In lpfc_abort_handler, the lock acquire order is
hbalock (irqsave), buf_lock (irq) and ring_lock (irq).
The issue is that in two places the locks are released
out of order - the buf_lock and the hbalock - resulting
in the cpu preemption/lock flags getting restored out of
order and deadlocking the cpu.
Fix the unlock order by fully releasing the hbalocks as well.
CC: Zhangguanghui <zhang.guanghui@xxxxxxx>
Signed-off-by: Dick Kennedy <dick.kennedy@xxxxxxxxxxxx>
Signed-off-by: James Smart <jsmart2021@xxxxxxxxx>
---
Original patch sent by Zhangguanghui <zhang.guanghui@xxxxxxx>
Modified slightly as merged and tested at Emulex.
---
drivers/scsi/lpfc/lpfc_scsi.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c
index 13e3e14b43f9..e4ec2b99b583 100644
--- a/drivers/scsi/lpfc/lpfc_scsi.c
+++ b/drivers/scsi/lpfc/lpfc_scsi.c
@@ -4848,20 +4848,21 @@ lpfc_abort_handler(struct scsi_cmnd *cmnd)
ret_val = __lpfc_sli_issue_iocb(phba, LPFC_FCP_RING,
abtsiocb, 0);
}
- /* no longer need the lock after this point */
- spin_unlock_irqrestore(&phba->hbalock, flags);
if (ret_val == IOCB_ERROR) {
/* Indicate the IO is not being aborted by the driver. */
iocb->iocb_flag &= ~LPFC_DRIVER_ABORTED;
lpfc_cmd->waitq = NULL;
spin_unlock(&lpfc_cmd->buf_lock);
+ spin_unlock_irqrestore(&phba->hbalock, flags);
lpfc_sli_release_iocbq(phba, abtsiocb);
ret = FAILED;
goto out;
}
+ /* no longer need the lock after this point */
spin_unlock(&lpfc_cmd->buf_lock);
+ spin_unlock_irqrestore(&phba->hbalock, flags);
if (phba->cfg_poll & DISABLE_FCP_RING_INT)
lpfc_sli_handle_fast_ring_event(phba,
--
2.13.7
