2.6.21.1 fails to suspend/resume to disk (sort of)

2.6.21.1 fails to suspend/resume to disk (sort of)

[Romano Giannetti]

Hi, 

I am trying to upgrade from my ubuntu kernel 2.6.17.3 to 2.6.21.1 and I
have quite a bunch of problems with suspend/resume, both to disk and to
ram. I have reported the s2ram problem in a separate message (search for
a follow-up to the thread with subject "[patch 00/69] -stable review".
http://lkml.org/lkml/2007/5/23/38

Now for the suspend to disk. With default setting (/sys/power/disk ==
platform) my laptop (vaio PGC-FX701) fails to poweroff; it stays with
backlight on, a white screen, and if I power it off manually afterward
it did not resume.

By doing echo shutdown > /sys/power/disk the laptop *do* power off (it's
very slow, but I think that's related with the 60 secs delay commented
in the other post). Then it resume --- text console did resume ok ---
but the virtual console with X do not come back. I had to do a
ctrl-alt-backspace thing. 

Any hints? More data?

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Rafael J. Wysocki]

Hi,

On Wednesday, 23 May 2007 09:25, Romano Giannetti wrote:
> Hi, 
> 
> I am trying to upgrade from my ubuntu kernel 2.6.17.3 to 2.6.21.1 and I
> have quite a bunch of problems with suspend/resume, both to disk and to
> ram. I have reported the s2ram problem in a separate message (search for
> a follow-up to the thread with subject "[patch 00/69] -stable review".
> http://lkml.org/lkml/2007/5/23/38

Please see http://bugzilla.kernel.org/show_bug.cgi?id=8456
That seems to resemble the symptoms you describe.

> Now for the suspend to disk. With default setting (/sys/power/disk ==
> platform) my laptop (vaio PGC-FX701) fails to poweroff; it stays with
> backlight on, a white screen, and if I power it off manually afterward
> it did not resume.
> 
> By doing echo shutdown > /sys/power/disk the laptop *do* power off (it's
> very slow, but I think that's related with the 60 secs delay commented
> in the other post). Then it resume --- text console did resume ok ---
> but the virtual console with X do not come back. I had to do a
> ctrl-alt-backspace thing. 
> 
> Any hints? More data?

Could you please test the 2.6.22-rc2 kernel?  Also, if it still doesn't work
properly, could you apply the hibernation/suspend patchset available from
http://www.sisk.pl/kernel/hibernation_and_suspend/2.6.22-rc2/patches
on top of it and see if that helps?

Rafael

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Romano Giannetti]

On Wed, 2007-05-23 at 11:05 +0200, Rafael J. Wysocki wrote:
> Hi,
> 
> On Wednesday, 23 May 2007 09:25, Romano Giannetti wrote:
> > http://lkml.org/lkml/2007/5/23/38
> 
> Please see http://bugzilla.kernel.org/show_bug.cgi?id=8456
> That seems to resemble the symptoms you describe.
> 

No, I don't think. The delay I observe is on resume (suspend is very
fast). Moreover, I have no SATA. It seems that my problem came from
pcmcia.

> > Now for the suspend to disk. With default setting (/sys/power/disk ==
> > platform) my laptop (vaio PGC-FX701) fails to poweroff; it stays with
> > backlight on, a white screen, and if I power it off manually afterward
> > it did not resume.
> > 
> > By doing echo shutdown > /sys/power/disk the laptop *do* power off (it's
> > very slow, but I think that's related with the 60 secs delay commented
> > in the other post). Then it resume --- text console did resume ok ---
> > but the virtual console with X do not come back. I had to do a
> > ctrl-alt-backspace thing. 
> > 
> > Any hints? More data?
> 
> Could you please test the 2.6.22-rc2 kernel?  Also, if it still doesn't work
> properly, could you apply the hibernation/suspend patchset available from
> http://www.sisk.pl/kernel/hibernation_and_suspend/2.6.22-rc2/patches
> on top of it and see if that helps?
> 

Uf, I will try to find the time. I am on my way to try to compile
2.6.21.2 now. Problem is that a kernel compile "the ubuntu way" last
hours here. If there is an expert on make-kpkg here, I could use some
advise. I'd prefer to test patches over a 2.6.21 stable release, to
avoid mixing symptoms from an experimental -rc...

Romano

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Rafael J. Wysocki]

On Wednesday, 23 May 2007 11:57, Romano Giannetti wrote:
> On Wed, 2007-05-23 at 11:05 +0200, Rafael J. Wysocki wrote:
> > Hi,
> > 
> > On Wednesday, 23 May 2007 09:25, Romano Giannetti wrote:
> > > http://lkml.org/lkml/2007/5/23/38
> > 
> > Please see http://bugzilla.kernel.org/show_bug.cgi?id=8456
> > That seems to resemble the symptoms you describe.
> > 
> 
> No, I don't think. The delay I observe is on resume (suspend is very
> fast). Moreover, I have no SATA. It seems that my problem came from
> pcmcia.

Hmm, there also is this patch: http://lkml.org/lkml/2007/5/22/434

> > > Now for the suspend to disk. With default setting (/sys/power/disk ==
> > > platform) my laptop (vaio PGC-FX701) fails to poweroff; it stays with
> > > backlight on, a white screen, and if I power it off manually afterward
> > > it did not resume.
> > > 
> > > By doing echo shutdown > /sys/power/disk the laptop *do* power off (it's
> > > very slow, but I think that's related with the 60 secs delay commented
> > > in the other post). Then it resume --- text console did resume ok ---
> > > but the virtual console with X do not come back. I had to do a
> > > ctrl-alt-backspace thing. 
> > > 
> > > Any hints? More data?
> > 
> > Could you please test the 2.6.22-rc2 kernel?  Also, if it still doesn't work
> > properly, could you apply the hibernation/suspend patchset available from
> > http://www.sisk.pl/kernel/hibernation_and_suspend/2.6.22-rc2/patches
> > on top of it and see if that helps?
> > 
> 
> Uf, I will try to find the time. I am on my way to try to compile
> 2.6.21.2 now. Problem is that a kernel compile "the ubuntu way" last
> hours here. If there is an expert on make-kpkg here, I could use some
> advise. I'd prefer to test patches over a 2.6.21 stable release, to
> avoid mixing symptoms from an experimental -rc...

OK, I'll try to see which ones apply on top of 2.6.21 cleanly.

Greetings,
Rafael

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Romano Giannetti]

On Wed, 2007-05-23 at 18:57 +0200, Rafael J. Wysocki wrote:
> On Wednesday, 23 May 2007 11:57, Romano Giannetti wrote:
> > On Wed, 2007-05-23 at 11:05 +0200, Rafael J. Wysocki wrote:
> > > Hi,
> > > 
> > > On Wednesday, 23 May 2007 09:25, Romano Giannetti wrote:
> > > > http://lkml.org/lkml/2007/5/23/38
> > > 
> > > Please see http://bugzilla.kernel.org/show_bug.cgi?id=8456
> > > That seems to resemble the symptoms you describe.
> > > 
> > 
> > No, I don't think. The delay I observe is on resume (suspend is very
> > fast). Moreover, I have no SATA. It seems that my problem came from
> > pcmcia.
> 
> Hmm, there also is this patch: http://lkml.org/lkml/2007/5/22/434
> 
Neither this is related. My report was for a delay of 60 seconds on
resume, not an infinite delay. And I have no bluetooth here.

I will try to add some data to the bug and then will open a bugzilla
report.

Romano

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Rafael J. Wysocki]

On Thursday, 24 May 2007 07:20, Romano Giannetti wrote:
> On Wed, 2007-05-23 at 18:57 +0200, Rafael J. Wysocki wrote:
> > On Wednesday, 23 May 2007 11:57, Romano Giannetti wrote:
> > > On Wed, 2007-05-23 at 11:05 +0200, Rafael J. Wysocki wrote:
> > > > Hi,
> > > > 
> > > > On Wednesday, 23 May 2007 09:25, Romano Giannetti wrote:
> > > > > http://lkml.org/lkml/2007/5/23/38
> > > > 
> > > > Please see http://bugzilla.kernel.org/show_bug.cgi?id=8456
> > > > That seems to resemble the symptoms you describe.
> > > > 
> > > 
> > > No, I don't think. The delay I observe is on resume (suspend is very
> > > fast). Moreover, I have no SATA. It seems that my problem came from
> > > pcmcia.
> > 
> > Hmm, there also is this patch: http://lkml.org/lkml/2007/5/22/434
> > 
> Neither this is related. My report was for a delay of 60 seconds on
> resume, not an infinite delay. And I have no bluetooth here.
> 
> I will try to add some data to the bug and then will open a bugzilla
> report.

OK, please add my address to the bugzilla entry's CC list.

Greetings,
Rafael

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Romano Giannetti]

On Thu, 2007-05-24 at 21:27 +0200, Rafael J. Wysocki wrote:
> On Thursday, 24 May 2007 07:20, Romano Giannetti wrote:
> > On Wed, 2007-05-23 at 18:57 +0200, Rafael J. Wysocki wrote:
> > > On Wednesday, 23 May 2007 11:57, Romano Giannetti wrote:
> > > > On Wed, 2007-05-23 at 11:05 +0200, Rafael J. Wysocki wrote:
> > > > > Hi,
> > > > > 
> > > > > On Wednesday, 23 May 2007 09:25, Romano Giannetti wrote:
> > > > > > http://lkml.org/lkml/2007/5/23/38
> > > > > 
> > > > > Please see http://bugzilla.kernel.org/show_bug.cgi?id=8456
> > > > > That seems to resemble the symptoms you describe.
> > > > > 
> > > > 
> > > > No, I don't think. The delay I observe is on resume (suspend is very
> > > > fast). Moreover, I have no SATA. It seems that my problem came from
> > > > pcmcia.
> > > 
> > > Hmm, there also is this patch: http://lkml.org/lkml/2007/5/22/434
> > > 
> > Neither this is related. My report was for a delay of 60 seconds on
> > resume, not an infinite delay. And I have no bluetooth here.
> > 
> > I will try to add some data to the bug and then will open a bugzilla
> > report.
> 
> OK, please add my address to the bugzilla entry's CC list.
> 

Good news. I tried it in 2.6.21.2, vanilla, without the pcmcia 3COM card
(see the other thread) and it works. With the default platform config. 

Thanks.

Romano

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Antonino Ingargiola]

Hi,

2007/5/23, Romano Giannetti <romano.giannetti@gmail.com>:
[cut]
> Uf, I will try to find the time. I am on my way to try to compile
> 2.6.21.2 now. Problem is that a kernel compile "the ubuntu way" last
> hours here. If there is an expert on make-kpkg here, I could use some
> advise. I'd prefer to test patches over a 2.6.21 stable release, to
> avoid mixing symptoms from an experimental -rc...

Is not the ubuntu or debian way to take hours to compile. Is that you
have to trim the config and enable the only modules you need for your
hardware, then a rebuild cycle could last 10-15 min.

You will spend one hour once to cut down the configuration, but then
you can reuse the .config for new kernels too with make oldconfig.

To find the minimum configuration for you hardware use lspci -v,
menuconfig help, and look at chapter 7 and 8 of Linux Kernel in a
Nutshell:

http://www.kroah.com/lkn/

Also, pasting "lspci -n" output in the following page can be of some help:

http://kmuto.jp/debian/hcl/

Hope this helps.

> Romano

    ~ Antonio

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Paolo Ornati]

On Thu, 24 May 2007 10:01:21 +0200
"Antonino Ingargiola" <tritemio@gmail.com> wrote:

> Is not the ubuntu or debian way to take hours to compile. Is that you
> have to trim the config and enable the only modules you need for your
> hardware, then a rebuild cycle could last 10-15 min.

Even 3-5 min if you strip it down a lot and with good hardware :)

Also ccache helps.

-- 
	Paolo Ornati
	Linux 2.6.22-rc2-gd2579053 on x86_64

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Romano Giannetti]

On Thu, 2007-05-24 at 10:01 +0200, Antonino Ingargiola wrote:
> Hi,
>
> 2007/5/23, Romano Giannetti <romano.giannetti@gmail.com>:
> [cut]
> > Uf, I will try to find the time. I am on my way to try to compile
> > 2.6.21.2 now. Problem is that a kernel compile "the ubuntu way" last
> > hours here. If there is an expert on make-kpkg here, I could use some
> > advise.

> Is not the ubuntu or debian way to take hours to compile. Is that you
> have to trim the config and enable the only modules you need for your
> hardware, then a rebuild cycle could last 10-15 min.

That is a problem, yes. But I was referring to the need of a make-kpkg
clean every time you recompile changing the EXTRAVERSION, wich forces a
make clean. I am trying to discover how to build manually and initramfs
that works...

>
> You will spend one hour once to cut down the configuration, but then
> you can reuse the .config for new kernels too with make oldconfig.
>

Yep, done that. The kernel does not boot (in the initramfs, there is
a /dev/hda1 but it refuses to mount; and yes, I have ext3 and ext2
compiled in, checked a lot of things, no joy. Will retry).

> To find the minimum configuration for you hardware use lspci -v,
> menuconfig help, and look at chapter 7 and 8 of Linux Kernel in a
> Nutshell:
>
> http://www.kroah.com/lkn/
>
> Also, pasting "lspci -n" output in the following page can be of some help:
>
> http://kmuto.jp/debian/hcl/
>
> Hope this helps.
>

What would be great is a tool that generate a config starting from the
lsmod list, to start from there and _enable_ more thing... but I know,
I'm dreaming.

Romano


--
Romano Giannetti --- romano.giannetti@gmail.com
Sorry for the following disclaimer, it's attached by our otugoing server
and I cannot shut it up.



--
La presente comunicación tiene carácter confidencial y es para el exclusivo uso del destinatario indicado en la misma. Si Ud. no es el destinatario indicado, le informamos que cualquier forma de distribución, reproducción o uso de esta comunicación y/o de la información contenida en la misma están estrictamente prohibidos por la ley. Si Ud. ha recibido esta comunicación por error, por favor, notifíquelo inmediatamente al remitente contestando a este mensaje y proceda a continuación a destruirlo. Gracias por su colaboración.

This communication contains confidential information. It is for the exclusive use of the intended addressee. If you are not the intended addressee, please note that any form of distribution, copying or use of this communication or the information in it is strictly prohibited by law. If you have received this communication in error, please immediately notify the sender by reply e-mail and destroy this message. Thank you for your cooperation.

Re: 2.6.21.1 fails to suspend/resume to disk (sort of)

[Ray Lee]

On 5/24/07, Romano Giannetti <romanol@upcomillas.es> wrote:
> On Thu, 2007-05-24 at 10:01 +0200, Antonino Ingargiola wrote:
> > 2007/5/23, Romano Giannetti <romano.giannetti@gmail.com>:
> > [cut]
> > > Uf, I will try to find the time. I am on my way to try to compile
> > > 2.6.21.2 now. Problem is that a kernel compile "the ubuntu way" last
> > > hours here. If there is an expert on make-kpkg here, I could use some
> > > advise.
>
>
> > Is not the ubuntu or debian way to take hours to compile. Is that you
> > have to trim the config and enable the only modules you need for your
> > hardware, then a rebuild cycle could last 10-15 min.
>
> That is a problem, yes. But I was referring to the need of a make-kpkg
> clean every time you recompile changing the EXTRAVERSION, wich forces a
> make clean. I am trying to discover how to build manually and initramfs
> that works...

Yeah, the make-kpkg clean is highly annoying, and eventually made me
give up on using that tool. Instead, I've been using the deb-pkg
target that ships with the kernel. When bisecting, I use a script like
the one below to test a new kernel with a minimum of interaction. It
should give you a good idea of how to do all the steps by hand.

And yeah, paring down the .config is critical :-).

#!/bin/bash
sudo -v
make oldconfig
fakeroot make -j2 deb-pkg
sudo dpkg -i ../linux-2.6.20_2.6.20_amd64.deb
sudo update-initramfs -c -k 2.6.20