nV News Forums

 
 

nV News Forums (http://www.nvnews.net/vbulletin/index.php)
-   NVIDIA Linux (http://www.nvnews.net/vbulletin/forumdisplay.php?f=14)
-   -   Segfault inside driver, GF4MX 61.11 (http://www.nvnews.net/vbulletin/showthread.php?t=35958)

ggambett 08-31-04 10:38 PM

Segfault inside driver, GF4MX 61.11
 
I'm getting a segfault using OpenGL, inside the driver :
Code:

#0  0x4049942f in memcpy () from /lib/libc.so.6
#1  0x40be17fc in _nv001155gl () from /usr/local/lib/libGLcore.so.1
#2  0x40a1dd34 in _nv000028gl () from /usr/local/lib/libGL.so.1
#3  0x830c408b in ?? ()
Cannot access memory at address 0x424448b

This is with drivers 61.11 just installed. Was doing almost the same with 44.96.

The app is a game I'm developing so I can offer some additional info. There is a terrain which uses ROAM, so the triangle count is variable and view-dependent, as is the segfault (it doesn't happen always, it depends on where the camera is). Besides, this started happening when I added a specific model to the scene. Removing the model stops the segfault. Keeping the model and removing the terrain stops the segfault. Keeping the terrain and using a model with similar index counts stops the segfault.

As far as I could see, when the segfault happened, the terrain engine was using 4097 vertexes in a vertex buffer (and about 21K indices in an index buffer - and yes, I have more than 4097 vertexes allocated). Not every time I used 4097 vertexes it crashed though. Is there any low limit in the size of the vertex or index buffers that I may be hitting?

ggambett 08-31-04 10:54 PM

Re: Segfault inside driver, GF4MX 61.11
 
2 Attachment(s)
More details...
System is Athlon XP 1800+, 512 MB RAM, Kernel 2.4.20, XFree86-4.2.0-72. Logs attached.

zander 09-01-04 04:43 AM

Re: Segfault inside driver, GF4MX 61.11
 
If you can, please provide a simple test case that reproduces the crash.

ggambett 09-01-04 09:52 AM

Re: Segfault inside driver, GF4MX 61.11
 
Quote:

Originally Posted by zander
If you can, please provide a simple test case that reproduces the crash.

Unfortunately, I can't - even when I make subtle changes (rendering the terrain at slightly lower quality) the bug goes away, so I think the conditions to trigger it are quite specific.

I could upload the whole project precompiled, you could be able to debug inside the driver code (which I can't). Do you want me to do that?

zander 09-01-04 10:07 AM

Re: Segfault inside driver, GF4MX 61.11
 
That would help; please send a bug report with all relevant information to linux-bugs@nvidia.com (including X/nvidia-bug-report log files, configuration details, ...).

ggambett 09-01-04 10:31 AM

Re: Segfault inside driver, GF4MX 61.11
 
Thank you, I'll do that. Just in case you can't reproduce it, will it help if I also include a core dump?

zander 09-01-04 10:57 AM

Re: Segfault inside driver, GF4MX 61.11
 
Please do.

ggambett 09-02-04 09:52 PM

Re: Segfault inside driver, GF4MX 61.11
 
An update : I found and fixed a bug inside the terrain engine of the rendering engine I'm using, and I'm pretty sure this was the cause of the segfault. I'm really sorry. Shall I delete this thread?


All times are GMT -5. The time now is 04:11 PM.

Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.