Re: 2.6.18 suspend regression on Intel Macs

["Frédéric Riss" <frederic.riss@gmail.com>]

Le mardi 10 octobre 2006 à 16:53 -0700, Linus Torvalds a écrit :
> 
> On Wed, 11 Oct 2006, Frédéric Riss wrote:
> >
> > I was about to send a patch doing exactly the same. It fixes the issue
> > for me. Thanks.
> 
> Hmm. My Mac Mini doesn't restore properly even with it, but I suspect it's 
> the old DRM "resume AGP in the wrong order" problem.

I should have mentioned that I've tested it by patching a 2.6.18 and not
the current git tree (current git wouldn't boot for me, need to
investigate). Maybe something got pulled recently that introduced new
issues?

Fred.

> When trying to verify that, though, I noticed that if I enable the "keep 
> console active over suspend", then it won't even suspend. It hangs after 
> printing "i801_smbus 0000:00:1f.3: suspend".
> 
> I'm wondering what Pavel does for debugging these things, since the claim 
> was that keeping printk() active would make debugging easier. As it is, it 
> just seems to break suspend exactly because it wants to access devices 
> that are turned off.
> 
> Pavel?
> 
> 		Linus