pmdisk/swusp1 (merged) with 2.6.8-rc2-mm1

pmdisk/swusp1 (merged) with 2.6.8-rc2-mm1

[Kevin Fenzi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


Greetings. 

I am a happy software suspend2 user, but with the recent merge of
pmdisk and swsusp1, I thought I would give it a try and see how far
along it's come. 

Using 2.6.8-rc2-mm1 (that has the merged pmdisk/swsusp) I booted
single user and unloaded all modules, then started a hibernate. 

It gets to: 

PM: Attempting to suspend to disk.

and then hangs. Machine has to be hard power cycled. 
Looking at the code the problem appears to be that my laptop is
reporting that it has "S4_bios" support. It doesn't, but swsup1 sees
the S4_bios in /proc/acpi/sleep and tries to call that instead of
swsusp1. 

Is there any way to disable detection of S4_bios?

Is there any way to get swsusp1 to use it's S4 instead of S4_bios?

I guess I can recompile with the calls to S4_bios removed and see how
things go. 

kevin
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>

iD8DBQFBB/Y53imCezTjY0ERAhleAJsEaaiXoC05j9Wm/w51G9YSKPEwmwCdE/Ux
0dxA8IXceHabVDCw5BJTqfI=
=2wIb
-----END PGP SIGNATURE-----

Re: pmdisk/swusp1 (merged) with 2.6.8-rc2-mm1

[Kevin Fenzi, Kevin Fenzi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

>>>>> "Kevin" == Kevin Fenzi <kevin-kernel@scrye.com> writes:

Kevin> Greetings.

Kevin> I am a happy software suspend2 user, but with the recent merge
Kevin> of pmdisk and swsusp1, I thought I would give it a try and see
Kevin> how far along it's come.

Kevin> Using 2.6.8-rc2-mm1 (that has the merged pmdisk/swsusp) I
Kevin> booted single user and unloaded all modules, then started a
Kevin> hibernate.

Kevin> It gets to:

Kevin> PM: Attempting to suspend to disk.

Kevin> and then hangs. Machine has to be hard power cycled. Looking at
Kevin> the code the problem appears to be that my laptop is reporting
Kevin> that it has "S4_bios" support. It doesn't, but swsup1 sees the
Kevin> S4_bios in /proc/acpi/sleep and tries to call that instead of
Kevin> swsusp1.

Kevin> Is there any way to disable detection of S4_bios?

Kevin> Is there any way to get swsusp1 to use it's S4 instead of
Kevin> S4_bios?

Kevin> I guess I can recompile with the calls to S4_bios removed and
Kevin> see how things go.

To followup on myself here, this can be set with: 

echo "shutdown" > /sys/power/disk

to tell it to use OS shutdown instead of firmware mode letting the
bios handle things. 

Using that, I get a suspend to happen ok, but on resume I get a kernel
panic. ;( 

Pavel/Patrick: Does your current implementation handle himem and/or
preempt? 

suspend2 handles both, so I had them enabled in my config, but I
wonder if that isn't the issue. The panic is in memory allocation it
seems like. 

I can re-compile with both those off. 

kevin
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>

iD8DBQFBCAhx3imCezTjY0ERAoYRAKCQ022MJ+znKAsK0YsiCypLEeRvMQCgkHob
6LQW1i/bqgblQSZJikohdjg=
=NmQD
-----END PGP SIGNATURE-----

Re: pmdisk/swusp1 (merged) with 2.6.8-rc2-mm1

[Pavel Machek]

Hi!

> To followup on myself here, this can be set with: 
> 
> echo "shutdown" > /sys/power/disk
> 
> to tell it to use OS shutdown instead of firmware mode letting the
> bios handle things. 
> 
> Using that, I get a suspend to happen ok, but on resume I get a kernel
> panic. ;( 
> 
> Pavel/Patrick: Does your current implementation handle himem and/or
> preempt? 
> 
> suspend2 handles both, so I had them enabled in my config, but I
> wonder if that isn't the issue. The panic is in memory allocation it
> seems like. 
> 
> I can re-compile with both those off. 

There was bug in himem support, it is now fixed.
Preempt still leads to ugly warning; I'll eventually fix it.
-- 
64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms