Bug#444182: pata disabled in ata_piix
Hi,
On Wed, Feb 04, 2009 at 06:14:56PM +0100, Bastian Blank wrote:
> > I was looking into the same hotswap issue and found to my surprise that
> > the debian kernel has the needed support _removed in the kernel sources_.
>
> Supporting and not breaking thousands of machines is more important that
> supporting one.
I agree with your statement. However, I am puzzled that you claim that
this only affects one machine. This patch removes quite a lot of IDs,
which already makes it a lot of machines, not only one and certainly not
only mine.
I agree that it would be best if removing this patch again would not
cause harm to other users. Could we try to eliminate this problem
instead of ignoring it?
If I understood correctly, the situation is:
- There are two drivers which claim to support the same hardware. This
should be fine in general.
- They implement this slitely differently (different device names). This
should also be fine in general.
- The main problem seems to be that the order at which those drivers
are assigned to the hardware is not fixed / cannot be specified.
(See bug #419458).
Do I understand this correctly? If I dont, it would be very nice to give
me a hint where I can find information about the reason why this patch
is included.
Assuming I got it right, the question is, how the assignement of drivers
can be specified or at least be fixed (as in not changing), fixing bug
#419458 without removing support for PATA from ata_piix. This patch was
introduced in 2.6.20. Are you sure that the original problem is still
present in 2.6.26 or later kernels?
> So a bug, but will not be fixed soon. Tagging correctly.
Is there anything I could do to get it fixed? I assume upstream or at
least other major linux distributions would have had the same problem
and must have found a way around it?
thanks, Frank Loeffler
Reply to: