There was an extensive discussion about VIA chipset locks on this list.
I turned out also to be a victim of this problem. Unfortunately none of
the solutions posted by different people worked for me. I managed to
find a fix (not a solution) which could be of certain interest.
Previous VIA chipset hangs discussion, see for example:
http://www.uwsg.indiana.edu/hypermail/linux/kernel/9908.3/0002.html
Generally problem reveal itself as a serious hang during extensive disk
I/O operation. Keyboard does not respond, no log message, system
completely stuck.
I'm almost certain that the problem is in initialization of VIA chipset,
especially it's IDE part. I read quite a few mails where people had
troubles with FIC PA-2013 or more generally VIA chipset, it looks like
there is some trick in initialization of IDE, which VIA provides only in
win98 patch.
Hardware
MB : FIC PA-2013 (revision 2.0)
CPU : K6-2-450 with huge fan on top of it.
HD : IBM 15 Gb DeskStar (UDMA66)
CD : CD-DVD Toshiba
Sound : SB live (value)
Video : AGP Diamond V770 TNT2
Modem : Actiontech PCI (works under Linux)
Linux : Debian Slink installation
Fix:
Load 2.0.36 kernel from Win98 (loadlin). No crashes under heavy load.
Just for reference things that I tried and they FAILED. BUT I must say
that they helped quite a few people to solve there hardware problems.
BIOS (award):
Turning off CPU - PCI cache
Disabling UDMA mode
Forcing PIO mode0
Kernel:
2.0.36
2.2.13 + PCI quirks
2.2.13 + recent IDE patch (includes big routine for VIA chipset)
hdparm:
turning off DMA (makes Linux lifetime longer, but it would still crash)
hdparm -X34 -d1 (Linux fails right away)
IBM disk ATA33 utility (make disk report UDMA33 - rather than UDMA66)
I apologize if none of this belongs to this list.
Sasha.
-
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/