Re: More on the IDE drive problems after resume. (2.1.122-2.1.125)

Andre M. Hedrick (hedrick@astro.dyer.vanderbilt.edu)
Thu, 15 Oct 1998 00:15:51 -0500 (CDT)


On 14 Oct 1998, Steve Dunham wrote:

> Steve Dunham <dunham@cse.msu.edu> writes:
>
> > After I resume my laptop, I'm getting the following error messages:
> >
> > kernel: ide0: unexpected interrupt, status=0x58, count=1
> > kernel: hda: status error: status=0x58 { DriveReadh SeekComplete DataRequest }
> > kernel: hda: drive not ready for command
>
> (Along with this I get a lot of segmentation faults, in any program
> that isn't in the disk cache.)
>
> More on the IDE/APM problems:
>
> For me these problems began with the 2.1.122 kernel. I managed to
> track this down to the fact that 2.1.122 enables multiple mode by
> default.

I made the changes that went into 2.1.122.

> If I boot a 2.1.125 kernel, and use hdparm to turn off multimode, then
> the problem goes away.
>
> Also, if I lower multimode to "8", the problem seems to go away. (The
> newer kernels set it to the drive's maximum of 16.) With multimode
> set to 16, the drive does work fine, until I suspend it for the first
> time.

The ide-driver now tests, learns, and sets "multimode" to the max value
if the hardware is compatable. Since the problem happens after you
suspend the drive and then wake it back up, there may be a combination of
problems to resolve.

Does APM need to disable all drive features down to PIO 0 before sleeping?
If so, then APM will need to have a kernel memset buffer to hold the
current drive settings prior to spindown. With spin-up being in PIO 0
and DMA transfers suspended, it sounds like this is acceptable to APM.
If this is the case, final procedures would be to return the drive and
hardware to pre-sleep status.

> I'm not sure if the problem is with a buggy APM implementation in the
> BIOS, a buggy hard drive, or a problem with the IDE driver in the
> kernel.
>
> The drive identifies itself as an "IBM-DSOA-20810". The machine is a
> Dell Latitude XPi. A few other people have noted a similar problem on
> the kernel list.
>
>
> Steve
> dunham@cse.msu.edu
>
>
> -
> 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.tux.org/lkml/
>

Cheers,
Andre Hedrick
The IDE-FNG for Linux
The APC UPS Specialist for Linux

http://www.dyer.vanderbilt.edu/server/udma/
http://www.dyer.vanderbilt.edu/server/apcupsd/

-
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.tux.org/lkml/