Hello again,
First of all, thanks to all the people who answered!
After commenting out the hdparm lines in rc.sysinit,
I have rebooted with the linux append="ide0=autotune"
option.
Running
hdparm -tT /dev/hda
in single user mode, before and after these changes,
doesn't show any difference in buffered disks reads:
6.04 MB/s vs. 6.05 MB/s
running sdiff -s on the corresponding dmesg reports
gives (as expected, right?) the following result:
2c2
Detected 350799006 Hz processor. |
Detected 350804250 Hz processor.
32c32
PCSP 1.3 measurement: maximal samplerate 91783 Hz, 18356 Hz u | PCSP
1.3 measurement: maximal samplerate 85227 Hz, 18356
72,76d71
hda: timeout waiting for DMA <
hda: irq timeout: status=0x59 { DriveReady SeekComplete DataR <
hda: irq timeout: error=0x84 { DriveStatusError BadCRC } <
hda: DMA disabled <
ide0: reset: success <
78a74,75
> nfsd:
terminating on signal 9
> nfsd:
last server exiting
The MB is a FIC PA-2013 with VIA APOLLO MVP3 chipset,
128 MB ECC SDRAM with parity, and a 8.4 GB HDD (/dev/hda)
whith UDMA-33 capabilities, and NEC CDROM on /dev/hdc (other
controller?).
I haven't tried yet to check the cables as someone suggested.
My questions now are:
1) Any explanation why the LILO option doesn't make any difference?
(or is it just that ~6 MB/s is already the limit for this MB/HDD
combination?)
2) I am going to read carefully the man page for hdparm as soon as I
can, but it would be nice to get some suggestions on
how to use it with my setup.
Thank you again,
Marco Fioretti
