On Wed, 9 Jun 2004, Manuel Bouyer wrote:
> On Tue, Jun 08, 2004 at 04:34:47PM +0100, Ian Fry wrote:
[...]
> > However, with all the 2.0F kernels I've tried, wd0 isn't detected at all,
> > and I get a 'lost interrupt' message from the IDE controller. Probing also
> > takes much longer than normal - an extra 5-10 seconds, I think, and the
> > CDROM seems to spin-up twice, rather than once.
>
> Can you set wdcdebug_mask to 0x10, and send me the result ?
> It's easier to do it with a serial console of course, as I assume your
> boot disk isn't functionnal.
> Another way to avoid writting the debug messages by hand is to use a boot
> floppy, or a NFS root.

maybe this is related to the problems I have with current kernels
on my Inspiron8200 (ICH3 based):

cd0 (piixide:0:0:1) is found, but wd0 (piixide:0:0:0), the boot disk,
is not. This does happen with several kernel configs including GENERIC.

However, if I disable piixide using userconf, the disk is found,
and the systems boots with wd0 (as pciide0:0:0).

------------------------------------------------------------
[success, sources Mar 25]
piixide0:0: before reset, st0=0x50, st1=0x0
wdciintr:inactive controllers
piixide0:1: before reset, st0=0x0, st1=0x0
atabusattach: ch_drive_flags 0x0 0x0
piixide0:0:0: after reset, sc=0x1 sn=0x1 cl=0x0 ch=0x0
piixide0:0:1: after reset, sc=0x1 sn=0x1 c=0x14 ch=0xeb
piixide0:0: wdcwait_reset() end, st0=0x50 st1=0x0
piixide0:0: after reset, ret_value=0x3
...

[failure, sources Jun x]
piixide0:0: before reset, st0=0x50, st1=0x0
wdcintr: inactive controller
piixide0:0:0: after reset, sc=0x80 sn=0x80 cl=0x80 ch=0x80
piixide0:0:1: after reset, sc=0x1 sn = 0x1 cl=0x14 ch=0xeb
piixide0:0: wdcwait_reset() end, st0=0x80 st1=0x0
piixide0:0: after_reset, ret_value=0x2
...
------------------------------------------------------------

So I guess the disk is never even probed, since the controller
only reports the second device on channel 0 after reset.


Holger