Re: [PATCH AUTOSEL 5.12 03/43] spi: Fix spi device unregister flow

[Lukas Wunner <lukas@xxxxxxxxx>]

On Thu, Jun 10, 2021 at 12:30:18PM -0700, Saravana Kannan wrote:
> On Thu, Jun 10, 2021 at 12:26 PM Lukas Wunner <lukas@xxxxxxxxx> wrote:
> >
> > On Thu, Jun 10, 2021 at 12:22:40PM -0700, Saravana Kannan wrote:
> > > On Thu, Jun 10, 2021 at 10:55 AM Sasha Levin <sashal@xxxxxxxxxx> wrote:
> > > > On Sun, Jun 06, 2021 at 01:10:28PM +0200, Lukas Wunner wrote:
> > > > >On Thu, Jun 03, 2021 at 01:06:53PM -0400, Sasha Levin wrote:
> > > > >> From: Saravana Kannan <saravanak@xxxxxxxxxx>
> > > > >>
> > > > >> [ Upstream commit c7299fea67696db5bd09d924d1f1080d894f92ef ]
> > > > >
> > > > >This commit shouldn't be backported to stable by itself, it requires
> > > > >that the following fixups are applied on top of it:
> > > > >
> > > > >* Upstream commit 27e7db56cf3d ("spi: Don't have controller clean up spi
> > > > >  device before driver unbind")
> > > > >
> > > > >* spi.git commit 2ec6f20b33eb ("spi: Cleanup on failure of initial setup")
> > > > >  https://git.kernel.org/broonie/spi/c/2ec6f20b33eb
> > > > >
> > > > >Note that the latter is queued for v5.13, but hasn't landed there yet.
> > > > >So you probably need to back out c7299fea6769 from the stable queue and
> > > > >wait for 2ec6f20b33eb to land in upstream.
> > > > >
> > > > >Since you've applied c7299fea6769 to v5.12, v5.10, v5.4, v4.14 and v4.19
> > > > >stable trees, the two fixups listed above need to be backported to all
> > > > >of them.
> > > >
> > > > I took those two patches into 5.12-5.4, but as they needed a more
> > > > complex backport for 4.14 and 4.19, I've dropped c7299fea67 from those
> > > > trees.
> > >
> > > Sounds good. Also, there was a subsequent "Fixes" for this patch and I
> > > think another "Fixes" for the "Fixes". So, before picking this up,
> > > maybe make sure those Fixes patches are pickable too?
> >
> > Aren't those the commits I've listed above?  Or did I miss any fixes?
> > I'm not aware of any others besides these two.
> 
> Ah, those are the ones. I didn't see them. My bad.

All good.  Sasha says that backporting the fixes is a little more
involved in the case of 4.14 and 4.19.  Do you consider the issue
critical enough that it should be addressed in those stable kernels
as well?  (I assume the issue concerns Android devices, not sure
in how far those are using 4.14 and 4.19?)

Thanks,

Lukas