2015-07-15, 12:36
2015-07-15, 13:10
(2015-07-15, 03:26)Milhouse Wrote: Testing your video with #0714 on RPi2, I tried "Auto" and "MMAL - Advanced" but this hangs Kodi on a frozen frame - it's not garbled like your screenshot however. It does require a reboot of OpenELECYes, I get this behaviour - no garbled video. It's actually an interaction with the HEVC qpu patch, and I have a fix for that.
Quote:I also found a sample 1080i video called "PAW Patrol.ts" - I think this video had problems a while ago with changes of frame rate (or was it number of audio channels?) at the advert transitions
Yes, after 45 seconds we start getting non-interlaced frames and I wasn't handling that correctly from qpu code path. I have a fix for that.
2015-07-15, 14:43
(2015-07-15, 12:36)popcornmix Wrote:(2015-07-14, 14:21)misa Wrote: gpu_freq=400
System is freezing and cec is not working anymore
Pi1 or Pi2?
Two possibilities.
Your power supply isn't sufficient.
You need to set over_voltage to make that overclock reliable.
My config on a Pi2 with the right adapter (working good for 1 year now without any problems)
arm_freq=1062
sdram_freq=480
core_freq=590
over_voltage=4
gpu_mem=320
hdmi_ignore_cec_init=1
over_voltage_sdram_c=0
over_voltage_sdram_i=0
over_voltage_sdram_p=2
avoid_pwm_pll=1
v3d_freq=400
force_turbo=1
# decode_MPG2=0x00000000
# decode_WVC1=0x00000000
# decode_DTS=0x00000000
# decode_DDP=0x00000000
Whit this my Pi2 is stable but when Iput in gpu_freq=400 it strarts to freeze
2015-07-15, 14:48
Stop overclocking sdram_freq - it gives practically no benefit, but can often result in instability with recent firmware.
I too had a completely stable Pi2 for months with sdram_freq=500, but recent firmware resulted in random crashes and freezes, which went away once I stopped overclocking sdram_freq.
You could possibly regain stability by bumping up the sdram voltage, but unlike the Pi1, the Pi2 benefits very little from overclocking sdram (mainly because the ARMv7 SoC has increased L2 cache) so ultimately the increased risk of instability is just not worth the hassle for the meagre improved performance (which you probably won't even notice).
I too had a completely stable Pi2 for months with sdram_freq=500, but recent firmware resulted in random crashes and freezes, which went away once I stopped overclocking sdram_freq.
You could possibly regain stability by bumping up the sdram voltage, but unlike the Pi1, the Pi2 benefits very little from overclocking sdram (mainly because the ARMv7 SoC has increased L2 cache) so ultimately the increased risk of instability is just not worth the hassle for the meagre improved performance (which you probably won't even notice).
2015-07-15, 14:55
Thanks Milhouse and popcornmix! I look forward to testing 1080i advanced deinterlacing in the next build(s). The PI just keeps getting better and better!
2015-07-15, 22:46
New OpenELEC Isengard build #0715b: RPi / RPi2
(Supercedes previous build)
Based on tip of OpenELEC master (daf850a7, changelog) and tip of XBMC master (d8c14987, changelog) with the following modifications:
(Supercedes previous build)
Code:
# uname -a
Linux rpi512 4.1.2 #1 Thu Jul 16 02:06:49 BST 2015 armv6l GNU/Linux
# vcgencmd version
Jul 15 2015 18:46:28
Copyright (c) 2012 Broadcom
version f381e529aeea3dc67be93fa5001a08e01c1961af (clean) (release)
# lsb_release
OpenELEC (Milhouse) - Version: devel-20150716020532-#0715b-gdaf850a [Build #0715b]
# vcdbg log msg 2>&1 | grep DTOK
001562.647: Kernel trailer DTOK property says yes
# Kernel device tree status: Enabled
Based on tip of OpenELEC master (daf850a7, changelog) and tip of XBMC master (d8c14987, changelog) with the following modifications:
- Includes newclock4 patches
- Excludes the OpenELEC fernetmenta patches due to conflicts with newclock4
- Excludes the OpenELEC linux-01-RPi_support patch in favour of sourcing these and possibly more recent patches directly from kernel branch rpi-4.1.y
- Default setting for "Show RSS Feed" changed to disabled (new installs only) [patch details]
- Disabled "Total Duration" in Confluence (see build #0221 for details)
- Includes latest dcadec master (2a9186e3, ahead +34)
- Includes latest kodi-platform master (33b6390b, ahead +1)
- Includes latest libcec master (d997b9aa, ahead +28)
- Includes latest libnfs master (68941809, ahead +25)
- Includes latest platform master (22e0f42b)
- Includes latest addons: pvr.argustv (b9e2b289, +16), pvr.demo (40efb477, +7), pvr.dvblink (820a9383, +6), pvr.dvbviewer (1e1da026, +6), pvr.filmon (c4858f87, +6), pvr.hts (0cff617b, +18), pvr.iptvsimple (10626b4b, +6), pvr.mediaportal.tvserver (100d1dfe, +20), pvr.mythtv (db1d3cd0, +25), pvr.nextpvr (e4a2d794, +6), pvr.njoy (45c23808, +6), pvr.pctv (efa92a78, +6), pvr.stalker (506c16f3, +6), pvr.vbox (f7e02322, +18), pvr.vdr.vnsi (f761709b, +7), pvr.vuplus (20edf9d9, +8), pvr.wmc (3907c8bd, +8)
- Exclude ffmpeg-99.0011-mpeg4video-Signal-unsupported-GMC-with-more-than-one.patch: Ignore, as already in newclock4
- Exclude kodi-999.20-PR7102.patch: Temporarily ignore dcadec settings and use newclock4 version
- Exclude kodi-999.21-configure.ac-dont-break-if-dcadec-is-missing-dcadec-.patch: Temporarily ignore due to breakage
- Include patch: Add config options for Sense HAT Joystick (raspberrypi/linux/pull/1043)
- Include patch: Temp fix PVR crashes after PR7268
- Include patch: See https://github.com/raspberrypi/linux/pull/1037
- Include patch: Add ffmpeg dependency and includes for HEVC optimisations
- Include patch: Temp fix after PR7268 until next Kodi/OE sync
- Include patch: Enable Wolfson 8804 I2C module
- Include patch: Add experimental splash video
- Include patch: Enable pvr addons, disable PVR updates
- Include PR:4228: ffmpeg: update to ffmpeg-2.7.1
- Include PR:7462: [backport-resurrect] added: also hide thumbs for unwatched episodes option to show plot is off
- New firmware, with fixes for hardware deinterlace
- #0715b: Fix for PVR add-on crashes
- Firmware (Jul 15):
- firmware: di_adv: Fix for non-interlaced frames with qpu deinterlace
- firmware: di_adv: Updated version of qpu code
- firmware: di_adv: Bump up thread priority when doing qpu deinterlace See: link
- firmware: di_adv: Fix for non-interlaced frames with qpu deinterlace
- OpenELEC:
- XBMC:
- [guilib] fix dirty regions because it evaluates always to true (PR:7504, 1 commit, 1 file changed)
- fix building with --disable-webserver (PR:7510, 5 commits, 8 files changed)
- Clean "using namespace std" under xbmc/ (PR:7496, 10 commits, 75 files changed)
- [cosmetics] some cleanup on spelling/capitalisation & xbmc -> kodi (PR:7392, 1 commit, 5 files changed)
- [baserenderer] make ViewModeStretch16x9Nonlin fall in range (PR:7506, 1 commit, 1 file changed)
- [dx11] Fixed creating device on GPU whose driver not supports DX11 vi… (PR:7514, 1 commit, 1 file changed)
- Reworked initialization for win32 (PR:6809, 1 commit, 14 files changed)
- [xbmc][cleanup] Clean up ApplicationMessenger (PR:7268, 2 commits, 110 files changed)
- exclude resource:// from texture caching as these are always local ad… (PR:7519, 1 commit, 1 file changed)
- [PVR] fix timer deletion (PR:7521, 1 commit, 1 file changed)
- Revert "dummy commit" (PR:7523, 1 commit, 1 file changed)
- [win32] add system/ssleay32.dll to .gitignore (cd2207c8)
- [Re-Touched] update (5c2df624)
- [guilib] fix dirty regions because it evaluates always to true (PR:7504, 1 commit, 1 file changed)
- newclock4:
- New commits in this build:
- squash: hevc: Avoid holding qpu lock when not required. Add optimised strength calculation (a50052fc)
- squash: hevc: Avoid holding qpu lock when not required. Add optimised strength calculation (a50052fc)
- New commits in this build:
- kernel 4.1.y:
- Additional commits/pull requests/changes not yet merged upstream:
2015-07-15, 23:06
I get a boot loop with this latest build. Log at http://sprunge.us/LJGE
It looks like it crashes when it tries to show a PVR message on boot. I guess PR:7268 is the likely culprit?
With live tv disabled there is no boot loop.
It looks like it crashes when it tries to show a PVR message on boot. I guess PR:7268 is the likely culprit?
With live tv disabled there is no boot loop.
2015-07-15, 23:27
I wonder if the PVR addons need updating (you're already using the latest versions, but maybe they all now need a bump following the ApplicationMessenger change).
2015-07-15, 23:38
(2015-07-15, 23:06)Leopold Wrote: I get a boot loop with this latest build.
Ditto I will see if I can log tomorrow
2015-07-15, 23:44
(2015-07-15, 23:06)Leopold Wrote: I get a boot loop with this latest build. Log at http://sprunge.us/LJGEI am seeing similar boot loop after installing latest version, I also use Live TV through TVheadend backend (Version you provided Millhouse) here is crashlog using ssh
It looks like it crashes when it tries to show a PVR message on boot. I guess PR:7268 is the likely culprit?
With live tv disabled there is no boot loop.
paste $(ls -1art /storage/.kodi/temp/kodi_crashlog* | tail -1)
http://sprunge.us/OaRN
And general log if its helpful
http://sprunge.us/jCIe
2015-07-16, 00:32
Well crud, live tv is my biggest use case for advanced 1080i deinterlacing
Guess I'll wait until that is sorted out.
Guess I'll wait until that is sorted out.
2015-07-16, 03:52
The latest build has been updated to build #0715b which now includes a fix for the PVR add-on crashes - thanks all for the crash logs.
2015-07-16, 05:01
Excellent guys! Working great on my 2 RPi2 systems for live TV using the MythTV plugin. Thanks for the continued improvements!
2015-07-16, 06:35
Just installed 0715b my open vpn no longer works i tried downgrading to #0714 with dev update and that's not working either had to manually downgrade
2015-07-16, 06:44
(2015-07-16, 06:35)M32134c╘ Wrote: Just installed 0715b my open vpn no longer works i tried downgrading to #0714 with dev update and that's not working either had to manually downgrade
You'll need to debug this yourself, although it sounds like your network probably isn't working as it should be (and maybe it's just a coincidence it stopped working with build #0715b?)
Edit: Not sure if the PPTP change could break OpenVPN, I've asked in the commit but maybe other OpenVPN users will be able to confirm if they're also experiencing problems.