lp:~afrantzis/mir/fix-1679591-display-config-observer-race

Branch merges

Propose for merging

No branches dependent on this one.

On hold for merging into lp:mir

Alexandros Frantzis (community): Needs Information on 2017-04-06

Mir CI Bot: Needs Fixing (continuous-integration) on 2017-04-06

Cemil Azizoglu (community): Approve on 2017-04-05

Alberto Aguirre (community): Approve on 2017-04-05

Chris Halse Rogers: Approve on 2017-04-05

Diff: 95 lines (+57/-3)

2 files modified

src/server/scene/mediating_display_changer.cpp (+6/-1)
tests/unit-tests/scene/test_mediating_display_changer.cpp (+51/-2)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Bug #1679591: Laptop suspends when closing lid with an external display plugged

High

Triaged

Link a bug report

Related blueprints

4141. By Alexandros Frantzis on 2017-04-05

tests: Fix gmock expectation complaints and leak

4140. By Alexandros Frantzis on 2017-04-04

server: Emit the DisplayConfigurationObserver initial_configuration notification through the server action queue

4139. By Daniel van Vugt on 2017-04-04

Generalise android's DisplayBuffer transformation passing.
. Fixes: https://bugs.launchpad.net/bugs/1639226.

Approved by Kevin DuBois, mir-ci-bot.

4138. By Daniel van Vugt on 2017-04-04

Simplify android::DisplayBuffer::view_area() so that it's also now more
correctly reflecting the logical view area (extents()) of its output. Fixes: https://bugs.launchpad.net/bugs/1639226.

Approved by mir-ci-bot, Alan Griffiths.

4137. By Chris Halse Rogers on 2017-04-04

Initial hybrid-output support for mesa-kms platform.

This is (mostly) what is known in the X world as reverse-prime - rendering on the integrated card, displaying on the external GPU.

This selects whatever card is first to initialise in DRM (which is [usually?] the boot VGA device) to do all shell rendering on, but exposes all outputs across all GPUs for display.

As a first cut, it misses out on hardware cursor support for hybrid systems, and performance on hybrid systems will be non-optimal, but this should not significantly change the code flow for non-hybrid systems.

Approved by Kevin DuBois, mir-ci-bot, Alan Griffiths, Cemil Azizoglu.

4136. By Daniel van Vugt on 2017-04-03

Start to implement eglstream-kms DisplayBuffer logical layout:
 * Use the logical size (extents) of the output, which takes into
   consideration transformations (presently orientation, later others).
 * Return the correct transformation from DisplayBuffer::transformation()
   so that the renderer knows what to do. Fixes: https://bugs.launchpad.net/bugs/1639226.

Approved by Chris Halse Rogers, Kevin DuBois, mir-ci-bot.

4135. By Alan Griffiths on 2017-04-03

Use struct consistently for InputDeviceInfo in public headers.

Approved by Brandon Schaefer, mir-ci-bot.

4134. By Daniel van Vugt on 2017-04-03

eglapp: Go fullscreen properly, using window states.
.

Approved by Alan Griffiths, mir-ci-bot.

4133. By Andreas Pokorny on 2017-03-31

Catch configuration exceptions during mi::DefaultDevice construction

On device removal, DefaultInputDeviceHub stores and later restores the configuration. That config might be invalid when the device appears again. I.e. the keymap requested might have been removed from the system... To ensure that the device is still usable those exceptions are caught.

Approved by mir-ci-bot, Alexandros Frantzis.

4132. By Daniel van Vugt on 2017-03-31

Generalise transformations supported by X::DisplayBuffer,
which is a prerequisite to completing LP: #1639226
. Fixes: https://bugs.launchpad.net/bugs/1639226.

Approved by Alexandros Frantzis, mir-ci-bot.