pcbnew opengl crash on "big" file, legacy & cairo works, debug output

[User12020310012]

Hello,

pcbnew stops working in accelerated mode (OpenGl) when board has to many features to render
works in legacy and cairo mode using cpu

Log file, interesting part:
i965: Failed to submit batchbuffer: Input/output error
pcbnew: /usr/include/boost/interprocess/sync/posix/mutex.hpp:91: boost::interprocess::ipcdetail::posix_mutex::~posix_mutex(): Assertion `res == 0' failed.

Application: kicad
Version: (6.0.0-rc1-dev-313-g8db3618), debug build
Libraries:
    wxWidgets 3.0.2
    libcurl/7.47.0 GnuTLS/3.4.10 zlib/1.2.8 libidn/1.32 librtmp/2.3
Platform: Linux 4.15.0-30-generic x86_64, 64 bit, Little endian, wxGTK
Build Info:
    wxWidgets: 3.0.2 (wchar_t,wx containers,compatible with 2.8) GTK+ 2.24
    Boost: 1.58.0
    Curl: 7.47.0
    Compiler: GCC 5.4.0 with C++ ABI 1009

Build settings:
    USE_WX_GRAPHICS_CONTEXT=OFF
    USE_WX_OVERLAY=OFF
    KICAD_SCRIPTING=ON
    KICAD_SCRIPTING_MODULES=ON
    KICAD_SCRIPTING_WXPYTHON=ON
    KICAD_SCRIPTING_ACTION_MENU=OFF
    BUILD_GITHUB_PLUGIN=ON
    KICAD_USE_OCE=OFF
    KICAD_USE_OCC=OFF
    KICAD_SPICE=ON

[Tomasz Wlostowski (twlostow)]

How much RAM/what graphics card do you have?

Tom

[User12020310012 (user12020310012-deactivatedaccount)]

RAM 16GB,
Integrated Intel HD 4000
https://ark.intel.com/products/64896/Intel-Core-i5-3320M-Processor-3M-Cache-up-to-3_30-GHz

same thing is happening on Windows 10 64bit, latest drivers (have dual boot)

[User12020310012 (user12020310012-deactivatedaccount)]

glxinfo output if it is of any help

[User12020310012 (user12020310012-deactivatedaccount)]

dmesg

[21695.310058] i915 0000:00:02.0: Resetting chip after gpu hang
[21703.310044] i915 0000:00:02.0: Resetting chip after gpu hang
[21711.310011] i915 0000:00:02.0: Resetting chip after gpu hang
[21714.413924] asynchronous wait on fence i915:[global]:1532ad timed out
[21719.309988] i915 0000:00:02.0: Resetting chip after gpu hang
[21727.341952] i915 0000:00:02.0: Resetting chip after gpu hang
[25074.324916] i915 0000:00:02.0: Resetting chip after gpu hang
[25082.325330] i915 0000:00:02.0: Resetting chip after gpu hang
[25090.325340] i915 0000:00:02.0: Resetting chip after gpu hang
[25098.325328] i915 0000:00:02.0: Resetting chip after gpu hang
[25106.292908] i915 0000:00:02.0: Resetting chip after gpu hang

[Tomasz Wlostowski (twlostow)]

Hi, it crashes on my PC (also intel graphics). In the case of this particular board, the workaround is to disable showing of pad clearances (they're drawn in very GPU-memory-consuming way).

Working on a proper fix.
Tom

[jean-pierre charras (jp-charras)]

This board looks strange to me: there are 28000 pads to create a very basic copper zone.
And these 27000 of 28000 pads are illegal SMD pads on all layers, that cannot be created by pcbnew.

28000 pads have certainly a *very high cost* in matter of memory and calculation time.
So I am not sure there is a bug.

[User12020310012 (user12020310012-deactivatedaccount)]

New design uses Kicad v5 pad instead of multiple per piece.
https://github.com/electroniceel/protoboard/pull/3

Will convert this .kicad_pcb file and see if crash persists.

[User12020310012 (user12020310012-deactivatedaccount)]

@jean oh I see what you mean. both F.Cu and B.Cu layers enabled on SMD footprint. Will suggest going to 2 separate layers

[Tomasz Wlostowski (twlostow)]

This looks like an issue with Intel graphics drivers (crash in glCopyBufferSubData()). Could someone check if this board causes a crash on Nvidia/Ati drivers?

Tom

[Seth Hillbrand (sethh)]

NVidia here. Does not crash.

[Wayne Stambaugh (stambaughw)]

I can confirm the nvidia drivers on linux does not crash.

[Tomasz Wlostowski (twlostow)]

Thanks Wayne, the bug looks specific to Intel drivers. I'll make a workaround disabling hardware framebuffer copying on intel devices.

[KiCad Janitor (kicad-janitor)]

Fixed in revision aab40ee9e8dbe3ba8d0cb69ff37f298c29a9d7a3
https://git.launchpad.net/kicad/patch/?id=aab40ee9e8dbe3ba8d0cb69ff37f298c29a9d7a3