[Fedora] pcbnew crash when trying to view 3D

[Alain Portal]

This bug was first reported on Fedora for the the kicad-2010-03-14 release:
https://bugzilla.redhat.com/show_bug.cgi?id=592047
but still present on 2010-05-09.

I can reproduce it on my laptop, not on my work PC, twice are x86_64, using wxGTK-2.8.11.

To reproduce, run pcbnew, with or without a board loaded, open "3D visu" menu and just click -> crash

Unfortunately, gdb is useless:

This GDB was configured as "x86_64-redhat-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/bin/pcbnew...(no debugging symbols found)...done.
(gdb) run
Starting program: /usr/bin/pcbnew
warning: "/usr/lib/debug/usr/lib64/kicad/libbitmaps.so.debug": separate debug info file has no debug info
[Thread debugging using libthread_db enabled]
drmRadeonCmdBuffer: -22. Kernel failed to parse or rejected command stream. See dmesg for more info.

Program exited with code 0352.
(gdb) bt
No stack.

dmesg said:
[drm:r100_cs_track_check] *ERROR* [drm] No buffer for z buffer !
[drm:radeon_cs_ioctl] *ERROR* Invalid command stream !

[Alain Portal (alain-portal-univ-montp2)]

This bug is probably video driver relative.

[Jerry Jacobs (jerryjacobs-deactivatedaccount)]

Same problem for me under Fedora 12, this seems to work under Ubuntu 10.04. Probably its a problem with the OpenGL layer. Under OS X seems to work fine either.

Hope this will be fixed under Fedora 13.
Close it for now to let it in the archive.

[Alain Portal (alain-portal-univ-montp2)]

I changed status to "invalid" just to say it wasn't a kicad bug.
Could you tell me which is you video card?
Which video driver (name-version-release) it uses?
Same question for ubuntu.

A workaround for me is to add "radeon.modeset=0" on boot command line.

[Jerry Jacobs (jerryjacobs-deactivatedaccount)]

I use NVIDIA-Linux-x86-195.36.24, it seems last week OpenGL had a
update on Fedora and tested it again and it seems to work fine now.
But I installed the NVIDIA blob myself by hand because there where
problems after updating that the rpmfusion repository did'nt supplied
the
kernel module for the latest Fedora kernel.

So,
Seems to be fixed for me now.

On Mon, May 24, 2010 at 1:13 PM, Alain Portal
<email address hidden> wrote:
> I changed status to "invalid" just to say it wasn't a kicad bug.
> Could you tell me which is you video card?
> Which video driver (name-version-release) it uses?
> Same question for ubuntu.
>
> A workaround for me is to add "radeon.modeset=0" on boot command line.
>
> --
> [Fedora] pcbnew crash when trying to view 3D
> https://bugs.launchpad.net/bugs/582997
> You received this bug notification because you are a member of KiCad Bug
> Squad, which is subscribed to KiCad.
>

--
Kind Regards,
Met vriendelijke groet,
Jerry Jacobs
- http://www.xor-gate.org

--------------
"One of the main causes of the fall of the Roman Empire was that,
lacking zero, they had no way to indicate successful termination of
their C programs" - Robert Firth
Not everything that counts can be counted, and not everything that can
be counted counts. - Albert Einstein

[Stephen Eaton (seaton)]

marking as "won't fix" as appears to be driver related

[Fabio Varesano (fabio-varesano)]

Having the same problem. Usually when a bug is caused by another FLOSS software is a good thing to report it to the broken software.

So, I reported it to Xorg radeon devs at https://bugs.freedesktop.org/show_bug.cgi?id=29833

[Peter Clifton (pcjc2)]

Won't fix is not the right attitude here, and the bug is either in Wx, or KiCad's use / initialisation of GL. Working on one machine / driver does not mean you have the code right with GL. (I know that from experience).

[Alain Portal (alain-portal-univ-montp2)]

Can somebody change the status of this bug from "Won't fix" to "Fix commited" in rev 2767?

[Clyde R. Shappee (clydes)]

3D viewer crashes PCBNEW on Ubuntu 10.10, current release candidate and several revisions back.

This happens on an IBM Thinkpad with ATi Mobility Radio 7500 Graphics and on a desktop with an Intel D865GVHGZ motherboard with Intel 8565GV Extreme graphics 2 controller.

Wings 3D modeling software and MeshLab run fine on both machines.

[Fabio Varesano (fabio-varesano)]

Clyde, the status has been set as "Fix Committed". This means that a fix has been committed to the source tree.
That happened ~ one week ago, so it's very unlikely that the fix has been incorporated into the release candidate yet. If it would the status should be "Fix Released".

That said, you might wanna get KiCAD sources from bzr and compile your own.

I experienced this bug from the first time with KiCAD and, for me, it now works great!

[Clyde R. Shappee (clydes)]

Fair enough. As I am not a developer I did fully understand the meaning of what was said in the thread. I added my $0.02 to be sure it was known that the bug was not to specific to one video card or Linux distribution. I will read more carefully next time.

Also, I do not have the skills yet to compile my own, so I rely upon others (thankfully).

Right now most of my use is done on a Windows machine, but I am learning and exploring Ubuntu. So, in that OS, I can afford to wait.

Clyde

[Marius B. Kotsbak (mariusko)]

I also see this in Ubuntu.

[Marius B. Kotsbak (mariusko)]

This is the backtrace on Ubuntu:

Program received signal SIGSEGV, Segmentation fault.
0x01f30d23 in ?? () from /usr/lib/dri/libdricore.so
(gdb) backtrace
#0 0x01f30d23 in ?? () from /usr/lib/dri/libdricore.so
#1 0x01f53731 in ?? () from /usr/lib/dri/libdricore.so
#2 0x08254b57 in SetGLColor(int) ()
#3 0x08256aba in Pcb3D_GLCanvas::CreateDrawGL_List() ()
#4 0x08257250 in Pcb3D_GLCanvas::Redraw(bool) ()
#5 0x082532a4 in Pcb3D_GLCanvas::OnPaint(wxPaintEvent&) ()
#6 0x0085396f in wxAppConsole::HandleEvent(wxEvtHandler*, void (wxEvtHandler::*)(wxEvent&), wxEvent&) const () from /usr/lib/libwx_baseu-2.8.so.0
#7 0x008e4ac9 in wxEvtHandler::ProcessEventIfMatches(wxEventTableEntryBase const&, wxEvtHandler*, wxEvent&) () from /usr/lib/libwx_baseu-2.8.so.0
#8 0x008e4c44 in wxEventHashTable::HandleEvent(wxEvent&, wxEvtHandler*) ()
   from /usr/lib/libwx_baseu-2.8.so.0
#9 0x008e4fe7 in wxEvtHandler::ProcessEvent(wxEvent&) ()
   from /usr/lib/libwx_baseu-2.8.so.0
#10 0x00304863 in ?? () from /usr/lib/libwx_gtk2u_gl-2.8.so.0
#11 0x0134648c in g_cclosure_marshal_VOID__VOID ()
   from /usr/lib/i386-linux-gnu/libgobject-2.0.so.0
#12 0x0132a372 in g_closure_invoke ()
   from /usr/lib/i386-linux-gnu/libgobject-2.0.so.0
#13 0x0133d048 in ?? () from /usr/lib/i386-linux-gnu/libgobject-2.0.so.0
#14 0x01345b29 in g_signal_emit_valist ()
   from /usr/lib/i386-linux-gnu/libgobject-2.0.so.0
---Type <return> to continue, or q <return> to quit---

Is it the same issue that is solved?

[Fabio Varesano (fabio-varesano)]

Marius, which version of KiCAD are you running? Please upgrade to the most recent KiCAD version.

[Marius B. Kotsbak (mariusko)]

I still see the above stacktrace in kicad-2011-07-08-BZR3044-stable-UBUNTU_10.10_full_with_components_doc.tgz package.

[Fabio Varesano (fabio-varesano)]

Ok, are you able to compile KiCAD from the most recent sources from bzr?