IRC log of #cubox of Sun 25 Nov 2012. All times are in CET < Back to index

12:02 rabee 12:02 * rabeeh has a working xbmc frodo with 5.1 passthrough
12:02 rabeeh thanks to shesselba and rmk_
12:23 rabeeh anyone has 7.1 truehd to try out?
12:26 N30N what do we need to change to get this working (patches avalible)?
12:42 shesselba rabeeh: np :) .. are you aware of any player that can do truehd pass-through?
13:10 shesselba rabeeh: http://www.demo-world.eu/trailers/high-definition-trailers.php
13:10 rabeeh N30N it's mainly changing shesselba's kernel driver to be SPDIF instead of S/PDIF, and then place the Kirkwood-SPDIF.conf in the alsa cards directory
13:13 rabeeh shesselba: do you have geexbox?
13:19 rabeeh flac 96khz works too - http://www.solid-run.com/phpbb/viewtopic.php?f=2&t=1006#p5531
13:25 _rmk_ yea, / in names used for directory names doesn't work all that well
13:27 dv_ who came up with "S/PDIF" anyway
13:27 dv_ so impractical
13:28 dv_ so _rmk_'s patches got merged to rabeeh's kernel?
13:29 rabeeh dv_: nop
13:29 rabeeh still using older code; but only changed S/PDIF thing.
13:29 dv_ so rmk's patches were solely about spdif ?
13:29 rabeeh to be honest i can't keep up with all the work and hoping someone can maintain the off-mainline kernel
13:30 rabeeh dV_: rmk's patches where against mainline, most of it contains patches written by him, shesselba and me
13:31 dv_ okay, and what is the difference between your kernel and the mainline? how much does it deviate?
13:31 rabeeh (mainline + rmk patches) is almost like my kernel
13:31 dv_ because right now I find the situation somewhat confusing. for example, archlinuxarm apparently uses mainline (plus a few patches?), but then there is your off-mainline kernel
13:32 dv_ ah, okay
13:32 rabeeh my main intention of having off-mainline tree was gettinig into the 3.x area (where the original kernel was 2.6.32.9)
13:32 dv_ so if rmk's patches are pushed upstream, then you almost dont have to maintain yours anymore?
13:33 rabeeh i was hoping to get that part of the kernel as soon as possible; but then arm-kernel got the DT as mandatory to get into mainline, and then generic clock too
13:33 dv_ oh :/
13:34 rabeeh dv_: somehow we got into a situation that we have random work going in/out on mainling kernel
13:35 rabeeh now we have lots of things ready (again thanks shesselba and rmk) and need people upstreaming things.
13:35 rabeeh notice that shesselba got u-boot head ported to CuBox too :)
13:35 dv_ yay
13:35 dv_ unfortunately, i have no experience with kernel stuff, and not much time overall, otherwise i'd try to help a bit
13:36 dv_ still, I want to set up several rootfs for several devices (for a variety of reasons), I could help with testing patched kernels perhaps
13:37 _rmk_ rabeeh: well, I have your kernel + 3.6 + my stuff, and I'm planning when 3.7 comes around to move it forward to 3.7
13:38 _rmk_ but I tend to keep this git tree in an 'unstable' state; I run my git trees more like a quilt patch set than a set of stable git commits
13:38 _rmk_ so the 'my stuff' will become yours + 3.6 + 3.7 with my stuff on top of that
13:39 rabeeh _rmk_: so; what should i be doing now? i can pull to newer kernel tree and get my stuff ported back again
13:40 rabeeh _rmk_: but then i know some key kernel-arm maintainer that can get things pushed straight upstreamed :)
13:41 rabeeh or; lets have it in different way; what do you think the best way to get things straightened for mainlining everything minus NXP driver minus GPU driver?
13:41 _rmk_ well, I'm hoping that over the course of 3.8/3.9, we'll have just the problematic bits outstanding, which will be vmeta, bmm, drm, nxp and galcore
13:41 dv_ actually, are there any known issues with the 3d accelerator or the video decoding engine?
13:41 dv_ i mean, the audio driver had severe issues prior to rmk's patches. would you consider 3d and video to be in a better shape?
13:42 _rmk_ I think we may be able to get vmeta/bmm in, provided it doesn't descend into a licensing flamewar (it really shouldn't)
13:43 _rmk_ as for galcore, I think that requires a serious amount of rework and we need to take a big decision there: do we do one last update against Marvell stuff, and then put significant cleanup effort into it to bring it to a state where it can be mainlined, or do we persist with it being out of tree.
13:43 rabeeh _rmk_: even getting stable CuBox with zero multimedia stuff is an achievement
13:44 rabeeh _rmk_: there is potentially a good chance of getting galcore cleaned up both by CuBox and OLPC devs (jnettlet)
13:44 rabeeh hardware wise it's different versions of the same IP; but share same galcore driver
13:45 _rmk_ well, cjb sent me an OLPC to help with this
13:45 _rmk_ remember that I already have a significant patch set for galcore against your tree
13:47 _rmk_ and OLPC already have a later version than what's in your tree
13:47 _rmk_ (though with much the same bugs)
13:47 rabeeh yes. the 4.6.4
13:48 rabeeh _rmk_: if i would redo all the patches to get them straightened for upstreaming, against which tree is best?
13:48 _rmk_ not published; they have some 0.8.0.2xxx version (I need to look at firefox on a different machine where I have it loaded up to give the exact version)
13:48 rabeeh asoc-linux?
13:49 _rmk_ if you can wait until next week I'll see about publishing some of my tree...
13:49 _rmk_ however, what I have in my tree is different from the patches I'm upstreaming (though the end result is essentially the same)
13:49 _rmk_ that's one of the problems of working from different starting points :(
13:51 _rmk_ what I would suggest is waiting until we've got mainline mostly sorted first before looking to move stuff forward, otherwise I predict lots of pain.
13:54 rabeeh _rmk_: ok.
13:59 _rmk 13:59 * _rmk_ needs to look at the damage the weather has brought to our garden (we have a tree down)
14:04 rabeeh google maps ?
14:04 _rmk_ naa, they only update once every three years
20:06 rabee 20:06 * rabeeh added geexbox-armhf-devel to the installer
20:06 rabeeh thanks tomlohave for the support