=== JanC is now known as Guest31516 === JanC_ is now known as JanC === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === dandrader is now known as dandrader|afk === chihchun is now known as chihchun_afk === dandrader|afk is now known as dandrader === chihchun_afk is now known as chihchun === dandrader is now known as dandrader|afk [14:46] tjaalton, Xmir branch at https://git.launchpad.net/~xmir-team/xorg-server/+git/xmir/?h=xmir-1.19 now builds and tests cleanly on Zesty if you can find the time to integrate into xorg-server [14:56] bregma: sweet, thanks! will do that next week when I'm back === dandrader|afk is now known as dandrader [15:16] alf_: On https://bugs.launchpad.net/mir/+bug/1667001 I think repowerd is innocent, suspect Mir/kms [15:16] Ubuntu bug 1667001 in unity8 (Ubuntu) "[amd] screen-unblanking seems to flicker backlight on/off 5 times before display resumes" [Undecided,Incomplete] [15:17] I've attached the log you requested [15:17] greyback: thanks [15:18] alf_: the machine is by me if there's anything else I can offer you [15:20] greyback: if you are willing to recompile repowerd we may get some more hints about what's going wrong [15:20] alf_: Sure. I know it's not a high priority bug though, so only if you want [15:21] greyback: sure, let's do it now that it's hot :) [15:21] greyback: git clone lp:repowerd (assuming you have set the lp: shortcut for git) [15:21] I have [15:23] greyback: build repowerd the usual cmake way (no special options needed) [15:23] greyback: get the deps first of course [15:25] greyback: ... "sudo systemctl stop repowerd" to stop the system one, and from the build dir run the git version with 'sudo REPOWERD_LOG=console bin/repowerd' [15:25] greyback: make sure it works and you can see the same issue as before [15:25] ok [15:28] alf_: my build machine != my test machine, can I just copy over the repowerd binary, or do I need the tools too? [15:29] greyback: not yet, perhaps later depending on the results of our experiments [15:32] alf_: repowerd binary is 10mb! [15:42] alf_: http://pastebin.ubuntu.com/24053613/ <- any idea? [15:45] greyback: do you have repowerd-2017.02 installed on that system? [15:45] alf_: no 2016.12+17.04.20161212.1-0ubuntu1 [15:46] greyback: ok, then update (and stop the system repowerd again) [15:46] ok [15:55] alf_: http://pastebin.ubuntu.com/24053671/ [15:55] same issue with the newer repowerd [15:56] greyback: good [15:57] greyback: one thing to change to see if it makes a difference: make the turn_on() call to USC synchronous [15:57] greyback: so in src/adapters/unity_display.cpp [15:58] alan_g, how do I get the top-left position of a miral::WindowInfo? [15:58] dandrader: wi.top_left() [15:58] alf_: done, trying [15:59] greyback: in ::turn_on() change the dbus call to _sync and remove one "nullptr," [15:59] alf_: already done ;) [15:59] greyback: ack [15:59] I am still surprised how big the binary is [16:00] alan_g, ah, you mean windowIndo.window().top_left() [16:00] *windowInfo [16:00] dandrader: oh, right. yes [16:00] * alan_g isn't a good autocomplete function [16:00] alf_: yes it seems to block on that TurnOn dbus call, as the screen flickers [16:02] greyback: It's all the debug info I guess? The released binary is much smaller (~700kb) [16:02] alf_: yeah probably === dandrader is now known as dandrader|afk [16:03] greyback: ok, so this strongly indicates that repowerd is not doing anything strange, but let's experiment a bit more [16:04] c++ debug symbols are BUG [16:04] greyback: with the screen on stop repowerd (ctrl-c) [16:05] greyback: i.e. I want the backlight to stay on [16:06] yep [16:07] greyback: now just turn USC rendering on/off manually to see what behavior you get: sudo gdbus call --system --dest com.canonical.Unity.Display --object-path /com/canonical/Unity/Display --method com.canonical.Unity.Display.TurnOff "all" [16:08] alf_: I've actually gone further. A while ago I made this tool: https://code.launchpad.net/~gerboland/+git/display-configurator [16:08] alf_: just extending the mir display config tool, adding display power on/off stuff [16:09] alf_: if I ask mir to turn off the display, and then on, the on causes the same flicker [16:09] greyback: ok, just wanted to test if the backlight state affects this at all [16:09] greyback: it seems not [16:09] alf_: ack [16:10] greyback: ok, so repowerd is out... anything interesting in /var/log/lightdm/unity-system-compositor.log [16:12] greyback: or in /var/log/syslog for that matter? Any drm kernel messages? [16:20] alf_: I tore it down to the simplest elements: mir_demo_server_basic and my test client. From Mir: http://pastebin.ubuntu.com/24053776/ [16:20] there's a small complaint from kms about invalid argument for Gamma setting [16:20] but I see nothing at all relevant in syslog or dmesg [16:22] actually I get the flicker when I turn off the display too [16:24] and setting mir power mode to 1 (on) does not enable the display again [16:24] I have to kill the client to get the display to turn back on [16:29] greyback: I just submitted a QtMir MP that ran into LTTng errors that seem unrelated. Is there a known problem? [16:29] alan_g: news to me [16:29] dandrader|afk: ^^ [16:31] Hmm, seems to be just my MP. [16:32] * alan_g can't ignore it [16:38] greyback: I am out of easy-to-test ideas. I guess next step would be to check if something wrong is going on at the Mir platform/backend level, and if everything is fine there, too, then it's probably a driver or hardware issue. [16:41] alf_: ack. I'm updating the bug with all I've found, but I expect it's driver/hardware [17:12] greyback: not sure what the cause is, but after "apt update" it is happening locally on a clean lp:qtmir. :( [17:13] * greyback trying === chihchun is now known as chihchun_afk === dandrader|afk is now known as dandrader