system memory leakage due to suspend;vmap allocation failed: use vmalloc=<size> to increase size

[gene]

I am on 9.04. Started noticing that after a few suspend-to-rams, when the memory usage builds up, I lose wireless connectivity. It turns out to be failure to load the module ath_pci. When I try loading it with

"sudo modprobe -i ath_pci"

I receive the error message : FATAL: Error inserting module such and such: Cannot allocate memory. Then I try unloading some other module like snd_hda_intel and loading ath_pci. It does, however I cannot load the sound module back in with the same complaint. Next time I resume after suspend, I lose the atheros module again.

Of course, rebooting fixes the issue. Surprisingly, just kill the X will also do it.

I suspect, that this is related to X and/or suspend mechanism. I go an ATI chip and use the radeon driver.

I appreciate any help and suggestions.

I will upload some logs and useful files today.
Thanks

[gene (eugenios)]

 uname -a
Linux 2.6.28-15-generic #49-Ubuntu SMP Tue Aug 18 18:40:08 UTC 2009 i686 GNU/Linux

With no memory issue (after X was restarted):
vmstat
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
 r b swpd free buff cache si so bi bo in cs us sy id wa
 3 0 25616 506392 45640 579564 0 1 21 12 47 42 6 2 92 1

free
             total used free shared buffers cached
Mem: 1413372 907236 506136 0 45640 579620
-/+ buffers/cache: 281976 1131396
Swap: 1670628 25616 1645012

I also noticed this in the dmesg:
Sep 25 01:18:01 jenshen kernel: [298573.363150] vmap allocation failed: use vmalloc=<size> to increase size

I am uploading the last dmesg log anyways.

[gene (eugenios)]

Hello (to myself, it seems)
I continue with the warning message from both the messages log and the kern.log.0:

vmap allocation failed: use vmalloc=<size> to increase size

The fact the time when it turns up is the same I encounter problems with modprobe is not a mere coincidence.

With a little bit of search, I found that some folks suggest adding a boot parameter "vmalloc=some_sizeM", which I might try and report the results in the nearest future.

However, it appears to me that the kernel is not the "bad guy" here. Something is related to the X. Since killing the X will fix it.

[gene (eugenios)]

Interestingly, I found a similar bug #380696. I am not setting this one as a duplicate yet though...

[gene (eugenios)]

Ok, I feel lonely here :)
I am sure, that a lot of people experience this bug. They might not see it or else...
I continue monitoring my dmesg.
Right now, my virtual memory is not filled up yet. So I get
grep 'vmap ' /var/log/kern.log.0 | less

ion information and exit
jenya@jenshen:~$ free -l
             total used free shared buffers cached
Mem: 1413372 1325188 88184 0 114848 521444
Low: 878324 791356 86968
High: 535048 533832 1216
-/+ buffers/cache: 688896 724476
Swap: 1670628 35196 1635432

.....Nothing

I did not try different values for the boot parameters vmalloc. I do not understand it.
Well, I guess, I need to learn some kernel internals then...

[Jeremy Foshee (jeremyfoshee)]

Hi gene,

This bug was reported a while ago and there hasn't been any activity in it recently. We were wondering if this is still an issue? Can you try with the latest development release of Ubuntu? ISO CD images are available from http://cdimage.ubuntu.com/releases/ .

If it remains an issue, could you run the following command from a Terminal (Applications->Accessories->Terminal). It will automatically gather and attach updated debug information to this report.

apport-collect -p linux 436665

Also, if you could test the latest upstream kernel available that would be great. It will allow additional upstream developers to examine the issue. Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Once you've tested the upstream kernel, please remove the 'needs-upstream-testing' tag. This can be done by clicking on the yellow pencil icon next to the tag located at the bottom of the bug description and deleting the 'needs-upstream-testing' text. Please let us know your results.

Thanks in advance.

    [This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[gene (eugenios)]

Thanks Jeremy,
Finally I got a feedback. I do not experience this bug and can't remember how I fixed it. I do not use wireless device anymore (which might be the "fix"). I guess, the culprit was in the compcache which I disabled by
"rm -f /usr/share/initramfs-tools/conf.d/compcache && update-initramfs -u"
I do remember that it was annoying to see 100% cpu usage at some point and because of this issue too, perhaps.

I guess you can dismiss this case for the time being. Before I tag it "invalid" i will try a couple tests filling up the memory and unloading-loading some modules.

Thanks

[gene (eugenios)]

I ran a test. Unloaded snd_hda_intel module, then opened quite a few html files with firefox and konqueror, plus about 30 pdf files with evince simultaneously bringing the system to its memory "knees". Nevertheless, I could load the sound module back without any problem.
I guess the issue was fixed either by disabling the compcache, or with some updates.
The bug may be closed now.
Thanks