Mir server crashes with "std::exception::what: Failed to get front buffer object" when trying to fullscreen a surface

[Daniel van Vugt]

Mir server crashes with "std::exception::what: Failed to get front buffer object" when trying to fullscreen/bypass a surface in multimonitor clone mode.

ERROR: /home/dan/bzr/mir/fullscreen/src/platform/fatal/fatal.cpp(55): Throw in function void mir::fatal_error_except(const char*, ...)
Dynamic exception type: N5boost16exception_detail10clone_implINS0_19error_info_injectorISt13runtime_errorEEEE
std::exception::what: Failed to get front buffer object

Mesa only has a fixed limit of (four?) front framebuffer objects available. So somehow the combination of fullscreen bypass and multimonitor is making us ask for more than that... (?)

Sorry I don't have a test case to reproduce it easily yet without using my own custom branch.

[Daniel van Vugt (vanvugt)]

I can't reproduce this now. But I'm sure someone will eventually.

[Daniel van Vugt (vanvugt)]

Test case (I have no idea why this works):

 0. Use branch: lp:~vanvugt/mir/managed-surface to get access to the required key combos.
 1. Start the demo shell
 2. mir_demo_client_egltriangle
 3. Resize the client rapidly using Alt+middlebutton
 4. Try to fullscreen it: Alt+F11

[Daniel van Vugt (vanvugt)]

That test case works because the resizing tips the dimensions over the 800x600 threshold and Mir then allocates scanout buffers. This leads it to run out of scanout buffer resources and bypass can't work.

[PS Jenkins bot (ps-jenkins)]

Fix committed into lp:mir at revision None, scheduled for release in mir, milestone 0.10.0

[Daniel van Vugt (vanvugt)]

mir (0.10.0+15.04.20150107.2-0ubuntu1) vivid; urgency=medium