X.org tight loop with SIGALRM

[greg]

Hi,

  is there a known bug with X.org and infinite SIGALRM looping?
I've googled for this and looked through every bug in your bugzilla that
mentions SIGALRM.

I'm seeing X using full cpu on some units in production.
If I do a strace of the X process I get this:

--- SIGALRM (Alarm clock) @ 0 (0) ---
sigreturn()                             = ? (mask now [])
--- SIGALRM (Alarm clock) @ 0 (0) ---
sigreturn()                             = ? (mask now [])

Repeating for ever. It looks like it gets a SIGALRM, handles it
then instantly gets another one.

This has so far happened with no discernable pattern for us.
I've also noticed a post on the debian mailing list from december
with pretty much identical symptoms
(http://lists.debian.org/debian-x/2005/12/msg00257.html

However he is using a Nvidia card and we are using a the via driver on a
via mini-itx with
01:00.0 VGA compatible controller: VIA Technologies, Inc. VT8623 [Apollo
CLE266] integrated CastleRock graphics (rev 03)
Although I've no reason to believe it's the graphics driver.

  thanks for any help you can provide,
     Greg