=== duflu_ is now known as duflu | ||
alan_g | alf__: I guess this is no linger needed? https://code.launchpad.net/~afrantzis/mir/always-consume-input-events-in-clients/+merge/209313 | 09:54 |
---|---|---|
alf__ | alan_g: affirmative | 09:54 |
hikiko | hello | 10:01 |
hikiko | I am trying to run mir on nvidia, I have loaded and installed the nouveau driver (checked with lsmod that there is no other nvidia driver) and when I run the mir server basic demo, I get a message that gbm failed to load the nouveau driver... (nvidia GF188 [GeForce GT 440]) | 10:03 |
hikiko | does anyone else have the problem? | 10:04 |
hikiko | is there anything I can do to fix it? | 10:04 |
RAOF | Hey! | 10:04 |
RAOF | I... think that card should be supported by nouveau? | 10:04 |
RAOF | Does X work properly? | 10:04 |
hikiko | yes RAOF | 10:05 |
hikiko | X is fine | 10:05 |
RAOF | And glxinfo under X shows that it's using nouveau? | 10:05 |
hikiko | let me check | 10:06 |
hikiko | well if I run: glxinfo | grep -i nouveau | 10:07 |
hikiko | I gett libGL error: failed to open drm device: Permission denied libGL error: failed to load driver nouveau | 10:08 |
hikiko | (i ran startx though so it might be just permissions let me check in unity7) | 10:09 |
hikiko | yep | 10:10 |
hikiko | when i start unity from lightdm | 10:10 |
hikiko | i get OpenGL vendor string: nouveau | 10:10 |
RAOF | Cool. Hm... | 10:11 |
duflu | Look over there! | 10:18 |
* duflu runs away | 10:18 | |
hikiko | :P | 10:20 |
hikiko | in /usr/lib/x86_64-linux-gnu/dri I have a nouveau_dri.so (and a nouveau_vieux_dri.so) if that helps | 10:23 |
hikiko | so it should find it | 10:23 |
=== alan_g is now known as alan_g|luch | ||
=== alan_g|luch is now known as alan_g|lunch | ||
=== jono is now known as Guest95937 | ||
=== alan_g|lunch is now known as alan_g | ||
=== dandrader is now known as dandrader|afk | ||
=== dandrader|afk is now known as dandrader | ||
alan_g | alf__: @fix-1260612 - it is an improvement. Should I approve and log a bug for the not-working scenario? Or do you have a clue how to fix? | 15:43 |
alf__ | alan_g: No clue, but I will keep investigating. Feel free to log a different bug if you like and assign it to me. | 15:47 |
alf__ | alan_g: hmm, you are running shell as host basic as nested, I am running basic as host shell as nested. | 15:49 |
alan_g | alf__: I've tried your scenario with "_shell" and "_basic" as the nested server. The latter shows the problem I saw. Not sure why there's a difference though. | 15:51 |
alan_g | I'll approve the MP - it seems to be a "_basic" bug | 15:52 |
alf__ | alan_g: well, basic doesn't support resizing or moving gestures, in the shell==host scenario we are moving/resizing the whole nested window, and I am not sure if resizing means scaling or cropping in this case | 15:56 |
alan_g | alf__: I think some /examples gardening is required - who can remember these differences. But explained to my satisfaction | 15:59 |
=== alan_g is now known as alan_g|tea | ||
alf__ | alan_g|tea: so, yes, we seem to be cropping in the shell==host scenario (if you "zoom in" hard enough you will see it) | 16:01 |
=== alan_g|tea is now known as alan_g | ||
kgunn | AlbertA: btw, i landed the usc change to have fuzzy time on screen blank...but i still see the app on unblank | 16:36 |
kgunn | will our plan to move all the screen state in usc better address this ? | 16:36 |
kgunn | AlbertA: what's weird is i know i saw it work once (e.g. saw only the lock screen) | 16:36 |
AlbertA | kgunn: well you need to add the delay to a configuration file | 16:37 |
AlbertA | kgunn: by default its 0 | 16:37 |
AlbertA | If you add /etc/xdg/unity-system-compositor.conf | 16:37 |
AlbertA | with a power-off-delay=500 maybe | 16:38 |
AlbertA | kgunn: but yes moving the screen state out of powerd will help here | 16:39 |
AlbertA | kgunn: also for the power-off-delay to be "seen" | 16:40 |
AlbertA | kgunn: you need this in powerd: https://code.launchpad.net/~albaguirre/powerd/let-display-turn-off-backlight/+merge/208667 | 16:41 |
=== tvoss is now known as tvoss|test | ||
=== tvoss|test is now known as tvoss | ||
=== dandrader is now known as dandrader|lunch | ||
mterry | alan_g, btw, I finally got a chance to test your nested-sessions-dont-post-buffers-until-something-happens branch (had to wait for 0.1.6 to settle first, was getting weird crashes before). It works like a charm, thanks! | 17:31 |
alan_g | mterry: \o/ | 17:32 |
ogra_ | hmm, so after using mirscreencast it seems that the screen stopped auto-suspending | 17:46 |
AlbertA | ogra_: hmm, interesting... I don't see why that would be the case from a mir perspective | 17:49 |
ogra_ | the screen was initially suspended ... running miscreencast properly woke it up too ... it just doesnt turn off again until i press the power button twice | 17:50 |
AlbertA | ogra_: there must be some resource we are holding on the screencast compositor | 17:50 |
ogra_ | well, is that solely screencast or do you perhaps invoke powerd somewhere to get the display up | 17:51 |
AlbertA | ogra_: there must be a display on - I think that was fixed recently - but not through powerd | 17:53 |
ogra_ | hmm, i could imagine that both processes start wrangling over owning the screen then or that powerd simply isnt aware the screen is on if you force it on from Mir | 17:54 |
ogra_ | (so it doesnt try to suspend it then) | 17:54 |
AlbertA | ogra_: Oh I see, you started mirscreencast while the screen was off, then yeah, powerd wouldn't know about that | 18:05 |
ogra_ | AlbertA, yeah, not a biggie, but it would be good if they could communicate to each other about this | 18:06 |
AlbertA | ogra_: In any case, we are ripping the screen power state out of powerd and putting it in usc | 18:06 |
ogra_ | ah, k | 18:06 |
ogra_ | thats will be fine then | 18:06 |
AlbertA | ogra_: Also the forcing of display on will be removed if this lands: | 18:06 |
AlbertA | https://code.launchpad.net/~robertcarr/mir/remove-ensure-display-powered/+merge/209734 | 18:06 |
ogra_ | hmm, tests need this function | 18:07 |
ogra_ | you should talk to the QA guys about that, autopilot needs a way to force the screen on before starting its tests | 18:07 |
AlbertA | racarr: ^ | 18:09 |
racarr_ | ogra_: AlbertA: I think my branch is unrelated | 18:26 |
racarr_ | you guys need to be able to to force screen on externally | 18:27 |
racarr_ | like powerd-ctl or hatever it was called | 18:27 |
racarr_ | does | 18:27 |
racarr_ | which should still work for now? but maybe they got changed out | 18:27 |
AlbertA | racarr_: well I think it's kinda related in that right now the screencast turns the screen on | 18:27 |
racarr_ | my branch is just about, clients which have explicitly turned the screen off | 18:28 |
AlbertA | racarr_: I believe your MP will avoid that | 18:28 |
racarr_ | there was a behavior that automatically turns it back on when they swap buffers | 18:28 |
racarr_ | hmm | 18:28 |
racarr_ | maybe! | 18:28 |
AlbertA | racarr_: which is good | 18:28 |
racarr_ | mm | 18:29 |
=== dandrader|lunch is now known as dandrader | ||
=== bschaefer_ is now known as bschaefer | ||
=== dandrader|afk is now known as dandrader | ||
=== seb128_ is now known as seb128 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!