Re: [PATCH 0/2] kthread_worker: Fix race between kthread_mod_delayed_work()

[Greg KH <gregkh@linuxfoundation.org>]

On Wed, Jul 07, 2021 at 12:55:30PM +0200, Petr Mladek wrote:
> On Mon 2021-07-05 09:10:58, Greg KH wrote:
> > On Wed, Jun 30, 2021 at 01:31:45PM +0200, Petr Mladek wrote:
> > > On Wed 2021-06-30 13:06:39, Greg KH wrote:
> > > > On Wed, Jun 30, 2021 at 01:01:47PM +0200, Petr Mladek wrote:
> > > > > This is backport of the series for the following stable trees:
> > > > > 
> > > > >    + 4.9
> > > > >    + 4.14
> > > > >    + 4.19
> > > > > 
> > > > > The orignal series did not apply because of a conflict with the commit
> > > > > ("kthread: Convert worker lock to raw spinlock").
> > > > > 
> > > > > Petr Mladek (2):
> > > > >   kthread_worker: split code for canceling the delayed work timer
> > > > >   kthread: prevent deadlock when kthread_mod_delayed_work() races with
> > > > >     kthread_cancel_delayed_work_sync()
> > > > > 
> > > > >  kernel/kthread.c | 77 ++++++++++++++++++++++++++++++++----------------
> > > > >  1 file changed, 51 insertions(+), 26 deletions(-)
> > > > > 
> > > > > -- 
> > > > > 2.26.2
> > > > > 
> > > > 
> > > > What is the original git commit ids of these patches in Linus's tree?
> > > 
> > > commit 34b3d5344719d14fd2185b2d9459b3abcb8cf9d8 ("kthread_worker: split code for canceling the delayed work timer")
> > > commit 5fa54346caf67b4b1b10b1f390316ae466da4d53 ("kthread: prevent deadlock when kthread_mod_delayed_work() races with kthread_cancel_delayed_work_sync())"
> > > 
> > > The original commits have already been taken for the newer stable trees.
> > > 
> > > I am sorry for the inconvenience.
> > 
> > What about backports for 5.4, 5.10, and 5.12?  We can not take patches
> > only for older kernels, otherwise people upgrading to newer ones would
> > have a regression.  Please can you submit these patches for those trees
> > too?  Until that, I can not take these.
> 
> The patches for 5.4, 5.10, and 5.12 did not need any changes. My
> understanding is that the original commits have already been queued
> for these never code trees. See the mail in stable-commits:
> 
> for 5.4:
> 
> https://www.spinics.net/lists/stable-commits/msg203482.html
> https://www.spinics.net/lists/stable-commits/msg203481.html
> 
> for 5.10:
> 
> https://www.spinics.net/lists/stable-commits/msg203486.html
> https://www.spinics.net/lists/stable-commits/msg203485.html
> 
> for 5.12:
> 
> https://www.spinics.net/lists/stable-commits/msg203494.html
> https://www.spinics.net/lists/stable-commits/msg203493.html
> 
> I am sorry, I should have mentioned this in the cover letter.

No worries, all now queued up, thanks.

greg k-h