HAL quirks tbd for Toshiba P100 resume from STR

[Yann]

Binary package hint: hal

Toshiba P100 (PSPA3E) fails to suspend (and so, resume) properly, approximatively 50% tries.

/var/log/messages showing:
1) For successful resumes:
(...)
May 13 23:15:44 marsupilami gnome-power-manager: (yann) Mise en veille de l'ordinateur. Raison : Le bouton de mise en veille a été appuyé.
May 13 23:15:44 marsupilami kernel: [ 954.761484] atkbd.c: Unknown key released (translated set 2, code 0x55 on isa0060/serio0).
May 13 23:15:44 marsupilami kernel: [ 954.761492] atkbd.c: Use 'setkeycodes 55 <keycode>' to make it known.
May 13 23:15:46 marsupilami kernel: [ 957.035467] ehci_hcd 0000:00:1d.7: remove, state 4
May 13 23:15:46 marsupilami kernel: [ 957.035480] usb usb5: USB disconnect, address 1
May 13 23:15:46 marsupilami kernel: [ 957.039685] ehci_hcd 0000:00:1d.7: USB bus 5 deregistered
May 13 23:15:46 marsupilami kernel: [ 957.039739] ACPI: PCI interrupt for device 0000:00:1d.7 disabled
May 13 23:15:47 marsupilami kernel: [ 957.225345] ACPI: PCI interrupt for device 0000:02:00.0 disabled
May 13 23:15:47 marsupilami kernel: [ 957.490361] ACPI: PCI interrupt for device 0000:03:00.0 disabled
May 13 23:16:02 marsupilami kernel: [ 958.707658] Syncing filesystems ... done.
May 13 23:16:02 marsupilami kernel: [ 958.707695] Freezing user space processes ... (elapsed 0.00 seconds) done.
May 13 23:16:02 marsupilami kernel: [ 958.708397] Freezing remaining freezable tasks ... (elapsed 0.09 seconds) done.
May 13 23:16:02 marsupilami kernel: [ 958.805121] Suspending console(s)
May 13 23:16:02 marsupilami kernel: [ 958.805162] sd 0:0:0:0: [sda] Synchronizing SCSI cache
May 13 23:16:02 marsupilami kernel: [ 958.874304] sd 0:0:0:0: [sda] Stopping disk
May 13 23:16:02 marsupilami kernel: [ 960.079840] ACPI: PCI interrupt for device 0000:0a:04.3 disabled
May 13 23:16:02 marsupilami kernel: [ 960.095343] ACPI: PCI interrupt for device 0000:0a:04.2 disabled
May 13 23:16:02 marsupilami kernel: [ 960.127683] NVRM: RmPowerManagement: 4
May 13 23:16:02 marsupilami kernel: [ 961.691691] ACPI: PCI interrupt for device 0000:00:1f.2 disabled
May 13 23:16:02 marsupilami kernel: [ 961.706395] ACPI: PCI interrupt for device 0000:00:1d.3 disabled
May 13 23:16:02 marsupilami kernel: [ 961.706437] ACPI: PCI interrupt for device 0000:00:1d.2 disabled
May 13 23:16:02 marsupilami kernel: [ 961.706478] ACPI: PCI interrupt for device 0000:00:1d.1 disabled
May 13 23:16:02 marsupilami kernel: [ 961.706519] ACPI: PCI interrupt for device 0000:00:1d.0 disabled
May 13 23:16:02 marsupilami kernel: [ 961.722803] ACPI: PCI interrupt for device 0000:00:1b.0 disabled
May 13 23:16:02 marsupilami kernel: [ 961.764973] Disabling non-boot CPUs ...
May 13 23:16:02 marsupilami kernel: [ 961.878071] CPU 1 is now offline
May 13 23:16:02 marsupilami kernel: [ 961.878074] SMP alternatives: switching to UP code
May 13 23:16:02 marsupilami kernel: [ 961.879683] CPU1 is down
May 13 23:16:02 marsupilami kernel: [ 0.338622] Enabling non-boot CPUs ...
May 13 23:16:02 marsupilami kernel: [ 0.348801] SMP alternatives: switching to SMP code
May 13 23:16:02 marsupilami kernel: [ 0.349907] Booting processor 1/1 eip 3000
(...)

2) For unsuccessul ones:
(...)
May 13 06:34:02 marsupilami gnome-power-manager: (yann) Mise en veille de l'ordinateur. Raison : Le bouton de mise en veille a été appuyé.
May 13 06:34:02 marsupilami kernel: [ 205.768427] atkbd.c: Unknown key released (translated set 2, code 0x55 on isa0060/serio0).
May 13 06:34:02 marsupilami kernel: [ 205.768437] atkbd.c: Use 'setkeycodes 55 <keycode>' to make it known.
May 13 06:34:04 marsupilami kernel: [ 207.478772] ehci_hcd 0000:00:1d.7: remove, state 4
May 13 06:34:04 marsupilami kernel: [ 207.478784] usb usb5: USB disconnect, address 1
May 13 06:34:04 marsupilami kernel: [ 207.482936] ehci_hcd 0000:00:1d.7: USB bus 5 deregistered
May 13 06:34:04 marsupilami kernel: [ 207.482996] ACPI: PCI interrupt for device 0000:00:1d.7 disabled
May 13 06:34:04 marsupilami kernel: [ 207.668975] ACPI: PCI interrupt for device 0000:02:00.0 disabled
May 13 06:34:04 marsupilami kernel: [ 207.916748] ACPI: PCI interrupt for device 0000:03:00.0 disabled
May 13 06:37:56 marsupilami syslogd 1.5.0#1ubuntu1: restart.
May 13 06:37:56 marsupilami kernel: Inspecting /boot/System.map-2.6.24-16-generic
May 13 06:37:56 marsupilami kernel: Loaded 27704 symbols from /boot/System.map-2.6.24-16-generic.
May 13 06:37:56 marsupilami kernel: Symbols match kernel version 2.6.24.
(...)

So, we an see that when unsucessul, suspend process stop just before syncing disks, also causing FS checks at reboot (always with some journal retrieve/orphan inodes cleared, hope I'll not have problems).

It seems the solution to this bug is to had a quirk in the HAL, which previously had to be done in grub ("acpi_sleep=s3_mode,s3_bios" with my previous Dapper+Powersaved configuration I just ugraded to Hardy). A google search show this quirk have already be pushed in HAL since april and with pm-utils it seems to override grub boot line options (nothing changed using them??!!).

See:
http://lists.freedesktop.org/archives/hal-commit/2008-April/004400.html

Verified adding "P100" here to my /usr/share/hal/fdi/information/10freedesktop/20-video-quirk-pm-toshiba.fdi
to get april hal-commit like config on Hardy:
        <match key="system.hardware.product" contains_outof="M35X;P100;P200; A100; A135; L30;M70">
          <merge key="power_management.quirk.s3_bios" type="bool">true</merge>
          <merge key="power_management.quirk.s3_mode" type="bool">true</merge>
        </match>

And it seems to work so far... inch'Allah...

The only issues I still have with suspend is sometimes a blank screen using nvidia proprietary drivers (7600 Go GPU). Tried a few quircks but without success so far (any hint welcome!). Resuming was not causing such problems with dapper/powersaved and boot line adds. Also, as I Dapper, I still have no sound after resume (should be nice to fix this also, any hints welcome again!).

But this remaining blank screen issue (only backlight seems to be on) is far less critical: If you enter your password it resumes with normal screen!

Should be nice to push this asap in updates!

Regards

[Yann (lostec)]

Is there someone reading bugs on Launchpad?
So much reports still waiting corrections since months or years, even when the solution is given!

Really, I hope Ubuntu is not disconnecting from it's users with it's growong success...

Regards

[Sense Egbert Hofstede (sense)]

Thank you for reporting this bug.
However, it doesn't contain enough information to pass it on to the developers. Can you please follow the procedure described at <https://wiki.ubuntu.com/DebuggingHal>?

Ubuntu's not losing contact with the community in my eyes. There are just so many bugs that one easily is forgotten. If you have suggestions to improve the number of bugs triaged, please mail them to
<email address hidden>.

[Sense Egbert Hofstede (sense)]

I confused this bug report with another. Can you please _also_ attach the output of the command 'dmesg' and the file /var/log/kern.log after you've tried to suspend/hibernate.

Thanks in advance.

[Yann (lostec)]

Here is kern.log from a failled suspend hanging at:
May 27 18:24:01 marsupilami kernel: [ 760.032888] ACPI: PCI interrupt for device 0000:03:00.0 disabled

Just before FS syncing => orphaned inodes on restart...

[Yann (lostec)]

Here is a lspci -vvv result because suspend seems to fail on a pci device??
dmesg is a subset of kern.log and does not show more informations...

Regards

[Yann (lostec)]

After last kernel update (2.6.24-17), suspend NEVER work!
One more regression on this subject: Hereupper quirks are still there, but no more making suspend to ram work.

I don't know if Linux will one day have a reliable suspend?

[Yann (lostec)]

Please also note that PCI device IRQ disabling after which suspend hangs is wifi 3945 ABG. It seems driver have been changed since my old Dapper (was using Intel driver)? Is the new one buggy like hell for suspend? Is it possible to use old one?

Regards

[Sense Egbert Hofstede (sense)]

It seems that the Intel Corporation 82801G USB Controller is the cause. Can you please try to suspend/resume again after you've disabled all USB devices in the BIOS? Thanks in advance.

[Yann (lostec)]

Hi,

I checked my bios and the only USB related option is to disable an "USB emulation" (related to HDD, CD... possible USB boot devices I think)... and it is already disabled.

I also already have a custom DSDT for this laptop, related to acpi bios bugs (GPU was not cooling, sound irq routing problems...). One advice: For Linux, don't buy Toshiba!

For USB+suspend, I've read there is another bug related to USB and I've already this:
~$ more /etc/pm/config.d/unload_modules
SUSPEND_MODULES="ehci_hcd"

Something could be added for USB to wake-up correctly?

[Sense Egbert Hofstede (sense)]

I've found something that might be worth a shot. Can you please add 'blacklist usbcore' to /etc/modules and try again.
Can you also please tell what USB devices are connected?
Thanks in advance.

[Yann (lostec)]

I was on vacations, I try this asap. Only one usb device is always connected: A M$ usb laptop mini-mouse... Maybe Billy-bug is to blame ;-)

[Yann (lostec)]

Hello,

I tried blacklisting usbcore in /etc/modules but see:

lsmod |grep usb
usbhid 31872 0
hid 38784 1 usbhid
usbcore 146028 4 usbhid,ehci_hcd,uhci_hcd

usbcore is still there ?!

So I tried in /etc/modprobe.d/blacklist

Same result: usbcore is still there!

It seems impossible to blacklist this module??!!

Regards

[Sense Egbert Hofstede (sense)]

It shouldn't be impossible to blacklist if I'm right. Are you sure you added 'blacklist' before 'usbcore'? Even in /etc/modprobe.d/blacklist ?

[Yann (lostec)]

Yes, I was sure... but I retried and there is no possible error either on the module name or blacklist keyword.

Maybe it's blacklisted, but mandatory on top of hid/hcd modules and these ones, not blacklisted, load usbcore on their own?

But if I remove these ones, I'll get no USB at all... Maybe I could try them in pm-utils SUSPEND_MODULES list?

Regards...

[Sense Egbert Hofstede (sense)]

It could be that bug 220607 describes the same problem. If that's true your problem will be solved in the 2.6.25 kernel and higher. However, you probably want this fixed now, and not when you upgrade to 8.10. I'm not sure what to do next, actually. When your bugs turns out to be a duplicate, it will be marked as such. Otherwise I'll try again to ask for someone who knows more about the kernel than I do to have a look at this bug.

[Yann (lostec)]

I'm not sure it's the same problem. Symptoms differ.

As many others, my suspend was "almost working" in 2.6.24-16, broken in -17/-18 kernels and partially restored with -19.

Now, my Toshiba P100 (core-duo T2300 based) can suspend 2 times... third time It cannot resume... even with the HAL Quirks.

Concerning updating to Intrepid... I really do not plan to do so: I stick with Ubuntu because LTS versions exists and allow update rates between 2-3 years (2 years between release and 1 year to upgrade before previous version end of life)! This is the good "New features attraction" versus "upgrade problems" compromise for me (and many others I think).

So if someone could have suspend fixed in this long lasting version, that would really be great for laptop users.

Regards

[Sense Egbert Hofstede (sense)]

In the hardy-proposed repositories there is already a newer kernel, version 2.6.24-20, which has a long changelog. I'm not sure when it will be accepted for the masses, but there is a large change it will fix your problems.
Could you tell me if the new kernel completely fixes your issues, either by waiting until the update comes or by enabling hardy-proposed?

Thanks in advance.

[Yann (lostec)]

It seems there is some problems with this kernel and 3945 based wifi, see:
https://bugs.launchpad.net/ubuntu/+bug/250520

So I should probably wait until this issue is solved in another "proposed" iteration of -20? I've read some bug reports about nvidia drivers that also seems problematic with this one?

Regards.

[Sense Egbert Hofstede (sense)]

There have been a lot of kernel updates lately. Does the bug still occur at your system? Otherwise I think I should forward it upstream.

[Leann Ogasawara (leannogasawara)]

The Ubuntu Kernel Team is planning to move to the 2.6.27 kernel for the upcoming Intrepid Ibex 8.10 release. As a result, the kernel team would appreciate it if you could please test this newer 2.6.27 Ubuntu kernel. There are one of two ways you should be able to test:

1) If you are comfortable installing packages on your own, the linux-image-2.6.27-* package is currently available for you to install and test.

--or--

2) The upcoming Alpha5 for Intrepid Ibex 8.10 will contain this newer 2.6.27 Ubuntu kernel. Alpha5 is set to be released Thursday Sept 4. Please watch http://www.ubuntu.com/testing for Alpha5 to be announced. You should then be able to test via a LiveCD.

Please let us know immediately if this newer 2.6.27 kernel resolves the bug reported here or if the issue remains. More importantly, please open a new bug report for each new bug/regression introduced by the 2.6.27 kernel and tag the bug report with 'linux-2.6.27'. Also, please specifically note if the issue does or does not appear in the 2.6.26 kernel. Thanks again, we really appreicate your help and feedback.

[Yann (lostec)]

@ Sense:
Yes, it's still there on an updated machine today (but without -proposed activated, but I got -19 updates recently)...

@ Leann:
I only install LTS because I don't want to upgrade every 6 month... So I could only check using a liveCD when I have some time.

But my question is: If 2.6.27 works, will current hardy kernel be corrected or changed? As Debian now start making kernel version upgrades inside a Debian version, maybe ubuntu could start doing the same for long lasting LTS versions.

[Leann Ogasawara (leannogasawara)]

Hi Yann,

Please let us know your results if you are able to test Alpha5.

Regarding Hardy eventually updating it's kernel, it's actually been discussed a little. As you've pointed out, Hardy is an LTS release, so it may prove beneficial to update the kernel at some point. However, these types of discussions are extremely preliminary and nothing concrete has been decided.

[Yann (lostec)]

Hi,

Resume from suspend now seems to work every times since first week of november, maybe related to hardy kernel modules update!

So this is good news! The only remaining pb is this Connexant soundcard not working after resume even if:

cat /dev/sndstat
cat /proc/asound/cards
grep ^Codec /proc/asound/card?/codec*
grep '.*' /sys/module/snd_hda_intel/parameters/*

commands return the same outputs after and before resume...

But that's indeed a good progress!

[Chow Loong Jin (hyperair)]

On Tue, 2008-11-18 at 23:29 +0000, Yann wrote:
> Hi,
>
> Resume from suspend now seems to work every times since first week of
> november, maybe related to hardy kernel modules update!
>
> So this is good news! The only remaining pb is this Connexant soundcard
> not working after resume even if:
>
> cat /dev/sndstat
> cat /proc/asound/cards
> grep ^Codec /proc/asound/card?/codec*
> grep '.*' /sys/module/snd_hda_intel/parameters/*
>
> commands return the same outputs after and before resume...
>
> But that's indeed a good progress!
>
That's probably related to Bug 202089. This is a different bug. Marking
as fix released.

  status fixreleased

--
Chow Loong Jin

[Chow Loong Jin (hyperair)]

Not a pm-utils bug.