Bug 742032
| Summary: | [abrt] kernel: WARNING: at arch/x86/xen/mmu.c:475 xen_make_pte+0x32/0x8e(): TAINTED ---------W | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Jerry Amundson <jamundso> | ||||||
| Component: | kernel | Assignee: | Kernel Maintainer List <kernel-maint> | ||||||
| Status: | CLOSED ERRATA | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||||
| Severity: | unspecified | Docs Contact: | |||||||
| Priority: | unspecified | ||||||||
| Version: | 16 | CC: | Dmitry.Bogomolov, drjones, gansalmon, itamar, jonathan, jonathanjstevens, kernel-maint, ketuzsezr, madhu.chinakonda, raphoszap | ||||||
| Target Milestone: | --- | ||||||||
| Target Release: | --- | ||||||||
| Hardware: | x86_64 | ||||||||
| OS: | Unspecified | ||||||||
| Whiteboard: | abrt_hash:1df95ad1faa006524cc8fbf050f78320479f15a5 | ||||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||||
| Doc Text: | Story Points: | --- | |||||||
| Clone Of: | Environment: | ||||||||
| Last Closed: | 2012-07-17 12:40:28 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: | |||||||||
| Bug Depends On: | |||||||||
| Bug Blocks: | 741684 | ||||||||
| Attachments: |
|
||||||||
|
Description
Jerry Amundson
2011-09-28 20:07:52 UTC
Can you pass 'nopat' on the Linux command line and see if that makes the warning go away? What is the ATI/AMD card you have in your machine? (In reply to comment #1) > Can you pass 'nopat' on the Linux command line and see if that makes the > warning go away? OK, will do when I have access to it on Tuesday. > What is the ATI/AMD card you have in your machine? OTTOMH, I think it's an RV610. Package: kernel Architecture: x86_64 OS Release: Fedora release 16 (Verne) Comment ----- Sorry, I'm just responding to the ABRT alert and reporting the bug cos I'm nice. abrt reported a xen kernel crash. system is still running OK though (not currently running any VMs though). I was just browsing the net at the time, only Firefox running (In reply to comment #1) > Can you pass 'nopat' on the Linux command line and see if that makes the > warning go away? I've confirmed that it does. > What is the ATI/AMD card you have in your machine? RV610: [ 23.850] (--) RADEON(0): Chipset: "ATI Radeon HD 2400 XT" (ChipID = 0x94c1) (In reply to comment #1) > Can you pass 'nopat' on the Linux command line and see if that makes the > warning go away? Konrad, as Jerry reported 'nopat' makes this go away. Do you know if there is an upstream change pending in this area or have any further ideas? Created attachment 529979 [details]
pvops: make pte_flags() go via pvops
Hey Josh,
There is this patch that will make the warning (and the problem as far as I can remember) go away. But ... I *think* I tested it when I saw this with my Radeon box and while the warnings did go away I saw ext4 and userspace programs dying with weird errors. The reason for all of this is that pages are being returned by TTM to the main memory management which then eventually tries to recycle them (and coalesce collections of them in 2MB pages). Anyhow, some of those pages end up being used by applications, other kernel subsystems, etc.. except that some of them seem to have Write Combined flags set! Which is really bad for some applications.
It also could be that when I compiled a kernel with this patch I had booted the wrong kernel and in effect never tested the patch :-(
So what I am saying is - yes there is a patch, but I am not yet sure if this the right cure for the problem. Need to investigate this more.
Created attachment 532138 [details]
Use pte_val instead of pte_flags to retrieve the page flags when changing caching.
The previous patch did make it work, but it is such a big hammer that I am not comfortable with it. The patch attached here solves the problem for me. I would appreciate if somebody tested it besides me just to make sure it works.
Note, the patch has not yet been posted on LKML, but will shortly.
This seems like 2/3 in this series: http://thread.gmane.org/gmane.comp.emulators.xen.devel/115063/focus=1212837 which is (hopefully?) going to get queued up for -stable. Package: kernel Architecture: x86_64 OS Release: Fedora release 16 (Verne) Comment ----- Install xen *** Bug 765604 has been marked as a duplicate of this bug. *** A temporary fix went into 3.3 with commit 8eaffa67b43e99ae581622c5133e20b0f48bcef1. Konrad, is that going to be brought back to 3.2 stable? *** Bug 787403 has been marked as a duplicate of this bug. *** *** Bug 802532 has been marked as a duplicate of this bug. *** [mass update] kernel-3.3.0-4.fc16 has been pushed to the Fedora 16 stable repository. Please retest with this update. [mass update] kernel-3.3.0-4.fc16 has been pushed to the Fedora 16 stable repository. Please retest with this update. [mass update] kernel-3.3.0-4.fc16 has been pushed to the Fedora 16 stable repository. Please retest with this update. (In reply to comment #11) > A temporary fix went into 3.3 with commit > 8eaffa67b43e99ae581622c5133e20b0f48bcef1. > > Konrad, is that going to be brought back to 3.2 stable? I should do that. Let me email Greg so that he can queue them up. Thanks *** Bug 745574 has been marked as a duplicate of this bug. *** This was fixed in 3.3/3.4 |