Hello Alex,
The workaround is to get matlab to use software opengl. We added the
following to the matlab startup script at line 497 (version 2009b):
>
> #####################################################
> # Fix for SunRay issue running 3D Graphics via OpenGL
> # Make sure that matlab uses software OpenGL
> # 07/12/09, Lee Heagney
> #####################################################
> ##################################################
> # If $UTAUDIODEV is not set, then set it to "NONE"
> ##################################################
> : ${UTAUDIODEV:="NONE"}
> if [ $UTAUDIODEV != "NONE" ]; then
> #################################################
> # The host is a SunRay, so enable software openGL
> #################################################
> usemesalinux=1
> fi
> ############
> # End of Fix
> ############
This forces matlab to link in its own MESA library.
Wilkinson, Alex wrote:
> Hi all,
>
> Using RHEL 5.3 (2.6.18-128) + SRSS 4.2 - x86_64.
>
> SunRay-2FS + Firmware Revision 4.2_140993-01_2010.01.21.17.35.
>
> Problem
> -------
>
> When running the "bench" command in Matlab2009a (x86_64) Xnewt segfaults.
> This ONLY happens when the '3D rendering' part of the benchmark kicks in.
>
> The result - session completely dies and resets.
>
> dmesg(1) output
>
> Xnewt[8330]: segfault at 00000000000166e0 rip 00000000082ba0e3 rsp
> 00000000ffca55a4 error 4
>
> /var/opt/SUNWut/log/messages output
>
> Apr 15 11:10:35 stlsux05 gdm[6253]: Handling message: 'XPID 12744 0'
> Apr 15 11:10:35 stlsux05 gdm[6253]: Got XPID == 0
> Apr 15 11:10:35 stlsux05 gdm[6253]: Handling message: 'SESSPID 12744 0'
> Apr 15 11:10:35 stlsux05 gdm[6253]: Got SESSPID == 0
> Apr 15 11:10:35 stlsux05 gconfd (myusername-13562): Received signal 15,
> shutting down cleanly
> Apr 15 11:10:35 stlsux05 gdm[12744]: slave_waitpid: done_waiting
> Apr 15 11:10:35 stlsux05 gdm[12744]: Session: start_time: 1271300046
> end_time: 1271301035
> Apr 15 11:10:35 stlsux05 gdm[12744]: Sending SESSPID == 0 for slave 12744
> Apr 15 11:10:35 stlsux05 gdm[6253]: Handling message: 'SESSPID 12744 0'
> Apr 15 11:10:35 stlsux05 gdm[6253]: Got SESSPID == 0
> Apr 15 11:10:35 stlsux05 gdm[12744]: gdm_slave_session_stop: myusername
> on :11
> Apr 15 11:10:35 stlsux05 gdm[12744]: Fatal X error detected. Ignoring
> same during session shut down.
> Apr 15 11:10:35 stlsux05 gconfd (myusername-13562): Exiting
> Apr 15 11:10:35 stlsux05 gdm[12744]: gdm_slave_session_stop: back here
> from xioerror
> Apr 15 11:10:35 stlsux05 gdm[12744]: gdm_slave_session_stop: Running post
> session script
> Apr 15 11:10:35 stlsux05 gdm[12744]: gdm_auth_user_remove: Removing
> cookie from /home/myusername/.Xauthority (0)
> Apr 15 11:10:35 stlsux05 gdm[12744]: gdm_auth_purge: :11
> Apr 15 11:10:35 stlsux05 gdm[12744]: Running gdm_verify_cleanup and pamh
> != NULL
> Apr 15 11:10:35 stlsux05 gdm[12744]: Running pam_close_session
> Apr 15 11:10:35 stlsux05 gdm[12744]: Running pam_setcred with
> PAM_DELETE_CRED
>
> gdb(1) backtrace (looks to be related to MESA).
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x082ba0e3 in xmesa_check_and_update_buffer_size (xmctx=0x0,
> drawBuffer=0xaf48220) at xm_api.c:1887
> 1887 xm_api.c: No such file or directory. in xm_api.c
> (gdb) bt
> #0 0x082ba0e3 in xmesa_check_and_update_buffer_size (xmctx=0x0,
> drawBuffer=0xaf48220) at xm_api.c:1887
> #1 0x082baeb3 in XMesaResizeBuffers (b=0xaf48220) at xm_api.c:2539
> #2 0x08267a81 in __glXMesaDrawableResize (base=0xa7c4228) at
> xf86glx.c:113
> #3 0x082282a1 in PositionWindow (pWin=0xa805f80, x=519, y=104) at
> glxscreens.c:230
> #4 0x080a3d5a in ResizeChildrenWinSize (pWin=0xa8055f0, dx=0, dy=0,
> dw=0, dh=-25) at window.c:1867
> #5 0x081f31f1 in miSlideAndSizeWindow (pWin=0xa8055f0, x=514, y=25,
> w=570, h=479, pSib=0x0) at miwindow.c:728
> #6 0x080a5134 in ConfigureWindow (pWin=0xa8055f0, mask=15,
> vlist=0xaf2e604, client=0xa7e30f0) at window.c:2492
> #7 0x08077813 in ProcConfigureWindow (client=0xa7e30f0) at dispatch.c:764
> #8 0x081ad760 in XaceCatchDispatchProc (client=0xa7e30f0) at xace.c:281
> #9 0x08076df1 in Dispatch () at dispatch.c:457
> #10 0x0809355c in main (argc=11, argv=0xff9b2304, envp=0xff9b2334) at
> main.c:445
> (gdb) quit
> The program is running. Quit anyway (and detach it)? (y or n) y
> LND: Sending signal 11 to Thread 0xf7f05b00 (LWP 14093)
> Detaching from program: /opt/SUNWut/lib/Xnewt, process 14093
>
> Has anyone else seen this ? Are there any patches available for this ?
>
> Thanks!
>
> -- --
>
> /* Please Dont Blame Me For The Below Text */
>
> IMPORTANT: This email remains the property of the Australian Defence
> Organisation and is subject to the jurisdiction of section 70 of the
> CRIMES ACT 1914. If you have received this email in error, you are
> requested to contact the sender and delete the email.
>
> -- --
> _______________________________________________
> SunRay-Users mailing list
> [email protected]
> http://www.filibeto.org/mailman/listinfo/sunray-users
--
Tom.
--------------------------------------------------------------------------
Tom Crummey, Infrastructure Manager, EMAIL: [email protected]
Dept. of Electronic and Electrical Engineering,
University College London, Roberts Building, TEL: +44 (0)20 7679 3898
Torrington Place, FAX: +44 (0)20 7388 9325
London, UK, WC1E 7JE.
--------------------------------------------------------------------------
_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users
