Linux-BTRFS Archive on lore.kernel.org
 help / color / Atom feed
From: David Sterba <dsterba@suse.com>
To: linux-btrfs@vger.kernel.org
Cc: wqu@suse.com
Subject: BUG at fs/btrfs/relocation.c:794!
Date: Wed,  1 Jul 2020 00:10:06 +0200
Message-ID: <20200630221006.17585-1-dsterba@suse.com> (raw)

Hi,

I've hit a crash in relocation I've never seen before.

[ 2129.210066] kernel BUG at fs/btrfs/relocation.c:794!
[ 2129.215268] invalid opcode: 0000 [#1] PREEMPT SMP
[ 2129.220114] CPU: 1 PID: 3303 Comm: btrfs Not tainted 5.8.0-rc3-git+ #638
[ 2129.220116] Hardware name: empty empty/S3993, BIOS PAQEX0-3 02/24/2008
[ 2129.220265] RIP: 0010:create_reloc_root+0x214/0x260 [btrfs]
[ 2129.258760] RSP: 0018:ffffbe1e809b38b8 EFLAGS: 00010282
[ 2129.258763] RAX: 00000000ffffffef RBX: ffff988d577f9000 RCX: 0000000000000000
[ 2129.258765] RDX: 0000000000000001 RSI: ffffffff8e2a2580 RDI: ffff988d64aaa6a8
[ 2129.258766] RBP: ffff988d5dfcdc00 R08: 0000000000000000 R09: 0000000000000000
[ 2129.258767] R10: 0000000000000001 R11: 0000000000000000 R12: ffff988d0e02fa78
[ 2129.258769] R13: 0000000000000005 R14: ffff988d64fe8000 R15: ffff988d0e02fa78
[ 2129.258771] FS:  00007f82a612e8c0(0000) GS:ffff988d67000000(0000) knlGS:0000000000000000
[ 2129.258772] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 2129.258774] CR2: 000000000559d028 CR3: 000000020b289000 CR4: 00000000000006e0
[ 2129.258775] Call Trace:
[ 2129.258825]  btrfs_init_reloc_root+0xe8/0x120 [btrfs]
[ 2129.258862]  record_root_in_trans+0xae/0xd0 [btrfs]
[ 2129.258901]  btrfs_record_root_in_trans+0x51/0x70 [btrfs]
[ 2129.340388]  select_reloc_root+0x94/0x340 [btrfs]
[ 2129.340433]  do_relocation+0xda/0x7b0 [btrfs]
[ 2129.349854]  ? _raw_spin_unlock+0x1f/0x40
[ 2129.349898]  relocate_tree_blocks+0x336/0x670 [btrfs]
[ 2129.359325]  relocate_block_group+0x2f6/0x600 [btrfs]
[ 2129.359365]  btrfs_relocate_block_group+0x15e/0x340 [btrfs]
[ 2129.359408]  btrfs_relocate_chunk+0x38/0x110 [btrfs]
[ 2129.375494]  __btrfs_balance+0x42c/0xce0 [btrfs]
[ 2129.375553]  btrfs_balance+0x66a/0xbe0 [btrfs]
[ 2129.375562]  ? kmem_cache_alloc_trace+0x19c/0x330
[ 2129.389852]  btrfs_ioctl_balance+0x298/0x350 [btrfs]
[ 2129.389887]  btrfs_ioctl+0x304/0x2490 [btrfs]
[ 2129.389898]  ? do_user_addr_fault+0x221/0x49c
[ 2129.404070]  ? sched_clock_cpu+0x15/0x140
[ 2129.404073]  ? do_user_addr_fault+0x221/0x49c
[ 2129.404079]  ? up_read+0x18/0x240
[ 2129.404086]  ? ksys_ioctl+0x68/0xa0
[ 2129.404091]  ksys_ioctl+0x68/0xa0
[ 2129.423308]  __x64_sys_ioctl+0x16/0x20
[ 2129.423312]  do_syscall_64+0x50/0xe0
[ 2129.423315]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[ 2129.423318] RIP: 0033:0x7f82a51c6327
[ 2129.423319] Code: Bad RIP value.
[ 2129.423348] RSP: 002b:00007ffd32cf6218 EFLAGS: 00000206 ORIG_RAX: 0000000000000010
[ 2129.423367] RAX: ffffffffffffffda RBX: 0000000000000001 RCX: 00007f82a51c6327
[ 2129.423368] RDX: 00007ffd32cf62a0 RSI: 00000000c4009420 RDI: 0000000000000003
[ 2129.423372] RBP: 0000000000000003 R08: 0000000000000000 R09: 0000000000000000
[ 2129.423377] R10: 000000000fa99fa0 R11: 0000000000000206 R12: 00007ffd32cf8823
[ 2129.423379] R13: 00007ffd32cf62a0 R14: 0000000000000001 R15: 0000000000000000

Relevant code called from create_reloc_root:

        ret = btrfs_insert_root(trans, fs_info->tree_root,
                                &root_key, root_item);
        BUG_ON(ret)

and according to EAX, ret is -17 which is EEXIST.

I don't have a reproducer, the testing image has been filled by random git
checkouts, deduplicated by BEES, then tons of snapshots created until the
metadata got exhausted, some file deletion and balances.

This is the same image that led to the patch "btrfs: allow use of global block
reserve for balance item deletion", so this could have left it in some
intermediate state where the balance item was not removed and the reloc tree as
well.

There were a few unsuccessful mounts due to relocation recovery, that was
trying to debug but then it started to work.

The error happened with this 'fi df' saved after the balance start:

# btrfs fi df mnt
Data, single: total=80.01GiB, used=38.67GiB
System, single: total=4.00MiB, used=16.00KiB
Metadata, single: total=19.99GiB, used=19.46GiB
GlobalReserve, single: total=512.00MiB, used=44.00KiB

The error looks like a repeated relocation tree creation, which would point to
the unsuccesful balances or inconsistent state (balance item, reloc trees).
It's not a "typical" mix of operations but I'd appreciate any insights here.

                 reply index

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200630221006.17585-1-dsterba@suse.com \
    --to=dsterba@suse.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=wqu@suse.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-BTRFS Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-btrfs/0 linux-btrfs/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-btrfs linux-btrfs/ https://lore.kernel.org/linux-btrfs \
		linux-btrfs@vger.kernel.org
	public-inbox-index linux-btrfs

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-btrfs


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git