Re: [PATCH 5.10.y 00/15] timers: Provide timer_shutdown[_sync]()

[Jeongjun Park]

Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On Fri, Feb 20, 2026 at 02:12:55AM +0900, Jeongjun Park wrote:
> > The "timers: Provide timer_shutdown[_sync]()" patch series implemented a
> > useful feature that addresses various bugs caused by attempts to rearm
> > shutdown timers.
> >
> > https://lore.kernel.org/all/20221123201306.823305113@xxxxxxxxxxxxx/
> >
> > However, this patch series was not fully backported to versions prior to
> > 6.2, requiring separate patches for older kernels if these bugs were
> > encountered.
> >
> > The biggest problem with this is that even if these bugs were discovered
> > and patched in the upstream kernel, if the maintainer or author didn't
> > create a separate backport patch for versions prior to 6.2, the bugs would
> > remain untouched in older kernels.
> >
> > Therefore, to reduce the hassle of having to write a separate patch, we
> > should backport the remaining unbackported commits from the
> > "timers: Provide timer_shutdown[_sync]()" patch series to versions prior
> > to 6.2.
> >
> > ---
> >  Documentation/RCU/Design/Requirements/Requirements.rst      |   2 +-
> >  Documentation/core-api/local_ops.rst                        |   2 +-
> >  Documentation/kernel-hacking/locking.rst                    |  17 ++---
> >  Documentation/timers/hrtimers.rst                           |   2 +-
> >  Documentation/translations/it_IT/kernel-hacking/locking.rst |  14 ++---
> >  arch/arm/mach-spear/time.c                                  |   8 +--
> >  drivers/bluetooth/hci_qca.c                                 |  10 ++-
> >  drivers/char/tpm/tpm-dev-common.c                           |   4 +-
> >  drivers/clocksource/arm_arch_timer.c                        |  12 ++--
> >  drivers/clocksource/timer-sp804.c                           |   6 +-
> >  drivers/staging/wlan-ng/hfa384x_usb.c                       |   4 +-
> >  drivers/staging/wlan-ng/prism2usb.c                         |   6 +-
> >  include/linux/timer.h                                       |  17 ++++-
> >  kernel/time/timer.c                                         | 316 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-----------------
> >  net/sunrpc/xprt.c                                           |   2 +-
> >  15 files changed, 322 insertions(+), 100 deletions(-)
> >
>
> Ugh, I got the following build error for this series:
> ../drivers/misc/sgi-xp/xpc_partition.c: In function 'xpc_partition_disengaged':
> ../drivers/misc/sgi-xp/xpc_partition.c:294:25: error: implicit declaration of function 'del_singleshot_timer_sync' [-Werror=implicit-function-declaration]
>   294 |                         del_singleshot_timer_sync(&part->disengage_timer);
>       |                         ^~~~~~~~~~~~~~~~~~~~~~~~~
>

Oh dear. This issue occurred because commit 997754f114ef ("misc/sgi-xp:
Replace in_interrupt() usage") was merged into version 5.11-rc1 and was
therefore not backported to 5.10.y.

Since this is a simple fix that only requires adding this commit to this
patch series, I will quickly write and send you the v2 patch.

https://lore.kernel.org/all/20201119103151.ppo45mj53ulbxjx4@xxxxxxxxxxxxx/

>
> Don't know what happened, but I'll go and drop them all now.
>
> Do you _REALLY_ need these in the 5.10.y kernel?  Who is going to use
> them?
>

You might think it is unnecessary, but I have seen bug patches related to
timer_shutdown[_sync]() being backported after I backported it, and I
believe it is well worth backporting if this feature allows various
bug-fixing patches to be backported smoothly.

> thanks,
>
> greg k-h

Regards,
Jeongjun Park