Mir server crashes with "std::exception::what: Failed to get front buffer object" when trying to fullscreen a surface
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| | Mir |
Fix Released
|
High
|
Daniel van Vugt | |
| | mir (Ubuntu) |
High
|
Unassigned | ||
Bug Description
Mir server crashes with "std::exception
ERROR: /home/dan/
Dynamic exception type: N5boost16except
std::exception:
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.
Related branches
- Kevin DuBois (community): Approve on 2014-12-17
- Cemil Azizoglu (community): Approve on 2014-12-17
- Alan Griffiths: Approve on 2014-12-17
- PS Jenkins bot (community): Approve (continuous-integration) on 2014-12-17
-
Diff: 106 lines (+38/-7)3 files modifiedsrc/platforms/mesa/display_buffer.cpp (+15/-6)
src/platforms/mesa/display_buffer.h (+1/-1)
tests/unit-tests/graphics/mesa/test_display_buffer.cpp (+22/-0)
- Cemil Azizoglu (community): Approve on 2014-12-18
- Alan Griffiths: Approve on 2014-12-18
- PS Jenkins bot (community): Approve (continuous-integration) on 2014-12-18
-
Diff: 127 lines (+49/-12)2 files modifiedsrc/platforms/mesa/display_buffer.cpp (+3/-8)
tests/unit-tests/graphics/mesa/test_display_buffer.cpp (+46/-4)
| description: | updated |
| tags: | added: fullscreen |
| summary: |
Mir server crashes with "std::exception::what: Failed to get front - buffer object" when trying to fullscreen/bypass a surface in clone mode + buffer object" when trying to fullscreen a surface |
| Daniel van Vugt (vanvugt) wrote : | #2 |
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_
3. Resize the client rapidly using Alt+middlebutton
4. Try to fullscreen it: Alt+F11
| Changed in mir: | |
| status: | Incomplete → Triaged |
| Daniel van Vugt (vanvugt) wrote : | #3 |
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.
| Changed in mir: | |
| assignee: | nobody → Daniel van Vugt (vanvugt) |
| milestone: | none → 0.10.0 |
| status: | Triaged → In Progress |
| PS Jenkins bot (ps-jenkins) wrote : | #4 |
Fix committed into lp:mir at revision None, scheduled for release in mir, milestone 0.10.0
| Changed in mir: | |
| status: | In Progress → Fix Committed |
| Changed in mir: | |
| status: | Fix Committed → Fix Released |
| Daniel van Vugt (vanvugt) wrote : | #6 |
mir (0.10.0+
| Changed in mir (Ubuntu): | |
| importance: | Undecided → High |
| status: | New → Fix Released |
I can't reproduce this now. But I'm sure someone will eventually.