KASAN: slab-out-of-bounds Read in hfa384x_usbin_callback

[Dmitry Vyukov]

On Thu, May 7, 2020 at 5:56 PM 'Andrey Konovalov' via syzkaller-bugs
<syzkaller-bugs at googlegroups.com> wrote:
>
> On Wed, May 6, 2020 at 1:50 PM Andrey Konovalov <andreyknvl at google.com> wrote:
> >
> > On Wed, May 6, 2020 at 10:54 AM Oliver Neukum <oneukum at suse.com> wrote:
> > >
> > > Am Freitag, den 20.03.2020, 12:28 -0700 schrieb syzbot:
> > > > Hello,
> > > >
> > > > syzbot found the following crash on:
> > > >
> > > > HEAD commit:    e17994d1 usb: core: kcov: collect coverage from usb comple..
> > > > git tree:       https://github.com/google/kasan.git usb-fuzzer
> > > > console output: https://syzkaller.appspot.com/x/log.txt?x=11d74573e00000
> > > > kernel config:  https://syzkaller.appspot.com/x/.config?x=5d64370c438bc60
> > > > dashboard link: https://syzkaller.appspot.com/bug?extid=7d42d68643a35f71ac8a
> > > > compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> > > > syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=15fa561de00000
> > > > C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=15d74573e00000
> > > >
> > > > IMPORTANT: if you fix the bug, please add the following tag to the commit:
> > > > Reported-by: syzbot+7d42d68643a35f71ac8a at syzkaller.appspotmail.com
> > > >
> > >
> > > Hi,
> > >
> > > is this bug still active and can a test be run on it? I requested one
> > > yesterday. If my analysis is correct this bug has security
> > > implications, so it is kind of important.
> >
> > I see your request in the queue and it's been registered and
> > completed, but for some reason syzbot didn't send an email with a
> > response.
> >
> > Let me try this once again:
> >
> > #syz test: https://github.com/google/kasan.git e17994d1
>
> Still no response. Dmitry, any idea what could be wrong here?

I suspect it has something to do with the fact that the bug is already
fixed (has a fixing commit).

...right, it was broken by:
https://github.com/google/syzkaller/commit/f8368f999a1964df6d39a225cd3f5ab3942dd755
and we lack a test for this scenario. It was supposed to only disable
mailing of bisection jobs.