Comment 13 for bug 1059276
Revision history for this message
|
|
#13 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
a60fb26
(Get the code!)
(In reply to comment #5)
> Does it always hang inside of XRRGetCrtcInfo? That function doesn't do anything
> special and I don't really see why this should indefinitely wait for a reply
> from the X11 server.
No, it hangs inside all sorts of different functions. Sometimes it's just a direct _XReply call (not sure how that happens). The most common parent (from maybe 10-20 runs) is:
6 0xf6f332d5 in libgl.so.1 (+0x8f2d4) (0x022de488)
> Also, unlocking the Xlib display before polling for the
> reply shouldn't make any difference on this at all. :-/
I'm not sure what you mean here, do you mean in attachment 66895 where I remove the unlock so that it stays locked while it goes through xcb_wait_for_reply? I suspect what's happening is that one thread is depleting the fd of events, so the other thread hangs inside the poll() virtually indefinitely. (So if there's a lock around the poll() and recv() then that's not possible)