Bug 9204

Summary:	Segfault after glXMakeCurrent
Product:	Mesa	Reporter:	Miha Tomšič <miha.tomsic>
Component:	GLX	Assignee:	mesa-dev
Status:	RESOLVED INVALID	QA Contact:
Severity:	normal
Priority:	high	CC:	idr
Version:	6.5	Keywords:	NEEDINFO
Hardware:	x86 (IA32)
OS:	Linux (All)
Whiteboard:
i915 platform:		i915 features:

Description Miha Tomšič 2006-11-30 01:03:19 UTC

When using pd with Gem OpenGL extension I get segfault when destroying or
closing the GL window. I am using Fedora Core 6, with mesa 6.5.1 and Radeon
Mobility 9000. There is no problem closing or destroying the GL window using
mesa 6.4.2 version of libGL.

Here's the gdb backtrace:
(gdb) run
Starting program: /usr/local/bin/pd -lib Gem
[Thread debugging using libthread_db enabled]
[New Thread -1208682816 (LWP 9600)]
Vertex3f: 2
Normal3f: 2

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1208682816 (LWP 9600)]
0x017f2880 in ?? ()
(gdb) bt
#0  0x017f2880 in ?? ()
#1  0x0376dd3c in BindContextWrapper (dpy=0x8cc5a30, gc=0x8d23118,
draw=<value optimized out>, read=39845892) at glxext.c:1542
#2  0x0376ff1f in MakeContextCurrent (dpy=0x8cc5a30, draw=39845892,
read=39845892, gc=0x8d23118) at glxext.c:1630
#3  0x037701c3 in glXMakeCurrent (dpy=0x8cc5a30, draw=39845892,
gc=0x8d23118) at glxext.c:1792
#4  0x00779046 in GemMan::destroyWindow () at GemMan.cpp:1303
#5  0x007797a5 in GemMan::render () at GemMan.cpp:749
#6  0x080b194b in sched_tick (next_sys_time=1511854080) at m_sched.c:355
#7  0x080b266e in m_scheduler () at m_sched.c:460
#8  0x080b580a in sys_main (argc=3, argv=0xbffaef44) at s_main.c:316
#9  0x080bea0f in main (argc=Cannot access memory at address 0x0
) at s_entry.c:27
(gdb)

Comment 1 Roland Scheidegger 2006-12-01 02:43:41 UTC

Some cleanup happened with MakeCurrentContext which might be related
(https://bugs.freedesktop.org/show_bug.cgi?id=8443). Could you try Mesa 6.5.2 (rc)?

Comment 2 Miha Tomšič 2006-12-05 00:12:24 UTC

Installed mesa 6.5.2 from CVS and still I get the segfault. And I don't get any
renderig in Gem GL window. 

Checked with other openGL apps, and some are OK, but others have problems
displaying text.

The backtrace at segfault is almost the same. 

Starting program: /usr/local/bin/pd -lib Gem
[Thread debugging using libthread_db enabled]
[New Thread -1209014592 (LWP 29828)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1209014592 (LWP 29828)]
0x059e7570 in ?? ()
(gdb) bt
#0  0x059e7570 in ?? ()
#1  0x00610c60 in MakeContextCurrent (dpy=0x8959518, draw=44040196,
read=44040196, gc=0x8982400) at glxext.c:1554
#2  0x00610f53 in glXMakeCurrent (dpy=0x8959518, draw=44040196, gc=0x8982400) at
glxext.c:1730
#3  0x00321106 in GemMan::destroyWindow () at GemMan.cpp:1303
#4  0x00321865 in GemMan::render () at GemMan.cpp:749
#5  0x080b194b in sched_tick (next_sys_time=133877760) at m_sched.c:355
#6  0x080b266e in m_scheduler () at m_sched.c:460
#7  0x080b580a in sys_main (argc=3, argv=0xbfe36de4) at s_main.c:316
#8  0x080bea0f in main (argc=143752376, argv=0x89594b8) at s_entry.c:27
(gdb)

Comment 3 Ian Romanick 2009-09-17 16:34:08 UTC

Are you able to reproduce this with a more recent version of Mesa?  Especially 7.5.1?  A lot has changed in these code paths over the last 2+ years.

Comment 4 Kristian Høgsberg 2010-05-07 10:33:07 UTC

ping timeout... closing bug.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.