another software suspend issue, seems to be caused by i915/KMS

another software suspend issue, seems to be caused by i915/KMS

[Willi Mann]

Hi!

On my Thinkpad T500 with Debian-compiled kernel 2.6.35~rc6-1~experimental.1, 
I still get OOpses on suspend. 

The following OOPS has been found be using the platform method. I had to 
hard-reboot as SysRQ didn't work. (although I have also seen similar suspend 
OOPSes where it worked). The suspend has been initiated by echo disk > 
/sys/power/state from virtual terminal. 

http://wserver.wm1.at/~willi/linux/suspend_issue/suspend_error.jpg

However, I don't get the same OOPS every time, and the problem does not 
occur at every suspend. I haven't managed to get more than one suspend 
successfully finished (since a cold-boot). Particularly, it seems that if 
the first suspend succeeds, the problem always occurs at the second attempt. 

Also note that if I initiated the suspend from X, I often got a blank screen 
(except for a blinking cursor) instead of the OOPS. (The fact that the 
screen was blank could also be related to the fact that when tested from X, 
I hadn't raised the log levels)

When I disable the i915 module (I renamed it), I can't reproduce the 
problem.

lspci -v -s 00:02.0:
00:02.0 VGA compatible controller: Intel Corporation Mobile 4 Series Chipset 
Integrated Graphics Controller (rev 07) (prog-if 00 [VGA controller])
        Subsystem: Lenovo Device 2114
        Flags: bus master, fast devsel, latency 0, IRQ 48
        Memory at f4400000 (64-bit, non-prefetchable) [size=4M]
        Memory at d0000000 (64-bit, prefetchable) [size=256M]
        I/O ports at 1800 [size=8]
        Expansion ROM at <unassigned> [disabled]
        Capabilities: [90] MSI: Enable+ Count=1/1 Maskable- 64bit-
        Capabilities: [d0] Power Management version 3
        Kernel driver in use: i915

Please let me know what additional infos are required to resolve this 
problem. 

WM 

Re: another software suspend issue, seems to be caused by i915/KMS

[Rafael J. Wysocki]

On Sunday, August 01, 2010, Willi Mann wrote:
> Hi!
> 
> On my Thinkpad T500 with Debian-compiled kernel 2.6.35~rc6-1~experimental.1, 
> I still get OOpses on suspend. 

Did it happen with 2.6.34 or previous kernels?

> The following OOPS has been found be using the platform method. I had to 
> hard-reboot as SysRQ didn't work. (although I have also seen similar suspend 
> OOPSes where it worked). The suspend has been initiated by echo disk > 
> /sys/power/state from virtual terminal. 
> 
> http://wserver.wm1.at/~willi/linux/suspend_issue/suspend_error.jpg
> 
> However, I don't get the same OOPS every time, and the problem does not 
> occur at every suspend. I haven't managed to get more than one suspend 
> successfully finished (since a cold-boot). Particularly, it seems that if 
> the first suspend succeeds, the problem always occurs at the second attempt. 
> 
> Also note that if I initiated the suspend from X, I often got a blank screen 
> (except for a blinking cursor) instead of the OOPS. (The fact that the 
> screen was blank could also be related to the fact that when tested from X, 
> I hadn't raised the log levels)
> 
> When I disable the i915 module (I renamed it), I can't reproduce the 
> problem.
> 
> lspci -v -s 00:02.0:
> 00:02.0 VGA compatible controller: Intel Corporation Mobile 4 Series Chipset 
> Integrated Graphics Controller (rev 07) (prog-if 00 [VGA controller])
>         Subsystem: Lenovo Device 2114
>         Flags: bus master, fast devsel, latency 0, IRQ 48
>         Memory at f4400000 (64-bit, non-prefetchable) [size=4M]
>         Memory at d0000000 (64-bit, prefetchable) [size=256M]
>         I/O ports at 1800 [size=8]
>         Expansion ROM at <unassigned> [disabled]
>         Capabilities: [90] MSI: Enable+ Count=1/1 Maskable- 64bit-
>         Capabilities: [d0] Power Management version 3
>         Kernel driver in use: i915
> 
> Please let me know what additional infos are required to resolve this 
> problem. 

Rafael

Re: [linux-pm] another software suspend issue, seems to be caused by i915/KMS

[Rafael J. Wysocki]

On Sunday, August 01, 2010, Willi Mann wrote:
> Hi!
> 
> On my Thinkpad T500 with Debian-compiled kernel 2.6.35~rc6-1~experimental.1, 
> I still get OOpses on suspend. 

Did it happen with 2.6.34 or previous kernels?

> The following OOPS has been found be using the platform method. I had to 
> hard-reboot as SysRQ didn't work. (although I have also seen similar suspend 
> OOPSes where it worked). The suspend has been initiated by echo disk > 
> /sys/power/state from virtual terminal. 
> 
> http://wserver.wm1.at/~willi/linux/suspend_issue/suspend_error.jpg
> 
> However, I don't get the same OOPS every time, and the problem does not 
> occur at every suspend. I haven't managed to get more than one suspend 
> successfully finished (since a cold-boot). Particularly, it seems that if 
> the first suspend succeeds, the problem always occurs at the second attempt. 
> 
> Also note that if I initiated the suspend from X, I often got a blank screen 
> (except for a blinking cursor) instead of the OOPS. (The fact that the 
> screen was blank could also be related to the fact that when tested from X, 
> I hadn't raised the log levels)
> 
> When I disable the i915 module (I renamed it), I can't reproduce the 
> problem.
> 
> lspci -v -s 00:02.0:
> 00:02.0 VGA compatible controller: Intel Corporation Mobile 4 Series Chipset 
> Integrated Graphics Controller (rev 07) (prog-if 00 [VGA controller])
>         Subsystem: Lenovo Device 2114
>         Flags: bus master, fast devsel, latency 0, IRQ 48
>         Memory at f4400000 (64-bit, non-prefetchable) [size=4M]
>         Memory at d0000000 (64-bit, prefetchable) [size=256M]
>         I/O ports at 1800 [size=8]
>         Expansion ROM at <unassigned> [disabled]
>         Capabilities: [90] MSI: Enable+ Count=1/1 Maskable- 64bit-
>         Capabilities: [d0] Power Management version 3
>         Kernel driver in use: i915
> 
> Please let me know what additional infos are required to resolve this 
> problem. 

Rafael

Re: another software suspend issue, seems to be caused by i915/KMS

[Rafael J. Wysocki]

On Sunday, August 01, 2010, Willi Mann wrote:
> Rafael J. Wysocki wrote:
> 
> > On Sunday, August 01, 2010, Willi Mann wrote:
> >> Hi!
> >> 
> >> On my Thinkpad T500 with Debian-compiled kernel
> >> 2.6.35~rc6-1~experimental.1, I still get OOpses on suspend.
> > 
> > Did it happen with 2.6.34 or previous kernels?
> 
> Yes, the suspend/resume also failed with 2.6.34, although I only tried 
> suspending from X, which almost always leads to a blank screen, so I can't 
> say whether the symptoms are the same. Of course I can test that if you 
> think this is useful. However, rc5 and rc6 contained a fix for i915 related 
> to suspend/resume, which might hide this remaining problem. So it might not 
> be particularly useful to test 2.6.34.
> 
> As far as I remember, KMS always led to problems related to suspend/resume, 
> but the both the highly patched Debian Kernel 2.6.32-18 (Debian version) and 
> 2.6.35-rc6 fail very often on suspend. I'd like to avoid testing earlier 
> 2.6.32 kernels, because some of them even cause file system corruption on 
> suspend. (If really required, I could set up a small test partition, and 
> hope that the filesystem corruption was no arbitrary write on the disk, but 
> only on the partition.)
> 
> The last time I had no problems with suspend/resume was with some kernel 
> 2.6.29 version, which of course used User Mode Setting. I actually didn't 
> cold-boot my notebook for 180 days back then.

Thanks for the reply, but please keep the CC list when replying next time.

Rafael

Re: another software suspend issue, seems to be caused by i915/KMS

[Rafael J. Wysocki]

On Sunday, August 01, 2010, Willi Mann wrote:
> Rafael J. Wysocki wrote:
> 
> > On Sunday, August 01, 2010, Willi Mann wrote:
> >> Hi!
> >> 
> >> On my Thinkpad T500 with Debian-compiled kernel
> >> 2.6.35~rc6-1~experimental.1, I still get OOpses on suspend.
> > 
> > Did it happen with 2.6.34 or previous kernels?
> 
> Yes, the suspend/resume also failed with 2.6.34, although I only tried 
> suspending from X, which almost always leads to a blank screen, so I can't 
> say whether the symptoms are the same. Of course I can test that if you 
> think this is useful. However, rc5 and rc6 contained a fix for i915 related 
> to suspend/resume, which might hide this remaining problem. So it might not 
> be particularly useful to test 2.6.34.
> 
> As far as I remember, KMS always led to problems related to suspend/resume, 
> but the both the highly patched Debian Kernel 2.6.32-18 (Debian version) and 
> 2.6.35-rc6 fail very often on suspend. I'd like to avoid testing earlier 
> 2.6.32 kernels, because some of them even cause file system corruption on 
> suspend. (If really required, I could set up a small test partition, and 
> hope that the filesystem corruption was no arbitrary write on the disk, but 
> only on the partition.)
> 
> The last time I had no problems with suspend/resume was with some kernel 
> 2.6.29 version, which of course used User Mode Setting. I actually didn't 
> cold-boot my notebook for 180 days back then.

Thanks for the reply, but please keep the CC list when replying next time.

Rafael