Wizzup | dsc: arno: this is known until we merge tp module | 04:18 |
---|---|---|
Wizzup | freemangordon: I'm getting more music and videos in OMP now :) | 05:43 |
tmlind | got the mz617 usb modem flakeyness sorted out, will try to push out updated branch over the weekend. will take a look at enabling mz609 lcd next, then i guess it's time to rebase the n_gsm pathes for gps :) hmm maybe uvos already rebased the n_gsm stuff to v6.6-rc series? | 06:47 |
tmlind | i only really care about wlan + lcd + touch + gps on these devices to use with maps with m-l fyi.. most of the other devices should be similar to d4 though | 06:49 |
tmlind | presumably the gps is still on the modem | 06:50 |
Wizzup | tmlind: this is super great | 06:52 |
Wizzup | I am at openfest not this weekend but next weekend and want to see if I can show off mz617 with m-l | 06:52 |
Wizzup | I think uvos tried to rebase but he said something was missing | 06:52 |
Wizzup | I don't remember what | 06:52 |
tmlind | ok, mz609 has different backlight and touchscreen so no idea how long that might take | 06:54 |
tmlind | with mz617 i noticed that i could just mkfs ext4 on the big vfat partition, then reboot to android and it gets still mounted an reinitialized with the default directories | 06:55 |
tmlind | so that can be used both by anrdoid and m-l if anybody cares about the stock android support still | 06:55 |
tmlind | it's not mkfs.ext4 on the android root though, i think it's mke2fs where ext4 needs to be specified to reformat (and lose all the user data naturally) | 06:56 |
Wizzup | yes, that makes sense | 06:57 |
tmlind | hmm i think it was make_ext4fs on the stock android partition i used to reformat userdata partition mmcblk1p20 | 07:00 |
Wizzup | freemangordon: I'm thinking of quickly adding support for the menu arrow in qt5, I think we can just decide whether to render it or not in QXcbWindow::show ? | 07:11 |
Wizzup | there is an old ifdef 0'd version in qmenu_bar.cpp that just needs to be adapter for xcb mostly | 07:11 |
Wizzup | adapted | 07:12 |
Wizzup | I don't know if the xgetwmprotocols hackery is really necessary | 07:12 |
Wizzup | It might just need to set the property | 07:13 |
Wizzup | _HILDON_WM_WINDOW_MENU_INDICATOR | 07:13 |
Wizzup | I don't know what we need the _NET_WM_CONTEXT_CUSTOM for | 07:13 |
Wizzup | freemangordon: like this https://github.com/maemo-leste/qt-platform-maemo/commit/bceb0413873b15c87f06deca1fafe0b769d0eb2a | 07:20 |
freemangordon | Wizzup: yes, now the tracker does not crash... | 07:27 |
freemangordon | sorry, just woke up, need some coffee first | 07:28 |
uvos__ | tmlind: i did not rebase the gps or modem patches, i rebased all of my/leste stuff, noticed the broken omap dai and gave up | 09:52 |
uvos__ | iic omap-dai should work again now so ill continue when i have the time | 09:52 |
uvos__ | tmlind: ^^^ | 09:53 |
tmlind | uvos__: ok i'll take a look at rebasing the modem patches, i think audio might need that wip mcbsp patch, maybe it's only needed with some configurations though | 09:55 |
tmlind | uvos__: this one, not sure what caused that to start happening https://lore.kernel.org/linux-omap/20231018082142.5b7d3ad5@aktux/T/#m64db662085f60290447b13f09c96cdb38cd4f4fd | 09:56 |
uvos__ | tmlind: ok great yeah only the modem is still needed | 10:00 |
uvos__ | everything else is rebased | 10:00 |
tmlind | nice | 10:00 |
tmlind | did you rebase the pvr stuff too? | 10:00 |
tmlind | it will be interesting to see if the ants are still there on mz619 as it's a video mode panel.. | 10:01 |
* tmlind needs some lunch bbl | 10:03 | |
uvos__ | tmlind: yeah i have pvr | 10:08 |
uvos__ | tmlind: there was just some minor drm api changes | 10:08 |
Wizzup | uvos__: tmlind: ok, good to hear | 10:14 |
Wizzup | I have the mz619 on my desk here, brought it to asia :) | 10:14 |
tmlind | ok great less stuff for me to rebase | 10:48 |
uvos__ | i was pretty suprised how pvr rebased with almost no fuss | 14:25 |
uvos__ | and worked too | 14:26 |
uvos__ | the pvr code is pretty scary | 14:26 |
uvos__ | btw we should probubly update our usermode binaries and move the omap3/4 onto mesa-amber | 14:27 |
uvos__ | atm we cant futher update mesa (for the pp for instance) | 14:27 |
uvos__ | since any further mesa updates would permanenly break pvr | 14:27 |
dsc_ | walking antssss | 14:29 |
uvos__ | since the ants are in textures its more like ants in amber :P | 14:29 |
uvos__ | the ants situation is no different in 6,6 unfortinatly | 14:30 |
Wizzup | uvos__: please lmk if you have a kernel for the mz617, would love to try | 16:12 |
uvos__ | Wizzup: https://github.com/IMbackK/droid4-linux/tree/maemo-6.6 | 16:23 |
uvos__ | i mean this should boot all the way to accelerated hildon on m617, but i have not tried it on the tablet besides checking if it boots to console | 16:23 |
uvos__ | on d4 it works fine, besides the modem and sound | 16:24 |
freemangordon | new definition of 'boots fine' :p | 16:24 |
freemangordon | sorry, "works fine" | 16:24 |
freemangordon | BTW, it seems omapdrm driver does not support color keying and alpha blending :( | 16:25 |
freemangordon | or rather, the code is in the driver, but userspace cannot reach it | 16:26 |
freemangordon | omapfb supports it, but not omapdrm | 16:27 |
uvos__ | why would we want color keying? we dont need accelerated overlays in a compositing window manager | 16:28 |
freemangordon | so we want compositing to happen during fullscreen video playback? | 16:30 |
freemangordon | like, 'pause', 'play' etc buttons? | 16:30 |
freemangordon | or, I am missing somthign | 16:30 |
freemangordon | uvos__: like, how am I supposed to put button over XV playback? | 16:31 |
Wizzup | uvos__: hm, no modem and sound, I guess that is what you said was fixed now? | 16:31 |
uvos__ | Wizzup: modem ist just not rebased yet, sound is probably fixed with the patch tmlind posted | 16:32 |
uvos__ | freemangordon: you dont | 16:32 |
freemangordon | hmm? | 16:32 |
uvos__ | freemangordon: with xv you either do wat mpv dose (insert the widget into the stream) | 16:32 |
Wizzup | ok | 16:32 |
uvos__ | or you forget about xv and use gl (wich is what everyone has doen | 16:32 |
uvos__ | ) | 16:32 |
freemangordon | or, use color keying | 16:32 |
uvos__ | except everyhone is dropping support | 16:33 |
uvos__ | (incl xv in the first place) | 16:33 |
freemangordon | IIRC, this does not mean overlay | 16:33 |
uvos__ | so its not really a sollution | 16:33 |
freemangordon | ok, but we don;t have glimagesink that work with gles2 :( | 16:33 |
freemangordon | *works | 16:33 |
uvos__ | hm ok :( | 16:33 |
freemangordon | also, I don;t see what is the issue with using DSS HW to do what we want at no cost, but have to write shaders and whatnot | 16:35 |
freemangordon | that's suboptimal | 16:35 |
freemangordon | BTW, drm subsystem has support for color keying | 16:36 |
uvos__ | no issue, its just how everyone else in the stack is railroading you too | 16:36 |
freemangordon | not ,really | 16:36 |
freemangordon | kernel has support | 16:36 |
uvos__ | sure the kernel | 16:36 |
freemangordon | it is that particular driver that does not have | 16:36 |
freemangordon | uvos__: see https://www.kernel.org/doc/html/latest/gpu/drm-kms.html, "Existing KMS Properties" | 16:39 |
freemangordon | armada, nv and radeon support it | 16:40 |
freemangordon | BTW, qt somehow manages to draw on top of the video | 16:41 |
freemangordon | it is just that we are missing transparency | 16:42 |
uvos__ | qt is gl accelerated internally, presumably it just renders the video to a texure and composes its own window | 16:44 |
uvos__ | or do you mean it manages while using xv | 16:44 |
freemangordon | yes | 16:44 |
uvos__ | hmm | 16:45 |
freemangordon | video playback in openmediaplayer is done by mafw-gst-renderer (separate process), which uses xvimagesink | 16:45 |
freemangordon | so omp passes playback widget window xid to mafw | 16:45 |
uvos__ | hmm ok | 16:46 |
dsc_ | did someone say qt | 16:46 |
freemangordon | and mafw-gst-renderer gets 'embeded' | 16:46 |
freemangordon | yes, I am in love with it :) | 16:46 |
uvos__ | xembed yeah | 16:46 |
uvos__ | but thow it renders over xv | 16:46 |
uvos__ | no idea | 16:46 |
freemangordon | me neither | 16:46 |
uvos__ | without hw overlay | 16:46 |
freemangordon | but it works fine | 16:46 |
freemangordon | besides the buttons that are in the middle (play, pause, etc) that have funny backgrounds | 16:47 |
freemangordon | maybe xorg does something sane, dunno | 16:47 |
freemangordon | and no, qt does not use gl for QWidget, afaik | 16:47 |
uvos__ | it sure dose for qmultimedia surfaces | 16:48 |
uvos__ | which you arnt using | 16:48 |
freemangordon | yeah | 16:48 |
freemangordon | maybe somehow the clipping region is correct | 16:48 |
uvos__ | these things are fundamentally impossible in wayland btw, jfyi | 16:51 |
uvos__ | so porting of that will be fun | 16:51 |
uvos__ | (or rather you would just ditch this for qmulitmedia) | 16:51 |
freemangordon | yeah | 16:52 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!