suspend/resume improved?

[Nate Lawson]

Eric Anderson wrote:
> Nate Lawson wrote:
> 
>> I made a commit over the weekend to -current that may fix some 
>> suspend/resume driver issues for people.  Please test and let me know 
>> if it helps.
> 
> Will these changes also apply to -stable users?

In about a day when I MFC.

> Nate - I'd like to help you debug some S3 issues - I've read the ACPI 
> debugging page and I'll start playing with what I can - but is there 
> anything in particular I should work with?  I'm running 5.3-STABLE with 
> the acpi patches that were posted about a month ago to the -acpi list.

It's pretty straightforward but arduous work.  First strip down your 
system, removing all drivers except for the hard drive and keyboard (no 
USB, network, etc.)  Don't run X.  Try S3.  If it works, add back in 
drivers until it fails.

If it fails, try to find out where it fails.  Does it truly make it to 
sleep or does it immediately resume?  Add the beep code Warner posted a 
while back to the resume code and see if you get a beep.  If you get a 
beep but a dead system, it's something in driver resume.  Try to enable 
the network driver and ssh into your system after resume.  If it works, 
it's a video driver problem and you need to mess with resuming the VESA 
BIOS.  Don't run DRM with X.

A lot of this is in the ACPI section of the handbook (which no one 
appears to have seen for some reason).  More info about specific issues 
can be found on my web page:

http://www.root.org/~nate/freebsd/acpi-todo.html

-- 
Nate