On Fri, Jan 23, 2004 at 10:47:26AM +1300, Carl Cerecke wrote:
> # hdparm -d -u /dev/hdc
[ snipped output]
Looks fine.
> I think the * next to udma5 in the output of hdparm -i also indicates
> that it is using udma mode 5, no? (That's why I didn't run the command
That's right. But the transfer mode (i.e. udma5) is not related to
whether the transfers are using DMA, which is why I asked for that
additional hdparm output.
> VP_IDE: IDE controller at PCI slot 00:11.1
> VP_IDE: chipset revision 6
> VP_IDE: not 100% native mode: will probe irqs later
> VP_IDE: VIA vt8233 (rev 00) IDE UDMA100 controller on pci00:11.1
[snipped additional dmesg]
Looks fine--your kernel has the correct driver for your IDE controller.
Next step; what kernel are you running? It's worth checking the
changelogs between the kernel you're running and whatever is current
(assuming you're not running the latest and greatest) to see if there
have been any improvements in the VIA IDE drivers.
If you've got time to muck around, you could try compiling a 2.6 series
kernel and seeing how it runs. Enable CONFIG_PREEMPT and make sure the
VIA IDE drivers are included. You probably need to set aside a few
hours to get a 2.6 kernel working correctly, because you can't use your
old 2.4 .config verbatim, and it's likely you'll end up with a partially
broken machine on your first attempt.
Can we assume you're not seeing any IDE or disk related errors in your
kernel logs?
What does 'hdparm -tT /dev/hdc' say? You need to run it at least 3
times to get a decent set of results. This doesn't reflect the
real-world speed of your disk, but it's useful for locating certain
types of performance problems.
Does your machine feel sluggish if you run something like 'dd
if=/dev/zero of=/path/to/somewhere/on/hdc/testfile bs=32k count=10k'[0]?
How about when running 'find / > /dev/null'?
[0] This will create a 312MB file named testfile. Don't forgot to
remove it afterwards.
Cheers,
-mjg
--
Matthew Gregan |/
/| [EMAIL PROTECTED]
