[PATCH 5.10.y 0/2] Fix incorrect backport of nvme-fc ioerr_work cancel_work_sync()

Jaskaran Singh posted 2 patches 1 month, 3 weeks ago
There is a newer version of this series
drivers/nvme/host/fc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
[PATCH 5.10.y 0/2] Fix incorrect backport of nvme-fc ioerr_work cancel_work_sync()
Posted by Jaskaran Singh 1 month, 3 weeks ago
The backport of upstream commit 0a2c5495b6d1 ("nvme: nvme-fc: Ensure
->ioerr_work is cancelled in nvme_fc_delete_ctrl()") to linux-5.10.y
was incorrectly applied as commit 3d78e8e01251.

The original upstream fix moves the cancel_work_sync(&ctrl->ioerr_work)
call within nvme_fc_delete_ctrl() to after nvme_fc_delete_association(),
so that ->ioerr_work is not running when the nvme_fc_ctrl object is
freed. However, the stable backport mistakenly placed the
cancel_work_sync() call in nvme_fc_reset_ctrl_work() instead of
nvme_fc_delete_ctrl(), leaving the original bug unfixed while
introducing an unnecessary change to the reset path.

This series reverts the broken backport and then applies the fix
correctly.

Jaskaran Singh (2):
  Revert "nvme: nvme-fc: Ensure ->ioerr_work is cancelled in nvme_fc_delete_ctrl()"
  nvme: nvme-fc: Ensure ->ioerr_work is cancelled in nvme_fc_delete_ctrl()

 drivers/nvme/host/fc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

-- 
2.43.7
Re: [PATCH 5.10.y 0/2] Fix incorrect backport of nvme-fc ioerr_work cancel_work_sync()
Posted by Justin Tee 1 month, 1 week ago
For this patch set,

Reviewed-by: Justin Tee <justin.tee@broadcom.com>

Regards,
Justin
Re: [PATCH 5.10.y 0/2] Fix incorrect backport of nvme-fc ioerr_work cancel_work_sync()
Posted by Jaskaran Singh 1 month, 2 weeks ago
Gentle ping.

Thanks,
Jaskaran.

On Mon, Feb 23, 2026 at 10:52 PM Jaskaran Singh <jsingh@cloudlinux.com> wrote:
>
> The backport of upstream commit 0a2c5495b6d1 ("nvme: nvme-fc: Ensure
> ->ioerr_work is cancelled in nvme_fc_delete_ctrl()") to linux-5.10.y
> was incorrectly applied as commit 3d78e8e01251.
>
> The original upstream fix moves the cancel_work_sync(&ctrl->ioerr_work)
> call within nvme_fc_delete_ctrl() to after nvme_fc_delete_association(),
> so that ->ioerr_work is not running when the nvme_fc_ctrl object is
> freed. However, the stable backport mistakenly placed the
> cancel_work_sync() call in nvme_fc_reset_ctrl_work() instead of
> nvme_fc_delete_ctrl(), leaving the original bug unfixed while
> introducing an unnecessary change to the reset path.
>
> This series reverts the broken backport and then applies the fix
> correctly.
>
> Jaskaran Singh (2):
>   Revert "nvme: nvme-fc: Ensure ->ioerr_work is cancelled in nvme_fc_delete_ctrl()"
>   nvme: nvme-fc: Ensure ->ioerr_work is cancelled in nvme_fc_delete_ctrl()
>
>  drivers/nvme/host/fc.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> --
> 2.43.7
>
Re: [PATCH 5.10.y 0/2] Fix incorrect backport of nvme-fc ioerr_work cancel_work_sync()
Posted by Greg KH 1 month, 1 week ago
On Tue, Mar 03, 2026 at 11:16:34AM +0530, Jaskaran Singh wrote:
> Gentle ping.

It's been just over 1 week, please relax, we are staring down loads of
submissions at the moment...