Linux 2.6.32-rc8 - T500 fails to suspend to RAM

Linux 2.6.32-rc8 - T500 fails to suspend to RAM

[Lior Dotan]

Hi,

Ever since I've upgraded to rc8, my T500 fails to suspend to RAM for the
second time. Suspending for the first time works ok but when I resume and
later try to suspend again the screen gets blank and the moon LED blinks
forever. rc6 & rc7 as well as 2.6.31 worked great with out a problem.

Lior.

Re: Linux 2.6.32-rc8 - T500 fails to suspend to RAM

[Rafael J. Wysocki]

On Tuesday 01 December 2009, Lior Dotan wrote:
> Hi,
> 
> Ever since I've upgraded to rc8, my T500 fails to suspend to RAM for the
> second time. Suspending for the first time works ok but when I resume and
> later try to suspend again the screen gets blank and the moon LED blinks
> forever. rc6 & rc7 as well as 2.6.31 worked great with out a problem.

Well, I don't see any particular change that might cause the symptoms you
describe between -rc7 and -rc8.  If this is 100% reproducible, I guess
bisecting is the only option at the moment.

Thanks,
Rafael

Re: Linux 2.6.32-rc8 - T500 fails to suspend to RAM

[Lior Dotan]

Rafael J. Wysocki wrote:
> On Tuesday 01 December 2009, Lior Dotan wrote:
>   
>> Hi,
>>
>> Ever since I've upgraded to rc8, my T500 fails to suspend to RAM for the
>> second time. Suspending for the first time works ok but when I resume and
>> later try to suspend again the screen gets blank and the moon LED blinks
>> forever. rc6 & rc7 as well as 2.6.31 worked great with out a problem.
>>     
>
> Well, I don't see any particular change that might cause the symptoms you
> describe between -rc7 and -rc8.  If this is 100% reproducible, I guess
> bisecting is the only option at the moment.
>   
It is very easy to reproduce, I will try to bisect it.

Re: Linux 2.6.32-rc8 - T500 fails to suspend to RAM

[Lior Dotan]

Rafael J. Wysocki wrote:
> On Tuesday 01 December 2009, Lior Dotan wrote:
>   
>> Hi,
>>
>> Ever since I've upgraded to rc8, my T500 fails to suspend to RAM for the
>> second time. Suspending for the first time works ok but when I resume and
>> later try to suspend again the screen gets blank and the moon LED blinks
>> forever. rc6 & rc7 as well as 2.6.31 worked great with out a problem.
>>     
>
> Well, I don't see any particular change that might cause the symptoms you
> describe between -rc7 and -rc8.  If this is 100% reproducible, I guess
> bisecting is the only option at the moment.
>   
I tried bisecting it, not sure if it's any good but this is what I got:

ldot@topaz ~/linux/linux-2.6 $ git bisect bad
Bisecting: 0 revisions left to test after this (roughly 0 steps)
[1456edbb1476db735715ddcc7ac306de587024eb] kconfig: Make zconf.y work 
with current bison
ldot@topaz ~/linux/linux-2.6 $ git bisect log
git bisect start
# bad: [648f4e3e50c4793d9dbf9a09afa193631f76fa26] Linux 2.6.32-rc8
git bisect bad 648f4e3e50c4793d9dbf9a09afa193631f76fa26
# good: [156171c71a0dc4bce12b4408bb1591f8fe32dc1a] Linux 2.6.32-rc7
git bisect good 156171c71a0dc4bce12b4408bb1591f8fe32dc1a
# bad: [b6b24c73d56a166a829318a1f06803ff9b54857b] spi: error status 
should be negative
git bisect bad b6b24c73d56a166a829318a1f06803ff9b54857b
# bad: [82abc2a97a4a3c7635c40e6dff571816062ec506] Merge branch 
'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ryusuke/nilfs2
git bisect bad 82abc2a97a4a3c7635c40e6dff571816062ec506
# good: [24dfb2b5867df24ba03b6c4418312e23b1300aa8] Merge branch 
'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
git bisect good 24dfb2b5867df24ba03b6c4418312e23b1300aa8
# bad: [e8e63cbf9a339c972eeb5ccf8777c8067bdfd869] oss: Mark loadhex 
static in hex2hex.c
git bisect bad e8e63cbf9a339c972eeb5ccf8777c8067bdfd869
# bad: [ad0cd62ab632afe9b562b5f09343c21e1521505d] 
arch/alpha/boot/tools/objstrip.c: Mark "usage" static
git bisect bad ad0cd62ab632afe9b562b5f09343c21e1521505d
# bad: [65166571800af0f63f21e5576d7db12d2bf73790] kconfig: Mark various 
internal functions static
git bisect bad 65166571800af0f63f21e5576d7db12d2bf73790

Cheers,
Lior

Re: Linux 2.6.32-rc8 - T500 fails to suspend to RAM

[Rafael J. Wysocki]

On Wednesday 02 December 2009, Lior Dotan wrote:
> Rafael J. Wysocki wrote:
> > On Tuesday 01 December 2009, Lior Dotan wrote:
> >   
> >> Hi,
> >>
> >> Ever since I've upgraded to rc8, my T500 fails to suspend to RAM for the
> >> second time. Suspending for the first time works ok but when I resume and
> >> later try to suspend again the screen gets blank and the moon LED blinks
> >> forever. rc6 & rc7 as well as 2.6.31 worked great with out a problem.
> >>     
> >
> > Well, I don't see any particular change that might cause the symptoms you
> > describe between -rc7 and -rc8.  If this is 100% reproducible, I guess
> > bisecting is the only option at the moment.
> >   
> I tried bisecting it, not sure if it's any good but this is what I got:

Thanks for the effort.

> ldot@topaz ~/linux/linux-2.6 $ git bisect bad
> Bisecting: 0 revisions left to test after this (roughly 0 steps)
> [1456edbb1476db735715ddcc7ac306de587024eb] kconfig: Make zconf.y work 
> with current bison
> ldot@topaz ~/linux/linux-2.6 $ git bisect log
> git bisect start
> # bad: [648f4e3e50c4793d9dbf9a09afa193631f76fa26] Linux 2.6.32-rc8
> git bisect bad 648f4e3e50c4793d9dbf9a09afa193631f76fa26
> # good: [156171c71a0dc4bce12b4408bb1591f8fe32dc1a] Linux 2.6.32-rc7
> git bisect good 156171c71a0dc4bce12b4408bb1591f8fe32dc1a
> # bad: [b6b24c73d56a166a829318a1f06803ff9b54857b] spi: error status 
> should be negative
> git bisect bad b6b24c73d56a166a829318a1f06803ff9b54857b
> # bad: [82abc2a97a4a3c7635c40e6dff571816062ec506] Merge branch 
> 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ryusuke/nilfs2
> git bisect bad 82abc2a97a4a3c7635c40e6dff571816062ec506
> # good: [24dfb2b5867df24ba03b6c4418312e23b1300aa8] Merge branch 
> 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
> git bisect good 24dfb2b5867df24ba03b6c4418312e23b1300aa8
> # bad: [e8e63cbf9a339c972eeb5ccf8777c8067bdfd869] oss: Mark loadhex 
> static in hex2hex.c
> git bisect bad e8e63cbf9a339c972eeb5ccf8777c8067bdfd869
> # bad: [ad0cd62ab632afe9b562b5f09343c21e1521505d] 
> arch/alpha/boot/tools/objstrip.c: Mark "usage" static
> git bisect bad ad0cd62ab632afe9b562b5f09343c21e1521505d
> # bad: [65166571800af0f63f21e5576d7db12d2bf73790] kconfig: Mark various 
> internal functions static
> git bisect bad 65166571800af0f63f21e5576d7db12d2bf73790

This commit is not very likely to be the source of the issue.

Perhaps one of the commits you marked as 'good' was in fact a bad one.

Best,
Rafael

Re: Linux 2.6.32-rc8 - T500 fails to suspend to RAM

[Andy Isaacson]

On Tue, Dec 01, 2009 at 03:53:38PM +0200, Lior Dotan wrote:
> Ever since I've upgraded to rc8, my T500 fails to suspend to RAM for the
> second time. Suspending for the first time works ok but when I resume and
> later try to suspend again the screen gets blank and the moon LED blinks
> forever. rc6 & rc7 as well as 2.6.31 worked great with out a problem.

How are you suspending -- pm-suspend, gnome power manager, closing the
lid, ... ?

Could you compare the dmesg between rc7 and rc8 after the first suspend
and look for any likely differences?  (I use "diff -W160 --side-by-side"
for this.)

-andy

Re: Linux 2.6.32-rc8 - T500 fails to suspend to RAM

[Andy Isaacson]

On Thu, Dec 03, 2009 at 10:23:36AM -0800, Andy Isaacson wrote:
> On Tue, Dec 01, 2009 at 03:53:38PM +0200, Lior Dotan wrote:
> > Ever since I've upgraded to rc8, my T500 fails to suspend to RAM for the
> > second time. Suspending for the first time works ok but when I resume and
> > later try to suspend again the screen gets blank and the moon LED blinks
> > forever. rc6 & rc7 as well as 2.6.31 worked great with out a problem.
> 
> How are you suspending -- pm-suspend, gnome power manager, closing the
> lid, ... ?

For what it's worth, I'm testing suspend-to-ram on Ubuntu 9.10 x86_64
with 2.6.32 final on a T500 and it's been pretty reliable -- I've
suspended by closing the lid at least a dozen times and it hasn't locked
up yet.

Things to check:

1. are you using DMAR (VT-d)?  I'm not, and I had problems on another
   machine (also Core 2 Duo with GM45 graphics) when I tried suspending
   with CONFIG_DMAR=y.
2. are you running into http://bugzilla.kernel.org/show_bug.cgi?id=14484
   or a similar problem?
3. (if you have GMA graphics) Are you up to date on the xorg-video-intel
   driver?  I'm using KMS and xorg-video-intel 2.9.0.  2.8.x is missing
   some relevant bugfixes.  (Of course if you have ATI graphics then
   this isn't relevant.)

-andy

Re: Linux 2.6.32-rc8 - T500 fails to suspend to RAM

[Andy Isaacson]

Re-adding LKML.

On Fri, Dec 04, 2009 at 11:39:25PM +0200, Lior Dotan wrote:
> On Fri, Dec 4, 2009 at 10:24 PM, Andy Isaacson <adi@hexapodia.org> wrote:
> > For what it's worth, I'm testing suspend-to-ram on Ubuntu 9.10 x86_64
> > with 2.6.32 final on a T500 and it's been pretty reliable -- I've
> > suspended by closing the lid at least a dozen times and it hasn't locked
> > up yet.
> 
> I'm using either KDE menu to suspend or hibernate-ram from hibernate-script.
> I will try to use the lid and see if it makes any difference.
> The strange thing is that I tried using rc6 & rc7 and they don't work now
> even though they used to, so I now I need to see what broke it.

I mentioned the lid because of BZ #14484 which is only on Intel GMA plus
using the lid switch to trigger suspend.  Since neither is relevant to
you it's just a red herring. :)

> > 1. are you using DMAR (VT-d)?  I'm not, and I had problems on another
> >   machine (also Core 2 Duo with GM45 graphics) when I tried suspending
> >   with CONFIG_DMAR=y.
>
>  I disabled DMAR but still no luck.

It's getting better but DMAR has still been a little problematic for me
the last few months.  Updating to the very latest BIOS does help
sometimes.

> > 2. are you running into http://bugzilla.kernel.org/show_bug.cgi?id=14484
> >   or a similar problem?
> > 3. (if you have GMA graphics) Are you up to date on the xorg-video-intel
> >   driver?  I'm using KMS and xorg-video-intel 2.9.0.  2.8.x is missing
> >   some relevant bugfixes.  (Of course if you have ATI graphics then
> >   this isn't relevant.)
> >
> I'm using the Radeon GPU so no KMS for me yet.

Sounds like your Radeon or fglrx driver got updated and broke
suspend/resume.

-andy

Re: Linux 2.6.32-rc8 - T500 fails to suspend to RAM

[Lior Dotan]

Andy Isaacson wrote:
> Re-adding LKML.
>
> On Fri, Dec 04, 2009 at 11:39:25PM +0200, Lior Dotan wrote:
>   
>> On Fri, Dec 4, 2009 at 10:24 PM, Andy Isaacson <adi@hexapodia.org> wrote:
>>     
>>> For what it's worth, I'm testing suspend-to-ram on Ubuntu 9.10 x86_64
>>> with 2.6.32 final on a T500 and it's been pretty reliable -- I've
>>> suspended by closing the lid at least a dozen times and it hasn't locked
>>> up yet.
>>>       
>> I'm using either KDE menu to suspend or hibernate-ram from hibernate-script.
>> I will try to use the lid and see if it makes any difference.
>> The strange thing is that I tried using rc6 & rc7 and they don't work now
>> even though they used to, so I now I need to see what broke it.
>>     
>
> I mentioned the lid because of BZ #14484 which is only on Intel GMA plus
> using the lid switch to trigger suspend.  Since neither is relevant to
> you it's just a red herring. :)
>
>   
>>> 1. are you using DMAR (VT-d)?  I'm not, and I had problems on another
>>>   machine (also Core 2 Duo with GM45 graphics) when I tried suspending
>>>   with CONFIG_DMAR=y.
>>>       
>>  I disabled DMAR but still no luck.
>>     
>
> It's getting better but DMAR has still been a little problematic for me
> the last few months.  Updating to the very latest BIOS does help
> sometimes.
>
>   
>>> 2. are you running into http://bugzilla.kernel.org/show_bug.cgi?id=14484
>>>   or a similar problem?
>>> 3. (if you have GMA graphics) Are you up to date on the xorg-video-intel
>>>   driver?  I'm using KMS and xorg-video-intel 2.9.0.  2.8.x is missing
>>>   some relevant bugfixes.  (Of course if you have ATI graphics then
>>>   this isn't relevant.)
>>>
>>>       
>> I'm using the Radeon GPU so no KMS for me yet.
>>     
>
> Sounds like your Radeon or fglrx driver got updated and broke
> suspend/resume.
>   
I got some progess, when using the Intel GMA, the suspend works fine. 
The problem are only when using the radeon driver.
I am trying to use KMS for radeon but it fails to load the firmware and 
the system boots without KMS. I will disable KMS and revert the 
xf86-video-ati driver to the latest released version (I am now using the 
git version).

> -andy
>   
Thanks for your help,
Lior.