Bad CRC [was Debian freezes]

[Robert Storey <y2kbug@ms25.hinet.net>]

On Sat, 8 Mar 2003 06:49:50 -0500
Subba Rao <subba9@cablespeed.com> wrote:
 
> I am clueless as to find why the X is freezing.  Sometimes it uses 99%
> CPU while other times it simple freezes the system. 
> Cntrl+Alt+Backspace will do nothing.  I have to power down the system
> and in worst cases (when the system did not respond to the "power"
> button) I had to pull the power cord, to restart the system.

Sorry to hijack your thread, Subba, but this might be the source of your
problem...

I started experiencing freezing too, and at first I blamed my Debian
installation. I did a reinstall, but the problem continued, and indeed
seemed to grow gradually worse. I also have a (seldom-used) Windows
partition - I started using it just to see what would happen, and sure
enough, experienced the freezing there too.

I noticed during Debian's boot-up that an error message would flash by
(too fast to read). I explored his further, like this:

   bob@sonic:~$ dmesg | grep error
   hda: dma_intr: error=0x84 { DriveStatusError BadCRC }
   hda: dma_intr: error=0x84 { DriveStatusError BadCRC }
   hda: dma_intr: error=0x84 { DriveStatusError BadCRC }
   hda: dma_intr: error=0x84 { DriveStatusError BadCRC }

After doing some searching on Google, I found that a lot of people had
reported these errors. The consensus of opinion seemed to be that this
could be a fading hard disk or bad controller, but many people suggested
installing a new hard disk cable before trashing the hard disk itself. I
did install a new cable, and the freezing problem has gone away.
However, the "BadCRC" errors are still being reported during bootup.

There were suggestions to turn off dma with hdparm, like this:

# hdparm -d0 /dev/hda

I tried that but it didn't stop the error messages, so I've turned dma
back on again.

# hdparm -d1 /dev/hda

Another suggestion was to turn off dma by inserting this into GRUB's
menu.lst configuration file:

hda=nodma

This led to a "kernel panic" error. I removed the offending line from
menu.lst, and my system boots up again as normal, but the CRC error
messages remain. The machine is usable and no more freezes since I
replaced the cable. But I wonder if I'm taking a risk just doing nothing
- should I be prepared to deploy my hard disk for use as a paperweight,
or should I just muddle along and not worry about this? I'd be
interested to know what others who have encountered these error messages
and/or freezing problem have to say.

regards,
Robert