| Summary: | sys-kernel/gentoo-sources-3.4.4 - radeon driver with ColorTiling2D causes display corruption | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Marcus Becker <marcus.disi> |
| Component: | [OLD] Core system | Assignee: | Gentoo Kernel Bug Wranglers and Kernel Maintainers <kernel> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | x11 |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | AMD64 | ||
| OS: | Linux | ||
| Whiteboard: | linux-3.5 libdrm-2.4.37 xorg-server-1.12.99.903 mesa-8.1_rc1 empty-xorg-config | ||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
emerge --info
kernel config gentoo-sources 3.4.4 Xorg.0.log log during bisect git bisect log Xorg.0.log with 9999 drivers display corruption with mirrored displays |
||
|
Description
Marcus Becker
2012-07-01 12:41:27 UTC
Created attachment 316861 [details]
emerge --info
Created attachment 316863 [details]
kernel config gentoo-sources 3.4.4
I created it using 'make oldconfig' on the 3.3.8 config and chose default values.
Created attachment 316889 [details]
Xorg.0.log
I first tried the pre setup of my existing user, logged into gdm and the screen got scrambled
created a new user 'peng' and he can login with mirrored displays, if I switched to extend the screen gets scrambled
switching back to 3.3.8 fixes the problem
I also rebuild xorg-server and all x11 drivers with the new kernel headers and kernel
This happens since version 3.4.0...
Can you do a git bisect between vanilla-sources-3.3.8 and 3.4.0 to see if we can narrow the commit that caused the error? http://www.reactivated.net/weblog/archives/2006/01/using-git-bisect-to-find-buggy-kernel-patches/ Not today :( Will try tomorrow... btw. http://wiki.gentoo.org/wiki/Kernel_git-bisect ;) Created attachment 318598 [details]
log during bisect
Created attachment 318600 [details]
git bisect log
I had to do it between 3.4.1 and 3.3.8, it wouldn't take 3.4.0 or 3.4
hope that helps :)
drm/radeon: add support for evergreen/ni tiling informations v11 http://lkml.indiana.edu/hypermail/linux/kernel/1206.0/02997.html This is their solution, have to try it out: Update your libdrm to libdrm from git should fix your issue. Created attachment 318754 [details]
Xorg.0.log with 9999 drivers
Exactly the same result as before on gentoo-sources-3.4.5 with xf86-video-ati-9999 and libdrm-9999
It works fine with a single display or mirror but not to extend on the external monitor, which scrambles the whole display. Without ColorTiling2D I can extend to the external monitor but popups are scrambled.
Related packages:
[ebuild R ] x11-libs/libdrm-2.4.33 USE="libkms -static-libs" VIDEO_CARDS="radeon -intel -nouveau (-omap) -vmware" 0 kB
[ebuild R ] x11-apps/xrandr-1.3.5 0 kB
[ebuild R ] sys-kernel/gentoo-sources-3.4.5 USE="-build -deblob -symlink" 0 kB
[ebuild R ] media-libs/mesa-8.0.3-r1 USE="egl g3dvl gallium gbm gles1 gles2 llvm nptl openvg shared-glapi vdpau xa xvmc -bindist -classic -d3d -debug -osmesa -pax_kernel -pic (-selinux) -shared-dricore -wayland" VIDEO_CARDS="radeon -i915 -i965 -intel -nouveau -r100 -r200 -r300 -r600 -vmware" 0 kB
[ebuild R ] x11-base/xorg-server-1.12.3 USE="ipv6 kdrive nptl udev xorg -dmx -doc -minimal (-selinux) -static-libs -tslib -xnest -xvfb" 0 kB
[ebuild R ] x11-drivers/xf86-video-ati-6.14.4-r1 0 kB
Where is works just fine pre linux-3.4
Created attachment 318756 [details]
display corruption with mirrored displays
If there is a popup window
With everything turned off and default settings (empty xorg.conf) it works fine on gentoo-sources-3.5 with x11-drivers/xf86-video-ati-6.14.6-r1 and x11-libs/libdrm-2.4.37 and x11-base/xorg-server-1.12.99.903 and media-libs/mesa-8.1_rc1_pre20120724::x-portage Seems the issue has been mitigated then. Between 3.3.8 and 3.5.7 there's only 3.4.9 now which sounds like a sufficient upgrade from 3.4.4 and might even not experience this anymore; and if it wasn't the kernel, then there have been quite some newer libdrm versions as well. If current stable versions still reproduce this for you or others, it might be worth attempting to stabilize the newer working versions... Another option is to undo the patch, although I don't know if that has consequences. |
