RE: problem with 2.6.13.[0-2]
From: Parag Warudkar
Date: Tue Sep 27 2005 - 21:56:50 EST
DMA is on in both 2.6.12.6 and in 2.6.13.2. Here's what hdparm
/dev/hda gives:
in 2.6.12.6:
/dev/hda:
multcount = 16 (on)
IO_support = 0 (default 16-bit)
unmaskirq = 0 (off)
using_dma = 1 (on)
keepsettings = 0 (off)
readonly = 0 (off)
readahead = 256 (on)
geometry = 65535/16/63, sectors = 195371568, start = 0
in 2.6.13.2 it's the same, except for
multcount = 0 (off)
Could this be the problem, and if so, would setting multcount to on
fix it? (I take it hdparm -m 16 /dev/hda would do it)
I doubt if multcount will make a difference, but you can try.
Probably next best thing to do is profile the slow kernel to find where it
is spending time. (pass profile=2 on kernel command line and use
readprofile or use oprofile).
Did you try hdparm -tT /dev/hda? It's going to be slow, thats for sure
from what you described but try that and then post dmesg (for both
kernels) and profile results - may be someone will get a clue from that.
Parag
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/