[mythtv-users] MythTV Raspberry Pi2 frontend testers

Fri Jan 8 09:13:00 UTC 2016

On Fri, 2016-01-08 at 05:05 +0000, Thomas Mashos wrote:
> On Mon, Dec 7, 2015 at 7:52 AM Thomas Mashos <thomas at mashos.com> wrote:
> 
> > On Sat, Dec 5, 2015 at 1:06 PM Thomas Mashos <thomas at mashos.com> wrote:
> >
> >> On Sat, Dec 5, 2015 at 1:31 AM Lawrence Rust <lvr at softsystem.co.uk>
> >> wrote:
> >>
> >>> On Sat, 2015-12-05 at 00:17 +0000, Thomas Mashos wrote:
> >>> > On Fri, Dec 4, 2015 at 9:36 AM Lawrence Rust <lvr at softsystem.co.uk>
> >>> wrote:
> >>> >
> >>> > > On Fri, 2015-12-04 at 15:18 +0000, Thomas Mashos wrote:
> >>> > > > On Thu, Dec 3, 2015 at 6:50 AM Peter Bennett (cats22) <
> >>> > > cats22 at comcast.net>
> >>> > > > wrote:
> >>> > > >
> >>> > > > > On 12/01/2015 08:58 AM, Lawrence Rust wrote:
> >>> > > > > > I believe Kodi borrowed from omxplayer which uses a single
> >>> OpenMAX
> >>> > > > > > pipeline from decoder to video renderer which shares buffers
> >>> and
> >>> > > avoids
> >>> > > > > > copying.  Myth currently requires the separation of the
> >>> decoder from
> >>> > > the
> >>> > > > > > renderer and so causes two video buffer copies - one each for
> >>> the GPU
> >>> > > > > > and CPU.  For a 1920x1080 frame this is a significant burden
> >>> that
> >>> > > causes
> >>> > > > > > memory contention.
> >>> > > > > >
> >>> > > > > > I will investigate ways of integrating the decoder and
> >>> renderer so
> >>> > > that
> >>> > > > > > they use a single OpenMAX pipeline.  Unfortunately this is not
> >>> > > trivial
> >>> > > > > > because of the current assumption that ffmpeg is the only
> >>> provider of
> >>> > > > > > hardware accelerated decoding i.e. VDPAU, VAAPI, DXVA2 etc.
> >>> > > > > >
> >>> > > > > >
> >>> > > > > Lawrence
> >>> > > > >
> >>> > > > > One thing to bear in mind, in case you are not aware. Kodi has
> >>> some
> >>> > > > > significant audio sync problems when playing the MPG files that
> >>> mythtv
> >>> > > > > produces in my setup. Especially after skipping forward or back,
> >>> the
> >>> > > > > audio sync is sometimes far off. In this regard Mythtv is
> >>> perfect. So
> >>> > > > > maybe that is part of the reason for the mythtv design. I would
> >>> not
> >>> > > want
> >>> > > > > to do something that lands up with the audio sync problems of
> >>> Kodi.
> >>> > > > >
> >>> > > > > Note that I tested the Kodi audio sync problem with the mythtv
> >>> plugin
> >>> > > > > and also playing the mythtv MPG files as videos in Kodi. Both
> >>> ways
> >>> > > > > exhibit the problem and the author of the kodi mythtv plugin
> >>> confirms
> >>> > > > > that this is a bug with kodi itself, not the plugin.
> >>> > > > >
> >>> > > > > Peter
> >>> > > > >
> >>> > > > > _______________________________________________
> >>> > > > > mythtv-users mailing list
> >>> > > > > mythtv-users at mythtv.org
> >>> > > > > http://lists.mythtv.org/mailman/listinfo/mythtv-users
> >>> > > > > http://wiki.mythtv.org/Mailing_List_etiquette
> >>> > > > > MythTV Forums: https://forum.mythtv.org
> >>> > > >
> >>> > > >
> >>> > > > Lawrence,
> >>> > > >
> >>> > > > I've decided to build this only for 0.28, since we already have our
> >>> > > > packaging setup for Qt 5. However we're back to failing builds
> >>> again. Any
> >>> > > > idea whats causing this?
> >>> > > >
> >>> > > >
> >>> > >
> >>> https://launchpadlibrarian.net/228619903/buildlog_ubuntu-wily-armhf.mythtv_2%3A0.28.0~master.20151203.989baa7-0ubuntu1_BUILDING.txt.gz
> >>> > >
> >>> > > The linker is missing symbols from libEGL so either add it to your
> >>> > > Mythtv configure command line:
> >>> > >
> >>> > > --extra-libs=-lEGL
> >>> > >
> >>> > > or to libmythtv.pro, after -lGLESv2.
> >>> > >
> >>> > > -- Lawrence Rust
> >>> > >
> >>> > > _______________________________________________
> >>> > > mythtv-users mailing list
> >>> > > mythtv-users at mythtv.org
> >>> > > http://lists.mythtv.org/mailman/listinfo/mythtv-users
> >>> > > http://wiki.mythtv.org/Mailing_List_etiquette
> >>> > > MythTV Forums: https://forum.mythtv.org
> >>> >
> >>> >
> >>> >  Thanks Lawrence. I tried doing this and starting a build but I'm
> >>> getting
> >>> > patch issues now. When is the last time you rebased against upstream? I
> >>> > think some commit in the last day is causing trouble
> >>> > https://github.com/MythTV/mythtv/commits/master but I haven't had a
> >>> chance
> >>> > to look into it yet (my builds grab the latest commits from upstream
> >>> and
> >>> > then add your patches on top)
> >>>
> >>> The master branch has just been updated and conflicts with one of my
> >>> OpenGL patches.  I rebased the master-rpi2 tree (to keep the patches
> >>> grouped above master/HEAD) so you will need to 'git reset --hard
> >>> softsystem/master-rpi2' after fetching.
> >>>
> >>> -- Lawrence Rust
> >>>
> >>> _______________________________________________
> >>> mythtv-users mailing list
> >>> mythtv-users at mythtv.org
> >>> http://lists.mythtv.org/mailman/listinfo/mythtv-users
> >>> http://wiki.mythtv.org/Mailing_List_etiquette
> >>> MythTV Forums: https://forum.mythtv.org
> >>
> >>
> >>  Hi Lawrence,
> >>
> >> Thanks for your work on this, and thanks for rebasing to keep all your
> >> stuff after the other commits, it's making it a lot easier to create a
> >> patch to apply to the builds (which I'm just doing by running)
> >>
> >> git format-patch bea7056a9672dc4435cdbc502f4d012660bc4b4b..HEAD --stdout
> >> > rpi28.patch
> >>
> >> New build is building right now at
> >> https://launchpad.net/~mythbuntu/+archive/ubuntu/stagingpatches/+build/8397888
> >>
> >> --
> >> Thomas Mashos
> >>
> >
> > Just a quick update, I was able to build fine it appears but I'm having
> > trouble running this on Ubuntu Mate for the Pi2 in both xcb and eglfs
> > modes. xcb mode just hangs after mythfrontend prints 'QT_QPA_PLATFORM=xcb"
> > while eglfs mode throws the following error.
> >
> > 2015-12-07 07:50:03.414068 I  Binding to UDP 192.168.0.255:6948
> > EGL Error : Could not create the egl surface: error = 0x300b
> >
> > 2015-12-07 07:50:03.734615 I  Using Frameless Window
> > 2015-12-07 07:50:03.734649 I  Using Full Screen Window
> > Handling Aborted
> > ./launch.sh: line 3:  3346 Aborted                 (core dumped)
> > QT_QPA_EGLFS_FORCE888=1 QT_QPA_PLATFORM=eglfs
> > QT_PLUGIN_PATH=/usr/lib/arm-linux-gnueabihf/qt5/plugins/platforms
> > LD_LIBRARY_PATH=/usr/lib /usr/bin/mythfrontend.real
> >
> >
> > I also tried the egl test in pi3d, which failed with
> >
> > For ref
> >   EGL_FALSE=0x0000
> >   EGL_BAD_DISPLAY=0x3008
> >   EGL_BAD_ATTRIBUTE=0x3004
> >   EGL_NOT_INITIALIZED=0x3001
> >   EGL_BAD_PARAMETER=0x300C
> >   EGL_WINDOW_BIT=0x0004
> >
> > Traceback (most recent call last):
> >   File "test_egl.py", line 60, in <module>
> >     d3dcompiler = ctypes.WinDLL(os.path.join(path, "d3dcompiler_47.dll"))
> > AttributeError: 'module' object has no attribute 'WinDLL'
> >
> >
> > So obviously EGLFS isn't working/installed right, but I'm not sure why the
> > xcb mode just appears to hang.
> >
> >
> > --
> > Thomas Mashos
> >
> 
> Lawrence,
> 
> After the merge to upstream master, I'm having difficulty building again.
> Any idea what is going on here
> https://launchpadlibrarian.net/233401745/buildlog_ubuntu-wily-armhf.mythtv_2%3A0.28.0~master.20160107.ff03753-0ubuntu5_BUILDING.txt.gz

This commit:
c5352a3 TV: OpenMAX compatibility with the Bellagio library implementation

made it possible to detect & support OpenMAX headers & libs from
providers other than Broadcom.  An important change was to remove the
'IL/' prefix from the OpenMAX headers.  Your configure line already
includes:

--extra-cflags="-I/opt/vc/include/IL" --extra-cxxflags="-I/opt/vc/include/IL"

so you just need to add the path to Broadcom's bcm_host.h:

--extra-cflags="-I/opt/vc/include" --extra-cxxflags="-I/opt/vc/include"

-- 
Lawrence Rust