Suspend not working on Sony Vaio SZ650N, Intel 965 display adapter

With hardy alpha-2 installed to disk, suspending to ram works, but restarting results in a very dim desktop - seems that the backlight has been set to minimum. You can get the backlight back by logging out and back in again.

Thanks Michael. Can you attempt a Suspend/Resume and then capture your dmesg output and attach? Also, the following might be helpful: https://wiki.ubuntu.com/DebuggingKernelSuspend . Thanks!

Suspend-resume seems to be working well in Gutsy now - to do with
linux-backports-modules? Still a problem in Hardy alpha-2 though.

These are dmesg outputs after suspend-resume in Gutsy and Hardy on the
Intel graphics adapter (GM965). Have not tried on the nVidia adapter as
it is still unstable.

Leann Ogasawara wrote:
> Thanks Michael. Can you attempt a Suspend/Resume and then capture your
> dmesg output and attach? Also, the following might be helpful:
> https://wiki.ubuntu.com/DebuggingKernelSuspend . Thanks!
>
> ** Changed in: linux (Ubuntu)
> Status: New => Incomplete
>

Attached are some logs from Gutsy failing to resume fully from suspend-to-RAM. Suspend appeared to work perfectly, but resume just gave a blank screen.

Just suspended and hibernated successfully with Hardy alpha 4, using the intel 965 display adapter. dmesg attached.

Hi Michael,

Thanks for testing and the update. I'm glad to hear things appear to be working better. Per your last comment I'll go ahead and close this report. Thanks!

Suspend and hibernate work when using intel adapter (GM 965; stamina mode),
but do not work (fail to resume) when using the nVidia adapter (speed mode;
GeForce 8400M GS, both nv and nvidia drivers), so perhaps rather than
closing this bug it could be modified to specify that there is still a
problem with support for the nVidia card.

On 05/02/2008, Leann Ogasawara <email address hidden> wrote:
>
> Hi Michael,
>
> Thanks for testing and the update. I'm glad to hear things appear to be
> working better. Per your last comment I'll go ahead and close this
> report. Thanks!
>
> ** Changed in: linux (Ubuntu)
> Status: Triaged => Fix Released
>
> --
> Suspend not working on Sony Vaio SZ650N
> https://bugs.launchpad.net/bugs/178286
> You received this bug notification because you are a direct subscriber
> of the bug.
>

Leann - looks like I spoke too soon. Just had a failure to suspend on Hardy alpha-4, have marked it as complete while I collect logs.

Hi Michael,

Thanks for collecting the logs. I didn't realize until I saw your other bug report regarding Hibernation that you had issues with both Intel and Nvidia adapters. Thanks for adding the comment here. It would actually be good if we could target this report against one of those adapters. If you could maybe update the title and description of this report (see link in Actions area on the left) with the adapter you want to target that would be helpful. The reason I say that is because the symptom of this particular bug, failure to Suspend, is specific to the hardware used.

With your Intel adapter, care to maybe try:

echo 3 > /proc/sys/kernel/acpi_video_flags
/etc/acpi/sleep.sh force

Thanks.

OK, changed description as requested, will limit this bug to the Intel 965.

Running:

echo 3 > /proc/sys/kernel/acpi_video_flags
/etc/acpi/sleep.sh force

Resulted in a successful suspend, which resumed correctly.

The attached log is from a broken suspend on Hardy alpha 5

Hi Michael,

Care to actually test the latest 2.6.24-10 kernel which was released after Alpha 5 came out. Some changes went in that may have resolved the issue you are seeing. See bug 189260 for a reference. Let us know your results. Thanks!

Leann

Just had a successful suspend-resume with kernel 2.6.24-10, though 'one swallow doesn't make a summer', as they say. If I have a failure, I'll be sure to post my logs here.

Mike

Thanks Mike,

I'm going to mark this "Fix Released" for now. But obviously if you notice any regressions please feel free to reopen. Thanks!

Suspend just failed using Hardy Beta, amd64. Logs coming...

And twice again, after the latest updates were applied...

$ uname -a
Linux doris 2.6.24-16-generic #1 SMP Thu Apr 10 12:47:45 UTC 2008 x86_64 GNU/Linux

in that last kern.log, see the entries starting here:

Apr 15 00:30:03 doris kernel: [ 8080.438909] PM: Writing back config space on device 0000:06:00.0 at offset 1 (was 100002, writing 100006)

I had another suspend failure, which dumped some possibly useful information in kern.log, see entries starting here:

Apr 23 13:19:03 doris kernel: [ 2491.512095] sky2 eth0: disabling interface

To my great surprise, I have just had a successful suspend-resume cycle. See kern.log starting here:

Apr 28 13:04:39 doris kernel: [13001.894395] ACPI: PCI interrupt for device 0000:06:00.0 disabled

The only unusual things about today's session are that bluetooth has not been activated (possibly because of the prior WinXP session disabling it) so the blue LED is not glowing, and that I disabled WiFi with the kill switch about 3 minutes prior to suspending.

I haven't knowingly installed or uninstalled any kernel packages; I am still using linux-backports-modules-hardy.

After an Ubuntu->Ubuntu reboot, suspend failed again. Booting into XP (where bluetooth is disabled), then restarting into Ubuntu allowed a normal suspend-resume cycle in Hardy. The only outstanding feature that I can see is that the Bluetooth LED is disabled, and the device is not listed by lsusb.

I've captured dmesg output from a Win->Ubuntu restart and then booting to Ubuntu from cold. The way that USB is handled is quite different: Bluetooth is not started after restarting from Win, uvcvideo is handled differently and there are a bunch of other little differences. I'm not sure which of these might be significant, but it makes the difference between being able to suspend and not.

Last dmesg is booting from cold into Ubuntu, this one is restarting from Win XP into Ubuntu:

Suspend behaviour at the moment seems to be dependent on the boot sequence; If I boot from cold into Hardy, suspend will not work. If I restart Hardy->Hardy, suspend will not work. If I restart WinXP->Hardy or Gutsy->Hardy, suspend works.

Additionally, I have got suspend to work today after removing linux-backports-modules and doing a Hardy->Hardy restart, which is promising.

I've got the same laptop, and had the same issue. I find if I start the machine down with wireless set to off, it successfully suspends, even if I turn the wireless back on later. No luck on hiberation though.

Hardy amd64.

Just tried booting into Hardy with the kill switch on (i.e. radio off) and suspended successfully while the radio was off but failed to suspend with radio on.

I was able to isolate the issue down to partially to a specific driver: hci_usb which is loaded as part of the bluetooth driver initialization. If this driver is unloaded, and the machine is started with the wireless kill switch set to off, then turned on, and then put back to off, it will successfully suspend/resume.

I've managed to get it to fully work by adding hci_usb to the modules list in /etc/default/acpi-support. As long as I have wireless switched to off when I start the machine, and I turn it off before suspending it works properly.

The switch causes a bluetooth USB dongle to appear in dmesg when flipped on, and it disappears when flipped off.The onboard WWAN card may also be loading/unloading with this switch.

I was able to resolve the issue:

YOu need to actually blacklist uvcvideo (since the onboard video doesn't work anyway, no loss of functionality), and install linux-backports-hardy, to get an updated iwl4963 (if successful, your WLAN light should work). You may need to add iwl4963 to the MODULES list in /etc/default/acpi, but it works.

That works for me, thanks!
I have both hci_usb and iwl4965 listed in /etc/default/acpi-support MODULES list, and blacklisted uvcvideo in /etc/modprobe.d/blacklist.

Interestingly, cheese no longer works, but Skype video works - a direct reversal of the situation after a Gutsy->Hardy restart.

uvcvideo never worked with me, but I just read you have the SZ650, and I have a SZ660 :-P

If you want cheese to work, you'll probably have to reload uvcvideo into the kernel, and then remember to remove it before suspending; listing it in MODULES didn't work for me.

Skype to my knowledge doesn't use the video4linux API, which is why that works. Beyond that, no idea

I'm still getting errors when I suspend with WiFi on, having booted with WiFi off. kern.log attached: looks like there is some problem stopping iwlwifi or iwl4965.

After applying this fix, I can boot with wifi on and leave it on. I haven't tried booting with it off ...

This issue been resolved in intrepid, I do not need to blacklist any modules to get sleep/suspoend working out of the box for me now.

Intrepid is working for me too (with no special modifications), but rebooting from Intrepid to Hardy has resulted in 3 failed suspends under Hardy, with the kernel Aiee printed to screen once, a la bug #252270.

Hardy, which has Long Term Support, has not had this bug fixed - a pain, because it is my work system and will be for the next 2 years. So, while a fix may have been released for Intrepid, it has not been released for Hardy.

I just went to suspend again in Intrepid, and the Suspend button is missing. WTF?

Hmmmm OK. Maybe this was a problem with the latest Hardy kernel in proposed, 2.6.24-21. I just tried rmmod - ing iwl4965 from a running system and got the same total system hang, flashing led's unresponsive power button etc. So perhaps this is mainly a problem with the way the kernel inserts and removes iwl4965?

Suspend seems to be working OK on linux-2.6.24-20.

Confirmed. Suspend disappears with 2.6.27 on intrepid. Reappears with the 2.6.26 kernel. Investigating now.

Suspend has worked for me fine on 2 attempts on Intrepid alpha 6 (2.6.27-3). Have just installed 2.6.27-4, will retest shortly.

Suspend is working fine with linux-2.6.27-4.

I think the cheese issue is actually in the r5u870 driver not making itself known to HAL (bug #219252), so isn't relevant to this discussion.

For me, Linux convinced there is no suspend functionality (doesn't even attempt to enter at all), so possibly a separate issue. Will debug.

Michael,

Suspend worked fine for me on a fresh install of Intrepid alpha 6,
with a new user created so there is no 'carry-over' of cruft from
earlier installs. The disappearing Suspend button problem seemed to
only occur on alpha 5.

Suspend has been working fine for a while now (a month or longer?) but today threw a "general protection fault" when I attempted to suspend. I've been having intermittent problems with network functionality (WiFi mainly) after resuming from boot and thought that may be due to the new iwlagn driver, which incorporates iwl4965. I now specify in /etc/default/acpi-support to unload pretty much all wifi-related modules; perhaps that is the problem?

/etc/default/acpi-support
snip------
MODULES="hci_usb lbm_iwl_mac80211 iwl4965 iwlagn iwlcore"
snip------

OK, removed all but hci_usb and iwl4965 and suspend is working agin (so far...)

/etc/default/acpi-support
snip------
MODULES="hci_usb iwl4965"
snip------

Per a decision made by the Ubuntu Kernel Team, bugs will longer be assigned to the ubuntu-kernel-team in Launchpad as part of the bug triage process. The ubuntu-kernel-team is being unassigned from this bug report. Refer to https://wiki.ubuntu.com/KernelTeamBugPolicies for more information. Thanks.