Re: [ACPI] Call for help: list of machines with working S3 (fwd)

Re: [ACPI] Call for help: list of machines with working S3 (fwd)

[Pavel Machek]

Hi!

Sorry, I was too fast with my  mail client. Please cc: Len in your
replies.


To: Len Brown <len.brown@intel.com>
Subject: Re: [ACPI] Call for help: list of machines with working S3
X-Warning: Reading this can be dangerous to your mental health.

Hi!

> > I'm not sure if you can push the whole industry at once.
> 
> The goal is to know what to tell the system vendors
> interested in supporting Linux what they should do
> with their BIOS on future platforms.
> 
> I believe our message should be:
> 1. BIOS should save/restore video in S3

Actually, that'd expect too much of BIOS writers. I believe right
solution is "POST video as you do during normal boot in S3 resume".

> 2. Use Intel's ACPICA ASL compiler -- if not for production,
> then at least as a static source code checker for validation.

3. Try to boot linux (here's live cd). If it complains about bios bugs
(dmesg | grep ...), try to see if it is not indeed your bug.
-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!

----- End forwarded message -----

-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!

Re: [ACPI] Call for help: list of machines with working S3 (fwd)

[Carl-Daniel Hailfinger]

Pavel Machek schrieb:
> 
>>>I'm not sure if you can push the whole industry at once.
>>
>>The goal is to know what to tell the system vendors
>>interested in supporting Linux what they should do
>>with their BIOS on future platforms.
>>
>>I believe our message should be:
>>1. BIOS should save/restore video in S3
> 
> Actually, that'd expect too much of BIOS writers. I believe right
> solution is "POST video as you do during normal boot in S3 resume".

Why not leave the choice to the BIOS writers? If some are able to
save/restore video state by themselves, we shoudln't stop them.
As long as the state after resume accepts setting the mode without
lockup, we should be fine.


>>2. Use Intel's ACPICA ASL compiler -- if not for production,
>>then at least as a static source code checker for validation.
> 
> 
> 3. Try to boot linux (here's live cd). If it complains about bios bugs
> (dmesg | grep ...), try to see if it is not indeed your bug.

That could even be automated more. A live cd which boots, performs
a few tests, displays the results on screen and offers an option
to save these results to usbstick/network/disk/whatever.
After saving the boot results, it will perform a S3 suspend and
resume and display/save the results of that, too. If Intel have
enough money, they could provide the functionality on a hard
disk and it would have the benefit that S4 could be tested as well.
If the disk has a FAT32 partition, the results can be saved there.


Regards,
Carl-Daniel
-- 
http://www.hailfinger.org/