sata kernel panic while dd'ing disks

[mrbigglesworth]

hi folks,

please excuse the cross post from the kernel/hardware list, but due to no answers there... i thought i bother you guys with my little problem here....maybe some of you have seen something similar.

during a dd of the active disk to a spare unmounted one, the box suddenly freezes. now after quiet some time i seem to have found the reason:

root@mojo mirror # dd bs=16384 if=/dev/sda2 of=/dev/sdb2
Unable to handle kernel NULL pointer dereference at 0000000000000048 RIP:
<ffffffff8011a3f2>{dma_unmap_sg+50}
PGD 2d9b9067 PUD 3dd9e067 PMD 0
Oops: 0000 [1] PREEMPT
CPU 0
Modules linked in:
Pid: 14567, comm: dd Tainted: G M 2.6.11-gentoo-r11
RIP: 0010:[<ffffffff8011a3f2>] <ffffffff8011a3f2>{dma_unmap_sg+50}
RSP: 0018:ffffffff804a1378 EFLAGS: 00010083
RAX: ffff81003fdccc80 RBX: 0000000000000000 RCX: 0000000000000001
RDX: 0000000000000080 RSI: ffff8100091b5000 RDI: ffff81003feb3070
RBP: 0000000000000080 R08: ffffffff801785d0 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000001
R13: ffff8100091b5000 R14: ffff81003feb3070 R15: 0000000000000000
FS: 0000000000560ae0(0000) GS:ffffffff804e1180(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000048 CR3: 000000003e72c000 CR4: 00000000000006e0
Process dd (pid: 14567, threadinfo ffff81003dd20000, task ffff810032ba6f90)
Stack: ffff81003fdd87f8 ffff8100091b5000 ffff81003fdd82d8 0000000000000001
ffff81003fdd87f8 ffffffff802e1f3e 00000000000001 ffff81003fdd87f8
0000000000000050 0000000000000050
Call Trace:<IRQ> <ffffffff802e1f3e>{ata_sg_clean+222} <ffffffff802e3112>{ata_qc_complete+130}
<ffffffff802e3556>{ata_interrupt+230} <ffffffff8014e8bc>{handle_IRQ_event+44}
<ffffffff8014e9d5>{__do_IRQ+229} <ffffffff8010fdca>{do_IRQ+74}
<ffffffff8010d6c1>{ret_from_intr+0} <EOI> <ffffffff80174229>{__block_prepare_write+329}
<ffffffff801785d0>{blkdev_get_block+0} <ffffffff80243ae2>{copy_user_generic_c+8}
<ffffffff80151e19>{generic_file_buffered_write+713}
<ffffffff80135923>{current_fs_time+99} <ffffffff8018ce5e>{inode_update_time+62}
<ffffffff801524a6>{__generic_file_aio_write_nolock+950}
<ffffffff801509a6>{do_generic_mapping_read+1366} <ffffffff80152524>{generic_file_aio_write_nolock+52}
<ffffffff80152705>{generic_file_write_nolock+165} <ffffffff80150d72>{generic_file_read+194}
<ffffffff80146c00>{autoremove_wake_function+0} <ffffffff8017967a>{blkdev_file_write+26}
<ffffffff80170d40>{vfs_write+176} <ffffffff80170ea3>{sys_write+83}
<ffffffff8010d11a>{system_call+126}

Code: 48 63 c3 48 c1 e0 05 4a 8d 04 28 8b 50 18 85 d2 74 20 44 8b
RIP <ffffffff8011a3f2>{dma_unmap_sg+50} RSP <ffffffff804a1378>
CR2: 0000000000000048
<0>Kernel panic - not syncing: Aiee, killing interrupt handler!


i can do very large compile jobs like a charm, but as soon as im doing a dd or a emerge sync ... ... it crashes.

the box itself is a k8t800 with 1 gig of ram and a VIA SATA controller on it:

ata1: SATA max UDMA/133 cmd 0xEC00 ctl 0xE402 bmdma 0xD800 irq 177
ata2: SATA max UDMA/133 cmd 0xE000 ctl 0xDC02 bmdma 0xD808 irq 177
ata1: dev 0 cfg 49:2f00 82:346b 83:7f61 84:4003 85:3469 86:3e41 87:4003 88:407f
ata1: dev 0 ATA, max UDMA/133, 390721968 sectors: lba48
ata1: dev 0 configured for UDMA/133
scsi0 : sata_via
ata2: dev 0 cfg 49:2f00 82:346b 83:7f61 84:4003 85:3469 86:3e41 87:4003 88:407f
ata2: dev 0 ATA, max UDMA/133, 390721968 sectors: lba48
ata2: dev 0 configured for UDMA/133
scsi1 : sata_via
Vendor: ATA Model: WDC WD2000JD-00H Rev: 08.0
Type: Direct-Access ANSI SCSI revision: 05
Vendor: ATA Model: WDC WD2000JD-00H Rev: 08.0
Type: Direct-Access ANSI SCSI revision: 05
SCSI device sda: 390721968 512-byte hdwr sectors (200050 MB)
SCSI device sda: drive cache: write back
SCSI device sda: 390721968 512-byte hdwr sectors (200050 MB)
SCSI device sda: drive cache: write back
/dev/scsi/host0/bus0/target0/lun0: p1 p2 p3
Attached scsi disk sda at scsi0, channel 0, id 0, lun 0
SCSI device sdb: 390721968 512-byte hdwr sectors (200050 MB)
SCSI device sdb: drive cache: write back
SCSI device sdb: 390721968 512-byte hdwr sectors (200050 MB)
SCSI device sdb: drive cache: write back
/dev/scsi/host1/bus0/target0/lun0: p1 p2 p3
Attached scsi disk sdb at scsi1, channel 0, id 0, lun 0

i ran on 2.6.7 wit the via8237 driver and a very similar behave now on 2.6.11r11 on the sata driver.

any ideas ?

cheers,
hannes

[crazycat]

I suppose your cflags are ok. Try newer kernel. You also might try to boot with noapic option. Then I would try to compile kernel without preemtion.

[mrbigglesworth]

thanks for your hints

well.. the cflags are pretty conservative :

CFLAGS="-march=athlon64 -O2 -pipe"
CHOST="x86_64-pc-linux-gnu"
CXXFLAGS="${CFLAGS}"

i'll try the noapic and the non preemptible kernel.. thanks for your help anyway

hannes