| dansuf | kgunn: Hi, I discovered something interesting. When I launch mir server demo shell and then "turn off" the display by clicking power button then launch some of the client demos it doesn't have this genlock error | 20:10 |
|---|---|---|
| dansuf | kgunn: I forgot to say that it normally renders everything after turning on the display | 20:11 |
| kgunn | dansuf: that's interesting....i wouldn't have thot that... | 20:11 |
| dansuf | I still don't know how to turn off the display on unity8 | 20:11 |
| kgunn | so you turn the display off with pressing the pwr button, but then it "comes back on" upon mir demo rendering ? | 20:11 |
| dansuf | no, i turn it off then launch demo then turn on | 20:12 |
| dansuf | and it also worked when i turned off then on when an app is running but I'm not sure it always works | 20:12 |
| kgunn | dansuf: are you looking for a way to control the display ?...if you are...then check this out | 20:12 |
| kgunn | https://wiki.ubuntu.com/powerd | 20:12 |
| kgunn | you can simply (from another terminal) get on the device shell (adb shell)...and then run those cli commands | 20:13 |
| kgunn | dansuf: i would say that the display interaction is really kind of undefined | 20:14 |
| kgunn | we're controlling display state with unity-system-compositor | 20:14 |
| kgunn | well... | 20:15 |
| kgunn | maybe it is that mir is at least turning on in the render case...so not completely undefined | 20:15 |
| dansuf | kgunn: I'm trying to figure out how to turn off the display with these commands. On the other hand pressing pwr in demo just makes the screen blank, not powered off completely and with genlock error only first frame is shown with 1 FPS | 20:20 |
| kgunn | dansuf: sorry, what is your goal of "turning off" the display ? | 20:28 |
| dansuf | kgunn: with mir launched (which is giving me genlock error) I wanted to turn the creen off then on with hope it will magically work | 20:29 |
| dansuf | kgunn: running unity8 with --offscreen seems to run fine | 20:34 |
| dansuf | so it's definitely sth with the display | 20:35 |
| kgunn | dansuf: problem is, if the display is off then you won't have a vsync...which you need to "consume buffers", and pump the rendering | 20:36 |
| kgunn | unless you configure mir to ignore vsycn (swapinterval 0) | 20:37 |
| dansuf | kgunn: I'm afraid my hardware drivers doesn't support vsync | 20:38 |
| kgunn | dansuf: even if its not a "real" vsynch...it must be providing some sort of heartbeat or swapping buffer scenario | 20:39 |
| kgunn | e.g. even if your display is "command mode" | 20:39 |
| dansuf | kgunn: So how can I configure mir to set swapinterval to 0? | 20:41 |
| kgunn | dansuf: which i guess is _exactly_ what you're trying to tell me :)....you get a genlock error...but the genlock is exactly the thing mimicking vsync | 20:42 |
| kgunn | actually...right after i typed swapinterval 0...i'm thinking we don't have that functinoality in our mobile(android) platform...only mesa | 20:42 |
| kgunn | dansuf: plus, i guess the real answer is to determine what might be the issue with genlock to begin | 20:43 |
| kgunn | dansuf: https://github.com/lgics/cm_device_lge_p350/issues/2 | 20:46 |
| kgunn | might check the post from RonGokhale about a year ago | 20:46 |
| kgunn | only a guess...but kind of sounds like what you're experiencing | 20:47 |
| dansuf | kgunn: my error is different if it does count | 20:52 |
| dansuf | E/qdgenlock( 1612): perform_lock_unlock_operation: GENLOCK_IOC_DREADLOCK failed (lockType0x1,err=Connection timed out fd=73) | 20:52 |
| dansuf | if not, I can try applying the patches | 20:52 |
| dansuf | kgunn: I will leave the chat soon, thank you for your help, I'll try to find the problem either in genlock or in mir or maybe do some hacks in mir | 20:56 |
| kgunn | ack | 20:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!