Re: [syzbot] KASAN: vmalloc-out-of-bounds Write in ringbuf_map_alloc

[Marco Elver]

On Wed, 2 Feb 2022 at 15:36, syzbot
<syzbot+5ad567a418794b9b5983@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> syzbot has bisected this issue to:
>
> commit c34cdf846c1298de1c0f7fbe04820fe96c45068c
> Author: Andrey Konovalov <andreyknvl@xxxxxxxxxx>
> Date:   Wed Feb 2 01:04:27 2022 +0000
>
>     kasan, vmalloc: unpoison VM_ALLOC pages after mapping

Is this a case of a new bug surfacing due to KASAN improvements? But
it's not quite clear to me why this commit.

Andrey, any thoughts?

> bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=128cb900700000
> start commit:   6abab1b81b65 Add linux-next specific files for 20220202
> git tree:       linux-next
> final oops:     https://syzkaller.appspot.com/x/report.txt?x=118cb900700000
> console output: https://syzkaller.appspot.com/x/log.txt?x=168cb900700000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=b8d8750556896349
> dashboard link: https://syzkaller.appspot.com/bug?extid=5ad567a418794b9b5983
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=1450d9f0700000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=130ef35bb00000
>
> Reported-by: syzbot+5ad567a418794b9b5983@xxxxxxxxxxxxxxxxxxxxxxxxx
> Fixes: c34cdf846c12 ("kasan, vmalloc: unpoison VM_ALLOC pages after mapping")
>
> For information about bisection process see: https://goo.gl/tpsmEJ#bisection