diff mbox

[21/43] qla2xxx: Remove calling cancel_work_sync()

Message ID 20171220065644.21511-22-himanshu.madhani@cavium.com (mailing list archive)
State Changes Requested
Headers show

Commit Message

Madhani, Himanshu Dec. 20, 2017, 6:56 a.m. UTC 
From: Sawan Chandak <sawan.chandak@cavium.com>

This is blocking call and issue is seen, when called in interrupt context.

__cancel_work_timer+0x140/0x210
? ql_dbg+0xcb/0x110 [qla2xxx]
cancel_work_sync+0x10/0x20
qlt_schedule_sess_for_deletion+0x89/0x170 [qla2xxx]
qla24xx_handle_gpnid_event+0x146/0x4a0 [qla2xxx]
qla2x00_fcport_event_handler+0x14b/0x290 [qla2xxx]
qla2x00_async_gpnid_sp_done+0x118/0x240 [qla2xxx]
qla24xx_els_ct_entry.isra.15+0x1df/0x2b0 [qla2xxx]
? qla24xx_msix_rsp_q+0x39/0xf0 [qla2xxx]
qla24xx_process_response_queue+0xae/0x270 [qla2xxx]
qla24xx_msix_rsp_q+0x8a/0xf0 [qla2xxx]
__handle_irq_event_percpu+0x3c/0x350
handle_irq_event_percpu+0x32/0x80
handle_irq_event+0x39/0x60
handle_edge_irq+0x8c/0x140
handle_irq+0xab/0x130
? _local_bh_enable+0x21/0x50
do_IRQ+0x5e/0x120
common_interrupt+0x9d/0x9d
RIP: 0010:cpuidle_enter_state+0xe9/0x320

Signed-off-by: Sawan Chandak <sawan.chandak@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
---
 drivers/scsi/qla2xxx/qla_target.c | 2 --
 1 file changed, 2 deletions(-)

Comments

Bart Van Assche Dec. 20, 2017, 4:56 p.m. UTC | #1 
On Tue, 2017-12-19 at 22:56 -0800, Himanshu Madhani wrote:
> @@ -1234,8 +1234,6 @@ void qlt_schedule_sess_for_deletion(struct fc_port *sess,

>  	ql_dbg(ql_dbg_tgt, sess->vha, 0xe001,

>  	    "Scheduling sess %p for deletion\n", sess);

>  

> -	/* use cancel to push work element through before re-queue */

> -	cancel_work_sync(&sess->del_work);

>  	INIT_WORK(&sess->del_work, qla24xx_delete_sess_fn);

>  	queue_work(sess->vha->hw->wq, &sess->del_work);


Why had this cancel_work_sync() call been introduced in
qlt_schedule_sess_for_deletion() and why do you think it is safe to remove
this call? Both questions should have been answered in the patch description.

Thanks,

Bart.
Madhani, Himanshu Dec. 20, 2017, 9:41 p.m. UTC | #2 
> On Dec 20, 2017, at 10:56 AM, Bart Van Assche <bart.vanassche@wdc.com> wrote:
> 
> On Tue, 2017-12-19 at 22:56 -0800, Himanshu Madhani wrote:
>> @@ -1234,8 +1234,6 @@ void qlt_schedule_sess_for_deletion(struct fc_port *sess,
>> 	ql_dbg(ql_dbg_tgt, sess->vha, 0xe001,
>> 	    "Scheduling sess %p for deletion\n", sess);
>> 
>> -	/* use cancel to push work element through before re-queue */
>> -	cancel_work_sync(&sess->del_work);
>> 	INIT_WORK(&sess->del_work, qla24xx_delete_sess_fn);
>> 	queue_work(sess->vha->hw->wq, &sess->del_work);
> 
> Why had this cancel_work_sync() call been introduced in
> qlt_schedule_sess_for_deletion() and why do you think it is safe to remove
> this call? Both questions should have been answered in the patch description.
> 
> Thanks,
> 
> Bart.

I will drop this patch from current series and will update commit message with details
as suggested and submit in next series.

Thanks,
- Himanshu
diff mbox

Patch

diff --git a/drivers/scsi/qla2xxx/qla_target.c b/drivers/scsi/qla2xxx/qla_target.c
index 71be1a95ba86..002fe05dd344 100644
--- a/drivers/scsi/qla2xxx/qla_target.c
+++ b/drivers/scsi/qla2xxx/qla_target.c
@@ -1234,8 +1234,6 @@  void qlt_schedule_sess_for_deletion(struct fc_port *sess,
 	ql_dbg(ql_dbg_tgt, sess->vha, 0xe001,
 	    "Scheduling sess %p for deletion\n", sess);
 
-	/* use cancel to push work element through before re-queue */
-	cancel_work_sync(&sess->del_work);
 	INIT_WORK(&sess->del_work, qla24xx_delete_sess_fn);
 	queue_work(sess->vha->hw->wq, &sess->del_work);
 }