Tuesday, 2014-07-01

*** situ has quit IRC01:50
*** situ has joined #sailfishos-porters03:26
*** vgrade_ has quit IRC04:11
*** sledges has quit IRC04:11
*** Aquilum has quit IRC04:11
*** netchip has quit IRC04:11
*** Stskeeps has quit IRC04:11
*** MSameer has quit IRC04:11
*** MSameer has joined #sailfishos-porters04:22
*** vgrade_ has joined #sailfishos-porters04:22
*** sledges has joined #sailfishos-porters04:22
*** Aquilum has joined #sailfishos-porters04:22
*** netchip has joined #sailfishos-porters04:22
*** Stskeeps has joined #sailfishos-porters04:22
*** slate has quit IRC06:51
*** slate has joined #sailfishos-porters06:53
vgrade_situ: are we picking the right plugin in the sucks ks07:05
*** Sazpaimon__ has joined #sailfishos-porters07:18
*** Sazpaimon_ has quit IRC07:21
*** slate has quit IRC07:37
*** slate has joined #sailfishos-porters07:54
*** slate has quit IRC08:11
*** slate has joined #sailfishos-porters08:15
morphisvgrade_, situ: so you get things rendered on the screen with hwcomposer 1.3, right?08:35
situvgrade_: No, i had to downgrade qt5-qpa-hwcomposer-plugin08:44
situmorphis: Yes, lipstick comes up fine everytime. With the fauly commit, lipstick was crashing at leat 4 out of 5 times.08:46
morphishm, I have troubles getting a similar stack up on a HP touchpad08:51
situmorphis: But we see rendering related issues without that commit.09:00
morphissitu: what kind of rendering issues?09:06
situmorphis: After 30 seconds or so of using UI, screen doesn't refresh properly and you will see tearing.09:08
*** vgrade has joined #sailfishos-porters09:23
vgradesitu: does lipstick crash straight away?  I did not seen any crashes on Saturday when I was testing u7 image from out ks09:25
situvgrade: Yes, it does crash.09:27
vgradesitu: does it show anything before crashing09:29
situvgrade: Nope.09:29
vgradeonce I'm home we can compare package version09:29
situvgrade: did you try restarting lipstick several times ?10:18
*** phdeswer has quit IRC10:50
vgradeyes but I'll try again tonight10:51
situHere some errors from logcat http://pastebin.com/ppMbq6js (may be related to vsync issue)11:03
morphissitu: that looks very familar too me11:05
morphissitu: if you're interested: http://pastebin.com/NXSAsFQT are the changes I have for the hwcomposer QPA so far11:07
morphisit's on top of master11:11
situmorphis: did you see any vsync issues ?11:16
morphissitu: it doesn't even render something11:25
morphissitu: the log output looks similar but is not the same11:34
morphissitu: http://bpaste.net/show/Fz2LSRFcAOoQS8wvPSdh/12:05
*** situ has quit IRC12:08
*** situ has joined #sailfishos-porters12:11
situmorphis: May be we can look into this https://android.googlesource.com/platform/frameworks/native/+/android-4.4.4_r1/services/surfaceflinger/DisplayHardware/HWComposer.cpp ?13:15
morphissitu: I am already doing :)13:16
morphisI am pretty deep in hardware/qcom/display to find out what is wrong13:16
morphissitu: I fear my problem is called QCOM_BSP13:27
morphisit does ... sizeof(hwc_layer_1_t) = 112 vs sizeof(hwc_layer_1_t) = 9613:30
situvgrade: E/qdhwcomposer( 1120): hwc_sync: ioctl MSMFB_BUFFER_SYNC failed, err=Operation not permitted13:33
situLooks like another permission related issue13:33
situmorphis: I believe my problems are related to permissions.13:34
morphissitu: you can check the kernel path for MSMFB_BUFFER_SYNC and look what and why returns EPERM13:38
morphislook in drivers/video/msm and grap for MSMFB_BUFFER_SYNC13:38
morphisah, that looks much better: http://bpaste.net/show/4h9Xg8hJOutyHxTpxcHT/13:40
situmorphis: Is there a correct way to disable virtual display altogether ?13:42
morphisdon't know13:45
situopen("/dev/graphics/fb2", O_RDWR|O_LARGEFILE) = -1 ENOTSUPP (Unknown error 524)13:48
situshouldn't happen13:49
situLater it's trying to do ioctl on -1 (and I guess it shows a permission error in logcat output).13:53
situioctl(-1, 0xc1546d87, 0xab1bb44c) = -1 EBADF (Bad file descriptor)13:53
morphisis there a /dev/graphics/fb2 node?13:56
situmorphis: Yes13:56
situI don't know what could cause ENOTSUPP13:56
morphissitu: I get "Failed to call ioctl MSMFB_OVERLAY_SET err=Operation not permitted" when I start the compositor a second time13:59
morphiswhen starting it the first time this never happens13:59
morphishowever I only get one frame on the screen13:59
situmorphis: did you try to strace lipstick ?13:59
morphissitu: I am not using lipstick13:59
morphisbut a stack which is pretty similar14:00
situok, whichever compositor you are using.14:00
*** phdeswer has joined #sailfishos-porters14:09
morphissitu: https://github.com/webOS-ports/luna-next14:20
MSameerI recall there is/was a test composer app in hybris?15:10
situMSameer: Which never works. :P15:11
MSameeri recall i saw it working :)15:25
*** slate has quit IRC16:19
*** slate has joined #sailfishos-porters16:37
*** vgrade has quit IRC16:58

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!