Fw: [Bugme-new] [Bug 3125] New: Can't read data from zip250 reiserfs partition

Fw: [Bugme-new] [Bug 3125] New: Can't read data from zip250 reiserfs partition

[Andrew Morton]

Looks like a repeatable crash in the imm.c driver.

Begin forwarded message:

Date: Tue, 27 Jul 2004 12:27:49 -0700
From: bugme-daemon@osdl.org
To: bugme-new@lists.osdl.org
Subject: [Bugme-new] [Bug 3125] New: Can't read data from zip250 reiserfs partition


http://bugme.osdl.org/show_bug.cgi?id=3125

           Summary: Can't read data from zip250 reiserfs partition
    Kernel Version: 2.6.7
            Status: NEW
          Severity: normal
             Owner: reiserfs-dev@namesys.com
         Submitter: wwh20610@cmsu2.cmsu.edu


Distribution: My Own
Hardware Environment: i686 pentium3 i386
Software Environment: GNU/linux

Gnu C                  3.4.1
Gnu make               3.80
binutils               2.15.91.0.1
util-linux             2.12a
mount                  2.12a
module-init-tools      3.0
e2fsprogs              1.35
Linux C Library        2.3.4
Dynamic linker (ldd)   2.3.4
Linux C++ Library      6.0.1
Procps                 3.2.1
Net-tools              1.60
Kbd                    1.12
Sh-utils               5.2.1
Modules Loaded

Problem Description: Not able to read data from reiserfs partition on zip250 
disk.  Using imm driver in scsi low level drivers in the kernel.

Steps to reproduce:

Mount a zip250 with a reiserfs partition on it not created from the host 
machine.  Try to cat, move, copy, or anything with the file and the kernel is 
not happy.

Here's the information:

SCSI device sdc: 489532 512-byte hdwr sectors (251 MB)
sdc: Write Protect is off
sdc: Mode Sense: 31 00 00 08
SCSI device sdc: drive cache: write back
SCSI device sdc: 489532 512-byte hdwr sectors (251 MB)
sdc: Write Protect is off
sdc: Mode Sense: 31 00 00 08
SCSI device sdc: drive cache: write back
 sdc: sdc4
ReiserFS: sdc4: found reiserfs format "3.6" with standard journal
ReiserFS: sdc4: using ordered data mode
ReiserFS: sdc4: journal params: device sdc4, size 8192, journal first block 18, 
max trans len 1024, max batch 900, max commit age 30, max trans age 30
ReiserFS: sdc4: checking transaction log (sdc4)
ReiserFS: sdc4: Using r5 hash to sort names
Unable to handle kernel NULL pointer dereference at virtual address 00000000
 printing eip:
c02bef2b
*pde = 00000000
Oops: 0002 [#1]
PREEMPT SMP 
Modules linked in:
CPU:    1
EIP:    0060:[<c02bef2b>]    Not tainted
EFLAGS: 00010202   (2.6.7) 
EIP is at imm_byte_in+0x3b/0x60
eax: 000000d0   ebx: 00000200   ecx: 0000037a   edx: 00000378
esi: 00000001   edi: 00000378   ebp: 00000000   esp: c2071e80
ds: 007b   es: 007b   ss: 0068
Process events/1 (pid: 7, threadinfo=c2071000 task=c20fcb70)
Stack: f7fc4758 00000378 00000000 00000379 c02bf232 00000378 00000000 00000200 
       c2022be0 00000200 f528cb80 f7fc4758 00000379 c02bf6ef f7fc4758 00000000 
       00000200 0000037a 00000001 004b1b45 000007d0 00000006 f528cb80 f7fc4758 
Call Trace:
 [<c02bf232>] imm_in+0x152/0x190
 [<c02bf6ef>] imm_completion+0x1cf/0x220
 [<c02bfc93>] imm_engine+0x403/0x480
 [<c02bf777>] imm_interrupt+0x37/0x150
 [<c0130686>] worker_thread+0x1f6/0x2d0
 [<c02bf740>] imm_interrupt+0x0/0x150
 [<c011b130>] default_wake_function+0x0/0x20
 [<c011b130>] default_wake_function+0x0/0x20 [<c0102185>] 
kernel_thread_helper+0x5/0x10

Code: 88 45 00 89 ca b0 25 ee 4b 75 ea 8b 1c 24 b8 01 00 00 00 8b 
 <6>parport0: imm already owner
parport0: imm already owner
scsi: Device offlined - not ready after error recovery: host 1 channel 0 id 6 
lun 0
SCSI error : <1 0 6 0> return code = 0x70000
end_request: I/O error, dev sdc, sector 76560
scsi1 (6:0): rejecting I/O to offline device
scsi1 (6:0): rejecting I/O to offline device
Buffer I/O error on device sdc4, logical block 95
lost page write due to I/O error on sdc4
Buffer I/O error on device sdc4, logical block 96
REISERFS: panic (device sdc4): journal-601, buffer write failed

------------[ cut here ]------------
kernel BUG at fs/reiserfs/prints.c:362!
invalid operand: 0000 [#2]
PREEMPT SMP 
Modules linked in:
CPU:    0
EIP:    0060:[<c01b95c2>]    Not tainted
EFLAGS: 00010286   (2.6.7) 
lost page write due to I/O error on sdc4
EIP is at reiserfs_panic+0x52/0x80
eax: 00000044   ebx: c03ae800   ecx: c03eba14   edx: f7d13000
esi: f7cb4c00   edi: f7cb4d44   ebp: f7cb4c00   esp: f7d13df8
ds: 007b   es: 007b   ss: 0068
Process pdflush (pid: 13, threadinfo=f7d13000 task=f7ceb6d0)
Stack: c03b3220 f7cb4d44 c04ceaa0 f3b58000 f7f9ad58 00000000 c01c6b83 f7cb4c00 
       c03b42f4 00001000 fabe6070 f3b58000 00000001 f7f9ad70 f58bc83c f7cb4c00 
       f7f9ad58 00000000 f7f9ad58 c01cbd98 f7cb4c00 f7f9ad58 00000001 00000006 
Call Trace:
 [<c01c6b83>] flush_commit_list+0x443/0x470
 [<c01cbd98>] do_journal_end+0xb18/0xb99
 [<c01ca718>] journal_end_sync+0x48/0xb0
 [<c01b5a09>] reiserfs_sync_fs+0x79/0xe0
 [<c0168d1d>] put_super+0x2d/0x70
 [<c0169490>] sync_supers+0x120/0x130
 [<c0145415>] wb_kupdate+0x35/0x130
 [<c0394c6b>] schedule+0x1cb/0x670
 [<c014612b>] __pdflush+0x10b/0x240
 [<c0146260>] pdflush+0x0/0x30
 [<c0146286>] pdflush+0x26/0x30
 [<c01453e0>] wb_kupdate+0x0/0x130
 [<c0146260>] pdflush+0x0/0x30
 [<c013478c>] kthread+0xbc/0x100
 [<c01346d0>] kthread+0x0/0x100
 [<c0102185>] kernel_thread_helper+0x5/0x10

Code: 0f 0b 6a 01 cb ec 3a c0 c7 04 24 44 32 3b c0 85 f6 be a0 ea 
 <3>scsi1 (6:0): rejecting I/O to offline device
Buffer I/O error on device sdc4, logical block 9566
 [<c0130490>] worker_thread+0x0/0x2d0
 [<c013478c>] kthread+0xbc/0x100
 [<c01346d0>] kthread+0x0/0x100

Here is what I have done to confirm my hardware is ok.  I installed slackware 
8.1 using the 2.4.18 kernel and it read the data from the zip250 disk I'm 
having problems with on 2.6.7.  I have another zip250 disk I deleted the 
partitions, cleared the disk, and then made a partition on it and a formatted 
it with reiserfs with the 2.6.7 machine.  I was able to mount and copy and read 
files to the partition.  I took that same disk I made with the 2.6.7 machine 
and took it to the 2.4.18 machine and it could read and write files to it.

With the 2.4.18 machine, I copied the files I needed from the zip250 disk I'm 
having problems with and then copied the files to the zip250 disk made on the 
2.6.7 machine.  I took that disk back to the 2.6.7 machine, mounted it and 
tried to copy a file from it and the kernel gave messages.

My problem, I can't share zipdisks between older kernels and 2.6.7  I do not 
know if it affects kernels in the 2.6 before 2.6.7.  Somehow, writing from the 
2.4.18 kernel messed it up enough that the 2.6.7 kernel has problems reading 
the data.  I can get directory listings no problem in the 2.6.7 kernel.

2.6.7 can work with its own but not with the 2.4.18-22 (havne't confirmed with 
23-26)

2.4.18 can work with it's own and other kernels

The original zip250 disk I'm having problems with was created on a 2.4.22 
machine.

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

Re: Fw: [Bugme-new] [Bug 3125] New: Can't read data from zip250 reiserfs partition

[viro]

On Tue, Jul 27, 2004 at 12:57:12PM -0700, Andrew Morton wrote:
> 
> Looks like a repeatable crash in the imm.c driver.

Can't reproduce it here.  OTOH, randy had been able to; looks like we
might get somewhere tonight (it's on his home box).

It looks like we somehow get a SCSI command with NULL ->request_buffer
and non-zero ->request_buflen.  Which, AFAICS, should break a lot of
SCSI drivers...

Re: Fw: [Bugme-new] [Bug 3125] New: Can't read data from zip250 reiserfs partition

[viro]

On Mon, Aug 02, 2004 at 10:33:23PM +0100, viro@parcelfarce.linux.theplanet.co.uk wrote:
> On Tue, Jul 27, 2004 at 12:57:12PM -0700, Andrew Morton wrote:
> > 
> > Looks like a repeatable crash in the imm.c driver.
> 
> Can't reproduce it here.  OTOH, randy had been able to; looks like we
> might get somewhere tonight (it's on his home box).

OK...  Here's what's going on: imm is PIO-only (hardly a surprise, since it's
a parport ZIP drive).  If we have highmem enabled, eventually we get a request
with sglist refering to unmapped page.  Normally we wouldn't give a fsck -
we know that it's accessible to DMA and that's it.  However, in this driver
we really need CPU access to the damn thing.

So imm.c and ppa.c are screwed on highmem kernels and I would expect even more
fun on non-x86.  AFAICS, other non-DMA drivers are no better - e.g. eata_pio()
is broken in the same way, ditto for fdomain.c, etc.

Any suggestions re clean fix?  I'm seriously tempted to do atomic kmap of
pages we are accessing, ugly as it is...