Hello!
I have troubles using IDE disk with recent versions of Linux kernel.
System configuration:
- MB: Chaintech PA-2013 with VIA Apollo VP3 chipset
- CPU: AMD K6-III 350 Mhz
- RAM: 128Mb PC-100 SDRAM
- HDD: IBM DPTA-371360
Kernel versions tried:
1. 2.2.15 with ide.2.2.15.20000509 patch from Andre Hedrick
and reiserfs (versions 3.5.21 and 3.5.22 tried)
2. 2.2.16 with reiserfs-3.5.22
3. 2.2.16 with reiserfs-3.5.22 and ide.2.2.16.20000619 patch
4. 2.4.0-test1-ac22-riel++ with reiserfs-3.6.9
5. 2.4.0-test2-ac2 with reiserfs-3.6.9
I use ALSA sound driver 0.5.8a for my SB AWE 32.
I use "hdparm -d1 -X66 -m16 /dev/hda" to tune IDE subsystem at startup.
Kernels (2) and (3) generate error messages pretty soon after startup
( before I log in and see bash prompt )
Last time I tried it was (from kernel(2)):
> Jul 9 04:26:11 tentacle kernel: hda: timeout waiting for DMA
> Jul 9 04:26:11 tentacle kernel: hda: irq timeout: status=0x58 { DriveReady Seek
> Complete DataRequest }
> Jul 9 04:26:21 tentacle kernel: hda: timeout waiting for DMA
> Jul 9 04:26:21 tentacle kernel: hda: irq timeout: status=0x58 { DriveReady Seek
> Complete DataRequest }
> Jul 9 04:26:31 tentacle kernel: hda: timeout waiting for DMA
> Jul 9 04:26:31 tentacle kernel: hda: irq timeout: status=0x58 { DriveReady Seek
> Complete DataRequest }
> Jul 9 04:26:31 tentacle kernel: hda: status timeout: status=0xd0 { Busy }
> Jul 9 04:26:31 tentacle kernel: hda: DMA disabled
> Jul 9 04:26:31 tentacle kernel: hda: drive not ready for command
> Jul 9 04:26:31 tentacle kernel: ide0: reset: success
Similar message were produced by kernel (3).
Kernel (5) hanged with HDD LED turned on and last message in log
about cron starting updatedb. IIRC there was no message on console.
Machine is currently running kernel (1). It could reach more than 20 days
uptime without error message in the log. But some days ago I noticed
that sometimes data corruption occured when reading large files.
I prepared test directories with large files both on ext2 and reiserfs
partitions and ran md5sum on them several times. I got aprox. one error
per 100Mb of data. After reboot I could not reproduce that for 2 days,
but now it started to happen again. Switching DMA off did help, but perfomance
was much worse even with umaskirq enabled. I managed to read several gigabytes
without an error 3 or 4 times. After re-enabling DMA read errors started
to appear again. No kernel messages were generated on read errors.
The system is otherwise stable. It could compile a kernel more than 100
times in a row after I discovered read errors.
Various system information attached:
kernel configuration
output of hdparm -i
output of lspci -v
kernel boot-up messages
I used similar config for all 2.2 kernels. I tried turning VIA82C586
support on and off when using 2.2.16 w/o ide patches.
:wq
With best regards,
Vladimir Savkin.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Jul 15 2000 - 21:00:14 EST