Re: Major 2.1.x problem index

Jamie Lokier (lkd@tantalophile.demon.co.uk)
Mon, 20 Jul 1998 04:34:27 +0100

On Wed, Jul 08, 1998 Jamie Lokier wrote:
> Ok, I haven't been getting freezes at bootup, but in general use with
> DMA instead. I'm trying something similar now:
>
> At LILO, `ide0=nodma'
> In /etc/rc.d/rc.local, `hdparm -X34 -d1 /dev/hda'
>
> The results from `hdparm -t -f /dev/hda' are consistent with DMA working
> (~9.30 MB/sec).
>
> I've had this going since Sunday, no lockups so far, though the computer
> hasn't been on for most of that time. This isn't long enough to be
> conclusive yet. (I've had to wait longer for a freeze before, even with
> the computer on all the time, doing all the cron things, downloading
> news in the night etc.).
>
> I'm going to stick with the same kernel (2.1.105) and these settings for
> a while, and see if the freezes have gone away.

Ok, I ran the computer for 4 days and had no crash:
- At LILO, id0=nodma
- In /etc/rc.d/rc.local, `hdparm -X34 -d1 /dev/hda'

Then I ran it for 2.5 days like this, and it crashed with a total freeze once:
- At LILO, no special options
- Nothing in /etc/rc.d/rc.local either

Since then (8-9 more days) I've run it like this, with no further crashes:
- At LILO, no special options (DMA on by default)
- In /etc/rc.d/rc.local, `hdparm -X34 -d1 /dev/hda'

All of these were with the computer only switched on for part of the
day, or (although left on), in the suspended power saving state.
I have been using the computer every day.

The first 4 days are inconclusive, because 4 days is not enough. With
the final 8 days as well, giving a total 12 days no crashing with UDMA
on and -X34 on, and the crash after 2.5 days with exactly the same
kernel but no -X34, I can say:

I'm now quite confident that `hdparm -X34 /dev/hda' fixes the freezing
problem, and that I can use UDMA reliably. I'll get back to you if this
turns out to be wrong.

Note that I do _not_ have an SiS chipset (I saw the comment in
patch-2.1.109). I have:

> [ I have an AMD K6-233 (without the bug), 64Mb SDRAM, an FIC PA-2011
> motherboard (with one of the VIA chipsets), a 6.4G Quantum Fireball ST
> disk (IDE), a Matrox Millenium II, and a Creative Labs SoundBlaster 16.

The obvious thought that occurs is, is there a way for the kernel to
determine when it should do -X34 automatically?

-- Jamie

-
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.altern.org/andrebalsa/doc/lkml-faq.html