Bug 15873
Summary: | Panic: Compress_read returned -22 on some resume attempts | ||
---|---|---|---|
Product: | Power Management | Reporter: | Martin Steigerwald (Martin) |
Component: | Hibernation/Suspend | Assignee: | Nigel Cunningham (nigel) |
Status: | CLOSED INSUFFICIENT_DATA | ||
Severity: | normal | CC: | gmazyland, martin.steigerwald, nigel, rui.zhang |
Priority: | P1 | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | http://www.tuxonice.net:80/ | ||
Kernel Version: | 2.6.33.2-tp42-toi-3.1-lowmem-free-991-992-04964-gf00c7ec-dirty | Subsystem: | |
Regression: | Yes | Bisected commit-id: | |
Attachments: |
First of the mentioned patches from Nigel, version 2 of it
Second of the mentioned patches from Nigel tuxonice config for hibernate script happened again with 2.6.34.1 and tuxonice 3.1.1.1, photo of backtrace error still occurs with 2.6.36-rc3 + TOI 3.2-rc1, photo of backtrace |
Description
Martin Steigerwald
2010-04-29 07:33:34 UTC
Created attachment 26178 [details]
First of the mentioned patches from Nigel, version 2 of it
Created attachment 26179 [details] Second of the mentioned patches from Nigel Reason for these are Radeon KMS induced problems to free enough low memory pages: http://lists.tuxonice.net/pipermail/tuxonice-devel/2010-April/006075.html Created attachment 26180 [details]
tuxonice config for hibernate script
I am using 1.99-1.1 of debian hibernate package for squeeze/sid, without text user interface and with no_console_suspend kernel parameter to gather any messages of the kernel during snapshot cycles.
It's definitely a TuxOnIce issue - a known one. I just haven't managed to find the time to put into find into finding the cause (sorry!) Now I had this on my ThinkPad T42 with: martin@deepdance:~> cat /proc/version Linux version 2.6.34.1-tp23-toi-3.1.1.1-04990-g3a7d1f4 (martin@deepdance) (gcc version 4.4.4 (Debian 4.4.4-5) ) #2 PREEMPT Tue Jul 6 20:27:13 CEST 2010 I didn't see this yet on my ThinkPad T42, and with 2.6.33 I never saw it on my T23 as far as I remember. Hmm, I might should add, that I tried two more resumes of the same image and the machine always just rebooted after having read 150 MB of the caches. Created attachment 27078 [details]
happened again with 2.6.34.1 and tuxonice 3.1.1.1, photo of backtrace
It happened again with that kernel. This time I have a photo of the backtrace. Hope it helps. Thanks.
After 3-4 attempts and about 7 days uptime it happened again with my T23. It definately happens more often with 2.6.34 than with 2.6.33. I now switched compressor to LZF. Maybe that helps. I switched compression on my ThinkPad T23 from LZO to LZF. And since then I didn't get the error anymore, but with only 5 attempts so far, so I am not sure whether switching to LZF "fixed" it: deepdance:~> cat /sys/power/tuxonice/debug_info TuxOnIce debugging info: - TuxOnIce core : 3.1.1.1 - Kernel Version : 2.6.34.1-tp23-toi-3.1.1.1-04990-g3a7d1f4 - Compiler vers. : 4.4 - Attempt number : 5 - Parameters : 0 667656 0 1 0 0 - Overall expected compression percentage: 0. - Checksum method is 'md4'. 0 pages resaved in atomic copy. - Compressor is 'lzf'. Compressed 776593408 bytes into 359897499 (53 percent compression). - Block I/O active. - Max outstanding reads 714. Max writes 5. Memory_needed: 1024 x (4096 + 200 + 76) = 4476928 bytes. Free mem throttle point reached 983. - Swap Allocator enabled. Swap available for image: 229016 pages. - File Allocator active. Storage available for image: 0 pages. - I/O speed: Write 28 MB/s, Read 33 MB/s. - Extra pages : 26 used/500. - Result : Succeeded. I will have this one running for at least 10 or 15 attempts, let's see how it goes. Pedro has some more information on this bug: Date Thu, 29 Jul 2010 02:30:37 +0100 Subject kcryptd oops when resuming with TuxOnIce with KDB oops afterwards From Pedro Ribeiro I hit a bug when resuming with TuxOnIce. At the middle of a resume, it says Compress Read -22 and locks up. I caught the stack trace with kdb and took photos of that. I'm running 2.6.35-rc6 on a Lenovo T400. I have an encrypted LUKS partition (aes-cbc-essiv-128) which contains an LVM2 with my root, swap and home partitions inside. http://lkml.org/lkml/2010/7/28/478 Today I had this with a Dell Dimension 5100 workstation. I switched this one from lzo to lzf compression as well, as I didn't have another failure on my ThinkPad T23 after the switch. So it might really be something related to lzo compression in combination with TuxOnIce. I found the cause yesterday - it was a locking issue in TuxOnIce. I'm not sure why it's triggered more easily with LZO, but LZO isn't the cause. The fix hasn't yet been committed to my git trees, but will be soon. This should be fixed with the current release (3.2-rc2). Created attachment 29472 [details]
error still occurs with 2.6.36-rc3 + TOI 3.2-rc1, photo of backtrace
Hi Nigel, unfortunately this still happens with 2.6.36-rc3-tp42-toi-3.2-rc1-vmembase-0-05032-g60140c1-dirty.
In the interest of having a stable kernel I will switch to LZF as so far with LZF it didn't happen to me. In case you have any idea, on what debug information to gather, please tell me. I am on the jump, thus this time I haven't taken time for additional steps.
It's great that kernel bugzilla is back. can you please verify if the problem still exists in the latest upstream kernel? Thanks for the reminder, Zhang. This is soo old and not even related to upstream code. I just close it. I did not use TuxOnIce for the last time, I think I like to use it again, but then I can have a look again. |