_BRIDGE, and a type confusion for a DSA_DB_PORT bridge. However, this bug is without consequences, for 2 reasons: - dsa_port_standalone_host_fdb_add() is only called from code which is (in)directly guarded by dsa_switch_supports_uc_filtering(ds), and that function only returns true if ds->fdb_isolation is set. So, the code only executed for DSA_DB_BRIDGE databases. - Even if the code was not dead for DSA_DB_PORT, we have the following memory layout: struct dsa_bridge { struct net_device *dev; unsigned int num; bool tx_fwd_offload; refcount_t refcount; }; struct dsa_db { enum dsa_db_type type; union { const struct dsa_port *dp; // DSA_DB_PORT struct dsa_lag lag; struct dsa_bridge bridge; // DSA_DB_BRIDGE }; }; So, the zeroization of dsa_db :: bridge :: num on a dsa_db structure of type DSA_DB_PORT would access memory which is unused, because we only use dsa_db :: dp for DSA_DB_PORT, and this is mapped at the same address with dsa_db :: dev for DSA_DB_BRIDGE, thanks to the union definition. It is correct to fix up dsa_db :: bridge :: num only from code paths that come from the bridge / switchdev, so move these there. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Simon Horman <simon.horman@corigine.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Link: https://lore.kernel.org/r/20230329133819.697642-1-vladimir.oltean@nxp.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2023-03-29 13:38:19 +00:00
if (!dp->ds->fdb_isolation)
db.bridge.num = 0;
if (conduit->priv_flags & IFF_UNICAST_FLT) {
err = dev_uc_del(conduit, addr);
net: dsa: avoid call to __dev_set_promiscuity() while rtnl_mutex isn't held If the DSA master doesn't support IFF_UNICAST_FLT, then the following call path is possible: dsa_slave_switchdev_event_work -> dsa_port_host_fdb_add -> dev_uc_add -> __dev_set_rx_mode -> __dev_set_promiscuity Since the blamed commit, dsa_slave_switchdev_event_work() no longer holds rtnl_lock(), which triggers the ASSERT_RTNL() from __dev_set_promiscuity(). Taking rtnl_lock() around dev_uc_add() is impossible, because all the code paths that call dsa_flush_workqueue() do so from contexts where the rtnl_mutex is already held - so this would lead to an instant deadlock. dev_uc_add() in itself doesn't require the rtnl_mutex for protection. There is this comment in __dev_set_rx_mode() which assumes so: /* Unicast addresses changes may only happen under the rtnl, * therefore calling __dev_set_promiscuity here is safe. */ but it is from commit 4417da668c00 ("[NET]: dev: secondary unicast address support") dated June 2007, and in the meantime, commit f1f28aa3510d ("netdev: Add addr_list_lock to struct net_device."), dated July 2008, has added &dev->addr_list_lock to protect this instead of the global rtnl_mutex. Nonetheless, __dev_set_promiscuity() does assume rtnl_mutex protection, but it is the uncommon path of what we typically expect dev_uc_add() to do. So since only the uncommon path requires rtnl_lock(), just check ahead of time whether dev_uc_add() would result into a call to __dev_set_promiscuity(), and handle that condition separately. DSA already configures the master interface to be promiscuous if the tagger requires this. We can extend this to also cover the case where the master doesn't handle dev_uc_add() (doesn't support IFF_UNICAST_FLT), and on the premise that we'd end up making it promiscuous during operation anyway, either if a DSA slave has a non-inherited MAC address, or if the bridge notifies local FDB entries for its own MAC address, the address of a station learned on a foreign port, etc. Fixes: 0faf890fc519 ("net: dsa: drop rtnl_lock from dsa_slave_switchdev_event_work") Reported-by: Oleksij Rempel <o.rempel@pengutronix.de> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-18 12:13:02 +00:00
if (err)
return err;
}
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
return dsa_port_host_fdb_del(dp, addr, vid, db);
}
net: dsa: support FDB events on offloaded LAG interfaces This change introduces support for installing static FDB entries towards a bridge port that is a LAG of multiple DSA switch ports, as well as support for filtering towards the CPU local FDB entries emitted for LAG interfaces that are bridge ports. Conceptually, host addresses on LAG ports are identical to what we do for plain bridge ports. Whereas FDB entries _towards_ a LAG can't simply be replicated towards all member ports like we do for multicast, or VLAN. Instead we need new driver API. Hardware usually considers a LAG to be a "logical port", and sets the entire LAG as the forwarding destination. The physical egress port selection within the LAG is made by hashing policy, as usual. To represent the logical port corresponding to the LAG, we pass by value a copy of the dsa_lag structure to all switches in the tree that have at least one port in that LAG. To illustrate why a refcounted list of FDB entries is needed in struct dsa_lag, it is enough to say that: - a LAG may be a bridge port and may therefore receive FDB events even while it isn't yet offloaded by any DSA interface - DSA interfaces may be removed from a LAG while that is a bridge port; we don't want FDB entries lingering around, but we don't want to remove entries that are still in use, either For all the cases below to work, the idea is to always keep an FDB entry on a LAG with a reference count equal to the DSA member ports. So: - if a port joins a LAG, it requests the bridge to replay the FDB, and the FDB entries get created, or their refcount gets bumped by one - if a port leaves a LAG, the FDB replay deletes or decrements refcount by one - if an FDB is installed towards a LAG with ports already present, that entry is created (if it doesn't exist) and its refcount is bumped by the amount of ports already present in the LAG echo "Adding FDB entry to bond with existing ports" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond, then removing ports one by one" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link set swp1 nomaster ip link set swp2 nomaster ip link del br0 ip link del bond0 Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:53 +00:00
int dsa_port_lag_fdb_add(struct dsa_port *dp, const unsigned char *addr,
u16 vid)
{
struct dsa_notifier_lag_fdb_info info = {
.lag = dp->lag,
.addr = addr,
.vid = vid,
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
.db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
},
net: dsa: support FDB events on offloaded LAG interfaces This change introduces support for installing static FDB entries towards a bridge port that is a LAG of multiple DSA switch ports, as well as support for filtering towards the CPU local FDB entries emitted for LAG interfaces that are bridge ports. Conceptually, host addresses on LAG ports are identical to what we do for plain bridge ports. Whereas FDB entries _towards_ a LAG can't simply be replicated towards all member ports like we do for multicast, or VLAN. Instead we need new driver API. Hardware usually considers a LAG to be a "logical port", and sets the entire LAG as the forwarding destination. The physical egress port selection within the LAG is made by hashing policy, as usual. To represent the logical port corresponding to the LAG, we pass by value a copy of the dsa_lag structure to all switches in the tree that have at least one port in that LAG. To illustrate why a refcounted list of FDB entries is needed in struct dsa_lag, it is enough to say that: - a LAG may be a bridge port and may therefore receive FDB events even while it isn't yet offloaded by any DSA interface - DSA interfaces may be removed from a LAG while that is a bridge port; we don't want FDB entries lingering around, but we don't want to remove entries that are still in use, either For all the cases below to work, the idea is to always keep an FDB entry on a LAG with a reference count equal to the DSA member ports. So: - if a port joins a LAG, it requests the bridge to replay the FDB, and the FDB entries get created, or their refcount gets bumped by one - if a port leaves a LAG, the FDB replay deletes or decrements refcount by one - if an FDB is installed towards a LAG with ports already present, that entry is created (if it doesn't exist) and its refcount is bumped by the amount of ports already present in the LAG echo "Adding FDB entry to bond with existing ports" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond, then removing ports one by one" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link set swp1 nomaster ip link set swp2 nomaster ip link del br0 ip link del bond0 Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:53 +00:00
};
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
if (!dp->ds->fdb_isolation)
info.db.bridge.num = 0;
net: dsa: support FDB events on offloaded LAG interfaces This change introduces support for installing static FDB entries towards a bridge port that is a LAG of multiple DSA switch ports, as well as support for filtering towards the CPU local FDB entries emitted for LAG interfaces that are bridge ports. Conceptually, host addresses on LAG ports are identical to what we do for plain bridge ports. Whereas FDB entries _towards_ a LAG can't simply be replicated towards all member ports like we do for multicast, or VLAN. Instead we need new driver API. Hardware usually considers a LAG to be a "logical port", and sets the entire LAG as the forwarding destination. The physical egress port selection within the LAG is made by hashing policy, as usual. To represent the logical port corresponding to the LAG, we pass by value a copy of the dsa_lag structure to all switches in the tree that have at least one port in that LAG. To illustrate why a refcounted list of FDB entries is needed in struct dsa_lag, it is enough to say that: - a LAG may be a bridge port and may therefore receive FDB events even while it isn't yet offloaded by any DSA interface - DSA interfaces may be removed from a LAG while that is a bridge port; we don't want FDB entries lingering around, but we don't want to remove entries that are still in use, either For all the cases below to work, the idea is to always keep an FDB entry on a LAG with a reference count equal to the DSA member ports. So: - if a port joins a LAG, it requests the bridge to replay the FDB, and the FDB entries get created, or their refcount gets bumped by one - if a port leaves a LAG, the FDB replay deletes or decrements refcount by one - if an FDB is installed towards a LAG with ports already present, that entry is created (if it doesn't exist) and its refcount is bumped by the amount of ports already present in the LAG echo "Adding FDB entry to bond with existing ports" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond, then removing ports one by one" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link set swp1 nomaster ip link set swp2 nomaster ip link del br0 ip link del bond0 Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:53 +00:00
return dsa_port_notify(dp, DSA_NOTIFIER_LAG_FDB_ADD, &info);
}
int dsa_port_lag_fdb_del(struct dsa_port *dp, const unsigned char *addr,
u16 vid)
{
struct dsa_notifier_lag_fdb_info info = {
.lag = dp->lag,
.addr = addr,
.vid = vid,
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
.db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
},
net: dsa: support FDB events on offloaded LAG interfaces This change introduces support for installing static FDB entries towards a bridge port that is a LAG of multiple DSA switch ports, as well as support for filtering towards the CPU local FDB entries emitted for LAG interfaces that are bridge ports. Conceptually, host addresses on LAG ports are identical to what we do for plain bridge ports. Whereas FDB entries _towards_ a LAG can't simply be replicated towards all member ports like we do for multicast, or VLAN. Instead we need new driver API. Hardware usually considers a LAG to be a "logical port", and sets the entire LAG as the forwarding destination. The physical egress port selection within the LAG is made by hashing policy, as usual. To represent the logical port corresponding to the LAG, we pass by value a copy of the dsa_lag structure to all switches in the tree that have at least one port in that LAG. To illustrate why a refcounted list of FDB entries is needed in struct dsa_lag, it is enough to say that: - a LAG may be a bridge port and may therefore receive FDB events even while it isn't yet offloaded by any DSA interface - DSA interfaces may be removed from a LAG while that is a bridge port; we don't want FDB entries lingering around, but we don't want to remove entries that are still in use, either For all the cases below to work, the idea is to always keep an FDB entry on a LAG with a reference count equal to the DSA member ports. So: - if a port joins a LAG, it requests the bridge to replay the FDB, and the FDB entries get created, or their refcount gets bumped by one - if a port leaves a LAG, the FDB replay deletes or decrements refcount by one - if an FDB is installed towards a LAG with ports already present, that entry is created (if it doesn't exist) and its refcount is bumped by the amount of ports already present in the LAG echo "Adding FDB entry to bond with existing ports" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond, then removing ports one by one" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link set swp1 nomaster ip link set swp2 nomaster ip link del br0 ip link del bond0 Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:53 +00:00
};
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
if (!dp->ds->fdb_isolation)
info.db.bridge.num = 0;
net: dsa: support FDB events on offloaded LAG interfaces This change introduces support for installing static FDB entries towards a bridge port that is a LAG of multiple DSA switch ports, as well as support for filtering towards the CPU local FDB entries emitted for LAG interfaces that are bridge ports. Conceptually, host addresses on LAG ports are identical to what we do for plain bridge ports. Whereas FDB entries _towards_ a LAG can't simply be replicated towards all member ports like we do for multicast, or VLAN. Instead we need new driver API. Hardware usually considers a LAG to be a "logical port", and sets the entire LAG as the forwarding destination. The physical egress port selection within the LAG is made by hashing policy, as usual. To represent the logical port corresponding to the LAG, we pass by value a copy of the dsa_lag structure to all switches in the tree that have at least one port in that LAG. To illustrate why a refcounted list of FDB entries is needed in struct dsa_lag, it is enough to say that: - a LAG may be a bridge port and may therefore receive FDB events even while it isn't yet offloaded by any DSA interface - DSA interfaces may be removed from a LAG while that is a bridge port; we don't want FDB entries lingering around, but we don't want to remove entries that are still in use, either For all the cases below to work, the idea is to always keep an FDB entry on a LAG with a reference count equal to the DSA member ports. So: - if a port joins a LAG, it requests the bridge to replay the FDB, and the FDB entries get created, or their refcount gets bumped by one - if a port leaves a LAG, the FDB replay deletes or decrements refcount by one - if an FDB is installed towards a LAG with ports already present, that entry is created (if it doesn't exist) and its refcount is bumped by the amount of ports already present in the LAG echo "Adding FDB entry to bond with existing ports" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link del br0 ip link del bond0 echo "Adding FDB entry to empty bond, then removing ports one by one" ip link del bond0 ip link add bond0 type bond mode 802.3ad ip link del br0 ip link add br0 type bridge ip link set bond0 master br0 bridge fdb add dev bond0 00:01:02:03:04:05 master static ip link set swp1 down && ip link set swp1 master bond0 && ip link set swp1 up ip link set swp2 down && ip link set swp2 master bond0 && ip link set swp2 up ip link set swp1 nomaster ip link set swp2 nomaster ip link del br0 ip link del bond0 Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-02-23 14:00:53 +00:00
return dsa_port_notify(dp, DSA_NOTIFIER_LAG_FDB_DEL, &info);
}
int dsa_port_fdb_dump(struct dsa_port *dp, dsa_fdb_dump_cb_t *cb, void *data)
{
struct dsa_switch *ds = dp->ds;
int port = dp->index;
if (!ds->ops->port_fdb_dump)
return -EOPNOTSUPP;
return ds->ops->port_fdb_dump(ds, port, cb, data);
}
int dsa_port_mdb_add(const struct dsa_port *dp,
net: switchdev: remove the transaction structure from port object notifiers Since the introduction of the switchdev API, port objects were transmitted to drivers for offloading using a two-step transactional model, with a prepare phase that was supposed to catch all errors, and a commit phase that was supposed to never fail. Some classes of failures can never be avoided, like hardware access, or memory allocation. In the latter case, merely attempting to move the memory allocation to the preparation phase makes it impossible to avoid memory leaks, since commit 91cf8eceffc1 ("switchdev: Remove unused transaction item queue") which has removed the unused mechanism of passing on the allocated memory between one phase and another. It is time we admit that separating the preparation from the commit phase is something that is best left for the driver to decide, and not something that should be baked into the API, especially since there are no switchdev callers that depend on this. This patch removes the struct switchdev_trans member from switchdev port object notifier structures, and converts drivers to not look at this member. Where driver conversion is trivial (like in the case of the Marvell Prestera driver, NXP DPAA2 switch, TI CPSW, and Rocker drivers), it is done in this patch. Where driver conversion needs more attention (DSA, Mellanox Spectrum), the conversion is left for subsequent patches and here we only fake the prepare/commit phases at a lower level, just not in the switchdev notifier itself. Where the code has a natural structure that is best left alone as a preparation and a commit phase (as in the case of the Ocelot switch), that structure is left in place, just made to not depend upon the switchdev transactional model. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Jiri Pirko <jiri@nvidia.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-09 00:01:48 +00:00
const struct switchdev_obj_port_mdb *mdb)
{
struct dsa_notifier_mdb_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.mdb = mdb,
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
.db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
},
};
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
if (!dp->ds->fdb_isolation)
info.db.bridge.num = 0;
return dsa_port_notify(dp, DSA_NOTIFIER_MDB_ADD, &info);
}
int dsa_port_mdb_del(const struct dsa_port *dp,
const struct switchdev_obj_port_mdb *mdb)
{
struct dsa_notifier_mdb_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.mdb = mdb,
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
.db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
},
};
net: dsa: request drivers to perform FDB isolation For DSA, to encourage drivers to perform FDB isolation simply means to track which bridge does each FDB and MDB entry belong to. It then becomes the driver responsibility to use something that makes the FDB entry from one bridge not match the FDB lookup of ports from other bridges. The top-level functions where the bridge is determined are: - dsa_port_fdb_{add,del} - dsa_port_host_fdb_{add,del} - dsa_port_mdb_{add,del} - dsa_port_host_mdb_{add,del} aka the pre-crosschip-notifier functions. Changing the API to pass a reference to a bridge is not superfluous, and looking at the passed bridge argument is not the same as having the driver look at dsa_to_port(ds, port)->bridge from the ->port_fdb_add() method. DSA installs FDB and MDB entries on shared (CPU and DSA) ports as well, and those do not have any dp->bridge information to retrieve, because they are not in any bridge - they are merely the pipes that serve the user ports that are in one or multiple bridges. The struct dsa_bridge associated with each FDB/MDB entry is encapsulated in a larger "struct dsa_db" database. Although only databases associated to bridges are notified for now, this API will be the starting point for implementing IFF_UNICAST_FLT in DSA. There, the idea is to install FDB entries on the CPU port which belong to the corresponding user port's port database. These are supposed to match only when the port is standalone. It is better to introduce the API in its expected final form than to introduce it for bridges first, then to have to change drivers which may have made one or more assumptions. Drivers can use the provided bridge.num, but they can also use a different numbering scheme that is more convenient. DSA must perform refcounting on the CPU and DSA ports by also taking into account the bridge number. So if two bridges request the same local address, DSA must notify the driver twice, once for each bridge. In fact, if the driver supports FDB isolation, DSA must perform refcounting per bridge, but if the driver doesn't, DSA must refcount host addresses across all bridges, otherwise it would be telling the driver to delete an FDB entry for a bridge and the driver would delete it for all bridges. So introduce a bool fdb_isolation in drivers which would make all bridge databases passed to the cross-chip notifier have the same number (0). This makes dsa_mac_addr_find() -> dsa_db_equal() say that all bridge databases are the same database - which is essentially the legacy behavior. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-25 09:22:22 +00:00
if (!dp->ds->fdb_isolation)
info.db.bridge.num = 0;
return dsa_port_notify(dp, DSA_NOTIFIER_MDB_DEL, &info);
}
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
static int dsa_port_host_mdb_add(const struct dsa_port *dp,
const struct switchdev_obj_port_mdb *mdb,
struct dsa_db db)
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
{
struct dsa_notifier_mdb_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
.mdb = mdb,
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
.db = db,
};
return dsa_port_notify(dp, DSA_NOTIFIER_HOST_MDB_ADD, &info);
}
int dsa_port_standalone_host_mdb_add(const struct dsa_port *dp,
const struct switchdev_obj_port_mdb *mdb)
{
struct dsa_db db = {
.type = DSA_DB_PORT,
.dp = dp,
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
};
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
return dsa_port_host_mdb_add(dp, mdb, db);
}
int dsa_port_bridge_host_mdb_add(const struct dsa_port *dp,
const struct switchdev_obj_port_mdb *mdb)
{
struct net_device *conduit = dsa_port_to_conduit(dp);
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
struct dsa_db db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
};
int err;
net: dsa: fix db type confusion in host fdb/mdb add/del We have the following code paths: Host FDB (unicast RX filtering): dsa_port_standalone_host_fdb_add() dsa_port_bridge_host_fdb_add() | | +--------------+ +------------+ | | v v dsa_port_host_fdb_add() dsa_port_standalone_host_fdb_del() dsa_port_bridge_host_fdb_del() | | +--------------+ +------------+ | | v v dsa_port_host_fdb_del() Host MDB (multicast RX filtering): dsa_port_standalone_host_mdb_add() dsa_port_bridge_host_mdb_add() | | +--------------+ +------------+ | | v v dsa_port_host_mdb_add() dsa_port_standalone_host_mdb_del() dsa_port_bridge_host_mdb_del() | | +--------------+ +------------+ | | v v dsa_port_host_mdb_del() The logic added by commit 5e8a1e03aa4d ("net: dsa: install secondary unicast and multicast addresses as host FDB/MDB") zeroes out db.bridge.num if the switch doesn't support ds->fdb_isolation (the majority doesn't). This is done for a reason explained in commit c26933639b54 ("net: dsa: request drivers to perform FDB isolation"). Taking a single code path as example - dsa_port_host_fdb_add() - the others are similar - the problem is that this function handles: - DSA_DB_PORT databases, when called from dsa_port_standalone_host_fdb_add() - DSA_DB_BRIDGE databases, when called from dsa_port_bridge_host_fdb_add() So, if dsa_port_host_fdb_add() were to make any change on the "bridge.num" attribute of the database, this would only be correct for a DSA_DB_BRIDGE, and a type confusion for a DSA_DB_PORT bridge. However, this bug is without consequences, for 2 reasons: - dsa_port_standalone_host_fdb_add() is only called from code which is (in)directly guarded by dsa_switch_supports_uc_filtering(ds), and that function only returns true if ds->fdb_isolation is set. So, the code only executed for DSA_DB_BRIDGE databases. - Even if the code was not dead for DSA_DB_PORT, we have the following memory layout: struct dsa_bridge { struct net_device *dev; unsigned int num; bool tx_fwd_offload; refcount_t refcount; }; struct dsa_db { enum dsa_db_type type; union { const struct dsa_port *dp; // DSA_DB_PORT struct dsa_lag lag; struct dsa_bridge bridge; // DSA_DB_BRIDGE }; }; So, the zeroization of dsa_db :: bridge :: num on a dsa_db structure of type DSA_DB_PORT would access memory which is unused, because we only use dsa_db :: dp for DSA_DB_PORT, and this is mapped at the same address with dsa_db :: dev for DSA_DB_BRIDGE, thanks to the union definition. It is correct to fix up dsa_db :: bridge :: num only from code paths that come from the bridge / switchdev, so move these there. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Simon Horman <simon.horman@corigine.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Link: https://lore.kernel.org/r/20230329133819.697642-1-vladimir.oltean@nxp.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2023-03-29 13:38:19 +00:00
if (!dp->ds->fdb_isolation)
db.bridge.num = 0;
err = dev_mc_add(conduit, mdb->addr);
if (err)
return err;
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
return dsa_port_host_mdb_add(dp, mdb, db);
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
}
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
static int dsa_port_host_mdb_del(const struct dsa_port *dp,
const struct switchdev_obj_port_mdb *mdb,
struct dsa_db db)
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
{
struct dsa_notifier_mdb_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
.mdb = mdb,
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
.db = db,
};
return dsa_port_notify(dp, DSA_NOTIFIER_HOST_MDB_DEL, &info);
}
int dsa_port_standalone_host_mdb_del(const struct dsa_port *dp,
const struct switchdev_obj_port_mdb *mdb)
{
struct dsa_db db = {
.type = DSA_DB_PORT,
.dp = dp,
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
};
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
return dsa_port_host_mdb_del(dp, mdb, db);
}
int dsa_port_bridge_host_mdb_del(const struct dsa_port *dp,
const struct switchdev_obj_port_mdb *mdb)
{
struct net_device *conduit = dsa_port_to_conduit(dp);
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
struct dsa_db db = {
.type = DSA_DB_BRIDGE,
.bridge = *dp->bridge,
};
int err;
net: dsa: fix db type confusion in host fdb/mdb add/del We have the following code paths: Host FDB (unicast RX filtering): dsa_port_standalone_host_fdb_add() dsa_port_bridge_host_fdb_add() | | +--------------+ +------------+ | | v v dsa_port_host_fdb_add() dsa_port_standalone_host_fdb_del() dsa_port_bridge_host_fdb_del() | | +--------------+ +------------+ | | v v dsa_port_host_fdb_del() Host MDB (multicast RX filtering): dsa_port_standalone_host_mdb_add() dsa_port_bridge_host_mdb_add() | | +--------------+ +------------+ | | v v dsa_port_host_mdb_add() dsa_port_standalone_host_mdb_del() dsa_port_bridge_host_mdb_del() | | +--------------+ +------------+ | | v v dsa_port_host_mdb_del() The logic added by commit 5e8a1e03aa4d ("net: dsa: install secondary unicast and multicast addresses as host FDB/MDB") zeroes out db.bridge.num if the switch doesn't support ds->fdb_isolation (the majority doesn't). This is done for a reason explained in commit c26933639b54 ("net: dsa: request drivers to perform FDB isolation"). Taking a single code path as example - dsa_port_host_fdb_add() - the others are similar - the problem is that this function handles: - DSA_DB_PORT databases, when called from dsa_port_standalone_host_fdb_add() - DSA_DB_BRIDGE databases, when called from dsa_port_bridge_host_fdb_add() So, if dsa_port_host_fdb_add() were to make any change on the "bridge.num" attribute of the database, this would only be correct for a DSA_DB_BRIDGE, and a type confusion for a DSA_DB_PORT bridge. However, this bug is without consequences, for 2 reasons: - dsa_port_standalone_host_fdb_add() is only called from code which is (in)directly guarded by dsa_switch_supports_uc_filtering(ds), and that function only returns true if ds->fdb_isolation is set. So, the code only executed for DSA_DB_BRIDGE databases. - Even if the code was not dead for DSA_DB_PORT, we have the following memory layout: struct dsa_bridge { struct net_device *dev; unsigned int num; bool tx_fwd_offload; refcount_t refcount; }; struct dsa_db { enum dsa_db_type type; union { const struct dsa_port *dp; // DSA_DB_PORT struct dsa_lag lag; struct dsa_bridge bridge; // DSA_DB_BRIDGE }; }; So, the zeroization of dsa_db :: bridge :: num on a dsa_db structure of type DSA_DB_PORT would access memory which is unused, because we only use dsa_db :: dp for DSA_DB_PORT, and this is mapped at the same address with dsa_db :: dev for DSA_DB_BRIDGE, thanks to the union definition. It is correct to fix up dsa_db :: bridge :: num only from code paths that come from the bridge / switchdev, so move these there. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Simon Horman <simon.horman@corigine.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Link: https://lore.kernel.org/r/20230329133819.697642-1-vladimir.oltean@nxp.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2023-03-29 13:38:19 +00:00
if (!dp->ds->fdb_isolation)
db.bridge.num = 0;
err = dev_mc_del(conduit, mdb->addr);
if (err)
return err;
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
net: dsa: install secondary unicast and multicast addresses as host FDB/MDB In preparation of disabling flooding towards the CPU in standalone ports mode, identify the addresses requested by upper interfaces and use the new API for DSA FDB isolation to request the hardware driver to offload these as FDB or MDB objects. The objects belong to the user port's database, and are installed pointing towards the CPU port. Because dev_uc_add()/dev_mc_add() is VLAN-unaware, we offload to the port standalone database addresses with VID 0 (also VLAN-unaware). So this excludes switches with global VLAN filtering from supporting unicast filtering, because there, it is possible for a port of a switch to join a VLAN-aware bridge, and this changes the VLAN awareness of standalone ports, requiring VLAN-aware standalone host FDB entries. For the same reason, hellcreek, which requires VLAN awareness in standalone mode, is also exempted from unicast filtering. We create "standalone" variants of dsa_port_host_fdb_add() and dsa_port_host_mdb_add() (and the _del coresponding functions). We also create a separate work item type for handling deferred standalone host FDB/MDB entries compared to the switchdev one. This is done for the purpose of clarity - the procedure for offloading a bridge FDB entry is different than offloading a standalone one, and the switchdev event work handles only FDBs anyway, not MDBs. Deferral is needed for standalone entries because ndo_set_rx_mode runs in atomic context. We could probably optimize things a little by first queuing up all entries that need to be offloaded, and scheduling the work item just once, but the data structures that we can pass through __dev_uc_sync() and __dev_mc_sync() are limiting (there is nothing like a void *priv), so we'd have to keep the list of queued events somewhere in struct dsa_switch, and possibly a lock for it. Too complicated for now. Adding the address to the master is handled by dev_uc_sync(), adding it to the hardware is handled by __dev_uc_sync(). So this is the reason why dsa_port_standalone_host_fdb_add() does not call dev_uc_add(). Not that it had the rtnl_mutex anyway - ndo_set_rx_mode has it, but is atomic. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-03-02 19:14:10 +00:00
return dsa_port_host_mdb_del(dp, mdb, db);
net: dsa: introduce a separate cross-chip notifier type for host MDBs Commit abd49535c380 ("net: dsa: execute dsa_switch_mdb_add only for routing port in cross-chip topologies") does a surprisingly good job even for the SWITCHDEV_OBJ_ID_HOST_MDB use case, where DSA simply translates a switchdev object received on dp into a cross-chip notifier for dp->cpu_dp. To visualize how that works, imagine the daisy chain topology below and consider a SWITCHDEV_OBJ_ID_HOST_MDB object emitted on sw2p0. How does the cross-chip notifier know to match on all the right ports (sw0p4, the dedicated CPU port, sw1p4, an upstream DSA link, and sw2p4, another upstream DSA link)? | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ user ] [ dsa ] [ ] [ ] [ ] [ ] [ x ] The answer is simple: the dedicated CPU port of sw2p0 is sw0p4, and dsa_routing_port returns the upstream port for all switches. That is fine, but there are other topologies where this does not work as well. There are trees with "H" topologies in the wild, where there are 2 or more switches with DSA links between them, but every switch has its dedicated CPU port. For these topologies, it seems stupid for the neighbor switches to install an MDB entry on the routing port, since these multicast addresses are fundamentally different than the usual ones we support (and that is the justification for this patch, to introduce the concept of a termination plane multicast MAC address, as opposed to a forwarding plane multicast MAC address). For example, when a SWITCHDEV_OBJ_ID_HOST_MDB would get added to sw0p0, without this patch, it would get treated as a regular port MDB on sw0p2 and it would match on the ports below (including the sw1p3 routing port). | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ x ] [ ] [ ] [ ] With the patch, the host MDB notifier on sw0p0 matches only on the local switch, which is what we want for a termination plane address. | | sw0p0 sw0p1 sw0p2 sw0p3 sw1p3 sw1p2 sw1p1 sw1p0 [ user ] [ user ] [ cpu ] [ dsa ] [ dsa ] [ cpu ] [ user ] [ user ] [ ] [ ] [ x ] [ ] ---- [ ] [ ] [ ] [ ] Name this new matching function "dsa_switch_host_address_match" since we will be reusing it soon for host FDB entries as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-06-29 14:06:49 +00:00
}
int dsa_port_vlan_add(struct dsa_port *dp,
const struct switchdev_obj_port_vlan *vlan,
struct netlink_ext_ack *extack)
{
struct dsa_notifier_vlan_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.vlan = vlan,
.extack = extack,
};
return dsa_port_notify(dp, DSA_NOTIFIER_VLAN_ADD, &info);
}
int dsa_port_vlan_del(struct dsa_port *dp,
const struct switchdev_obj_port_vlan *vlan)
{
struct dsa_notifier_vlan_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
.vlan = vlan,
};
return dsa_port_notify(dp, DSA_NOTIFIER_VLAN_DEL, &info);
}
net: dsa: add explicit support for host bridge VLANs Currently, DSA programs VLANs on shared (DSA and CPU) ports each time it does so on user ports. This is good for basic functionality but has several limitations: - the VLAN group which must reach the CPU may be radically different from the VLAN group that must be autonomously forwarded by the switch. In other words, the admin may want to isolate noisy stations and avoid traffic from them going to the control processor of the switch, where it would just waste useless cycles. The bridge already supports independent control of VLAN groups on bridge ports and on the bridge itself, and when VLAN-aware, it will drop packets in software anyway if their VID isn't added as a 'self' entry towards the bridge device. - Replaying host FDB entries may depend, for some drivers like mv88e6xxx, on replaying the host VLANs as well. The 2 VLAN groups are approximately the same in most regular cases, but there are corner cases when timing matters, and DSA's approximation of replicating VLANs on shared ports simply does not work. - If a user makes the bridge (implicitly the CPU port) join a VLAN by accident, there is no way for the CPU port to isolate itself from that noisy VLAN except by rebooting the system. This is because for each VLAN added on a user port, DSA will add it on shared ports too, but for each VLAN deletion on a user port, it will remain installed on shared ports, since DSA has no good indication of whether the VLAN is still in use or not. Now that the bridge driver emits well-balanced SWITCHDEV_OBJ_ID_PORT_VLAN addition and removal events, DSA has a simple and straightforward task of separating the bridge port VLANs (these have an orig_dev which is a DSA slave interface, or a LAG interface) from the host VLANs (these have an orig_dev which is a bridge interface), and to keep a simple reference count of each VID on each shared port. Forwarding VLANs must be installed on the bridge ports and on all DSA ports interconnecting them. We don't have a good view of the exact topology, so we simply install forwarding VLANs on all DSA ports, which is what has been done until now. Host VLANs must be installed primarily on the dedicated CPU port of each bridge port. More subtly, they must also be installed on upstream-facing and downstream-facing DSA ports that are connecting the bridge ports and the CPU. This ensures that the mv88e6xxx's problem (VID of host FDB entry may be absent from VTU) is still addressed even if that switch is in a cross-chip setup, and it has no local CPU port. Therefore: - user ports contain only bridge port (forwarding) VLANs, and no refcounting is necessary - DSA ports contain both forwarding and host VLANs. Refcounting is necessary among these 2 types. - CPU ports contain only host VLANs. Refcounting is also necessary. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-15 17:02:17 +00:00
int dsa_port_host_vlan_add(struct dsa_port *dp,
const struct switchdev_obj_port_vlan *vlan,
struct netlink_ext_ack *extack)
{
struct net_device *conduit = dsa_port_to_conduit(dp);
net: dsa: add explicit support for host bridge VLANs Currently, DSA programs VLANs on shared (DSA and CPU) ports each time it does so on user ports. This is good for basic functionality but has several limitations: - the VLAN group which must reach the CPU may be radically different from the VLAN group that must be autonomously forwarded by the switch. In other words, the admin may want to isolate noisy stations and avoid traffic from them going to the control processor of the switch, where it would just waste useless cycles. The bridge already supports independent control of VLAN groups on bridge ports and on the bridge itself, and when VLAN-aware, it will drop packets in software anyway if their VID isn't added as a 'self' entry towards the bridge device. - Replaying host FDB entries may depend, for some drivers like mv88e6xxx, on replaying the host VLANs as well. The 2 VLAN groups are approximately the same in most regular cases, but there are corner cases when timing matters, and DSA's approximation of replicating VLANs on shared ports simply does not work. - If a user makes the bridge (implicitly the CPU port) join a VLAN by accident, there is no way for the CPU port to isolate itself from that noisy VLAN except by rebooting the system. This is because for each VLAN added on a user port, DSA will add it on shared ports too, but for each VLAN deletion on a user port, it will remain installed on shared ports, since DSA has no good indication of whether the VLAN is still in use or not. Now that the bridge driver emits well-balanced SWITCHDEV_OBJ_ID_PORT_VLAN addition and removal events, DSA has a simple and straightforward task of separating the bridge port VLANs (these have an orig_dev which is a DSA slave interface, or a LAG interface) from the host VLANs (these have an orig_dev which is a bridge interface), and to keep a simple reference count of each VID on each shared port. Forwarding VLANs must be installed on the bridge ports and on all DSA ports interconnecting them. We don't have a good view of the exact topology, so we simply install forwarding VLANs on all DSA ports, which is what has been done until now. Host VLANs must be installed primarily on the dedicated CPU port of each bridge port. More subtly, they must also be installed on upstream-facing and downstream-facing DSA ports that are connecting the bridge ports and the CPU. This ensures that the mv88e6xxx's problem (VID of host FDB entry may be absent from VTU) is still addressed even if that switch is in a cross-chip setup, and it has no local CPU port. Therefore: - user ports contain only bridge port (forwarding) VLANs, and no refcounting is necessary - DSA ports contain both forwarding and host VLANs. Refcounting is necessary among these 2 types. - CPU ports contain only host VLANs. Refcounting is also necessary. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-15 17:02:17 +00:00
struct dsa_notifier_vlan_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
net: dsa: add explicit support for host bridge VLANs Currently, DSA programs VLANs on shared (DSA and CPU) ports each time it does so on user ports. This is good for basic functionality but has several limitations: - the VLAN group which must reach the CPU may be radically different from the VLAN group that must be autonomously forwarded by the switch. In other words, the admin may want to isolate noisy stations and avoid traffic from them going to the control processor of the switch, where it would just waste useless cycles. The bridge already supports independent control of VLAN groups on bridge ports and on the bridge itself, and when VLAN-aware, it will drop packets in software anyway if their VID isn't added as a 'self' entry towards the bridge device. - Replaying host FDB entries may depend, for some drivers like mv88e6xxx, on replaying the host VLANs as well. The 2 VLAN groups are approximately the same in most regular cases, but there are corner cases when timing matters, and DSA's approximation of replicating VLANs on shared ports simply does not work. - If a user makes the bridge (implicitly the CPU port) join a VLAN by accident, there is no way for the CPU port to isolate itself from that noisy VLAN except by rebooting the system. This is because for each VLAN added on a user port, DSA will add it on shared ports too, but for each VLAN deletion on a user port, it will remain installed on shared ports, since DSA has no good indication of whether the VLAN is still in use or not. Now that the bridge driver emits well-balanced SWITCHDEV_OBJ_ID_PORT_VLAN addition and removal events, DSA has a simple and straightforward task of separating the bridge port VLANs (these have an orig_dev which is a DSA slave interface, or a LAG interface) from the host VLANs (these have an orig_dev which is a bridge interface), and to keep a simple reference count of each VID on each shared port. Forwarding VLANs must be installed on the bridge ports and on all DSA ports interconnecting them. We don't have a good view of the exact topology, so we simply install forwarding VLANs on all DSA ports, which is what has been done until now. Host VLANs must be installed primarily on the dedicated CPU port of each bridge port. More subtly, they must also be installed on upstream-facing and downstream-facing DSA ports that are connecting the bridge ports and the CPU. This ensures that the mv88e6xxx's problem (VID of host FDB entry may be absent from VTU) is still addressed even if that switch is in a cross-chip setup, and it has no local CPU port. Therefore: - user ports contain only bridge port (forwarding) VLANs, and no refcounting is necessary - DSA ports contain both forwarding and host VLANs. Refcounting is necessary among these 2 types. - CPU ports contain only host VLANs. Refcounting is also necessary. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-15 17:02:17 +00:00
.vlan = vlan,
.extack = extack,
};
int err;
err = dsa_port_notify(dp, DSA_NOTIFIER_HOST_VLAN_ADD, &info);
if (err && err != -EOPNOTSUPP)
return err;
vlan_vid_add(conduit, htons(ETH_P_8021Q), vlan->vid);
net: dsa: add explicit support for host bridge VLANs Currently, DSA programs VLANs on shared (DSA and CPU) ports each time it does so on user ports. This is good for basic functionality but has several limitations: - the VLAN group which must reach the CPU may be radically different from the VLAN group that must be autonomously forwarded by the switch. In other words, the admin may want to isolate noisy stations and avoid traffic from them going to the control processor of the switch, where it would just waste useless cycles. The bridge already supports independent control of VLAN groups on bridge ports and on the bridge itself, and when VLAN-aware, it will drop packets in software anyway if their VID isn't added as a 'self' entry towards the bridge device. - Replaying host FDB entries may depend, for some drivers like mv88e6xxx, on replaying the host VLANs as well. The 2 VLAN groups are approximately the same in most regular cases, but there are corner cases when timing matters, and DSA's approximation of replicating VLANs on shared ports simply does not work. - If a user makes the bridge (implicitly the CPU port) join a VLAN by accident, there is no way for the CPU port to isolate itself from that noisy VLAN except by rebooting the system. This is because for each VLAN added on a user port, DSA will add it on shared ports too, but for each VLAN deletion on a user port, it will remain installed on shared ports, since DSA has no good indication of whether the VLAN is still in use or not. Now that the bridge driver emits well-balanced SWITCHDEV_OBJ_ID_PORT_VLAN addition and removal events, DSA has a simple and straightforward task of separating the bridge port VLANs (these have an orig_dev which is a DSA slave interface, or a LAG interface) from the host VLANs (these have an orig_dev which is a bridge interface), and to keep a simple reference count of each VID on each shared port. Forwarding VLANs must be installed on the bridge ports and on all DSA ports interconnecting them. We don't have a good view of the exact topology, so we simply install forwarding VLANs on all DSA ports, which is what has been done until now. Host VLANs must be installed primarily on the dedicated CPU port of each bridge port. More subtly, they must also be installed on upstream-facing and downstream-facing DSA ports that are connecting the bridge ports and the CPU. This ensures that the mv88e6xxx's problem (VID of host FDB entry may be absent from VTU) is still addressed even if that switch is in a cross-chip setup, and it has no local CPU port. Therefore: - user ports contain only bridge port (forwarding) VLANs, and no refcounting is necessary - DSA ports contain both forwarding and host VLANs. Refcounting is necessary among these 2 types. - CPU ports contain only host VLANs. Refcounting is also necessary. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-15 17:02:17 +00:00
return err;
}
int dsa_port_host_vlan_del(struct dsa_port *dp,
const struct switchdev_obj_port_vlan *vlan)
{
struct net_device *conduit = dsa_port_to_conduit(dp);
net: dsa: add explicit support for host bridge VLANs Currently, DSA programs VLANs on shared (DSA and CPU) ports each time it does so on user ports. This is good for basic functionality but has several limitations: - the VLAN group which must reach the CPU may be radically different from the VLAN group that must be autonomously forwarded by the switch. In other words, the admin may want to isolate noisy stations and avoid traffic from them going to the control processor of the switch, where it would just waste useless cycles. The bridge already supports independent control of VLAN groups on bridge ports and on the bridge itself, and when VLAN-aware, it will drop packets in software anyway if their VID isn't added as a 'self' entry towards the bridge device. - Replaying host FDB entries may depend, for some drivers like mv88e6xxx, on replaying the host VLANs as well. The 2 VLAN groups are approximately the same in most regular cases, but there are corner cases when timing matters, and DSA's approximation of replicating VLANs on shared ports simply does not work. - If a user makes the bridge (implicitly the CPU port) join a VLAN by accident, there is no way for the CPU port to isolate itself from that noisy VLAN except by rebooting the system. This is because for each VLAN added on a user port, DSA will add it on shared ports too, but for each VLAN deletion on a user port, it will remain installed on shared ports, since DSA has no good indication of whether the VLAN is still in use or not. Now that the bridge driver emits well-balanced SWITCHDEV_OBJ_ID_PORT_VLAN addition and removal events, DSA has a simple and straightforward task of separating the bridge port VLANs (these have an orig_dev which is a DSA slave interface, or a LAG interface) from the host VLANs (these have an orig_dev which is a bridge interface), and to keep a simple reference count of each VID on each shared port. Forwarding VLANs must be installed on the bridge ports and on all DSA ports interconnecting them. We don't have a good view of the exact topology, so we simply install forwarding VLANs on all DSA ports, which is what has been done until now. Host VLANs must be installed primarily on the dedicated CPU port of each bridge port. More subtly, they must also be installed on upstream-facing and downstream-facing DSA ports that are connecting the bridge ports and the CPU. This ensures that the mv88e6xxx's problem (VID of host FDB entry may be absent from VTU) is still addressed even if that switch is in a cross-chip setup, and it has no local CPU port. Therefore: - user ports contain only bridge port (forwarding) VLANs, and no refcounting is necessary - DSA ports contain both forwarding and host VLANs. Refcounting is necessary among these 2 types. - CPU ports contain only host VLANs. Refcounting is also necessary. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-15 17:02:17 +00:00
struct dsa_notifier_vlan_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
net: dsa: add explicit support for host bridge VLANs Currently, DSA programs VLANs on shared (DSA and CPU) ports each time it does so on user ports. This is good for basic functionality but has several limitations: - the VLAN group which must reach the CPU may be radically different from the VLAN group that must be autonomously forwarded by the switch. In other words, the admin may want to isolate noisy stations and avoid traffic from them going to the control processor of the switch, where it would just waste useless cycles. The bridge already supports independent control of VLAN groups on bridge ports and on the bridge itself, and when VLAN-aware, it will drop packets in software anyway if their VID isn't added as a 'self' entry towards the bridge device. - Replaying host FDB entries may depend, for some drivers like mv88e6xxx, on replaying the host VLANs as well. The 2 VLAN groups are approximately the same in most regular cases, but there are corner cases when timing matters, and DSA's approximation of replicating VLANs on shared ports simply does not work. - If a user makes the bridge (implicitly the CPU port) join a VLAN by accident, there is no way for the CPU port to isolate itself from that noisy VLAN except by rebooting the system. This is because for each VLAN added on a user port, DSA will add it on shared ports too, but for each VLAN deletion on a user port, it will remain installed on shared ports, since DSA has no good indication of whether the VLAN is still in use or not. Now that the bridge driver emits well-balanced SWITCHDEV_OBJ_ID_PORT_VLAN addition and removal events, DSA has a simple and straightforward task of separating the bridge port VLANs (these have an orig_dev which is a DSA slave interface, or a LAG interface) from the host VLANs (these have an orig_dev which is a bridge interface), and to keep a simple reference count of each VID on each shared port. Forwarding VLANs must be installed on the bridge ports and on all DSA ports interconnecting them. We don't have a good view of the exact topology, so we simply install forwarding VLANs on all DSA ports, which is what has been done until now. Host VLANs must be installed primarily on the dedicated CPU port of each bridge port. More subtly, they must also be installed on upstream-facing and downstream-facing DSA ports that are connecting the bridge ports and the CPU. This ensures that the mv88e6xxx's problem (VID of host FDB entry may be absent from VTU) is still addressed even if that switch is in a cross-chip setup, and it has no local CPU port. Therefore: - user ports contain only bridge port (forwarding) VLANs, and no refcounting is necessary - DSA ports contain both forwarding and host VLANs. Refcounting is necessary among these 2 types. - CPU ports contain only host VLANs. Refcounting is also necessary. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-15 17:02:17 +00:00
.vlan = vlan,
};
int err;
err = dsa_port_notify(dp, DSA_NOTIFIER_HOST_VLAN_DEL, &info);
if (err && err != -EOPNOTSUPP)
return err;
vlan_vid_del(conduit, htons(ETH_P_8021Q), vlan->vid);
net: dsa: add explicit support for host bridge VLANs Currently, DSA programs VLANs on shared (DSA and CPU) ports each time it does so on user ports. This is good for basic functionality but has several limitations: - the VLAN group which must reach the CPU may be radically different from the VLAN group that must be autonomously forwarded by the switch. In other words, the admin may want to isolate noisy stations and avoid traffic from them going to the control processor of the switch, where it would just waste useless cycles. The bridge already supports independent control of VLAN groups on bridge ports and on the bridge itself, and when VLAN-aware, it will drop packets in software anyway if their VID isn't added as a 'self' entry towards the bridge device. - Replaying host FDB entries may depend, for some drivers like mv88e6xxx, on replaying the host VLANs as well. The 2 VLAN groups are approximately the same in most regular cases, but there are corner cases when timing matters, and DSA's approximation of replicating VLANs on shared ports simply does not work. - If a user makes the bridge (implicitly the CPU port) join a VLAN by accident, there is no way for the CPU port to isolate itself from that noisy VLAN except by rebooting the system. This is because for each VLAN added on a user port, DSA will add it on shared ports too, but for each VLAN deletion on a user port, it will remain installed on shared ports, since DSA has no good indication of whether the VLAN is still in use or not. Now that the bridge driver emits well-balanced SWITCHDEV_OBJ_ID_PORT_VLAN addition and removal events, DSA has a simple and straightforward task of separating the bridge port VLANs (these have an orig_dev which is a DSA slave interface, or a LAG interface) from the host VLANs (these have an orig_dev which is a bridge interface), and to keep a simple reference count of each VID on each shared port. Forwarding VLANs must be installed on the bridge ports and on all DSA ports interconnecting them. We don't have a good view of the exact topology, so we simply install forwarding VLANs on all DSA ports, which is what has been done until now. Host VLANs must be installed primarily on the dedicated CPU port of each bridge port. More subtly, they must also be installed on upstream-facing and downstream-facing DSA ports that are connecting the bridge ports and the CPU. This ensures that the mv88e6xxx's problem (VID of host FDB entry may be absent from VTU) is still addressed even if that switch is in a cross-chip setup, and it has no local CPU port. Therefore: - user ports contain only bridge port (forwarding) VLANs, and no refcounting is necessary - DSA ports contain both forwarding and host VLANs. Refcounting is necessary among these 2 types. - CPU ports contain only host VLANs. Refcounting is also necessary. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-02-15 17:02:17 +00:00
return err;
}
int dsa_port_mrp_add(const struct dsa_port *dp,
const struct switchdev_obj_mrp *mrp)
{
struct dsa_switch *ds = dp->ds;
if (!ds->ops->port_mrp_add)
return -EOPNOTSUPP;
return ds->ops->port_mrp_add(ds, dp->index, mrp);
}
int dsa_port_mrp_del(const struct dsa_port *dp,
const struct switchdev_obj_mrp *mrp)
{
struct dsa_switch *ds = dp->ds;
if (!ds->ops->port_mrp_del)
return -EOPNOTSUPP;
return ds->ops->port_mrp_del(ds, dp->index, mrp);
}
int dsa_port_mrp_add_ring_role(const struct dsa_port *dp,
const struct switchdev_obj_ring_role_mrp *mrp)
{
struct dsa_switch *ds = dp->ds;
if (!ds->ops->port_mrp_add_ring_role)
return -EOPNOTSUPP;
return ds->ops->port_mrp_add_ring_role(ds, dp->index, mrp);
}
int dsa_port_mrp_del_ring_role(const struct dsa_port *dp,
const struct switchdev_obj_ring_role_mrp *mrp)
{
struct dsa_switch *ds = dp->ds;
if (!ds->ops->port_mrp_del_ring_role)
return -EOPNOTSUPP;
return ds->ops->port_mrp_del_ring_role(ds, dp->index, mrp);
}
static int dsa_port_assign_conduit(struct dsa_port *dp,
struct net_device *conduit,
struct netlink_ext_ack *extack,
bool fail_on_err)
{
struct dsa_switch *ds = dp->ds;
int port = dp->index, err;
err = ds->ops->port_change_conduit(ds, port, conduit, extack);
if (err && !fail_on_err)
dev_err(ds->dev, "port %d failed to assign conduit %s: %pe\n",
port, conduit->name, ERR_PTR(err));
if (err && fail_on_err)
return err;
dp->cpu_dp = conduit->dsa_ptr;
dp->cpu_port_in_lag = netif_is_lag_master(conduit);
return 0;
}
/* Change the dp->cpu_dp affinity for a user port. Note that both cross-chip
* notifiers and drivers have implicit assumptions about user-to-CPU-port
* mappings, so we unfortunately cannot delay the deletion of the objects
* (switchdev, standalone addresses, standalone VLANs) on the old CPU port
* until the new CPU port has been set up. So we need to completely tear down
* the old CPU port before changing it, and restore it on errors during the
* bringup of the new one.
*/
int dsa_port_change_conduit(struct dsa_port *dp, struct net_device *conduit,
struct netlink_ext_ack *extack)
{
struct net_device *bridge_dev = dsa_port_bridge_dev_get(dp);
struct net_device *old_conduit = dsa_port_to_conduit(dp);
struct net_device *dev = dp->user;
struct dsa_switch *ds = dp->ds;
bool vlan_filtering;
int err, tmp;
/* Bridges may hold host FDB, MDB and VLAN objects. These need to be
* migrated, so dynamically unoffload and later reoffload the bridge
* port.
*/
if (bridge_dev) {
dsa_port_pre_bridge_leave(dp, bridge_dev);
dsa_port_bridge_leave(dp, bridge_dev);
}
/* The port might still be VLAN filtering even if it's no longer
* under a bridge, either due to ds->vlan_filtering_is_global or
* ds->needs_standalone_vlan_filtering. In turn this means VLANs
* on the CPU port.
*/
vlan_filtering = dsa_port_is_vlan_filtering(dp);
if (vlan_filtering) {
err = dsa_user_manage_vlan_filtering(dev, false);
if (err) {
NL_SET_ERR_MSG_MOD(extack,
"Failed to remove standalone VLANs");
goto rewind_old_bridge;
}
}
/* Standalone addresses, and addresses of upper interfaces like
* VLAN, LAG, HSR need to be migrated.
*/
dsa_user_unsync_ha(dev);
err = dsa_port_assign_conduit(dp, conduit, extack, true);
if (err)
goto rewind_old_addrs;
dsa_user_sync_ha(dev);
if (vlan_filtering) {
err = dsa_user_manage_vlan_filtering(dev, true);
if (err) {
NL_SET_ERR_MSG_MOD(extack,
"Failed to restore standalone VLANs");
goto rewind_new_addrs;
}
}
if (bridge_dev) {
err = dsa_port_bridge_join(dp, bridge_dev, extack);
if (err && err == -EOPNOTSUPP) {
NL_SET_ERR_MSG_MOD(extack,
"Failed to reoffload bridge");
goto rewind_new_vlan;
}
}
return 0;
rewind_new_vlan:
if (vlan_filtering)
dsa_user_manage_vlan_filtering(dev, false);
rewind_new_addrs:
dsa_user_unsync_ha(dev);
dsa_port_assign_conduit(dp, old_conduit, NULL, false);
/* Restore the objects on the old CPU port */
rewind_old_addrs:
dsa_user_sync_ha(dev);
if (vlan_filtering) {
tmp = dsa_user_manage_vlan_filtering(dev, true);
if (tmp) {
dev_err(ds->dev,
"port %d failed to restore standalone VLANs: %pe\n",
dp->index, ERR_PTR(tmp));
}
}
rewind_old_bridge:
if (bridge_dev) {
tmp = dsa_port_bridge_join(dp, bridge_dev, extack);
if (tmp) {
dev_err(ds->dev,
"port %d failed to rejoin bridge %s: %pe\n",
dp->index, bridge_dev->name, ERR_PTR(tmp));
}
}
return err;
}
net: dsa: allow changing the tag protocol via the "tagging" device attribute Currently DSA exposes the following sysfs: $ cat /sys/class/net/eno2/dsa/tagging ocelot which is a read-only device attribute, introduced in the kernel as commit 98cdb4807123 ("net: dsa: Expose tagging protocol to user-space"), and used by libpcap since its commit 993db3800d7d ("Add support for DSA link-layer types"). It would be nice if we could extend this device attribute by making it writable: $ echo ocelot-8021q > /sys/class/net/eno2/dsa/tagging This is useful with DSA switches that can make use of more than one tagging protocol. It may be useful in dsa_loop in the future too, to perform offline testing of various taggers, or for changing between dsa and edsa on Marvell switches, if that is desirable. In terms of implementation, drivers can support this feature by implementing .change_tag_protocol, which should always leave the switch in a consistent state: either with the new protocol if things went well, or with the old one if something failed. Teardown of the old protocol, if necessary, must be handled by the driver. Some things remain as before: - The .get_tag_protocol is currently only called at probe time, to load the initial tagging protocol driver. Nonetheless, new drivers should report the tagging protocol in current use now. - The driver should manage by itself the initial setup of tagging protocol, no later than the .setup() method, as well as destroying resources used by the last tagger in use, no earlier than the .teardown() method. For multi-switch DSA trees, error handling is a bit more complicated, since e.g. the 5th out of 7 switches may fail to change the tag protocol. When that happens, a revert to the original tag protocol is attempted, but that may fail too, leaving the tree in an inconsistent state despite each individual switch implementing .change_tag_protocol transactionally. Since the intersection between drivers that implement .change_tag_protocol and drivers that support D in DSA is currently the empty set, the possibility for this error to happen is ignored for now. Testing: $ insmod mscc_felix.ko [ 79.549784] mscc_felix 0000:00:00.5: Adding to iommu group 14 [ 79.565712] mscc_felix 0000:00:00.5: Failed to register DSA switch: -517 $ insmod tag_ocelot.ko $ rmmod mscc_felix.ko $ insmod mscc_felix.ko [ 97.261724] libphy: VSC9959 internal MDIO bus: probed [ 97.267363] mscc_felix 0000:00:00.5: Found PCS at internal MDIO address 0 [ 97.274998] mscc_felix 0000:00:00.5: Found PCS at internal MDIO address 1 [ 97.282561] mscc_felix 0000:00:00.5: Found PCS at internal MDIO address 2 [ 97.289700] mscc_felix 0000:00:00.5: Found PCS at internal MDIO address 3 [ 97.599163] mscc_felix 0000:00:00.5 swp0 (uninitialized): PHY [0000:00:00.3:10] driver [Microsemi GE VSC8514 SyncE] (irq=POLL) [ 97.862034] mscc_felix 0000:00:00.5 swp1 (uninitialized): PHY [0000:00:00.3:11] driver [Microsemi GE VSC8514 SyncE] (irq=POLL) [ 97.950731] mscc_felix 0000:00:00.5 swp0: configuring for inband/qsgmii link mode [ 97.964278] 8021q: adding VLAN 0 to HW filter on device swp0 [ 98.146161] mscc_felix 0000:00:00.5 swp2 (uninitialized): PHY [0000:00:00.3:12] driver [Microsemi GE VSC8514 SyncE] (irq=POLL) [ 98.238649] mscc_felix 0000:00:00.5 swp1: configuring for inband/qsgmii link mode [ 98.251845] 8021q: adding VLAN 0 to HW filter on device swp1 [ 98.433916] mscc_felix 0000:00:00.5 swp3 (uninitialized): PHY [0000:00:00.3:13] driver [Microsemi GE VSC8514 SyncE] (irq=POLL) [ 98.485542] mscc_felix 0000:00:00.5: configuring for fixed/internal link mode [ 98.503584] mscc_felix 0000:00:00.5: Link is Up - 2.5Gbps/Full - flow control rx/tx [ 98.527948] device eno2 entered promiscuous mode [ 98.544755] DSA: tree 0 setup $ ping 10.0.0.1 PING 10.0.0.1 (10.0.0.1): 56 data bytes 64 bytes from 10.0.0.1: seq=0 ttl=64 time=2.337 ms 64 bytes from 10.0.0.1: seq=1 ttl=64 time=0.754 ms ^C - 10.0.0.1 ping statistics - 2 packets transmitted, 2 packets received, 0% packet loss round-trip min/avg/max = 0.754/1.545/2.337 ms $ cat /sys/class/net/eno2/dsa/tagging ocelot $ cat ./test_ocelot_8021q.sh #!/bin/bash ip link set swp0 down ip link set swp1 down ip link set swp2 down ip link set swp3 down ip link set swp5 down ip link set eno2 down echo ocelot-8021q > /sys/class/net/eno2/dsa/tagging ip link set eno2 up ip link set swp0 up ip link set swp1 up ip link set swp2 up ip link set swp3 up ip link set swp5 up $ ./test_ocelot_8021q.sh ./test_ocelot_8021q.sh: line 9: echo: write error: Protocol not available $ rmmod tag_ocelot.ko rmmod: can't unload module 'tag_ocelot': Resource temporarily unavailable $ insmod tag_ocelot_8021q.ko $ ./test_ocelot_8021q.sh $ cat /sys/class/net/eno2/dsa/tagging ocelot-8021q $ rmmod tag_ocelot.ko $ rmmod tag_ocelot_8021q.ko rmmod: can't unload module 'tag_ocelot_8021q': Resource temporarily unavailable $ ping 10.0.0.1 PING 10.0.0.1 (10.0.0.1): 56 data bytes 64 bytes from 10.0.0.1: seq=0 ttl=64 time=0.953 ms 64 bytes from 10.0.0.1: seq=1 ttl=64 time=0.787 ms 64 bytes from 10.0.0.1: seq=2 ttl=64 time=0.771 ms $ rmmod mscc_felix.ko [ 645.544426] mscc_felix 0000:00:00.5: Link is Down [ 645.838608] DSA: tree 0 torn down $ rmmod tag_ocelot_8021q.ko Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-01-29 01:00:06 +00:00
void dsa_port_set_tag_protocol(struct dsa_port *cpu_dp,
const struct dsa_device_ops *tag_ops)
{
cpu_dp->rcv = tag_ops->rcv;
cpu_dp->tag_ops = tag_ops;
}
static struct phy_device *dsa_port_get_phy_device(struct dsa_port *dp)
{
struct device_node *phy_dn;
struct phy_device *phydev;
phy_dn = of_parse_phandle(dp->dn, "phy-handle", 0);
if (!phy_dn)
return NULL;
phydev = of_phy_find_device(phy_dn);
if (!phydev) {
of_node_put(phy_dn);
return ERR_PTR(-EPROBE_DEFER);
}
net: dsa: fix a leaked reference by adding missing of_node_put The call to of_parse_phandle returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage. Detected by coccinelle with the following warnings: ./net/dsa/port.c:294:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 284, but without a corresponding object release within this function. ./net/dsa/dsa2.c:627:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 618, but without a corresponding object release within this function. ./net/dsa/dsa2.c:630:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 618, but without a corresponding object release within this function. ./net/dsa/dsa2.c:636:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 618, but without a corresponding object release within this function. ./net/dsa/dsa2.c:639:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 618, but without a corresponding object release within this function. Signed-off-by: Wen Yang <wen.yang99@zte.com.cn> Reviewed-by: Vivien Didelot <vivien.didelot@gmail.com> Cc: Andrew Lunn <andrew@lunn.ch> Cc: Vivien Didelot <vivien.didelot@gmail.com> Cc: Florian Fainelli <f.fainelli@gmail.com> Cc: "David S. Miller" <davem@davemloft.net> Cc: Vivien Didelot <vivien.didelot@gmail.com> Cc: netdev@vger.kernel.org Cc: linux-kernel@vger.kernel.org Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-25 07:22:19 +00:00
of_node_put(phy_dn);
return phydev;
}
static struct phylink_pcs *
dsa_port_phylink_mac_select_pcs(struct phylink_config *config,
phy_interface_t interface)
{
struct dsa_port *dp = container_of(config, struct dsa_port, pl_config);
struct phylink_pcs *pcs = ERR_PTR(-EOPNOTSUPP);
struct dsa_switch *ds = dp->ds;
if (ds->ops->phylink_mac_select_pcs)
pcs = ds->ops->phylink_mac_select_pcs(ds, dp->index, interface);
return pcs;
}
static int dsa_port_phylink_mac_prepare(struct phylink_config *config,
unsigned int mode,
phy_interface_t interface)
{
struct dsa_port *dp = container_of(config, struct dsa_port, pl_config);
struct dsa_switch *ds = dp->ds;
int err = 0;
if (ds->ops->phylink_mac_prepare)
err = ds->ops->phylink_mac_prepare(ds, dp->index, mode,
interface);
return err;
}
static void dsa_port_phylink_mac_config(struct phylink_config *config,
unsigned int mode,
const struct phylink_link_state *state)
{
struct dsa_port *dp = container_of(config, struct dsa_port, pl_config);
struct dsa_switch *ds = dp->ds;
if (!ds->ops->phylink_mac_config)
return;
ds->ops->phylink_mac_config(ds, dp->index, mode, state);
}
static int dsa_port_phylink_mac_finish(struct phylink_config *config,
unsigned int mode,
phy_interface_t interface)
{
struct dsa_port *dp = container_of(config, struct dsa_port, pl_config);
struct dsa_switch *ds = dp->ds;
int err = 0;
if (ds->ops->phylink_mac_finish)
err = ds->ops->phylink_mac_finish(ds, dp->index, mode,
interface);
return err;
}
static void dsa_port_phylink_mac_link_down(struct phylink_config *config,
unsigned int mode,
phy_interface_t interface)
{
struct dsa_port *dp = container_of(config, struct dsa_port, pl_config);
struct phy_device *phydev = NULL;
struct dsa_switch *ds = dp->ds;
if (dsa_port_is_user(dp))
phydev = dp->user->phydev;
if (!ds->ops->phylink_mac_link_down) {
if (ds->ops->adjust_link && phydev)
ds->ops->adjust_link(ds, dp->index, phydev);
return;
}
ds->ops->phylink_mac_link_down(ds, dp->index, mode, interface);
}
static void dsa_port_phylink_mac_link_up(struct phylink_config *config,
struct phy_device *phydev,
unsigned int mode,
phy_interface_t interface,
int speed, int duplex,
bool tx_pause, bool rx_pause)
{
struct dsa_port *dp = container_of(config, struct dsa_port, pl_config);
struct dsa_switch *ds = dp->ds;
if (!ds->ops->phylink_mac_link_up) {
if (ds->ops->adjust_link && phydev)
ds->ops->adjust_link(ds, dp->index, phydev);
return;
}
ds->ops->phylink_mac_link_up(ds, dp->index, mode, interface, phydev,
speed, duplex, tx_pause, rx_pause);
}
static const struct phylink_mac_ops dsa_port_phylink_mac_ops = {
.mac_select_pcs = dsa_port_phylink_mac_select_pcs,
.mac_prepare = dsa_port_phylink_mac_prepare,
.mac_config = dsa_port_phylink_mac_config,
.mac_finish = dsa_port_phylink_mac_finish,
.mac_link_down = dsa_port_phylink_mac_link_down,
.mac_link_up = dsa_port_phylink_mac_link_up,
};
int dsa_port_phylink_create(struct dsa_port *dp)
{
struct dsa_switch *ds = dp->ds;
phy_interface_t mode;
net: dsa: don't leave dangling pointers in dp->pl when failing There is a desire to simplify the dsa_port registration path with devlink, and this involves reworking a bit how user ports which fail to connect to their PHY (because it's missing) get reinitialized as UNUSED devlink ports. The desire is for the change to look something like this; basically dsa_port_setup() has failed, we just change dp->type and call dsa_port_setup() again. -/* Destroy the current devlink port, and create a new one which has the UNUSED - * flavour. - */ -static int dsa_port_reinit_as_unused(struct dsa_port *dp) +static int dsa_port_setup_as_unused(struct dsa_port *dp) { - dsa_port_devlink_teardown(dp); dp->type = DSA_PORT_TYPE_UNUSED; - return dsa_port_devlink_setup(dp); + return dsa_port_setup(dp); } For an UNUSED port, dsa_port_setup() mostly only calls dsa_port_devlink_setup() anyway, so we could get away with calling just that. But if we call the full blown dsa_port_setup(dp) (which will be needed to properly set dp->setup = true), the callee will have the tendency to go through this code block too, and call dsa_port_disable(dp): switch (dp->type) { case DSA_PORT_TYPE_UNUSED: dsa_port_disable(dp); break; That is not very good, because dsa_port_disable() has this hidden inside of it: if (dp->pl) phylink_stop(dp->pl); Fact is, we are not prepared to handle a call to dsa_port_disable() with a struct dsa_port that came from a previous (and failed) call to dsa_port_setup(). We do not clean up dp->pl, and this will make the second call to dsa_port_setup() call phylink_stop() on a dangling dp->pl pointer. Solve this by creating an API for phylink destruction which is symmetric to the phylink creation, and never leave dp->pl set to anything except NULL or a valid phylink structure. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-29 07:29:00 +00:00
struct phylink *pl;
int err;
err = of_get_phy_mode(dp->dn, &mode);
if (err)
mode = PHY_INTERFACE_MODE_NA;
if (ds->ops->phylink_get_caps) {
ds->ops->phylink_get_caps(ds, dp->index, &dp->pl_config);
} else {
/* For legacy drivers */
if (mode != PHY_INTERFACE_MODE_NA) {
__set_bit(mode, dp->pl_config.supported_interfaces);
} else {
__set_bit(PHY_INTERFACE_MODE_INTERNAL,
dp->pl_config.supported_interfaces);
__set_bit(PHY_INTERFACE_MODE_GMII,
dp->pl_config.supported_interfaces);
}
}
net: dsa: don't leave dangling pointers in dp->pl when failing There is a desire to simplify the dsa_port registration path with devlink, and this involves reworking a bit how user ports which fail to connect to their PHY (because it's missing) get reinitialized as UNUSED devlink ports. The desire is for the change to look something like this; basically dsa_port_setup() has failed, we just change dp->type and call dsa_port_setup() again. -/* Destroy the current devlink port, and create a new one which has the UNUSED - * flavour. - */ -static int dsa_port_reinit_as_unused(struct dsa_port *dp) +static int dsa_port_setup_as_unused(struct dsa_port *dp) { - dsa_port_devlink_teardown(dp); dp->type = DSA_PORT_TYPE_UNUSED; - return dsa_port_devlink_setup(dp); + return dsa_port_setup(dp); } For an UNUSED port, dsa_port_setup() mostly only calls dsa_port_devlink_setup() anyway, so we could get away with calling just that. But if we call the full blown dsa_port_setup(dp) (which will be needed to properly set dp->setup = true), the callee will have the tendency to go through this code block too, and call dsa_port_disable(dp): switch (dp->type) { case DSA_PORT_TYPE_UNUSED: dsa_port_disable(dp); break; That is not very good, because dsa_port_disable() has this hidden inside of it: if (dp->pl) phylink_stop(dp->pl); Fact is, we are not prepared to handle a call to dsa_port_disable() with a struct dsa_port that came from a previous (and failed) call to dsa_port_setup(). We do not clean up dp->pl, and this will make the second call to dsa_port_setup() call phylink_stop() on a dangling dp->pl pointer. Solve this by creating an API for phylink destruction which is symmetric to the phylink creation, and never leave dp->pl set to anything except NULL or a valid phylink structure. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-29 07:29:00 +00:00
pl = phylink_create(&dp->pl_config, of_fwnode_handle(dp->dn),
mode, &dsa_port_phylink_mac_ops);
if (IS_ERR(pl)) {
pr_err("error creating PHYLINK: %ld\n", PTR_ERR(pl));
net: dsa: don't leave dangling pointers in dp->pl when failing There is a desire to simplify the dsa_port registration path with devlink, and this involves reworking a bit how user ports which fail to connect to their PHY (because it's missing) get reinitialized as UNUSED devlink ports. The desire is for the change to look something like this; basically dsa_port_setup() has failed, we just change dp->type and call dsa_port_setup() again. -/* Destroy the current devlink port, and create a new one which has the UNUSED - * flavour. - */ -static int dsa_port_reinit_as_unused(struct dsa_port *dp) +static int dsa_port_setup_as_unused(struct dsa_port *dp) { - dsa_port_devlink_teardown(dp); dp->type = DSA_PORT_TYPE_UNUSED; - return dsa_port_devlink_setup(dp); + return dsa_port_setup(dp); } For an UNUSED port, dsa_port_setup() mostly only calls dsa_port_devlink_setup() anyway, so we could get away with calling just that. But if we call the full blown dsa_port_setup(dp) (which will be needed to properly set dp->setup = true), the callee will have the tendency to go through this code block too, and call dsa_port_disable(dp): switch (dp->type) { case DSA_PORT_TYPE_UNUSED: dsa_port_disable(dp); break; That is not very good, because dsa_port_disable() has this hidden inside of it: if (dp->pl) phylink_stop(dp->pl); Fact is, we are not prepared to handle a call to dsa_port_disable() with a struct dsa_port that came from a previous (and failed) call to dsa_port_setup(). We do not clean up dp->pl, and this will make the second call to dsa_port_setup() call phylink_stop() on a dangling dp->pl pointer. Solve this by creating an API for phylink destruction which is symmetric to the phylink creation, and never leave dp->pl set to anything except NULL or a valid phylink structure. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-29 07:29:00 +00:00
return PTR_ERR(pl);
}
net: dsa: don't leave dangling pointers in dp->pl when failing There is a desire to simplify the dsa_port registration path with devlink, and this involves reworking a bit how user ports which fail to connect to their PHY (because it's missing) get reinitialized as UNUSED devlink ports. The desire is for the change to look something like this; basically dsa_port_setup() has failed, we just change dp->type and call dsa_port_setup() again. -/* Destroy the current devlink port, and create a new one which has the UNUSED - * flavour. - */ -static int dsa_port_reinit_as_unused(struct dsa_port *dp) +static int dsa_port_setup_as_unused(struct dsa_port *dp) { - dsa_port_devlink_teardown(dp); dp->type = DSA_PORT_TYPE_UNUSED; - return dsa_port_devlink_setup(dp); + return dsa_port_setup(dp); } For an UNUSED port, dsa_port_setup() mostly only calls dsa_port_devlink_setup() anyway, so we could get away with calling just that. But if we call the full blown dsa_port_setup(dp) (which will be needed to properly set dp->setup = true), the callee will have the tendency to go through this code block too, and call dsa_port_disable(dp): switch (dp->type) { case DSA_PORT_TYPE_UNUSED: dsa_port_disable(dp); break; That is not very good, because dsa_port_disable() has this hidden inside of it: if (dp->pl) phylink_stop(dp->pl); Fact is, we are not prepared to handle a call to dsa_port_disable() with a struct dsa_port that came from a previous (and failed) call to dsa_port_setup(). We do not clean up dp->pl, and this will make the second call to dsa_port_setup() call phylink_stop() on a dangling dp->pl pointer. Solve this by creating an API for phylink destruction which is symmetric to the phylink creation, and never leave dp->pl set to anything except NULL or a valid phylink structure. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-29 07:29:00 +00:00
dp->pl = pl;
return 0;
}
net: dsa: don't leave dangling pointers in dp->pl when failing There is a desire to simplify the dsa_port registration path with devlink, and this involves reworking a bit how user ports which fail to connect to their PHY (because it's missing) get reinitialized as UNUSED devlink ports. The desire is for the change to look something like this; basically dsa_port_setup() has failed, we just change dp->type and call dsa_port_setup() again. -/* Destroy the current devlink port, and create a new one which has the UNUSED - * flavour. - */ -static int dsa_port_reinit_as_unused(struct dsa_port *dp) +static int dsa_port_setup_as_unused(struct dsa_port *dp) { - dsa_port_devlink_teardown(dp); dp->type = DSA_PORT_TYPE_UNUSED; - return dsa_port_devlink_setup(dp); + return dsa_port_setup(dp); } For an UNUSED port, dsa_port_setup() mostly only calls dsa_port_devlink_setup() anyway, so we could get away with calling just that. But if we call the full blown dsa_port_setup(dp) (which will be needed to properly set dp->setup = true), the callee will have the tendency to go through this code block too, and call dsa_port_disable(dp): switch (dp->type) { case DSA_PORT_TYPE_UNUSED: dsa_port_disable(dp); break; That is not very good, because dsa_port_disable() has this hidden inside of it: if (dp->pl) phylink_stop(dp->pl); Fact is, we are not prepared to handle a call to dsa_port_disable() with a struct dsa_port that came from a previous (and failed) call to dsa_port_setup(). We do not clean up dp->pl, and this will make the second call to dsa_port_setup() call phylink_stop() on a dangling dp->pl pointer. Solve this by creating an API for phylink destruction which is symmetric to the phylink creation, and never leave dp->pl set to anything except NULL or a valid phylink structure. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-29 07:29:00 +00:00
void dsa_port_phylink_destroy(struct dsa_port *dp)
{
phylink_destroy(dp->pl);
dp->pl = NULL;
}
static int dsa_shared_port_setup_phy_of(struct dsa_port *dp, bool enable)
{
struct dsa_switch *ds = dp->ds;
struct phy_device *phydev;
int port = dp->index;
int err = 0;
phydev = dsa_port_get_phy_device(dp);
if (!phydev)
return 0;
if (IS_ERR(phydev))
return PTR_ERR(phydev);
if (enable) {
err = genphy_resume(phydev);
if (err < 0)
goto err_put_dev;
err = genphy_read_status(phydev);
if (err < 0)
goto err_put_dev;
} else {
err = genphy_suspend(phydev);
if (err < 0)
goto err_put_dev;
}
if (ds->ops->adjust_link)
ds->ops->adjust_link(ds, port, phydev);
dev_dbg(ds->dev, "enabled port's phy: %s", phydev_name(phydev));
err_put_dev:
put_device(&phydev->mdio.dev);
return err;
}
static int dsa_shared_port_fixed_link_register_of(struct dsa_port *dp)
{
struct device_node *dn = dp->dn;
struct dsa_switch *ds = dp->ds;
struct phy_device *phydev;
int port = dp->index;
phy_interface_t mode;
int err;
err = of_phy_register_fixed_link(dn);
if (err) {
dev_err(ds->dev,
"failed to register the fixed PHY of port %d\n",
port);
return err;
}
phydev = of_phy_find_device(dn);
err = of_get_phy_mode(dn, &mode);
if (err)
mode = PHY_INTERFACE_MODE_NA;
phydev->interface = mode;
genphy_read_status(phydev);
if (ds->ops->adjust_link)
ds->ops->adjust_link(ds, port, phydev);
put_device(&phydev->mdio.dev);
return 0;
}
static int dsa_shared_port_phylink_register(struct dsa_port *dp)
{
struct dsa_switch *ds = dp->ds;
struct device_node *port_dn = dp->dn;
int err;
dp->pl_config.dev = ds->dev;
dp->pl_config.type = PHYLINK_DEV;
err = dsa_port_phylink_create(dp);
if (err)
return err;
err = phylink_of_phy_connect(dp->pl, port_dn, 0);
if (err && err != -ENODEV) {
pr_err("could not attach to PHY: %d\n", err);
goto err_phy_connect;
}
return 0;
err_phy_connect:
net: dsa: don't leave dangling pointers in dp->pl when failing There is a desire to simplify the dsa_port registration path with devlink, and this involves reworking a bit how user ports which fail to connect to their PHY (because it's missing) get reinitialized as UNUSED devlink ports. The desire is for the change to look something like this; basically dsa_port_setup() has failed, we just change dp->type and call dsa_port_setup() again. -/* Destroy the current devlink port, and create a new one which has the UNUSED - * flavour. - */ -static int dsa_port_reinit_as_unused(struct dsa_port *dp) +static int dsa_port_setup_as_unused(struct dsa_port *dp) { - dsa_port_devlink_teardown(dp); dp->type = DSA_PORT_TYPE_UNUSED; - return dsa_port_devlink_setup(dp); + return dsa_port_setup(dp); } For an UNUSED port, dsa_port_setup() mostly only calls dsa_port_devlink_setup() anyway, so we could get away with calling just that. But if we call the full blown dsa_port_setup(dp) (which will be needed to properly set dp->setup = true), the callee will have the tendency to go through this code block too, and call dsa_port_disable(dp): switch (dp->type) { case DSA_PORT_TYPE_UNUSED: dsa_port_disable(dp); break; That is not very good, because dsa_port_disable() has this hidden inside of it: if (dp->pl) phylink_stop(dp->pl); Fact is, we are not prepared to handle a call to dsa_port_disable() with a struct dsa_port that came from a previous (and failed) call to dsa_port_setup(). We do not clean up dp->pl, and this will make the second call to dsa_port_setup() call phylink_stop() on a dangling dp->pl pointer. Solve this by creating an API for phylink destruction which is symmetric to the phylink creation, and never leave dp->pl set to anything except NULL or a valid phylink structure. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-29 07:29:00 +00:00
dsa_port_phylink_destroy(dp);
return err;
}
net: dsa: make phylink-related OF properties mandatory on DSA and CPU ports Early DSA drivers were kind of simplistic in that they assumed a fairly narrow hardware layout. User ports would have integrated PHYs at an internal MDIO address that is derivable from the port number, and shared (DSA and CPU) ports would have an MII-style (serial or parallel) connection to another MAC. Phylib and then phylink were used to drive the internal PHYs, and this needed little to no description through the platform data structures. Bringing up the shared ports at the maximum supported link speed was the responsibility of the drivers. As a result of this, when these early drivers were converted from platform data to the new DSA OF bindings, there was no link information translated into the first DT bindings. https://lore.kernel.org/all/YtXFtTsf++AeDm1l@lunn.ch/ Later, phylink was adopted for shared ports as well, and today we have a workaround in place, introduced by commit a20f997010c4 ("net: dsa: Don't instantiate phylink for CPU/DSA ports unless needed"). There, DSA checks for the presence of phy-handle/fixed-link/managed OF properties, and if missing, phylink registration would be skipped. This is because phylink is optional for some drivers (the shared ports already work without it), but the process of starting to register a port with phylink is irreversible: if phylink_create() fails to find the fwnode properties it needs, it bails out and it leaves the ports inoperational (because phylink expects ports to be initially down, so DSA necessarily takes them down, and doesn't know how to put them back up again). DSA being a common framework, new drivers opt into this workaround willy-nilly, but the ideal behavior from the DSA core's side would have been to not interfere with phylink's process of failing at all. This isn't possible because of regression concerns with pre-phylink DT blobs, but at least DSA should put a stop to the proliferation of more of such cases that rely on the workaround to skip phylink registration, and sanitize the environment that new drivers work in. To that end, create a list of compatible strings for which the workaround is preserved, and don't apply the workaround for any drivers outside that list (this includes new drivers). In some cases, we make the assumption that even existing drivers don't rely on DSA's workaround, and we do this by looking at the device trees in which they appear. We can't fully know what is the situation with downstream DT blobs, but we can guess the overall trend by studying the DT blobs that were submitted upstream. If there are upstream blobs that have lacking descriptions, we take it as very likely that there are many more downstream blobs that do so too. If all upstream blobs have complete descriptions, we take that as a hint that the driver is a candidate for enforcing strict DT bindings (considering that most bindings are copy-pasted). If there are no upstream DT blobs, we take the conservative route of allowing the workaround, unless the driver maintainer instructs us otherwise. The driver situation is as follows: ar9331 ~~~~~~ compatible strings: - qca,ar9331-switch 1 occurrence in mainline device trees, part of SoC dtsi (arch/mips/boot/dts/qca/ar9331.dtsi), description is not problematic. Verdict: opt into strict DT bindings and out of workarounds. b53 ~~~ compatible strings: - brcm,bcm5325 - brcm,bcm53115 - brcm,bcm53125 - brcm,bcm53128 - brcm,bcm5365 - brcm,bcm5389 - brcm,bcm5395 - brcm,bcm5397 - brcm,bcm5398 - brcm,bcm53010-srab - brcm,bcm53011-srab - brcm,bcm53012-srab - brcm,bcm53018-srab - brcm,bcm53019-srab - brcm,bcm5301x-srab - brcm,bcm11360-srab - brcm,bcm58522-srab - brcm,bcm58525-srab - brcm,bcm58535-srab - brcm,bcm58622-srab - brcm,bcm58623-srab - brcm,bcm58625-srab - brcm,bcm88312-srab - brcm,cygnus-srab - brcm,nsp-srab - brcm,omega-srab - brcm,bcm3384-switch - brcm,bcm6328-switch - brcm,bcm6368-switch - brcm,bcm63xx-switch I've found at least these mainline DT blobs with problems: arch/arm/boot/dts/bcm47094-linksys-panamera.dts - lacks phy-mode arch/arm/boot/dts/bcm47189-tenda-ac9.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47081-luxul-xap-1410.dts arch/arm/boot/dts/bcm47081-luxul-xwr-1200.dts arch/arm/boot/dts/bcm47081-buffalo-wzr-600dhp2.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47094-luxul-xbr-4500.dts arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts arch/arm/boot/dts/bcm4708-luxul-xap-1510.dts arch/arm/boot/dts/bcm953012er.dts arch/arm/boot/dts/bcm4708-netgear-r6250.dts arch/arm/boot/dts/bcm4708-buffalo-wzr-1166dhp-common.dtsi arch/arm/boot/dts/bcm4708-luxul-xwc-1000.dts arch/arm/boot/dts/bcm47094-luxul-abr-4500.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm53016-meraki-mr32.dts - lacks phy-mode Verdict: opt into DSA workarounds. bcm_sf2 ~~~~~~~ compatible strings: - brcm,bcm4908-switch - brcm,bcm7445-switch-v4.0 - brcm,bcm7278-switch-v4.0 - brcm,bcm7278-switch-v4.8 A single occurrence in mainline (arch/arm64/boot/dts/broadcom/bcm4908/bcm4908.dtsi), part of a SoC dtsi, valid description. Florian Fainelli explains that most of the bcm_sf2 device trees lack a full description for the internal IMP ports. Verdict: opt the BCM4908 into strict DT bindings, and opt the rest into the workarounds. Note that even though BCM4908 has strict DT bindings, it still does not register with phylink on the IMP port due to it implementing ->adjust_link(). hellcreek ~~~~~~~~~ compatible strings: - hirschmann,hellcreek-de1soc-r1 No occurrence in mainline device trees. Kurt Kanzenbach explains that the downstream device trees lacked phy-mode and fixed link, and needed work, but were fixed in the meantime. Verdict: opt into strict DT bindings and out of workarounds. lan9303 ~~~~~~~ compatible strings: - smsc,lan9303-mdio - smsc,lan9303-i2c 1 occurrence in mainline device trees: arch/arm/boot/dts/imx53-kp-hsc.dts - no phy-mode, no fixed-link Verdict: opt out of strict DT bindings and into workarounds. lantiq_gswip ~~~~~~~~~~~~ compatible strings: - lantiq,xrx200-gswip - lantiq,xrx300-gswip - lantiq,xrx330-gswip No occurrences in mainline device trees. Martin Blumenstingl confirms that the downstream OpenWrt device trees lack a proper fixed-link and need work, and that the incomplete description can even be seen in the example from Documentation/devicetree/bindings/net/dsa/lantiq-gswip.txt. Verdict: opt out of strict DT bindings and into workarounds. microchip ksz ~~~~~~~~~~~~~ compatible strings: - microchip,ksz8765 - microchip,ksz8794 - microchip,ksz8795 - microchip,ksz8863 - microchip,ksz8873 - microchip,ksz9477 - microchip,ksz9897 - microchip,ksz9893 - microchip,ksz9563 - microchip,ksz8563 - microchip,ksz9567 - microchip,lan9370 - microchip,lan9371 - microchip,lan9372 - microchip,lan9373 - microchip,lan9374 5 occurrences in mainline device trees, all descriptions are valid. But we had a snafu for the ksz8795 and ksz9477 drivers where the phy-mode property would be expected to be located directly under the 'switch' node rather than under a port OF node. It was fixed by commit edecfa98f602 ("net: dsa: microchip: look for phy-mode in port nodes"). The driver still has compatibility with the old DT blobs. The lan937x support was added later than the above snafu was fixed, and even though it has support for the broken DT blobs by virtue of sharing a common probing function, I'll take it that its DT blobs are correct. Verdict: opt lan937x into strict DT bindings, and the others out. mt7530 ~~~~~~ compatible strings - mediatek,mt7621 - mediatek,mt7530 - mediatek,mt7531 Multiple occurrences in mainline device trees, one is part of an SoC dtsi (arch/mips/boot/dts/ralink/mt7621.dtsi), all descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. mv88e6060 ~~~~~~~~~ compatible string: - marvell,mv88e6060 no occurrences in mainline, nobody knows anybody who uses it. Verdict: opt out of strict DT bindings and into workarounds. mv88e6xxx ~~~~~~~~~ compatible strings: - marvell,mv88e6085 - marvell,mv88e6190 - marvell,mv88e6250 Device trees that have incomplete descriptions of CPU or DSA ports: arch/arm64/boot/dts/freescale/imx8mq-zii-ultra.dtsi - lacks phy-mode arch/arm64/boot/dts/marvell/cn9130-crb.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-spu3.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-mv88f6281gtw-ge.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-spb4.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-cfu1.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-dev-rev-c.dts - lacks phy-mode on CPU port, fixed-link on DSA ports arch/arm/boot/dts/vf610-zii-dev-rev-b.dts - lacks phy-mode on CPU port arch/arm/boot/dts/armada-381-netgear-gs110emx.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-scu4-aib.dts - lacks fixed-link on xgmii DSA ports and/or in-band-status on 2500base-x DSA ports, and phy-mode on CPU port arch/arm/boot/dts/imx6qdl-gw5904.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-385-clearfog-gtr-l8.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-dtu.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-dir665.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-rd88f6281.dtsi - lacks phy-mode arch/arm/boot/dts/orion5x-netgear-wnr854t.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-388-clearfog.dts - lacks phy-mode arch/arm/boot/dts/armada-xp-linksys-mamba.dts - lacks phy-mode arch/arm/boot/dts/armada-385-linksys.dtsi - lacks phy-mode arch/arm/boot/dts/imx6q-b450v3.dts arch/arm/boot/dts/imx6q-b850v3.dts - has a phy-handle but not a phy-mode? arch/arm/boot/dts/armada-370-rd.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-linksys-viper.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-rdu1.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-scu2-mezz.dts - lacks phy-mode arch/arm/boot/dts/imx6qdl-zii-rdu2.dtsi - lacks phy-mode arch/arm/boot/dts/armada-385-clearfog-gtr-s4.dts - lacks phy-mode and fixed-link Verdict: opt out of strict DT bindings and into workarounds. ocelot ~~~~~~ compatible strings: - mscc,vsc9953-switch - felix (arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi) is a PCI device, has no compatible string 2 occurrences in mainline, both are part of SoC dtsi and complete. Verdict: opt into strict DT bindings and out of workarounds. qca8k ~~~~~ compatible strings: - qca,qca8327 - qca,qca8328 - qca,qca8334 - qca,qca8337 5 occurrences in mainline device trees, none of the descriptions are problematic. Verdict: opt into strict DT bindings and out of workarounds. realtek ~~~~~~~ compatible strings: - realtek,rtl8366rb - realtek,rtl8365mb 2 occurrences in mainline, both descriptions are fine, additionally rtl8365mb.c has a comment "The device tree firmware should also specify the link partner of the extension port - either via a fixed-link or other phy-handle." Verdict: opt into strict DT bindings and out of workarounds. rzn1_a5psw ~~~~~~~~~~ compatible strings: - renesas,rzn1-a5psw One single occurrence, part of SoC dtsi (arch/arm/boot/dts/r9a06g032.dtsi), description is fine. Verdict: opt into strict DT bindings and out of workarounds. sja1105 ~~~~~~~ Driver already validates its port OF nodes in sja1105_parse_ports_node(). Verdict: opt into strict DT bindings and out of workarounds. vsc73xx ~~~~~~~ compatible strings: - vitesse,vsc7385 - vitesse,vsc7388 - vitesse,vsc7395 - vitesse,vsc7398 2 occurrences in mainline device trees, both descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. xrs700x ~~~~~~~ compatible strings: - arrow,xrs7003e - arrow,xrs7003f - arrow,xrs7004e - arrow,xrs7004f no occurrences in mainline, we don't know. Verdict: opt out of strict DT bindings and into workarounds. Because there is a pattern where newly added switches reuse existing drivers more often than introducing new ones, I've opted for deciding who gets to opt into the workaround based on an OF compatible match table in the DSA core. The alternative would have been to add another boolean property to struct dsa_switch, like configure_vlan_while_not_filtering. But this avoids situations where sometimes driver maintainers obfuscate what goes on by sharing a common probing function, and therefore making new switches inherit old quirks. Side note, we also warn about missing properties for drivers that rely on the workaround. This isn't an indication that we'll break compatibility with those DT blobs any time soon, but is rather done to raise awareness about the change, for future DT blob authors. Cc: Rob Herring <robh+dt@kernel.org> Cc: Frank Rowand <frowand.list@gmail.com> Acked-by: Alvin Šipraga <alsi@bang-olufsen.dk> # realtek Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-08-18 11:55:00 +00:00
/* During the initial DSA driver migration to OF, port nodes were sometimes
* added to device trees with no indication of how they should operate from a
* link management perspective (phy-handle, fixed-link, etc). Additionally, the
* phy-mode may be absent. The interpretation of these port OF nodes depends on
* their type.
*
* User ports with no phy-handle or fixed-link are expected to connect to an
* internal PHY located on the ds->user_mii_bus at an MDIO address equal to
net: dsa: make phylink-related OF properties mandatory on DSA and CPU ports Early DSA drivers were kind of simplistic in that they assumed a fairly narrow hardware layout. User ports would have integrated PHYs at an internal MDIO address that is derivable from the port number, and shared (DSA and CPU) ports would have an MII-style (serial or parallel) connection to another MAC. Phylib and then phylink were used to drive the internal PHYs, and this needed little to no description through the platform data structures. Bringing up the shared ports at the maximum supported link speed was the responsibility of the drivers. As a result of this, when these early drivers were converted from platform data to the new DSA OF bindings, there was no link information translated into the first DT bindings. https://lore.kernel.org/all/YtXFtTsf++AeDm1l@lunn.ch/ Later, phylink was adopted for shared ports as well, and today we have a workaround in place, introduced by commit a20f997010c4 ("net: dsa: Don't instantiate phylink for CPU/DSA ports unless needed"). There, DSA checks for the presence of phy-handle/fixed-link/managed OF properties, and if missing, phylink registration would be skipped. This is because phylink is optional for some drivers (the shared ports already work without it), but the process of starting to register a port with phylink is irreversible: if phylink_create() fails to find the fwnode properties it needs, it bails out and it leaves the ports inoperational (because phylink expects ports to be initially down, so DSA necessarily takes them down, and doesn't know how to put them back up again). DSA being a common framework, new drivers opt into this workaround willy-nilly, but the ideal behavior from the DSA core's side would have been to not interfere with phylink's process of failing at all. This isn't possible because of regression concerns with pre-phylink DT blobs, but at least DSA should put a stop to the proliferation of more of such cases that rely on the workaround to skip phylink registration, and sanitize the environment that new drivers work in. To that end, create a list of compatible strings for which the workaround is preserved, and don't apply the workaround for any drivers outside that list (this includes new drivers). In some cases, we make the assumption that even existing drivers don't rely on DSA's workaround, and we do this by looking at the device trees in which they appear. We can't fully know what is the situation with downstream DT blobs, but we can guess the overall trend by studying the DT blobs that were submitted upstream. If there are upstream blobs that have lacking descriptions, we take it as very likely that there are many more downstream blobs that do so too. If all upstream blobs have complete descriptions, we take that as a hint that the driver is a candidate for enforcing strict DT bindings (considering that most bindings are copy-pasted). If there are no upstream DT blobs, we take the conservative route of allowing the workaround, unless the driver maintainer instructs us otherwise. The driver situation is as follows: ar9331 ~~~~~~ compatible strings: - qca,ar9331-switch 1 occurrence in mainline device trees, part of SoC dtsi (arch/mips/boot/dts/qca/ar9331.dtsi), description is not problematic. Verdict: opt into strict DT bindings and out of workarounds. b53 ~~~ compatible strings: - brcm,bcm5325 - brcm,bcm53115 - brcm,bcm53125 - brcm,bcm53128 - brcm,bcm5365 - brcm,bcm5389 - brcm,bcm5395 - brcm,bcm5397 - brcm,bcm5398 - brcm,bcm53010-srab - brcm,bcm53011-srab - brcm,bcm53012-srab - brcm,bcm53018-srab - brcm,bcm53019-srab - brcm,bcm5301x-srab - brcm,bcm11360-srab - brcm,bcm58522-srab - brcm,bcm58525-srab - brcm,bcm58535-srab - brcm,bcm58622-srab - brcm,bcm58623-srab - brcm,bcm58625-srab - brcm,bcm88312-srab - brcm,cygnus-srab - brcm,nsp-srab - brcm,omega-srab - brcm,bcm3384-switch - brcm,bcm6328-switch - brcm,bcm6368-switch - brcm,bcm63xx-switch I've found at least these mainline DT blobs with problems: arch/arm/boot/dts/bcm47094-linksys-panamera.dts - lacks phy-mode arch/arm/boot/dts/bcm47189-tenda-ac9.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47081-luxul-xap-1410.dts arch/arm/boot/dts/bcm47081-luxul-xwr-1200.dts arch/arm/boot/dts/bcm47081-buffalo-wzr-600dhp2.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47094-luxul-xbr-4500.dts arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts arch/arm/boot/dts/bcm4708-luxul-xap-1510.dts arch/arm/boot/dts/bcm953012er.dts arch/arm/boot/dts/bcm4708-netgear-r6250.dts arch/arm/boot/dts/bcm4708-buffalo-wzr-1166dhp-common.dtsi arch/arm/boot/dts/bcm4708-luxul-xwc-1000.dts arch/arm/boot/dts/bcm47094-luxul-abr-4500.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm53016-meraki-mr32.dts - lacks phy-mode Verdict: opt into DSA workarounds. bcm_sf2 ~~~~~~~ compatible strings: - brcm,bcm4908-switch - brcm,bcm7445-switch-v4.0 - brcm,bcm7278-switch-v4.0 - brcm,bcm7278-switch-v4.8 A single occurrence in mainline (arch/arm64/boot/dts/broadcom/bcm4908/bcm4908.dtsi), part of a SoC dtsi, valid description. Florian Fainelli explains that most of the bcm_sf2 device trees lack a full description for the internal IMP ports. Verdict: opt the BCM4908 into strict DT bindings, and opt the rest into the workarounds. Note that even though BCM4908 has strict DT bindings, it still does not register with phylink on the IMP port due to it implementing ->adjust_link(). hellcreek ~~~~~~~~~ compatible strings: - hirschmann,hellcreek-de1soc-r1 No occurrence in mainline device trees. Kurt Kanzenbach explains that the downstream device trees lacked phy-mode and fixed link, and needed work, but were fixed in the meantime. Verdict: opt into strict DT bindings and out of workarounds. lan9303 ~~~~~~~ compatible strings: - smsc,lan9303-mdio - smsc,lan9303-i2c 1 occurrence in mainline device trees: arch/arm/boot/dts/imx53-kp-hsc.dts - no phy-mode, no fixed-link Verdict: opt out of strict DT bindings and into workarounds. lantiq_gswip ~~~~~~~~~~~~ compatible strings: - lantiq,xrx200-gswip - lantiq,xrx300-gswip - lantiq,xrx330-gswip No occurrences in mainline device trees. Martin Blumenstingl confirms that the downstream OpenWrt device trees lack a proper fixed-link and need work, and that the incomplete description can even be seen in the example from Documentation/devicetree/bindings/net/dsa/lantiq-gswip.txt. Verdict: opt out of strict DT bindings and into workarounds. microchip ksz ~~~~~~~~~~~~~ compatible strings: - microchip,ksz8765 - microchip,ksz8794 - microchip,ksz8795 - microchip,ksz8863 - microchip,ksz8873 - microchip,ksz9477 - microchip,ksz9897 - microchip,ksz9893 - microchip,ksz9563 - microchip,ksz8563 - microchip,ksz9567 - microchip,lan9370 - microchip,lan9371 - microchip,lan9372 - microchip,lan9373 - microchip,lan9374 5 occurrences in mainline device trees, all descriptions are valid. But we had a snafu for the ksz8795 and ksz9477 drivers where the phy-mode property would be expected to be located directly under the 'switch' node rather than under a port OF node. It was fixed by commit edecfa98f602 ("net: dsa: microchip: look for phy-mode in port nodes"). The driver still has compatibility with the old DT blobs. The lan937x support was added later than the above snafu was fixed, and even though it has support for the broken DT blobs by virtue of sharing a common probing function, I'll take it that its DT blobs are correct. Verdict: opt lan937x into strict DT bindings, and the others out. mt7530 ~~~~~~ compatible strings - mediatek,mt7621 - mediatek,mt7530 - mediatek,mt7531 Multiple occurrences in mainline device trees, one is part of an SoC dtsi (arch/mips/boot/dts/ralink/mt7621.dtsi), all descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. mv88e6060 ~~~~~~~~~ compatible string: - marvell,mv88e6060 no occurrences in mainline, nobody knows anybody who uses it. Verdict: opt out of strict DT bindings and into workarounds. mv88e6xxx ~~~~~~~~~ compatible strings: - marvell,mv88e6085 - marvell,mv88e6190 - marvell,mv88e6250 Device trees that have incomplete descriptions of CPU or DSA ports: arch/arm64/boot/dts/freescale/imx8mq-zii-ultra.dtsi - lacks phy-mode arch/arm64/boot/dts/marvell/cn9130-crb.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-spu3.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-mv88f6281gtw-ge.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-spb4.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-cfu1.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-dev-rev-c.dts - lacks phy-mode on CPU port, fixed-link on DSA ports arch/arm/boot/dts/vf610-zii-dev-rev-b.dts - lacks phy-mode on CPU port arch/arm/boot/dts/armada-381-netgear-gs110emx.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-scu4-aib.dts - lacks fixed-link on xgmii DSA ports and/or in-band-status on 2500base-x DSA ports, and phy-mode on CPU port arch/arm/boot/dts/imx6qdl-gw5904.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-385-clearfog-gtr-l8.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-dtu.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-dir665.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-rd88f6281.dtsi - lacks phy-mode arch/arm/boot/dts/orion5x-netgear-wnr854t.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-388-clearfog.dts - lacks phy-mode arch/arm/boot/dts/armada-xp-linksys-mamba.dts - lacks phy-mode arch/arm/boot/dts/armada-385-linksys.dtsi - lacks phy-mode arch/arm/boot/dts/imx6q-b450v3.dts arch/arm/boot/dts/imx6q-b850v3.dts - has a phy-handle but not a phy-mode? arch/arm/boot/dts/armada-370-rd.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-linksys-viper.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-rdu1.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-scu2-mezz.dts - lacks phy-mode arch/arm/boot/dts/imx6qdl-zii-rdu2.dtsi - lacks phy-mode arch/arm/boot/dts/armada-385-clearfog-gtr-s4.dts - lacks phy-mode and fixed-link Verdict: opt out of strict DT bindings and into workarounds. ocelot ~~~~~~ compatible strings: - mscc,vsc9953-switch - felix (arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi) is a PCI device, has no compatible string 2 occurrences in mainline, both are part of SoC dtsi and complete. Verdict: opt into strict DT bindings and out of workarounds. qca8k ~~~~~ compatible strings: - qca,qca8327 - qca,qca8328 - qca,qca8334 - qca,qca8337 5 occurrences in mainline device trees, none of the descriptions are problematic. Verdict: opt into strict DT bindings and out of workarounds. realtek ~~~~~~~ compatible strings: - realtek,rtl8366rb - realtek,rtl8365mb 2 occurrences in mainline, both descriptions are fine, additionally rtl8365mb.c has a comment "The device tree firmware should also specify the link partner of the extension port - either via a fixed-link or other phy-handle." Verdict: opt into strict DT bindings and out of workarounds. rzn1_a5psw ~~~~~~~~~~ compatible strings: - renesas,rzn1-a5psw One single occurrence, part of SoC dtsi (arch/arm/boot/dts/r9a06g032.dtsi), description is fine. Verdict: opt into strict DT bindings and out of workarounds. sja1105 ~~~~~~~ Driver already validates its port OF nodes in sja1105_parse_ports_node(). Verdict: opt into strict DT bindings and out of workarounds. vsc73xx ~~~~~~~ compatible strings: - vitesse,vsc7385 - vitesse,vsc7388 - vitesse,vsc7395 - vitesse,vsc7398 2 occurrences in mainline device trees, both descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. xrs700x ~~~~~~~ compatible strings: - arrow,xrs7003e - arrow,xrs7003f - arrow,xrs7004e - arrow,xrs7004f no occurrences in mainline, we don't know. Verdict: opt out of strict DT bindings and into workarounds. Because there is a pattern where newly added switches reuse existing drivers more often than introducing new ones, I've opted for deciding who gets to opt into the workaround based on an OF compatible match table in the DSA core. The alternative would have been to add another boolean property to struct dsa_switch, like configure_vlan_while_not_filtering. But this avoids situations where sometimes driver maintainers obfuscate what goes on by sharing a common probing function, and therefore making new switches inherit old quirks. Side note, we also warn about missing properties for drivers that rely on the workaround. This isn't an indication that we'll break compatibility with those DT blobs any time soon, but is rather done to raise awareness about the change, for future DT blob authors. Cc: Rob Herring <robh+dt@kernel.org> Cc: Frank Rowand <frowand.list@gmail.com> Acked-by: Alvin Šipraga <alsi@bang-olufsen.dk> # realtek Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-08-18 11:55:00 +00:00
* the port number. This description is still actively supported.
*
* Shared (CPU and DSA) ports with no phy-handle or fixed-link are expected to
* operate at the maximum speed that their phy-mode is capable of. If the
* phy-mode is absent, they are expected to operate using the phy-mode
* supported by the port that gives the highest link speed. It is unspecified
* if the port should use flow control or not, half duplex or full duplex, or
* if the phy-mode is a SERDES link, whether in-band autoneg is expected to be
* enabled or not.
*
* In the latter case of shared ports, omitting the link management description
* from the firmware node is deprecated and strongly discouraged. DSA uses
* phylink, which rejects the firmware nodes of these ports for lacking
* required properties.
*
* For switches in this table, DSA will skip enforcing validation and will
* later omit registering a phylink instance for the shared ports, if they lack
* a fixed-link, a phy-handle, or a managed = "in-band-status" property.
* It becomes the responsibility of the driver to ensure that these ports
* operate at the maximum speed (whatever this means) and will interoperate
* with the DSA conduit or other cascade port, since phylink methods will not be
net: dsa: make phylink-related OF properties mandatory on DSA and CPU ports Early DSA drivers were kind of simplistic in that they assumed a fairly narrow hardware layout. User ports would have integrated PHYs at an internal MDIO address that is derivable from the port number, and shared (DSA and CPU) ports would have an MII-style (serial or parallel) connection to another MAC. Phylib and then phylink were used to drive the internal PHYs, and this needed little to no description through the platform data structures. Bringing up the shared ports at the maximum supported link speed was the responsibility of the drivers. As a result of this, when these early drivers were converted from platform data to the new DSA OF bindings, there was no link information translated into the first DT bindings. https://lore.kernel.org/all/YtXFtTsf++AeDm1l@lunn.ch/ Later, phylink was adopted for shared ports as well, and today we have a workaround in place, introduced by commit a20f997010c4 ("net: dsa: Don't instantiate phylink for CPU/DSA ports unless needed"). There, DSA checks for the presence of phy-handle/fixed-link/managed OF properties, and if missing, phylink registration would be skipped. This is because phylink is optional for some drivers (the shared ports already work without it), but the process of starting to register a port with phylink is irreversible: if phylink_create() fails to find the fwnode properties it needs, it bails out and it leaves the ports inoperational (because phylink expects ports to be initially down, so DSA necessarily takes them down, and doesn't know how to put them back up again). DSA being a common framework, new drivers opt into this workaround willy-nilly, but the ideal behavior from the DSA core's side would have been to not interfere with phylink's process of failing at all. This isn't possible because of regression concerns with pre-phylink DT blobs, but at least DSA should put a stop to the proliferation of more of such cases that rely on the workaround to skip phylink registration, and sanitize the environment that new drivers work in. To that end, create a list of compatible strings for which the workaround is preserved, and don't apply the workaround for any drivers outside that list (this includes new drivers). In some cases, we make the assumption that even existing drivers don't rely on DSA's workaround, and we do this by looking at the device trees in which they appear. We can't fully know what is the situation with downstream DT blobs, but we can guess the overall trend by studying the DT blobs that were submitted upstream. If there are upstream blobs that have lacking descriptions, we take it as very likely that there are many more downstream blobs that do so too. If all upstream blobs have complete descriptions, we take that as a hint that the driver is a candidate for enforcing strict DT bindings (considering that most bindings are copy-pasted). If there are no upstream DT blobs, we take the conservative route of allowing the workaround, unless the driver maintainer instructs us otherwise. The driver situation is as follows: ar9331 ~~~~~~ compatible strings: - qca,ar9331-switch 1 occurrence in mainline device trees, part of SoC dtsi (arch/mips/boot/dts/qca/ar9331.dtsi), description is not problematic. Verdict: opt into strict DT bindings and out of workarounds. b53 ~~~ compatible strings: - brcm,bcm5325 - brcm,bcm53115 - brcm,bcm53125 - brcm,bcm53128 - brcm,bcm5365 - brcm,bcm5389 - brcm,bcm5395 - brcm,bcm5397 - brcm,bcm5398 - brcm,bcm53010-srab - brcm,bcm53011-srab - brcm,bcm53012-srab - brcm,bcm53018-srab - brcm,bcm53019-srab - brcm,bcm5301x-srab - brcm,bcm11360-srab - brcm,bcm58522-srab - brcm,bcm58525-srab - brcm,bcm58535-srab - brcm,bcm58622-srab - brcm,bcm58623-srab - brcm,bcm58625-srab - brcm,bcm88312-srab - brcm,cygnus-srab - brcm,nsp-srab - brcm,omega-srab - brcm,bcm3384-switch - brcm,bcm6328-switch - brcm,bcm6368-switch - brcm,bcm63xx-switch I've found at least these mainline DT blobs with problems: arch/arm/boot/dts/bcm47094-linksys-panamera.dts - lacks phy-mode arch/arm/boot/dts/bcm47189-tenda-ac9.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47081-luxul-xap-1410.dts arch/arm/boot/dts/bcm47081-luxul-xwr-1200.dts arch/arm/boot/dts/bcm47081-buffalo-wzr-600dhp2.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47094-luxul-xbr-4500.dts arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts arch/arm/boot/dts/bcm4708-luxul-xap-1510.dts arch/arm/boot/dts/bcm953012er.dts arch/arm/boot/dts/bcm4708-netgear-r6250.dts arch/arm/boot/dts/bcm4708-buffalo-wzr-1166dhp-common.dtsi arch/arm/boot/dts/bcm4708-luxul-xwc-1000.dts arch/arm/boot/dts/bcm47094-luxul-abr-4500.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm53016-meraki-mr32.dts - lacks phy-mode Verdict: opt into DSA workarounds. bcm_sf2 ~~~~~~~ compatible strings: - brcm,bcm4908-switch - brcm,bcm7445-switch-v4.0 - brcm,bcm7278-switch-v4.0 - brcm,bcm7278-switch-v4.8 A single occurrence in mainline (arch/arm64/boot/dts/broadcom/bcm4908/bcm4908.dtsi), part of a SoC dtsi, valid description. Florian Fainelli explains that most of the bcm_sf2 device trees lack a full description for the internal IMP ports. Verdict: opt the BCM4908 into strict DT bindings, and opt the rest into the workarounds. Note that even though BCM4908 has strict DT bindings, it still does not register with phylink on the IMP port due to it implementing ->adjust_link(). hellcreek ~~~~~~~~~ compatible strings: - hirschmann,hellcreek-de1soc-r1 No occurrence in mainline device trees. Kurt Kanzenbach explains that the downstream device trees lacked phy-mode and fixed link, and needed work, but were fixed in the meantime. Verdict: opt into strict DT bindings and out of workarounds. lan9303 ~~~~~~~ compatible strings: - smsc,lan9303-mdio - smsc,lan9303-i2c 1 occurrence in mainline device trees: arch/arm/boot/dts/imx53-kp-hsc.dts - no phy-mode, no fixed-link Verdict: opt out of strict DT bindings and into workarounds. lantiq_gswip ~~~~~~~~~~~~ compatible strings: - lantiq,xrx200-gswip - lantiq,xrx300-gswip - lantiq,xrx330-gswip No occurrences in mainline device trees. Martin Blumenstingl confirms that the downstream OpenWrt device trees lack a proper fixed-link and need work, and that the incomplete description can even be seen in the example from Documentation/devicetree/bindings/net/dsa/lantiq-gswip.txt. Verdict: opt out of strict DT bindings and into workarounds. microchip ksz ~~~~~~~~~~~~~ compatible strings: - microchip,ksz8765 - microchip,ksz8794 - microchip,ksz8795 - microchip,ksz8863 - microchip,ksz8873 - microchip,ksz9477 - microchip,ksz9897 - microchip,ksz9893 - microchip,ksz9563 - microchip,ksz8563 - microchip,ksz9567 - microchip,lan9370 - microchip,lan9371 - microchip,lan9372 - microchip,lan9373 - microchip,lan9374 5 occurrences in mainline device trees, all descriptions are valid. But we had a snafu for the ksz8795 and ksz9477 drivers where the phy-mode property would be expected to be located directly under the 'switch' node rather than under a port OF node. It was fixed by commit edecfa98f602 ("net: dsa: microchip: look for phy-mode in port nodes"). The driver still has compatibility with the old DT blobs. The lan937x support was added later than the above snafu was fixed, and even though it has support for the broken DT blobs by virtue of sharing a common probing function, I'll take it that its DT blobs are correct. Verdict: opt lan937x into strict DT bindings, and the others out. mt7530 ~~~~~~ compatible strings - mediatek,mt7621 - mediatek,mt7530 - mediatek,mt7531 Multiple occurrences in mainline device trees, one is part of an SoC dtsi (arch/mips/boot/dts/ralink/mt7621.dtsi), all descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. mv88e6060 ~~~~~~~~~ compatible string: - marvell,mv88e6060 no occurrences in mainline, nobody knows anybody who uses it. Verdict: opt out of strict DT bindings and into workarounds. mv88e6xxx ~~~~~~~~~ compatible strings: - marvell,mv88e6085 - marvell,mv88e6190 - marvell,mv88e6250 Device trees that have incomplete descriptions of CPU or DSA ports: arch/arm64/boot/dts/freescale/imx8mq-zii-ultra.dtsi - lacks phy-mode arch/arm64/boot/dts/marvell/cn9130-crb.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-spu3.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-mv88f6281gtw-ge.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-spb4.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-cfu1.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-dev-rev-c.dts - lacks phy-mode on CPU port, fixed-link on DSA ports arch/arm/boot/dts/vf610-zii-dev-rev-b.dts - lacks phy-mode on CPU port arch/arm/boot/dts/armada-381-netgear-gs110emx.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-scu4-aib.dts - lacks fixed-link on xgmii DSA ports and/or in-band-status on 2500base-x DSA ports, and phy-mode on CPU port arch/arm/boot/dts/imx6qdl-gw5904.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-385-clearfog-gtr-l8.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-dtu.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-dir665.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-rd88f6281.dtsi - lacks phy-mode arch/arm/boot/dts/orion5x-netgear-wnr854t.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-388-clearfog.dts - lacks phy-mode arch/arm/boot/dts/armada-xp-linksys-mamba.dts - lacks phy-mode arch/arm/boot/dts/armada-385-linksys.dtsi - lacks phy-mode arch/arm/boot/dts/imx6q-b450v3.dts arch/arm/boot/dts/imx6q-b850v3.dts - has a phy-handle but not a phy-mode? arch/arm/boot/dts/armada-370-rd.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-linksys-viper.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-rdu1.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-scu2-mezz.dts - lacks phy-mode arch/arm/boot/dts/imx6qdl-zii-rdu2.dtsi - lacks phy-mode arch/arm/boot/dts/armada-385-clearfog-gtr-s4.dts - lacks phy-mode and fixed-link Verdict: opt out of strict DT bindings and into workarounds. ocelot ~~~~~~ compatible strings: - mscc,vsc9953-switch - felix (arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi) is a PCI device, has no compatible string 2 occurrences in mainline, both are part of SoC dtsi and complete. Verdict: opt into strict DT bindings and out of workarounds. qca8k ~~~~~ compatible strings: - qca,qca8327 - qca,qca8328 - qca,qca8334 - qca,qca8337 5 occurrences in mainline device trees, none of the descriptions are problematic. Verdict: opt into strict DT bindings and out of workarounds. realtek ~~~~~~~ compatible strings: - realtek,rtl8366rb - realtek,rtl8365mb 2 occurrences in mainline, both descriptions are fine, additionally rtl8365mb.c has a comment "The device tree firmware should also specify the link partner of the extension port - either via a fixed-link or other phy-handle." Verdict: opt into strict DT bindings and out of workarounds. rzn1_a5psw ~~~~~~~~~~ compatible strings: - renesas,rzn1-a5psw One single occurrence, part of SoC dtsi (arch/arm/boot/dts/r9a06g032.dtsi), description is fine. Verdict: opt into strict DT bindings and out of workarounds. sja1105 ~~~~~~~ Driver already validates its port OF nodes in sja1105_parse_ports_node(). Verdict: opt into strict DT bindings and out of workarounds. vsc73xx ~~~~~~~ compatible strings: - vitesse,vsc7385 - vitesse,vsc7388 - vitesse,vsc7395 - vitesse,vsc7398 2 occurrences in mainline device trees, both descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. xrs700x ~~~~~~~ compatible strings: - arrow,xrs7003e - arrow,xrs7003f - arrow,xrs7004e - arrow,xrs7004f no occurrences in mainline, we don't know. Verdict: opt out of strict DT bindings and into workarounds. Because there is a pattern where newly added switches reuse existing drivers more often than introducing new ones, I've opted for deciding who gets to opt into the workaround based on an OF compatible match table in the DSA core. The alternative would have been to add another boolean property to struct dsa_switch, like configure_vlan_while_not_filtering. But this avoids situations where sometimes driver maintainers obfuscate what goes on by sharing a common probing function, and therefore making new switches inherit old quirks. Side note, we also warn about missing properties for drivers that rely on the workaround. This isn't an indication that we'll break compatibility with those DT blobs any time soon, but is rather done to raise awareness about the change, for future DT blob authors. Cc: Rob Herring <robh+dt@kernel.org> Cc: Frank Rowand <frowand.list@gmail.com> Acked-by: Alvin Šipraga <alsi@bang-olufsen.dk> # realtek Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-08-18 11:55:00 +00:00
* invoked for them.
*
* If you are considering expanding this table for newly introduced switches,
* think again. It is OK to remove switches from this table if there aren't DT
* blobs in circulation which rely on defaulting the shared ports.
*/
static const char * const dsa_switches_apply_workarounds[] = {
#if IS_ENABLED(CONFIG_NET_DSA_XRS700X)
"arrow,xrs7003e",
"arrow,xrs7003f",
"arrow,xrs7004e",
"arrow,xrs7004f",
#endif
#if IS_ENABLED(CONFIG_B53)
"brcm,bcm5325",
"brcm,bcm53115",
"brcm,bcm53125",
"brcm,bcm53128",
"brcm,bcm5365",
"brcm,bcm5389",
"brcm,bcm5395",
"brcm,bcm5397",
"brcm,bcm5398",
"brcm,bcm53010-srab",
"brcm,bcm53011-srab",
"brcm,bcm53012-srab",
"brcm,bcm53018-srab",
"brcm,bcm53019-srab",
"brcm,bcm5301x-srab",
"brcm,bcm11360-srab",
"brcm,bcm58522-srab",
"brcm,bcm58525-srab",
"brcm,bcm58535-srab",
"brcm,bcm58622-srab",
"brcm,bcm58623-srab",
"brcm,bcm58625-srab",
"brcm,bcm88312-srab",
"brcm,cygnus-srab",
"brcm,nsp-srab",
"brcm,omega-srab",
"brcm,bcm3384-switch",
"brcm,bcm6328-switch",
"brcm,bcm6368-switch",
"brcm,bcm63xx-switch",
#endif
#if IS_ENABLED(CONFIG_NET_DSA_BCM_SF2)
"brcm,bcm7445-switch-v4.0",
"brcm,bcm7278-switch-v4.0",
"brcm,bcm7278-switch-v4.8",
#endif
#if IS_ENABLED(CONFIG_NET_DSA_LANTIQ_GSWIP)
"lantiq,xrx200-gswip",
"lantiq,xrx300-gswip",
"lantiq,xrx330-gswip",
#endif
#if IS_ENABLED(CONFIG_NET_DSA_MV88E6060)
"marvell,mv88e6060",
#endif
#if IS_ENABLED(CONFIG_NET_DSA_MV88E6XXX)
"marvell,mv88e6085",
"marvell,mv88e6190",
"marvell,mv88e6250",
#endif
#if IS_ENABLED(CONFIG_NET_DSA_MICROCHIP_KSZ_COMMON)
"microchip,ksz8765",
"microchip,ksz8794",
"microchip,ksz8795",
"microchip,ksz8863",
"microchip,ksz8873",
"microchip,ksz9477",
"microchip,ksz9897",
"microchip,ksz9893",
"microchip,ksz9563",
"microchip,ksz8563",
"microchip,ksz9567",
#endif
#if IS_ENABLED(CONFIG_NET_DSA_SMSC_LAN9303_MDIO)
"smsc,lan9303-mdio",
#endif
#if IS_ENABLED(CONFIG_NET_DSA_SMSC_LAN9303_I2C)
"smsc,lan9303-i2c",
#endif
NULL,
};
static void dsa_shared_port_validate_of(struct dsa_port *dp,
bool *missing_phy_mode,
bool *missing_link_description)
{
struct device_node *dn = dp->dn, *phy_np;
struct dsa_switch *ds = dp->ds;
phy_interface_t mode;
*missing_phy_mode = false;
*missing_link_description = false;
if (of_get_phy_mode(dn, &mode)) {
*missing_phy_mode = true;
dev_err(ds->dev,
"OF node %pOF of %s port %d lacks the required \"phy-mode\" property\n",
dn, dsa_port_is_cpu(dp) ? "CPU" : "DSA", dp->index);
}
/* Note: of_phy_is_fixed_link() also returns true for
* managed = "in-band-status"
*/
if (of_phy_is_fixed_link(dn))
return;
phy_np = of_parse_phandle(dn, "phy-handle", 0);
if (phy_np) {
of_node_put(phy_np);
return;
}
*missing_link_description = true;
dev_err(ds->dev,
"OF node %pOF of %s port %d lacks the required \"phy-handle\", \"fixed-link\" or \"managed\" properties\n",
dn, dsa_port_is_cpu(dp) ? "CPU" : "DSA", dp->index);
}
int dsa_shared_port_link_register_of(struct dsa_port *dp)
{
struct dsa_switch *ds = dp->ds;
net: dsa: make phylink-related OF properties mandatory on DSA and CPU ports Early DSA drivers were kind of simplistic in that they assumed a fairly narrow hardware layout. User ports would have integrated PHYs at an internal MDIO address that is derivable from the port number, and shared (DSA and CPU) ports would have an MII-style (serial or parallel) connection to another MAC. Phylib and then phylink were used to drive the internal PHYs, and this needed little to no description through the platform data structures. Bringing up the shared ports at the maximum supported link speed was the responsibility of the drivers. As a result of this, when these early drivers were converted from platform data to the new DSA OF bindings, there was no link information translated into the first DT bindings. https://lore.kernel.org/all/YtXFtTsf++AeDm1l@lunn.ch/ Later, phylink was adopted for shared ports as well, and today we have a workaround in place, introduced by commit a20f997010c4 ("net: dsa: Don't instantiate phylink for CPU/DSA ports unless needed"). There, DSA checks for the presence of phy-handle/fixed-link/managed OF properties, and if missing, phylink registration would be skipped. This is because phylink is optional for some drivers (the shared ports already work without it), but the process of starting to register a port with phylink is irreversible: if phylink_create() fails to find the fwnode properties it needs, it bails out and it leaves the ports inoperational (because phylink expects ports to be initially down, so DSA necessarily takes them down, and doesn't know how to put them back up again). DSA being a common framework, new drivers opt into this workaround willy-nilly, but the ideal behavior from the DSA core's side would have been to not interfere with phylink's process of failing at all. This isn't possible because of regression concerns with pre-phylink DT blobs, but at least DSA should put a stop to the proliferation of more of such cases that rely on the workaround to skip phylink registration, and sanitize the environment that new drivers work in. To that end, create a list of compatible strings for which the workaround is preserved, and don't apply the workaround for any drivers outside that list (this includes new drivers). In some cases, we make the assumption that even existing drivers don't rely on DSA's workaround, and we do this by looking at the device trees in which they appear. We can't fully know what is the situation with downstream DT blobs, but we can guess the overall trend by studying the DT blobs that were submitted upstream. If there are upstream blobs that have lacking descriptions, we take it as very likely that there are many more downstream blobs that do so too. If all upstream blobs have complete descriptions, we take that as a hint that the driver is a candidate for enforcing strict DT bindings (considering that most bindings are copy-pasted). If there are no upstream DT blobs, we take the conservative route of allowing the workaround, unless the driver maintainer instructs us otherwise. The driver situation is as follows: ar9331 ~~~~~~ compatible strings: - qca,ar9331-switch 1 occurrence in mainline device trees, part of SoC dtsi (arch/mips/boot/dts/qca/ar9331.dtsi), description is not problematic. Verdict: opt into strict DT bindings and out of workarounds. b53 ~~~ compatible strings: - brcm,bcm5325 - brcm,bcm53115 - brcm,bcm53125 - brcm,bcm53128 - brcm,bcm5365 - brcm,bcm5389 - brcm,bcm5395 - brcm,bcm5397 - brcm,bcm5398 - brcm,bcm53010-srab - brcm,bcm53011-srab - brcm,bcm53012-srab - brcm,bcm53018-srab - brcm,bcm53019-srab - brcm,bcm5301x-srab - brcm,bcm11360-srab - brcm,bcm58522-srab - brcm,bcm58525-srab - brcm,bcm58535-srab - brcm,bcm58622-srab - brcm,bcm58623-srab - brcm,bcm58625-srab - brcm,bcm88312-srab - brcm,cygnus-srab - brcm,nsp-srab - brcm,omega-srab - brcm,bcm3384-switch - brcm,bcm6328-switch - brcm,bcm6368-switch - brcm,bcm63xx-switch I've found at least these mainline DT blobs with problems: arch/arm/boot/dts/bcm47094-linksys-panamera.dts - lacks phy-mode arch/arm/boot/dts/bcm47189-tenda-ac9.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47081-luxul-xap-1410.dts arch/arm/boot/dts/bcm47081-luxul-xwr-1200.dts arch/arm/boot/dts/bcm47081-buffalo-wzr-600dhp2.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47094-luxul-xbr-4500.dts arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts arch/arm/boot/dts/bcm4708-luxul-xap-1510.dts arch/arm/boot/dts/bcm953012er.dts arch/arm/boot/dts/bcm4708-netgear-r6250.dts arch/arm/boot/dts/bcm4708-buffalo-wzr-1166dhp-common.dtsi arch/arm/boot/dts/bcm4708-luxul-xwc-1000.dts arch/arm/boot/dts/bcm47094-luxul-abr-4500.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm53016-meraki-mr32.dts - lacks phy-mode Verdict: opt into DSA workarounds. bcm_sf2 ~~~~~~~ compatible strings: - brcm,bcm4908-switch - brcm,bcm7445-switch-v4.0 - brcm,bcm7278-switch-v4.0 - brcm,bcm7278-switch-v4.8 A single occurrence in mainline (arch/arm64/boot/dts/broadcom/bcm4908/bcm4908.dtsi), part of a SoC dtsi, valid description. Florian Fainelli explains that most of the bcm_sf2 device trees lack a full description for the internal IMP ports. Verdict: opt the BCM4908 into strict DT bindings, and opt the rest into the workarounds. Note that even though BCM4908 has strict DT bindings, it still does not register with phylink on the IMP port due to it implementing ->adjust_link(). hellcreek ~~~~~~~~~ compatible strings: - hirschmann,hellcreek-de1soc-r1 No occurrence in mainline device trees. Kurt Kanzenbach explains that the downstream device trees lacked phy-mode and fixed link, and needed work, but were fixed in the meantime. Verdict: opt into strict DT bindings and out of workarounds. lan9303 ~~~~~~~ compatible strings: - smsc,lan9303-mdio - smsc,lan9303-i2c 1 occurrence in mainline device trees: arch/arm/boot/dts/imx53-kp-hsc.dts - no phy-mode, no fixed-link Verdict: opt out of strict DT bindings and into workarounds. lantiq_gswip ~~~~~~~~~~~~ compatible strings: - lantiq,xrx200-gswip - lantiq,xrx300-gswip - lantiq,xrx330-gswip No occurrences in mainline device trees. Martin Blumenstingl confirms that the downstream OpenWrt device trees lack a proper fixed-link and need work, and that the incomplete description can even be seen in the example from Documentation/devicetree/bindings/net/dsa/lantiq-gswip.txt. Verdict: opt out of strict DT bindings and into workarounds. microchip ksz ~~~~~~~~~~~~~ compatible strings: - microchip,ksz8765 - microchip,ksz8794 - microchip,ksz8795 - microchip,ksz8863 - microchip,ksz8873 - microchip,ksz9477 - microchip,ksz9897 - microchip,ksz9893 - microchip,ksz9563 - microchip,ksz8563 - microchip,ksz9567 - microchip,lan9370 - microchip,lan9371 - microchip,lan9372 - microchip,lan9373 - microchip,lan9374 5 occurrences in mainline device trees, all descriptions are valid. But we had a snafu for the ksz8795 and ksz9477 drivers where the phy-mode property would be expected to be located directly under the 'switch' node rather than under a port OF node. It was fixed by commit edecfa98f602 ("net: dsa: microchip: look for phy-mode in port nodes"). The driver still has compatibility with the old DT blobs. The lan937x support was added later than the above snafu was fixed, and even though it has support for the broken DT blobs by virtue of sharing a common probing function, I'll take it that its DT blobs are correct. Verdict: opt lan937x into strict DT bindings, and the others out. mt7530 ~~~~~~ compatible strings - mediatek,mt7621 - mediatek,mt7530 - mediatek,mt7531 Multiple occurrences in mainline device trees, one is part of an SoC dtsi (arch/mips/boot/dts/ralink/mt7621.dtsi), all descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. mv88e6060 ~~~~~~~~~ compatible string: - marvell,mv88e6060 no occurrences in mainline, nobody knows anybody who uses it. Verdict: opt out of strict DT bindings and into workarounds. mv88e6xxx ~~~~~~~~~ compatible strings: - marvell,mv88e6085 - marvell,mv88e6190 - marvell,mv88e6250 Device trees that have incomplete descriptions of CPU or DSA ports: arch/arm64/boot/dts/freescale/imx8mq-zii-ultra.dtsi - lacks phy-mode arch/arm64/boot/dts/marvell/cn9130-crb.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-spu3.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-mv88f6281gtw-ge.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-spb4.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-cfu1.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-dev-rev-c.dts - lacks phy-mode on CPU port, fixed-link on DSA ports arch/arm/boot/dts/vf610-zii-dev-rev-b.dts - lacks phy-mode on CPU port arch/arm/boot/dts/armada-381-netgear-gs110emx.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-scu4-aib.dts - lacks fixed-link on xgmii DSA ports and/or in-band-status on 2500base-x DSA ports, and phy-mode on CPU port arch/arm/boot/dts/imx6qdl-gw5904.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-385-clearfog-gtr-l8.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-dtu.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-dir665.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-rd88f6281.dtsi - lacks phy-mode arch/arm/boot/dts/orion5x-netgear-wnr854t.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-388-clearfog.dts - lacks phy-mode arch/arm/boot/dts/armada-xp-linksys-mamba.dts - lacks phy-mode arch/arm/boot/dts/armada-385-linksys.dtsi - lacks phy-mode arch/arm/boot/dts/imx6q-b450v3.dts arch/arm/boot/dts/imx6q-b850v3.dts - has a phy-handle but not a phy-mode? arch/arm/boot/dts/armada-370-rd.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-linksys-viper.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-rdu1.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-scu2-mezz.dts - lacks phy-mode arch/arm/boot/dts/imx6qdl-zii-rdu2.dtsi - lacks phy-mode arch/arm/boot/dts/armada-385-clearfog-gtr-s4.dts - lacks phy-mode and fixed-link Verdict: opt out of strict DT bindings and into workarounds. ocelot ~~~~~~ compatible strings: - mscc,vsc9953-switch - felix (arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi) is a PCI device, has no compatible string 2 occurrences in mainline, both are part of SoC dtsi and complete. Verdict: opt into strict DT bindings and out of workarounds. qca8k ~~~~~ compatible strings: - qca,qca8327 - qca,qca8328 - qca,qca8334 - qca,qca8337 5 occurrences in mainline device trees, none of the descriptions are problematic. Verdict: opt into strict DT bindings and out of workarounds. realtek ~~~~~~~ compatible strings: - realtek,rtl8366rb - realtek,rtl8365mb 2 occurrences in mainline, both descriptions are fine, additionally rtl8365mb.c has a comment "The device tree firmware should also specify the link partner of the extension port - either via a fixed-link or other phy-handle." Verdict: opt into strict DT bindings and out of workarounds. rzn1_a5psw ~~~~~~~~~~ compatible strings: - renesas,rzn1-a5psw One single occurrence, part of SoC dtsi (arch/arm/boot/dts/r9a06g032.dtsi), description is fine. Verdict: opt into strict DT bindings and out of workarounds. sja1105 ~~~~~~~ Driver already validates its port OF nodes in sja1105_parse_ports_node(). Verdict: opt into strict DT bindings and out of workarounds. vsc73xx ~~~~~~~ compatible strings: - vitesse,vsc7385 - vitesse,vsc7388 - vitesse,vsc7395 - vitesse,vsc7398 2 occurrences in mainline device trees, both descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. xrs700x ~~~~~~~ compatible strings: - arrow,xrs7003e - arrow,xrs7003f - arrow,xrs7004e - arrow,xrs7004f no occurrences in mainline, we don't know. Verdict: opt out of strict DT bindings and into workarounds. Because there is a pattern where newly added switches reuse existing drivers more often than introducing new ones, I've opted for deciding who gets to opt into the workaround based on an OF compatible match table in the DSA core. The alternative would have been to add another boolean property to struct dsa_switch, like configure_vlan_while_not_filtering. But this avoids situations where sometimes driver maintainers obfuscate what goes on by sharing a common probing function, and therefore making new switches inherit old quirks. Side note, we also warn about missing properties for drivers that rely on the workaround. This isn't an indication that we'll break compatibility with those DT blobs any time soon, but is rather done to raise awareness about the change, for future DT blob authors. Cc: Rob Herring <robh+dt@kernel.org> Cc: Frank Rowand <frowand.list@gmail.com> Acked-by: Alvin Šipraga <alsi@bang-olufsen.dk> # realtek Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-08-18 11:55:00 +00:00
bool missing_link_description;
bool missing_phy_mode;
int port = dp->index;
net: dsa: make phylink-related OF properties mandatory on DSA and CPU ports Early DSA drivers were kind of simplistic in that they assumed a fairly narrow hardware layout. User ports would have integrated PHYs at an internal MDIO address that is derivable from the port number, and shared (DSA and CPU) ports would have an MII-style (serial or parallel) connection to another MAC. Phylib and then phylink were used to drive the internal PHYs, and this needed little to no description through the platform data structures. Bringing up the shared ports at the maximum supported link speed was the responsibility of the drivers. As a result of this, when these early drivers were converted from platform data to the new DSA OF bindings, there was no link information translated into the first DT bindings. https://lore.kernel.org/all/YtXFtTsf++AeDm1l@lunn.ch/ Later, phylink was adopted for shared ports as well, and today we have a workaround in place, introduced by commit a20f997010c4 ("net: dsa: Don't instantiate phylink for CPU/DSA ports unless needed"). There, DSA checks for the presence of phy-handle/fixed-link/managed OF properties, and if missing, phylink registration would be skipped. This is because phylink is optional for some drivers (the shared ports already work without it), but the process of starting to register a port with phylink is irreversible: if phylink_create() fails to find the fwnode properties it needs, it bails out and it leaves the ports inoperational (because phylink expects ports to be initially down, so DSA necessarily takes them down, and doesn't know how to put them back up again). DSA being a common framework, new drivers opt into this workaround willy-nilly, but the ideal behavior from the DSA core's side would have been to not interfere with phylink's process of failing at all. This isn't possible because of regression concerns with pre-phylink DT blobs, but at least DSA should put a stop to the proliferation of more of such cases that rely on the workaround to skip phylink registration, and sanitize the environment that new drivers work in. To that end, create a list of compatible strings for which the workaround is preserved, and don't apply the workaround for any drivers outside that list (this includes new drivers). In some cases, we make the assumption that even existing drivers don't rely on DSA's workaround, and we do this by looking at the device trees in which they appear. We can't fully know what is the situation with downstream DT blobs, but we can guess the overall trend by studying the DT blobs that were submitted upstream. If there are upstream blobs that have lacking descriptions, we take it as very likely that there are many more downstream blobs that do so too. If all upstream blobs have complete descriptions, we take that as a hint that the driver is a candidate for enforcing strict DT bindings (considering that most bindings are copy-pasted). If there are no upstream DT blobs, we take the conservative route of allowing the workaround, unless the driver maintainer instructs us otherwise. The driver situation is as follows: ar9331 ~~~~~~ compatible strings: - qca,ar9331-switch 1 occurrence in mainline device trees, part of SoC dtsi (arch/mips/boot/dts/qca/ar9331.dtsi), description is not problematic. Verdict: opt into strict DT bindings and out of workarounds. b53 ~~~ compatible strings: - brcm,bcm5325 - brcm,bcm53115 - brcm,bcm53125 - brcm,bcm53128 - brcm,bcm5365 - brcm,bcm5389 - brcm,bcm5395 - brcm,bcm5397 - brcm,bcm5398 - brcm,bcm53010-srab - brcm,bcm53011-srab - brcm,bcm53012-srab - brcm,bcm53018-srab - brcm,bcm53019-srab - brcm,bcm5301x-srab - brcm,bcm11360-srab - brcm,bcm58522-srab - brcm,bcm58525-srab - brcm,bcm58535-srab - brcm,bcm58622-srab - brcm,bcm58623-srab - brcm,bcm58625-srab - brcm,bcm88312-srab - brcm,cygnus-srab - brcm,nsp-srab - brcm,omega-srab - brcm,bcm3384-switch - brcm,bcm6328-switch - brcm,bcm6368-switch - brcm,bcm63xx-switch I've found at least these mainline DT blobs with problems: arch/arm/boot/dts/bcm47094-linksys-panamera.dts - lacks phy-mode arch/arm/boot/dts/bcm47189-tenda-ac9.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47081-luxul-xap-1410.dts arch/arm/boot/dts/bcm47081-luxul-xwr-1200.dts arch/arm/boot/dts/bcm47081-buffalo-wzr-600dhp2.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47094-luxul-xbr-4500.dts arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts arch/arm/boot/dts/bcm4708-luxul-xap-1510.dts arch/arm/boot/dts/bcm953012er.dts arch/arm/boot/dts/bcm4708-netgear-r6250.dts arch/arm/boot/dts/bcm4708-buffalo-wzr-1166dhp-common.dtsi arch/arm/boot/dts/bcm4708-luxul-xwc-1000.dts arch/arm/boot/dts/bcm47094-luxul-abr-4500.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm53016-meraki-mr32.dts - lacks phy-mode Verdict: opt into DSA workarounds. bcm_sf2 ~~~~~~~ compatible strings: - brcm,bcm4908-switch - brcm,bcm7445-switch-v4.0 - brcm,bcm7278-switch-v4.0 - brcm,bcm7278-switch-v4.8 A single occurrence in mainline (arch/arm64/boot/dts/broadcom/bcm4908/bcm4908.dtsi), part of a SoC dtsi, valid description. Florian Fainelli explains that most of the bcm_sf2 device trees lack a full description for the internal IMP ports. Verdict: opt the BCM4908 into strict DT bindings, and opt the rest into the workarounds. Note that even though BCM4908 has strict DT bindings, it still does not register with phylink on the IMP port due to it implementing ->adjust_link(). hellcreek ~~~~~~~~~ compatible strings: - hirschmann,hellcreek-de1soc-r1 No occurrence in mainline device trees. Kurt Kanzenbach explains that the downstream device trees lacked phy-mode and fixed link, and needed work, but were fixed in the meantime. Verdict: opt into strict DT bindings and out of workarounds. lan9303 ~~~~~~~ compatible strings: - smsc,lan9303-mdio - smsc,lan9303-i2c 1 occurrence in mainline device trees: arch/arm/boot/dts/imx53-kp-hsc.dts - no phy-mode, no fixed-link Verdict: opt out of strict DT bindings and into workarounds. lantiq_gswip ~~~~~~~~~~~~ compatible strings: - lantiq,xrx200-gswip - lantiq,xrx300-gswip - lantiq,xrx330-gswip No occurrences in mainline device trees. Martin Blumenstingl confirms that the downstream OpenWrt device trees lack a proper fixed-link and need work, and that the incomplete description can even be seen in the example from Documentation/devicetree/bindings/net/dsa/lantiq-gswip.txt. Verdict: opt out of strict DT bindings and into workarounds. microchip ksz ~~~~~~~~~~~~~ compatible strings: - microchip,ksz8765 - microchip,ksz8794 - microchip,ksz8795 - microchip,ksz8863 - microchip,ksz8873 - microchip,ksz9477 - microchip,ksz9897 - microchip,ksz9893 - microchip,ksz9563 - microchip,ksz8563 - microchip,ksz9567 - microchip,lan9370 - microchip,lan9371 - microchip,lan9372 - microchip,lan9373 - microchip,lan9374 5 occurrences in mainline device trees, all descriptions are valid. But we had a snafu for the ksz8795 and ksz9477 drivers where the phy-mode property would be expected to be located directly under the 'switch' node rather than under a port OF node. It was fixed by commit edecfa98f602 ("net: dsa: microchip: look for phy-mode in port nodes"). The driver still has compatibility with the old DT blobs. The lan937x support was added later than the above snafu was fixed, and even though it has support for the broken DT blobs by virtue of sharing a common probing function, I'll take it that its DT blobs are correct. Verdict: opt lan937x into strict DT bindings, and the others out. mt7530 ~~~~~~ compatible strings - mediatek,mt7621 - mediatek,mt7530 - mediatek,mt7531 Multiple occurrences in mainline device trees, one is part of an SoC dtsi (arch/mips/boot/dts/ralink/mt7621.dtsi), all descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. mv88e6060 ~~~~~~~~~ compatible string: - marvell,mv88e6060 no occurrences in mainline, nobody knows anybody who uses it. Verdict: opt out of strict DT bindings and into workarounds. mv88e6xxx ~~~~~~~~~ compatible strings: - marvell,mv88e6085 - marvell,mv88e6190 - marvell,mv88e6250 Device trees that have incomplete descriptions of CPU or DSA ports: arch/arm64/boot/dts/freescale/imx8mq-zii-ultra.dtsi - lacks phy-mode arch/arm64/boot/dts/marvell/cn9130-crb.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-spu3.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-mv88f6281gtw-ge.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-spb4.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-cfu1.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-dev-rev-c.dts - lacks phy-mode on CPU port, fixed-link on DSA ports arch/arm/boot/dts/vf610-zii-dev-rev-b.dts - lacks phy-mode on CPU port arch/arm/boot/dts/armada-381-netgear-gs110emx.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-scu4-aib.dts - lacks fixed-link on xgmii DSA ports and/or in-band-status on 2500base-x DSA ports, and phy-mode on CPU port arch/arm/boot/dts/imx6qdl-gw5904.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-385-clearfog-gtr-l8.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-dtu.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-dir665.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-rd88f6281.dtsi - lacks phy-mode arch/arm/boot/dts/orion5x-netgear-wnr854t.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-388-clearfog.dts - lacks phy-mode arch/arm/boot/dts/armada-xp-linksys-mamba.dts - lacks phy-mode arch/arm/boot/dts/armada-385-linksys.dtsi - lacks phy-mode arch/arm/boot/dts/imx6q-b450v3.dts arch/arm/boot/dts/imx6q-b850v3.dts - has a phy-handle but not a phy-mode? arch/arm/boot/dts/armada-370-rd.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-linksys-viper.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-rdu1.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-scu2-mezz.dts - lacks phy-mode arch/arm/boot/dts/imx6qdl-zii-rdu2.dtsi - lacks phy-mode arch/arm/boot/dts/armada-385-clearfog-gtr-s4.dts - lacks phy-mode and fixed-link Verdict: opt out of strict DT bindings and into workarounds. ocelot ~~~~~~ compatible strings: - mscc,vsc9953-switch - felix (arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi) is a PCI device, has no compatible string 2 occurrences in mainline, both are part of SoC dtsi and complete. Verdict: opt into strict DT bindings and out of workarounds. qca8k ~~~~~ compatible strings: - qca,qca8327 - qca,qca8328 - qca,qca8334 - qca,qca8337 5 occurrences in mainline device trees, none of the descriptions are problematic. Verdict: opt into strict DT bindings and out of workarounds. realtek ~~~~~~~ compatible strings: - realtek,rtl8366rb - realtek,rtl8365mb 2 occurrences in mainline, both descriptions are fine, additionally rtl8365mb.c has a comment "The device tree firmware should also specify the link partner of the extension port - either via a fixed-link or other phy-handle." Verdict: opt into strict DT bindings and out of workarounds. rzn1_a5psw ~~~~~~~~~~ compatible strings: - renesas,rzn1-a5psw One single occurrence, part of SoC dtsi (arch/arm/boot/dts/r9a06g032.dtsi), description is fine. Verdict: opt into strict DT bindings and out of workarounds. sja1105 ~~~~~~~ Driver already validates its port OF nodes in sja1105_parse_ports_node(). Verdict: opt into strict DT bindings and out of workarounds. vsc73xx ~~~~~~~ compatible strings: - vitesse,vsc7385 - vitesse,vsc7388 - vitesse,vsc7395 - vitesse,vsc7398 2 occurrences in mainline device trees, both descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. xrs700x ~~~~~~~ compatible strings: - arrow,xrs7003e - arrow,xrs7003f - arrow,xrs7004e - arrow,xrs7004f no occurrences in mainline, we don't know. Verdict: opt out of strict DT bindings and into workarounds. Because there is a pattern where newly added switches reuse existing drivers more often than introducing new ones, I've opted for deciding who gets to opt into the workaround based on an OF compatible match table in the DSA core. The alternative would have been to add another boolean property to struct dsa_switch, like configure_vlan_while_not_filtering. But this avoids situations where sometimes driver maintainers obfuscate what goes on by sharing a common probing function, and therefore making new switches inherit old quirks. Side note, we also warn about missing properties for drivers that rely on the workaround. This isn't an indication that we'll break compatibility with those DT blobs any time soon, but is rather done to raise awareness about the change, for future DT blob authors. Cc: Rob Herring <robh+dt@kernel.org> Cc: Frank Rowand <frowand.list@gmail.com> Acked-by: Alvin Šipraga <alsi@bang-olufsen.dk> # realtek Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-08-18 11:55:00 +00:00
dsa_shared_port_validate_of(dp, &missing_phy_mode,
&missing_link_description);
if ((missing_phy_mode || missing_link_description) &&
!of_device_compatible_match(ds->dev->of_node,
dsa_switches_apply_workarounds))
return -EINVAL;
if (!ds->ops->adjust_link) {
net: dsa: make phylink-related OF properties mandatory on DSA and CPU ports Early DSA drivers were kind of simplistic in that they assumed a fairly narrow hardware layout. User ports would have integrated PHYs at an internal MDIO address that is derivable from the port number, and shared (DSA and CPU) ports would have an MII-style (serial or parallel) connection to another MAC. Phylib and then phylink were used to drive the internal PHYs, and this needed little to no description through the platform data structures. Bringing up the shared ports at the maximum supported link speed was the responsibility of the drivers. As a result of this, when these early drivers were converted from platform data to the new DSA OF bindings, there was no link information translated into the first DT bindings. https://lore.kernel.org/all/YtXFtTsf++AeDm1l@lunn.ch/ Later, phylink was adopted for shared ports as well, and today we have a workaround in place, introduced by commit a20f997010c4 ("net: dsa: Don't instantiate phylink for CPU/DSA ports unless needed"). There, DSA checks for the presence of phy-handle/fixed-link/managed OF properties, and if missing, phylink registration would be skipped. This is because phylink is optional for some drivers (the shared ports already work without it), but the process of starting to register a port with phylink is irreversible: if phylink_create() fails to find the fwnode properties it needs, it bails out and it leaves the ports inoperational (because phylink expects ports to be initially down, so DSA necessarily takes them down, and doesn't know how to put them back up again). DSA being a common framework, new drivers opt into this workaround willy-nilly, but the ideal behavior from the DSA core's side would have been to not interfere with phylink's process of failing at all. This isn't possible because of regression concerns with pre-phylink DT blobs, but at least DSA should put a stop to the proliferation of more of such cases that rely on the workaround to skip phylink registration, and sanitize the environment that new drivers work in. To that end, create a list of compatible strings for which the workaround is preserved, and don't apply the workaround for any drivers outside that list (this includes new drivers). In some cases, we make the assumption that even existing drivers don't rely on DSA's workaround, and we do this by looking at the device trees in which they appear. We can't fully know what is the situation with downstream DT blobs, but we can guess the overall trend by studying the DT blobs that were submitted upstream. If there are upstream blobs that have lacking descriptions, we take it as very likely that there are many more downstream blobs that do so too. If all upstream blobs have complete descriptions, we take that as a hint that the driver is a candidate for enforcing strict DT bindings (considering that most bindings are copy-pasted). If there are no upstream DT blobs, we take the conservative route of allowing the workaround, unless the driver maintainer instructs us otherwise. The driver situation is as follows: ar9331 ~~~~~~ compatible strings: - qca,ar9331-switch 1 occurrence in mainline device trees, part of SoC dtsi (arch/mips/boot/dts/qca/ar9331.dtsi), description is not problematic. Verdict: opt into strict DT bindings and out of workarounds. b53 ~~~ compatible strings: - brcm,bcm5325 - brcm,bcm53115 - brcm,bcm53125 - brcm,bcm53128 - brcm,bcm5365 - brcm,bcm5389 - brcm,bcm5395 - brcm,bcm5397 - brcm,bcm5398 - brcm,bcm53010-srab - brcm,bcm53011-srab - brcm,bcm53012-srab - brcm,bcm53018-srab - brcm,bcm53019-srab - brcm,bcm5301x-srab - brcm,bcm11360-srab - brcm,bcm58522-srab - brcm,bcm58525-srab - brcm,bcm58535-srab - brcm,bcm58622-srab - brcm,bcm58623-srab - brcm,bcm58625-srab - brcm,bcm88312-srab - brcm,cygnus-srab - brcm,nsp-srab - brcm,omega-srab - brcm,bcm3384-switch - brcm,bcm6328-switch - brcm,bcm6368-switch - brcm,bcm63xx-switch I've found at least these mainline DT blobs with problems: arch/arm/boot/dts/bcm47094-linksys-panamera.dts - lacks phy-mode arch/arm/boot/dts/bcm47189-tenda-ac9.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47081-luxul-xap-1410.dts arch/arm/boot/dts/bcm47081-luxul-xwr-1200.dts arch/arm/boot/dts/bcm47081-buffalo-wzr-600dhp2.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47094-luxul-xbr-4500.dts arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts arch/arm/boot/dts/bcm4708-luxul-xap-1510.dts arch/arm/boot/dts/bcm953012er.dts arch/arm/boot/dts/bcm4708-netgear-r6250.dts arch/arm/boot/dts/bcm4708-buffalo-wzr-1166dhp-common.dtsi arch/arm/boot/dts/bcm4708-luxul-xwc-1000.dts arch/arm/boot/dts/bcm47094-luxul-abr-4500.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm53016-meraki-mr32.dts - lacks phy-mode Verdict: opt into DSA workarounds. bcm_sf2 ~~~~~~~ compatible strings: - brcm,bcm4908-switch - brcm,bcm7445-switch-v4.0 - brcm,bcm7278-switch-v4.0 - brcm,bcm7278-switch-v4.8 A single occurrence in mainline (arch/arm64/boot/dts/broadcom/bcm4908/bcm4908.dtsi), part of a SoC dtsi, valid description. Florian Fainelli explains that most of the bcm_sf2 device trees lack a full description for the internal IMP ports. Verdict: opt the BCM4908 into strict DT bindings, and opt the rest into the workarounds. Note that even though BCM4908 has strict DT bindings, it still does not register with phylink on the IMP port due to it implementing ->adjust_link(). hellcreek ~~~~~~~~~ compatible strings: - hirschmann,hellcreek-de1soc-r1 No occurrence in mainline device trees. Kurt Kanzenbach explains that the downstream device trees lacked phy-mode and fixed link, and needed work, but were fixed in the meantime. Verdict: opt into strict DT bindings and out of workarounds. lan9303 ~~~~~~~ compatible strings: - smsc,lan9303-mdio - smsc,lan9303-i2c 1 occurrence in mainline device trees: arch/arm/boot/dts/imx53-kp-hsc.dts - no phy-mode, no fixed-link Verdict: opt out of strict DT bindings and into workarounds. lantiq_gswip ~~~~~~~~~~~~ compatible strings: - lantiq,xrx200-gswip - lantiq,xrx300-gswip - lantiq,xrx330-gswip No occurrences in mainline device trees. Martin Blumenstingl confirms that the downstream OpenWrt device trees lack a proper fixed-link and need work, and that the incomplete description can even be seen in the example from Documentation/devicetree/bindings/net/dsa/lantiq-gswip.txt. Verdict: opt out of strict DT bindings and into workarounds. microchip ksz ~~~~~~~~~~~~~ compatible strings: - microchip,ksz8765 - microchip,ksz8794 - microchip,ksz8795 - microchip,ksz8863 - microchip,ksz8873 - microchip,ksz9477 - microchip,ksz9897 - microchip,ksz9893 - microchip,ksz9563 - microchip,ksz8563 - microchip,ksz9567 - microchip,lan9370 - microchip,lan9371 - microchip,lan9372 - microchip,lan9373 - microchip,lan9374 5 occurrences in mainline device trees, all descriptions are valid. But we had a snafu for the ksz8795 and ksz9477 drivers where the phy-mode property would be expected to be located directly under the 'switch' node rather than under a port OF node. It was fixed by commit edecfa98f602 ("net: dsa: microchip: look for phy-mode in port nodes"). The driver still has compatibility with the old DT blobs. The lan937x support was added later than the above snafu was fixed, and even though it has support for the broken DT blobs by virtue of sharing a common probing function, I'll take it that its DT blobs are correct. Verdict: opt lan937x into strict DT bindings, and the others out. mt7530 ~~~~~~ compatible strings - mediatek,mt7621 - mediatek,mt7530 - mediatek,mt7531 Multiple occurrences in mainline device trees, one is part of an SoC dtsi (arch/mips/boot/dts/ralink/mt7621.dtsi), all descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. mv88e6060 ~~~~~~~~~ compatible string: - marvell,mv88e6060 no occurrences in mainline, nobody knows anybody who uses it. Verdict: opt out of strict DT bindings and into workarounds. mv88e6xxx ~~~~~~~~~ compatible strings: - marvell,mv88e6085 - marvell,mv88e6190 - marvell,mv88e6250 Device trees that have incomplete descriptions of CPU or DSA ports: arch/arm64/boot/dts/freescale/imx8mq-zii-ultra.dtsi - lacks phy-mode arch/arm64/boot/dts/marvell/cn9130-crb.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-spu3.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-mv88f6281gtw-ge.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-spb4.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-cfu1.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-dev-rev-c.dts - lacks phy-mode on CPU port, fixed-link on DSA ports arch/arm/boot/dts/vf610-zii-dev-rev-b.dts - lacks phy-mode on CPU port arch/arm/boot/dts/armada-381-netgear-gs110emx.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-scu4-aib.dts - lacks fixed-link on xgmii DSA ports and/or in-band-status on 2500base-x DSA ports, and phy-mode on CPU port arch/arm/boot/dts/imx6qdl-gw5904.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-385-clearfog-gtr-l8.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-dtu.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-dir665.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-rd88f6281.dtsi - lacks phy-mode arch/arm/boot/dts/orion5x-netgear-wnr854t.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-388-clearfog.dts - lacks phy-mode arch/arm/boot/dts/armada-xp-linksys-mamba.dts - lacks phy-mode arch/arm/boot/dts/armada-385-linksys.dtsi - lacks phy-mode arch/arm/boot/dts/imx6q-b450v3.dts arch/arm/boot/dts/imx6q-b850v3.dts - has a phy-handle but not a phy-mode? arch/arm/boot/dts/armada-370-rd.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-linksys-viper.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-rdu1.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-scu2-mezz.dts - lacks phy-mode arch/arm/boot/dts/imx6qdl-zii-rdu2.dtsi - lacks phy-mode arch/arm/boot/dts/armada-385-clearfog-gtr-s4.dts - lacks phy-mode and fixed-link Verdict: opt out of strict DT bindings and into workarounds. ocelot ~~~~~~ compatible strings: - mscc,vsc9953-switch - felix (arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi) is a PCI device, has no compatible string 2 occurrences in mainline, both are part of SoC dtsi and complete. Verdict: opt into strict DT bindings and out of workarounds. qca8k ~~~~~ compatible strings: - qca,qca8327 - qca,qca8328 - qca,qca8334 - qca,qca8337 5 occurrences in mainline device trees, none of the descriptions are problematic. Verdict: opt into strict DT bindings and out of workarounds. realtek ~~~~~~~ compatible strings: - realtek,rtl8366rb - realtek,rtl8365mb 2 occurrences in mainline, both descriptions are fine, additionally rtl8365mb.c has a comment "The device tree firmware should also specify the link partner of the extension port - either via a fixed-link or other phy-handle." Verdict: opt into strict DT bindings and out of workarounds. rzn1_a5psw ~~~~~~~~~~ compatible strings: - renesas,rzn1-a5psw One single occurrence, part of SoC dtsi (arch/arm/boot/dts/r9a06g032.dtsi), description is fine. Verdict: opt into strict DT bindings and out of workarounds. sja1105 ~~~~~~~ Driver already validates its port OF nodes in sja1105_parse_ports_node(). Verdict: opt into strict DT bindings and out of workarounds. vsc73xx ~~~~~~~ compatible strings: - vitesse,vsc7385 - vitesse,vsc7388 - vitesse,vsc7395 - vitesse,vsc7398 2 occurrences in mainline device trees, both descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. xrs700x ~~~~~~~ compatible strings: - arrow,xrs7003e - arrow,xrs7003f - arrow,xrs7004e - arrow,xrs7004f no occurrences in mainline, we don't know. Verdict: opt out of strict DT bindings and into workarounds. Because there is a pattern where newly added switches reuse existing drivers more often than introducing new ones, I've opted for deciding who gets to opt into the workaround based on an OF compatible match table in the DSA core. The alternative would have been to add another boolean property to struct dsa_switch, like configure_vlan_while_not_filtering. But this avoids situations where sometimes driver maintainers obfuscate what goes on by sharing a common probing function, and therefore making new switches inherit old quirks. Side note, we also warn about missing properties for drivers that rely on the workaround. This isn't an indication that we'll break compatibility with those DT blobs any time soon, but is rather done to raise awareness about the change, for future DT blob authors. Cc: Rob Herring <robh+dt@kernel.org> Cc: Frank Rowand <frowand.list@gmail.com> Acked-by: Alvin Šipraga <alsi@bang-olufsen.dk> # realtek Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-08-18 11:55:00 +00:00
if (missing_link_description) {
dev_warn(ds->dev,
"Skipping phylink registration for %s port %d\n",
dsa_port_is_cpu(dp) ? "CPU" : "DSA", dp->index);
} else {
if (ds->ops->phylink_mac_link_down)
ds->ops->phylink_mac_link_down(ds, port,
MLO_AN_FIXED, PHY_INTERFACE_MODE_NA);
net: dsa: make phylink-related OF properties mandatory on DSA and CPU ports Early DSA drivers were kind of simplistic in that they assumed a fairly narrow hardware layout. User ports would have integrated PHYs at an internal MDIO address that is derivable from the port number, and shared (DSA and CPU) ports would have an MII-style (serial or parallel) connection to another MAC. Phylib and then phylink were used to drive the internal PHYs, and this needed little to no description through the platform data structures. Bringing up the shared ports at the maximum supported link speed was the responsibility of the drivers. As a result of this, when these early drivers were converted from platform data to the new DSA OF bindings, there was no link information translated into the first DT bindings. https://lore.kernel.org/all/YtXFtTsf++AeDm1l@lunn.ch/ Later, phylink was adopted for shared ports as well, and today we have a workaround in place, introduced by commit a20f997010c4 ("net: dsa: Don't instantiate phylink for CPU/DSA ports unless needed"). There, DSA checks for the presence of phy-handle/fixed-link/managed OF properties, and if missing, phylink registration would be skipped. This is because phylink is optional for some drivers (the shared ports already work without it), but the process of starting to register a port with phylink is irreversible: if phylink_create() fails to find the fwnode properties it needs, it bails out and it leaves the ports inoperational (because phylink expects ports to be initially down, so DSA necessarily takes them down, and doesn't know how to put them back up again). DSA being a common framework, new drivers opt into this workaround willy-nilly, but the ideal behavior from the DSA core's side would have been to not interfere with phylink's process of failing at all. This isn't possible because of regression concerns with pre-phylink DT blobs, but at least DSA should put a stop to the proliferation of more of such cases that rely on the workaround to skip phylink registration, and sanitize the environment that new drivers work in. To that end, create a list of compatible strings for which the workaround is preserved, and don't apply the workaround for any drivers outside that list (this includes new drivers). In some cases, we make the assumption that even existing drivers don't rely on DSA's workaround, and we do this by looking at the device trees in which they appear. We can't fully know what is the situation with downstream DT blobs, but we can guess the overall trend by studying the DT blobs that were submitted upstream. If there are upstream blobs that have lacking descriptions, we take it as very likely that there are many more downstream blobs that do so too. If all upstream blobs have complete descriptions, we take that as a hint that the driver is a candidate for enforcing strict DT bindings (considering that most bindings are copy-pasted). If there are no upstream DT blobs, we take the conservative route of allowing the workaround, unless the driver maintainer instructs us otherwise. The driver situation is as follows: ar9331 ~~~~~~ compatible strings: - qca,ar9331-switch 1 occurrence in mainline device trees, part of SoC dtsi (arch/mips/boot/dts/qca/ar9331.dtsi), description is not problematic. Verdict: opt into strict DT bindings and out of workarounds. b53 ~~~ compatible strings: - brcm,bcm5325 - brcm,bcm53115 - brcm,bcm53125 - brcm,bcm53128 - brcm,bcm5365 - brcm,bcm5389 - brcm,bcm5395 - brcm,bcm5397 - brcm,bcm5398 - brcm,bcm53010-srab - brcm,bcm53011-srab - brcm,bcm53012-srab - brcm,bcm53018-srab - brcm,bcm53019-srab - brcm,bcm5301x-srab - brcm,bcm11360-srab - brcm,bcm58522-srab - brcm,bcm58525-srab - brcm,bcm58535-srab - brcm,bcm58622-srab - brcm,bcm58623-srab - brcm,bcm58625-srab - brcm,bcm88312-srab - brcm,cygnus-srab - brcm,nsp-srab - brcm,omega-srab - brcm,bcm3384-switch - brcm,bcm6328-switch - brcm,bcm6368-switch - brcm,bcm63xx-switch I've found at least these mainline DT blobs with problems: arch/arm/boot/dts/bcm47094-linksys-panamera.dts - lacks phy-mode arch/arm/boot/dts/bcm47189-tenda-ac9.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47081-luxul-xap-1410.dts arch/arm/boot/dts/bcm47081-luxul-xwr-1200.dts arch/arm/boot/dts/bcm47081-buffalo-wzr-600dhp2.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm47094-luxul-xbr-4500.dts arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts arch/arm/boot/dts/bcm4708-luxul-xap-1510.dts arch/arm/boot/dts/bcm953012er.dts arch/arm/boot/dts/bcm4708-netgear-r6250.dts arch/arm/boot/dts/bcm4708-buffalo-wzr-1166dhp-common.dtsi arch/arm/boot/dts/bcm4708-luxul-xwc-1000.dts arch/arm/boot/dts/bcm47094-luxul-abr-4500.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/bcm53016-meraki-mr32.dts - lacks phy-mode Verdict: opt into DSA workarounds. bcm_sf2 ~~~~~~~ compatible strings: - brcm,bcm4908-switch - brcm,bcm7445-switch-v4.0 - brcm,bcm7278-switch-v4.0 - brcm,bcm7278-switch-v4.8 A single occurrence in mainline (arch/arm64/boot/dts/broadcom/bcm4908/bcm4908.dtsi), part of a SoC dtsi, valid description. Florian Fainelli explains that most of the bcm_sf2 device trees lack a full description for the internal IMP ports. Verdict: opt the BCM4908 into strict DT bindings, and opt the rest into the workarounds. Note that even though BCM4908 has strict DT bindings, it still does not register with phylink on the IMP port due to it implementing ->adjust_link(). hellcreek ~~~~~~~~~ compatible strings: - hirschmann,hellcreek-de1soc-r1 No occurrence in mainline device trees. Kurt Kanzenbach explains that the downstream device trees lacked phy-mode and fixed link, and needed work, but were fixed in the meantime. Verdict: opt into strict DT bindings and out of workarounds. lan9303 ~~~~~~~ compatible strings: - smsc,lan9303-mdio - smsc,lan9303-i2c 1 occurrence in mainline device trees: arch/arm/boot/dts/imx53-kp-hsc.dts - no phy-mode, no fixed-link Verdict: opt out of strict DT bindings and into workarounds. lantiq_gswip ~~~~~~~~~~~~ compatible strings: - lantiq,xrx200-gswip - lantiq,xrx300-gswip - lantiq,xrx330-gswip No occurrences in mainline device trees. Martin Blumenstingl confirms that the downstream OpenWrt device trees lack a proper fixed-link and need work, and that the incomplete description can even be seen in the example from Documentation/devicetree/bindings/net/dsa/lantiq-gswip.txt. Verdict: opt out of strict DT bindings and into workarounds. microchip ksz ~~~~~~~~~~~~~ compatible strings: - microchip,ksz8765 - microchip,ksz8794 - microchip,ksz8795 - microchip,ksz8863 - microchip,ksz8873 - microchip,ksz9477 - microchip,ksz9897 - microchip,ksz9893 - microchip,ksz9563 - microchip,ksz8563 - microchip,ksz9567 - microchip,lan9370 - microchip,lan9371 - microchip,lan9372 - microchip,lan9373 - microchip,lan9374 5 occurrences in mainline device trees, all descriptions are valid. But we had a snafu for the ksz8795 and ksz9477 drivers where the phy-mode property would be expected to be located directly under the 'switch' node rather than under a port OF node. It was fixed by commit edecfa98f602 ("net: dsa: microchip: look for phy-mode in port nodes"). The driver still has compatibility with the old DT blobs. The lan937x support was added later than the above snafu was fixed, and even though it has support for the broken DT blobs by virtue of sharing a common probing function, I'll take it that its DT blobs are correct. Verdict: opt lan937x into strict DT bindings, and the others out. mt7530 ~~~~~~ compatible strings - mediatek,mt7621 - mediatek,mt7530 - mediatek,mt7531 Multiple occurrences in mainline device trees, one is part of an SoC dtsi (arch/mips/boot/dts/ralink/mt7621.dtsi), all descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. mv88e6060 ~~~~~~~~~ compatible string: - marvell,mv88e6060 no occurrences in mainline, nobody knows anybody who uses it. Verdict: opt out of strict DT bindings and into workarounds. mv88e6xxx ~~~~~~~~~ compatible strings: - marvell,mv88e6085 - marvell,mv88e6190 - marvell,mv88e6250 Device trees that have incomplete descriptions of CPU or DSA ports: arch/arm64/boot/dts/freescale/imx8mq-zii-ultra.dtsi - lacks phy-mode arch/arm64/boot/dts/marvell/cn9130-crb.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-spu3.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-mv88f6281gtw-ge.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-spb4.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-cfu1.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-dev-rev-c.dts - lacks phy-mode on CPU port, fixed-link on DSA ports arch/arm/boot/dts/vf610-zii-dev-rev-b.dts - lacks phy-mode on CPU port arch/arm/boot/dts/armada-381-netgear-gs110emx.dts - lacks phy-mode arch/arm/boot/dts/vf610-zii-scu4-aib.dts - lacks fixed-link on xgmii DSA ports and/or in-band-status on 2500base-x DSA ports, and phy-mode on CPU port arch/arm/boot/dts/imx6qdl-gw5904.dtsi - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-385-clearfog-gtr-l8.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/vf610-zii-ssmb-dtu.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-dir665.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-rd88f6281.dtsi - lacks phy-mode arch/arm/boot/dts/orion5x-netgear-wnr854t.dts - lacks phy-mode and fixed-link arch/arm/boot/dts/armada-388-clearfog.dts - lacks phy-mode arch/arm/boot/dts/armada-xp-linksys-mamba.dts - lacks phy-mode arch/arm/boot/dts/armada-385-linksys.dtsi - lacks phy-mode arch/arm/boot/dts/imx6q-b450v3.dts arch/arm/boot/dts/imx6q-b850v3.dts - has a phy-handle but not a phy-mode? arch/arm/boot/dts/armada-370-rd.dts - lacks phy-mode arch/arm/boot/dts/kirkwood-linksys-viper.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-rdu1.dts - lacks phy-mode arch/arm/boot/dts/imx51-zii-scu2-mezz.dts - lacks phy-mode arch/arm/boot/dts/imx6qdl-zii-rdu2.dtsi - lacks phy-mode arch/arm/boot/dts/armada-385-clearfog-gtr-s4.dts - lacks phy-mode and fixed-link Verdict: opt out of strict DT bindings and into workarounds. ocelot ~~~~~~ compatible strings: - mscc,vsc9953-switch - felix (arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi) is a PCI device, has no compatible string 2 occurrences in mainline, both are part of SoC dtsi and complete. Verdict: opt into strict DT bindings and out of workarounds. qca8k ~~~~~ compatible strings: - qca,qca8327 - qca,qca8328 - qca,qca8334 - qca,qca8337 5 occurrences in mainline device trees, none of the descriptions are problematic. Verdict: opt into strict DT bindings and out of workarounds. realtek ~~~~~~~ compatible strings: - realtek,rtl8366rb - realtek,rtl8365mb 2 occurrences in mainline, both descriptions are fine, additionally rtl8365mb.c has a comment "The device tree firmware should also specify the link partner of the extension port - either via a fixed-link or other phy-handle." Verdict: opt into strict DT bindings and out of workarounds. rzn1_a5psw ~~~~~~~~~~ compatible strings: - renesas,rzn1-a5psw One single occurrence, part of SoC dtsi (arch/arm/boot/dts/r9a06g032.dtsi), description is fine. Verdict: opt into strict DT bindings and out of workarounds. sja1105 ~~~~~~~ Driver already validates its port OF nodes in sja1105_parse_ports_node(). Verdict: opt into strict DT bindings and out of workarounds. vsc73xx ~~~~~~~ compatible strings: - vitesse,vsc7385 - vitesse,vsc7388 - vitesse,vsc7395 - vitesse,vsc7398 2 occurrences in mainline device trees, both descriptions are fine. Verdict: opt into strict DT bindings and out of workarounds. xrs700x ~~~~~~~ compatible strings: - arrow,xrs7003e - arrow,xrs7003f - arrow,xrs7004e - arrow,xrs7004f no occurrences in mainline, we don't know. Verdict: opt out of strict DT bindings and into workarounds. Because there is a pattern where newly added switches reuse existing drivers more often than introducing new ones, I've opted for deciding who gets to opt into the workaround based on an OF compatible match table in the DSA core. The alternative would have been to add another boolean property to struct dsa_switch, like configure_vlan_while_not_filtering. But this avoids situations where sometimes driver maintainers obfuscate what goes on by sharing a common probing function, and therefore making new switches inherit old quirks. Side note, we also warn about missing properties for drivers that rely on the workaround. This isn't an indication that we'll break compatibility with those DT blobs any time soon, but is rather done to raise awareness about the change, for future DT blob authors. Cc: Rob Herring <robh+dt@kernel.org> Cc: Frank Rowand <frowand.list@gmail.com> Acked-by: Alvin Šipraga <alsi@bang-olufsen.dk> # realtek Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-08-18 11:55:00 +00:00
return dsa_shared_port_phylink_register(dp);
}
return 0;
}
dev_warn(ds->dev,
"Using legacy PHYLIB callbacks. Please migrate to PHYLINK!\n");
if (of_phy_is_fixed_link(dp->dn))
return dsa_shared_port_fixed_link_register_of(dp);
else
return dsa_shared_port_setup_phy_of(dp, true);
}
void dsa_shared_port_link_unregister_of(struct dsa_port *dp)
{
struct dsa_switch *ds = dp->ds;
if (!ds->ops->adjust_link && dp->pl) {
rtnl_lock();
phylink_disconnect_phy(dp->pl);
rtnl_unlock();
net: dsa: don't leave dangling pointers in dp->pl when failing There is a desire to simplify the dsa_port registration path with devlink, and this involves reworking a bit how user ports which fail to connect to their PHY (because it's missing) get reinitialized as UNUSED devlink ports. The desire is for the change to look something like this; basically dsa_port_setup() has failed, we just change dp->type and call dsa_port_setup() again. -/* Destroy the current devlink port, and create a new one which has the UNUSED - * flavour. - */ -static int dsa_port_reinit_as_unused(struct dsa_port *dp) +static int dsa_port_setup_as_unused(struct dsa_port *dp) { - dsa_port_devlink_teardown(dp); dp->type = DSA_PORT_TYPE_UNUSED; - return dsa_port_devlink_setup(dp); + return dsa_port_setup(dp); } For an UNUSED port, dsa_port_setup() mostly only calls dsa_port_devlink_setup() anyway, so we could get away with calling just that. But if we call the full blown dsa_port_setup(dp) (which will be needed to properly set dp->setup = true), the callee will have the tendency to go through this code block too, and call dsa_port_disable(dp): switch (dp->type) { case DSA_PORT_TYPE_UNUSED: dsa_port_disable(dp); break; That is not very good, because dsa_port_disable() has this hidden inside of it: if (dp->pl) phylink_stop(dp->pl); Fact is, we are not prepared to handle a call to dsa_port_disable() with a struct dsa_port that came from a previous (and failed) call to dsa_port_setup(). We do not clean up dp->pl, and this will make the second call to dsa_port_setup() call phylink_stop() on a dangling dp->pl pointer. Solve this by creating an API for phylink destruction which is symmetric to the phylink creation, and never leave dp->pl set to anything except NULL or a valid phylink structure. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-09-29 07:29:00 +00:00
dsa_port_phylink_destroy(dp);
return;
}
if (of_phy_is_fixed_link(dp->dn))
of_phy_deregister_fixed_link(dp->dn);
else
dsa_shared_port_setup_phy_of(dp, false);
}
int dsa_port_hsr_join(struct dsa_port *dp, struct net_device *hsr,
struct netlink_ext_ack *extack)
{
struct dsa_switch *ds = dp->ds;
int err;
if (!ds->ops->port_hsr_join)
return -EOPNOTSUPP;
dp->hsr_dev = hsr;
err = ds->ops->port_hsr_join(ds, dp->index, hsr, extack);
if (err)
dp->hsr_dev = NULL;
return err;
}
void dsa_port_hsr_leave(struct dsa_port *dp, struct net_device *hsr)
{
struct dsa_switch *ds = dp->ds;
int err;
dp->hsr_dev = NULL;
if (ds->ops->port_hsr_leave) {
err = ds->ops->port_hsr_leave(ds, dp->index, hsr);
if (err)
dev_err(dp->ds->dev,
"port %d failed to leave HSR %s: %pe\n",
dp->index, hsr->name, ERR_PTR(err));
}
}
net: dsa: tag_8021q: add proper cross-chip notifier support The big problem which mandates cross-chip notifiers for tag_8021q is this: | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] When the user runs: ip link add br0 type bridge ip link set sw0p0 master br0 ip link set sw2p0 master br0 It doesn't work. This is because dsa_8021q_crosschip_bridge_join() assumes that "ds" and "other_ds" are at most 1 hop away from each other, so it is sufficient to add the RX VLAN of {ds, port} into {other_ds, other_port} and vice versa and presto, the cross-chip link works. When there is another switch in the middle, such as in this case switch 1 with its DSA links sw1p3 and sw1p4, somebody needs to tell it about these VLANs too. Which is exactly why the problem is quadratic: when a port joins a bridge, for each port in the tree that's already in that same bridge we notify a tag_8021q VLAN addition of that port's RX VLAN to the entire tree. It is a very complicated web of VLANs. It must be mentioned that currently we install tag_8021q VLANs on too many ports (DSA links - to be precise, on all of them). For example, when sw2p0 joins br0, and assuming sw1p0 was part of br0 too, we add the RX VLAN of sw2p0 on the DSA links of switch 0 too, even though there isn't any port of switch 0 that is a member of br0 (at least yet). In theory we could notify only the switches which sit in between the port joining the bridge and the port reacting to that bridge_join event. But in practice that is impossible, because of the way 'link' properties are described in the device tree. The DSA bindings require DT writers to list out not only the real/physical DSA links, but in fact the entire routing table, like for example switch 0 above will have: sw0p3: port@3 { link = <&sw1p4 &sw2p4>; }; This was done because: /* TODO: ideally DSA ports would have a single dp->link_dp member, * and no dst->rtable nor this struct dsa_link would be needed, * but this would require some more complex tree walking, * so keep it stupid at the moment and list them all. */ but it is a perfect example of a situation where too much information is actively detrimential, because we are now in the position where we cannot distinguish a real DSA link from one that is put there to avoid the 'complex tree walking'. And because DT is ABI, there is not much we can change. And because we do not know which DSA links are real and which ones aren't, we can't really know if DSA switch A is in the data path between switches B and C, in the general case. So this is why tag_8021q RX VLANs are added on all DSA links, and probably why it will never change. On the other hand, at least the number of additions/deletions is well balanced, and this means that once we implement reference counting at the cross-chip notifier level a la fdb/mdb, there is absolutely zero need for a struct dsa_8021q_crosschip_link, it's all self-managing. In fact, with the tag_8021q notifiers emitted from the bridge join notifiers, it becomes so generic that sja1105 does not need to do anything anymore, we can just delete its implementation of the .crosschip_bridge_{join,leave} methods. Among other things we can simply delete is the home-grown implementation of sja1105_notify_crosschip_switches(). The reason why that is wrong is because it is not quadratic - it only covers remote switches to which we have a cross-chip bridging link and that does not cover in-between switches. This deletion is part of the same patch because sja1105 used to poke deep inside the guts of the tag_8021q context in order to do that. Because the cross-chip links went away, so needs the sja1105 code. Last but not least, dsa_8021q_setup_port() is simplified (and also renamed). Because our TAG_8021Q_VLAN_ADD notifier is designed to react on the CPU port too, the four dsa_8021q_vid_apply() calls: - 1 for RX VLAN on user port - 1 for the user port's RX VLAN on the CPU port - 1 for TX VLAN on user port - 1 for the user port's TX VLAN on the CPU port now get squashed into only 2 notifier calls via dsa_port_tag_8021q_vlan_add. And because the notifiers to add and to delete a tag_8021q VLAN are distinct, now we finally break up the port setup and teardown into separate functions instead of relying on a "bool enabled" flag which tells us what to do. Arguably it should have been this way from the get go. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-19 17:14:52 +00:00
net: dsa: tag_8021q: don't broadcast during setup/teardown Currently, on my board with multiple sja1105 switches in disjoint trees described in commit f66a6a69f97a ("net: dsa: permit cross-chip bridging between all trees in the system"), rebooting the board triggers the following benign warnings: [ 12.345566] sja1105 spi2.0: port 0 failed to notify tag_8021q VLAN 1088 deletion: -ENOENT [ 12.353804] sja1105 spi2.0: port 0 failed to notify tag_8021q VLAN 2112 deletion: -ENOENT [ 12.362019] sja1105 spi2.0: port 1 failed to notify tag_8021q VLAN 1089 deletion: -ENOENT [ 12.370246] sja1105 spi2.0: port 1 failed to notify tag_8021q VLAN 2113 deletion: -ENOENT [ 12.378466] sja1105 spi2.0: port 2 failed to notify tag_8021q VLAN 1090 deletion: -ENOENT [ 12.386683] sja1105 spi2.0: port 2 failed to notify tag_8021q VLAN 2114 deletion: -ENOENT Basically switch 1 calls dsa_tag_8021q_unregister, and switch 1's TX and RX VLANs cannot be found on switch 2's CPU port. But why would switch 2 even attempt to delete switch 1's TX and RX tag_8021q VLANs from its CPU port? Well, because we use dsa_broadcast, and it is supposed that it had added those VLANs in the first place (because in dsa_port_tag_8021q_vlan_match, all CPU ports match regardless of their tree index or switch index). The two trees probe asynchronously, and when switch 1 probed, it called dsa_broadcast which did not notify the tree of switch 2, because that didn't probe yet. But during unbind, switch 2's tree _is_ probed, so it _is_ notified of the deletion. Before jumping to introduce a synchronization mechanism between the probing across disjoint switch trees, let's take a step back and see whether we _need_ to do that in the first place. The RX and TX VLANs of switch 1 would be needed on switch 2's CPU port only if switch 1 and 2 were part of a cross-chip bridge. And dsa_tag_8021q_bridge_join takes care precisely of that (but if probing was synchronous, the bridge_join would just end up bumping the VLANs' refcount, because they are already installed by the setup path). Since by the time the ports are bridged, all DSA trees are already set up, and we don't need the tag_8021q VLANs of one switch installed on the other switches during probe time, the answer is that we don't need to fix the synchronization issue. So make the setup and teardown code paths call dsa_port_notify, which notifies only the local tree, and the bridge code paths call dsa_broadcast, which let the other trees know as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-11 13:46:06 +00:00
int dsa_port_tag_8021q_vlan_add(struct dsa_port *dp, u16 vid, bool broadcast)
net: dsa: tag_8021q: add proper cross-chip notifier support The big problem which mandates cross-chip notifiers for tag_8021q is this: | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] When the user runs: ip link add br0 type bridge ip link set sw0p0 master br0 ip link set sw2p0 master br0 It doesn't work. This is because dsa_8021q_crosschip_bridge_join() assumes that "ds" and "other_ds" are at most 1 hop away from each other, so it is sufficient to add the RX VLAN of {ds, port} into {other_ds, other_port} and vice versa and presto, the cross-chip link works. When there is another switch in the middle, such as in this case switch 1 with its DSA links sw1p3 and sw1p4, somebody needs to tell it about these VLANs too. Which is exactly why the problem is quadratic: when a port joins a bridge, for each port in the tree that's already in that same bridge we notify a tag_8021q VLAN addition of that port's RX VLAN to the entire tree. It is a very complicated web of VLANs. It must be mentioned that currently we install tag_8021q VLANs on too many ports (DSA links - to be precise, on all of them). For example, when sw2p0 joins br0, and assuming sw1p0 was part of br0 too, we add the RX VLAN of sw2p0 on the DSA links of switch 0 too, even though there isn't any port of switch 0 that is a member of br0 (at least yet). In theory we could notify only the switches which sit in between the port joining the bridge and the port reacting to that bridge_join event. But in practice that is impossible, because of the way 'link' properties are described in the device tree. The DSA bindings require DT writers to list out not only the real/physical DSA links, but in fact the entire routing table, like for example switch 0 above will have: sw0p3: port@3 { link = <&sw1p4 &sw2p4>; }; This was done because: /* TODO: ideally DSA ports would have a single dp->link_dp member, * and no dst->rtable nor this struct dsa_link would be needed, * but this would require some more complex tree walking, * so keep it stupid at the moment and list them all. */ but it is a perfect example of a situation where too much information is actively detrimential, because we are now in the position where we cannot distinguish a real DSA link from one that is put there to avoid the 'complex tree walking'. And because DT is ABI, there is not much we can change. And because we do not know which DSA links are real and which ones aren't, we can't really know if DSA switch A is in the data path between switches B and C, in the general case. So this is why tag_8021q RX VLANs are added on all DSA links, and probably why it will never change. On the other hand, at least the number of additions/deletions is well balanced, and this means that once we implement reference counting at the cross-chip notifier level a la fdb/mdb, there is absolutely zero need for a struct dsa_8021q_crosschip_link, it's all self-managing. In fact, with the tag_8021q notifiers emitted from the bridge join notifiers, it becomes so generic that sja1105 does not need to do anything anymore, we can just delete its implementation of the .crosschip_bridge_{join,leave} methods. Among other things we can simply delete is the home-grown implementation of sja1105_notify_crosschip_switches(). The reason why that is wrong is because it is not quadratic - it only covers remote switches to which we have a cross-chip bridging link and that does not cover in-between switches. This deletion is part of the same patch because sja1105 used to poke deep inside the guts of the tag_8021q context in order to do that. Because the cross-chip links went away, so needs the sja1105 code. Last but not least, dsa_8021q_setup_port() is simplified (and also renamed). Because our TAG_8021Q_VLAN_ADD notifier is designed to react on the CPU port too, the four dsa_8021q_vid_apply() calls: - 1 for RX VLAN on user port - 1 for the user port's RX VLAN on the CPU port - 1 for TX VLAN on user port - 1 for the user port's TX VLAN on the CPU port now get squashed into only 2 notifier calls via dsa_port_tag_8021q_vlan_add. And because the notifiers to add and to delete a tag_8021q VLAN are distinct, now we finally break up the port setup and teardown into separate functions instead of relying on a "bool enabled" flag which tells us what to do. Arguably it should have been this way from the get go. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-19 17:14:52 +00:00
{
struct dsa_notifier_tag_8021q_vlan_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
net: dsa: tag_8021q: add proper cross-chip notifier support The big problem which mandates cross-chip notifiers for tag_8021q is this: | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] When the user runs: ip link add br0 type bridge ip link set sw0p0 master br0 ip link set sw2p0 master br0 It doesn't work. This is because dsa_8021q_crosschip_bridge_join() assumes that "ds" and "other_ds" are at most 1 hop away from each other, so it is sufficient to add the RX VLAN of {ds, port} into {other_ds, other_port} and vice versa and presto, the cross-chip link works. When there is another switch in the middle, such as in this case switch 1 with its DSA links sw1p3 and sw1p4, somebody needs to tell it about these VLANs too. Which is exactly why the problem is quadratic: when a port joins a bridge, for each port in the tree that's already in that same bridge we notify a tag_8021q VLAN addition of that port's RX VLAN to the entire tree. It is a very complicated web of VLANs. It must be mentioned that currently we install tag_8021q VLANs on too many ports (DSA links - to be precise, on all of them). For example, when sw2p0 joins br0, and assuming sw1p0 was part of br0 too, we add the RX VLAN of sw2p0 on the DSA links of switch 0 too, even though there isn't any port of switch 0 that is a member of br0 (at least yet). In theory we could notify only the switches which sit in between the port joining the bridge and the port reacting to that bridge_join event. But in practice that is impossible, because of the way 'link' properties are described in the device tree. The DSA bindings require DT writers to list out not only the real/physical DSA links, but in fact the entire routing table, like for example switch 0 above will have: sw0p3: port@3 { link = <&sw1p4 &sw2p4>; }; This was done because: /* TODO: ideally DSA ports would have a single dp->link_dp member, * and no dst->rtable nor this struct dsa_link would be needed, * but this would require some more complex tree walking, * so keep it stupid at the moment and list them all. */ but it is a perfect example of a situation where too much information is actively detrimential, because we are now in the position where we cannot distinguish a real DSA link from one that is put there to avoid the 'complex tree walking'. And because DT is ABI, there is not much we can change. And because we do not know which DSA links are real and which ones aren't, we can't really know if DSA switch A is in the data path between switches B and C, in the general case. So this is why tag_8021q RX VLANs are added on all DSA links, and probably why it will never change. On the other hand, at least the number of additions/deletions is well balanced, and this means that once we implement reference counting at the cross-chip notifier level a la fdb/mdb, there is absolutely zero need for a struct dsa_8021q_crosschip_link, it's all self-managing. In fact, with the tag_8021q notifiers emitted from the bridge join notifiers, it becomes so generic that sja1105 does not need to do anything anymore, we can just delete its implementation of the .crosschip_bridge_{join,leave} methods. Among other things we can simply delete is the home-grown implementation of sja1105_notify_crosschip_switches(). The reason why that is wrong is because it is not quadratic - it only covers remote switches to which we have a cross-chip bridging link and that does not cover in-between switches. This deletion is part of the same patch because sja1105 used to poke deep inside the guts of the tag_8021q context in order to do that. Because the cross-chip links went away, so needs the sja1105 code. Last but not least, dsa_8021q_setup_port() is simplified (and also renamed). Because our TAG_8021Q_VLAN_ADD notifier is designed to react on the CPU port too, the four dsa_8021q_vid_apply() calls: - 1 for RX VLAN on user port - 1 for the user port's RX VLAN on the CPU port - 1 for TX VLAN on user port - 1 for the user port's TX VLAN on the CPU port now get squashed into only 2 notifier calls via dsa_port_tag_8021q_vlan_add. And because the notifiers to add and to delete a tag_8021q VLAN are distinct, now we finally break up the port setup and teardown into separate functions instead of relying on a "bool enabled" flag which tells us what to do. Arguably it should have been this way from the get go. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-19 17:14:52 +00:00
.vid = vid,
};
net: dsa: tag_8021q: don't broadcast during setup/teardown Currently, on my board with multiple sja1105 switches in disjoint trees described in commit f66a6a69f97a ("net: dsa: permit cross-chip bridging between all trees in the system"), rebooting the board triggers the following benign warnings: [ 12.345566] sja1105 spi2.0: port 0 failed to notify tag_8021q VLAN 1088 deletion: -ENOENT [ 12.353804] sja1105 spi2.0: port 0 failed to notify tag_8021q VLAN 2112 deletion: -ENOENT [ 12.362019] sja1105 spi2.0: port 1 failed to notify tag_8021q VLAN 1089 deletion: -ENOENT [ 12.370246] sja1105 spi2.0: port 1 failed to notify tag_8021q VLAN 2113 deletion: -ENOENT [ 12.378466] sja1105 spi2.0: port 2 failed to notify tag_8021q VLAN 1090 deletion: -ENOENT [ 12.386683] sja1105 spi2.0: port 2 failed to notify tag_8021q VLAN 2114 deletion: -ENOENT Basically switch 1 calls dsa_tag_8021q_unregister, and switch 1's TX and RX VLANs cannot be found on switch 2's CPU port. But why would switch 2 even attempt to delete switch 1's TX and RX tag_8021q VLANs from its CPU port? Well, because we use dsa_broadcast, and it is supposed that it had added those VLANs in the first place (because in dsa_port_tag_8021q_vlan_match, all CPU ports match regardless of their tree index or switch index). The two trees probe asynchronously, and when switch 1 probed, it called dsa_broadcast which did not notify the tree of switch 2, because that didn't probe yet. But during unbind, switch 2's tree _is_ probed, so it _is_ notified of the deletion. Before jumping to introduce a synchronization mechanism between the probing across disjoint switch trees, let's take a step back and see whether we _need_ to do that in the first place. The RX and TX VLANs of switch 1 would be needed on switch 2's CPU port only if switch 1 and 2 were part of a cross-chip bridge. And dsa_tag_8021q_bridge_join takes care precisely of that (but if probing was synchronous, the bridge_join would just end up bumping the VLANs' refcount, because they are already installed by the setup path). Since by the time the ports are bridged, all DSA trees are already set up, and we don't need the tag_8021q VLANs of one switch installed on the other switches during probe time, the answer is that we don't need to fix the synchronization issue. So make the setup and teardown code paths call dsa_port_notify, which notifies only the local tree, and the bridge code paths call dsa_broadcast, which let the other trees know as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-11 13:46:06 +00:00
if (broadcast)
return dsa_broadcast(DSA_NOTIFIER_TAG_8021Q_VLAN_ADD, &info);
return dsa_port_notify(dp, DSA_NOTIFIER_TAG_8021Q_VLAN_ADD, &info);
net: dsa: tag_8021q: add proper cross-chip notifier support The big problem which mandates cross-chip notifiers for tag_8021q is this: | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] When the user runs: ip link add br0 type bridge ip link set sw0p0 master br0 ip link set sw2p0 master br0 It doesn't work. This is because dsa_8021q_crosschip_bridge_join() assumes that "ds" and "other_ds" are at most 1 hop away from each other, so it is sufficient to add the RX VLAN of {ds, port} into {other_ds, other_port} and vice versa and presto, the cross-chip link works. When there is another switch in the middle, such as in this case switch 1 with its DSA links sw1p3 and sw1p4, somebody needs to tell it about these VLANs too. Which is exactly why the problem is quadratic: when a port joins a bridge, for each port in the tree that's already in that same bridge we notify a tag_8021q VLAN addition of that port's RX VLAN to the entire tree. It is a very complicated web of VLANs. It must be mentioned that currently we install tag_8021q VLANs on too many ports (DSA links - to be precise, on all of them). For example, when sw2p0 joins br0, and assuming sw1p0 was part of br0 too, we add the RX VLAN of sw2p0 on the DSA links of switch 0 too, even though there isn't any port of switch 0 that is a member of br0 (at least yet). In theory we could notify only the switches which sit in between the port joining the bridge and the port reacting to that bridge_join event. But in practice that is impossible, because of the way 'link' properties are described in the device tree. The DSA bindings require DT writers to list out not only the real/physical DSA links, but in fact the entire routing table, like for example switch 0 above will have: sw0p3: port@3 { link = <&sw1p4 &sw2p4>; }; This was done because: /* TODO: ideally DSA ports would have a single dp->link_dp member, * and no dst->rtable nor this struct dsa_link would be needed, * but this would require some more complex tree walking, * so keep it stupid at the moment and list them all. */ but it is a perfect example of a situation where too much information is actively detrimential, because we are now in the position where we cannot distinguish a real DSA link from one that is put there to avoid the 'complex tree walking'. And because DT is ABI, there is not much we can change. And because we do not know which DSA links are real and which ones aren't, we can't really know if DSA switch A is in the data path between switches B and C, in the general case. So this is why tag_8021q RX VLANs are added on all DSA links, and probably why it will never change. On the other hand, at least the number of additions/deletions is well balanced, and this means that once we implement reference counting at the cross-chip notifier level a la fdb/mdb, there is absolutely zero need for a struct dsa_8021q_crosschip_link, it's all self-managing. In fact, with the tag_8021q notifiers emitted from the bridge join notifiers, it becomes so generic that sja1105 does not need to do anything anymore, we can just delete its implementation of the .crosschip_bridge_{join,leave} methods. Among other things we can simply delete is the home-grown implementation of sja1105_notify_crosschip_switches(). The reason why that is wrong is because it is not quadratic - it only covers remote switches to which we have a cross-chip bridging link and that does not cover in-between switches. This deletion is part of the same patch because sja1105 used to poke deep inside the guts of the tag_8021q context in order to do that. Because the cross-chip links went away, so needs the sja1105 code. Last but not least, dsa_8021q_setup_port() is simplified (and also renamed). Because our TAG_8021Q_VLAN_ADD notifier is designed to react on the CPU port too, the four dsa_8021q_vid_apply() calls: - 1 for RX VLAN on user port - 1 for the user port's RX VLAN on the CPU port - 1 for TX VLAN on user port - 1 for the user port's TX VLAN on the CPU port now get squashed into only 2 notifier calls via dsa_port_tag_8021q_vlan_add. And because the notifiers to add and to delete a tag_8021q VLAN are distinct, now we finally break up the port setup and teardown into separate functions instead of relying on a "bool enabled" flag which tells us what to do. Arguably it should have been this way from the get go. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-19 17:14:52 +00:00
}
net: dsa: tag_8021q: don't broadcast during setup/teardown Currently, on my board with multiple sja1105 switches in disjoint trees described in commit f66a6a69f97a ("net: dsa: permit cross-chip bridging between all trees in the system"), rebooting the board triggers the following benign warnings: [ 12.345566] sja1105 spi2.0: port 0 failed to notify tag_8021q VLAN 1088 deletion: -ENOENT [ 12.353804] sja1105 spi2.0: port 0 failed to notify tag_8021q VLAN 2112 deletion: -ENOENT [ 12.362019] sja1105 spi2.0: port 1 failed to notify tag_8021q VLAN 1089 deletion: -ENOENT [ 12.370246] sja1105 spi2.0: port 1 failed to notify tag_8021q VLAN 2113 deletion: -ENOENT [ 12.378466] sja1105 spi2.0: port 2 failed to notify tag_8021q VLAN 1090 deletion: -ENOENT [ 12.386683] sja1105 spi2.0: port 2 failed to notify tag_8021q VLAN 2114 deletion: -ENOENT Basically switch 1 calls dsa_tag_8021q_unregister, and switch 1's TX and RX VLANs cannot be found on switch 2's CPU port. But why would switch 2 even attempt to delete switch 1's TX and RX tag_8021q VLANs from its CPU port? Well, because we use dsa_broadcast, and it is supposed that it had added those VLANs in the first place (because in dsa_port_tag_8021q_vlan_match, all CPU ports match regardless of their tree index or switch index). The two trees probe asynchronously, and when switch 1 probed, it called dsa_broadcast which did not notify the tree of switch 2, because that didn't probe yet. But during unbind, switch 2's tree _is_ probed, so it _is_ notified of the deletion. Before jumping to introduce a synchronization mechanism between the probing across disjoint switch trees, let's take a step back and see whether we _need_ to do that in the first place. The RX and TX VLANs of switch 1 would be needed on switch 2's CPU port only if switch 1 and 2 were part of a cross-chip bridge. And dsa_tag_8021q_bridge_join takes care precisely of that (but if probing was synchronous, the bridge_join would just end up bumping the VLANs' refcount, because they are already installed by the setup path). Since by the time the ports are bridged, all DSA trees are already set up, and we don't need the tag_8021q VLANs of one switch installed on the other switches during probe time, the answer is that we don't need to fix the synchronization issue. So make the setup and teardown code paths call dsa_port_notify, which notifies only the local tree, and the bridge code paths call dsa_broadcast, which let the other trees know as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-11 13:46:06 +00:00
void dsa_port_tag_8021q_vlan_del(struct dsa_port *dp, u16 vid, bool broadcast)
net: dsa: tag_8021q: add proper cross-chip notifier support The big problem which mandates cross-chip notifiers for tag_8021q is this: | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] When the user runs: ip link add br0 type bridge ip link set sw0p0 master br0 ip link set sw2p0 master br0 It doesn't work. This is because dsa_8021q_crosschip_bridge_join() assumes that "ds" and "other_ds" are at most 1 hop away from each other, so it is sufficient to add the RX VLAN of {ds, port} into {other_ds, other_port} and vice versa and presto, the cross-chip link works. When there is another switch in the middle, such as in this case switch 1 with its DSA links sw1p3 and sw1p4, somebody needs to tell it about these VLANs too. Which is exactly why the problem is quadratic: when a port joins a bridge, for each port in the tree that's already in that same bridge we notify a tag_8021q VLAN addition of that port's RX VLAN to the entire tree. It is a very complicated web of VLANs. It must be mentioned that currently we install tag_8021q VLANs on too many ports (DSA links - to be precise, on all of them). For example, when sw2p0 joins br0, and assuming sw1p0 was part of br0 too, we add the RX VLAN of sw2p0 on the DSA links of switch 0 too, even though there isn't any port of switch 0 that is a member of br0 (at least yet). In theory we could notify only the switches which sit in between the port joining the bridge and the port reacting to that bridge_join event. But in practice that is impossible, because of the way 'link' properties are described in the device tree. The DSA bindings require DT writers to list out not only the real/physical DSA links, but in fact the entire routing table, like for example switch 0 above will have: sw0p3: port@3 { link = <&sw1p4 &sw2p4>; }; This was done because: /* TODO: ideally DSA ports would have a single dp->link_dp member, * and no dst->rtable nor this struct dsa_link would be needed, * but this would require some more complex tree walking, * so keep it stupid at the moment and list them all. */ but it is a perfect example of a situation where too much information is actively detrimential, because we are now in the position where we cannot distinguish a real DSA link from one that is put there to avoid the 'complex tree walking'. And because DT is ABI, there is not much we can change. And because we do not know which DSA links are real and which ones aren't, we can't really know if DSA switch A is in the data path between switches B and C, in the general case. So this is why tag_8021q RX VLANs are added on all DSA links, and probably why it will never change. On the other hand, at least the number of additions/deletions is well balanced, and this means that once we implement reference counting at the cross-chip notifier level a la fdb/mdb, there is absolutely zero need for a struct dsa_8021q_crosschip_link, it's all self-managing. In fact, with the tag_8021q notifiers emitted from the bridge join notifiers, it becomes so generic that sja1105 does not need to do anything anymore, we can just delete its implementation of the .crosschip_bridge_{join,leave} methods. Among other things we can simply delete is the home-grown implementation of sja1105_notify_crosschip_switches(). The reason why that is wrong is because it is not quadratic - it only covers remote switches to which we have a cross-chip bridging link and that does not cover in-between switches. This deletion is part of the same patch because sja1105 used to poke deep inside the guts of the tag_8021q context in order to do that. Because the cross-chip links went away, so needs the sja1105 code. Last but not least, dsa_8021q_setup_port() is simplified (and also renamed). Because our TAG_8021Q_VLAN_ADD notifier is designed to react on the CPU port too, the four dsa_8021q_vid_apply() calls: - 1 for RX VLAN on user port - 1 for the user port's RX VLAN on the CPU port - 1 for TX VLAN on user port - 1 for the user port's TX VLAN on the CPU port now get squashed into only 2 notifier calls via dsa_port_tag_8021q_vlan_add. And because the notifiers to add and to delete a tag_8021q VLAN are distinct, now we finally break up the port setup and teardown into separate functions instead of relying on a "bool enabled" flag which tells us what to do. Arguably it should have been this way from the get go. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-19 17:14:52 +00:00
{
struct dsa_notifier_tag_8021q_vlan_info info = {
net: dsa: make cross-chip notifiers more efficient for host events To determine whether a given port should react to the port targeted by the notifier, dsa_port_host_vlan_match() and dsa_port_host_address_match() look at the positioning of the switch port currently executing the notifier relative to the switch port for which the notifier was emitted. To maintain stylistic compatibility with the other match functions from switch.c, the host address and host VLAN match functions take the notifier information about targeted port, switch and tree indices as argument. However, these functions only use that information to retrieve the struct dsa_port *targeted_dp, which is an invariant for the outer loop that calls them. So it makes more sense to calculate the targeted dp only once, and pass it to them as argument. But furthermore, the targeted dp is actually known at the time the call to dsa_port_notify() is made. It is just that we decide to only save the indices of the port, switch and tree in the notifier structure, just to retrace our steps and find the dp again using dsa_switch_find() and dsa_to_port(). But both the above functions are relatively expensive, since they need to iterate through lists. It appears more straightforward to make all notifiers just pass the targeted dp inside their info structure, and have the code that needs the indices to look at info->dp->index instead of info->port, or info->dp->ds->index instead of info->sw_index, or info->dp->ds->dst->index instead of info->tree_index. For the sake of consistency, all cross-chip notifiers are converted to pass the "dp" directly. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2022-04-15 15:46:22 +00:00
.dp = dp,
net: dsa: tag_8021q: add proper cross-chip notifier support The big problem which mandates cross-chip notifiers for tag_8021q is this: | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] When the user runs: ip link add br0 type bridge ip link set sw0p0 master br0 ip link set sw2p0 master br0 It doesn't work. This is because dsa_8021q_crosschip_bridge_join() assumes that "ds" and "other_ds" are at most 1 hop away from each other, so it is sufficient to add the RX VLAN of {ds, port} into {other_ds, other_port} and vice versa and presto, the cross-chip link works. When there is another switch in the middle, such as in this case switch 1 with its DSA links sw1p3 and sw1p4, somebody needs to tell it about these VLANs too. Which is exactly why the problem is quadratic: when a port joins a bridge, for each port in the tree that's already in that same bridge we notify a tag_8021q VLAN addition of that port's RX VLAN to the entire tree. It is a very complicated web of VLANs. It must be mentioned that currently we install tag_8021q VLANs on too many ports (DSA links - to be precise, on all of them). For example, when sw2p0 joins br0, and assuming sw1p0 was part of br0 too, we add the RX VLAN of sw2p0 on the DSA links of switch 0 too, even though there isn't any port of switch 0 that is a member of br0 (at least yet). In theory we could notify only the switches which sit in between the port joining the bridge and the port reacting to that bridge_join event. But in practice that is impossible, because of the way 'link' properties are described in the device tree. The DSA bindings require DT writers to list out not only the real/physical DSA links, but in fact the entire routing table, like for example switch 0 above will have: sw0p3: port@3 { link = <&sw1p4 &sw2p4>; }; This was done because: /* TODO: ideally DSA ports would have a single dp->link_dp member, * and no dst->rtable nor this struct dsa_link would be needed, * but this would require some more complex tree walking, * so keep it stupid at the moment and list them all. */ but it is a perfect example of a situation where too much information is actively detrimential, because we are now in the position where we cannot distinguish a real DSA link from one that is put there to avoid the 'complex tree walking'. And because DT is ABI, there is not much we can change. And because we do not know which DSA links are real and which ones aren't, we can't really know if DSA switch A is in the data path between switches B and C, in the general case. So this is why tag_8021q RX VLANs are added on all DSA links, and probably why it will never change. On the other hand, at least the number of additions/deletions is well balanced, and this means that once we implement reference counting at the cross-chip notifier level a la fdb/mdb, there is absolutely zero need for a struct dsa_8021q_crosschip_link, it's all self-managing. In fact, with the tag_8021q notifiers emitted from the bridge join notifiers, it becomes so generic that sja1105 does not need to do anything anymore, we can just delete its implementation of the .crosschip_bridge_{join,leave} methods. Among other things we can simply delete is the home-grown implementation of sja1105_notify_crosschip_switches(). The reason why that is wrong is because it is not quadratic - it only covers remote switches to which we have a cross-chip bridging link and that does not cover in-between switches. This deletion is part of the same patch because sja1105 used to poke deep inside the guts of the tag_8021q context in order to do that. Because the cross-chip links went away, so needs the sja1105 code. Last but not least, dsa_8021q_setup_port() is simplified (and also renamed). Because our TAG_8021Q_VLAN_ADD notifier is designed to react on the CPU port too, the four dsa_8021q_vid_apply() calls: - 1 for RX VLAN on user port - 1 for the user port's RX VLAN on the CPU port - 1 for TX VLAN on user port - 1 for the user port's TX VLAN on the CPU port now get squashed into only 2 notifier calls via dsa_port_tag_8021q_vlan_add. And because the notifiers to add and to delete a tag_8021q VLAN are distinct, now we finally break up the port setup and teardown into separate functions instead of relying on a "bool enabled" flag which tells us what to do. Arguably it should have been this way from the get go. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-19 17:14:52 +00:00
.vid = vid,
};
int err;
net: dsa: tag_8021q: don't broadcast during setup/teardown Currently, on my board with multiple sja1105 switches in disjoint trees described in commit f66a6a69f97a ("net: dsa: permit cross-chip bridging between all trees in the system"), rebooting the board triggers the following benign warnings: [ 12.345566] sja1105 spi2.0: port 0 failed to notify tag_8021q VLAN 1088 deletion: -ENOENT [ 12.353804] sja1105 spi2.0: port 0 failed to notify tag_8021q VLAN 2112 deletion: -ENOENT [ 12.362019] sja1105 spi2.0: port 1 failed to notify tag_8021q VLAN 1089 deletion: -ENOENT [ 12.370246] sja1105 spi2.0: port 1 failed to notify tag_8021q VLAN 2113 deletion: -ENOENT [ 12.378466] sja1105 spi2.0: port 2 failed to notify tag_8021q VLAN 1090 deletion: -ENOENT [ 12.386683] sja1105 spi2.0: port 2 failed to notify tag_8021q VLAN 2114 deletion: -ENOENT Basically switch 1 calls dsa_tag_8021q_unregister, and switch 1's TX and RX VLANs cannot be found on switch 2's CPU port. But why would switch 2 even attempt to delete switch 1's TX and RX tag_8021q VLANs from its CPU port? Well, because we use dsa_broadcast, and it is supposed that it had added those VLANs in the first place (because in dsa_port_tag_8021q_vlan_match, all CPU ports match regardless of their tree index or switch index). The two trees probe asynchronously, and when switch 1 probed, it called dsa_broadcast which did not notify the tree of switch 2, because that didn't probe yet. But during unbind, switch 2's tree _is_ probed, so it _is_ notified of the deletion. Before jumping to introduce a synchronization mechanism between the probing across disjoint switch trees, let's take a step back and see whether we _need_ to do that in the first place. The RX and TX VLANs of switch 1 would be needed on switch 2's CPU port only if switch 1 and 2 were part of a cross-chip bridge. And dsa_tag_8021q_bridge_join takes care precisely of that (but if probing was synchronous, the bridge_join would just end up bumping the VLANs' refcount, because they are already installed by the setup path). Since by the time the ports are bridged, all DSA trees are already set up, and we don't need the tag_8021q VLANs of one switch installed on the other switches during probe time, the answer is that we don't need to fix the synchronization issue. So make the setup and teardown code paths call dsa_port_notify, which notifies only the local tree, and the bridge code paths call dsa_broadcast, which let the other trees know as well. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-08-11 13:46:06 +00:00
if (broadcast)
err = dsa_broadcast(DSA_NOTIFIER_TAG_8021Q_VLAN_DEL, &info);
else
err = dsa_port_notify(dp, DSA_NOTIFIER_TAG_8021Q_VLAN_DEL, &info);
net: dsa: tag_8021q: add proper cross-chip notifier support The big problem which mandates cross-chip notifiers for tag_8021q is this: | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] When the user runs: ip link add br0 type bridge ip link set sw0p0 master br0 ip link set sw2p0 master br0 It doesn't work. This is because dsa_8021q_crosschip_bridge_join() assumes that "ds" and "other_ds" are at most 1 hop away from each other, so it is sufficient to add the RX VLAN of {ds, port} into {other_ds, other_port} and vice versa and presto, the cross-chip link works. When there is another switch in the middle, such as in this case switch 1 with its DSA links sw1p3 and sw1p4, somebody needs to tell it about these VLANs too. Which is exactly why the problem is quadratic: when a port joins a bridge, for each port in the tree that's already in that same bridge we notify a tag_8021q VLAN addition of that port's RX VLAN to the entire tree. It is a very complicated web of VLANs. It must be mentioned that currently we install tag_8021q VLANs on too many ports (DSA links - to be precise, on all of them). For example, when sw2p0 joins br0, and assuming sw1p0 was part of br0 too, we add the RX VLAN of sw2p0 on the DSA links of switch 0 too, even though there isn't any port of switch 0 that is a member of br0 (at least yet). In theory we could notify only the switches which sit in between the port joining the bridge and the port reacting to that bridge_join event. But in practice that is impossible, because of the way 'link' properties are described in the device tree. The DSA bindings require DT writers to list out not only the real/physical DSA links, but in fact the entire routing table, like for example switch 0 above will have: sw0p3: port@3 { link = <&sw1p4 &sw2p4>; }; This was done because: /* TODO: ideally DSA ports would have a single dp->link_dp member, * and no dst->rtable nor this struct dsa_link would be needed, * but this would require some more complex tree walking, * so keep it stupid at the moment and list them all. */ but it is a perfect example of a situation where too much information is actively detrimential, because we are now in the position where we cannot distinguish a real DSA link from one that is put there to avoid the 'complex tree walking'. And because DT is ABI, there is not much we can change. And because we do not know which DSA links are real and which ones aren't, we can't really know if DSA switch A is in the data path between switches B and C, in the general case. So this is why tag_8021q RX VLANs are added on all DSA links, and probably why it will never change. On the other hand, at least the number of additions/deletions is well balanced, and this means that once we implement reference counting at the cross-chip notifier level a la fdb/mdb, there is absolutely zero need for a struct dsa_8021q_crosschip_link, it's all self-managing. In fact, with the tag_8021q notifiers emitted from the bridge join notifiers, it becomes so generic that sja1105 does not need to do anything anymore, we can just delete its implementation of the .crosschip_bridge_{join,leave} methods. Among other things we can simply delete is the home-grown implementation of sja1105_notify_crosschip_switches(). The reason why that is wrong is because it is not quadratic - it only covers remote switches to which we have a cross-chip bridging link and that does not cover in-between switches. This deletion is part of the same patch because sja1105 used to poke deep inside the guts of the tag_8021q context in order to do that. Because the cross-chip links went away, so needs the sja1105 code. Last but not least, dsa_8021q_setup_port() is simplified (and also renamed). Because our TAG_8021Q_VLAN_ADD notifier is designed to react on the CPU port too, the four dsa_8021q_vid_apply() calls: - 1 for RX VLAN on user port - 1 for the user port's RX VLAN on the CPU port - 1 for TX VLAN on user port - 1 for the user port's TX VLAN on the CPU port now get squashed into only 2 notifier calls via dsa_port_tag_8021q_vlan_add. And because the notifiers to add and to delete a tag_8021q VLAN are distinct, now we finally break up the port setup and teardown into separate functions instead of relying on a "bool enabled" flag which tells us what to do. Arguably it should have been this way from the get go. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-19 17:14:52 +00:00
if (err)
dev_err(dp->ds->dev,
"port %d failed to notify tag_8021q VLAN %d deletion: %pe\n",
dp->index, vid, ERR_PTR(err));
net: dsa: tag_8021q: add proper cross-chip notifier support The big problem which mandates cross-chip notifiers for tag_8021q is this: | sw0p0 sw0p1 sw0p2 sw0p3 sw0p4 [ user ] [ user ] [ user ] [ dsa ] [ cpu ] | +---------+ | sw1p0 sw1p1 sw1p2 sw1p3 sw1p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] | +---------+ | sw2p0 sw2p1 sw2p2 sw2p3 sw2p4 [ user ] [ user ] [ user ] [ dsa ] [ dsa ] When the user runs: ip link add br0 type bridge ip link set sw0p0 master br0 ip link set sw2p0 master br0 It doesn't work. This is because dsa_8021q_crosschip_bridge_join() assumes that "ds" and "other_ds" are at most 1 hop away from each other, so it is sufficient to add the RX VLAN of {ds, port} into {other_ds, other_port} and vice versa and presto, the cross-chip link works. When there is another switch in the middle, such as in this case switch 1 with its DSA links sw1p3 and sw1p4, somebody needs to tell it about these VLANs too. Which is exactly why the problem is quadratic: when a port joins a bridge, for each port in the tree that's already in that same bridge we notify a tag_8021q VLAN addition of that port's RX VLAN to the entire tree. It is a very complicated web of VLANs. It must be mentioned that currently we install tag_8021q VLANs on too many ports (DSA links - to be precise, on all of them). For example, when sw2p0 joins br0, and assuming sw1p0 was part of br0 too, we add the RX VLAN of sw2p0 on the DSA links of switch 0 too, even though there isn't any port of switch 0 that is a member of br0 (at least yet). In theory we could notify only the switches which sit in between the port joining the bridge and the port reacting to that bridge_join event. But in practice that is impossible, because of the way 'link' properties are described in the device tree. The DSA bindings require DT writers to list out not only the real/physical DSA links, but in fact the entire routing table, like for example switch 0 above will have: sw0p3: port@3 { link = <&sw1p4 &sw2p4>; }; This was done because: /* TODO: ideally DSA ports would have a single dp->link_dp member, * and no dst->rtable nor this struct dsa_link would be needed, * but this would require some more complex tree walking, * so keep it stupid at the moment and list them all. */ but it is a perfect example of a situation where too much information is actively detrimential, because we are now in the position where we cannot distinguish a real DSA link from one that is put there to avoid the 'complex tree walking'. And because DT is ABI, there is not much we can change. And because we do not know which DSA links are real and which ones aren't, we can't really know if DSA switch A is in the data path between switches B and C, in the general case. So this is why tag_8021q RX VLANs are added on all DSA links, and probably why it will never change. On the other hand, at least the number of additions/deletions is well balanced, and this means that once we implement reference counting at the cross-chip notifier level a la fdb/mdb, there is absolutely zero need for a struct dsa_8021q_crosschip_link, it's all self-managing. In fact, with the tag_8021q notifiers emitted from the bridge join notifiers, it becomes so generic that sja1105 does not need to do anything anymore, we can just delete its implementation of the .crosschip_bridge_{join,leave} methods. Among other things we can simply delete is the home-grown implementation of sja1105_notify_crosschip_switches(). The reason why that is wrong is because it is not quadratic - it only covers remote switches to which we have a cross-chip bridging link and that does not cover in-between switches. This deletion is part of the same patch because sja1105 used to poke deep inside the guts of the tag_8021q context in order to do that. Because the cross-chip links went away, so needs the sja1105 code. Last but not least, dsa_8021q_setup_port() is simplified (and also renamed). Because our TAG_8021Q_VLAN_ADD notifier is designed to react on the CPU port too, the four dsa_8021q_vid_apply() calls: - 1 for RX VLAN on user port - 1 for the user port's RX VLAN on the CPU port - 1 for TX VLAN on user port - 1 for the user port's TX VLAN on the CPU port now get squashed into only 2 notifier calls via dsa_port_tag_8021q_vlan_add. And because the notifiers to add and to delete a tag_8021q VLAN are distinct, now we finally break up the port setup and teardown into separate functions instead of relying on a "bool enabled" flag which tells us what to do. Arguably it should have been this way from the get go. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-07-19 17:14:52 +00:00
}