Re: [syzbot] WARNING in btrfs_block_rsv_release
From: Hawkins Jiawei
Date: Fri Oct 28 2022 - 07:14:40 EST
Hi,
> Hello,
>
> syzbot found the following issue on:
>
> HEAD commit: 493ffd6605b2 Merge tag 'ucount-rlimits-cleanups-for-v5.19'..
> git tree: upstream
> console+strace: https://syzkaller.appspot.com/x/log.txt?x=1025dd72880000
> kernel config: https://syzkaller.appspot.com/x/.config?x=d19f5d16783f901
> dashboard link: https://syzkaller.appspot.com/bug?extid=dde7e853812ed57835ea
> compiler: Debian clang version 13.0.1-++20220126092033+75e33f71c2da-1~exp1~20220126212112.63, GNU ld (GNU Binutils for Debian) 2.35.2
> syz repro: https://syzkaller.appspot.com/x/repro.syz?x=17d16e6e880000
> C reproducer: https://syzkaller.appspot.com/x/repro.c?x=1672873c880000
>
> Downloadable assets:
> disk image: https://storage.googleapis.com/syzbot-assets/f1ff6481e26f/disk-493ffd66.raw.xz
> vmlinux: https://storage.googleapis.com/syzbot-assets/101bd3c7ae47/vmlinux-493ffd66.xz
> mounted in repro: https://storage.googleapis.com/syzbot-assets/df89d50ed284/mount_0.gz
>
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+dde7e853812ed57835ea@xxxxxxxxxxxxxxxxxxxxxxxxx
>
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 3612 at fs/btrfs/space-info.h:122 btrfs_space_info_free_bytes_may_use fs/btrfs/space-info.h:154 [inline]
> WARNING: CPU: 0 PID: 3612 at fs/btrfs/space-info.h:122 block_rsv_release_bytes fs/btrfs/block-rsv.c:151 [inline]
> WARNING: CPU: 0 PID: 3612 at fs/btrfs/space-info.h:122 btrfs_block_rsv_release+0x5d1/0x730 fs/btrfs/block-rsv.c:295
> Modules linked in:
> CPU: 0 PID: 3612 Comm: syz-executor894 Not tainted 6.0.0-syzkaller-09423-g493ffd6605b2 #0
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/22/2022
> RIP: 0010:btrfs_space_info_update_bytes_may_use fs/btrfs/space-info.h:122 [inline]
> RIP: 0010:btrfs_space_info_free_bytes_may_use fs/btrfs/space-info.h:154 [inline]
> RIP: 0010:block_rsv_release_bytes fs/btrfs/block-rsv.c:151 [inline]
> RIP: 0010:btrfs_block_rsv_release+0x5d1/0x730 fs/btrfs/block-rsv.c:295
> Code: 8b 7c 24 10 74 08 4c 89 f7 e8 2b 94 33 fe 49 8b 1e 48 89 df 48 8b 2c 24 48 89 ee e8 a9 2b e0 fd 48 39 eb 73 0b e8 5f 29 e0 fd <0f> 0b 31 db eb 25 e8 54 29 e0 fd 48 b8 00 00 00 00 00 fc ff df 41
> RSP: 0000:ffffc90003baf9e8 EFLAGS: 00010293
> RAX: ffffffff83a657f1 RBX: 00000000000d0000 RCX: ffff888020c59d80
> RDX: 0000000000000000 RSI: 00000000000e0000 RDI: 00000000000d0000
> RBP: 00000000000e0000 R08: ffffffff83a657e7 R09: fffffbfff1c19fde
> R10: fffffbfff1c19fde R11: 1ffffffff1c19fdd R12: 1ffff11004f2190c
> R13: 00000000000e0000 R14: ffff88802790c860 R15: 0000000000000000
> FS: 000055555651b300(0000) GS:ffff8880b9a00000(0000) knlGS:0000000000000000
> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 00007f207ed54000 CR3: 0000000026ea2000 CR4: 00000000003506f0
> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> Call Trace:
> <TASK>
> btrfs_release_global_block_rsv+0x2f/0x250 fs/btrfs/block-rsv.c:463
> btrfs_free_block_groups+0xb67/0xfd0 fs/btrfs/block-group.c:4053
> close_ctree+0x6c5/0xbde fs/btrfs/disk-io.c:4710
> generic_shutdown_super+0x130/0x310 fs/super.c:491
> kill_anon_super+0x36/0x60 fs/super.c:1085
> btrfs_kill_super+0x3d/0x50 fs/btrfs/super.c:2441
> deactivate_locked_super+0xa7/0xf0 fs/super.c:331
> cleanup_mnt+0x4ce/0x560 fs/namespace.c:1186
> task_work_run+0x146/0x1c0 kernel/task_work.c:177
> ptrace_notify+0x29a/0x340 kernel/signal.c:2354
> ptrace_report_syscall include/linux/ptrace.h:420 [inline]
> ptrace_report_syscall_exit include/linux/ptrace.h:482 [inline]
> syscall_exit_work+0x8c/0xe0 kernel/entry/common.c:249
> syscall_exit_to_user_mode_prepare+0x63/0xc0 kernel/entry/common.c:276
> __syscall_exit_to_user_mode_work kernel/entry/common.c:281 [inline]
> syscall_exit_to_user_mode+0xa/0x60 kernel/entry/common.c:294
> do_syscall_64+0x49/0xb0 arch/x86/entry/common.c:86
> entry_SYSCALL_64_after_hwframe+0x63/0xcd
> RIP: 0033:0x7f694614c2f7
> Code: 07 00 48 83 c4 08 5b 5d c3 66 2e 0f 1f 84 00 00 00 00 00 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 b8 a6 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48
> RSP: 002b:00007ffee1dcd8e8 EFLAGS: 00000206 ORIG_RAX: 00000000000000a6
> RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00007f694614c2f7
> RDX: 00007ffee1dcd9a9 RSI: 000000000000000a RDI: 00007ffee1dcd9a0
> RBP: 00007ffee1dcd9a0 R08: 00000000ffffffff R09: 00007ffee1dcd780
> R10: 000055555651c653 R11: 0000000000000206 R12: 00007ffee1dcea20
> R13: 000055555651c5f0 R14: 00007ffee1dcd910 R15: 0000000000000004
> </TASK>
According to my analysis, this bug seems to be related to
btrfs_free_reserved_bytes()(Please correct me if I am wrong).
To be more specific, in btrfs_new_extent_direct(), kernel will
reserves space for extent by btrfs_reserve_extent(), and
frees those space by btrfs_free_reserved_extent() if
btrfs_create_dio_extent() fails(such as inject_fault in the syzkaller)
Yet the problem is that when reserving space for extent, kernel will
converts it from * ->bytes_may_use to ->bytes_reserved, in
btrfs_add_reserved_bytes(). But in freeing those space,
kernel does not convert it from ->bytes_reserved back to
* ->bytes_may_use in btrfs_free_reserved_bytes(),
which triggers the above warning.(Please correct me if I am wrong)
So I think we can convert space from ->bytes_reserved back to
* ->bytes_may_use in btrfs_free_reserved_bytes() to solve this bug,
as below:
#syz test https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
master
diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index deebc8ddbd93..cb74fcbf7aaf 100644
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -3415,6 +3415,11 @@ void btrfs_free_reserved_bytes(struct btrfs_block_group *cache,
space_info->bytes_reserved -= num_bytes;
space_info->max_extent_size = 0;
+ trace_btrfs_space_reservation(cache->fs_info, "space_info",
+ space_info->flags, -num_bytes, 1);
+ btrfs_space_info_update_bytes_may_use(cache->fs_info,
+ space_info, num_bytes);
+
if (delalloc)
cache->delalloc_bytes -= num_bytes;
spin_unlock(&cache->lock);