Weird display problem in recent current

[Kevin Oberman]

> From: Randy Bush <randy at psg.com>
> Date: Sat, 16 Apr 2005 08:03:09 -1000
> 
> kevin, what was the last cvsup date which gave you a workable current
> on your thinkpad.  i will revert to it and see if it addresses the
> inability to start my window system on a t41

Randy,

I thought I knew the answer to this, but decided to confirm it before
replying. Good thing!

I can state without a doubt that it worked on Mar. 31 with the Scott's
ATAPICAM patches. (They were not committed until late on 4/5.) The
failure certainly did show up by 4/8. I am currently building a system
at 4/5 just after the last ATAPICAM patch of the day was committed. 

Doug,

ktrace on the hung process shows nothing. Other than context switches,
there are no kernel events that appear to originate from gkrellm. Since
the problem can take hours to show up, starting ktrace and waiting to see
what has been captured may take a while and fill a LOT of disk.

If I can get a better handle on just when it happened, I may try to fill
a backup disk with ktrace data and see if I can find anything.

Two notes: I am running ULE on the system and it seems like it might
tend to happen when the CPU is throttled back to about 300 MHz by TCC
and ST. (This is just conjecture at this point.)

I have only had X freeze on initialization once. What Randy is seeing my
be a totally separate problem from what I am seeing with gkrellm (or
not).

Still hunting.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman at es.net			Phone: +1 510 486-8634