Replace deprecated create_singlethread_workqueue with
alloc_ordered_workqueue.
Work items include getting tx/rx frame sizes, resetting MPI processor,
setting asic recovery bit so ordering seems necessary as only one work
item should be in queue/executing at any given time, hence the use of
alloc_ordered_workqueue.
WQ_MEM_RECLAIM flag has been set since ethernet devices seem to sit in
memory reclaim path, so to guarantee forward progress regardless of
memory pressure.
Signed-off-by: Amitoj Kaur Chawla <[email protected]>
Acked-by: Tejun Heo <[email protected]>
---
Not sure if the assumption of requiring ordering of work items
is correct.
drivers/net/ethernet/qlogic/qlge/qlge_main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qlogic/qlge/qlge_main.c b/drivers/net/ethernet/qlogic/qlge/qlge_main.c
index b28e73e..83d7210 100644
--- a/drivers/net/ethernet/qlogic/qlge/qlge_main.c
+++ b/drivers/net/ethernet/qlogic/qlge/qlge_main.c
@@ -4687,7 +4687,7 @@ static int ql_init_device(struct pci_dev *pdev, struct net_device *ndev,
/*
* Set up the operating parameters.
*/
- qdev->workqueue = create_singlethread_workqueue(ndev->name);
+ qdev->workqueue = alloc_ordered_workqueue(ndev->name, WQ_MEM_RECLAIM);
INIT_DELAYED_WORK(&qdev->asic_reset_work, ql_asic_reset_work);
INIT_DELAYED_WORK(&qdev->mpi_reset_work, ql_mpi_reset_work);
INIT_DELAYED_WORK(&qdev->mpi_work, ql_mpi_work);
--
1.9.1
On Sat, Apr 09, 2016 at 05:27:45PM +0530, Amitoj Kaur Chawla wrote:
> Replace deprecated create_singlethread_workqueue with
> alloc_ordered_workqueue.
>
> Work items include getting tx/rx frame sizes, resetting MPI processor,
> setting asic recovery bit so ordering seems necessary as only one work
> item should be in queue/executing at any given time, hence the use of
> alloc_ordered_workqueue.
>
> WQ_MEM_RECLAIM flag has been set since ethernet devices seem to sit in
> memory reclaim path, so to guarantee forward progress regardless of
> memory pressure.
>
> Signed-off-by: Amitoj Kaur Chawla <[email protected]>
> Acked-by: Tejun Heo <[email protected]>
Ping?
--
tejun
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf
> Of Tejun Heo
> Sent: Wednesday, April 13, 2016 11:14 PM
> To: Amitoj Kaur Chawla <[email protected]>
> Cc: Sudarsana Kalluru <[email protected]>; netdev
> <[email protected]>; linux-kernel <[email protected]>; Dept-
> Eng Linux Driver <[email protected]>; Harish Patil
> <[email protected]>; Dept-GE Linux NIC Dev <Dept-
> [email protected]>
> Subject: Re: [PATCH] qlge: Replace create_singlethread_workqueue with
> alloc_ordered_workqueue
>
> On Sat, Apr 09, 2016 at 05:27:45PM +0530, Amitoj Kaur Chawla wrote:
> > Replace deprecated create_singlethread_workqueue with
> > alloc_ordered_workqueue.
> >
> > Work items include getting tx/rx frame sizes, resetting MPI processor,
> > setting asic recovery bit so ordering seems necessary as only one work
> > item should be in queue/executing at any given time, hence the use of
> > alloc_ordered_workqueue.
> >
> > WQ_MEM_RECLAIM flag has been set since ethernet devices seem to sit in
> > memory reclaim path, so to guarantee forward progress regardless of
> > memory pressure.
> >
> > Signed-off-by: Amitoj Kaur Chawla <[email protected]>
> > Acked-by: Tejun Heo <[email protected]>
>
> Ping?
>
Hi,
Just want to confirm that __WQ_LEGACY flag is not necessary here as this is removed
with this change ?
Thanks
Hello, Manish.
On Thu, Apr 14, 2016 at 07:25:15AM +0000, Manish Chopra wrote:
> Just want to confirm that __WQ_LEGACY flag is not necessary here as this is removed
> with this change ?
Yeah, that should be fine. That only affects locking dependency
tracking which can fire spuriously due to workqueues created with the
old interface having WQ_MEM_RECLAIM unconditionally. In this case, we
actually want WQ_MEM_RECLAIM and thus we want the dependency tracking
too.
Thanks.
--
tejun
From: Amitoj Kaur Chawla <[email protected]>
Date: Sat, 9 Apr 2016 17:27:45 +0530
> Replace deprecated create_singlethread_workqueue with
> alloc_ordered_workqueue.
>
> Work items include getting tx/rx frame sizes, resetting MPI processor,
> setting asic recovery bit so ordering seems necessary as only one work
> item should be in queue/executing at any given time, hence the use of
> alloc_ordered_workqueue.
>
> WQ_MEM_RECLAIM flag has been set since ethernet devices seem to sit in
> memory reclaim path, so to guarantee forward progress regardless of
> memory pressure.
>
> Signed-off-by: Amitoj Kaur Chawla <[email protected]>
> Acked-by: Tejun Heo <[email protected]>
I'll apply this to net-next, thanks.