Re: [syzbot] [btrfs?] kernel BUG in set_state_bits

[David Sterba <dsterba@suse.cz>]

On Tue, Aug 06, 2024 at 03:18:45PM -0700, Darrick J. Wong wrote:
> On Tue, Aug 06, 2024 at 12:25:03PM -0700, syzbot wrote:
> > syzbot suspects this issue was fixed by commit:
> > 
> > commit 33336c1805d3a03240afda0bfb8c8d20395fb1d3
> > Author: Boris Burkov <boris@bur.io>
> > Date:   Thu Jun 20 17:33:10 2024 +0000
> > 
> >     btrfs: preallocate ulist memory for qgroup rsv
> > 
> > bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=165cd373980000
> > start commit:   9fdfb15a3dbf Merge tag 'net-6.6-rc2' of git://git.kernel.o..
> > git tree:       upstream
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=9681c105d52b0a72
> > dashboard link: https://syzkaller.appspot.com/bug?extid=b9d2e54d2301324657ed
> > syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=148ba274680000
> > C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=14ff46c2680000
> > 
> > If the result looks correct, please mark the issue as fixed by replying with:
> > 
> > #syz fix: btrfs: preallocate ulist memory for qgroup rsv
> > 
> > For information about bisection process see: https://goo.gl/tpsmEJ#bisection
> 
> I don't get it, why am I being cc'd on some random btrfs bug?

Because for some reason bisection blames your patch

  Cause bisection: introduced by (bisect log) :
  commit 05fd9564e9faf0f23b4676385e27d9405cef6637
  Author: Darrick J. Wong <djwong@kernel.org>
  Date: Mon Mar 14 17:55:32 2022 +0000

    btrfs: fix fallocate to use file_modified to update permissions consistently


but this is wrong, the issue syzbot hits is an injected ENOMEM in a path
that does not handle it. This is known and we count on memory allocator
not to fail such cases. The subsystem for the syzbot issue is set
correctly, you got CCed because another automatic bisection round was
started after a year.