xfs
[Top] [All Lists]

Re: [PATCH 00/30] xfsprogs: Initial CRC support

To: Dave Chinner <david@xxxxxxxxxxxxx>
Subject: Re: [PATCH 00/30] xfsprogs: Initial CRC support
From: "Michael L. Semon" <mlsemon35@xxxxxxxxx>
Date: Fri, 17 May 2013 16:54:47 -0400
Cc: xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=message-id:date:from:user-agent:mime-version:to:cc:subject :references:in-reply-to:content-type:content-transfer-encoding; bh=TBbV58DptiwbvP9f0KQHL7K7RnFrFcKKMOhz9jDr+wc=; b=x5am4Dt45+3jEk+65sQLORWSInlYRbikUIp8CtcMjK308RAt0c9+JpMn2QzCf84qTF 6/n2mAgDAlsJs1NWqLF14nfPXBxKggwL/hRWBpebQkEa+EDW/FFtTNu/3cGZYvoz6NOl uDor6Q1et+kbAvPmpfLBPbAyOHdq0/EG0D6ddqqg4KRDhUgAPg6HXuHL8NP+jK/oKoGb SK7hxIJLB4qwYk890uTukrn64vaaYvAsnkQu6ZgIg7pCVM/mU1Ls7CdYGZArlFtp9HTi JOv0Mon60Eulg5R3gRFxu0y/dxi2c9eWfpkLin43WzKImZ8zl4bmvub2U2CwWofJNdrv c42Q==
In-reply-to: <1368789205-19969-1-git-send-email-david@xxxxxxxxxxxxx>
References: <1368789205-19969-1-git-send-email-david@xxxxxxxxxxxxx>
User-agent: Mozilla/5.0 (X11; Linux i686; rv:17.0) Gecko/20130509 Thunderbird/17.0.6
On 05/17/2013 07:12 AM, Dave Chinner wrote:
Hi Folks,

This is the first real "works ok" CRC patchset for xfsprogs. It
provides full support for mkfs.xfs and xfs_repair, and partial
read-only support for xfs_db.

For mkfs.xfs, it does everything properly, and filesystems that are
freshly made also run cleanly through xfs_repair and mount and run
just fine.

For xfs_repair, it reads and writes all metadata with CRC checks,
calculations and validation just like the kernel code does, but it
currently silently ignores the validation done in the IO layer.
Enabling that is future work - it involves adding buffer error checking to
every libxfs_readbuf() call that is made, and we do none of that
right now. It does, however, fully validate all the non-CRC format
metadata just as it does for non-CRC filesystems, and so the
coverage it has is the same for both CRC and non-CRC filesystems.

For xfs_db, there is read-only support for looking at the filesystem
as the xfs_db IO stack does not support CRCs at all. We need to
convert xfs_db to use the libxfs infrastructure to enable that.
Apart from that, xfs_db has partial support for the extended
metadata fields - the directory/attribute blocks don't have extended
support yet, but everything else does.

xfs_check is made special. It currently detects a version 5
superblock, and immediately exits with success. Hence it always says
CRC enabled filesystems are OK. This is a temporary change that
enables running xfstests without full support in xfs_db for all the
new metadata structures (like headers in remote symlink and
attribute blocks). Depending on if we want to keep xfs-check useful
for xfstests, we can revisit this bypass hack once xfs_db has been
converted to use the libxfs IO engine.

Overall, xfstests is now running enough to start to find bugs in the
kernel CRC code - I'm mainly hitting remote attribute block bugs
right now (generic/117!) but there's certainly less problems being
reported than I expected.

Oh, and I've tested it with external log devices and real time
devices, too.

Comments, thoughts, flames, and testing all welcome!

Cheers,

Dave.

OK. The basics look good so far. The patchset applied without need for additional work with vi and patch. Whitespace errors were reported for Patches 8, 14, 16, 17, 24, 25, and 27. xfsprogs built with no additional errors over a normal xfsprogs build.

That all stated, the `tar -xvf qt-source.tar.xz` still fails on a CRC-enabled filesystem. Worse, until I return home, I won't be able to do serial-console capture of hard oopses. However, the initial oops I got was a soft one, so it is included after my closing. The kernel is this...

last night's kernel git

last night's xfs-oss/master

some of your recent patches (didn't apply your 6_5 patch yet)

J. Liu's most recent patchset + 2 older bitness patches

Chandra's v8 pquota/gquota patchset + one E-mail fix

Shaggy's JFS patch to make it through the old xfstests #068 on JFS

an NILFS2 patch to address broken bmap handling, lurked from the NILFS2 mailing list

one local removed assert to make it through the old xfstests #111

maybe one or two XFS patches beyond this

...all on a 32-bit Pentium 4.

What I'm trying to state is that a lot is in there, but the PC is spinning like a top, and xfstests results are really good right now. However, if I feel the need to provide a fresh environment, patch management is taking some time.

Great job on a fine patchset so far, and good luck!

Michael

[ 6188.126012] XFS: Assertion failed: first <= last && last < BBTOB(bp->b_length), file: fs/xfs/xfs_trans_buf.c, line: 569
[ 6188.137663] ------------[ cut here ]------------
[ 6188.143109] kernel BUG at fs/xfs/xfs_message.c:108!
[ 6188.147632] invalid opcode: 0000 [#1]
[ 6188.147632] CPU: 0 PID: 12676 Comm: tar Not tainted 3.10.0-rc1+ #25
[ 6188.147632] Hardware name: Dell Computer Corporation Dimension 2350/07W080, BIOS A01 12/17/2002
[ 6188.147632] task: e0ef53e0 ti: ea330000 task.ti: ea330000
[ 6188.147632] EIP: 0060:[<c1170664>] EFLAGS: 00010282 CPU: 0
[ 6188.147632] EIP is at assfail+0x26/0x28
[ 6188.147632] EAX: 0000006b EBX: ea042630 ECX: 00000000 EDX: c1689820
[ 6188.147632] ESI: d94a3900 EDI: ea347aa8 EBP: ea331bf0 ESP: ea331bdc
[ 6188.147632]  DS: 007b ES: 007b FS: 0000 GS: 00e0 SS: 0068
[ 6188.147632] CR0: 8005003b CR2: b74f0000 CR3: 1964e000 CR4: 000007d0
[ 6188.147632] DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
[ 6188.147632] DR6: ffff0ff0 DR7: 00000400
[ 6188.147632] Stack:
[ 6188.147632] 00000000 c160a688 c1613b74 c15fe01d 00000239 ea331c08 c11c6d67 00001ff6 [ 6188.147632] d94a5ffc d94a5000 d94a3900 ea331c24 c11a0653 00001ff7 ea347aa8 d94a5ffc [ 6188.147632] 00000ffc d94a5000 ea331c54 c11a0e45 d94a3900 ea347aa8 01ff0034 d94a5030
[ 6188.147632] Call Trace:
[ 6188.147632]  [<c11c6d67>] xfs_trans_log_buf+0x64/0x11b
[ 6188.147632]  [<c11a0653>] xfs_dir2_data_log_unused+0x7b/0x83
[ 6188.147632]  [<c11a0e45>] xfs_dir2_data_use_free+0x1bf/0x41a
[ 6188.147632]  [<c11a308b>] xfs_dir2_leaf_addname+0x307/0x6f2
[ 6188.147632]  [<c119d32f>] xfs_dir_createname+0x113/0x129
[ 6188.147632]  [<c1174633>] xfs_create+0x3e0/0x4fb
[ 6188.147632]  [<c116e555>] xfs_vn_mknod+0x8f/0x15a
[ 6188.147632]  [<c116e620>] ? xfs_vn_mknod+0x15a/0x15a
[ 6188.147632]  [<c116e635>] xfs_vn_create+0x15/0x17
[ 6188.147632]  [<c109eb31>] vfs_create+0x68/0xeb
[ 6188.147632]  [<c109f410>] do_last+0x85c/0xc45
[ 6188.147632]  [<c109cf53>] ? inode_permission+0x11/0x3e
[ 6188.147632]  [<c109d955>] ? link_path_walk+0x4f/0x67f
[ 6188.147632]  [<c109f898>] path_openat+0x9f/0x38d
[ 6188.147632]  [<c109fbac>] do_filp_open+0x26/0x6b
[ 6188.147632]  [<c12ad0c6>] ? tty_write_unlock+0x2d/0x33
[ 6188.147632]  [<c109cc94>] ? getname_flags+0x86/0x118
[ 6188.147632]  [<c1095157>] do_sys_open+0xf0/0x1ae
[ 6188.147632]  [<c109524f>] SyS_openat+0x1b/0x1d
[ 6188.147632]  [<c14d5d4c>] syscall_call+0x7/0xb
[ 6188.147632]  [<c14d0000>] ? spurious_fault+0xbf/0xc2
[ 6188.147632] Code: 56 eb ff c9 c3 55 89 e5 83 ec 14 89 4c 24 10 89 54 24 0c 89 44 24 08 c7 44 24 04 88 a6 60 c1 c7 04 24 00 00 00 00 e8 e4 fd ff ff <0f> 0b 55 89 e5 83 ec 14 c7 44 24 10 01 00 00 00 89 54 24 0c 89
[ 6188.147632] EIP: [<c1170664>] assfail+0x26/0x28 SS:ESP 0068:ea331bdc
[ 6188.415714] ---[ end trace c213c626812e5949 ]---
[ 6211.447695] XFS (sdb5): Version 5 superblock detected. This kernel has EXPERIMENTAL support enabled!
[ 6211.447695] Use of these features in this kernel is at your own risk!


<Prev in Thread] Current Thread [Next in Thread>