Re: [PATCH v8 0/2] gpiolib: don't allow user-space to crash the kernel with hot-unplugs

[Bartosz Golaszewski <brgl@xxxxxxxx>]

On Mon, Dec 5, 2022 at 2:01 PM Andy Shevchenko
<andriy.shevchenko@xxxxxxxxxxxxxxx> wrote:
>
> On Mon, Dec 05, 2022 at 02:59:42PM +0200, Andy Shevchenko wrote:
> > On Mon, Dec 05, 2022 at 01:39:01PM +0100, Bartosz Golaszewski wrote:
> > > From: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxx>
> > >
> > > Linus Torvalds pointed out that using trylock here is wrong. This iteration
> > > drops it in favor of unconditional locks but keeps all the fixes that came
> > > later.
> > >
> > > I will also not send it for this release but make it part of the updates PR
> > > for v6.2 to give it some time in next.
> > >
> > > v7 -> v8:
> > > - don't use down_read_trylock(), just go straight for a full lock
> >
> > Yep, it was a good wrap-up explanation.
>
> But he also suggested to fold NULL check into call_*_locked(). Any points why
> you decided not to go that way?
>

Yes! Every read(), ioctl() and poll() variant extract a different
structure from file->private_data. I'm afraid we need to keep those in
the callbacks down the stack. I too thought it was a good idea but
that won't work.

Bart