(2018-01-26, 14:35)popcornmix Wrote: [ -> ]Does the leak seem to be once per play? i.e. if you play/stop after a few seconds repeatedly does memory exhaust much faster than playing a single file that lasts an hour? (I'd guess it probably would)
I tested #0805 again, playing/stopping Elysium after about 5 seconds, and the leak seems to be quite small - 11 plays and the RAM deficit is only 48MB (compared with 120MB+ after 2 complete plays), so the leak seems to occur during/throughout playback, possibly the 1-2MB/s allocations are not being released at the end of playback?
(2018-01-26, 14:35)popcornmix Wrote: [ -> ]How hard would it be to build head of tree with older gcc/glibc? Does that crash with out-of-memory after repeated plays?
(I suspect that #0805 is really the culprit, but the evidence is currently a bit unclear).
BTW John is on holiday this week so we won't get an immediate fix, but would be nice to be sure the bug is in the hevc code when he returns.
I'm currently building latest LE master/Kodi master with gcc-6.2.0/glibc 2.25 (rather than gcc-7.1.0/glibc-2.26) however a lot of other packages were also updated to work with gcc-7.x so rolling back just gcc/glibc might not be a success (and several of the updates packages have received subsequent updated so rolling back the entire PR is tricky). Anyway... I'll let you know.
(2018-01-26, 18:51)ksooo Wrote: [ -> ]Applied the PR. Does not make a difference. Still working. Problem with custom keymaps? Just guessing.
Hmmm, no custom keymap here. Also tested with keyboard by pressing ENTER, everythings fine.
With Harmony remote no reaction.
(2018-01-26, 19:36)debutanker Wrote: [ -> ] (2018-01-26, 18:51)ksooo Wrote: [ -> ]Applied the PR. Does not make a difference. Still working. Problem with custom keymaps? Just guessing.
Hmmm, no custom keymap here. Also tested with keyboard by pressing ENTER, everythings fine.
With Harmony remote no reaction.
We're getting closer! Keyboard works, remote not! I tested with keyboard only.
(2018-01-26, 19:38)ksooo Wrote: [ -> ]We're getting closer! Keyboard works, remote not! I tested with keyboard only.
Does this mean you've been able to reproduce with a non-LibreELEC builds?
I've just uploaded build #0125b:
RPi2
This build does not include:
(2018-01-26, 02:19)Milhouse Wrote: [ -> ]Build Details:
- Additional commits/pull requests/changes not yet merged upstream:
(2018-01-26, 19:42)Milhouse Wrote: [ -> ] (2018-01-26, 19:38)ksooo Wrote: [ -> ]We're getting closer! Keyboard works, remote not! I tested with keyboard only.
Does this mean you've been able to reproduce with a non-LibreELEC builds?
No.
@
debutanker wrote that keyboard works, but Harmony not. I don't own a Harmony.
(2018-01-26, 19:55)debutanker Wrote: [ -> ]Yes, but as i already mentioned in a few posts before ( https://forum.kodi.tv/showthread.php?tid...pid2696356 ), those behavior has been tested with NUC and GENERIC Build #0125.
But as it seems, GENERIC and Pi shows pretty much the same behavior.
The point is that it was not clear to me so far that it works with keyboard and not with Harmony remote. Can you confirm this?
Yes, can confirm that keyboard is working without problems and Harmony not.
I assume the OK remote button is working elsewhere in the GUI, it's just not working in the Live TV? That might rule out the kernel change. I'll upload another build based on the older 4.14.14 kernel with all other changes the same as #0125.
Haven't tested with keyboard, only with remote, but with 0125b it isn't working too only in LiveTV
while playing a video the OK-button is working
I've just tested my Harmony One remote with RPi2 (MCE USB dongle) and Intel NUC (built-in IR) and the OK button (and all other buttons) are working normally with #0125. However I don't have a LiveTV setup to test so can't help test that.
Could reproduce this here, OK on remote is working fine and also brings up OSD on normal video playback. Only livetv is affected.
During livetv OK button hits kodi fine, but action is Select instead of OSD:
Code:
18:40:52.007 T:1941894144 DEBUG: LIRC: Update - NEW at 3238985:160 0 KEY_OK devinput (KEY_OK)
18:40:52.008 T:1941894144 DEBUG: HandleKey: 11 (0x0b, obc244) pressed, action is Select
With Enter on keyboard it looks fine (action OSD):
Code:
18:41:00.618 T:1941894144 DEBUG: Keyboard: scancode: 0x1c, sym: 0x000d, unicode: 0x0000, modifier: 0x0
18:41:00.619 T:1941894144 DEBUG: HandleKey: return (0xf00d) pressed, action is OSD
18:41:00.619 T:1941894144 DEBUG: ------ Window Init (VideoOSD.xml) ------
With 0124 OK button looks fine as well, action is OSD:
Code:
18:49:14.599 T:1942049792 DEBUG: LIRC: Update - NEW at 89509:160 0 KEY_OK devinput (KEY_OK)
18:49:14.599 T:1942049792 DEBUG: HandleKey: 11 (0x0b, obc244) pressed, action is OSD
18:49:14.599 T:1942049792 DEBUG: ------ Window Init (VideoOSD.xml) ------
so long,
Hias
Those with the OSD issue, please try build #0125c:
RPi2
This is the same as #0125 but with the 4.14.14 kernel instead of 4.14.15.
(2018-01-26, 20:53)HiassofT Wrote: [ -> ]During livetv OK button hits kodi fine, but action is Select instead of OSD:
I'll try reverting:
(2018-01-26, 02:19)Milhouse Wrote: [ -> ]
- XBMC:
- [gui] cleanups in WindowManager especially for getting the active window and dialog id (PR:13433, 5 commits, 23 files changed)
to see if that is responsible... after that I'm out of options/guesses.