I tried a self-compiled 3.5 kernel (compiled from patched source in the Debian experimental repository), but the problem persisted.
Looking through /var/log/messages (this under kernel 3.2.0-3-686-pae again), I see GPU lockups reported every minute or so: GPU lockup (waiting for 0x0000002E last fence id 0x0000002D) Sep 13 17:47:00 branci40 kernel: [ 829.917159] radeon 0000:01:00.0: GPU softreset Sep 13 17:47:00 branci40 kernel: [ 829.917162] radeon 0000:01:00.0: GRBM_STATUS=0xF5702828 Sep 13 17:47:00 branci40 kernel: [ 829.917165] radeon 0000:01:00.0: GRBM_STATUS_SE0=0xFC000005 Sep 13 17:47:00 branci40 kernel: [ 829.917168] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x00000007 Sep 13 17:47:00 branci40 kernel: [ 829.917171] radeon 0000:01:00.0: SRBM_STATUS=0x200000C0 Sep 13 17:47:00 branci40 kernel: [ 829.917196] radeon 0000:01:00.0: GRBM_SOFT_RESET=0x00007F6B Sep 13 17:47:00 branci40 kernel: [ 829.917300] radeon 0000:01:00.0: GRBM_STATUS=0x00003828 Sep 13 17:47:00 branci40 kernel: [ 829.917303] radeon 0000:01:00.0: GRBM_STATUS_SE0=0x00000007 Sep 13 17:47:00 branci40 kernel: [ 829.917306] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x00000007 Sep 13 17:47:00 branci40 kernel: [ 829.917309] radeon 0000:01:00.0: SRBM_STATUS=0x200000C0 Sep 13 17:47:00 branci40 kernel: [ 829.918304] radeon 0000:01:00.0: GPU reset succeed Jim -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: http://lists.debian.org/20120914005823.GA21979@ohlone
