scsi: lpfc: Fix SCSI I/O completion and abort handler deadlock

[ Upstream commit 03cbbd7c2f5ee288f648f4aeedc765a181188553 ]

During stress I/O tests with 500+ vports, hard LOCKUP call traces are
observed.

CPU A:
 native_queued_spin_lock_slowpath+0x192
 _raw_spin_lock_irqsave+0x32
 lpfc_handle_fcp_err+0x4c6
 lpfc_fcp_io_cmd_wqe_cmpl+0x964
 lpfc_sli4_fp_handle_cqe+0x266
 __lpfc_sli4_process_cq+0x105
 __lpfc_sli4_hba_process_cq+0x3c
 lpfc_cq_poll_hdler+0x16
 irq_poll_softirq+0x76
 __softirqentry_text_start+0xe4
 irq_exit+0xf7
 do_IRQ+0x7f

CPU B:
 native_queued_spin_lock_slowpath+0x5b
 _raw_spin_lock+0x1c
 lpfc_abort_handler+0x13e
 scmd_eh_abort_handler+0x85
 process_one_work+0x1a7
 worker_thread+0x30
 kthread+0x112
 ret_from_fork+0x1f

Diagram of lockup:

CPUA                            CPUB
----                            ----
lpfc_cmd->buf_lock
                            phba->hbalock
                            lpfc_cmd->buf_lock
phba->hbalock

Fix by reordering the taking of the lpfc_cmd->buf_lock and phba->hbalock in
lpfc_abort_handler routine so that it tries to take the lpfc_cmd->buf_lock
first before phba->hbalock.

Link: https://lore.kernel.org/r/20220412222008.126521-7-jsmart2021@gmail.com
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>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c
index 3d9175f1b6787823075d5ad1dbfb2d518b31370c..8c79264a935b4f6ec0ec20d060592189da070abf 100644 (file)
--- a/drivers/scsi/lpfc/lpfc_scsi.c
+++ b/drivers/scsi/lpfc/lpfc_scsi.c
@@ -5885,25 +5885,25 @@ lpfc_abort_handler(struct scsi_cmnd *cmnd)
        if (!lpfc_cmd)
                return ret;
 
-       spin_lock_irqsave(&phba->hbalock, flags);
+       /* Guard against IO completion being called at same time */
+       spin_lock_irqsave(&lpfc_cmd->buf_lock, flags);
+
+       spin_lock(&phba->hbalock);
        /* driver queued commands are in process of being flushed */
        if (phba->hba_flag & HBA_IOQ_FLUSH) {
                lpfc_printf_vlog(vport, KERN_WARNING, LOG_FCP,
                        "3168 SCSI Layer abort requested I/O has been "
                        "flushed by LLD.\n");
                ret = FAILED;
-               goto out_unlock;
+               goto out_unlock_hba;
        }
 
-       /* Guard against IO completion being called at same time */
-       spin_lock(&lpfc_cmd->buf_lock);
-
        if (!lpfc_cmd->pCmd) {
                lpfc_printf_vlog(vport, KERN_WARNING, LOG_FCP,
                         "2873 SCSI Layer I/O Abort Request IO CMPL Status "
                         "x%x ID %d LUN %llu\n",
                         SUCCESS, cmnd->device->id, cmnd->device->lun);
-               goto out_unlock_buf;
+               goto out_unlock_hba;
        }
 
        iocb = &lpfc_cmd->cur_iocbq;
@@ -5911,7 +5911,7 @@ lpfc_abort_handler(struct scsi_cmnd *cmnd)
                pring_s4 = phba->sli4_hba.hdwq[iocb->hba_wqidx].io_wq->pring;
                if (!pring_s4) {
                        ret = FAILED;
-                       goto out_unlock_buf;
+                       goto out_unlock_hba;
                }
                spin_lock(&pring_s4->ring_lock);
        }
@@ -5944,8 +5944,8 @@ lpfc_abort_handler(struct scsi_cmnd *cmnd)
                         "3389 SCSI Layer I/O Abort Request is pending\n");
                if (phba->sli_rev == LPFC_SLI_REV4)
                        spin_unlock(&pring_s4->ring_lock);
-               spin_unlock(&lpfc_cmd->buf_lock);
-               spin_unlock_irqrestore(&phba->hbalock, flags);
+               spin_unlock(&phba->hbalock);
+               spin_unlock_irqrestore(&lpfc_cmd->buf_lock, flags);
                goto wait_for_cmpl;
        }
 
@@ -5966,15 +5966,13 @@ lpfc_abort_handler(struct scsi_cmnd *cmnd)
        if (ret_val != IOCB_SUCCESS) {
                /* Indicate the IO is not being aborted by the driver. */
                lpfc_cmd->waitq = NULL;
-               spin_unlock(&lpfc_cmd->buf_lock);
-               spin_unlock_irqrestore(&phba->hbalock, flags);
                ret = FAILED;
-               goto out;
+               goto out_unlock_hba;
        }
 
        /* no longer need the lock after this point */
-       spin_unlock(&lpfc_cmd->buf_lock);
-       spin_unlock_irqrestore(&phba->hbalock, flags);
+       spin_unlock(&phba->hbalock);
+       spin_unlock_irqrestore(&lpfc_cmd->buf_lock, flags);
 
        if (phba->cfg_poll & DISABLE_FCP_RING_INT)
                lpfc_sli_handle_fast_ring_event(phba,
@@ -6009,10 +6007,9 @@ wait_for_cmpl:
 out_unlock_ring:
        if (phba->sli_rev == LPFC_SLI_REV4)
                spin_unlock(&pring_s4->ring_lock);
-out_unlock_buf:
-       spin_unlock(&lpfc_cmd->buf_lock);
-out_unlock:
-       spin_unlock_irqrestore(&phba->hbalock, flags);
+out_unlock_hba:
+       spin_unlock(&phba->hbalock);
+       spin_unlock_irqrestore(&lpfc_cmd->buf_lock, flags);
 out:
        lpfc_printf_vlog(vport, KERN_WARNING, LOG_FCP,
                         "0749 SCSI Layer I/O Abort Request Status x%x ID %d "