• 1
  • 129
  • 130
  • 131(current)
  • 132
  • 133
  • 168
OpenELEC Testbuilds for RaspberryPi (Kodi 16.0)
New OpenELEC Jarvis build #1105: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.3.0 #1 Fri Nov 6 01:12:21 GMT 2015 armv6l GNU/Linux

# vcgencmd version
Nov  1 2015 14:49:52
Copyright (c) 2012 Broadcom
version 8f1c8e72174c0415fd479c1c7be66b880fab3d79 (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20151106011130-#1105-ge1c98c8 [Build #1105]

# vcdbg log msg 2>&1 | grep DTOK
001690.168: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (e1c98c83, changelog) and tip of XBMC master (c4a9e5d1, changelog) with the following modifications: Build Highlights:
  1. Fix inversion of export library dialogs, fixes #16358
  2. Fix JSON output of album and artist mood to be array as in spec
Build Details:
  1. XBMC:
    • [iimage] Prevent possible access violations, especially for gifs. (PR:8348, 7 commits, 8 files changed)
    • [guilib] colordiffuse not applying to progresstextures (PR:8347, 1 commit, 1 file changed)
    • [guilib] fix togglebutton label not rendered properly (PR:8341, 1 commit, 2 files changed)
    • [addons] fix library path building (PR:8349, 1 commit, 1 file changed)
    • [music] don't use the directory cache when cleaning the database (fix… (PR:8352, 1 commit, 1 file changed)
  2. pvr.vdr.vnsi:
    • Crash fixes (PR:32, 5 commits, 7 files changed)
  3. newclock5:
    • New commits in this build:
      • renderer: promote interface up to BaseRenderer (1242858a)
      • fixup! filesystem: Make support of browsing into archives optional (d4e5b20e)
    • Commits no longer in build:
      • [pvr] Add progress bar with timeshift progress when timeshift is active (32e152fb)
  4. Additional commits/pull requests/changes not yet merged upstream:
    • Reverted: [pkg] PR:8348: [iimage] Prevent possible access violations, especially for gifs. (Reason: newclock5 rebase required)
    • Added: [pkg] PR:8356: Fix inversion of export library dialogs, fixes #16358
    • Added: [pkg] PR:8358: Fix JSON output of album and artist mood to be array as in spec
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.
Hi Millhouse,

I see you reverted back 3.0.1 libcec change, can I ask you why? There was some bug or something?

Thank you
Sound stopped working for Hifiberry Digi after the update above, alsa no longer can recognize Alsa output... it worked in previous Kodi 16 alpha 4. I'm using RPi 2.


(edit: if you follow https://www.hifiberry.com/guides/configu...-overlays/)
lsmod says:
Code:
Module                  Size  Used by
snd_soc_hifiberry_digi     2935  0
snd_soc_bcm2835_i2s     5548  2
regmap_mmio             2214  1 snd_soc_bcm2835_i2s
snd_soc_wm8804_i2c      1344  1
snd_soc_wm8804          6957  1 snd_soc_wm8804_i2c
snd_soc_core          104586  3 snd_soc_wm8804,snd_soc_bcm2835_i2s,snd_soc_hifiberry_digi
regmap_i2c              2753  1 snd_soc_wm8804_i2c
snd_compress            6979  1 snd_soc_core
snd_pcm_dmaengine       2815  1 snd_soc_core
snd_pcm                61657  3 snd_soc_wm8804,snd_soc_core,snd_pcm_dmaengine
snd_timer              16664  1 snd_pcm
snd                    40412  4 snd_soc_core,snd_timer,snd_pcm,snd_compress
bcm2835_gpiomem         2652  0
bcm2708_rng              925  0

and so aplay -l returns
Code:
**** List of PLAYBACK Hardware Devices ****
card 0: sndrpihifiberry [snd_rpi_hifiberry_digi], device 0: HifiBerry Digi HiFi wm8804-spdif-0 []
  Subdevices: 1/1
  Subdevice #0: subdevice #0

but still Kodi does not see this device
(2015-11-06, 14:09)JimmySmith2 Wrote: Hi Millhouse,

I see you reverted back 3.0.1 libcec change, can I ask you why? There was some bug or something?

Thank you

No, still pulling latest libcec master commits from top-of-tree:
(2015-11-06, 04:06)Milhouse Wrote:
  • Includes latest libcec master (2c675dac, ahead +31)

These builds are currently 31 libcec commits ahead of the libcec version used by OpenELEC master, and include the 3.0.1 bump.

What revert are you referring to?
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.
(2015-11-06, 20:15)Milhouse Wrote: These builds are currently 31 libcec commits ahead of the libcec version used by OpenELEC master, and include the 3.0.1 bump.

What revert are you referring to?

I only checked package with patch on github (https://github.com/MilhouseVH/OpenELEC.t...package.mk).. But it seems I missing something in the building process...
(2015-11-06, 21:02)JimmySmith2 Wrote:
(2015-11-06, 20:15)Milhouse Wrote: These builds are currently 31 libcec commits ahead of the libcec version used by OpenELEC master, and include the 3.0.1 bump.

What revert are you referring to?

I only checked package with patch on github (https://github.com/MilhouseVH/OpenELEC.t...package.mk).. But it seems I missing something in the building process...

Ah, that repo isn't used at all by my build process.
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.
Smile Apologies for dummy questions, I am newbie.. I am looking for a way to correctly implement 3.0.1 into OE6 sources. Or is it internal thing? There are voices that 3.0.0 version from OE6 stable has buggy behauvior with (some?) Samsung TVs..
(2015-11-06, 21:19)JimmySmith2 Wrote: Smile Apologies for dummy questions, I am newbie.. I am looking for a way to correctly implement 3.0.1 into OE6 sources. Or is it internal thing? There are voices that 3.0.0 version from OE6 stable has buggy behauvior with (some?) Samsung TVs..

You'll need to wait for a new OE6 build with libcec 3.0.1, or build your own OE6 and bump the libcec package yourself.
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.
Ok. Thank you anyway!
hi Milhouse, I feel like guilty - I have checked a couple of pages backwards here and the problem with audio output has been reported several times already. I this something you are planning to look into at a later stage?
(2015-11-06, 21:59)mlab Wrote: I this something you are planning to look into at a later stage?

Fixing Hifiberry is on the list... not really sure what the problem is right now but the backtraces in dmesg suggest a bug or incompatibility with kernel 4.3.0. If Hifiberry is important to you I'd suggest going back to the last working build (presumably build #1101 with kernel 4.1.2) until there's news of a fix. Unfortunately there always seems to be a problem with these Hifiberry devices after major kernel bumps, and it can take a little while to resolve the issues.
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.
Thanks for reply Smile

Indeed it could be a kernel bug, just since ALSA detects Hifiberry correctly - why it is not listed under Kodi audio devices? I would be happy to test newer builds, just it's not easy to do so without audio (in my case I use SPDIF output to external DAC).
New OpenELEC Jarvis build #1106: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.3.0 #1 Fri Nov 6 22:53:04 GMT 2015 armv6l GNU/Linux

# vcgencmd version
Nov  6 2015 21:55:08
Copyright (c) 2012 Broadcom
version 40eee6c6d44edb1db4302217927a66a8dffd9374 (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20151106225214-#1106-ge1c98c8 [Build #1106]

# vcdbg log msg 2>&1 | grep DTOK
001708.076: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (e1c98c83, changelog) and tip of XBMC master (1f102fd7, changelog) with the following modifications: Build Highlights:
  1. New firmware
  2. dcdadec: Improved DTSHD-MA (XLL) sync loss recovery and resilience to errors
Build Details:
  1. Firmware (Nov 6):
    • firmware: di_adv: Move to new scheme where qpu code handles top/bottom edges See: link
    • firmware: di_adv: Add config setting to add nop delays to shader. See: link
    • firmware: vcilcs: Avoid a potential deadlock when very threaded. See: link
    • firmware: vrf: Add spinlock around vrf acquire/release calls to avoid restoring an invalid p10 from ISR context
    • firmware: rpi_display: only ratelimit if the backlight is actually changed. See: link
    • firmware: di_adv: Support multiple instances of qpu deinterlace at SD resolution. See: link
    • linux: rpi-ft5406: Use interruptible sleep to avoid high load reported. See: link
    • linux: dts: Added overlay for Adafruit PiTFT 2.8 capacitive touch screen. See: link
    • linux: config: Add MCP320X. See: link
    • linux: Build i2c_gpio module and add a device tree overlay to configure it. See: link
  2. XBMC:
    • FIX: [amc] VC1 tweak (PR:8359, 1 commit, 1 file changed)
    • [coverity] fixes (PR:8346, 5 commits, 8 files changed)
    • Fix TRAC 16359: not filtering tvshows/movies already in tag (PR:8342, 1 commit, 1 file changed)
    • [PVR] Fix: Do not try to update read-only timers. (PR:8357, 2 commits, 2 files changed)
    • [texture/iimage] Pass the texture's width to iimage's Decode() method. (PR:8361, 1 commit, 1 file changed)
  3. dcadec:
    • Improve XLL sync loss recovery. (4de21178)
    • Clip XLL samples by default. (80258daa)
    • Correctly account for LSB section CRC bytes. (765cdb3a)
    • Improve XLL error resilience. (81121d11)
  4. pvr.vdr.vnsi:
    • Various code cleanup commits, preparing for more crash bug fixes. (PR:33, 19 commits, 13 files changed)
  5. newclock5:
    • New commits in this build:
      • gl: split output stage from yuv2rgb shader (98c673ee)
      • ffmpeg: Update AVPixelFormat and AV_PIX_FMT_* to compile with master (0a456688)
      • Merge pull request #313 from philipl/fernetmenta (79142168)
      • squash: Fixup prototype to decode (2adc5fd6)
    • Commits no longer in build:
      • renderer: promote interface up to BaseRenderer (1242858a)
  6. Additional commits/pull requests/changes not yet merged upstream:
    • Reverted: [pkg] PR:8346: [coverity] fixes (Reason: newclock5 rebase required)
    • Reverted: [pkg] PR:8361: [texture/iimage] Pass the texture's width to iimage's Decode() method. (Reason: newclock5 rebase required)
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.
(2015-11-05, 21:24)mule1 Wrote: Is it usual that navigating (using skips) in a 3D movie using the latest testbuilds (started testing from build #1029) is yet nearly impossible?
In rare cases after a skip the movie continues to play. But most times Kodi only fills the buffer (shown in the GUI) but never continues to play the movie or the movie continues to play but only with 2-4 frames per second or if it continues there are heavy stutters. Stopping and trying to resume shows the same results. ISO seems to be even worse than MKV files. Navigating in non 3d movies is working as it should.

No one else having these issues? Is navigating (skipping forward, backward, fast forward, rewind) in 3d movies working flawless for you?
(2015-11-04, 00:07)popcornmix Wrote: The Advanced deinterlace at 1080i60 is expensive. It generates a lot of sdram traffic.
DVB USB dongles generate a lot of data, and depending on the driver, it may need to read prompty to avoid losing data.

The suspicion is that the heavy sdram traffic delays the USB driver occasionally and it loses a packet, causing corruption in the video.

Now there may be ways of working around this. Forcing the deinterlace to spread out its sdram accesses the right amount may solve this.
E.g. currently deinterlace might take 13ms every 16ms. If I were to make the deinterlace pause for 10us every 100us, it would still likely get it's job done,
and hopefully the USB driver won't be held up during the pauses.

Unfortunately it's not something I can test (I have encrypted cable at home, and too weak a terrestrial signal at both work and home).
I may be able to produce a test build that attempts this (perhaps with a config setting to tweak how frequent/long the pauses are).

@bagofcrap24 @polo_joe

I've added a test setting to see if we can fix the artefacts with advanced 1080i deinterlave and locally attached DVB dongle.
Use latest Milhouse build and add:
Code:
hvs_priority=0xN0000000

Replace the N with a hex digit from 0 to f (i.e. 0,1,2,3,4,5,6,7,8,9,a,b,c,d,e,f). That number determines how long to pause for, with 0 meaning no change in behaviour, to f means lots of pausing.
Start with 8, and if you still have artefacts then increase it, and if no artefacts then decrease it. Try to find the minimum setting that doesn't have artefacts.
Too high a setting will cause lagging of the video, as the deinterlace will take longer. Hopefully there will be a value that fixes artefacts and doesn't have lagging.
  • 1
  • 129
  • 130
  • 131(current)
  • 132
  • 133
  • 168

Logout Mark Read Team Forum Stats Members Help
OpenELEC Testbuilds for RaspberryPi (Kodi 16.0)10