[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
Title: | open3d |
Notice: | Kits on notes 3 and 4; Documents note 223 |
Moderator: | WRKSYS::COULTER |
|
Created: | Wed Dec 09 1992 |
Last Modified: | Fri Jun 06 1997 |
Last Successful Update: | Fri Jun 06 1997 |
Number of topics: | 1306 |
Total number of notes: | 5260 |
1262.0. "Xserver crashes running ported OpenGL demo." by PEACHS::LAMPERT (Pat Lampert, UNIX Applications Support, 343-1050) Fri Mar 14 1997 18:15
A company called Analytical Methods in Seattle is in
the process of porting their OpenGL package to Alpha.
Currently they have an Alphastation 255/300 running DUX
4.0b with a zlxp-l2 and Open3D 4.2.
The demo they have ported came from their SGI machine and
compiled without changes on the Alpha. If they run this
demo on the SGI and display it to the Alpha it completes
just fine. Unfortunately when they run their demo on the
Alpha the Xserver crashes with the following traceback...
signal Segmentation fault at
warning: PC value 0x0 not valid, trying RA
> [_NValPB, 0x30006898f7c] ldah gp, 16454(r26)
(dbx) where
0 _NValPB(0x3ecccccd, 0xed, 0x1, 0x0, 0x1401eaef0) [0x30006898f78]
1 N_DispatchStream(0x1401eaef0, 0x3ff00000108, 0x4500000002, 0x1401eb060, 0x1
0800000108) [0x30006890e10]
2 glExecStreamDEC(0x3ffbf6cd2a0, 0x14019ed00, 0x1402ac864, 0x14, 0x1402ac86c)
[0x3ffbf6d2b30]
3 glProcRender(0x1402ac864, 0x10, 0x3fffff8df18, 0x0, 0x3fffff8dcd0) [0x3ffbf
6cd29c]
4 glProcDispatch(0x14014f500, 0x0, 0x3fffff8dcd0, 0x14019ed00, 0x14019ed00) [
0x3ffbf6be6dc]
5 Dispatch(0x3fffff80600, 0x3fffff8dcc8, 0x3fffff8dd20, 0x140014d40, 0xffffff
fffffffff8) [0x3ffbfeff3e4]
6 dix_main(0x3ffbff17998, 0x140004760, 0x100000000, 0x14000a240, 0x10000000a)
[0x3ffbff181dc]
If they run the demo from the Alpha and display
it to the SGI machine, it almost immediatly hangs.
This looks like some sort of client side porting problem,
so you probably will need the sources. Does the Xserver
traceback indicate anything here?
If we do need the sources to troubleshoot this the customer
has indicated he will provide them.
At this point all I have is an executable. The executable
produces the exact same crash on my as600 with zlxp-l1
graphics.
Pat
T.R | Title | User | Personal Name | Date | Lines |
---|
1262.1 | Humm... Problem seems to be timing or local transport | PEACHS::LAMPERT | Pat Lampert, UNIX Applications Support, 343-1050 | Wed Mar 19 1997 14:51 | 9 |
| If I set the DISPLAY to localnodename:0 then the demo
works fine! When the DISPLAY environment variable is
not used and the Xservers file has :0, the demo causes
the Xserver to consistently crash as noted in .0.
Since I have a workaround now. I will file a priority 3
IPMT on the problem.
Pat
|
1262.2 | It is dying in the 'stream' extension | VESPER::VESPER | OpenGL Alpha Geek | Wed Mar 19 1997 17:16 | 33 |
| It appears that there is a bug in the stream extension, a proprietary
and hidden extension that we implemented to speed up local display.
The only way that this is not a bug in the stream extension is if the
application is writing trash over our buffers.
Some environment variables can be set to manage this extension.
$ setenv NOGLSTREAM 1
This will turn off the stream extension completely. If this is not
defined, then we will try to use the stream extension.
$ setenv NOGLSHAREDMEMORY 1
This will put the stream extension in a mode where it uses the X11 transport
buffer. If this is not defined, then we will use a shared memory buffer.
$ setenv GLSTREAMBUFFERSIZE number-of-pages-for-buffer (default 10)
This will change the size of the shared memory buffer.
If you get different behavior changing only the stream buffer size and
shared memory flag, it is possible that the application is writing trash
in various places.
If only turning off the stream extension entirely allows the application
to complete, then it is even more likely to be a bug in the stream extension.
There are things we can do with just an executable, so be sure to include
that in the IPMT.
Andy V
|