linux/drivers/scsi/elx/libefc/efc_nport.c
James Smart fcd427303e scsi: elx: libefc: SLI and FC PORT state machine interfaces
Add library interface definitions for:

 - SLI and FC port (aka n_port_id) registration, allocation and
   deallocation.

Link: https://lore.kernel.org/r/20210601235512.20104-12-jsmart2021@gmail.com
Reviewed-by: Daniel Wagner <dwagner@suse.de>
Co-developed-by: Ram Vegesna <ram.vegesna@broadcom.com>
Signed-off-by: Ram Vegesna <ram.vegesna@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2021-06-15 23:39:30 -04:00

778 lines
18 KiB
C

// SPDX-License-Identifier: GPL-2.0
/*
* Copyright (C) 2021 Broadcom. All Rights Reserved. The term
* “Broadcom” refers to Broadcom Inc. and/or its subsidiaries.
*/
/*
* NPORT
*
* Port object for physical port and NPIV ports.
*/
/*
* NPORT REFERENCE COUNTING
*
* A nport reference should be taken when:
* - an nport is allocated
* - a vport populates associated nport
* - a remote node is allocated
* - a unsolicited frame is processed
* The reference should be dropped when:
* - the unsolicited frame processesing is done
* - the remote node is removed
* - the vport is removed
* - the nport is removed
*/
#include "efc.h"
void
efc_nport_cb(void *arg, int event, void *data)
{
struct efc *efc = arg;
struct efc_nport *nport = data;
unsigned long flags = 0;
efc_log_debug(efc, "nport event: %s\n", efc_sm_event_name(event));
spin_lock_irqsave(&efc->lock, flags);
efc_sm_post_event(&nport->sm, event, NULL);
spin_unlock_irqrestore(&efc->lock, flags);
}
static struct efc_nport *
efc_nport_find_wwn(struct efc_domain *domain, uint64_t wwnn, uint64_t wwpn)
{
struct efc_nport *nport = NULL;
/* Find a nport, given the WWNN and WWPN */
list_for_each_entry(nport, &domain->nport_list, list_entry) {
if (nport->wwnn == wwnn && nport->wwpn == wwpn)
return nport;
}
return NULL;
}
static void
_efc_nport_free(struct kref *arg)
{
struct efc_nport *nport = container_of(arg, struct efc_nport, ref);
kfree(nport);
}
struct efc_nport *
efc_nport_alloc(struct efc_domain *domain, uint64_t wwpn, uint64_t wwnn,
u32 fc_id, bool enable_ini, bool enable_tgt)
{
struct efc_nport *nport;
if (domain->efc->enable_ini)
enable_ini = 0;
/* Return a failure if this nport has already been allocated */
if ((wwpn != 0) || (wwnn != 0)) {
nport = efc_nport_find_wwn(domain, wwnn, wwpn);
if (nport) {
efc_log_err(domain->efc,
"NPORT %016llX %016llX already allocated\n",
wwnn, wwpn);
return NULL;
}
}
nport = kzalloc(sizeof(*nport), GFP_ATOMIC);
if (!nport)
return nport;
/* initialize refcount */
kref_init(&nport->ref);
nport->release = _efc_nport_free;
nport->efc = domain->efc;
snprintf(nport->display_name, sizeof(nport->display_name), "------");
nport->domain = domain;
xa_init(&nport->lookup);
nport->instance_index = domain->nport_count++;
nport->sm.app = nport;
nport->enable_ini = enable_ini;
nport->enable_tgt = enable_tgt;
nport->enable_rscn = (nport->enable_ini ||
(nport->enable_tgt && enable_target_rscn(nport->efc)));
/* Copy service parameters from domain */
memcpy(nport->service_params, domain->service_params,
sizeof(struct fc_els_flogi));
/* Update requested fc_id */
nport->fc_id = fc_id;
/* Update the nport's service parameters for the new wwn's */
nport->wwpn = wwpn;
nport->wwnn = wwnn;
snprintf(nport->wwnn_str, sizeof(nport->wwnn_str), "%016llX",
(unsigned long long)wwnn);
/*
* if this is the "first" nport of the domain,
* then make it the "phys" nport
*/
if (list_empty(&domain->nport_list))
domain->nport = nport;
INIT_LIST_HEAD(&nport->list_entry);
list_add_tail(&nport->list_entry, &domain->nport_list);
kref_get(&domain->ref);
efc_log_debug(domain->efc, "New Nport [%s]\n", nport->display_name);
return nport;
}
void
efc_nport_free(struct efc_nport *nport)
{
struct efc_domain *domain;
if (!nport)
return;
domain = nport->domain;
efc_log_debug(domain->efc, "[%s] free nport\n", nport->display_name);
list_del(&nport->list_entry);
/*
* if this is the physical nport,
* then clear it out of the domain
*/
if (nport == domain->nport)
domain->nport = NULL;
xa_destroy(&nport->lookup);
xa_erase(&domain->lookup, nport->fc_id);
if (list_empty(&domain->nport_list))
efc_domain_post_event(domain, EFC_EVT_ALL_CHILD_NODES_FREE,
NULL);
kref_put(&domain->ref, domain->release);
kref_put(&nport->ref, nport->release);
}
struct efc_nport *
efc_nport_find(struct efc_domain *domain, u32 d_id)
{
struct efc_nport *nport;
/* Find a nport object, given an FC_ID */
nport = xa_load(&domain->lookup, d_id);
if (!nport || !kref_get_unless_zero(&nport->ref))
return NULL;
return nport;
}
int
efc_nport_attach(struct efc_nport *nport, u32 fc_id)
{
int rc;
struct efc_node *node;
struct efc *efc = nport->efc;
unsigned long index;
/* Set our lookup */
rc = xa_err(xa_store(&nport->domain->lookup, fc_id, nport, GFP_ATOMIC));
if (rc) {
efc_log_err(efc, "Sport lookup store failed: %d\n", rc);
return rc;
}
/* Update our display_name */
efc_node_fcid_display(fc_id, nport->display_name,
sizeof(nport->display_name));
xa_for_each(&nport->lookup, index, node) {
efc_node_update_display_name(node);
}
efc_log_debug(nport->efc, "[%s] attach nport: fc_id x%06x\n",
nport->display_name, fc_id);
/* Register a nport, given an FC_ID */
rc = efc_cmd_nport_attach(efc, nport, fc_id);
if (rc < 0) {
efc_log_err(nport->efc,
"efc_hw_port_attach failed: %d\n", rc);
return -EIO;
}
return 0;
}
static void
efc_nport_shutdown(struct efc_nport *nport)
{
struct efc *efc = nport->efc;
struct efc_node *node;
unsigned long index;
xa_for_each(&nport->lookup, index, node) {
if (!(node->rnode.fc_id == FC_FID_FLOGI && nport->is_vport)) {
efc_node_post_event(node, EFC_EVT_SHUTDOWN, NULL);
continue;
}
/*
* If this is a vport, logout of the fabric
* controller so that it deletes the vport
* on the switch.
*/
/* if link is down, don't send logo */
if (efc->link_status == EFC_LINK_STATUS_DOWN) {
efc_node_post_event(node, EFC_EVT_SHUTDOWN, NULL);
continue;
}
efc_log_debug(efc, "[%s] nport shutdown vport, send logo\n",
node->display_name);
if (!efc_send_logo(node)) {
/* sent LOGO, wait for response */
efc_node_transition(node, __efc_d_wait_logo_rsp, NULL);
continue;
}
/*
* failed to send LOGO,
* go ahead and cleanup node anyways
*/
node_printf(node, "Failed to send LOGO\n");
efc_node_post_event(node, EFC_EVT_SHUTDOWN_EXPLICIT_LOGO, NULL);
}
}
static void
efc_vport_link_down(struct efc_nport *nport)
{
struct efc *efc = nport->efc;
struct efc_vport *vport;
/* Clear the nport reference in the vport specification */
list_for_each_entry(vport, &efc->vport_list, list_entry) {
if (vport->nport == nport) {
kref_put(&nport->ref, nport->release);
vport->nport = NULL;
break;
}
}
}
static void
__efc_nport_common(const char *funcname, struct efc_sm_ctx *ctx,
enum efc_sm_event evt, void *arg)
{
struct efc_nport *nport = ctx->app;
struct efc_domain *domain = nport->domain;
struct efc *efc = nport->efc;
switch (evt) {
case EFC_EVT_ENTER:
case EFC_EVT_REENTER:
case EFC_EVT_EXIT:
case EFC_EVT_ALL_CHILD_NODES_FREE:
break;
case EFC_EVT_NPORT_ATTACH_OK:
efc_sm_transition(ctx, __efc_nport_attached, NULL);
break;
case EFC_EVT_SHUTDOWN:
/* Flag this nport as shutting down */
nport->shutting_down = true;
if (nport->is_vport)
efc_vport_link_down(nport);
if (xa_empty(&nport->lookup)) {
/* Remove the nport from the domain's lookup table */
xa_erase(&domain->lookup, nport->fc_id);
efc_sm_transition(ctx, __efc_nport_wait_port_free,
NULL);
if (efc_cmd_nport_free(efc, nport)) {
efc_log_debug(nport->efc,
"efc_hw_port_free failed\n");
/* Not much we can do, free the nport anyways */
efc_nport_free(nport);
}
} else {
/* sm: node list is not empty / shutdown nodes */
efc_sm_transition(ctx,
__efc_nport_wait_shutdown, NULL);
efc_nport_shutdown(nport);
}
break;
default:
efc_log_debug(nport->efc, "[%s] %-20s %-20s not handled\n",
nport->display_name, funcname,
efc_sm_event_name(evt));
}
}
void
__efc_nport_allocated(struct efc_sm_ctx *ctx,
enum efc_sm_event evt, void *arg)
{
struct efc_nport *nport = ctx->app;
struct efc_domain *domain = nport->domain;
nport_sm_trace(nport);
switch (evt) {
/* the physical nport is attached */
case EFC_EVT_NPORT_ATTACH_OK:
WARN_ON(nport != domain->nport);
efc_sm_transition(ctx, __efc_nport_attached, NULL);
break;
case EFC_EVT_NPORT_ALLOC_OK:
/* ignore */
break;
default:
__efc_nport_common(__func__, ctx, evt, arg);
}
}
void
__efc_nport_vport_init(struct efc_sm_ctx *ctx,
enum efc_sm_event evt, void *arg)
{
struct efc_nport *nport = ctx->app;
struct efc *efc = nport->efc;
nport_sm_trace(nport);
switch (evt) {
case EFC_EVT_ENTER: {
__be64 be_wwpn = cpu_to_be64(nport->wwpn);
if (nport->wwpn == 0)
efc_log_debug(efc, "vport: letting f/w select WWN\n");
if (nport->fc_id != U32_MAX) {
efc_log_debug(efc, "vport: hard coding port id: %x\n",
nport->fc_id);
}
efc_sm_transition(ctx, __efc_nport_vport_wait_alloc, NULL);
/* If wwpn is zero, then we'll let the f/w assign wwpn*/
if (efc_cmd_nport_alloc(efc, nport, nport->domain,
nport->wwpn == 0 ? NULL :
(uint8_t *)&be_wwpn)) {
efc_log_err(efc, "Can't allocate port\n");
break;
}
break;
}
default:
__efc_nport_common(__func__, ctx, evt, arg);
}
}
void
__efc_nport_vport_wait_alloc(struct efc_sm_ctx *ctx,
enum efc_sm_event evt, void *arg)
{
struct efc_nport *nport = ctx->app;
struct efc *efc = nport->efc;
nport_sm_trace(nport);
switch (evt) {
case EFC_EVT_NPORT_ALLOC_OK: {
struct fc_els_flogi *sp;
sp = (struct fc_els_flogi *)nport->service_params;
if (nport->wwnn == 0) {
nport->wwnn = be64_to_cpu(nport->sli_wwnn);
nport->wwpn = be64_to_cpu(nport->sli_wwpn);
snprintf(nport->wwnn_str, sizeof(nport->wwnn_str),
"%016llX", nport->wwpn);
}
/* Update the nport's service parameters */
sp->fl_wwpn = cpu_to_be64(nport->wwpn);
sp->fl_wwnn = cpu_to_be64(nport->wwnn);
/*
* if nport->fc_id is uninitialized,
* then request that the fabric node use FDISC
* to find an fc_id.
* Otherwise we're restoring vports, or we're in
* fabric emulation mode, so attach the fc_id
*/
if (nport->fc_id == U32_MAX) {
struct efc_node *fabric;
fabric = efc_node_alloc(nport, FC_FID_FLOGI, false,
false);
if (!fabric) {
efc_log_err(efc, "efc_node_alloc() failed\n");
return;
}
efc_node_transition(fabric, __efc_vport_fabric_init,
NULL);
} else {
snprintf(nport->wwnn_str, sizeof(nport->wwnn_str),
"%016llX", nport->wwpn);
efc_nport_attach(nport, nport->fc_id);
}
efc_sm_transition(ctx, __efc_nport_vport_allocated, NULL);
break;
}
default:
__efc_nport_common(__func__, ctx, evt, arg);
}
}
void
__efc_nport_vport_allocated(struct efc_sm_ctx *ctx,
enum efc_sm_event evt, void *arg)
{
struct efc_nport *nport = ctx->app;
struct efc *efc = nport->efc;
nport_sm_trace(nport);
/*
* This state is entered after the nport is allocated;
* it then waits for a fabric node
* FDISC to complete, which requests a nport attach.
* The nport attach complete is handled in this state.
*/
switch (evt) {
case EFC_EVT_NPORT_ATTACH_OK: {
struct efc_node *node;
/* Find our fabric node, and forward this event */
node = efc_node_find(nport, FC_FID_FLOGI);
if (!node) {
efc_log_debug(efc, "can't find node %06x\n", FC_FID_FLOGI);
break;
}
/* sm: / forward nport attach to fabric node */
efc_node_post_event(node, evt, NULL);
efc_sm_transition(ctx, __efc_nport_attached, NULL);
break;
}
default:
__efc_nport_common(__func__, ctx, evt, arg);
}
}
static void
efc_vport_update_spec(struct efc_nport *nport)
{
struct efc *efc = nport->efc;
struct efc_vport *vport;
unsigned long flags = 0;
spin_lock_irqsave(&efc->vport_lock, flags);
list_for_each_entry(vport, &efc->vport_list, list_entry) {
if (vport->nport == nport) {
vport->wwnn = nport->wwnn;
vport->wwpn = nport->wwpn;
vport->tgt_data = nport->tgt_data;
vport->ini_data = nport->ini_data;
break;
}
}
spin_unlock_irqrestore(&efc->vport_lock, flags);
}
void
__efc_nport_attached(struct efc_sm_ctx *ctx,
enum efc_sm_event evt, void *arg)
{
struct efc_nport *nport = ctx->app;
struct efc *efc = nport->efc;
nport_sm_trace(nport);
switch (evt) {
case EFC_EVT_ENTER: {
struct efc_node *node;
unsigned long index;
efc_log_debug(efc,
"[%s] NPORT attached WWPN %016llX WWNN %016llX\n",
nport->display_name,
nport->wwpn, nport->wwnn);
xa_for_each(&nport->lookup, index, node)
efc_node_update_display_name(node);
efc->tt.new_nport(efc, nport);
/*
* Update the vport (if its not the physical nport)
* parameters
*/
if (nport->is_vport)
efc_vport_update_spec(nport);
break;
}
case EFC_EVT_EXIT:
efc_log_debug(efc,
"[%s] NPORT deattached WWPN %016llX WWNN %016llX\n",
nport->display_name,
nport->wwpn, nport->wwnn);
efc->tt.del_nport(efc, nport);
break;
default:
__efc_nport_common(__func__, ctx, evt, arg);
}
}
void
__efc_nport_wait_shutdown(struct efc_sm_ctx *ctx,
enum efc_sm_event evt, void *arg)
{
struct efc_nport *nport = ctx->app;
struct efc_domain *domain = nport->domain;
struct efc *efc = nport->efc;
nport_sm_trace(nport);
switch (evt) {
case EFC_EVT_NPORT_ALLOC_OK:
case EFC_EVT_NPORT_ALLOC_FAIL:
case EFC_EVT_NPORT_ATTACH_OK:
case EFC_EVT_NPORT_ATTACH_FAIL:
/* ignore these events - just wait for the all free event */
break;
case EFC_EVT_ALL_CHILD_NODES_FREE: {
/*
* Remove the nport from the domain's
* sparse vector lookup table
*/
xa_erase(&domain->lookup, nport->fc_id);
efc_sm_transition(ctx, __efc_nport_wait_port_free, NULL);
if (efc_cmd_nport_free(efc, nport)) {
efc_log_err(nport->efc, "efc_hw_port_free failed\n");
/* Not much we can do, free the nport anyways */
efc_nport_free(nport);
}
break;
}
default:
__efc_nport_common(__func__, ctx, evt, arg);
}
}
void
__efc_nport_wait_port_free(struct efc_sm_ctx *ctx,
enum efc_sm_event evt, void *arg)
{
struct efc_nport *nport = ctx->app;
nport_sm_trace(nport);
switch (evt) {
case EFC_EVT_NPORT_ATTACH_OK:
/* Ignore as we are waiting for the free CB */
break;
case EFC_EVT_NPORT_FREE_OK: {
/* All done, free myself */
efc_nport_free(nport);
break;
}
default:
__efc_nport_common(__func__, ctx, evt, arg);
}
}
static int
efc_vport_nport_alloc(struct efc_domain *domain, struct efc_vport *vport)
{
struct efc_nport *nport;
lockdep_assert_held(&domain->efc->lock);
nport = efc_nport_alloc(domain, vport->wwpn, vport->wwnn, vport->fc_id,
vport->enable_ini, vport->enable_tgt);
vport->nport = nport;
if (!nport)
return -EIO;
kref_get(&nport->ref);
nport->is_vport = true;
nport->tgt_data = vport->tgt_data;
nport->ini_data = vport->ini_data;
efc_sm_transition(&nport->sm, __efc_nport_vport_init, NULL);
return 0;
}
int
efc_vport_start(struct efc_domain *domain)
{
struct efc *efc = domain->efc;
struct efc_vport *vport;
struct efc_vport *next;
int rc = 0;
unsigned long flags = 0;
/* Use the vport spec to find the associated vports and start them */
spin_lock_irqsave(&efc->vport_lock, flags);
list_for_each_entry_safe(vport, next, &efc->vport_list, list_entry) {
if (!vport->nport) {
if (efc_vport_nport_alloc(domain, vport))
rc = -EIO;
}
}
spin_unlock_irqrestore(&efc->vport_lock, flags);
return rc;
}
int
efc_nport_vport_new(struct efc_domain *domain, uint64_t wwpn, uint64_t wwnn,
u32 fc_id, bool ini, bool tgt, void *tgt_data,
void *ini_data)
{
struct efc *efc = domain->efc;
struct efc_vport *vport;
int rc = 0;
unsigned long flags = 0;
if (ini && domain->efc->enable_ini == 0) {
efc_log_debug(efc, "driver initiator mode not enabled\n");
return -EIO;
}
if (tgt && domain->efc->enable_tgt == 0) {
efc_log_debug(efc, "driver target mode not enabled\n");
return -EIO;
}
/*
* Create a vport spec if we need to recreate
* this vport after a link up event
*/
vport = efc_vport_create_spec(domain->efc, wwnn, wwpn, fc_id, ini, tgt,
tgt_data, ini_data);
if (!vport) {
efc_log_err(efc, "failed to create vport object entry\n");
return -EIO;
}
spin_lock_irqsave(&efc->lock, flags);
rc = efc_vport_nport_alloc(domain, vport);
spin_unlock_irqrestore(&efc->lock, flags);
return rc;
}
int
efc_nport_vport_del(struct efc *efc, struct efc_domain *domain,
u64 wwpn, uint64_t wwnn)
{
struct efc_nport *nport;
struct efc_vport *vport;
struct efc_vport *next;
unsigned long flags = 0;
spin_lock_irqsave(&efc->vport_lock, flags);
/* walk the efc_vport_list and remove from there */
list_for_each_entry_safe(vport, next, &efc->vport_list, list_entry) {
if (vport->wwpn == wwpn && vport->wwnn == wwnn) {
list_del(&vport->list_entry);
kfree(vport);
break;
}
}
spin_unlock_irqrestore(&efc->vport_lock, flags);
if (!domain) {
/* No domain means no nport to look for */
return 0;
}
spin_lock_irqsave(&efc->lock, flags);
list_for_each_entry(nport, &domain->nport_list, list_entry) {
if (nport->wwpn == wwpn && nport->wwnn == wwnn) {
kref_put(&nport->ref, nport->release);
/* Shutdown this NPORT */
efc_sm_post_event(&nport->sm, EFC_EVT_SHUTDOWN, NULL);
break;
}
}
spin_unlock_irqrestore(&efc->lock, flags);
return 0;
}
void
efc_vport_del_all(struct efc *efc)
{
struct efc_vport *vport;
struct efc_vport *next;
unsigned long flags = 0;
spin_lock_irqsave(&efc->vport_lock, flags);
list_for_each_entry_safe(vport, next, &efc->vport_list, list_entry) {
list_del(&vport->list_entry);
kfree(vport);
}
spin_unlock_irqrestore(&efc->vport_lock, flags);
}
struct efc_vport *
efc_vport_create_spec(struct efc *efc, uint64_t wwnn, uint64_t wwpn,
u32 fc_id, bool enable_ini,
bool enable_tgt, void *tgt_data, void *ini_data)
{
struct efc_vport *vport;
unsigned long flags = 0;
/*
* walk the efc_vport_list and return failure
* if a valid(vport with non zero WWPN and WWNN) vport entry
* is already created
*/
spin_lock_irqsave(&efc->vport_lock, flags);
list_for_each_entry(vport, &efc->vport_list, list_entry) {
if ((wwpn && vport->wwpn == wwpn) &&
(wwnn && vport->wwnn == wwnn)) {
efc_log_err(efc,
"VPORT %016llX %016llX already allocated\n",
wwnn, wwpn);
spin_unlock_irqrestore(&efc->vport_lock, flags);
return NULL;
}
}
vport = kzalloc(sizeof(*vport), GFP_ATOMIC);
if (!vport) {
spin_unlock_irqrestore(&efc->vport_lock, flags);
return NULL;
}
vport->wwnn = wwnn;
vport->wwpn = wwpn;
vport->fc_id = fc_id;
vport->enable_tgt = enable_tgt;
vport->enable_ini = enable_ini;
vport->tgt_data = tgt_data;
vport->ini_data = ini_data;
INIT_LIST_HEAD(&vport->list_entry);
list_add_tail(&vport->list_entry, &efc->vport_list);
spin_unlock_irqrestore(&efc->vport_lock, flags);
return vport;
}