Hello,
On Sat, 9 Dec 2000, Victor J. Orlikowski wrote:
> After doing some googling....
> It would appear that, in Family 5, Model 8, Stepping 12 of the K6-2,
> AMD used a different CPU core, that was more similar to the K6-3, and
> that there is a slightly odd way of doing write-combining.
I'm now investigating the problem. It seems it is related
to XF86_SVGA (3.3.6):
S3VGEReset called from s3v_accel.c line 300
Hm, what can trigger this reset
On total lockup I can't activate ikd nor to use sysrq.
I can reproduce the temporary lockups afetr 1 minute of testing and
usually receive the above message. I have a ktrace output (not
sure if I catch the problem in this output) and this strace -r for
the XF86_SVGA program:
0.001019 brk(0x851a000) = 0x851a000
0.009577 gettimeofday({976474616, 24017}, NULL) = 0
4.865610 write(2, "\tS3VGEReset called from s3v_acce"..., 45) = 45
0.000410 nanosleep({0, 10000000}, NULL) = 0
0.012520 nanosleep({0, 10000000}, NULL) = 0
0.019712 nanosleep({0, 10000000}, NULL) = 0
0.019997 nanosleep({0, 10000000}, NULL) = 0
What means 4.8 seconds between gettimeofday() and write() ?
Can this be a problem raised from gettimeofday? I have CONFIG_RTC=y
in all tests.
> Perhaps this is the problem?
> Anyone with more knowledge on the AMD cores care to comment?
>
> Victor
So, I'm not sure whether the problem is in XFree or is
hardware/kernel related. IMO, it is not related to the MTRR
support. Any ideas for testing are welcome!
Regards
-- Julian Anastasov <ja@ssi.bg>- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Dec 15 2000 - 21:00:20 EST