Re: [PATCH] spi: fsl-dspi: fix NULL pointer dereference

[Krzysztof Kozlowski <krzk@xxxxxxxxxx>]

On Mon, 28 Sep 2020 at 09:29, Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote:
> > >
> > > This is because since this commit, the allocation of the drivers private
> > > data is done explicitly and in this case spi_alloc_master() won't set the
> > > correct pointer.
> > >
> > > Fixes: 530b5affc675 ("spi: fsl-dspi: fix use-after-free in remove path")
> > > Signed-off-by: Michael Walle <michael@xxxxxxxx>
> > > ---
> >
> > Sascha, how did you test commit 530b5affc675?
>
> Hi,
>
> I just hit it on my Vybrid systems as well. It fails on every boot, so
> I have doubts that it was actually tested. The fix was posted on 23rd
> and applied within a few hours... also no time for anyone else to test
> it.

The flow of this patch to mainline/RC reminds me what Sasha Levin was
saying here:
https://lwn.net/Articles/753329/
" - This means that -rc commits mostly end up replacing obvious bugs
with less obvious ones.
 - A merge window commit spent 50% more days, on average, in -next
than a -rc commit."

Best regards,
Krzysztof