With the 2.6.32-3 kernel (no .33 drm bits), I can boot into a KMS desktop with no 3d, and running the proper resolution.
When I try to do the same with the 2.6.32-4 (or even the 2.6.33 from experimental that I can't ultimately use without the installable headers that it currently lacks), I get a black screen on boot and then my monitor just shuts off.
If I try the .32-4 or .33 kernel without trying to enable kms, I get 3d and a maximum resolution of 1152x864 instead of 1600x1200.
I figure there's something I can do with an xorg.conf (like for starters creating one!) to fix the resolution issue, but it seems like there's always at least one upstream project that isn't current enough in Debian for me to not need fglrx anymore. The radeonBuildHowTo says that I shouldn't be able to use KMS on my rv740 without mesa 7.8-rc1, but I've been able to use with the 3d-less .32 kernel.
Apparently I can't even get a kernel built anymore either, my attempts to build-in the firmware files keep failing with zero useful information being output by the build process upon failure. "error 2" yeah that's useful!
When I try to do the same with the 2.6.32-4 (or even the 2.6.33 from experimental that I can't ultimately use without the installable headers that it currently lacks), I get a black screen on boot and then my monitor just shuts off.
If I try the .32-4 or .33 kernel without trying to enable kms, I get 3d and a maximum resolution of 1152x864 instead of 1600x1200.
I figure there's something I can do with an xorg.conf (like for starters creating one!) to fix the resolution issue, but it seems like there's always at least one upstream project that isn't current enough in Debian for me to not need fglrx anymore. The radeonBuildHowTo says that I shouldn't be able to use KMS on my rv740 without mesa 7.8-rc1, but I've been able to use with the 3d-less .32 kernel.
Apparently I can't even get a kernel built anymore either, my attempts to build-in the firmware files keep failing with zero useful information being output by the build process upon failure. "error 2" yeah that's useful!
Comment