Experiences with new PCI code

[Nate Lawson]

On Wed, 14 Apr 2004, M. Warner Losh wrote:
> In message: <20040414054152.GL58667 at pcwin002.win.tue.nl>
>             Stijn Hoop <stijn at win.tue.nl> writes:
> : On Tue, Apr 13, 2004 at 10:27:46PM -0700, Nate Lawson wrote:
> : > On Wed, 14 Apr 2004, Stijn Hoop wrote:
> : > > On Tue, Apr 13, 2004 at 09:23:39PM -0700, Nate Lawson wrote:
> : > > > The other major issue that I can't ever solve is drivers not correctly
> : > > > resuming themselves.  That is up to the driver authors and a lot are
> : > > > incomplete.
> : > >
> : > > Is there a list somewhere so that we end users don't report lots of
> : > > known problems?
> : >
> : > That's item one on the list:  make a list.  Care to start one?  :)
> :
> : Errrr... Can I tell by reading / grepping the source?
>
> Typically no.  You tell because on your otherwise working laptop, the
> Foo device doesn't resume.

Yep.  You can grep for drivers that don't have device_suspend in their
methods list.  But most have one.  The usual case is that the suspend
method is incomplete for some chipsets although it works ok on others.
For instance, snd_ich works pretty well for me but will stop generating
interrupts if shared with if_an on resume.  But for others its sample rate
is off after resume.  Now this problem may have been fixed by Warner's
power code so it would be a good time to test those drivers because if the
problem is still there, it's a driver problem and there's not much more
to be done outside the driver.

-Nate