Kodi Community Forum

Full Version: OpenELEC Testbuilds for RaspberryPi (Kodi 15.0) Part 2
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Quote:Everything tested here should eventually appear in future official OpenELEC builds, I've no interest in supporting out of tree features long term.

Just out of curiosity does that mean I should give up hope for frame sequential output? I'd help get it supported if I knew anything about coding... as always thanks for all your hard work. Much appreciated.
(2015-05-11, 17:36)MrMagic Wrote: [ -> ]Sorry, I forgot to report back. I tested with some BDMV rips that had problems before and unfortunately they persist with 0.8.0, so no improvement visible from my side.

Balls. I've just checked the build configuration and realised I only included the bump to 0.8.0 in a local test build, not the build I publish. It will definitely be in the next published build #0511... can you test again when it's available? Thanks.
New OpenELEC Isengard build #0511: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.0.2 #1 Mon May 11 23:52:43 BST 2015 armv6l GNU/Linux

# vcgencmd version
May  1 2015 18:56:06
Copyright (c) 2012 Broadcom
version f5f5ba5f2423c3dff6ccd4ca4b832264cd33f12a (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20150511235155-#0511-gd30d87f [Build #0511]

# vcdbg log msg 2>&1 | grep DTOK
001543.865: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (d30d87f0, changelog) and tip of XBMC master (c38f8fd2, changelog) with the following modifications: Build Highlights:
  1. New kernel 4.0.2
  2. Bump libbluray to 0.8.0 (for realz, this time)
  3. No longer include audiodecoder, audioencoder and visualization addons in the build - install from repo
Build Details:
  1. XBMC:
    • Fix grouping movies into sets in smartplaylists (PR:7095, 2 commits, 2 files changed)
    • Jenkins verbose (PR:7123, 1 commit, 9 files changed)
    • [renderer] Fix for number of surfaces being passed to codec (PR:7114, 1 commit, 1 file changed)
    • [SinkPi] Handle multichannel layout more like OMXAudio (PR:7100, 1 commit, 1 file changed)
    • [Confluence] - fix some invalid values / fonts (PR:7118, 2 commits, 18 files changed)
  2. platform:
    • readd sockets/cdevsocket.h, needed by IMX6 and TDA995x (PR:6, 1 commit, 4 files changed)
    • added guards for #defines on OS X and iOS for values already defined in CFPlugInCOM.h (7c02d9f2)
    • bump to 1.0.8 (caeac494)
  3. pvr.vbox:
    • add separate settings for internal and external connection parameters (0b658b67)
    • add ability to control the connection timeout (fixes #41) (310856b6)
    • bumped version to 1.1.0 (321c022c)
  4. newclock4:
    • Commits no longer in build:
      • [SinkPi] Handle multichannel layout more like OMXAudio (5a378d48)
  5. kernel 4.0.y:
    • Commits no longer in build:
      • bcm2835-mmc: Add option to disable MMC_QUIRK_BLK_NO_CMD23 (497dbba4)
      • 8192cu needs old wireless extensions (4c864893)
      • bcm2835-mmc: Add option to disable some delays (d2a4b3dd)
      • bcm2835-mmc: Default to disabling MMC_QUIRK_BLK_NO_CMD23 (488911bc)
      • vchiq: Allocation does not need to be atomic (67fd2c5a)
      • bcm2835-mmc: Add option to disable delays on some writeb calls (aa6b3e63)
      • bcm2835-mmc: warn on accesses without spinlock (39315a8e)
      • squash: sdhost support for faster core clocks, and force-pio mode (fdc5e429)
      • bcm2835-mmc/sdhost: of_dma_request_slave_channel isn't exported... (503ce7c5)
      • Merge pull request #954 from pelwell/rpi-4.0.y (75c5da71)
      • squash: bcm2709: Allow to build without SMP (a80e690c)
      • squash: entry_macro: Allow to build without SMP (12149d45)
      • squash: dwc_otg: Allow to build without SMP (ef265947)
      • squash: bcm2709: Fix build with SYSTEM_TIMER defined (121d87c8)
      • 8192cu: CONFIG_AP_MODE hardcoded in autoconf.h (cb97407b)
      • Revert "bcm2835-mmc: Add locks when accessing sdhost registers" (a7865cec)
      • Revert "bcm2835-mmc: warn on accesses without spinlock" (73f24007)
      • Revert "bcm2835-mmc: Add option to disable delays on some writeb calls" (95ae72d6)
      • Revert "bcm2835-mmc: Add option to disable some delays" (a52246c7)
      • Revert "bcm2835-mmc: Default to disabling MMC_QUIRK_BLK_NO_CMD23" (3fefacdc)
      • Revert "bcm2835-mmc: Add option to disable MMC_QUIRK_BLK_NO_CMD23" (4c0553ea)
      • Revert "bcm2835-mmc: Add range of debug options for slowing things down" (b8a19a3a)
      • vchiq: Fix wrong condition check (c7f63111)
      • Merge pull request #960 from pranith/rpi-4.0.y (41a5ce7c)
  6. Additional commits/pull requests/changes not yet merged upstream:
Strange sound problem may fixed #0511 test build. Smile
Hello,

I would use my raspi-openelec on the way, and control this about yatse.
For that, I will activate "tethering", but after a while the enabled is again disabled. ??

Can you say: How I can fix enabled tethering ?
Where are you enabling "tethering", in Yatse? I'm not familiar with this option if it's available in OpenELEC, or maybe you mean something else WiFi related? I'd suggest testing OE 6 Beta 1 to confirm if it's an upstream issue.
Hello,

I mean tethering in Kodi-Openelec-Settings.
But, I will test it in Beta 1.
Ah right, it's only visible once you enable WiFi (I've only wired here). As usual, what was the last working version, but confirming behaviour of Beta 1 will indicate whether it's an upstream issue.
(2015-05-11, 15:50)Kiralina Wrote: [ -> ]I updated from both 5.0.8 and 5.95.1 with the latest #0510 (on separate micro sdhc) and have no problems with my current add ons.
All works well and I had no reboots on my PI 2. On both cards I have turbo overclocking with no issues.

Thank you, but I do mean the technical aspect. Are there any differences in the final result.
Anyone else have a problem locking the settings/addons/etc with master lock code?
I set a 'numeric pass code' and checked lock music, pictures, programs & scripts, file manager, lock settings==all, lock addon-manager.
I can still access all of these things, including settings, after even a reboot, without entering my passcode.

It works on my Pi2 with #511, using the master profile, but on the Pi running #511 in my kids room using the second profile it doesn't.

I will try setting the pass code with master profile on it later and see if after switching to second profile the lock sticks.
Since build #0505 I have a problem with stuttering live tv video where the audio temporarily goes out of sync. This seems to occur only when OMXPlayer is enabled. Unfortunately I can't recreate the issue with a recorded live tv clip, it only happens when playing live.

#0504 and 5.95.1 appear to work perfectly.

Here's the log http://paste.ubuntu.com/11092275/
(2015-05-12, 00:55)Milhouse Wrote: [ -> ]
(2015-05-11, 17:36)MrMagic Wrote: [ -> ]Sorry, I forgot to report back. I tested with some BDMV rips that had problems before and unfortunately they persist with 0.8.0, so no improvement visible from my side.

Balls. I've just checked the build configuration and realised I only included the bump to 0.8.0 in a local test build, not the build I publish. It will definitely be in the next published build #0511... can you test again when it's available? Thanks.
I tested again and there's good news and bad news.

The good news is that the BDMV rip I tried (Transformers 4), no longer stops playing at the point where it previously did.

The bad news is that after the problem point, the movie is no longer watchable because it starts to display continuous digital artifacts.

Apparently libbluray is now more resilient against the "lost sync byte" error, but it doesn't get back in sync fully and starts to display artifacts.

As with my previous tests, playing the m2ts file directly still works just fine.
(2015-05-12, 19:36)Leopold Wrote: [ -> ]Since build #0505 I have a problem with stuttering live tv video where the audio temporarily goes out of sync. This seems to occur only when OMXPlayer is enabled. Unfortunately I can't recreate the issue with a recorded live tv clip, it only happens when playing live.

There are a couple of possibilities in #0505. The PVR addons got bumped, and this:
Quote:omxplayer: testing give preference to dts timestamp

was introduced. I'll revert the omxplayer commit for tonight's build (only - as that commit did fix dvd stutter issues), so please test that.

Any more info? Is playback fine with omxplayer acceleration disabled?
How long is audio out of sync for? Does it coincide with errors in the stream?
(2015-05-12, 19:48)MrMagic Wrote: [ -> ]I tested again and there's good news and bad news.

Thanks for the update, sorry it's not a complete fix. At least upstream has now bumped libbluray to 0.8.0 so this will be in all future official builds (including the next OE 6 Beta).
New OpenELEC Isengard build #0512: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.0.2 #1 Wed May 13 01:30:25 BST 2015 armv6l GNU/Linux

# vcgencmd version
May  1 2015 18:56:06
Copyright (c) 2012 Broadcom
version f5f5ba5f2423c3dff6ccd4ca4b832264cd33f12a (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20150513012934-#0512-gac669fc [Build #0512]

# vcdbg log msg 2>&1 | grep DTOK
001540.236: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (ac669fc7, changelog) and tip of XBMC master (e2793a3e, changelog) with the following modifications: Build Highlights:
  1. New Shadertoy visualization
    popcornmix Wrote:I've added a port of topfs2's shadertoy visualisation (http://forum.kodi.tv/showthread.php?tid=204991)

    Now desktop shaders aren't entirely suitable for an embedded platform, and the way shadertoy works is to do everything in the fragment shader without using the vertex shader. That means we are doing something closer to raytracing than 3D rendering. i.e. for every pixel the whole scene is considered to work out its contribution to that pixel.

    As such I've had to remove some shaders. I've had to render to a smaller offscreen buffer. Some that work are too slow to look good (although simplifying them may be an option). However some work quite nicely, and it's pretty simple to try others from the shadertoy website.

    The different effects come from the presets button. When playing audio go fullscreen and press enter and select the rightmost mountain shaped button.

    When using Confluence, it's also best to enable "Hide fanart in full screen visualistion" otherwise the visualisations will look "washed out".
  2. Revert: omxplayer: testing give preference to dts timestamp
  3. Added: patch: Enable the mailbox driver (new kernel options)
Build Details:
  1. OpenELEC:
    • bump audiodecoder.qsf (PR:4144, 1 commit, 1 file changed)
    • Add screensavers (PR:4142, 8 commits, 15 files changed)
    • sqlite: update to sqlite-autoconf-3081001 (4986022b)
    • mesa: update to mesa-10.5.5 (6f00a01a)
    • projects/imx6/patches/linux: fix dvbsky driver (e045e8e0)
    • libbluray: update to libbluray-0.8.0, enable UDF support (50d086b1)
    • libbdplus: update to libbdplus-0.1.2 (6cacd6f7)
    • platform: update to platform-1.0.8 (5f090ea3)
  2. XBMC:
    • [Translations] fix escaping of ' " ' in strings.po (PR:7126, 1 commit, 1 file changed)
    • [imx] RGB -> BGR for RenderCapture (PR:7124, 1 commit, 2 files changed)
    • [mmalrenderer] Fix for stereo view modes (PR:7115, 1 commit, 1 file changed)
  3. dcadec:
    • Take XXCH core speaker mask into account. (8b2e70bd)
    • Rename ‘ch’ suffix to ‘spkr’ on a function. (fe9c34ef)
    • Add ta_strdup() function. (68986e3c)
    • Improve WAV file writer. (a0d16229)
    • Change dcadec_stream_open() signature. (da9b3fc4)
    • Add ‘-m’ option to dcadec for mono WAV output. (d1d37065)
    • Update feature list. (415d9547)
    • Report number of skipped frames and clipped samples. (6b643aed)
    • Remove SOURCE_PCM_RES flag and ‘-s’ dcadec option. (8219a314)
    • Ignore CORE_SYNTH_X96 flag when setting profile. (9b35ec60)
    • Perform downmix clipping after filtering. (6f6126a8)
    • Skip extra XBR/X96 channel sets when downmixing. (10f875bc)
    • Fix dcacut.c compilation. (396e7565)
  4. libcec:
    • update libplatform ref (b7063a97)
    • rebranding: XBMC -> Kodi in the installer (e79f0e68)
    • updated create-installer.bat to use the new binaries (d957cdce)
    • fixed: errors in the installer if eventghost isn't installed. (e5a60b42)
    • sync ChangeLog (18555911)
  5. pvr.argustv:
    • [debian] fix packaging with multi-arch (PR:11, 1 commit, 1 file changed)
    • Option to group single recordings in folders (PR:10, 1 commit, 7 files changed)
  6. pvr.demo:
    • [debian] fix packaging with multi-arch and missing kodiplatform dependency (PR:7, 2 commits, 2 files changed)
  7. pvr.dvblink:
    • [debian] fix packaging with multi-arch (PR:10, 1 commit, 1 file changed)
  8. pvr.dvbviewer:
    • [debian] fix packaging with multi-arch and missing kodiplatform dependency (PR:5, 2 commits, 2 files changed)
  9. pvr.filmon:
    • [debian] fix packaging with multi-arch (PR:9, 1 commit, 1 file changed)
  10. pvr.hts:
    • [debian] fix packaging with multi-arch (PR:29, 1 commit, 1 file changed)
  11. pvr.iptvsimple:
    • [debian] fix packaging with multi-arch (PR:15, 1 commit, 1 file changed)
  12. pvr.mediaportal.tvserver:
    • [debian] fix packaging with multi-arch (PR:6, 1 commit, 1 file changed)
  13. pvr.mythtv:
    • [debian] fix packaging with multi-arch (PR:8, 1 commit, 1 file changed)
  14. pvr.nextpvr:
    • [debian] fix packaging with multi-arch (PR:8, 1 commit, 1 file changed)
  15. pvr.njoy:
    • [debian] fix packaging with multi-arch and missing kodiplatform dependency (PR:5, 2 commits, 2 files changed)
  16. pvr.pctv:
    • [debian] fix packaging with multi-arch (PR:9, 1 commit, 1 file changed)
  17. pvr.stalker:
    • [debian] fix packaging with multi-arch (PR:5, 1 commit, 1 file changed)
  18. pvr.vbox:
    • [debian] fix packaging with multi-arch (PR:45, 1 commit, 1 file changed)
  19. pvr.vuplus:
    • [debian] fix packaging with multi-arch and missing kodiplatform dependency (PR:5, 2 commits, 2 files changed)
  20. pvr.wmc:
    • [debian] fix packaging with multi-arch and missing kodiplatform dependency (PR:6, 2 commits, 2 files changed)
  21. newclock4:
    • New commits in this build:
      • - added contrib dir and kissfft sources (7634ff64)
      • - added kissfft based rfft implementation and unit tests for it (4b44ce70)
      • - replaced our buggy fft implementation with rfft (05dc8b3d)
      • introduce new force window activation which do not check for active modals (d6c936bf)
      • force activate window screen calibration in CGUIDialogVideoSettings (53bcb646)
      • force activate fullscreen window with CApplication::SwitchToFullScreen() (d0678f64)
      • corrects format + indentation (ee2846b8)
      • force activate fullscreen for TMSG_SWITCHTOFULLSCREEN (9aa4538e)
      • Revert: omxplayer: testing give preference to dts timestamp (ebe5ed44)
    • Commits no longer in build:
      • [mmalrenderer] Fix for stereo view modes (01aa2574)
      • fixed: broken rfft implementation used in audio visualizer api (2a9ea18c)
      • squash: Add rfft.cpp to utils lib (5586fd98)
  22. kernel 4.0.y:
    • New commits in this build:
      • BCM270x: Correct vcio device memory resource and add irq resource (117316a9)
      • video: fbdev: bcm2708_fb: Don't panic on error (6631cf53)
      • BCM2708: vcio: Fix checkpatch issues (c81b388c)
      • BCM2708: vcio: Remove unused code and compact comments (d2ca8b29)
      • BCM2708: vcio: Move character device teardown to driver remove (ae524a32)
      • BCM2708: vcio: Restructure error paths (64987b07)
      • BCM2708: vcio: Move some macros from header to driver (8729fd45)
      • BCM2708: vcio: Only store the register base address (8e02083d)
      • BCM2708: vcio: Do not print messages in module init (0c63a9ed)
      • BCM2708: vcio: Use device resources (34cc4116)
      • mailbox: bcm2708: Add bcm2708-vcio (9507617f)
      • BCM270x: power: Change initcall level to subsys (ad446d1d)
      • BCM270x: Use bcm2708-vcio (61a75aa0)
      • BCM270x: Remove arch driver vcio.c (db783a2d)
      • BCM270x_DT: Add mailbox bcm2708-vcio (75f995b0)
      • bcm2835: bcm2835_defconfig enable BCM2708_MBOX (77b2d215)
      • bcm2835: Add mailbox bcm2708-vcio to Device Tree (3f956a8a)
      • mailbox: bcm2708-vcio: Allocation does not need to be atomic (bbbbf670)
      • mailbox: bcm2708-vcio: Check the correct status register before writing (686928c9)
      • rpi-ft5406: Add touchscreen driver for pi LCD display (88a6d667)
  23. Additional commits/pull requests/changes not yet merged upstream:
    • Added: patch: Enable the mailbox driver (new kernel options)
    • Added: patch: [visualisation] Add shadertoy visualisation (alt. source)