[Linux-kernel-mentees] [PATCH v3] media/v4l2-core: Fix kernel-infoleak in video_put_user()

Arnd Bergmann arnd at arndb.de
Thu Jul 30 13:49:08 UTC 2020


On Thu, Jul 30, 2020 at 1:48 PM Andy Shevchenko
<andriy.shevchenko at linux.intel.com> wrote:
> On Thu, Jul 30, 2020 at 11:18:04AM +0200, Arnd Bergmann wrote:
> > The in_ia32_syscall() check should be completely reliable in telling whether
> > we are called from read() by an ia32 task or not, and we use the same
> > logic for input_event, which has a similar problem (on all compat architectures,
> > not just x86).
>
> By the way any reason why we have to have in_ia32_syscall() instead of
> in_compat_syscall()?

x86 is the only architecture that has different struct alignment between 32-bit
and 64-bit processes, so others don't have this particular problem. On top of
that, x86 also has two different 32-bit ABIs and only one of them needs the
workaround, while the other (x32) uses the same struct layout as x86-64 and
must use the normal code path.

       Arnd


More information about the Linux-kernel-mentees mailing list