[00:00] but on 3.19.0-12 it doesn't work at all [00:00] so, still a difference === tmpRAOF is now known as RAOF === tmpRAOF is now known as RAOF [14:31] furkan: can't reproduce the checkerboarding bug or the corruption. the dpms issue is real, testing a commit now [17:17] tjaalton: this is interesting... i unplugged my secondary monitor, and everything works much, much better [17:18] is the other monitor a 4k one? [17:19] no checkerboarding, no "snow" in the virtual terminal, no delay when switching between virtual terminals (with 2 monitors, it takes a few seconds), and no graphics corruption when initially logging in (i didn't mention that in the bug report because i thought that was "normal", i'll take a vid for you if you want) [17:19] i have just my 27" 2560x1440 monitor connected now [17:20] and the other one? [17:20] i unplugged it.. i'll plug it in and try it in landscape mode first [17:20] before switching to portrait, which is how i originally had it [17:20] the resolution? [17:20] oh right [17:20] * furkan tests [17:24] tried rotating the other one, still nothing [17:24] nope... killed my desktop [17:24] i took a pic [17:24] https://www.dropbox.com/s/fd1tdi67og8iz91/IMG_20150408_132316.jpg?dl=0 [17:24] that's what happened when i tried to change from 2560x1440 to 1920x1080 [17:25] so you have a dual-monitor setup too? [17:25] what are your resolutions? [17:25] 25x14, 19x10 [17:25] oh, so almost the same [17:26] hooked to the same monitor via dvi & hdmi [17:26] and you tried rotating the 19x10 monitor? [17:26] yes [17:26] wait, hooked to the same monitor? you mean 1 screen connected with 2 cables? [17:26] yup [17:26] i only have one [17:26] hmm [17:27] ok i will try connecting it again in landscape [17:33] ok so i have the monitor in landscape now [17:33] switching between vts is no longer instantaneous [17:33] but no checkerboarding/snow [17:33] after suspend/resume [17:34] changing resolution still trips it up, i need to switch into a VT and service lightdm restart [17:36] so i made the switch to portrait, right now i have checkerboarding, but no "snow" [17:37] even after 2 suspend/resume cycles (1 when i had the monitor in landscape, and 1 after i switched to portrait) [17:38] i bet it'll be back after i restart though [17:38] so which one is in portrait? [17:43] yeah so i restarted, at first no snow, then suspend/resume and the snow is back... so it seems to reproduce that bug i need to boot directly into that display configuration [17:43] i have 27" 2560x1440 in landscape mode [17:43] and 24" 1920x1200 in portrait mode, on the left-hand side [17:43] i suppose i could also try putting it on the right-hand side as a final test [17:44] interesting, this time my screen didn't become garbled after changing resolution [17:45] and VT switches are instantaneous [17:49] ok well after restarting, no difference [17:50] but at that moment, just after switching the screen from one side to the other, i was getting instantaneous VT switches, now i'm not [17:50] but snow/checkerboarding still there [17:50] have you tried 4.0? [17:51] https://www.dropbox.com/s/4koy953b7faj095/VID_20150408_134129.mp4?dl=0 [17:51] that's the corruption when initially logging in.. pretty minor so i hadn't thought much of it before, but it doesn't happen with a single monitor connected [17:51] and yeah i was using 4.0 for a while, and same issues [17:52] try #radeon then, and file a bug on bugs.freedesktop.org [17:53] but i didn't have the issue with Xorg 1.16, so what if it's an Xorg bug and not a kernel bug? [17:53] and when i downgrade to 3.16 i still have the issue [17:53] whereas before upgrading to 15.04, i didn't [17:53] so i think that rules out the kernel, no? [17:53] same devs [17:53] ah [17:53] pretty much [17:54] ok will do, then [17:54] thanks for the pointer [17:56] btw, another regression, unrelated to X: after rebooting or suspend/resume, i need to unplug my headphone and plug it back in for it to be recognized again [17:57] if i submit too many bug reports i feel like i'll become the "boy who cried wolf", but right about now i'm wishing that i stayed with 14.04 :P [18:00] well, can't reproduce that either, something wrong with jack detection I guess [18:00] file that against pulseaudio [18:00] for starters [19:54] tjaalton: i guess i understand the "snow" thing now [19:55] when on a VT, it sets the same resolution on both monitors [19:55] so since my smaller monitor is 1920x1200, it sets that resolution on both [19:56] and before suspending, it's scaled to full screen [19:56] after suspend/resume, you get the 1920x1200 terminal drawn on the top left corner of the screen, and the blank areas are white [19:58] in 4.0, the terminal isn't scaled to full screen but the blank areas are black (then after suspend/resume it turns white) [20:02] really not the worst out of the bugs i've seen so far, but something is really broken w/ dual monitor support on a radeon card [20:19] furkan: yep, saw the same, no corruption though, just white borders [21:00] tjaalton: i know this must be an upstream bug, but just to share the hilarity https://www.dropbox.com/s/caawhzn8zfycc00/VID_20150408_165520.mp4?dl=0 [21:01] the time difference between switching, where the secondary monitor is on the left vs. the right [21:01] *switching to a VT [21:03] it is what it is. [21:04] well with the open-source drivers w/ kernel modesetting, it's supposed to be instantaneous [21:04] whereas with the proprietary drivers it's not [21:04] so while the bug itself is no big deal, i think it's indicative that something is going wrong [22:37] tjaalton: do you use vanilla Ubuntu, or one of the derivatives?