Re: Error: freeing invalid memtype

From: Ryan Hill
Date: Wed Apr 08 2009 - 23:50:41 EST


On Tue, 24 Mar 2009 23:25:16 +0100
Ingo Molnar <mingo@xxxxxxx> wrote:

> No, there was no fix in this area mainline, so it's this patch. I've
> queued up the fix for the .30 merge window with a -stable backport
> tag - so it will show up in .29.1 or .29.2 if it remains
> problem-free.
>
> Thanks a lot for your testing! Below is the final commit.

I'm running 2.6.29.1 which has this patch but I'm still seeing an insane
number of these. I can't trigger them just by running glxinfo like
Kevin, however opening a few tabs containing large images in firefox does the
trick and also freezes my system for about a minute.

cat /debug/x86/pat_memtype_list before the bug gives me ~17000 lines, of the form

write-combining @ 0xeff70000-0xeff71000
write-combining @ 0xeff71000-0xeff72000
write-combining @ 0xeff72000-0xeff73000
write-combining @ 0xeff73000-0xeff74000
write-combining @ 0xeff74000-0xeff75000
write-combining @ 0xeff75000-0xeff76000
write-combining @ 0xeff76000-0xeff77000
write-combining @ 0xeff77000-0xeff78000
write-combining @ 0xeff78000-0xeff79000
write-combining @ 0xeff79000-0xeff7a000

checking pat_memtype_list after triggering the bug produced ~25000 lines, so
I'm thinking a diff isn't really helpful.

you can find my kernel log @
http://dev.gentoo.org/~dirtyepic/misc/logs/20090508-kern.log

what other info do you need?


--
gcc-porting, by design, by neglect
treecleaner, for a fact or just for effect
wxwidgets @ gentoo EFFD 380E 047A 4B51 D2BD C64F 8AA8 8346 F9A4 0662

Attachment: signature.asc
Description: PGP signature