USing XFree 4.3.0-r3 with kernel 2.6.0-test6 (mm2 and mm4) radeon DRM drivers and via-agp/agpgart causes 100% segfault. Problem disappears when I downgrade and returns when I upgrade. no clue to error in XFree log. Reproducible: Always Steps to Reproduce: 1. Emerge XFree 4.3.0-r3 2. access ANY opengl app (eg : glxinfo) 3. segfault Actual Results: segfault and app dies gracefully Expected Results: executed :) KT400 (epox 8k9a9I) onboard Via Rhine LAN Radeon 8500 LE 32MB SB Live! 512 MB PC2700 kernel 2.6.0-test6-mm4 kernel drm drivers for radeon
try using the kernel provided drm drivers, not sure if they work but worth a try
I *WAS* using the kernel provided DRM drivers in 2.6.0 test 6 mm4 (also tried mm2) I have reverted to the ati-drivers as a successful work around.
FWIW, I'm seeing the exact same problem; 2.6.0-test6 + XFree 4.3.0-r3.
I can confirm this. Please post workaround if one is known.
David, Sean, Harold: Is this the exact problem you're having? i.e., it only happens on: xfree-4.3.0-r3 via-agp 2.6-test6 kernel 1) Please test the latest development-sources and see if it still happens. 2) Also confirm whether it's an issue with xfree by testing xfree-4.3.0-r2 with latest development-sources. 3) and that all of you have Via AGP chipsets, and which one (lspci | grep AGP). 4) Also, which video cards do you all have? For Radeons, try Option "ForcePCIMode" "true" in Section "Device" of XF86Config to see if it's an AGP GART problem.
Donnie, I get the same result if I boot with 2.6.0-test4, so it isn't just 2.6.0-test6 -- the problem appears to be related to the XFree 4.3.0-r3 release, and there wasn't any problem with -r2. Also, I have an Intel 82845G/GL AGP chipset, not a Via. I'll try the ForcePCIMode, but I'm unlikely to try the 2.6.0-test7. To confirm: 2.6.0-test4 + 4.3.0-r2 : Worked 2.6.0-test4 + 4.3.0-r3 : Doesn't work 2.6.0-test6 + 4.3.0-r3 : Doesn't work Not a lot of data points, but it certainly looks like something that changed in XFree is causing this.
Same here : Via kt400 2.6.0-testX + 4.3.0-r2 : Worked 2.6.0-test4 + 4.3.0-r3 : Doesn't work 2.6.0-test6 + 4.3.0-r3 : Doesn't work 2.6.0-test7 + 4.3.0-r3 : Doesn't Work forcepci didnt help, neither did mm-sources for the same kernel levels in all combinations, using the kernel agpgart and the "ati-drivers-3.2.8" package was a successful work around. Cheers.
I have a radeon 8500SE (agp4x) 32MB Also trialled radeon 9200SE (agp8x) 128MB and had identical results.
oops; I have a kt400A mobo, not kt400!
Same exact problem here. xfree-4.3.0-r3 mm-sources-beta6-r4 agp = intel i845 vid = radeon mobility
In my opinion, this needs to get fixed ASAP, so a fix in -r4 is stable before 2.6 is released. Changing priority to Critical. Hopefully a new patchset I'm putting together right now will alleviate the problem.
Fixing summary.
Hi, I just posted a new patchset. On your next sync, do the following: grep ^PATCH_VER /usr/portage/x11-base/xfree/xfree-4.3.0-r3.ebuild Make sure you get 2.1.16. Please emerge this and see if you still have the trouble. If you do, I'll have to dig through patchsets for the differences between -r2 and -r3, and identifying the culprit could be interesting.
I reemerged xfree with the new patchset and I still have the same problem.
*** Bug 31033 has been marked as a duplicate of this bug. ***
I'm trying to track this down. Patches added/modified that are relevant seem to be: 1212_all_4.3.0-radeon-dpms-on-dvi-v2.patch 5117_all_4.3.0-radeon-support-backport-from-CVS-v2.patch 5100_all_4.3.0-radeon-support-from-ATI-backport-from-CVS-v2.patch 0100_all_4.3.0-xf-4_3-branch-10-17.patch 9132_all_4.3.0-Mesa-SSE-fixes-from-MesaCVS.patch After spotting bug #31033, it appears to be non-Radeon-specific, therefore I'm guessing 9132 is the culprit.
Created attachment 19387 [details, diff] Patch to ebuild to exclude patch #9132 Please try this patch to see if #9132 is the problem. # cd /usr/portage/x11-base/xfree # patch -p0 < /path/to/downloaded/patch # emerge xfree
followed instructions in comment above. new patch does seem to have fixed the problem ;) SSE on AMD problem? bus_drivr
I'd like confirmation from a few more people that the patch on Comment #17 fixes the issue before pushing it out.
Oh, yeah. That does it. GL works again (for me) with this patch on -r3. Sweeeet.
Just committed PATCH_VER 2.1.19, which pulled patch #9132. # grep ^PATCH_VER /usr/portage/x11-base/xfree/xfree-4.3.0-r3.ebuild Make sure you get the right version, and remerge xfree. Thanks for your help.
A minor correction to the way this works: We perform a compile-time check for a 2.4 kernel, and if that is the case, apply the patch. Otherwise, the patch isn't applied. Yes this screws things up when moving from a 2.4 to a 2.6 kernel, but it works for 2.6 people and it works for 2.4 people. After discussing some with Mike Harris, a runtime hack would probably be best. You may also be interested in following http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=105581.
*** Bug 32493 has been marked as a duplicate of this bug. ***
