•   
  • 1
  • 15
  • 16
  • 17(current)
  • 18
  • 19
  • 168
  •   
  Thread Closed
OpenELEC Testbuilds for RaspberryPi (Kodi 16.0)
I did that, watched an hour H264 show, when it ended, the clock remained at 250. It went between 250-333 all throughout the show, and remained at 250 after the video ended. Been at the desktop for 20 min, the H264 clock is still at 250.
(2015-07-24, 01:40)popcornmix Wrote:
(2015-07-23, 22:40)michidg Wrote: i searched the forum a bit and found this thread : http://forum.kodi.tv/showthread.php?tid=226933&page=2

Seems to be the same problem ! downloaded his test file and also stops after some seconds.

Okay, both samples have the same issue. I've determined that firmware from about six months ago can play these files.
I've bisected it and found the commit "video codec: refactor userdata release mec hanics in categoriser".

Now, I don't know a lot about this code, and the commit is too large to make immediate sense.
I really need the author (Deborah) to help but she's not currently around.

I have determined that reverting this commit does succeed and allows the samples to play. If you want to try it then I've built a test firmware:
https://dl.dropboxusercontent.com/u/3669...150724.zip

However I suspect the commit was added for a reason, so it could cause issues (probably with MVC). But give it a try and report back how it works.

made a short test, only some minutes on every problem dvd, and all are working again. Thanks !
Will make longer tests the next days.

What do you mean with "she's not currently around". Could this take weeks or month ?

Can I overwrite with your test firmware also future milhouse builds or could there be incompatibilities ?
(2015-07-24, 04:19)michidg Wrote: made a short test, only some minutes on every problem dvd, and all are working again. Thanks !
Will make longer tests the next days.

What do you mean with "she's not currently around". Could this take weeks or month ?

Can I overwrite with your test firmware also future milhouse builds or could there be incompatibilities ?

In general you can update to a new milhouse and then replace the firmware.
Very occasionally there may be a new firmware feature that kodi relies on, or a bug fix that you would benefit from.
In that case, let me know and I can build a new firmware with the patch reverted.

Deborah is off for health reasons. I hope she will be back, but we have no definite date.
I will have a close look at the patch to see if I can find a way of fixing it for these files, but I'm not certain if that will be successful.
The other possibility is to revert it for everyone, but that may cause regressions elsewhere (possibly with MVC which is when the patch was added).
If I can't find any regressions with my test files, I may revert it for testing purposes and see if any problems emerge.
(2015-07-24, 12:06)popcornmix Wrote: In general you can update to a new milhouse and then replace the firmware.
Very occasionally there may be a new firmware feature that kodi relies on, or a bug fix that you would benefit from.
In that case, let me know and I can build a new firmware with the patch reverted.

Deborah is off for health reasons. I hope she will be back, but we have no definite date.
I will have a close look at the patch to see if I can find a way of fixing it for these files, but I'm not certain if that will be successful.
The other possibility is to revert it for everyone, but that may cause regressions elsewhere (possibly with MVC which is when the patch was added).
If I can't find any regressions with my test files, I may revert it for testing purposes and see if any problems emerge.

ok thanks. wanted only to know, if there are already plans for changes, that would cause the firmware stop working with newer builds.
but if you can build then a new one, its ok.

Can you tell me where the commit is or which files where changed by this? I would like to take a look at it, simply because i am interested. have been c++ developer in my previous job.

I looked a bit at the more detailled component log and at the kodi source. CMMALVideo::m_output_ready is empty. So the dec_input and dec_output callbacks get never called again.
You said the commit is a change with "userdata". As i can see this is passed to the callbacks.
(2015-07-24, 15:03)michidg Wrote: Can you tell me where the commit is or which files where changed by this? I would like to take a look at it, simply because i am interested. have been c++ developer in my previous job.

This is in the gpu firmware which is not open source. Open sourcing the hardware video decode driver is complicated due to the codec licences...
ah ok. now i understand why i could only find header files.

test firmware was running now about 5 hours. no problems !
(2015-07-13, 18:29)invisable Wrote: Hi Millhouse, can I ask if your still including your latest librtmp in Openelec 16 the only reason I ask is because I noticed the test build release you made on May 16 (#0516) the librtmp is slightly larger then the ones on the recent builds ?
Many Thanks.

I've just noticed that the extra rtmpdump squash commit added to newclock4 on 16 May is not being processed correctly (ie. it is being silently ignored), so only the original KSV commit is being included in recent builds.

Both commits will be in tonight's build (a little delayed while I've been working this out...) - apologies for not noticing when you brought this up.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
New OpenELEC Jarvis build #0724: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.1.3 #1 Sat Jul 25 01:12:30 BST 2015 armv6l GNU/Linux

# vcgencmd version
Jul 24 2015 14:44:13
Copyright (c) 2012 Broadcom
version 0f482c5017ad4526a6beb77fe39c93189c664fd4 (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20150725010646-#0724-g0cfb546 [Build #0724]

# vcdbg log msg 2>&1 | grep DTOK
001577.773: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (0cfb546d, changelog) and tip of XBMC master (4b19f7aa, changelog) with the following modifications: Build Highlights:
  1. New firmware
  2. New kernel 4.1.3
  3. Include librtmp update
Build Details:
  1. Firmware (Jul 24):
    • firmware: video_decode: Don't wait for a valid timestamp to output frames. See: link
  2. XBMC:
    • [adsp] disable binary add-on until they build on all platforms (PR:7605, 1 commit, 3 files changed)
    • [gui] CGUIDialogYesNo warnings (PR:7598, 2 commits, 1 file changed)
    • [gitignore] don't clean Makefile in xbmc/messaging/ (PR:7608, 1 commit, 1 file changed)
    • cmake: unify output of build messages by using message(STATUS ...) (4b19f7aa)
  3. platform:
    • fixed: set libdir on linux when CMAKE_INSTALL_PREFIX != /usr (PR:11, 1 commit, 1 file changed)
  4. newclock4:
    • New commits in this build:
      • [omxhelper] Ignore embedded coverart image when looking for supported video streams (7c24e700)
      • [mmalcodec] Don't set initial unknown timestamps to zero (303b25f4)
      • [mmalcodec] Tidy up setting of timestamp flags (a734ede6)
      • [mmalrenderer] Block until vsync in RenderUpdate (1f5b4a04)
    • Commits no longer in build:
      • fix stuttering indruduced by 9c3adb5a9f390ad73c52469ed79de297dad68d80 (4ab32b45)
  5. kernel 4.1.y:
    • New commits in this build:
      • BCM270X_DT: Use i2c_arm for rtc and bmp085 overlays (c9dce515)
      • BCM2708_DT: CM dtparams for audio, watchdog and RNG (1126ca25)
      • vchiq: Use firmware API (e3508264)
      • thermal: bcm2835: Use firmware API (da339e29)
      • cpufreq: bcm2835: Use firmware API (b5f2f604)
      • fbdev: bcm2708: Use firmware API (b436501d)
      • bcm2835: Add firmware property to affected devices (9be804b3)
      • rpi-ft5406: Use firmware API (193a0cac)
      • irqchip: bcm2835: Add FIQ support (062df639)
      • dwc_otg: Add ARCH_BCM2835 support (5b2e3f5d)
      • bcm2835: Use DWC_OTG (cdd86992)
      • Merge pull request #1086 from notro/fiq (ab1bc204)
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Updated to 0724. While watching 1080i60 mpeg2 live TV, I am seeing some strange audio dropouts and the number of missed frames going up quickly. Rolled back to 0723 and the issue goes away. I only use MMAL, OMX is disabled.
Can you repeat with #0724 and upload a debug log?
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Sure! Sorry it is so large.

http://sprunge.us/GOUO
(2015-07-25, 03:39)zaphod24 Wrote: Updated to 0724. While watching 1080i60 mpeg2 live TV, I am seeing some strange audio dropouts and the number of missed frames going up quickly. Rolled back to 0723 and the issue goes away. I only use MMAL, OMX is disabled.

I saw similar issues with 724 but I was also seeing the audio not syncing in 723. I changed the following settings which fixed the audio sync issue. I'm using a HDHomerun Connect (OTA) so stations with good signal seem to be working well
.
In Playback I have Adjust display refresh rate turned off,and Sync playback to display off.
Under Acceleration I am using MMAL
For Deinterlace video I set it to Auto, Deinterlace method is Auto Select and View mode is normal.
In playback, both adjust display refresh rate and sync are enabled with the resample audio option. I'm using MMAL and for deinterlace I'm using MMAL-Advanced. Everything works great that way with 0723.
(2015-07-25, 03:39)zaphod24 Wrote: Updated to 0724. While watching 1080i60 mpeg2 live TV, I am seeing some strange audio dropouts and the number of missed frames going up quickly. Rolled back to 0723 and the issue goes away. I only use MMAL, OMX is disabled.

Could you test this build?
Should be a Milhouse style build with a fix for the problem introduced with dvdplayer/mmal playback. Would be useful to know if the fix is good for you.
The test build works great!
  •   
  • 1
  • 15
  • 16
  • 17(current)
  • 18
  • 19
  • 168
  •   
  Thread Closed
 
Thread Rating:
  • 10 Vote(s) - 5 Average



Logout Mark Read Team Forum Stats Members Help
OpenELEC Testbuilds for RaspberryPi (Kodi 16.0)510