Bug 106049

Summary: suspend to disk (both ACPI and APM) hangs a thinkpad T23
Product: [openSUSE] SUSE LINUX 10.0 Reporter: Markus Dellermann <liml>
Component: KernelAssignee: Pavel Machek <pavel>
Status: RESOLVED DUPLICATE QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P5 - None CC: behlert, trenn
Version: Beta 2   
Target Milestone: ---   
Hardware: i586   
OS: All   
Whiteboard:
Found By: Beta-Customer Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: screenshot of attempted suspend after normal bootup (acpi + vga=0x...)
screenshot of attempted suspend with vga=normal
screenshot of attempted suspend with acpi=off (APM "working")
/var/log/boot.msg
hwinfo "thinkpad-t23"

Description Markus Dellermann 2005-08-21 15:46:06 UTC
This has worked with 9.3. 
 
It hangs on suspending the system, not on wake-up (never get there). 
 
Will attach screenshots of several attempts: 1. normal boot options (ACPI and 
vga=0x...), 2. bootup with vga=normal and 3. bootup with acpi=off.  
 
Also will attach /var/log/boot.msg and hwinfo output. 
 
think:~ # cat /proc/version 
Linux version 2.6.13-rc6-git7-3-default (geeko@buildhost) (gcc version 4.0.2 
20050812 (prerelease) (SUSE Linux)) #1 Mon Aug 15 18:28:43 UTC 2005
Comment 1 Hartmut Meyer 2005-08-21 15:51:37 UTC
Created attachment 46785 [details]
screenshot of attempted suspend after normal bootup (acpi + vga=0x...)

quality of screenshot is poor.

Have a look at apm.jpg first as that shows a very similar screenshot but in
much better quality ...
Comment 2 Hartmut Meyer 2005-08-21 15:53:12 UTC
Created attachment 46786 [details]
screenshot of attempted suspend with vga=normal
Comment 3 Hartmut Meyer 2005-08-21 15:55:14 UTC
Created attachment 46787 [details]
screenshot of attempted suspend with acpi=off (APM "working")
Comment 4 Markus Dellermann 2005-08-21 15:55:56 UTC
Created attachment 46788 [details]
/var/log/boot.msg
Comment 5 Markus Dellermann 2005-08-21 15:57:24 UTC
Created attachment 46789 [details]
hwinfo "thinkpad-t23"
Comment 6 Hartmut Meyer 2005-08-21 15:57:43 UTC
Comment on attachment 46787 [details]
screenshot of attempted suspend with acpi=off (APM "working")

this is the screenshot with the best quality - have a look at this one first.
Comment 7 Thomas Renninger 2005-08-21 17:31:25 UTC
Kernel ooops in savagefb_suspend.
Pavel is/should this even be supported?
That's your area...
Comment 8 Hartmut Meyer 2005-08-21 17:42:51 UTC
Note that it also hangs when you avoid loading savagefb (boot with 
vga=normal). See attachment #2 [details]. 
 
In fact, I belive I seen a oops when trying this before (again with vga=normal 
and savagefb not loaded. 
 
If you think that this oops would help, then I'm sure markus could try and 
reproduce ... 
Comment 9 Pavel Machek 2005-08-21 20:57:32 UTC
What is tainted: G? Aha, proprietary module. Will not debug this one...

Try again with minimal drivers to see if it works at all (you are using swsusp,
right?) and please try to be little less trigger-happy with screenshots.
Comment 10 Pavel Machek 2005-08-21 21:06:39 UTC
Oops, sorry, tainted: G is "GPL only". Can you try with latest -linus kernel?
Comment 11 Hartmut Meyer 2005-08-23 16:40:57 UTC
I'm trying to build a new kernel (2.6.13-rc6-git13) but it fails. 
 
I suspect that I did something wrong in the patching process. This is what I 
did: 
 
download: 
 
  linux-2.6.12.5.tar.bz2 
  patch-2.6.13-rc6 
  patch-2.6.13-rc6-git13 
 
to /usr/src 
 
Then 
 
  tar xjvf linux-2.6.12.5.tar.bz2 
  bunzip patch-2.6.13-rc6* 
  cd linux-2.6.12.5 
  patch -p1 < ../patch-2.6.13-rc6  [NOT ok - see below] 
  patch -p1 < ../patch-2.6.13-rc6-git13 [ok] 
  zcat /proc/config.gz > .config 
  make oldconfig [ok] 
  make binrpm-pkg [NOT ok] 
 
It already gave me several warnings/errors: "Reversed (or previously applied) 
patch detected!" and also some "Chunk soandso failed". 
 
The "make binrpm-pkg" later errors out in drivers/char/tpm/tpm.c which I 
remember was one of the reported "Reversed patch detected files". 
 
I will now try to apply those two patches against linux-2.6.12 (without 
the .5) ... 
 
Was linux-2.6.12.5 the wrong starting point for patch-2.6.13-rc6? 
Comment 12 Hartmut Meyer 2005-08-23 17:38:22 UTC
Ok - it patched without complaints now (against linux-2.6.12). 
 
Compile is running but looks good ... 
Comment 13 Markus Dellermann 2005-08-23 23:58:58 UTC
Ok, I only could try with 2.6.12.5 
Both, "suspend to disk" and "suspend to ram" works fine with this kernel 
 
Comment 14 Pavel Machek 2005-08-25 10:01:48 UTC
We should not have many changes in that area... Could you test 2.6.13 when it
comes out (should be soon), or 2.6.13-rc7...

Yes, to update to -rc6, you need to go from 2.6.12, not 2.6.12.X. There's even
nice script called "ketchup" to handle this for you.
Comment 15 Forgotten User ZhJd0F0L3x 2005-08-26 09:01:55 UTC
Stefan, we should have a T2{1,2,3} in our repository. Can we check this?
Comment 16 Stefan Behlert 2005-08-26 09:55:24 UTC
There's a T20 - Danny is currently replacing the broken harddisk of that one. 
He'll install beta3 if the machine is working afterwards. 
Comment 17 Pavel Machek 2005-08-26 22:32:06 UTC

*** This bug has been marked as a duplicate of 113296 ***