Re: [syzbot] [xfs?] possible deadlock in xfs_ilock_attr_map_shared (2)

[Aleksandr Nogikh <nogikh@google.com>]

Hi Dave,

On Thu, Jul 10, 2025 at 12:13 AM 'Dave Chinner' via syzkaller-bugs
<syzkaller-bugs@googlegroups.com> wrote:
>
> On Wed, Jul 09, 2025 at 10:39:29AM -0700, syzbot wrote:
> > Hello,
> >
> > syzbot found the following issue on:
> >
> > HEAD commit:    733923397fd9 Merge tag 'pwm/for-6.16-rc6-fixes' of git://g..
> > git tree:       upstream
> > console output: https://syzkaller.appspot.com/x/log.txt?x=13f53582580000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=b309c907eaab29da
> > dashboard link: https://syzkaller.appspot.com/bug?extid=3470c9ffee63e4abafeb
> > compiler:       Debian clang version 20.1.7 (++20250616065708+6146a88f6049-1~exp1~20250616065826.132), Debian LLD 20.1.7
> >
> > Unfortunately, I don't have any reproducer for this issue yet.
> >
> > Downloadable assets:
> > disk image (non-bootable): https://storage.googleapis.com/syzbot-assets/d900f083ada3/non_bootable_disk-73392339.raw.xz
> > vmlinux: https://storage.googleapis.com/syzbot-assets/be7feaa77b8c/vmlinux-73392339.xz
> > kernel image: https://storage.googleapis.com/syzbot-assets/a663b3e31463/bzImage-73392339.xz
> >
> > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > Reported-by: syzbot+3470c9ffee63e4abafeb@syzkaller.appspotmail.com
> >
> > ======================================================
> > WARNING: possible circular locking dependency detected
> > 6.16.0-rc5-syzkaller-00038-g733923397fd9 #0 Not tainted
> > ------------------------------------------------------
> > syz.0.0/5339 is trying to acquire lock:
> > ffffffff8e247500 (fs_reclaim){+.+.}-{0:0}, at: might_alloc include/linux/sched/mm.h:318 [inline]
> > ffffffff8e247500 (fs_reclaim){+.+.}-{0:0}, at: prepare_alloc_pages+0x153/0x610 mm/page_alloc.c:4727
> >
> > but task is already holding lock:
> > ffff888053415098 (&xfs_nondir_ilock_class){++++}-{4:4}, at: xfs_ilock_attr_map_shared+0x92/0xd0 fs/xfs/xfs_inode.c:85
> >
> > which lock already depends on the new lock.
> >
> >
> > the existing dependency chain (in reverse order) is:
> >
> > -> #1 (&xfs_nondir_ilock_class){++++}-{4:4}:
> >        lock_acquire+0x120/0x360 kernel/locking/lockdep.c:5871
> >        down_write_nested+0x9d/0x200 kernel/locking/rwsem.c:1693
> >        xfs_reclaim_inode fs/xfs/xfs_icache.c:1045 [inline]
> >        xfs_icwalk_process_inode fs/xfs/xfs_icache.c:1737 [inline]
> >        xfs_icwalk_ag+0x12c5/0x1ab0 fs/xfs/xfs_icache.c:1819
> >        xfs_icwalk fs/xfs/xfs_icache.c:1867 [inline]
> >        xfs_reclaim_inodes_nr+0x1e3/0x260 fs/xfs/xfs_icache.c:1111
> >        super_cache_scan+0x41b/0x4b0 fs/super.c:228
> >        do_shrink_slab+0x6ec/0x1110 mm/shrinker.c:437
> >        shrink_slab+0xd74/0x10d0 mm/shrinker.c:664
> >        shrink_one+0x28a/0x7c0 mm/vmscan.c:4939
> >        shrink_many mm/vmscan.c:5000 [inline]
> >        lru_gen_shrink_node mm/vmscan.c:5078 [inline]
> >        shrink_node+0x314e/0x3760 mm/vmscan.c:6060
> >        kswapd_shrink_node mm/vmscan.c:6911 [inline]
> >        balance_pgdat mm/vmscan.c:7094 [inline]
> >        kswapd+0x147c/0x2830 mm/vmscan.c:7359
> >        kthread+0x70e/0x8a0 kernel/kthread.c:464
> >        ret_from_fork+0x3f9/0x770 arch/x86/kernel/process.c:148
> >        ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245
> >
> > -> #0 (fs_reclaim){+.+.}-{0:0}:
> >        check_prev_add kernel/locking/lockdep.c:3168 [inline]
> >        check_prevs_add kernel/locking/lockdep.c:3287 [inline]
> >        validate_chain+0xb9b/0x2140 kernel/locking/lockdep.c:3911
> >        __lock_acquire+0xab9/0xd20 kernel/locking/lockdep.c:5240
> >        lock_acquire+0x120/0x360 kernel/locking/lockdep.c:5871
> >        __fs_reclaim_acquire mm/page_alloc.c:4045 [inline]
> >        fs_reclaim_acquire+0x72/0x100 mm/page_alloc.c:4059
> >        might_alloc include/linux/sched/mm.h:318 [inline]
> >        prepare_alloc_pages+0x153/0x610 mm/page_alloc.c:4727
> >        __alloc_frozen_pages_noprof+0x123/0x370 mm/page_alloc.c:4948
> >        alloc_pages_mpol+0x232/0x4a0 mm/mempolicy.c:2419
> >        alloc_frozen_pages_noprof mm/mempolicy.c:2490 [inline]
> >        alloc_pages_noprof+0xa9/0x190 mm/mempolicy.c:2510
> >        get_free_pages_noprof+0xf/0x80 mm/page_alloc.c:5018
> >        __kasan_populate_vmalloc mm/kasan/shadow.c:362 [inline]
> >        kasan_populate_vmalloc+0x33/0x1a0 mm/kasan/shadow.c:417
> >        alloc_vmap_area+0xd51/0x1490 mm/vmalloc.c:2084
> >        __get_vm_area_node+0x1f8/0x300 mm/vmalloc.c:3179
> >        __vmalloc_node_range_noprof+0x301/0x12f0 mm/vmalloc.c:3845
> >        __vmalloc_node_noprof mm/vmalloc.c:3948 [inline]
> >        __vmalloc_noprof+0xb1/0xf0 mm/vmalloc.c:3962
> >        xfs_buf_alloc_backing_mem fs/xfs/xfs_buf.c:239 [inline]
>
> KASAN is still failing to pass through __GFP_NOLOCKDEP allocation
> context flags. It's also failing to pass through other important
> context restrictions like GFP_NOFS, GFP_NOIO, __GFP_NOFAIL, etc.
>
> Fundamentally, it's a bug to be doing nested GFP_KERNEL allocations
> inside an allocation context that has a more restricted allocation
> context...
>
> #syz set subsystems: kasan

Thanks for the analysis!

I've added the kasan-dev list to Cc.

-- 
Aleksandr

>
> --
> Dave Chinner
> david@fromorbit.com
>