On Wed, Jan 12, 2000 at 08:48:32PM -0800, you [Ryan Lackey] claimed:
> On a debian potato system running 2.3.37/2.3.39, I keep getting the following
> errors:
>
> hdc: write_intr: status=0x58 { DriveReady SeekComplete DataRequest }
> ide1: reset: success
> hdc: write_intr error2: nr_sectors=1, stat=0x58
> hdc: write_intr: status=0x58 { DriveReady SeekComplete DataRequest }
> ide1: reset: success
(snip)
> And then every once in a while, it remounts my root partition readonly and
> I end up rebooting. I've even had some disk corruption.
>
> Presumably this is related to the hpt366 driver; what's up? What should
> I do to fix it?
I'm running 2.2.14 (and 2.2.14pre15). I began to get the same stuff in my
logs friday morning. It had got a bunch of read errors and then reverted
to PIO mode. Even in PIO mode, I got read errors for some files, and stuff
that I wrote to disk I got back slightly changed. This completely hosed a
-o loop mounted reiserfs.
The problem first wouldn't appear in NT /with the Highpoint's driver), but
the drive had only a FAT32 partition, so I was unable to write to it. I
too suspected HPT366 or the IDE patch, but the drive/HPT/patch combo had
been running flawlessly for three weeks.
I tried never IDE-patch, 20000107 and 20000108, but it did not help.
The problem seemed to get worse all the time, and I was unable to delete
my stuff nevermind backing it up.
I tried a different cable, different HPT bus, taking stuff out the machine
(to ensure it wasn't a power supply problem.) No cure. When I moved to the
drive to PX440 IDE channel and not only saw the problem in Linux, but in
NT too, I began to suspect the drive. I moved the drive back to HPT, and
saw the problem in NT too, this time.
As a last resort I tried the drive on another pc, but nothing could revive
it. It just kept giving that funny thic-thic noise when I tried to read
certain sectors.
Yesterday, I fetched a new drive as a guarantee replacement, and
everything is fine now. Now I have only two problems left: bp6 lockups and
linux geforce opengl driver lockups...
I'm not sure whether IDE controller and/or its driver could even in theory
cause harddrive to fail, but it is almostcertain that it was just a faulty
drive in my case. I don't know about yours.
BTW: Is have the 1.21 HPT bios and I was unable to change the UDMA mode of
the drive to anyhing but UDMA/4. Or, I could change it, but when I booted
it was back to UDMA/4. Not that I need it enymore, but has anybody got
that working?
BTW2: I accidentally compiled my new kernel with pgcc-2.95.1 (I usually
use gcc-2.7.2.3). It doesn't seem to be too unstable. I know kernel
developers' position about this is something like "the newest kernels
should compile ok with the newest gcc, but do not use them in production"
and "do not use pgcc". My experiences are good - I even ran 2.0.35
compiled with an older pgcc for couple of months with no hick-ups. What do
you people use? egcs-1.1.2?
-- v --
[EMAIL PROTECTED]
--
=- To unsubscribe, email [EMAIL PROTECTED] with the -=
=- body of "unsubscribe linux-abit". -=
