PDA

View Full Version : Thinkpad T43 crashes after suspended (2139 kernel)



idiotprogrammer
2nd July 2006, 04:15 AM
After I upgraded to the 39 kernel, my laptop ALWAYS crashes after I suspend it.

When I wake it up (by opening the screen), the screen is messed up, with streaks of multiple colors.

it's not out of the realm of possibility that something else I updated on yum might have broken it. But the problem never existed until I upgraded to this kernel.

The kernel is/usr/src/kernels/2.6.17-1.2139_FC5-i586/

I have thinkpad t43, with ATI Mobility Radeon X300. I did not install the ATI drivers.

Note that when I reverted back to 2133 kernel, everything went fine.

I'm seriously thinking of filing a bug about this one. Unless anyone points out an obvious cause, I will do so.

rj

Tuxic
2nd July 2006, 05:15 AM
After I upgraded to the 39 kernel, my laptop ALWAYS crashes after I suspend it.

When I wake it up (by opening the screen), the screen is messed up, with streaks of multiple colors.

it's not out of the realm of possibility that something else I updated on yum might have broken it. But the problem never existed until I upgraded to this kernel.

The kernel is/usr/src/kernels/2.6.17-1.2139_FC5-i586/

I have thinkpad t43, with ATI Mobility Radeon X300. I did not install the ATI drivers.

Note that when I reverted back to 2133 kernel, everything went fine.

I'm seriously thinking of filing a bug about this one. Unless anyone points out an obvious cause, I will do so.

rj

I thought you might find this page (http://www.thinkwiki.org/wiki/Problems_with_ACPI_suspend-to-ram) helpful.

Tuxic

idiotprogrammer
2nd July 2006, 05:13 PM
wow, I forgot about thinkwiki! Thanks!

I'll look at it more closely later, but a cursory look doesn't show anything that jumps out.

Because I can always reproduce it whenever I use the 39 kernel, this one is screaming "file a bug."

mrunge
10th July 2006, 06:49 AM
Hello,

is anybody aware of a solution? It looks like, idiotprogrammer has the same problem like me. Video is not initialized again after suspend.

erico
10th July 2006, 07:03 AM
Just so you know that you are not alone, my Dell Latitude C510 has problems as well -- but I think it may be related to an Internet service. One service in particular has been visibly misbehaving -- NNTP. Even though I turn off network time synchronization, it still insists on attempting to reactive upon logging on. This delays the boot process considerably. Interestingly, a long delay will cause the screen to switch to a 16-color framebuffer as it displays SWAP partition accessing info. After several seconds, it returns. I have noticed, however, that if my screen is not using any visible applications, it performs as specified.

mrunge
10th July 2006, 08:41 AM
@enrico: Are you sure, NNTP is responsible for the problem of not initialising graphics card after suspend? This sounds really weird!
Will you try the following and disable NNTP (temporarily)?
chkconfig --level 5 nntp off (and on to re-enable)

Does that change anything?

idiotprogrammer
10th July 2006, 05:01 PM
bugs filed on it (including my own which is probably a duplicate)
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=169201
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=197562

I've noticed this also on the 45 kernel.

Interestingly, I have that exact same problem with NNTP.

mrunge
11th July 2006, 08:04 AM
At least the first bugzilla entry (SATA drives fail) is afaik unrelated to the video corruption after suspend. I had that bug too (in earlier kernels). It shows like this: press the resume button (e.g. fn), laptop wakes up (moon led is off) and screen shows. Then the kernel tries to access hd and the whole system locks up.

I'm not really sure, if the second one is X-Server related. If yes, the corruption shouldn't occur during text mode (init 3). Has anyone tested that? My laptop is currently at home and I am unable to verify. Are we sure, it is realted to the kernel? Is the system working with kernel 2.6.16-*? Has anybody such an old kernel to verify?

mrunge
12th July 2006, 09:48 AM
Just tested a self compiled 2.6.16.24 kernel (vanilla). No problems with suspend any more. I assume, (at least) RedHats FC Kernels are broken at this point. I'll stay at 2.6.16 until they fix this problem.