From: Dmitry Monakhov <dmonakhov@openvz.org>
To: Sasha Levin <sasha.levin@oracle.com>,
Theodore Ts'o <tytso@mit.edu>,
adilger.kernel@dilger.ca
Cc: linux-ext4@vger.kernel.org, LKML <linux-kernel@vger.kernel.org>,
Dave Jones <davej@redhat.com>
Subject: Re: ext4: kernel BUG at fs/ext4/inode.c:2959!
Date: Thu, 09 Oct 2014 14:06:46 +0400 [thread overview]
Message-ID: <874mvdblqh.fsf@openvz.org> (raw)
In-Reply-To: <54359749.7010208@oracle.com>
[-- Attachment #1: Type: text/plain, Size: 4486 bytes --]
Sasha Levin <sasha.levin@oracle.com> writes:
> Hi all,
>
> While fuzzing with trinity inside a KVM tools guest running the latest -next
> kernel, I've stumbled on the following spew:
Indeed this is a pure bug.
Race between write and fcntl(F_SETFL) which toggling O_DIRECT.
So even if we started with buffered path we may endup in direct one :)
See my testcase: http://marc.info/?l=linux-ext4&m=141284883914726&w=2
Unfortunately f_flags guarded by filp->f_lock so it looks like we
can not use generic __generic_file_write_iter() and need specific one
which use precached o_direct status.
>
> [ 409.233952] kernel BUG at fs/ext4/inode.c:2959!
> [ 409.234761] invalid opcode: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC KASAN
> [ 409.237192] Dumping ftrace buffer:
> [ 409.238424] (ftrace buffer empty)
> [ 409.239664] Modules linked in:
> [ 409.240706] CPU: 18 PID: 14016 Comm: trinity-c27 Not tainted 3.17.0-next-20141008-sasha-00051-g78b3c93-dirty #1364
> [ 409.242724] task: ffff88032bb00000 ti: ffff8802e78f8000 task.ti: ffff8802e78f8000
> [ 409.242724] RIP: ext4_direct_IO (fs/ext4/inode.c:2959 fs/ext4/inode.c:3098)
> [ 409.242724] RSP: 0018:ffff8802e78fbb98 EFLAGS: 00010246
> [ 409.242724] RAX: ffffda005cf1f7cf RBX: ffff88015f72f2f0 RCX: 0000000000000007
> [ 409.242724] RDX: dfffe90000000000 RSI: 1ffff1002e58c5ef RDI: ffff8802e78fbe78
> [ 409.242724] RBP: ffff8802e78fbc28 R08: dfffe90000000001 R09: ffff8802e78fbe5f
> [ 409.242724] R10: 1ffff1005cf1f7cb R11: 0000000000000001 R12: ffff8802e78fbe60
> [ 409.242724] R13: 000000000000000d R14: ffff8802e78fbc60 R15: ffff88015f72f2f0
> [ 409.242724] FS: 0000000000798860(0063) GS:ffff880851c00000(0000) knlGS:0000000000000000
> [ 409.242724] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
> [ 409.242724] CR2: 0000000002bd70f8 CR3: 000000032b542000 CR4: 00000000000006a0
> [ 409.242724] Stack:
> [ 409.242724] ffff88017380a1e8 0000000000000000 ffff8802e78fbbf8 ffffffff9937c5ff
> [ 409.242724] ffff88015f72f318 7fffffffffffffff 0000000000000000 ffff8802e78fbe78
> [ 409.242724] ffff88015f72f2f0 0000000000000000 000000000000000d 000000015f72f2f0
> [ 409.242724] Call Trace:
> [ 409.242724] ? __filemap_fdatawrite_range (mm/filemap.c:288)
> [ 409.242724] generic_file_direct_write (mm/filemap.c:2413)
> [ 409.242724] __generic_file_write_iter (mm/filemap.c:2595)
> [ 409.242724] ext4_file_write_iter (fs/ext4/file.c:176)
> [ 409.242724] ? get_parent_ip (kernel/sched/core.c:2606)
> [ 409.242724] ? iov_iter_init (mm/iov_iter.c:442)
> [ 409.242724] new_sync_write (fs/read_write.c:478)
> [ 409.242724] vfs_write (fs/read_write.c:534)
> [ 409.242724] SyS_pwrite64 (fs/read_write.c:627 fs/read_write.c:614)
> [ 409.242724] tracesys_phase2 (arch/x86/kernel/entry_64.S:529)
> [ 409.242724] Code: 00 00 65 8b 04 25 58 cb 00 00 a9 ff ff ff 7f 0f 85 94 fb ff ff e8 c1 cf ff ff f6 c4 02 0f 84 86 fb ff ff e9 5d fd ff ff 0f 1f 00 <0f> 0b e8 c9 4c c9 ff 84 c0 0f 84 85 fc ff ff 48 c7 c7 80 63 58
> All code
> ========
> 0: 00 00 add %al,(%rax)
> 2: 65 8b 04 25 58 cb 00 mov %gs:0xcb58,%eax
> 9: 00
> a: a9 ff ff ff 7f test $0x7fffffff,%eax
> f: 0f 85 94 fb ff ff jne 0xfffffffffffffba9
> 15: e8 c1 cf ff ff callq 0xffffffffffffcfdb
> 1a: f6 c4 02 test $0x2,%ah
> 1d: 0f 84 86 fb ff ff je 0xfffffffffffffba9
> 23: e9 5d fd ff ff jmpq 0xfffffffffffffd85
> 28: 0f 1f 00 nopl (%rax)
> 2b:* 0f 0b ud2 <-- trapping instruction
> 2d: e8 c9 4c c9 ff callq 0xffffffffffc94cfb
> 32: 84 c0 test %al,%al
> 34: 0f 84 85 fc ff ff je 0xfffffffffffffcbf
> 3a: 48 c7 c7 80 63 58 00 mov $0x586380,%rdi
>
> Code starting with the faulting instruction
> ===========================================
> 0: 0f 0b ud2
> 2: e8 c9 4c c9 ff callq 0xffffffffffc94cd0
> 7: 84 c0 test %al,%al
> 9: 0f 84 85 fc ff ff je 0xfffffffffffffc94
> f: 48 c7 c7 80 63 58 00 mov $0x586380,%rdi
> [ 409.242724] RIP ext4_direct_IO (fs/ext4/inode.c:2959 fs/ext4/inode.c:3098)
> [ 409.242724] RSP <ffff8802e78fbb98>
>
> Thanks,
> Sasha
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
[-- Attachment #2: Type: application/pgp-signature, Size: 818 bytes --]
prev parent reply other threads:[~2014-10-09 10:06 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-10-08 19:58 ext4: kernel BUG at fs/ext4/inode.c:2959! Sasha Levin
2014-10-09 10:06 ` Dmitry Monakhov [this message]
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=874mvdblqh.fsf@openvz.org \
--to=dmonakhov@openvz.org \
--cc=adilger.kernel@dilger.ca \
--cc=davej@redhat.com \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=sasha.levin@oracle.com \
--cc=tytso@mit.edu \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).