Ticket #901 (closed task: wontfix)
Sort out global mutex situation
| Reported by: | reportingsjr | Owned by: | |
|---|---|---|---|
| Priority: | High | Milestone: | beta3 |
| Component: | soy._internals | Version: | Trunk |
| Keywords: | Cc: |
Description
I'm on a core 2 duo, 64-bit, nvidia GeForce? 7300 GS with the nvidia accelerated graphics driver. My OS is ubuntu 7.10 64-bit.
Whenever I run anything having to do with a scene, which would be all pysoy files, I get a deadlock on the thread for the Scene thread.
If you need any more info just ask me on irc.
Change History
comment:1 Changed 5 years ago by ArcRiley
- Status changed from new to assigned
- Type changed from defect to task
- Summary changed from Threading deadlock to Sort out global mutex situation
comment:2 Changed 5 years ago by KirkMcDonald
Resolving this issue also has implications for the Callback action and its associated eventloop thread. Callbacks currently break if a callback makes certain (any?) calls back into PySoy. I am reasonably certain that this is a mutex issue.
comment:3 Changed 5 years ago by ArcRiley
- Owner ArcRiley deleted
- Status changed from assigned to new
I've done the most I can on this ticket without a local multicore system to test on.
I'm releasing this ticket to someone w/ multicore to resolve, or until I can purchase such a system myself.
comment:4 Changed 5 years ago by andar
I get this on a Core2Duo 6420 running Gentoo AMD64 stable with Cario 1.4.14.
Portage 2.1.4.4 (default-linux/amd64/2007.0, gcc-4.1.2, glibc-2.6.1-r0, 2.6.25.4 x86_64) ================================================================= System uname: 2.6.25.4 x86_64 Intel(R) Core(TM)2 CPU 6420 @ 2.13GHz
This is part of a greater issue of using bodies as our generic coreloop rendering lock.
Several new mutexes need to be added which coreloop cycles through and are held to perform certain tasks. These need to be well documented (in HACKING) and used in a uniform manner to ensure no deadlocks or segfaults arise on multi-core systems.