Created attachment 25993 [details] dmesg log of a KMS boot to runlevel 5 Booting current drm-radeon-testing (commit 3a44d81fb0e85b18484cae6a20f0661710b6c1db) on my AGP X1300 with KMS enabled reliably leads to a blank screen (monitor goes to sleep) when Xorg fires up. These two messages show up in dmesg: [drm:atom_op_jump] *ERROR* atombios stuck in loop for more than 1sec aborting [drm:atom_execute_table_locked] *ERROR* atombios stuck executing E1B9 (len 86, WS 4, PS 0) @ 0xE1EC This bug is here for a long long time but without the recent atombios loop detection logic every Xorg startup using KMS ended with a blank screen and locked up X (SSH working, X not responding to signals including SIGKILL, machine unable to shutdown properly). With this checking in place, X fortunately doesn't lock up and it's possible to kill it or switch VTs normally. KMS has never been working on this machine partly because of this issue and also because of the race reported in bug 15078 (however, that one seems to have vanished with recent kernels - without a relevant commit).
Created attachment 25994 [details] corresponding Xorg log
Created attachment 25995 [details] xorg.conf used Nothing important here IMHO, only accel disabled (the problem happens without xorg.conf, too).
Created attachment 25996 [details] video BIOS dump