Rotating display to 90 and 270 degrees crashes shell

I'd like to get this resolved, but we'd need more information yet.

You have shown a client (System Settings) crashed at 16:51:04, but that USC was still running fine 12 minutes later at 17:03.

Which log is this line from? System Settings?
2016-11-02 16:51:04,111 - WARNING - QEGLPlatformContext: eglSwapBuffers failed: 3000

If so then you will find the client crashed because Unity8 went away. So we need to see where/why Unity8 has crashed(?)

Also what command (or code) did you use to rotate the screen?

You can likely use silo 2140 (soon) to replicate this and get to the logs you need.

unity8 session on xenial+overlay is broken for me, so I can test this. Suggest someone who's able to start a unity8 session install silo 2140 and grab a unity8 log if they are able to reproduce.

Code is in silo 2140, but I basically use mir_connection_set_base_display_config.

> Also what command (or code) did you use to rotate the screen?

used mir_connection_set_base_display_config

Sounds like we should have a test tool for this kind of thing... or modify 'mirout' to be able to make changes to the base display config.

Also, I recently encountered the same Mir exception. It occurred in a client when my server crashed.

Hey, sorry about the delay... I'm looking at the API now (mir_connection_set_base_display_config) and it appears that one is deprecated so you should not use it.

Unfortunately mir-team has done a poor job of documenting that it's a deprecated API. The only hint that it's being deprecated is in the docs for mir_connection_create_display_config() which you are probably using. And it's even worse that Mir's out mirout tool (out.c) still uses the old API which will eventually go away.

So you should probably only use functions which take the MirDisplayConfig type, and not the MirDisplayConfiguration type. Information about the newer API using MirDisplayConfig is in:
  /usr/include/mirclient/mir_toolkit/mir_display_configuration.h
  /usr/include/mirclient/mir_toolkit/mir_connection.h

More confusing again is that the "new" function names and file name actually matches the "old" type name better than the new API it represents.

Sorry about the bad news... I'm only looking at this new API in detail myself for the first time.

Using the new API I've built a test tool which among other things replicates what you're doing:
https://code.launchpad.net/~vanvugt/mir/mirout-basic-commands/+merge/311372

I've found that enable/disable can crash the server (bug 1643446), but rotation works fine (albeit distorted - bug 1643488).

So it sounds like just Unity8 is crashing. We'll need a stack trace from it.

@jgdx
When you have a chance please re-test this using the new api specified above

Tested this with latest mir 0.26 and silo 2140 https://bileto.ubuntu.com/#/ticket/2140 for ubuntu-system-settings. when I try rotating, I get the following crash from unity8

Thread 27 "QSGRenderThread" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fec84b6e700 (LWP 8251)]
0x00007fec94716d07 in Screen::makeCurrent() ()
   from /usr/lib/x86_64-linux-gnu/qt5/plugins/platforms/libqpa-mirserver.so
(gdb) bt
#0 0x00007fec94716d07 in Screen::makeCurrent() ()
   from /usr/lib/x86_64-linux-gnu/qt5/plugins/platforms/libqpa-mirserver.so
#1 0x00007fec9471d516 in MirOpenGLContext::makeCurrent(QPlatformSurface*) ()
   from /usr/lib/x86_64-linux-gnu/qt5/plugins/platforms/libqpa-mirserver.so
#2 0x00007fec9d51b81c in QOpenGLContext::makeCurrent(QSurface*) ()
   from /usr/lib/x86_64-linux-gnu/libQt5Gui.so.5
#3 0x00007fec9e6ba69b in ?? () from /usr/lib/x86_64-linux-gnu/libQt5Quick.so.5
#4 0x00007fec9cb447e8 in ?? () from /usr/lib/x86_64-linux-gnu/libQt5Core.so.5
#5 0x00007fec9bc176ba in start_thread (arg=0x7fec84b6e700) at pthread_create.c:333
#6 0x00007fec9c23c82d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:109
(gdb)

@vanvugt
please see above trace and confirm problem not in mir.. thanks

Thanks Bill. I can't see any Mir code in that stack trace. And Gerry is indeed the best person to look at it since he recently fixed another (related?) GL context crash that looked similar.

It's worth noting that Unity8 has historically (on phablet) done its own screen rotation independently of Mir's. I'm actually not sure if Unity8 supports Mir's screen rotation settings (via display config).

Oh! Good news: Gerry's fix hasn't landed yet so this bug may get fixed by that very soon:

lp:~gerboland/qtmir/fix-glcontext-loss-on-reconfigure

(bug 1521403)

This bug was fixed in the package qtmir - 0.5.1+17.04.20170215.1-0ubuntu1

---------------
qtmir (0.5.1+17.04.20170215.1-0ubuntu1) zesty; urgency=medium

  [ Michał Sawicz ]
  * We're at provides 26 already (LP: #1662608)

  [ Alan Griffiths ]
  * Identify the code that depends directly on mirserver-dev headers

  [ Albert Astals Cid ]
  * Check we provide the same unity-application-impl that we require

  [ Daniel d'Andrada ]
  * Resolve mir cursor names using mir symbols instead of plain strings
    (LP: #1662827)

  [ Gerry Boland ]
  * ScreenModel: Only expose windows on displays that are turned on (LP:
    #1521403, #1638611, #1656250)
  * Restore lost LTTng tracepoints, and delete unused ones (LP:
    #1658084)

  [ Nick Dedekind ]
  * Added Extended Display Information Data (EDID) parsing.

  [ Alan Griffiths, Nick Dedekind ]
  * Iteration 0 of miral::PersistDisplayConfig. This does nothing yet
    (and breaks nothing in the process). This MP creates a place (miral-
    prototypes) to build prototype miral features and sketches out what
    will need to be implemented for PersistDisplayConfig. (LP: #1644189)

 -- Michał Sawicz <email address hidden> Wed, 15 Feb 2017 13:23:17 +0000