Package: linux-image
Version: 2.6.26-1-amd64
Severity: important
I have a SuperMicro AOC-SATA-MV8 card, which is an 8-port PCI-X SATA
controller. Looks like:
# lspci -nv -s 5:1
05:01.0 0100: 11ab:6081 (rev 09)
Subsystem: 11ab:11ab
Flags: bus master, 66MHz, medium devsel, latency 32, IRQ 16
Memory at d3200000 (64-bit, non-prefetchable) [size=1M]
I/O ports at 3000 [size=256]
Capabilities: [40] Power Management version 2
Capabilities: [50] Message Signalled Interrupts: Mask- 64bit+
Queue=0/0 Enable-
Capabilities: [60] PCI-X non-bridge device
Kernel driver in use: sata_mv
Kernel modules: sata_mv
It worked fine with the sata_mv driver and Lenny RC1's linux-2.6.26-1-
amd64, until I added a fifth disk. At that point, under any kind of
load at all, I would get frequent timeouts -- say, 2-3 seconds of I/O
at 200+ MB/s, then 30 seconds at 0, until the error handler resets the
port in question and the cycle repeats. `dmesg` from one timeout event:
[ 52.101826] ata3.00: exception Emask 0x0 SAct 0x1 SErr 0x0 action
0x6 frozen
[ 52.109814] ata3.00: cmd 61/08:00:d9:58:70/00:00:74:00:00/40 tag 0
ncq 4096 out
[ 52.109815] res 40/00:00:00:00:00/00:00:00:00:00/00 Emask
0x4 (timeout)
[ 52.117795] ata3.00: status: { DRDY }
[ 52.121792] ata3: hard resetting link
[ 52.611832] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 52.627842] ata3.00: max_sectors limited to 256 for NCQ
[ 52.643839] ata3.00: max_sectors limited to 256 for NCQ
[ 52.643841] ata3.00: configured for UDMA/133
[ 52.643853] ata3: EH complete
[ 52.643793] sd 2:0:0:0: [sda] 1953525168 512-byte hardware sectors
(1000205 MB)
[ 52.643793] sd 2:0:0:0: [sda] Write Protect is off
[ 52.643793] sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 52.643793] sd 2:0:0:0: [sda] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
Drive info:
# smartctl -a /dev/sda
smartctl version 5.38 [x86_64-unknown-linux-gnu] Copyright (C) 2002-8
Bruce Allen
Home page is http://smartmontools.sourceforge.net/
=== START OF INFORMATION SECTION ===
Device Model: WDC WD1001FALS-00J7B0
Serial Number: WD-WMATV0413539
Firmware Version: 05.00K05
User Capacity: 1,000,204,886,016 bytes
Device is: Not in smartctl database [for details use: -P showall]
ATA Version is: 8
ATA Standard is: Exact ATA specification draft version not indicated
Local Time is: Wed Feb 4 13:19:14 2009 CST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled
<snip>
My configuration uses software RAID via md with LVM on top.
Experimentation revealed that the first four ports were affected,
regardless of which disks or cables were attached. This problem
persisted under a fresh build of 2.6.28.2, the then-current release.
Various BIOS settings and kernel options had no effect.
Further research brought me to RedHat bug #462425, where others
reported the exact same issue with SATA cards from the same controller
family. Included in that thread is a one-line patch from Mark Lord.
After applying the patch to 2.6.28.2, timeouts on my system went away
completely.
Patch references:
Originated at: https://bugzilla.redhat.com/show_bug.cgi?id=462425#c55
Submission to linux-ide:
http://www.spinics.net/lists/linux-ide/msg29192.html
linux-ide request for Linus pull:
http://patchwork.kernel.org/patch/4081/
2.6.29-rc3 inclusion: http://lkml.org/lkml/2009/1/28/277
I suggest that, at a minimum, this particular patch be applied to the
appropriate Debian tree. Without the patch, an 8-port controller is
really only usable with 4 ports. Past that, there have been a number
of other recent sata_mv fixes; perhaps the driver should be completely
updated?
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
