bencoh | nice blog update :) | 00:48 |
---|---|---|
tmlind | yeah great :) nice that terminal supports the vol keys now for font size, i did not know that | 06:23 |
mighty17[m] | hey, is there anything special needed to make usb-otg work on omap4? also usb only works when connected before booting the device, ie if once device booted and then i connect it to my pc, pc doesnt detect it :( here's my dts https://paste.debian.net/1218005/ | 07:25 |
freemangordon | oh, VRFB was never ported to omapdrm :( | 07:30 |
freemangordon | luckily the code is still there for omapfb | 07:33 |
tmlind | mighty17[m]: you need to configure the usb phy driver for it, probably the legacy drivers/usb/phy driver still for that pmic | 08:15 |
mighty17[m] | something like https://github.com/drebrez/linux/commit/55b64defc79f6c6ba4c8a9b95ad3fb6e96ac63fe?diff=split#diff-f4ce6be0e44247b3d208cb32128980f30a78e7577da984908b2630dfabd28ca7R197-R199 ? | 08:16 |
tmlind | mighty17[m]: compare to the pandaboard dts files assuming you have the same twl pmic and usb phy, i guess it's possible you have some separate usb phy for the otg interface | 08:19 |
Wizzup | freemangordon: what is VRFB ? | 09:26 |
freemangordon | virtual rotated frame-buffer | 09:56 |
freemangordon | this is what is used to rotate display on omap3 | 09:56 |
freemangordon | and maybe omap4, I am not sure it has dmm | 09:56 |
freemangordon | tmlind may know | 09:57 |
tmlind | omap4 has dmm with tiler for sure | 09:59 |
tmlind | omap3 has vrfb instead afaik | 10:01 |
tmlind | i think both provide separate addresses for various rotations, looks like vrfb configuration is missing from drivers/gpu/drm/omapdrm? | 10:03 |
tmlind | less than 400loc for drivers/video/fbdev/omap2/omapfb/vrfb.c it seems | 10:04 |
freemangordon | yeah | 10:06 |
freemangordon | also: https://patchwork.kernel.org/project/dri-devel/patch/1493893412-12178-7-git-send-email-tomi.valkeinen@ti.com/ | 10:06 |
tmlind | ok | 10:09 |
freemangordon | not really 'ok', but yeah. I would port that, but I'll need some hints where to start | 10:09 |
tmlind | no idea on that one, seems like it just needs to be configured, then saved and restored for idle states | 10:14 |
uvos | so if tiler rotation is supported by omapdss in practice atm | 10:16 |
uvos | what prevents it from working in modesetting | 10:16 |
uvos | do we know? | 10:16 |
freemangordon | no | 10:18 |
freemangordon | uvos: also, we need vrfb rotation form omap3 (n900, n9/50) | 10:18 |
tmlind | maybe nothing reconfigures the changed tiler/vrfb addresses right now on rotation change? | 10:18 |
tmlind | maybe there was some ioctl in the old ddk-1.9 setup doing that | 10:19 |
freemangordon | yes, there is | 10:19 |
uvos | freemangordon: sure but knowing why tiler is not working helps in understanding how to implement vrfb | 10:19 |
tmlind | although it sounds like the rotation change should be a standard ioctl.. | 10:19 |
uvos | freemangordon: since in theroy ms should use it | 10:19 |
freemangordon | not really | 10:19 |
freemangordon | at least IIUC | 10:19 |
tmlind | my guess is it all should be simpler than earlier anyways | 10:19 |
freemangordon | when you create BO you need to omap_bo_new_tiled() | 10:20 |
freemangordon | but, in GBM case I don;t see how this can be done | 10:21 |
uvos | no | 10:21 |
freemangordon | https://github.com/robclark/libdrm/blob/master/omap/omap_drm.c#L246 | 10:21 |
freemangordon | uvos: yes, in order to use TILER roatation, a special BO is needed | 10:22 |
uvos | sure but really the kernel should do this more or less transperantly if you setup kms/drm correctly | 10:22 |
freemangordon | yes, I agree | 10:22 |
uvos | at least thats how the interfaces are supposed to work | 10:22 |
freemangordon | but maybe this is not implemented | 10:22 |
* tmlind bbl | 10:22 | |
uvos | sure | 10:22 |
freemangordon | lemme check if omapdrm dma_buf implementation supports that, somehow | 10:23 |
freemangordon | uvos: I don;t see a way of requesting dma_buf with TILER flags | 10:28 |
freemangordon | https://elixir.bootlin.com/linux/latest/source/drivers/gpu/drm/omapdrm/omap_gem.c#L1300 | 10:29 |
freemangordon | see flags https://elixir.bootlin.com/linux/latest/source/drivers/gpu/drm/omapdrm/omap_gem.c#L1185 | 10:31 |
freemangordon | amd because modesetting uses gbm (dma_buf that is), bo's never have OMAP_BO_TILED_XXX set on them | 10:32 |
freemangordon | so, rotation does not work | 10:32 |
uvos | freemangordon: right but really omapdrm should allow the primary display plane to be rotatied | 10:51 |
uvos | by just calling the rotation transform on it in the kms interface | 10:51 |
uvos | and then the kernel should just do whatever to rotate the plane, ie move it to tiler space and configure it | 10:51 |
uvos | this appears not implemented atm | 10:52 |
uvos | but this is the right/generic way of doing it. | 10:52 |
Wizzup | MartijnBraam[m]: btw I answered the tweakers guy late last night, thanks for the refer, hope I was able to help | 10:54 |
Wizzup | uvos: freemangordon: seems like it would be quite helpful to have the omapdrm folks here | 11:23 |
Wizzup | or for us to move some discussion to a medium where they read along | 11:23 |
parazyd | uvos: This is what's left in -devel, http://sprunge.us/urFWSH | 11:25 |
parazyd | uvos: Anything from here I should promote? | 11:26 |
Wizzup | what about ke-recv ? | 11:26 |
parazyd | It's in the list. I dunno, so that's why I'm asking. | 11:27 |
uvos | you promoted leste-config you have to promote hildon-desktop or rotation breaks | 11:27 |
Wizzup | I think I just changed some logging in ke-recv | 11:28 |
uvos | also please make sure that hildon meta in stable no longer wants hildon-rotation-support | 11:28 |
uvos | as this also needs to go | 11:28 |
uvos | simple-brightness-applet is fine for promotion | 11:29 |
uvos | so is hildon-home | 11:30 |
uvos | hildon-theme-alpha | 11:30 |
uvos | hildon-theme-beta | 11:30 |
uvos | hildon-theme-devel | 11:30 |
uvos | these being in devel at all seams wierd | 11:30 |
uvos | ofono can be promoted | 11:31 |
uvos | thats just the motmdm bugfix | 11:31 |
uvos | that looks like its it | 11:32 |
parazyd | Thanks | 11:32 |
uvos | add profilesx to hildon-meta while at it | 11:34 |
uvos | if Wizzup dident allready | 11:34 |
Wizzup | that wasn't done yet | 11:35 |
Wizzup | but I agree it should be | 11:35 |
parazyd | uvos: We have a Conflicts and Provides for hildon-desktop-rotation-support | 11:36 |
parazyd | uvos: That should be fine | 11:36 |
uvos | ok | 11:36 |
uvos | anny one have had any trouble with new osso-applet-display? | 11:37 |
uvos | if no thats fine to i gues | 11:37 |
uvos | to promote | 11:37 |
Wizzup | it works for me | 11:37 |
freemangordon | Wizzup: right, but I doubt tomi will join here :) | 12:28 |
freemangordon | back then there was linux-omap, but I doubt it is anymore | 12:29 |
freemangordon | uvos: agree | 12:30 |
parazyd | uvos: What about droid4-battery-calibration? | 12:48 |
uvos | parazyd: what about it | 12:49 |
parazyd | Should I also move it to stable? | 12:49 |
uvos | i dont recall any changes | 12:49 |
parazyd | ah nvm, it's my changes to install it in sysinit runlevel instead of default | 12:50 |
uvos | and you renamed it from maserati to droid4 | 12:50 |
uvos | not that i mind | 12:50 |
uvos | but mapphone would have been more correct | 12:50 |
uvos | but no big deal | 12:51 |
parazyd | Wizzup: What about qt5 stuff? Leave that alone or move? | 13:00 |
uvos | this is me removing the color workaround | 13:01 |
uvos | its broken either way | 13:01 |
uvos | if you keep it osso_caluclator is broken, if you move it qcalendar is broken | 13:02 |
uvos | (amoung other things) | 13:02 |
Wizzup | freemangordon: maybe he would? | 13:11 |
Wizzup | parazyd: not sure @ qt | 13:12 |
parazyd | ok, then all is done now | 14:24 |
parazyd | Let's see if next week's images are ok | 14:24 |
parazyd | About hildon-theme-*, I didn't promote them yet, as I think there's some packaging bug, but I haven't confirmed it. | 14:25 |
parazyd | So that's gonna wait a bit | 14:25 |
Wizzup | dsc_: \o | 17:06 |
mighty17[m] | <tmlind> "mighty17: compare to the..." <- i suppose i need this commit https://github.com/torvalds/linux/commit/5bd2100ed2a76147ddec95499214eb606b799a22#diff-52c1876ad7abb628d8d7de325b4d82a4e4d1870ff81cc01d0af532fdb2188f88 ? | 17:07 |
mighty17[m] | but what about usb only working when connected before boot? | 17:08 |
dsc_ | hi | 17:13 |
dsc_ | does maemo-leste run on my iphone 13 pro max? | 17:14 |
Wizzup | dsc_: ;-) | 17:16 |
Wizzup | It runs on a few devices, but not iphones | 17:16 |
dreamer | maybe you can ask steve jobs' ghost | 17:19 |
dsc_ | :P | 17:23 |
uvos | mighty17[m]: well you dident explain what "usb working" or "usb not working" means and its very ambigous | 18:03 |
uvos | mighty17[m]: but if it "works" when coccentcted before boot its because your bootloader is initalizing usb for itself and the hw state carrys through | 18:04 |
mighty17[m] | uvos: yeah thats a possible reason | 18:04 |
uvos | the bootloader dose so to allow android fastboot or rather whatever the proprietary samsung equivalent is to work | 18:05 |
mighty17[m] | the thing is my pc doesnt show anything in dmesg, but on device i get `twl6030_irq: Unmapped PIH ISR 20 detected` when i do not let bootloader initialize it | 18:05 |
uvos | so some plug/usb interrupt fires and your not handling it | 18:07 |
mighty17[m] | any example of it? | 18:09 |
uvos | as tmlind said look at how it works on some other device with twl6030 pmic | 18:10 |
uvos | (what you have to do is device dependant) | 18:10 |
mighty17[m] | mine is twl6032, they're different | 18:11 |
mighty17[m] | afaik my device has some usb_sel1 thingy | 18:11 |
uvos | well your going to have to figure out what the differences are and ammend the usb phy driver | 18:11 |
mighty17[m] | oh, now thats weird, it worked now | 18:12 |
mighty17[m] | ie without bootloader initialisation | 18:12 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!