Re:

Previous thread: none

Next thread: [PATCH] x86: generic versions of find_first_(zero_)bit, convert i386 by Alexander van Heukelum on Monday, March 31, 2008 - 10:15 am. (34 messages)
From: Didier Raboud
Date: Monday, March 31, 2008 - 9:23 am

--Boundary-01=_GAR8HpowvbkYvMZ
Content-Type: text/plain;
  charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Hi,=20

I'm using the latest Debian 2.6.25-rc7 kernel [0] and I have big troubles w=
ith=20
hibernation. With this version, the hibernation (suspend on disk) is=20
sometimes not resuming correctly. The only thing I get is a blinking cursor=
=20
at the top-left of the screen. The problem is that it is not easily=20
reproducible=C2=A0: sometimes it works, sometimes not.

AFAIK, 2.6.24 is not affected, but 2.6.25-rc6 and -rc7 (at least) are.

I'm using a Dell Latitude D630. (lspci attached).

Attached are a 2.6.25-rc7 and a 2.6.24 dmesg's after two successful=20
hibernate/resume cycles.

Regards,=20

Didier Raboud, aka OdyX

N.B. Please keep me CC'ed as I'm not subscribed to the lists.

[0] http://wiki.debian.org/DebianKernel

--Boundary-01=_GAR8HpowvbkYvMZ
Content-Type: text/plain;
  charset="utf-8";
  name="d630_lspci"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
	filename="d630_lspci"

00:00.0 Host bridge: Intel Corporation Mobile PM965/GM965/GL960 Memory Controller Hub (rev 0c)
00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)
00:02.1 Display controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)
00:1a.0 USB Controller: Intel Corporation 82801H (ICH8 Family) USB UHCI Contoller #4 (rev 02)
00:1a.1 USB Controller: Intel Corporation 82801H (ICH8 Family) USB UHCI Controller #5 (rev 02)
00:1a.7 USB Controller: Intel Corporation 82801H (ICH8 Family) USB2 EHCI Controller #2 (rev 02)
00:1b.0 Audio device: Intel Corporation 82801H (ICH8 Family) HD Audio Controller (rev 02)
00:1c.0 PCI bridge: Intel Corporation 82801H (ICH8 Family) PCI Express Port 1 (rev 02)
00:1c.1 PCI bridge: Intel Corporation 82801H (ICH8 Family) PCI Express Port 2 (rev 02)
00:1c.5 PCI bridge: Intel Corporation 82801H (ICH8 Family) PCI ...
From: Rafael J. Wysocki
Subject: Re:
Date: Monday, March 31, 2008 - 9:33 am

Could you please check if the problem is still present if you boot with
acpi_new_pts_ordering kernel command line argument?

Thanks,
Rafael
--

From: Didier Raboud
Subject: Re:
Date: Monday, March 31, 2008 - 10:03 am

Hi.=20

It did not help. I had to test 4 hibernation/resume cycles to see the resum=
e=20
fail.

By the way, I just added acp_new_pts_ordering at the end of the Grub line=20
beginning by "linux (hd0,1)/vmlinuz(...)". Is that correct=A0?

Regards,=20

Didier
From: maximilian attems
Subject: Re:
Date: Monday, March 31, 2008 - 10:24 am

On Mon, Mar 31, 2008 at 07:03:13PM +0200, Didier Raboud wrote:

seems strange you should add it to the kopt line and then kick as root
update-grub

# kopt=root=/dev/sda1 ro
it should land 
kernel          /boot/vmlinuz-2.6.25-rc7-amd64 root=/dev/sda1 ro
--

From: Didier Raboud
Subject: Re:
Date: Monday, March 31, 2008 - 10:36 am

Well... I'm using Grub2 (grub-pc). I altered the boot option at boot time b=
ut=20
forgot to check /proc/cmdline. Will try again and report back.

Regards,=20

Didier
From: Didier Raboud
Subject: Re:
Date: Monday, March 31, 2008 - 11:22 am

Hi,=20

I tried again and ensured that /proc/cmdline 's end was=20
acpi_new_pts_ordering . It did work for 2-3 hibernation, but then failed.

Regards,=20

Didier
From: Rafael J. Wysocki
Subject: Re:
Date: Monday, March 31, 2008 - 1:29 pm

Well, there were quite a lot of suspend-related changes between 2.6.24 and
now.

Hmm, do you have CONFIG_USB_SUSPEND set in .config by chance?

Rafael
--

From: Didier Raboud
Subject: Re:
Date: Monday, March 31, 2008 - 2:03 pm

I do (Debian default option)... I will build a kernel without it and retry.=
=20

Didier
From: Didier Raboud
Subject: Re:
Date: Monday, March 31, 2008 - 11:18 pm

Hi again.=20

I compiled the same kernel with CONFIG_USB_SUSPEND not set. After 6=20
hibernations, I haven't had a problem for now.

Regards,

Didier
From: Didier Raboud
Subject: Re:
Date: Monday, March 31, 2008 - 11:41 pm

Hi again again,

It hung after the 10th hibernation...

I wonder if the problem would not be of Debian's uswsusp binary, because tt=
y1=20
tells me that the image is successfully loaded but I just can't act on my b=
ox=20
afterwards.

There is a bug open in the Debian BTS requesting a packaging of the new=20
upstream version (0.8) but the maintainer is told to be too busy to handle =
it=20
for now.

Don't know...

Regards,=20

Didier
From: Rafael J. Wysocki
Subject: Re:
Date: Tuesday, April 1, 2008 - 2:11 pm

I don't really think so.  s2disk doesn't really do anything that could hang a

Can you please test the in-kernel hibernation, then?

Rafael
--

From: Didier Raboud
Subject: Re:
Date: Tuesday, April 1, 2008 - 2:18 pm

Hi,

would like to. How=A0?=20

# echo -n disk > /sys/power/state=20

?

Regards,=20

Didier
From: Rafael J. Wysocki
Subject: Re:
Date: Tuesday, April 1, 2008 - 2:22 pm

Yes, should work without the "-n", too.  Of course, you have to use the
'resume=' kernel command line argument for this to work.

Perhaps increase the console log level before that.

Thanks,
Rafael
--

From: Didier Raboud
Subject: Re:
Date: Wednesday, April 2, 2008 - 5:06 am

Ok. I found a "better" way to do it=A0: just uninstall uswsusp and let=20
kpowersave use hal to use suspend :-)

Anyway, so far so good (with USB_SUSPEND not set), no hang. I will test wit=
h=20
USB_SUSPEND set and the incoming 2.6.25-rc8.

Regards,=20

Didier
From: Didier Raboud
Subject: Re:
Date: Thursday, April 3, 2008 - 6:41 am

Le mercredi, 2 avril 2008 14.06:29 Didier Raboud, vous avez =C3=A9crit=C2=

So far so good too with "stock" Debian Kernel (2.7.25-rc7-amd64) with=20
USB_SUSPEND set. No hang.

The kernel suspend is less user-friendly than uswsusp though...

So I bet there is a bug in uswsusp's kernel part ?

Regards,=20

Didier
From: Rafael J. Wysocki
Subject: Re:
Date: Thursday, April 3, 2008 - 8:48 am

It's almost the same as the in-kernel one, especially as far as the resume is
concerned.  The simplest explanation coming to mind is that the image is
somehow damaged in the uswsusp case.

You can use the uswsusp's checksumming to verify that, though.

Thanks,
Rafael
--

From: Didier Raboud
Subject: Re:
Date: Thursday, April 10, 2008 - 4:30 am

Hi,=20

I upgraded my uswsusp to 0.8 and my kernel to 2.6.25-rc8 from Debian. I=20
activated the checksumming. I have hangs _again_ (I probably always had...).

Here's my /etc/uswsusp.conf=A0:
# --------
# /etc/uswsusp.conf(8) -- Configuration file for s2disk/s2both
resume device =3D /dev/mapper/Tamino-swap_1
compress =3D y
early writeout =3D y
image size =3D 967540654
shutdown method =3D platform
compute checksum =3D y
# --------

I don't know how I can help further... I will try without "early writeout".

Any help would be greatly appreciated=A0!

Best regards,=20

Didier

=2D-=20
Didier Raboud
Cornalles 2
1802 Corseaux
didier@raboud.com
079 480 67 82
From: Pavel Machek
Date: Tuesday, April 1, 2008 - 12:37 pm

Just do echo disk > /sys/power/state to use swsusp and bypass uswsusp?


-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
--

From: Pavel Machek
Date: Tuesday, April 1, 2008 - 12:38 pm

Try to unload modules before suspend/reload them after resume?

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
--

Previous thread: none

Next thread: [PATCH] x86: generic versions of find_first_(zero_)bit, convert i386 by Alexander van Heukelum on Monday, March 31, 2008 - 10:15 am. (34 messages)