On Wed, 21 Mar 2007, Roland Scheidegger wrote:

> malc wrote:
>> Hello,
>>
>> Sometimes (in my case quite frequently) Mesa code triggers a hardware
>> exception and the process receives a SIGFAULT/ILL/whatever, normally
>> this is not much of the problem unless it was done while preceding
>> code held a lock, which sadly happens a lot to me.
>>
>> For instance calling glCopyTexSubImage2D while having
>> GL_FRAGMENT_SHADER_ATI
>> enabled goes through swrast path and crashes becasue one of the pointers
>> to function field is NULL and R200 did a LOCK_HARDWARE before.
> You could do what the r300 driver does (which has configurable span
> locking). If it crashes when doing a swrast fallback, the lock was in
> r200SpanRenderStart, you can just call UNLOCK_HARDWARE immidiately after
> LOCK_HARDWARE, and remove the UNLOCK_HARDWARE in r200SpanRenderFinish.

Sure this would help me with this particular problem. I have several
machines over here they are all connected to a hub (not even a router)
and more often then not i have the main one receive IP address from a
cable modem, to pin point this particular swrast problem i had to
configure the IP statically so that i can ssh to this box run the
offending code via SSH and explore the results in gdb, but sometimes
something inside Mesa faults while i'm on DHCP lease and i can't even
connect to the box to kill X remotelly or attach gdb to the failed
process so the only solution is to power off.

Perhaps there's something i can do to avoid it?

-- 
vale

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Mesa3d-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mesa3d-dev

Reply via email to