Re: PIIX oops

Andre Hedrick (andre@suse.com)
Sun, 17 Oct 1999 00:23:29 -0700 (PDT)

There is a problem else where or with your BIOS.
If the BIOS tells linux that there is a disk, it will attempt to find one.
Thus there are plans to ROOT-WAD the BIOS INT13 calls.....
If this is not the case then there is gobal kernel bug in the IDE code and
since this does not happen on any of the boxes that I have, I wonder if
your BIOS is doing funny things.

The patch you offer kind of late in the init processs, IMHO.
By the time chipset code is called, the driver should have "identified"
the device or marked it not present.

"piix_config_drive_for_dma" is called from ide.c's
"ide_register_subdriver" back to the ide_dmaproc and any callouts to the
chipset code.

I will look into this, but I thing something is may be kookie in your
hardware, IMHO.

Andre Hedrick
The Linux IDE guy

On Thu, 14 Oct 1999, Peter Samuelson wrote:

> A few weeks ago I sent linux-kernel an oops trace caused by boot-time
> IDE probing weirdness. Specifically Linux boots with
>
> hdb: non-IDE drive, CHS=622/64/63
>
> (hdb does not, in fact, exist at all), and in recent 2.3.x, this
> misconception causes an oops+hang just after. I still don't know why
> hdb is being "found", but I did manage to prevent the oops:
>
> --- drivers/block/piix.c.orig Fri Oct 8 06:52:02 1999
> +++ drivers/block/piix.c Fri Oct 8 08:52:39 1999
> @@ -166,6 +166,9 @@
> int a_speed = 2 << (drive_number * 4);
> int u_flag = 1 << drive_number;
>
> + if (!id)
> + return ide_dma_off_quietly;
> +
> pci_read_config_word(dev, maslave, &reg4042);
> sitre = (reg4042 & 0x4000) ? 1 : 0;
> pci_read_config_word(dev, 0x44, &reg44);
>
> (That was against 2.3.18ac8 or so, I believe.) So: does anyone want to
> try and debug why my piix3 seems to be telling Linux it has an hdb? I
> am happy to run test patches but I don't know enough about Linux, the
> PC BIOS or IDE to do much more by myself.
>
> --
> Peter Samuelson
> <sampo.creighton.edu!psamuels>
>

-
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/