radeon stall

[Thomas Klausner <tk%giga.or.at@localhost>]

Hi!

With a kernel from Jan 7, I had it happen two times so far that X eats
99% CPU and syslog is flooded with:

radeon0: autoconfiguration error: error: ring 0 stalled for more than 492240msec
radeon0: autoconfiguration error: warn: GPU lockup (current fence id 0x00000000001fe0ac last fence id 0x00000000001fe0ae on ring 0)

about once a second.

I can kill -9 X, but the kernel messages continue.

Locking issue?
 Thomas