X.org tight loop with SIGALRM

[Adam Jackson]

On Friday 27 January 2006 07:27, Felix Schulte wrote:
> On 1/27/06, greg <greg at lincor.com> wrote:
> > Hi,
> > 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 [])

The X server uses SIGALRM to effect client scheduling.  If you're seeing 
multiple in a row, then the server is hung doing something else.

Try getting a backtrace from gdb instead of using strace.

> Try whether turning off the "smart scheduler" fixes the problem. If
> this works then you hit
> https://bugs.freedesktop.org/show_bug.cgi?id=467 again... :-(

If disabling the smart scheduler fixes the bug I would be quite surprised.  
It's far more likely that the server is hung either in some system call or in 
an infinite loop in some rendering path.

- ajax
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.x.org/archives/xorg/attachments/20060130/0dc6983e/attachment.pgp>