[PATCH v2 5/6] kasan: add tests for workqueue stack recording

Andrey Konovalov andreyknvl at google.com
Mon Aug 24 09:45:56 EDT 2020


On Mon, Aug 24, 2020 at 1:49 PM Marco Elver <elver at google.com> wrote:
>
> On Mon, 24 Aug 2020 at 10:14, Walter Wu <walter-zh.wu at mediatek.com> wrote:
> >
> > Adds a test to verify workqueue stack recording and print it in
> > KASAN report.
> >
> > The KASAN report was as follows(cleaned up slightly):
> >
> >  BUG: KASAN: use-after-free in kasan_workqueue_uaf
> >
> >  Freed by task 54:
> >   kasan_save_stack+0x24/0x50
> >   kasan_set_track+0x24/0x38
> >   kasan_set_free_info+0x20/0x40
> >   __kasan_slab_free+0x10c/0x170
> >   kasan_slab_free+0x10/0x18
> >   kfree+0x98/0x270
> >   kasan_workqueue_work+0xc/0x18
> >
> >  Last potentially related work creation:
> >   kasan_save_stack+0x24/0x50
> >   kasan_record_wq_stack+0xa8/0xb8
> >   insert_work+0x48/0x288
> >   __queue_work+0x3e8/0xc40
> >   queue_work_on+0xf4/0x118
> >   kasan_workqueue_uaf+0xfc/0x190
> >
> > Signed-off-by: Walter Wu <walter-zh.wu at mediatek.com>
> > Cc: Andrey Ryabinin <aryabinin at virtuozzo.com>
> > Cc: Dmitry Vyukov <dvyukov at google.com>
> > Cc: Alexander Potapenko <glider at google.com>
> > Cc: Matthias Brugger <matthias.bgg at gmail.com>
> > Cc: Andrey Konovalov <andreyknvl at google.com>
> > ---
> >  lib/test_kasan.c | 29 +++++++++++++++++++++++++++++
> >  1 file changed, 29 insertions(+)
>
> These will majorly conflict with the KASAN-test KUnit rework, which I
> don't know what the status is. AFAIK, these are not yet in -mm tree.

I've asked Andrew to take those in 5.9, but that didn't happen.
Perhaps we should ping him again after Plumbers.

> I think the KASAN-test KUnit rework has priority, as rebasing that
> work on top of this patch is going to be difficult. So maybe these
> test additions can be declared optional if there are conflicts coming,
> and if that'll be the case you'll have to rebase and resend the test.

Yeah, either waiting for KASAN+Kunit or separating the tests sounds
like plausible approaches.



More information about the Linux-mediatek mailing list