* [PATCH 1/8] lpfc: Revert LOG_TRACE_EVENT back to LOG_INIT prior to driver_resource_setup
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
@ 2021-10-20 21:14 ` James Smart
2021-10-20 21:14 ` [PATCH 2/8] lpfc: Wait for successful restart of SLI3 adapter during host sg_reset James Smart
` (8 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: James Smart @ 2021-10-20 21:14 UTC (permalink / raw)
To: linux-scsi; +Cc: James Smart, Justin Tee, Zheyu Ma
In cases when lpfc_enable_pci_dev fails, lpfc_printf_log with
LOG_TRACE_EVENT set will call lpfc_dmp_dbg which uses the
phba->port_list_lock.
However, phba->port_list_lock does not get initialized until
lpfc_setup_driver_resource_phase1. Thus, any initialization routine with
LOG_TRACE_EVENT log message prior to lpfc_setup_driver_resource_phase1
will crash.
Revert LOG_TRACE_EVENT back to LOG_INIT for all log messages in routines
prior to lpfc_setup_driver_resource_phase1.
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
CC: Zheyu Ma <zheyuma97@gmail.com>
---
drivers/scsi/lpfc/lpfc_init.c | 6 +++---
drivers/scsi/lpfc/lpfc_scsi.c | 2 +-
drivers/scsi/lpfc/lpfc_sli.c | 4 ++--
3 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index b7de66f89b3b..5e959c4f05c2 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -7385,7 +7385,7 @@ lpfc_enable_pci_dev(struct lpfc_hba *phba)
out_disable_device:
pci_disable_device(pdev);
out_error:
- lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
+ lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"1401 Failed to enable pci device\n");
return -ENODEV;
}
@@ -8428,7 +8428,7 @@ lpfc_init_api_table_setup(struct lpfc_hba *phba, uint8_t dev_grp)
phba->lpfc_stop_port = lpfc_stop_port_s4;
break;
default:
- lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
+ lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"1431 Invalid HBA PCI-device group: 0x%x\n",
dev_grp);
return -ENODEV;
@@ -11661,7 +11661,7 @@ lpfc_sli4_pci_mem_setup(struct lpfc_hba *phba)
/* There is no SLI3 failback for SLI4 devices. */
if (bf_get(lpfc_sli_intf_valid, &phba->sli4_hba.sli_intf) !=
LPFC_SLI_INTF_VALID) {
- lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
+ lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"2894 SLI_INTF reg contents invalid "
"sli_intf reg 0x%x\n",
phba->sli4_hba.sli_intf.word0);
diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c
index 83ad03d22a49..f35246024988 100644
--- a/drivers/scsi/lpfc/lpfc_scsi.c
+++ b/drivers/scsi/lpfc/lpfc_scsi.c
@@ -5115,7 +5115,7 @@ lpfc_scsi_api_table_setup(struct lpfc_hba *phba, uint8_t dev_grp)
phba->lpfc_scsi_prep_cmnd_buf = lpfc_scsi_prep_cmnd_buf_s4;
break;
default:
- lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
+ lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"1418 Invalid HBA PCI-device group: 0x%x\n",
dev_grp);
return -ENODEV;
diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c
index d800c6a69c95..244e7d68428e 100644
--- a/drivers/scsi/lpfc/lpfc_sli.c
+++ b/drivers/scsi/lpfc/lpfc_sli.c
@@ -10026,7 +10026,7 @@ lpfc_mbox_api_table_setup(struct lpfc_hba *phba, uint8_t dev_grp)
phba->lpfc_sli_brdready = lpfc_sli_brdready_s4;
break;
default:
- lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
+ lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"1420 Invalid HBA PCI-device group: 0x%x\n",
dev_grp);
return -ENODEV;
@@ -11194,7 +11194,7 @@ lpfc_sli_api_table_setup(struct lpfc_hba *phba, uint8_t dev_grp)
phba->__lpfc_sli_issue_fcp_io = __lpfc_sli_issue_fcp_io_s4;
break;
default:
- lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
+ lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"1419 Invalid HBA PCI-device group: 0x%x\n",
dev_grp);
return -ENODEV;
--
2.26.2
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 2/8] lpfc: Wait for successful restart of SLI3 adapter during host sg_reset
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
2021-10-20 21:14 ` [PATCH 1/8] lpfc: Revert LOG_TRACE_EVENT back to LOG_INIT prior to driver_resource_setup James Smart
@ 2021-10-20 21:14 ` James Smart
2021-10-20 21:14 ` [PATCH 3/8] lpfc: Correct sysfs reporting of loop support after SFP status change James Smart
` (7 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: James Smart @ 2021-10-20 21:14 UTC (permalink / raw)
To: linux-scsi; +Cc: James Smart, Justin Tee
A prior patch introduced HBA_NEEDS_CFG_PORT flag logic, but in
lpfc_sli_brdrestart_s3 code path, right after HBA_NEEDS_CFG_PORT is set,
the phba->hba_flag is cleared in lpfc_sli_brdreset.
Fix by calling lpfc_sli_chipset_init to wait for successful restart of
the HBA in lpfc_host_reset_handler after lpfc_sli_brdrestart.
lpfc_sli_chipset_init sets the HBA_NEEDS_CFG_PORT flag so that the
lpfc_sli_hba_setup routine from lpfc_online will execute
lpfc_sli_config_port initialization step when the brdrestart is
successful.
Fixes: d2f2547efd39 ("scsi: lpfc: Fix auto sli_mode and its effect on CONFIG_PORT for SLI3")
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
---
drivers/scsi/lpfc/lpfc_scsi.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c
index f35246024988..0dce4b51ca1e 100644
--- a/drivers/scsi/lpfc/lpfc_scsi.c
+++ b/drivers/scsi/lpfc/lpfc_scsi.c
@@ -6648,6 +6648,13 @@ lpfc_host_reset_handler(struct scsi_cmnd *cmnd)
if (rc)
goto error;
+ /* Wait for successful restart of adapter */
+ if (phba->sli_rev < LPFC_SLI_REV4) {
+ rc = lpfc_sli_chipset_init(phba);
+ if (rc)
+ goto error;
+ }
+
rc = lpfc_online(phba);
if (rc)
goto error;
--
2.26.2
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 3/8] lpfc: Correct sysfs reporting of loop support after SFP status change
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
2021-10-20 21:14 ` [PATCH 1/8] lpfc: Revert LOG_TRACE_EVENT back to LOG_INIT prior to driver_resource_setup James Smart
2021-10-20 21:14 ` [PATCH 2/8] lpfc: Wait for successful restart of SLI3 adapter during host sg_reset James Smart
@ 2021-10-20 21:14 ` James Smart
2021-10-20 21:14 ` [PATCH 4/8] lpfc: Fix use-after-free in lpfc_unreg_rpi() routine James Smart
` (6 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: James Smart @ 2021-10-20 21:14 UTC (permalink / raw)
To: linux-scsi; +Cc: James Smart, Justin Tee
Applications determine loop support in part by querying the 'pls'
sysfs node. Reporting of 'pls' (Private Loop Support) is derived from
the descriptor returned by the COMMON_GET_SLI4_PARAMETERS mailbox
command, which is issued during initialization or after a reset.
The value of this field may change if there is a dynamic SFP change.
The driver currently will not pick up the change as there was no reset
scenario.
Rework to commonize the sending of the COMMON_GET_SLI4_PARAMETERS
command. Add the calling of the routine after receipt of an async event
indicating an SFP change.
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
---
drivers/scsi/lpfc/lpfc_crtn.h | 1 +
drivers/scsi/lpfc/lpfc_init.c | 50 +++++++++++++++++++++++++++++++++++
drivers/scsi/lpfc/lpfc_sli.c | 25 +++---------------
3 files changed, 55 insertions(+), 21 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_crtn.h b/drivers/scsi/lpfc/lpfc_crtn.h
index c512f4199142..f0bcfeecd161 100644
--- a/drivers/scsi/lpfc/lpfc_crtn.h
+++ b/drivers/scsi/lpfc/lpfc_crtn.h
@@ -205,6 +205,7 @@ void lpfc_delayed_disc_timeout_handler(struct lpfc_vport *);
int lpfc_config_port_prep(struct lpfc_hba *);
void lpfc_update_vport_wwn(struct lpfc_vport *vport);
int lpfc_config_port_post(struct lpfc_hba *);
+int lpfc_sli4_refresh_params(struct lpfc_hba *phba);
int lpfc_hba_down_prep(struct lpfc_hba *);
int lpfc_hba_down_post(struct lpfc_hba *);
void lpfc_hba_init(struct lpfc_hba *, uint32_t *);
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index 5e959c4f05c2..744da35b5074 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -662,6 +662,50 @@ lpfc_config_port_post(struct lpfc_hba *phba)
return 0;
}
+/**
+ * lpfc_sli4_refresh_params - update driver copy of params.
+ * @phba: Pointer to HBA context object.
+ *
+ * This is called to refresh driver copy of dynamic fields from the
+ * common_get_sli4_parameters descriptor.
+ **/
+int
+lpfc_sli4_refresh_params(struct lpfc_hba *phba)
+{
+ LPFC_MBOXQ_t *mboxq;
+ struct lpfc_mqe *mqe;
+ struct lpfc_sli4_parameters *mbx_sli4_parameters;
+ int length, rc;
+
+ mboxq = (LPFC_MBOXQ_t *)mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
+ if (!mboxq)
+ return -ENOMEM;
+
+ mqe = &mboxq->u.mqe;
+ /* Read the port's SLI4 Config Parameters */
+ length = (sizeof(struct lpfc_mbx_get_sli4_parameters) -
+ sizeof(struct lpfc_sli4_cfg_mhdr));
+ lpfc_sli4_config(phba, mboxq, LPFC_MBOX_SUBSYSTEM_COMMON,
+ LPFC_MBOX_OPCODE_GET_SLI4_PARAMETERS,
+ length, LPFC_SLI4_MBX_EMBED);
+
+ rc = lpfc_sli_issue_mbox(phba, mboxq, MBX_POLL);
+ if (unlikely(rc)) {
+ mempool_free(mboxq, phba->mbox_mem_pool);
+ return rc;
+ }
+ mbx_sli4_parameters = &mqe->un.get_sli4_parameters.sli4_parameters;
+ phba->sli4_hba.pc_sli4_params.mi_ver =
+ bf_get(cfg_mi_ver, mbx_sli4_parameters);
+ phba->sli4_hba.pc_sli4_params.cmf =
+ bf_get(cfg_cmf, mbx_sli4_parameters);
+ phba->sli4_hba.pc_sli4_params.pls =
+ bf_get(cfg_pvl, mbx_sli4_parameters);
+
+ mempool_free(mboxq, phba->mbox_mem_pool);
+ return rc;
+}
+
/**
* lpfc_hba_init_link - Initialize the FC link
* @phba: pointer to lpfc hba data structure.
@@ -6455,6 +6499,12 @@ lpfc_sli4_async_sli_evt(struct lpfc_hba *phba, struct lpfc_acqe_sli *acqe_sli)
"3194 Unable to retrieve supported "
"speeds, rc = 0x%x\n", rc);
}
+ rc = lpfc_sli4_refresh_params(phba);
+ if (rc) {
+ lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
+ "3174 Unable to update pls support, "
+ "rc x%x\n", rc);
+ }
vports = lpfc_create_vport_work_array(phba);
if (vports != NULL) {
for (i = 0; i <= phba->max_vports && vports[i] != NULL;
diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c
index 244e7d68428e..f82f809617a0 100644
--- a/drivers/scsi/lpfc/lpfc_sli.c
+++ b/drivers/scsi/lpfc/lpfc_sli.c
@@ -7891,36 +7891,19 @@ static int
lpfc_cmf_setup(struct lpfc_hba *phba)
{
LPFC_MBOXQ_t *mboxq;
- struct lpfc_mqe *mqe;
struct lpfc_dmabuf *mp;
struct lpfc_pc_sli4_params *sli4_params;
- struct lpfc_sli4_parameters *mbx_sli4_parameters;
- int length;
int rc, cmf, mi_ver;
+ rc = lpfc_sli4_refresh_params(phba);
+ if (unlikely(rc))
+ return rc;
+
mboxq = (LPFC_MBOXQ_t *)mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
if (!mboxq)
return -ENOMEM;
- mqe = &mboxq->u.mqe;
-
- /* Read the port's SLI4 Config Parameters */
- length = (sizeof(struct lpfc_mbx_get_sli4_parameters) -
- sizeof(struct lpfc_sli4_cfg_mhdr));
- lpfc_sli4_config(phba, mboxq, LPFC_MBOX_SUBSYSTEM_COMMON,
- LPFC_MBOX_OPCODE_GET_SLI4_PARAMETERS,
- length, LPFC_SLI4_MBX_EMBED);
-
- rc = lpfc_sli_issue_mbox(phba, mboxq, MBX_POLL);
- if (unlikely(rc)) {
- mempool_free(mboxq, phba->mbox_mem_pool);
- return rc;
- }
- /* Gather info on CMF and MI support */
sli4_params = &phba->sli4_hba.pc_sli4_params;
- mbx_sli4_parameters = &mqe->un.get_sli4_parameters.sli4_parameters;
- sli4_params->mi_ver = bf_get(cfg_mi_ver, mbx_sli4_parameters);
- sli4_params->cmf = bf_get(cfg_cmf, mbx_sli4_parameters);
/* Are we forcing MI off via module parameter? */
if (!phba->cfg_enable_mi)
--
2.26.2
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 4/8] lpfc: Fix use-after-free in lpfc_unreg_rpi() routine
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
` (2 preceding siblings ...)
2021-10-20 21:14 ` [PATCH 3/8] lpfc: Correct sysfs reporting of loop support after SFP status change James Smart
@ 2021-10-20 21:14 ` James Smart
2021-10-20 21:14 ` [PATCH 5/8] lpfc: Allow PLOGI retry if previous PLOGI was aborted James Smart
` (5 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: James Smart @ 2021-10-20 21:14 UTC (permalink / raw)
To: linux-scsi; +Cc: James Smart, Justin Tee
An error is detected with the following report when unloading the driver:
"KASAN: use-after-free in lpfc_unreg_rpi+0x1b1b"
The NLP_REG_LOGIN_SEND nlp_flag is set in lpfc_reg_fab_ctrl_node(), but
the flag is not cleared upon completion of the login.
This allows a second call to lpfc_unreg_rpi() to proceed with nlp_rpi
set to LPFC_RPI_ALLOW_ERROR. This results in a use after free access
when used as an rpi_ids array index.
Fix by clearing the NLP_REG_LOGIN_SEND nlp_flag in
lpfc_mbx_cmpl_fc_reg_login().
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
---
drivers/scsi/lpfc/lpfc_hbadisc.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/scsi/lpfc/lpfc_hbadisc.c b/drivers/scsi/lpfc/lpfc_hbadisc.c
index 0b1e1cc00e01..4c068fbb550a 100644
--- a/drivers/scsi/lpfc/lpfc_hbadisc.c
+++ b/drivers/scsi/lpfc/lpfc_hbadisc.c
@@ -4372,6 +4372,7 @@ lpfc_mbx_cmpl_fc_reg_login(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb)
ndlp->nlp_state);
ndlp->nlp_flag |= NLP_RPI_REGISTERED;
+ ndlp->nlp_flag &= ~NLP_REG_LOGIN_SEND;
ndlp->nlp_type |= NLP_FABRIC;
lpfc_nlp_set_state(vport, ndlp, NLP_STE_UNMAPPED_NODE);
--
2.26.2
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 5/8] lpfc: Allow PLOGI retry if previous PLOGI was aborted
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
` (3 preceding siblings ...)
2021-10-20 21:14 ` [PATCH 4/8] lpfc: Fix use-after-free in lpfc_unreg_rpi() routine James Smart
@ 2021-10-20 21:14 ` James Smart
2021-10-20 21:14 ` [PATCH 6/8] lpfc: Fix link down processing to address NULL pointer dereference James Smart
` (4 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: James Smart @ 2021-10-20 21:14 UTC (permalink / raw)
To: linux-scsi; +Cc: James Smart, Justin Tee
A remote nport can stop responding to PLOGI beyond the ELS IO timeout
under some fault conditions. When this happens, the non-response
triggers a dev_loss_tmo event from the transport which causes the driver
to abort the PLOGI and stop any retries. This was due to a policy in the
ELS completion handler whenever an ELS was terminated due to driver
request.
Revise the ELS completion path to detect PLOGI's that were aborted
and allow retries.
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
---
drivers/scsi/lpfc/lpfc_els.c | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c
index de38f4b886ca..746fe9772453 100644
--- a/drivers/scsi/lpfc/lpfc_els.c
+++ b/drivers/scsi/lpfc/lpfc_els.c
@@ -4577,6 +4577,19 @@ lpfc_els_retry(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
retry = 1;
delay = 100;
break;
+ case IOERR_SLI_ABORTED:
+ /* Retry ELS PLOGI command?
+ * Possibly the rport just wasn't ready.
+ */
+ if (cmd == ELS_CMD_PLOGI) {
+ /* No retry if state change */
+ if (ndlp &&
+ ndlp->nlp_state != NLP_STE_PLOGI_ISSUE)
+ goto out_retry;
+ retry = 1;
+ maxretry = 2;
+ }
+ break;
}
break;
--
2.26.2
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 6/8] lpfc: Fix link down processing to address NULL pointer dereference
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
` (4 preceding siblings ...)
2021-10-20 21:14 ` [PATCH 5/8] lpfc: Allow PLOGI retry if previous PLOGI was aborted James Smart
@ 2021-10-20 21:14 ` James Smart
2021-10-20 21:14 ` [PATCH 7/8] lpfc: Allow fabric node recovery if recovery is in progress before devloss James Smart
` (3 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: James Smart @ 2021-10-20 21:14 UTC (permalink / raw)
To: linux-scsi; +Cc: James Smart, Justin Tee
If an FC link down transition while PLOGIs are outstanding to
fabric well known addresses, outstanding ABTS requests may result
in a NULL pointer dereference. Driver unload requests may hang with
repeated "2878" log messages.
The Link down processing results in ABTS requests for outstanding ELS
requests. The Abort WQEs are sent for the ELS's before the driver had
set the link state to down. Thus the driver is sending the Abort with
the expectation that an ABTS will be sent on the wire. The Abort
request is stalled waiting for the link to come up. In some conditions
the driver may auto-complete the ELS's thus if the link does come up,
the Abort completions may reference an invalid structure.
Fix by ensuring that Abort set the flag to avoid link traffic if
issued due to conditions where the link failed.
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
---
drivers/scsi/lpfc/lpfc_sli.c | 14 +++++++-------
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c
index f82f809617a0..5dedb3de271d 100644
--- a/drivers/scsi/lpfc/lpfc_sli.c
+++ b/drivers/scsi/lpfc/lpfc_sli.c
@@ -12403,17 +12403,17 @@ lpfc_sli_issue_abort_iotag(struct lpfc_hba *phba, struct lpfc_sli_ring *pring,
/* ABTS WQE must go to the same WQ as the WQE to be aborted */
abtsiocbp->hba_wqidx = cmdiocb->hba_wqidx;
- if (cmdiocb->iocb_flag & LPFC_IO_FCP) {
- abtsiocbp->iocb_flag |= LPFC_IO_FCP;
- abtsiocbp->iocb_flag |= LPFC_USE_FCPWQIDX;
- }
+ if (cmdiocb->iocb_flag & LPFC_IO_FCP)
+ abtsiocbp->iocb_flag |= (LPFC_IO_FCP | LPFC_USE_FCPWQIDX);
if (cmdiocb->iocb_flag & LPFC_IO_FOF)
abtsiocbp->iocb_flag |= LPFC_IO_FOF;
- if (phba->link_state >= LPFC_LINK_UP)
- iabt->ulpCommand = CMD_ABORT_XRI_CN;
- else
+ if (phba->link_state < LPFC_LINK_UP ||
+ (phba->sli_rev == LPFC_SLI_REV4 &&
+ phba->sli4_hba.link_state.status == LPFC_FC_LA_TYPE_LINK_DOWN))
iabt->ulpCommand = CMD_CLOSE_XRI_CN;
+ else
+ iabt->ulpCommand = CMD_ABORT_XRI_CN;
if (cmpl)
abtsiocbp->iocb_cmpl = cmpl;
--
2.26.2
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 7/8] lpfc: Allow fabric node recovery if recovery is in progress before devloss
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
` (5 preceding siblings ...)
2021-10-20 21:14 ` [PATCH 6/8] lpfc: Fix link down processing to address NULL pointer dereference James Smart
@ 2021-10-20 21:14 ` James Smart
2021-10-20 21:14 ` [PATCH 8/8] lpfc: Update lpfc version to 14.0.0.3 James Smart
` (2 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: James Smart @ 2021-10-20 21:14 UTC (permalink / raw)
To: linux-scsi; +Cc: James Smart, Justin Tee
A link bounce to a slow fabric may observe FDISC response delays lasting
longer than devloss tmo. Current logic decrements the final fabric node
kref during a devloss tmo event. This results in a NULL ptr dereference
crash if the FDISC completes for that fabric node after devloss tmo.
Fix by adding the NLP_IN_RECOV_POST_DEV_LOSS flag, which is set when
devloss tmo triggers and we've noticed that fabric node recovery has
already started or finished in between the time lpfc_dev_loss_tmo_callbk
queues lpfc_dev_loss_tmo_handler. If fabric node recovery succeeds, then
the driver reverses the devloss tmo marked kref put with a kref get. If
fabric node recovery fails, then the final kref put relies
on the ELS timing out or the REG_LOGIN cmpl routine.
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
---
drivers/scsi/lpfc/lpfc_crtn.h | 2 +
drivers/scsi/lpfc/lpfc_disc.h | 12 +++-
drivers/scsi/lpfc/lpfc_els.c | 7 +-
drivers/scsi/lpfc/lpfc_hbadisc.c | 111 ++++++++++++++++++++++++++++++-
drivers/scsi/lpfc/lpfc_init.c | 12 ++--
drivers/scsi/lpfc/lpfc_scsi.c | 10 +--
6 files changed, 139 insertions(+), 15 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_crtn.h b/drivers/scsi/lpfc/lpfc_crtn.h
index f0bcfeecd161..1acff46b4c60 100644
--- a/drivers/scsi/lpfc/lpfc_crtn.h
+++ b/drivers/scsi/lpfc/lpfc_crtn.h
@@ -119,6 +119,8 @@ int lpfc_check_sli_ndlp(struct lpfc_hba *, struct lpfc_sli_ring *,
struct lpfc_nodelist *lpfc_nlp_init(struct lpfc_vport *vport, uint32_t did);
struct lpfc_nodelist *lpfc_nlp_get(struct lpfc_nodelist *);
int lpfc_nlp_put(struct lpfc_nodelist *);
+void lpfc_check_nlp_post_devloss(struct lpfc_vport *vport,
+ struct lpfc_nodelist *ndlp);
void lpfc_ignore_els_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
struct lpfc_iocbq *rspiocb);
int lpfc_nlp_not_used(struct lpfc_nodelist *ndlp);
diff --git a/drivers/scsi/lpfc/lpfc_disc.h b/drivers/scsi/lpfc/lpfc_disc.h
index 871b665bd72e..37a4b79010bf 100644
--- a/drivers/scsi/lpfc/lpfc_disc.h
+++ b/drivers/scsi/lpfc/lpfc_disc.h
@@ -85,6 +85,13 @@ enum lpfc_fc4_xpt_flags {
NLP_XPT_HAS_HH = 0x10
};
+enum lpfc_nlp_save_flags {
+ /* devloss occurred during recovery */
+ NLP_IN_RECOV_POST_DEV_LOSS = 0x1,
+ /* wait for outstanding LOGO to cmpl */
+ NLP_WAIT_FOR_LOGO = 0x2,
+};
+
struct lpfc_nodelist {
struct list_head nlp_listp;
struct serv_parm fc_sparam; /* buffer for service params */
@@ -144,8 +151,9 @@ struct lpfc_nodelist {
unsigned long *active_rrqs_xri_bitmap;
struct lpfc_scsicmd_bkt *lat_data; /* Latency data */
uint32_t fc4_prli_sent;
- u32 upcall_flags;
-#define NLP_WAIT_FOR_LOGO 0x2
+
+ /* flags to keep ndlp alive until special conditions are met */
+ enum lpfc_nlp_save_flags save_flags;
enum lpfc_fc4_xpt_flags fc4_xpt_flags;
diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c
index 746fe9772453..b940e0268f96 100644
--- a/drivers/scsi/lpfc/lpfc_els.c
+++ b/drivers/scsi/lpfc/lpfc_els.c
@@ -2905,9 +2905,9 @@ lpfc_cmpl_els_logo(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
irsp = &(rspiocb->iocb);
spin_lock_irq(&ndlp->lock);
ndlp->nlp_flag &= ~NLP_LOGO_SND;
- if (ndlp->upcall_flags & NLP_WAIT_FOR_LOGO) {
+ if (ndlp->save_flags & NLP_WAIT_FOR_LOGO) {
wake_up_waiter = 1;
- ndlp->upcall_flags &= ~NLP_WAIT_FOR_LOGO;
+ ndlp->save_flags &= ~NLP_WAIT_FOR_LOGO;
}
spin_unlock_irq(&ndlp->lock);
@@ -10735,6 +10735,9 @@ lpfc_cmpl_els_fdisc(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
irsp->ulpStatus, irsp->un.ulpWord[4]);
goto fdisc_failed;
}
+
+ lpfc_check_nlp_post_devloss(vport, ndlp);
+
spin_lock_irq(shost->host_lock);
vport->fc_flag &= ~FC_VPORT_CVL_RCVD;
vport->fc_flag &= ~FC_VPORT_LOGO_RCVD;
diff --git a/drivers/scsi/lpfc/lpfc_hbadisc.c b/drivers/scsi/lpfc/lpfc_hbadisc.c
index 4c068fbb550a..9fe6e5b386ce 100644
--- a/drivers/scsi/lpfc/lpfc_hbadisc.c
+++ b/drivers/scsi/lpfc/lpfc_hbadisc.c
@@ -209,7 +209,12 @@ lpfc_dev_loss_tmo_callbk(struct fc_rport *rport)
spin_lock_irqsave(&ndlp->lock, iflags);
ndlp->nlp_flag |= NLP_IN_DEV_LOSS;
- ndlp->nlp_flag &= ~NLP_NPR_2B_DISC;
+
+ /* If there is a PLOGI in progress, and we are in a
+ * NLP_NPR_2B_DISC state, don't turn off the flag.
+ */
+ if (ndlp->nlp_state != NLP_STE_PLOGI_ISSUE)
+ ndlp->nlp_flag &= ~NLP_NPR_2B_DISC;
/*
* The backend does not expect any more calls associated with this
@@ -340,6 +345,37 @@ static void lpfc_check_inactive_vmid(struct lpfc_hba *phba)
lpfc_destroy_vport_work_array(phba, vports);
}
+/**
+ * lpfc_check_nlp_post_devloss - Check to restore ndlp refcnt after devloss
+ * @vport: Pointer to vport object.
+ * @ndlp: Pointer to remote node object.
+ *
+ * If NLP_IN_RECOV_POST_DEV_LOSS flag was set due to outstanding recovery of
+ * node during dev_loss_tmo processing, then this function restores the nlp_put
+ * kref decrement from lpfc_dev_loss_tmo_handler.
+ **/
+void
+lpfc_check_nlp_post_devloss(struct lpfc_vport *vport,
+ struct lpfc_nodelist *ndlp)
+{
+ unsigned long iflags;
+
+ spin_lock_irqsave(&ndlp->lock, iflags);
+ if (ndlp->save_flags & NLP_IN_RECOV_POST_DEV_LOSS) {
+ ndlp->save_flags &= ~NLP_IN_RECOV_POST_DEV_LOSS;
+ spin_unlock_irqrestore(&ndlp->lock, iflags);
+ lpfc_nlp_get(ndlp);
+ lpfc_printf_vlog(vport, KERN_INFO, LOG_DISCOVERY | LOG_NODE,
+ "8438 Devloss timeout reversed on DID x%x "
+ "refcnt %d ndlp %p flag x%x "
+ "port_state = x%x\n",
+ ndlp->nlp_DID, kref_read(&ndlp->kref), ndlp,
+ ndlp->nlp_flag, vport->port_state);
+ spin_lock_irqsave(&ndlp->lock, iflags);
+ }
+ spin_unlock_irqrestore(&ndlp->lock, iflags);
+}
+
/**
* lpfc_dev_loss_tmo_handler - Remote node devloss timeout handler
* @ndlp: Pointer to remote node object.
@@ -358,6 +394,8 @@ lpfc_dev_loss_tmo_handler(struct lpfc_nodelist *ndlp)
uint8_t *name;
int warn_on = 0;
int fcf_inuse = 0;
+ bool recovering = false;
+ struct fc_vport *fc_vport = NULL;
unsigned long iflags;
vport = ndlp->vport;
@@ -394,6 +432,64 @@ lpfc_dev_loss_tmo_handler(struct lpfc_nodelist *ndlp)
/* Fabric nodes are done. */
if (ndlp->nlp_type & NLP_FABRIC) {
+ spin_lock_irqsave(&ndlp->lock, iflags);
+ /* In massive vport configuration settings, it's possible
+ * dev_loss_tmo fired during node recovery. So, check if
+ * fabric nodes are in discovery states outstanding.
+ */
+ switch (ndlp->nlp_DID) {
+ case Fabric_DID:
+ fc_vport = vport->fc_vport;
+ if (fc_vport &&
+ fc_vport->vport_state == FC_VPORT_INITIALIZING)
+ recovering = true;
+ break;
+ case Fabric_Cntl_DID:
+ if (ndlp->nlp_flag & NLP_REG_LOGIN_SEND)
+ recovering = true;
+ break;
+ case FDMI_DID:
+ fallthrough;
+ case NameServer_DID:
+ if (ndlp->nlp_state >= NLP_STE_PLOGI_ISSUE &&
+ ndlp->nlp_state <= NLP_STE_REG_LOGIN_ISSUE)
+ recovering = true;
+ break;
+ }
+ spin_unlock_irqrestore(&ndlp->lock, iflags);
+
+ /* Mark an NLP_IN_RECOV_POST_DEV_LOSS flag to know if reversing
+ * the following lpfc_nlp_put is necessary after fabric node is
+ * recovered.
+ */
+ if (recovering) {
+ lpfc_printf_vlog(vport, KERN_INFO,
+ LOG_DISCOVERY | LOG_NODE,
+ "8436 Devloss timeout marked on "
+ "DID x%x refcnt %d ndlp %p "
+ "flag x%x port_state = x%x\n",
+ ndlp->nlp_DID, kref_read(&ndlp->kref),
+ ndlp, ndlp->nlp_flag,
+ vport->port_state);
+ spin_lock_irqsave(&ndlp->lock, iflags);
+ ndlp->save_flags |= NLP_IN_RECOV_POST_DEV_LOSS;
+ spin_unlock_irqrestore(&ndlp->lock, iflags);
+ } else if (ndlp->nlp_state == NLP_STE_UNMAPPED_NODE) {
+ /* Fabric node fully recovered before this dev_loss_tmo
+ * queue work is processed. Thus, ignore the
+ * dev_loss_tmo event.
+ */
+ lpfc_printf_vlog(vport, KERN_INFO,
+ LOG_DISCOVERY | LOG_NODE,
+ "8437 Devloss timeout ignored on "
+ "DID x%x refcnt %d ndlp %p "
+ "flag x%x port_state = x%x\n",
+ ndlp->nlp_DID, kref_read(&ndlp->kref),
+ ndlp, ndlp->nlp_flag,
+ vport->port_state);
+ return fcf_inuse;
+ }
+
lpfc_nlp_put(ndlp);
return fcf_inuse;
}
@@ -423,6 +519,14 @@ lpfc_dev_loss_tmo_handler(struct lpfc_nodelist *ndlp)
ndlp->nlp_state, ndlp->nlp_rpi);
}
+ /* If we are devloss, but we are in the process of rediscovering the
+ * ndlp, don't issue a NLP_EVT_DEVICE_RM event.
+ */
+ if (ndlp->nlp_state >= NLP_STE_PLOGI_ISSUE &&
+ ndlp->nlp_state <= NLP_STE_PRLI_ISSUE) {
+ return fcf_inuse;
+ }
+
if (!(ndlp->fc4_xpt_flags & NVME_XPT_REGD))
lpfc_disc_state_machine(vport, ndlp, NULL, NLP_EVT_DEVICE_RM);
@@ -4363,6 +4467,8 @@ lpfc_mbx_cmpl_fc_reg_login(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb)
goto out;
}
+ lpfc_check_nlp_post_devloss(vport, ndlp);
+
if (phba->sli_rev < LPFC_SLI_REV4)
ndlp->nlp_rpi = mb->un.varWords[0];
@@ -4540,9 +4646,10 @@ lpfc_nlp_counters(struct lpfc_vport *vport, int state, int count)
void
lpfc_nlp_reg_node(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp)
{
-
unsigned long iflags;
+ lpfc_check_nlp_post_devloss(vport, ndlp);
+
spin_lock_irqsave(&ndlp->lock, iflags);
if (ndlp->fc4_xpt_flags & NLP_XPT_REGD) {
/* Already registered with backend, trigger rescan */
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index 744da35b5074..df3edbe2b2e5 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -3753,12 +3753,16 @@ lpfc_offline_prep(struct lpfc_hba *phba, int mbx_action)
lpfc_disc_state_machine(vports[i], ndlp,
NULL, NLP_EVT_DEVICE_RECOVERY);
- /* Don't remove the node unless the
+ /* Don't remove the node unless the node
* has been unregistered with the
- * transport. If so, let dev_loss
- * take care of the node.
+ * transport, and we're not in recovery
+ * before dev_loss_tmo triggered.
+ * Otherwise, let dev_loss take care of
+ * the node.
*/
- if (!(ndlp->fc4_xpt_flags &
+ if (!(ndlp->save_flags &
+ NLP_IN_RECOV_POST_DEV_LOSS) &&
+ !(ndlp->fc4_xpt_flags &
(NVME_XPT_REGD | SCSI_XPT_REGD)))
lpfc_disc_state_machine
(vports[i], ndlp,
diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c
index 0dce4b51ca1e..fe4e5f6ecc70 100644
--- a/drivers/scsi/lpfc/lpfc_scsi.c
+++ b/drivers/scsi/lpfc/lpfc_scsi.c
@@ -6475,28 +6475,28 @@ lpfc_target_reset_handler(struct scsi_cmnd *cmnd)
/* Issue LOGO, if no LOGO is outstanding */
spin_lock_irqsave(&pnode->lock, flags);
- if (!(pnode->upcall_flags & NLP_WAIT_FOR_LOGO) &&
+ if (!(pnode->save_flags & NLP_WAIT_FOR_LOGO) &&
!pnode->logo_waitq) {
pnode->logo_waitq = &waitq;
pnode->nlp_fcp_info &= ~NLP_FCP_2_DEVICE;
pnode->nlp_flag |= NLP_ISSUE_LOGO;
- pnode->upcall_flags |= NLP_WAIT_FOR_LOGO;
+ pnode->save_flags |= NLP_WAIT_FOR_LOGO;
spin_unlock_irqrestore(&pnode->lock, flags);
lpfc_unreg_rpi(vport, pnode);
wait_event_timeout(waitq,
- (!(pnode->upcall_flags &
+ (!(pnode->save_flags &
NLP_WAIT_FOR_LOGO)),
msecs_to_jiffies(dev_loss_tmo *
1000));
- if (pnode->upcall_flags & NLP_WAIT_FOR_LOGO) {
+ if (pnode->save_flags & NLP_WAIT_FOR_LOGO) {
lpfc_printf_vlog(vport, KERN_ERR, logit,
"0725 SCSI layer TGTRST "
"failed & LOGO TMO (%d, %llu) "
"return x%x\n",
tgt_id, lun_id, status);
spin_lock_irqsave(&pnode->lock, flags);
- pnode->upcall_flags &= ~NLP_WAIT_FOR_LOGO;
+ pnode->save_flags &= ~NLP_WAIT_FOR_LOGO;
} else {
spin_lock_irqsave(&pnode->lock, flags);
}
--
2.26.2
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 8/8] lpfc: Update lpfc version to 14.0.0.3
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
` (6 preceding siblings ...)
2021-10-20 21:14 ` [PATCH 7/8] lpfc: Allow fabric node recovery if recovery is in progress before devloss James Smart
@ 2021-10-20 21:14 ` James Smart
2021-10-21 3:34 ` [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 Martin K. Petersen
2021-10-27 4:00 ` Martin K. Petersen
9 siblings, 0 replies; 11+ messages in thread
From: James Smart @ 2021-10-20 21:14 UTC (permalink / raw)
To: linux-scsi; +Cc: James Smart, Justin Tee
Update lpfc version to 14.0.0.3
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
---
drivers/scsi/lpfc/lpfc_version.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/scsi/lpfc/lpfc_version.h b/drivers/scsi/lpfc/lpfc_version.h
index dec71775f677..5a4d3b24fbce 100644
--- a/drivers/scsi/lpfc/lpfc_version.h
+++ b/drivers/scsi/lpfc/lpfc_version.h
@@ -20,7 +20,7 @@
* included with this package. *
*******************************************************************/
-#define LPFC_DRIVER_VERSION "14.0.0.2"
+#define LPFC_DRIVER_VERSION "14.0.0.3"
#define LPFC_DRIVER_NAME "lpfc"
/* Used for SLI 2/3 */
--
2.26.2
^ permalink raw reply related [flat|nested] 11+ messages in thread* Re: [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
` (7 preceding siblings ...)
2021-10-20 21:14 ` [PATCH 8/8] lpfc: Update lpfc version to 14.0.0.3 James Smart
@ 2021-10-21 3:34 ` Martin K. Petersen
2021-10-27 4:00 ` Martin K. Petersen
9 siblings, 0 replies; 11+ messages in thread
From: Martin K. Petersen @ 2021-10-21 3:34 UTC (permalink / raw)
To: James Smart; +Cc: linux-scsi
James,
> Update lpfc to revision 14.0.0.3
Applied to 5.16/scsi-staging, thanks!
--
Martin K. Petersen Oracle Linux Engineering
^ permalink raw reply [flat|nested] 11+ messages in thread* Re: [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3
2021-10-20 21:14 [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 James Smart
` (8 preceding siblings ...)
2021-10-21 3:34 ` [PATCH 0/8] lpfc: Update lpfc to revision 14.0.0.3 Martin K. Petersen
@ 2021-10-27 4:00 ` Martin K. Petersen
9 siblings, 0 replies; 11+ messages in thread
From: Martin K. Petersen @ 2021-10-27 4:00 UTC (permalink / raw)
To: linux-scsi, James Smart; +Cc: Martin K . Petersen
On Wed, 20 Oct 2021 14:14:09 -0700, James Smart wrote:
> Update lpfc to revision 14.0.0.3
>
> This patch set contains several bug fixes.
>
> The patches were cut against Martin's 5.16/scsi-queue tree
>
>
> [...]
Applied to 5.16/scsi-queue, thanks!
[1/8] lpfc: Revert LOG_TRACE_EVENT back to LOG_INIT prior to driver_resource_setup
https://git.kernel.org/mkp/scsi/c/a516074c2026
[2/8] lpfc: Wait for successful restart of SLI3 adapter during host sg_reset
https://git.kernel.org/mkp/scsi/c/d305c253af69
[3/8] lpfc: Correct sysfs reporting of loop support after SFP status change
https://git.kernel.org/mkp/scsi/c/7a1dda943630
[4/8] lpfc: Fix use-after-free in lpfc_unreg_rpi() routine
https://git.kernel.org/mkp/scsi/c/79b20beccea3
[5/8] lpfc: Allow PLOGI retry if previous PLOGI was aborted
https://git.kernel.org/mkp/scsi/c/15af02d8a585
[6/8] lpfc: Fix link down processing to address NULL pointer dereference
https://git.kernel.org/mkp/scsi/c/1854f53ccd88
[7/8] lpfc: Allow fabric node recovery if recovery is in progress before devloss
https://git.kernel.org/mkp/scsi/c/af984c87293b
[8/8] lpfc: Update lpfc version to 14.0.0.3
https://git.kernel.org/mkp/scsi/c/83c3a7beaef7
--
Martin K. Petersen Oracle Linux Engineering
^ permalink raw reply [flat|nested] 11+ messages in thread