linux/net/rxrpc/call_event.c

548 lines
15 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/* Management of Tx window, Tx resend, ACKs and out-of-sequence reception
*
* Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/module.h>
#include <linux/circ_buf.h>
#include <linux/net.h>
#include <linux/skbuff.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
#include <linux/slab.h>
#include <linux/udp.h>
#include <net/sock.h>
#include <net/af_rxrpc.h>
#include "ar-internal.h"
/*
* Propose a PING ACK be sent.
*/
void rxrpc_propose_ping(struct rxrpc_call *call, u32 serial,
enum rxrpc_propose_ack_trace why)
{
unsigned long now = jiffies;
unsigned long ping_at = now + rxrpc_idle_ack_delay;
if (time_before(ping_at, call->ping_at)) {
WRITE_ONCE(call->ping_at, ping_at);
rxrpc_reduce_call_timer(call, ping_at, now,
rxrpc_timer_set_for_ping);
trace_rxrpc_propose_ack(call, why, RXRPC_ACK_PING, serial);
}
}
/*
* Propose a DELAY ACK be sent in the future.
*/
void rxrpc_propose_delay_ACK(struct rxrpc_call *call, rxrpc_serial_t serial,
enum rxrpc_propose_ack_trace why)
{
unsigned long expiry = rxrpc_soft_ack_delay;
unsigned long now = jiffies, ack_at;
call->ackr_serial = serial;
if (rxrpc_soft_ack_delay < expiry)
expiry = rxrpc_soft_ack_delay;
if (call->peer->srtt_us != 0)
ack_at = usecs_to_jiffies(call->peer->srtt_us >> 3);
else
ack_at = expiry;
ack_at += READ_ONCE(call->tx_backoff);
ack_at += now;
if (time_before(ack_at, call->delay_ack_at)) {
WRITE_ONCE(call->delay_ack_at, ack_at);
rxrpc_reduce_call_timer(call, ack_at, now,
rxrpc_timer_set_for_ack);
}
trace_rxrpc_propose_ack(call, why, RXRPC_ACK_DELAY, serial);
}
/*
* Queue an ACK for immediate transmission.
*/
void rxrpc_send_ACK(struct rxrpc_call *call, u8 ack_reason,
rxrpc_serial_t serial, enum rxrpc_propose_ack_trace why)
{
struct rxrpc_txbuf *txb;
if (test_bit(RXRPC_CALL_DISCONNECTED, &call->flags))
return;
rxrpc_inc_stat(call->rxnet, stat_tx_acks[ack_reason]);
txb = rxrpc_alloc_txbuf(call, RXRPC_PACKET_TYPE_ACK,
rcu_read_lock_held() ? GFP_ATOMIC | __GFP_NOWARN : GFP_NOFS);
if (!txb) {
kleave(" = -ENOMEM");
return;
}
txb->ack_why = why;
txb->wire.seq = 0;
txb->wire.type = RXRPC_PACKET_TYPE_ACK;
txb->wire.flags |= RXRPC_SLOW_START_OK;
txb->ack.bufferSpace = 0;
txb->ack.maxSkew = 0;
txb->ack.firstPacket = 0;
txb->ack.previousPacket = 0;
txb->ack.serial = htonl(serial);
txb->ack.reason = ack_reason;
txb->ack.nAcks = 0;
trace_rxrpc_send_ack(call, why, ack_reason, serial);
rxrpc_send_ack_packet(call, txb);
rxrpc_put_txbuf(txb, rxrpc_txbuf_put_ack_tx);
}
/*
* Handle congestion being detected by the retransmit timeout.
*/
static void rxrpc_congestion_timeout(struct rxrpc_call *call)
{
set_bit(RXRPC_CALL_RETRANS_TIMEOUT, &call->flags);
}
/*
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 10:10:12 +00:00
* Perform retransmission of NAK'd and unack'd packets.
*/
void rxrpc_resend(struct rxrpc_call *call, struct sk_buff *ack_skb)
{
struct rxrpc_ackpacket *ack = NULL;
rxrpc: Don't use a ring buffer for call Tx queue Change the way the Tx queueing works to make the following ends easier to achieve: (1) The filling of packets, the encryption of packets and the transmission of packets can be handled in parallel by separate threads, rather than rxrpc_sendmsg() allocating, filling, encrypting and transmitting each packet before moving onto the next one. (2) Get rid of the fixed-size ring which sets a hard limit on the number of packets that can be retained in the ring. This allows the number of packets to increase without having to allocate a very large ring or having variable-sized rings. [Note: the downside of this is that it's then less efficient to locate a packet for retransmission as we then have to step through a list and examine each buffer in the list.] (3) Allow the filler/encrypter to run ahead of the transmission window. (4) Make it easier to do zero copy UDP from the packet buffers. (5) Make it easier to do zero copy from userspace to the packet buffers - and thence to UDP (only if for unauthenticated connections). To that end, the following changes are made: (1) Use the new rxrpc_txbuf struct instead of sk_buff for keeping packets to be transmitted in. This allows them to be placed on multiple queues simultaneously. An sk_buff isn't really necessary as it's never passed on to lower-level networking code. (2) Keep the transmissable packets in a linked list on the call struct rather than in a ring. As a consequence, the annotation buffer isn't used either; rather a flag is set on the packet to indicate ackedness. (3) Use the RXRPC_CALL_TX_LAST flag to indicate that the last packet to be transmitted has been queued. Add RXRPC_CALL_TX_ALL_ACKED to indicate that all packets up to and including the last got hard acked. (4) Wire headers are now stored in the txbuf rather than being concocted on the stack and they're stored immediately before the data, thereby allowing zerocopy of a single span. (5) Don't bother with instant-resend on transmission failure; rather, leave it for a timer or an ACK packet to trigger. Signed-off-by: David Howells <dhowells@redhat.com> cc: Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org
2022-03-31 22:55:08 +00:00
struct rxrpc_txbuf *txb;
unsigned long resend_at;
rxrpc: Don't use a ring buffer for call Tx queue Change the way the Tx queueing works to make the following ends easier to achieve: (1) The filling of packets, the encryption of packets and the transmission of packets can be handled in parallel by separate threads, rather than rxrpc_sendmsg() allocating, filling, encrypting and transmitting each packet before moving onto the next one. (2) Get rid of the fixed-size ring which sets a hard limit on the number of packets that can be retained in the ring. This allows the number of packets to increase without having to allocate a very large ring or having variable-sized rings. [Note: the downside of this is that it's then less efficient to locate a packet for retransmission as we then have to step through a list and examine each buffer in the list.] (3) Allow the filler/encrypter to run ahead of the transmission window. (4) Make it easier to do zero copy UDP from the packet buffers. (5) Make it easier to do zero copy from userspace to the packet buffers - and thence to UDP (only if for unauthenticated connections). To that end, the following changes are made: (1) Use the new rxrpc_txbuf struct instead of sk_buff for keeping packets to be transmitted in. This allows them to be placed on multiple queues simultaneously. An sk_buff isn't really necessary as it's never passed on to lower-level networking code. (2) Keep the transmissable packets in a linked list on the call struct rather than in a ring. As a consequence, the annotation buffer isn't used either; rather a flag is set on the packet to indicate ackedness. (3) Use the RXRPC_CALL_TX_LAST flag to indicate that the last packet to be transmitted has been queued. Add RXRPC_CALL_TX_ALL_ACKED to indicate that all packets up to and including the last got hard acked. (4) Wire headers are now stored in the txbuf rather than being concocted on the stack and they're stored immediately before the data, thereby allowing zerocopy of a single span. (5) Don't bother with instant-resend on transmission failure; rather, leave it for a timer or an ACK packet to trigger. Signed-off-by: David Howells <dhowells@redhat.com> cc: Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org
2022-03-31 22:55:08 +00:00
rxrpc_seq_t transmitted = READ_ONCE(call->tx_transmitted);
rxrpc: Fix the excessive initial retransmission timeout rxrpc currently uses a fixed 4s retransmission timeout until the RTT is sufficiently sampled. This can cause problems with some fileservers with calls to the cache manager in the afs filesystem being dropped from the fileserver because a packet goes missing and the retransmission timeout is greater than the call expiry timeout. Fix this by: (1) Copying the RTT/RTO calculation code from Linux's TCP implementation and altering it to fit rxrpc. (2) Altering the various users of the RTT to make use of the new SRTT value. (3) Replacing the use of rxrpc_resend_timeout to use the calculated RTO value instead (which is needed in jiffies), along with a backoff. Notes: (1) rxrpc provides RTT samples by matching the serial numbers on outgoing DATA packets that have the RXRPC_REQUEST_ACK set and PING ACK packets against the reference serial number in incoming REQUESTED ACK and PING-RESPONSE ACK packets. (2) Each packet that is transmitted on an rxrpc connection gets a new per-connection serial number, even for retransmissions, so an ACK can be cross-referenced to a specific trigger packet. This allows RTT information to be drawn from retransmitted DATA packets also. (3) rxrpc maintains the RTT/RTO state on the rxrpc_peer record rather than on an rxrpc_call because many RPC calls won't live long enough to generate more than one sample. (4) The calculated SRTT value is in units of 8ths of a microsecond rather than nanoseconds. The (S)RTT and RTO values are displayed in /proc/net/rxrpc/peers. Fixes: 17926a79320a ([AF_RXRPC]: Provide secure RxRPC sockets for use by userspace and kernel both"") Signed-off-by: David Howells <dhowells@redhat.com>
2020-05-11 13:54:34 +00:00
ktime_t now, max_age, oldest, ack_ts;
rxrpc: Don't use a ring buffer for call Tx queue Change the way the Tx queueing works to make the following ends easier to achieve: (1) The filling of packets, the encryption of packets and the transmission of packets can be handled in parallel by separate threads, rather than rxrpc_sendmsg() allocating, filling, encrypting and transmitting each packet before moving onto the next one. (2) Get rid of the fixed-size ring which sets a hard limit on the number of packets that can be retained in the ring. This allows the number of packets to increase without having to allocate a very large ring or having variable-sized rings. [Note: the downside of this is that it's then less efficient to locate a packet for retransmission as we then have to step through a list and examine each buffer in the list.] (3) Allow the filler/encrypter to run ahead of the transmission window. (4) Make it easier to do zero copy UDP from the packet buffers. (5) Make it easier to do zero copy from userspace to the packet buffers - and thence to UDP (only if for unauthenticated connections). To that end, the following changes are made: (1) Use the new rxrpc_txbuf struct instead of sk_buff for keeping packets to be transmitted in. This allows them to be placed on multiple queues simultaneously. An sk_buff isn't really necessary as it's never passed on to lower-level networking code. (2) Keep the transmissable packets in a linked list on the call struct rather than in a ring. As a consequence, the annotation buffer isn't used either; rather a flag is set on the packet to indicate ackedness. (3) Use the RXRPC_CALL_TX_LAST flag to indicate that the last packet to be transmitted has been queued. Add RXRPC_CALL_TX_ALL_ACKED to indicate that all packets up to and including the last got hard acked. (4) Wire headers are now stored in the txbuf rather than being concocted on the stack and they're stored immediately before the data, thereby allowing zerocopy of a single span. (5) Don't bother with instant-resend on transmission failure; rather, leave it for a timer or an ACK packet to trigger. Signed-off-by: David Howells <dhowells@redhat.com> cc: Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org
2022-03-31 22:55:08 +00:00
bool unacked = false;
unsigned int i;
rxrpc: Don't use a ring buffer for call Tx queue Change the way the Tx queueing works to make the following ends easier to achieve: (1) The filling of packets, the encryption of packets and the transmission of packets can be handled in parallel by separate threads, rather than rxrpc_sendmsg() allocating, filling, encrypting and transmitting each packet before moving onto the next one. (2) Get rid of the fixed-size ring which sets a hard limit on the number of packets that can be retained in the ring. This allows the number of packets to increase without having to allocate a very large ring or having variable-sized rings. [Note: the downside of this is that it's then less efficient to locate a packet for retransmission as we then have to step through a list and examine each buffer in the list.] (3) Allow the filler/encrypter to run ahead of the transmission window. (4) Make it easier to do zero copy UDP from the packet buffers. (5) Make it easier to do zero copy from userspace to the packet buffers - and thence to UDP (only if for unauthenticated connections). To that end, the following changes are made: (1) Use the new rxrpc_txbuf struct instead of sk_buff for keeping packets to be transmitted in. This allows them to be placed on multiple queues simultaneously. An sk_buff isn't really necessary as it's never passed on to lower-level networking code. (2) Keep the transmissable packets in a linked list on the call struct rather than in a ring. As a consequence, the annotation buffer isn't used either; rather a flag is set on the packet to indicate ackedness. (3) Use the RXRPC_CALL_TX_LAST flag to indicate that the last packet to be transmitted has been queued. Add RXRPC_CALL_TX_ALL_ACKED to indicate that all packets up to and including the last got hard acked. (4) Wire headers are now stored in the txbuf rather than being concocted on the stack and they're stored immediately before the data, thereby allowing zerocopy of a single span. (5) Don't bother with instant-resend on transmission failure; rather, leave it for a timer or an ACK packet to trigger. Signed-off-by: David Howells <dhowells@redhat.com> cc: Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org
2022-03-31 22:55:08 +00:00
LIST_HEAD(retrans_queue);
rxrpc: Don't use a ring buffer for call Tx queue Change the way the Tx queueing works to make the following ends easier to achieve: (1) The filling of packets, the encryption of packets and the transmission of packets can be handled in parallel by separate threads, rather than rxrpc_sendmsg() allocating, filling, encrypting and transmitting each packet before moving onto the next one. (2) Get rid of the fixed-size ring which sets a hard limit on the number of packets that can be retained in the ring. This allows the number of packets to increase without having to allocate a very large ring or having variable-sized rings. [Note: the downside of this is that it's then less efficient to locate a packet for retransmission as we then have to step through a list and examine each buffer in the list.] (3) Allow the filler/encrypter to run ahead of the transmission window. (4) Make it easier to do zero copy UDP from the packet buffers. (5) Make it easier to do zero copy from userspace to the packet buffers - and thence to UDP (only if for unauthenticated connections). To that end, the following changes are made: (1) Use the new rxrpc_txbuf struct instead of sk_buff for keeping packets to be transmitted in. This allows them to be placed on multiple queues simultaneously. An sk_buff isn't really necessary as it's never passed on to lower-level networking code. (2) Keep the transmissable packets in a linked list on the call struct rather than in a ring. As a consequence, the annotation buffer isn't used either; rather a flag is set on the packet to indicate ackedness. (3) Use the RXRPC_CALL_TX_LAST flag to indicate that the last packet to be transmitted has been queued. Add RXRPC_CALL_TX_ALL_ACKED to indicate that all packets up to and including the last got hard acked. (4) Wire headers are now stored in the txbuf rather than being concocted on the stack and they're stored immediately before the data, thereby allowing zerocopy of a single span. (5) Don't bother with instant-resend on transmission failure; rather, leave it for a timer or an ACK packet to trigger. Signed-off-by: David Howells <dhowells@redhat.com> cc: Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org
2022-03-31 22:55:08 +00:00
_enter("{%d,%d}", call->acks_hard_ack, call->tx_top);
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
now = ktime_get_real();
max_age = ktime_sub_us(now, jiffies_to_usecs(call->peer->rto_j));
oldest = now;
if (list_empty(&call->tx_buffer))
goto no_resend;
if (list_empty(&call->tx_buffer))
goto no_further_resend;
trace_rxrpc_resend(call, ack_skb);
txb = list_first_entry(&call->tx_buffer, struct rxrpc_txbuf, call_link);
/* Scan the soft ACK table without dropping the lock and resend any
* explicitly NAK'd packets.
*/
if (ack_skb) {
ack = (void *)ack_skb->data + sizeof(struct rxrpc_wire_header);
for (i = 0; i < ack->nAcks; i++) {
rxrpc_seq_t seq;
if (ack->acks[i] & 1)
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 10:10:12 +00:00
continue;
seq = ntohl(ack->firstPacket) + i;
if (after(txb->seq, transmitted))
break;
if (after(txb->seq, seq))
continue; /* A new hard ACK probably came in */
list_for_each_entry_from(txb, &call->tx_buffer, call_link) {
if (txb->seq == seq)
goto found_txb;
}
goto no_further_resend;
found_txb:
if (after(ntohl(txb->wire.serial), call->acks_highest_serial))
continue; /* Ack point not yet reached */
rxrpc_see_txbuf(txb, rxrpc_txbuf_see_unacked);
if (list_empty(&txb->tx_link)) {
list_add_tail(&txb->tx_link, &retrans_queue);
set_bit(RXRPC_TXBUF_RESENT, &txb->flags);
}
trace_rxrpc_retransmit(call, txb->seq,
ktime_to_ns(ktime_sub(txb->last_sent,
max_age)));
if (list_is_last(&txb->call_link, &call->tx_buffer))
goto no_further_resend;
txb = list_next_entry(txb, call_link);
}
}
/* Fast-forward through the Tx queue to the point the peer says it has
* seen. Anything between the soft-ACK table and that point will get
* ACK'd or NACK'd in due course, so don't worry about it here; here we
* need to consider retransmitting anything beyond that point.
*
* Note that ACK for a packet can beat the update of tx_transmitted.
*/
if (after_eq(READ_ONCE(call->acks_prev_seq), READ_ONCE(call->tx_transmitted)))
goto no_further_resend;
list_for_each_entry_from(txb, &call->tx_buffer, call_link) {
if (before_eq(txb->seq, READ_ONCE(call->acks_prev_seq)))
continue;
if (after(txb->seq, READ_ONCE(call->tx_transmitted)))
break; /* Not transmitted yet */
if (ack && ack->reason == RXRPC_ACK_PING_RESPONSE &&
before(ntohl(txb->wire.serial), ntohl(ack->serial)))
goto do_resend; /* Wasn't accounted for by a more recent ping. */
if (ktime_after(txb->last_sent, max_age)) {
if (ktime_before(txb->last_sent, oldest))
oldest = txb->last_sent;
continue;
}
do_resend:
unacked = true;
if (list_empty(&txb->tx_link)) {
list_add_tail(&txb->tx_link, &retrans_queue);
set_bit(RXRPC_TXBUF_RESENT, &txb->flags);
rxrpc_inc_stat(call->rxnet, stat_tx_data_retrans);
}
}
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 10:10:12 +00:00
no_further_resend:
no_resend:
resend_at = nsecs_to_jiffies(ktime_to_ns(ktime_sub(now, oldest)));
rxrpc: Don't use a ring buffer for call Tx queue Change the way the Tx queueing works to make the following ends easier to achieve: (1) The filling of packets, the encryption of packets and the transmission of packets can be handled in parallel by separate threads, rather than rxrpc_sendmsg() allocating, filling, encrypting and transmitting each packet before moving onto the next one. (2) Get rid of the fixed-size ring which sets a hard limit on the number of packets that can be retained in the ring. This allows the number of packets to increase without having to allocate a very large ring or having variable-sized rings. [Note: the downside of this is that it's then less efficient to locate a packet for retransmission as we then have to step through a list and examine each buffer in the list.] (3) Allow the filler/encrypter to run ahead of the transmission window. (4) Make it easier to do zero copy UDP from the packet buffers. (5) Make it easier to do zero copy from userspace to the packet buffers - and thence to UDP (only if for unauthenticated connections). To that end, the following changes are made: (1) Use the new rxrpc_txbuf struct instead of sk_buff for keeping packets to be transmitted in. This allows them to be placed on multiple queues simultaneously. An sk_buff isn't really necessary as it's never passed on to lower-level networking code. (2) Keep the transmissable packets in a linked list on the call struct rather than in a ring. As a consequence, the annotation buffer isn't used either; rather a flag is set on the packet to indicate ackedness. (3) Use the RXRPC_CALL_TX_LAST flag to indicate that the last packet to be transmitted has been queued. Add RXRPC_CALL_TX_ALL_ACKED to indicate that all packets up to and including the last got hard acked. (4) Wire headers are now stored in the txbuf rather than being concocted on the stack and they're stored immediately before the data, thereby allowing zerocopy of a single span. (5) Don't bother with instant-resend on transmission failure; rather, leave it for a timer or an ACK packet to trigger. Signed-off-by: David Howells <dhowells@redhat.com> cc: Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org
2022-03-31 22:55:08 +00:00
resend_at += jiffies + rxrpc_get_rto_backoff(call->peer,
!list_empty(&retrans_queue));
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
WRITE_ONCE(call->resend_at, resend_at);
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 10:10:12 +00:00
if (unacked)
rxrpc_congestion_timeout(call);
/* If there was nothing that needed retransmission then it's likely
* that an ACK got lost somewhere. Send a ping to find out instead of
* retransmitting data.
*/
rxrpc: Don't use a ring buffer for call Tx queue Change the way the Tx queueing works to make the following ends easier to achieve: (1) The filling of packets, the encryption of packets and the transmission of packets can be handled in parallel by separate threads, rather than rxrpc_sendmsg() allocating, filling, encrypting and transmitting each packet before moving onto the next one. (2) Get rid of the fixed-size ring which sets a hard limit on the number of packets that can be retained in the ring. This allows the number of packets to increase without having to allocate a very large ring or having variable-sized rings. [Note: the downside of this is that it's then less efficient to locate a packet for retransmission as we then have to step through a list and examine each buffer in the list.] (3) Allow the filler/encrypter to run ahead of the transmission window. (4) Make it easier to do zero copy UDP from the packet buffers. (5) Make it easier to do zero copy from userspace to the packet buffers - and thence to UDP (only if for unauthenticated connections). To that end, the following changes are made: (1) Use the new rxrpc_txbuf struct instead of sk_buff for keeping packets to be transmitted in. This allows them to be placed on multiple queues simultaneously. An sk_buff isn't really necessary as it's never passed on to lower-level networking code. (2) Keep the transmissable packets in a linked list on the call struct rather than in a ring. As a consequence, the annotation buffer isn't used either; rather a flag is set on the packet to indicate ackedness. (3) Use the RXRPC_CALL_TX_LAST flag to indicate that the last packet to be transmitted has been queued. Add RXRPC_CALL_TX_ALL_ACKED to indicate that all packets up to and including the last got hard acked. (4) Wire headers are now stored in the txbuf rather than being concocted on the stack and they're stored immediately before the data, thereby allowing zerocopy of a single span. (5) Don't bother with instant-resend on transmission failure; rather, leave it for a timer or an ACK packet to trigger. Signed-off-by: David Howells <dhowells@redhat.com> cc: Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org
2022-03-31 22:55:08 +00:00
if (list_empty(&retrans_queue)) {
rxrpc_reduce_call_timer(call, resend_at, jiffies,
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
rxrpc_timer_set_for_resend);
ack_ts = ktime_sub(now, call->acks_latest_ts);
rxrpc: Fix the excessive initial retransmission timeout rxrpc currently uses a fixed 4s retransmission timeout until the RTT is sufficiently sampled. This can cause problems with some fileservers with calls to the cache manager in the afs filesystem being dropped from the fileserver because a packet goes missing and the retransmission timeout is greater than the call expiry timeout. Fix this by: (1) Copying the RTT/RTO calculation code from Linux's TCP implementation and altering it to fit rxrpc. (2) Altering the various users of the RTT to make use of the new SRTT value. (3) Replacing the use of rxrpc_resend_timeout to use the calculated RTO value instead (which is needed in jiffies), along with a backoff. Notes: (1) rxrpc provides RTT samples by matching the serial numbers on outgoing DATA packets that have the RXRPC_REQUEST_ACK set and PING ACK packets against the reference serial number in incoming REQUESTED ACK and PING-RESPONSE ACK packets. (2) Each packet that is transmitted on an rxrpc connection gets a new per-connection serial number, even for retransmissions, so an ACK can be cross-referenced to a specific trigger packet. This allows RTT information to be drawn from retransmitted DATA packets also. (3) rxrpc maintains the RTT/RTO state on the rxrpc_peer record rather than on an rxrpc_call because many RPC calls won't live long enough to generate more than one sample. (4) The calculated SRTT value is in units of 8ths of a microsecond rather than nanoseconds. The (S)RTT and RTO values are displayed in /proc/net/rxrpc/peers. Fixes: 17926a79320a ([AF_RXRPC]: Provide secure RxRPC sockets for use by userspace and kernel both"") Signed-off-by: David Howells <dhowells@redhat.com>
2020-05-11 13:54:34 +00:00
if (ktime_to_us(ack_ts) < (call->peer->srtt_us >> 3))
goto out;
rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
rxrpc_propose_ack_ping_for_lost_ack);
goto out;
}
/* Retransmit the queue */
rxrpc: Don't use a ring buffer for call Tx queue Change the way the Tx queueing works to make the following ends easier to achieve: (1) The filling of packets, the encryption of packets and the transmission of packets can be handled in parallel by separate threads, rather than rxrpc_sendmsg() allocating, filling, encrypting and transmitting each packet before moving onto the next one. (2) Get rid of the fixed-size ring which sets a hard limit on the number of packets that can be retained in the ring. This allows the number of packets to increase without having to allocate a very large ring or having variable-sized rings. [Note: the downside of this is that it's then less efficient to locate a packet for retransmission as we then have to step through a list and examine each buffer in the list.] (3) Allow the filler/encrypter to run ahead of the transmission window. (4) Make it easier to do zero copy UDP from the packet buffers. (5) Make it easier to do zero copy from userspace to the packet buffers - and thence to UDP (only if for unauthenticated connections). To that end, the following changes are made: (1) Use the new rxrpc_txbuf struct instead of sk_buff for keeping packets to be transmitted in. This allows them to be placed on multiple queues simultaneously. An sk_buff isn't really necessary as it's never passed on to lower-level networking code. (2) Keep the transmissable packets in a linked list on the call struct rather than in a ring. As a consequence, the annotation buffer isn't used either; rather a flag is set on the packet to indicate ackedness. (3) Use the RXRPC_CALL_TX_LAST flag to indicate that the last packet to be transmitted has been queued. Add RXRPC_CALL_TX_ALL_ACKED to indicate that all packets up to and including the last got hard acked. (4) Wire headers are now stored in the txbuf rather than being concocted on the stack and they're stored immediately before the data, thereby allowing zerocopy of a single span. (5) Don't bother with instant-resend on transmission failure; rather, leave it for a timer or an ACK packet to trigger. Signed-off-by: David Howells <dhowells@redhat.com> cc: Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org
2022-03-31 22:55:08 +00:00
while ((txb = list_first_entry_or_null(&retrans_queue,
struct rxrpc_txbuf, tx_link))) {
list_del_init(&txb->tx_link);
rxrpc_transmit_one(call, txb);
}
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 10:10:12 +00:00
out:
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 10:10:12 +00:00
_leave("");
}
/*
* Start transmitting the reply to a service. This cancels the need to ACK the
* request if we haven't yet done so.
*/
static void rxrpc_begin_service_reply(struct rxrpc_call *call)
{
unsigned long now = jiffies;
rxrpc_set_call_state(call, RXRPC_CALL_SERVER_SEND_REPLY);
WRITE_ONCE(call->delay_ack_at, now + MAX_JIFFY_OFFSET);
if (call->ackr_reason == RXRPC_ACK_DELAY)
call->ackr_reason = 0;
trace_rxrpc_timer(call, rxrpc_timer_init_for_send_reply, now);
}
/*
* Close the transmission phase. After this point there is no more data to be
* transmitted in the call.
*/
static void rxrpc_close_tx_phase(struct rxrpc_call *call)
{
_debug("________awaiting reply/ACK__________");
switch (__rxrpc_call_state(call)) {
case RXRPC_CALL_CLIENT_SEND_REQUEST:
rxrpc_set_call_state(call, RXRPC_CALL_CLIENT_AWAIT_REPLY);
break;
case RXRPC_CALL_SERVER_SEND_REPLY:
rxrpc_set_call_state(call, RXRPC_CALL_SERVER_AWAIT_ACK);
break;
default:
break;
}
}
static bool rxrpc_tx_window_has_space(struct rxrpc_call *call)
{
unsigned int winsize = min_t(unsigned int, call->tx_winsize,
call->cong_cwnd + call->cong_extra);
rxrpc_seq_t window = call->acks_hard_ack, wtop = window + winsize;
rxrpc_seq_t tx_top = call->tx_top;
int space;
space = wtop - tx_top;
return space > 0;
}
/*
* Decant some if the sendmsg prepared queue into the transmission buffer.
*/
static void rxrpc_decant_prepared_tx(struct rxrpc_call *call)
{
struct rxrpc_txbuf *txb;
if (!test_bit(RXRPC_CALL_EXPOSED, &call->flags)) {
if (list_empty(&call->tx_sendmsg))
return;
rxrpc_expose_client_call(call);
}
while ((txb = list_first_entry_or_null(&call->tx_sendmsg,
struct rxrpc_txbuf, call_link))) {
spin_lock(&call->tx_lock);
list_del(&txb->call_link);
spin_unlock(&call->tx_lock);
call->tx_top = txb->seq;
list_add_tail(&txb->call_link, &call->tx_buffer);
if (txb->wire.flags & RXRPC_LAST_PACKET)
rxrpc_close_tx_phase(call);
rxrpc_transmit_one(call, txb);
if (!rxrpc_tx_window_has_space(call))
break;
}
}
static void rxrpc_transmit_some_data(struct rxrpc_call *call)
{
switch (__rxrpc_call_state(call)) {
case RXRPC_CALL_SERVER_ACK_REQUEST:
if (list_empty(&call->tx_sendmsg))
return;
rxrpc_begin_service_reply(call);
fallthrough;
case RXRPC_CALL_SERVER_SEND_REPLY:
case RXRPC_CALL_CLIENT_SEND_REQUEST:
if (!rxrpc_tx_window_has_space(call))
return;
if (list_empty(&call->tx_sendmsg)) {
rxrpc_inc_stat(call->rxnet, stat_tx_data_underflow);
return;
}
rxrpc_decant_prepared_tx(call);
break;
default:
return;
}
}
/*
* Ping the other end to fill our RTT cache and to retrieve the rwind
* and MTU parameters.
*/
static void rxrpc_send_initial_ping(struct rxrpc_call *call)
{
if (call->peer->rtt_count < 3 ||
ktime_before(ktime_add_ms(call->peer->rtt_last_req, 1000),
ktime_get_real()))
rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
rxrpc_propose_ack_ping_for_params);
}
/*
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 10:10:12 +00:00
* Handle retransmission and deferred ACK/abort generation.
*/
bool rxrpc_input_call_event(struct rxrpc_call *call, struct sk_buff *skb)
{
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
unsigned long now, next, t;
rxrpc_serial_t ackr_serial;
bool resend = false, expired = false;
s32 abort_code;
rxrpc_see_call(call, rxrpc_call_see_input);
//printk("\n--------------------\n");
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 10:10:12 +00:00
_enter("{%d,%s,%lx}",
call->debug_id, rxrpc_call_states[__rxrpc_call_state(call)],
call->events);
if (__rxrpc_call_is_complete(call))
goto out;
/* Handle abort request locklessly, vs rxrpc_propose_abort(). */
abort_code = smp_load_acquire(&call->send_abort);
if (abort_code) {
rxrpc_abort_call(call, 0, call->send_abort, call->send_abort_err,
call->send_abort_why);
goto out;
}
if (skb && skb->mark == RXRPC_SKB_MARK_ERROR)
rxrpc: Don't hold a ref for call timer or workqueue Currently, rxrpc gives the call timer a ref on the call when it starts it and this is passed along to the workqueue by the timer expiration function. The problem comes when queue_work() fails (ie. the work item is already queued): the timer routine must put the ref - but this may cause the cleanup code to run. This has the unfortunate effect that the cleanup code may then be run in softirq context - which means that any spinlocks it might need to touch have to be guarded to disable softirqs (ie. they need a "_bh" suffix). Fix this by: (1) Don't give a ref to the timer. (2) Making the expiration function not do anything if the refcount is 0. Note that this is more of an optimisation. (3) Make sure that the cleanup routine waits for timer to complete. However, this has a consequence that timer cannot give a ref to the work item. Therefore the following fixes are also necessary: (4) Don't give a ref to the work item. (5) Make the work item return asap if it sees the ref count is 0. (6) Make sure that the cleanup routine waits for the work item to complete. Unfortunately, neither the timer nor the work item can simply get around the problem by just using refcount_inc_not_zero() as the waits would still have to be done, and there would still be the possibility of having to put the ref in the expiration function. Note the call work item is going to go away with the work being transferred to the I/O thread, so the wait in (6) will become obsolete. Signed-off-by: David Howells <dhowells@redhat.com> cc: Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org
2022-11-25 09:00:55 +00:00
goto out;
/* If we see our async-event poke, check for timeout trippage. */
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
now = jiffies;
t = READ_ONCE(call->expect_rx_by);
if (time_after_eq(now, t)) {
trace_rxrpc_timer(call, rxrpc_timer_exp_normal, now);
expired = true;
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
}
t = READ_ONCE(call->expect_req_by);
if (__rxrpc_call_state(call) == RXRPC_CALL_SERVER_RECV_REQUEST &&
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
time_after_eq(now, t)) {
trace_rxrpc_timer(call, rxrpc_timer_exp_idle, now);
expired = true;
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
}
t = READ_ONCE(call->expect_term_by);
if (time_after_eq(now, t)) {
trace_rxrpc_timer(call, rxrpc_timer_exp_hard, now);
expired = true;
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
}
t = READ_ONCE(call->delay_ack_at);
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
if (time_after_eq(now, t)) {
trace_rxrpc_timer(call, rxrpc_timer_exp_ack, now);
cmpxchg(&call->delay_ack_at, t, now + MAX_JIFFY_OFFSET);
ackr_serial = xchg(&call->ackr_serial, 0);
rxrpc_send_ACK(call, RXRPC_ACK_DELAY, ackr_serial,
rxrpc_propose_ack_ping_for_lost_ack);
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
}
t = READ_ONCE(call->ack_lost_at);
if (time_after_eq(now, t)) {
trace_rxrpc_timer(call, rxrpc_timer_exp_lost_ack, now);
cmpxchg(&call->ack_lost_at, t, now + MAX_JIFFY_OFFSET);
set_bit(RXRPC_CALL_EV_ACK_LOST, &call->events);
}
t = READ_ONCE(call->keepalive_at);
if (time_after_eq(now, t)) {
trace_rxrpc_timer(call, rxrpc_timer_exp_keepalive, now);
cmpxchg(&call->keepalive_at, t, now + MAX_JIFFY_OFFSET);
rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
rxrpc_propose_ack_ping_for_keepalive);
}
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
t = READ_ONCE(call->ping_at);
if (time_after_eq(now, t)) {
trace_rxrpc_timer(call, rxrpc_timer_exp_ping, now);
cmpxchg(&call->ping_at, t, now + MAX_JIFFY_OFFSET);
rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
rxrpc_propose_ack_ping_for_keepalive);
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
}
t = READ_ONCE(call->resend_at);
if (time_after_eq(now, t)) {
trace_rxrpc_timer(call, rxrpc_timer_exp_resend, now);
cmpxchg(&call->resend_at, t, now + MAX_JIFFY_OFFSET);
resend = true;
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
}
if (skb)
rxrpc_input_call_packet(call, skb);
rxrpc_transmit_some_data(call);
if (skb) {
struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
if (sp->hdr.type == RXRPC_PACKET_TYPE_ACK)
rxrpc_congestion_degrade(call);
}
if (test_and_clear_bit(RXRPC_CALL_EV_INITIAL_PING, &call->events))
rxrpc_send_initial_ping(call);
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
/* Process events */
if (expired) {
rxrpc: Fix handling of call quietly cancelled out on server Sometimes an in-progress call will stop responding on the fileserver when the fileserver quietly cancels the call with an internally marked abort (RX_CALL_DEAD), without sending an ABORT to the client. This causes the client's call to eventually expire from lack of incoming packets directed its way, which currently leads to it being cancelled locally with ETIME. Note that it's not currently clear as to why this happens as it's really hard to reproduce. The rotation policy implement by kAFS, however, doesn't differentiate between ETIME meaning we didn't get any response from the server and ETIME meaning the call got cancelled mid-flow. The latter leads to an oops when fetching data as the rotation partially resets the afs_read descriptor, which can result in a cleared page pointer being dereferenced because that page has already been filled. Handle this by the following means: (1) Set a flag on a call when we receive a packet for it. (2) Store the highest packet serial number so far received for a call (bearing in mind this may wrap). (3) If, when the "not received anything recently" timeout expires on a call, we've received at least one packet for a call and the connection as a whole has received packets more recently than that call, then cancel the call locally with ECONNRESET rather than ETIME. This indicates that the call was definitely in progress on the server. (4) In kAFS, if the rotation algorithm sees ECONNRESET rather than ETIME, don't try the next server, but rather abort the call. This avoids the oops as we don't try to reuse the afs_read struct. Rather, as-yet ungotten pages will be reread at a later data. Also: (5) Add an rxrpc tracepoint to log detection of the call being reset. Without this, I occasionally see an oops like the following: general protection fault: 0000 [#1] SMP PTI ... RIP: 0010:_copy_to_iter+0x204/0x310 RSP: 0018:ffff8800cae0f828 EFLAGS: 00010206 RAX: 0000000000000560 RBX: 0000000000000560 RCX: 0000000000000560 RDX: ffff8800cae0f968 RSI: ffff8800d58b3312 RDI: 0005080000000000 RBP: ffff8800cae0f968 R08: 0000000000000560 R09: ffff8800ca00f400 R10: ffff8800c36f28d4 R11: 00000000000008c4 R12: ffff8800cae0f958 R13: 0000000000000560 R14: ffff8800d58b3312 R15: 0000000000000560 FS: 00007fdaef108080(0000) GS:ffff8800ca680000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fb28a8fa000 CR3: 00000000d2a76002 CR4: 00000000001606e0 Call Trace: skb_copy_datagram_iter+0x14e/0x289 rxrpc_recvmsg_data.isra.0+0x6f3/0xf68 ? trace_buffer_unlock_commit_regs+0x4f/0x89 rxrpc_kernel_recv_data+0x149/0x421 afs_extract_data+0x1e0/0x798 ? afs_wait_for_call_to_complete+0xc9/0x52e afs_deliver_fs_fetch_data+0x33a/0x5ab afs_deliver_to_call+0x1ee/0x5e0 ? afs_wait_for_call_to_complete+0xc9/0x52e afs_wait_for_call_to_complete+0x12b/0x52e ? wake_up_q+0x54/0x54 afs_make_call+0x287/0x462 ? afs_fs_fetch_data+0x3e6/0x3ed ? rcu_read_lock_sched_held+0x5d/0x63 afs_fs_fetch_data+0x3e6/0x3ed afs_fetch_data+0xbb/0x14a afs_readpages+0x317/0x40d __do_page_cache_readahead+0x203/0x2ba ? ondemand_readahead+0x3a7/0x3c1 ondemand_readahead+0x3a7/0x3c1 generic_file_buffered_read+0x18b/0x62f __vfs_read+0xdb/0xfe vfs_read+0xb2/0x137 ksys_read+0x50/0x8c do_syscall_64+0x7d/0x1a0 entry_SYSCALL_64_after_hwframe+0x49/0xbe Note the weird value in RDI which is a result of trying to kmap() a NULL page pointer. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-06-03 01:17:39 +00:00
if (test_bit(RXRPC_CALL_RX_HEARD, &call->flags) &&
(int)call->conn->hi_serial - (int)call->rx_serial > 0) {
trace_rxrpc_call_reset(call);
rxrpc_abort_call(call, 0, RX_CALL_DEAD, -ECONNRESET,
rxrpc_abort_call_reset);
rxrpc: Fix handling of call quietly cancelled out on server Sometimes an in-progress call will stop responding on the fileserver when the fileserver quietly cancels the call with an internally marked abort (RX_CALL_DEAD), without sending an ABORT to the client. This causes the client's call to eventually expire from lack of incoming packets directed its way, which currently leads to it being cancelled locally with ETIME. Note that it's not currently clear as to why this happens as it's really hard to reproduce. The rotation policy implement by kAFS, however, doesn't differentiate between ETIME meaning we didn't get any response from the server and ETIME meaning the call got cancelled mid-flow. The latter leads to an oops when fetching data as the rotation partially resets the afs_read descriptor, which can result in a cleared page pointer being dereferenced because that page has already been filled. Handle this by the following means: (1) Set a flag on a call when we receive a packet for it. (2) Store the highest packet serial number so far received for a call (bearing in mind this may wrap). (3) If, when the "not received anything recently" timeout expires on a call, we've received at least one packet for a call and the connection as a whole has received packets more recently than that call, then cancel the call locally with ECONNRESET rather than ETIME. This indicates that the call was definitely in progress on the server. (4) In kAFS, if the rotation algorithm sees ECONNRESET rather than ETIME, don't try the next server, but rather abort the call. This avoids the oops as we don't try to reuse the afs_read struct. Rather, as-yet ungotten pages will be reread at a later data. Also: (5) Add an rxrpc tracepoint to log detection of the call being reset. Without this, I occasionally see an oops like the following: general protection fault: 0000 [#1] SMP PTI ... RIP: 0010:_copy_to_iter+0x204/0x310 RSP: 0018:ffff8800cae0f828 EFLAGS: 00010206 RAX: 0000000000000560 RBX: 0000000000000560 RCX: 0000000000000560 RDX: ffff8800cae0f968 RSI: ffff8800d58b3312 RDI: 0005080000000000 RBP: ffff8800cae0f968 R08: 0000000000000560 R09: ffff8800ca00f400 R10: ffff8800c36f28d4 R11: 00000000000008c4 R12: ffff8800cae0f958 R13: 0000000000000560 R14: ffff8800d58b3312 R15: 0000000000000560 FS: 00007fdaef108080(0000) GS:ffff8800ca680000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fb28a8fa000 CR3: 00000000d2a76002 CR4: 00000000001606e0 Call Trace: skb_copy_datagram_iter+0x14e/0x289 rxrpc_recvmsg_data.isra.0+0x6f3/0xf68 ? trace_buffer_unlock_commit_regs+0x4f/0x89 rxrpc_kernel_recv_data+0x149/0x421 afs_extract_data+0x1e0/0x798 ? afs_wait_for_call_to_complete+0xc9/0x52e afs_deliver_fs_fetch_data+0x33a/0x5ab afs_deliver_to_call+0x1ee/0x5e0 ? afs_wait_for_call_to_complete+0xc9/0x52e afs_wait_for_call_to_complete+0x12b/0x52e ? wake_up_q+0x54/0x54 afs_make_call+0x287/0x462 ? afs_fs_fetch_data+0x3e6/0x3ed ? rcu_read_lock_sched_held+0x5d/0x63 afs_fs_fetch_data+0x3e6/0x3ed afs_fetch_data+0xbb/0x14a afs_readpages+0x317/0x40d __do_page_cache_readahead+0x203/0x2ba ? ondemand_readahead+0x3a7/0x3c1 ondemand_readahead+0x3a7/0x3c1 generic_file_buffered_read+0x18b/0x62f __vfs_read+0xdb/0xfe vfs_read+0xb2/0x137 ksys_read+0x50/0x8c do_syscall_64+0x7d/0x1a0 entry_SYSCALL_64_after_hwframe+0x49/0xbe Note the weird value in RDI which is a result of trying to kmap() a NULL page pointer. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-06-03 01:17:39 +00:00
} else {
rxrpc_abort_call(call, 0, RX_CALL_TIMEOUT, -ETIME,
rxrpc_abort_call_timeout);
rxrpc: Fix handling of call quietly cancelled out on server Sometimes an in-progress call will stop responding on the fileserver when the fileserver quietly cancels the call with an internally marked abort (RX_CALL_DEAD), without sending an ABORT to the client. This causes the client's call to eventually expire from lack of incoming packets directed its way, which currently leads to it being cancelled locally with ETIME. Note that it's not currently clear as to why this happens as it's really hard to reproduce. The rotation policy implement by kAFS, however, doesn't differentiate between ETIME meaning we didn't get any response from the server and ETIME meaning the call got cancelled mid-flow. The latter leads to an oops when fetching data as the rotation partially resets the afs_read descriptor, which can result in a cleared page pointer being dereferenced because that page has already been filled. Handle this by the following means: (1) Set a flag on a call when we receive a packet for it. (2) Store the highest packet serial number so far received for a call (bearing in mind this may wrap). (3) If, when the "not received anything recently" timeout expires on a call, we've received at least one packet for a call and the connection as a whole has received packets more recently than that call, then cancel the call locally with ECONNRESET rather than ETIME. This indicates that the call was definitely in progress on the server. (4) In kAFS, if the rotation algorithm sees ECONNRESET rather than ETIME, don't try the next server, but rather abort the call. This avoids the oops as we don't try to reuse the afs_read struct. Rather, as-yet ungotten pages will be reread at a later data. Also: (5) Add an rxrpc tracepoint to log detection of the call being reset. Without this, I occasionally see an oops like the following: general protection fault: 0000 [#1] SMP PTI ... RIP: 0010:_copy_to_iter+0x204/0x310 RSP: 0018:ffff8800cae0f828 EFLAGS: 00010206 RAX: 0000000000000560 RBX: 0000000000000560 RCX: 0000000000000560 RDX: ffff8800cae0f968 RSI: ffff8800d58b3312 RDI: 0005080000000000 RBP: ffff8800cae0f968 R08: 0000000000000560 R09: ffff8800ca00f400 R10: ffff8800c36f28d4 R11: 00000000000008c4 R12: ffff8800cae0f958 R13: 0000000000000560 R14: ffff8800d58b3312 R15: 0000000000000560 FS: 00007fdaef108080(0000) GS:ffff8800ca680000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fb28a8fa000 CR3: 00000000d2a76002 CR4: 00000000001606e0 Call Trace: skb_copy_datagram_iter+0x14e/0x289 rxrpc_recvmsg_data.isra.0+0x6f3/0xf68 ? trace_buffer_unlock_commit_regs+0x4f/0x89 rxrpc_kernel_recv_data+0x149/0x421 afs_extract_data+0x1e0/0x798 ? afs_wait_for_call_to_complete+0xc9/0x52e afs_deliver_fs_fetch_data+0x33a/0x5ab afs_deliver_to_call+0x1ee/0x5e0 ? afs_wait_for_call_to_complete+0xc9/0x52e afs_wait_for_call_to_complete+0x12b/0x52e ? wake_up_q+0x54/0x54 afs_make_call+0x287/0x462 ? afs_fs_fetch_data+0x3e6/0x3ed ? rcu_read_lock_sched_held+0x5d/0x63 afs_fs_fetch_data+0x3e6/0x3ed afs_fetch_data+0xbb/0x14a afs_readpages+0x317/0x40d __do_page_cache_readahead+0x203/0x2ba ? ondemand_readahead+0x3a7/0x3c1 ondemand_readahead+0x3a7/0x3c1 generic_file_buffered_read+0x18b/0x62f __vfs_read+0xdb/0xfe vfs_read+0xb2/0x137 ksys_read+0x50/0x8c do_syscall_64+0x7d/0x1a0 entry_SYSCALL_64_after_hwframe+0x49/0xbe Note the weird value in RDI which is a result of trying to kmap() a NULL page pointer. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-06-03 01:17:39 +00:00
}
goto out;
}
if (test_and_clear_bit(RXRPC_CALL_EV_ACK_LOST, &call->events))
rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
rxrpc_propose_ack_ping_for_lost_ack);
if (resend && __rxrpc_call_state(call) != RXRPC_CALL_CLIENT_RECV_REPLY)
rxrpc_resend(call, NULL);
if (test_and_clear_bit(RXRPC_CALL_RX_IS_IDLE, &call->flags))
rxrpc_send_ACK(call, RXRPC_ACK_IDLE, 0,
rxrpc_propose_ack_rx_idle);
if (call->ackr_nr_unacked > 2) {
if (call->peer->rtt_count < 3)
rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
rxrpc_propose_ack_ping_for_rtt);
else if (ktime_before(ktime_add_ms(call->peer->rtt_last_req, 1000),
ktime_get_real()))
rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
rxrpc_propose_ack_ping_for_old_rtt);
else
rxrpc_send_ACK(call, RXRPC_ACK_IDLE, 0,
rxrpc_propose_ack_input_data);
}
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
/* Make sure the timer is restarted */
if (!__rxrpc_call_is_complete(call)) {
next = call->expect_rx_by;
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
#define set(T) { t = READ_ONCE(T); if (time_before(t, next)) next = t; }
set(call->expect_req_by);
set(call->expect_term_by);
set(call->delay_ack_at);
set(call->ack_lost_at);
set(call->resend_at);
set(call->keepalive_at);
set(call->ping_at);
rxrpc: Fix call timeouts Fix the rxrpc call expiration timeouts and make them settable from userspace. By analogy with other rx implementations, there should be three timeouts: (1) "Normal timeout" This is set for all calls and is triggered if we haven't received any packets from the peer in a while. It is measured from the last time we received any packet on that call. This is not reset by any connection packets (such as CHALLENGE/RESPONSE packets). If a service operation takes a long time, the server should generate PING ACKs at a duration that's substantially less than the normal timeout so is to keep both sides alive. This is set at 1/6 of normal timeout. (2) "Idle timeout" This is set only for a service call and is triggered if we stop receiving the DATA packets that comprise the request data. It is measured from the last time we received a DATA packet. (3) "Hard timeout" This can be set for a call and specified the maximum lifetime of that call. It should not be specified by default. Some operations (such as volume transfer) take a long time. Allow userspace to set/change the timeouts on a call with sendmsg, using a control message: RXRPC_SET_CALL_TIMEOUTS The data to the message is a number of 32-bit words, not all of which need be given: u32 hard_timeout; /* sec from first packet */ u32 idle_timeout; /* msec from packet Rx */ u32 normal_timeout; /* msec from data Rx */ This can be set in combination with any other sendmsg() that affects a call. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-24 10:18:41 +00:00
now = jiffies;
if (time_after_eq(now, next))
rxrpc_poke_call(call, rxrpc_call_poke_timer_now);
rxrpc_reduce_call_timer(call, next, now, rxrpc_timer_restart);
}
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 10:10:12 +00:00
out:
if (__rxrpc_call_is_complete(call)) {
del_timer_sync(&call->timer);
if (!test_bit(RXRPC_CALL_DISCONNECTED, &call->flags))
rxrpc_disconnect_call(call);
if (call->security)
call->security->free_call_crypto(call);
}
if (call->acks_hard_ack != call->tx_bottom)
rxrpc_shrink_call_tx_buffer(call);
_leave("");
return true;
}