scsi: lpfc: Fix missing wakeups on abort threads
authorJames Smart <jsmart2021@gmail.com>
Wed, 3 Apr 2019 18:10:34 +0000 (11:10 -0700)
committerMartin K. Petersen <martin.petersen@oracle.com>
Thu, 4 Apr 2019 03:35:09 +0000 (23:35 -0400)
Abort thread wakeups, on some wqe types, are not happening.  The thread
wakeup logic is dependent upon the LPFC_DRIVER_ABORTED flag. However, on
these wqes, the completion handler running prior to the io completion
routine ends up clearing the flag.

Rework the wakeup logic to look at a non-null waitq element which must be
set if the abort thread is waiting. This is reverting the change in the
indicated patch.

Fixes: c2017260eea2d ("scsi: lpfc: Rework locking on SCSI io completion")
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
drivers/scsi/lpfc/lpfc_scsi.c

index c98f264f1d83a030ea8a00678fd586ab70059218..a497b2c0cb798e07d240affc72b4621fa1503f34 100644 (file)
@@ -3878,10 +3878,9 @@ lpfc_scsi_cmd_iocb_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *pIocbIn,
         * wake up the thread.
         */
        spin_lock(&lpfc_cmd->buf_lock);
-       if (unlikely(lpfc_cmd->cur_iocbq.iocb_flag & LPFC_DRIVER_ABORTED)) {
-               lpfc_cmd->cur_iocbq.iocb_flag &= ~LPFC_DRIVER_ABORTED;
-               if (lpfc_cmd->waitq)
-                       wake_up(lpfc_cmd->waitq);
+       lpfc_cmd->cur_iocbq.iocb_flag &= ~LPFC_DRIVER_ABORTED;
+       if (lpfc_cmd->waitq) {
+               wake_up(lpfc_cmd->waitq);
                lpfc_cmd->waitq = NULL;
        }
        spin_unlock(&lpfc_cmd->buf_lock);