Re: [HEADSUP] amd64 suspend/resume code to be comitted

On Sunday 29 March 2009 04:13 pm, Gustau Perez wrote:
> > Then, it is something else, e.g., acpi_video(4).  You can try
> > setting debug.acpi.disabled="video" in /boot/loader.conf for
> > example.
> >
> > Jung-uk Kim
>
>     Sorry for the delay. Tried loading acpi_video in
> /boot/loader.conf. After having a terminal, I checked the available
> mibs with sysctl -a | grep acpi.video. Only get
> hw.acpi.reset_video. The man page didn't throw the waited ray of
> light :)

So, its BIOS does not support ACPI video extension. :-)

>    I also tried suspending from the terminal, without X's running.
> I tried vbetool (dpms off/on and post). I wasn't brave enought to
> try vbestate save and restore.

It seems many modern graphics boards do not support this function any 
way. :-(

>    Finally I decided it was time to use all the arsenal available
> :) so I decided to remove everything not necessary. So usb and
> friends, firewire, cbb and pccard, if_bge, if_rum, snd_hda and
> nouveau were removed from the kernel. Tried going single user mode,
> remove them and acpiconf -s 3. Guess what happened :)

Let me guess, it worked? ;-)

>    Checking messages doesn't show anything unsual. But if you want
> them or if I can try others things to debug, let me know.

If it worked fine, you can try each driver and tell us what breaks 
suspend/resume process.  If you can find the culprit, you can file a 
PR and you may able to work around it by adding 'kldunload whatever' 
and 'kldload whatever' from rc.suspend and rc.resume respectively.  
BTW, bge(4) is already known for suspend/resume problem if my memory 
serves.

Jung-uk Kim