On Thu, 2008-12-11 at 23:05 -0600, Michael Cronenworth wrote:
Has anyone else pulled this testing package? I was premature in
giving
this positive karma. I've had X.org crash three times after I gave it a
thumbs up. Every time I get this at the end of my Xorg.0.log or Xorg.1.log.
That crash doesn't hit the evdev driver at all. It's a bug in the
server if it's anything.
What sort of desktop are you running? This appears to be in the
save-under code, which I thought was quite rarely used these days.
Backtrace:
0: /usr/bin/Xorg(xorg_backtrace+0x26) [0x4e7a26]
1: /usr/bin/Xorg(xf86SigHandler+0x39) [0x47a679]
2: /lib64/libc.so.6 [0x3cf3e32f60]
3: /usr/bin/Xorg(miHandleValidateExposures+0x6c) [0x4dfa1c]
Frame #3 is the place where we fault. #2 is the signal trap in glibc,
and #1 and #0 are are the server's fault handler code, so they're not
the problem. So what's happening there? Well, after installing the
server debuginfo package:
[ajax@benzylpiperazine ~]$ echo 0x4dfa1c | eu-addr2line -e
/usr/lib/debug/usr/bin/Xorg.debug
/usr/src/debug/xorg-server-1.5.3/mi/miwindow.c:453
[ajax@benzylpiperazine ~]$ cat -n /usr/src/debug/xorg-server-1.5.3/mi/miwindow.c | head
-455 | tail -5
451 if ( (val = pChild->valdata) )
452 {
453 if (REGION_NOTEMPTY(pScreen, &val->after.borderExposed))
454 miPaintWindow(pChild, &val->after.borderExposed, PW_BORDER);
455 REGION_UNINIT(pScreen, &val->after.borderExposed);
Well, val should be non-null here, by the check on line 451, but I
suppose that doesn't necessarily imply that it's a valid pointer. Hmm.
- ajax