[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#463833: Bug#419458: ata_piix and piix claim same pci ids



Hi,

I am writing to you because you reported bug #419458:

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=419458

<cite>

Package: linux-image-2.6.20-1-686
Version: 2.6.20-1
Severity: grave

Upon upgrading to this kernel, about 1/3rd of the time when I boot,
the internal drive of my laptop becomes /dev/sda. The rest of the time,
it remains /dev/hda.

</cite>

As consequence of this, pata support for the ata_piix driver was
disabled in the Debian kernels and still is by means of a patch to the
Debian kernel sources. The piix driver is supposed to support that
hardware alone in Debian.

The problem now is that some hardware does not work with the piix
driver or does not provide full functionality (hot swapping): bugs
444182 and 463833, now merged. As consequence of the patch, those
users have to either (un)patch the Debian kernel after each package
update or use upstream kernels. Both is not ideal and a solution should
be found.

Before we now all go and dig into the issue it would be
interesting to know if the original bug is still present in recent
kernels. The bug was reported for 2.6.22-4 and about one and a half year
ago. Would it be possible for you to try to reproduce the bug with a
recent upstream kernel or a recent Debian kernel which does not contain
this patch?
As far as I can see, both drivers still claim to support some identical
ids. However, that should not be a problem. There are other examples of
different drivers for identical hardware.
The problem here seemed to be that the rule which driver gets
assigned to the hardware (in case both drivers are compiled in) is not
deterministic.

I CCed the bug report to keep a record of this in the data base.

I appologize if something in this email is not correct. Please correct
me in that case. I am still in the stage of trying to fully understand
the problem.

thanks, Frank Loeffler




Reply to: