Bug 551937
Summary: | Frequent "page allocation failure. order:2, mode:0x4020" with 2.6.31.5-122.fc12.x86_64 on X61 BTRFS Intel 4965 AGN GM965 video | ||||||
---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Gregory Maxwell <gmaxwell> | ||||
Component: | kernel | Assignee: | Stanislaw Gruszka <sgruszka> | ||||
Status: | CLOSED NEXTRELEASE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||
Severity: | medium | Docs Contact: | |||||
Priority: | low | ||||||
Version: | 12 | CC: | anton, dougsland, gansalmon, gjunk, itamar, kernel-maint, renard, sgruszka | ||||
Target Milestone: | --- | ||||||
Target Release: | --- | ||||||
Hardware: | x86_64 | ||||||
OS: | Linux | ||||||
Whiteboard: | |||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | Environment: | ||||||
Last Closed: | 2010-07-22 09:28:50 UTC | Type: | --- | ||||
Regression: | --- | Mount Type: | --- | ||||
Documentation: | --- | CRM: | |||||
Verified Versions: | Category: | --- | |||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
Cloudforms Team: | --- | Target Upstream Version: | |||||
Embargoed: | |||||||
Attachments: |
|
Description
Gregory Maxwell
2010-01-03 07:49:50 UTC
Ah, I should have mentioned that I have /tmp on tmpfs: tmpfs 1.5G 119M 1.4G 8% /tmp So I don't think the swap in use shouldn't have been indicative of any prior exhaustion. (nor have I seen any userspace evidence of memory exaustion) Hi Gregory The bug is not fixed upstream, fixing it in not so easy. We have upstream bug entry for it here: http://bugzilla.kernel.org/show_bug.cgi?id=14141 Are you able to reproduce? There are some upstream patches to test, could you test 2.6.33-rcX vanilla and help solving this bug on bugzilla.kernel.org ? Gregory, please note we have now koji build for 2.6.33 based kernels (see http://koji.fedoraproject.org/koji/packageinfo?packageID=8) If issue is not present in 2.6.33 and you can reproduce it in 2.6.32, I think we have to backport iwlwifi "rx paged" patch, fixing that memory management can be _very_ hard. I am seeing something very similar tho I do not use BTRFS ... just iwlagn. https://bugzilla.redhat.com/show_bug.cgi?id=569412 Is that a dup of this one ? Hi Gregory Please tell us if you are able to reproduce that bug. If so, I will prepare kernel for test. I'd just like to add a "me too". I'm running F12 x86_64 in a default install (no BTRFS or anything) on a Thinkpad T61p with 3 GB ram. Yesterday I got page allocation failures like the ones shown in the attached log file - I was pushing the memory limits and beginning to use some swap when the network connection started dropping out. I could apparently reproduce the problem by making sure that running processes took up a lot of memory and then doing something that causes network traffic like checking out a source tree from SVN. Created attachment 399277 [details]
Snippet from /var/log/messages
Could you please test kernel from. http://koji.fedoraproject.org/koji/taskinfo?taskID=2045788 Would be nice if you can assure that bug is reproducible fedora released kernel and not reproducible on kernel from above link. Bug should be fixed in F-13. |