Qt render gets blocked on EGLSwapBuffers [fka Upon upgrading to Qt5.2 the music app no longer plays the next song if the screen is off]

Marking this bug as medium since it hasn't been proven to also be seen in the trusty-proposed images.

I very much doubt this is a unity-mir bug, it doesn't know about when the song ends.

What's more, qmlscene uses CPU when the app stops playing, so the app is not stopped by the app lifecycle. This must be some app/Qt/QtMultimedia issue, or maybe android media services.

Additional data points: it's enough for the screen to turn on, no need to focus the app for it to resume playback.

Keeping the device active with `powerd-cli active` doesn't help - it does, however, help to keep the display on, even if locked (and hence the app is unfocused).

I wonder if it's some interaction with the compositor, maybe... like the app is prevented from rendering and Qt/QML gets confused and the playlist doesn't advance.

Suggested promotion blocker on mailing list. CCed a few that might be able to find the right direction to look here.

QA probably should do the call whether this is a blocker or not. If it is, we should ensure that we run a similar test "music lifecycle, screen off/on tests" for all kind of music/media related component landings and of course during image testing. Can we work this into the appropriate image and landing test plans?

I've related a branch that adds a test for this in the future for the music app.

I agree that this seems like a QtMultimedia issue. Jim Hodapp, could I get verification that the media hub's modified QtMultimedia MediaPlayer component isn't being included in the recent phone images.

@Victor Thompson: there is no media-hub code landed in the archive at all yet, so no this isn't media-hub affecting things. And there's the possibility that media-hub could in fact fix this issue. It sounds to me like the music-app is no longer in the foreground when the screen is off and so the music-app can't advance the playlist to the next song. If this is the case, this is why I think media-hub will fix this scenario with the background playlist support added.

Here goes: lp:~albaguirre/mir/test-hack-consume-buffers-display-off hacks in a way that lets apps render even though screen is off. Attached are armhf binaries with a script to push it to device (thanks AlbertA!).

tail -f /var/log/lightdm/unity-system-compositor.log will show you that composition happens every second (when app updates the time counter) even when screen is off.

This makes the app continue to next song on the playlist, as well as handle volume keys when screen is off.

This means that indeed Qt changed something that makes the events not go around when it can't render frames. Next step would be to find if that is the designed behaviour and why not, and how to fix it.

Status changed to 'Confirmed' because the bug affects multiple users.

Confirmed the suspicions and filed a bug upstream: https://bugreports.qt-project.org/browse/QTBUG-37677

I could emulate the rendering block by SIGSTOP'ing Xvfb to which qmlscene was rendering, see bug for details.

@rosamaria, only assign a bug to yourself if you're actively working on a fix, otherwise stop changing assignees. Thanks.

Once again. @rosa maria. Please don't assign yourself to the tasks, unless you're really working on them.

The commit responsible for this was the threaded render loop rewrite:

https://qt.gitorious.org/qt/qtdeclarative/commit/ebe8b9408cfcd953fae80514aa67e49221541bed

I'm marking this bug as a mir & unity-mir bug, since that's where we are going to solve the "problem"
short story, we're going to make mir not block on eglswapbuffers & rely on exposeEvent() in qt (via shell trigger)
please read this for detail
https://lists.launchpad.net/ubuntu-phone/msg07458.html

mir branch under test as of today
  https://code.launchpad.net/~afrantzis/mir/non-blocking-swap-buffers-spike
note, expose event change branch still needed, to be updated

FWIW, this still happens in image 276. Not surprising though, since no fix is committed.

just an update, associated branches up for MP. but at the moment experiencing the dreaded "fails ci testing, but when rerun ci testing locally no failures".

Fix committed into lp:mir/devel at revision 1545, scheduled for release in mir, milestone Unknown

Branches up for testing are built in silo3 for testing
https://launchpad.net/~ci-train-ppa-service/+archive/landing-003

Status changed to 'Confirmed' because the bug affects multiple users.

Reproduced on utopic image 2.

This bug was fixed in the package mir - 0.1.9+14.10.20140430.1-0ubuntu1

---------------
mir (0.1.9+14.10.20140430.1-0ubuntu1) utopic; urgency=medium

  [ Daniel van Vugt ]
  * New upstream release 0.1.9 (https://launchpad.net/mir/+milestone/0.1.9)
    - mirclient ABI unchanged, still at 7. Clients do not need rebuilding.
    - mirserver ABI bumped to 19. Shells need rebuilding.
    - More libmirserver class changes and reorganization, including;
      . Moving things from shell:: to scene::
      . Rewriting/refactoring surface factories.
    - Added an id() to Renderable.
    - Scene/Renderer interfaces:
      . Scene is no longer responsible for its own iteration (no for_each
        any more). Instead you should iterate over the list returned by
        Scene::generate_renderable_list().
    - Bugs fixed:
      . Stale socket issue. (LP: #1285215)
      . Qt render gets blocked on EGLSwapBuffers. (LP: #1292306)
      . Lock order violated found in helgrind (potential deadlock).
        (LP: #1296544)
      . [regression] SwitchingBundle in framedropping mode can hang.
        (LP: #1306464)
      . [DPMS] Display backlight turns back on almost immediately after
        being turned off. (LP: #1231857)
      . Wrong frame is seen on wake up/resume/unlock. (LP: #1233564)
      . Nested platform is not testable (LP: #1299101)
      . [regression] mir_demo_server_shell crashes on display resume.
        (LP: #1308941)
      . Multi-threaded composition is actually mostly serialized by
        SurfaceStack::guard. (LP: #1234018)
      . Mirscreencast slows down compositing and makes it very jerky.
        (LP: #1280938)
      . Mirscreencast can cause clients to render faster than the screen
        refresh rate. (LP: #1294361)
      . Screen turns on when a new session/surface appears. (LP: #1297876)
      . mir-doc package is >56MB in size, expands to >100MB of files.
        (LP: #1304998)
      . [regression] Clang: 'mir::test::doubles::MockSurface::visible'
        hides overloaded virtual function [-Woverloaded-virtual].
        (LP: #1301135)
      . [regression] GLRenderer* unit tests have recently become noisy.
        (LP: #1308905)
      . FocusController::set_focus_to() no longer seems to raise a session
        to the top. (LP: #1302689)

  [ Ubuntu daily release ]
  * New rebuild forced
 -- Ubuntu daily release <email address hidden> Wed, 30 Apr 2014 13:26:58 +0000

Confirmed this as fixed in Utopic image r7.