20:29 | _rmk_ | rabeeh: ping |
20:36 | mhoney_work | pong |
20:43 | _rmk_ | you don't look like a rabeeh :) |
20:46 | frilled | not really ;) |
20:48 | _rmk_ | does rabeeh not talk here very often anymore? |
20:51 | frilled | He's an vacation, IIRC until this weekend. Otherwise, he's very much present. |
20:51 | frilled | on, even |
20:52 | _rmk_ | in which case I will wait until after this weekend and then chase him for details about how the TDA19988 is connected to the Armada, particularly the sync/de signals |
20:52 | frilled | I guess he'll answer it straight out of his mind :) |
20:53 | _rmk_ | and also whether there's somewhere on the board I can pick those signals up without having to totally dismantle my cubox or whether he can answer what the default polarity of the signals are with the 'invert' bits in the armada at 0 |
20:54 | frilled | I know for sure he's very open to modifications or picking up something from the PCB if you want to. Just wait a couple of days |
20:55 | _rmk_ | the problem is, the tda19988 driver which was merged for 3.9 is being a real pain at the moment and misbehaving in all sorts of wonderful ways |
20:55 | _rmk_ | I can get 480p 576p and 720p out of it just fine, 1080i (my tv won't do 1080p) no chance except with the NXP driver |
20:56 | _rmk_ | and I've long wondered how the armada lcd controller behaves with its sync outputs - the marvell docs are pretty poor there :( |
20:56 | frilled | can't really help you there, but I get the feeling that just about everybody is reluctant with the newer kernels anyway |
20:57 | frilled | I guess I'll have to order another cubox for some bleeding edge stuff since my first one is so productive ;) |
20:58 | frilled | I'd kinda like to play around with newer kernels on the cubox |
20:59 | frilled | currently on 3.5.7 |
21:00 | frilled | somehow I lost my sound h/w with 3.5.7-11(arch) ;P |
21:09 | _rmk_ | my kernel is Rabeeh's kernel from a long time ago, plus 3.6, 3.7, *lots* of my patches, and finally 3.8. At some point I'll merge 3.9 in and rebase all my stuff, provided the 3.9 merge doesn't break it |
21:10 | frilled | any hint on what the problems are? could save me quite some tim I guess |
21:10 | _rmk_ | I do at least have audio working with the v3.9 tda19988 driver for some resolutions though, but there's currently a debate about whether my programming of it (which ties up with the nxp source) is correct by others who have worked on the tda19988 driver |
21:11 | _rmk_ | on merging the v3.x releases on top of rabeeh's tree? there's numerous conflicts I'm afraid. |
21:12 | frilled | mh, I guess I'll just have to check it out |
21:15 | _rmk_ | just looking |
21:16 | _rmk_ | my tree was based on Rabeeh's tree in August 2012 |
21:16 | frilled | yeah, like ancient |
21:16 | _rmk_ | top commit there was "avoid genirq: Flags mismatch on bootup, patch from rudi" |
21:17 | _rmk_ | v3.6 merged with conflicts in 4 files - sdhci-dove.c, plat-orion/common.c, mach-dove/common.c, mach-dove/include/mach/sdhci.h |
21:17 | _rmk_ | v3.7 has 17 conflicting files (including one subtle one) |
21:18 | frilled | hmk. I'll dig into it should I have some more time :-/ |
21:18 | _rmk_ | frilled: well, I suspect much of the changes in Rabeeh's tree would have been for graphics or audio... and its likely that the audio stuff is a result of me pushing it out of my tree |
21:19 | jnettlet | _rmk_, for the XO-4 we had to use a Marvell provided signal tuning board to fix our hdmi out signals. |
21:19 | frilled | push out why? |
21:19 | _rmk_ | frilled: into mainline which then would come back to rabeeh eventually |
21:20 | _rmk_ | jnettlet: any ideas what "tuning" was applied? |
21:20 | frilled | ok |
21:21 | jnettlet | _rmk_, there are a bunch of tunables on the our hdmi hardware. Let me see if I have the docs handy. |
21:22 | jnettlet | _rmk_, we don't have the same hdmi encoder as the dove platform. |
21:22 | frilled | rabeeh posted some nice stuff liek 2 weeks ago |
21:23 | _rmk_ | frilled: yea, uio stuff and bmm stuff from what I see on github |
21:23 | _rmk_ | I've rewritten all that here :) |
21:24 | frilled | ;) |
21:27 | jnettlet | _rmk_, it probably doesn't pertain to your use case, but I switched to using ION Memory Manager backed by CMA and it works quite well. One big bonus is that the memory can be used for other things when you aren't watching videos. More desktop/laptop oriented rather than media-center oriented |
21:28 | frilled | you're right, it doesn't. as of yet. The box ist just my workhorse for nas, dnsmasq, httpd, but I'm thinking about some more, like one for media center purposes, too |
21:30 | frilled | which is why I am basically interested in getting into that side of things soon |
21:31 | _rmk_ | one disappointing thing that I've found recently is with an eSATA drive connected, my wireless mouse/keyboard has become very picky about whether they work |
21:32 | _rmk_ | without the eSATA drive connected, they work from the room next door |
21:32 | _rmk_ | with the eSATA drive connected, 4ft becomes really touch and go |
21:32 | jnettlet | _rmk_, our tuning was for the physical transmitter. Basically some registers that changed the the physical signalling of the hardware. damp eamp, itune...couple of others. |
21:32 | jnettlet | I still haven't gotten audio working with 1080p on our hardware yet either. |
21:33 | frilled | eh?? Mh. I have eSATA connected, but no wireless stuff, but no problerms whatsoever. sound like a badly shielded esata cable ;) |
21:33 | _rmk_ | frilled: maybe... I'm thinking about wrapping the eSATA cable up in tin foil to test that theory :) |
21:34 | frilled | connect to ground on one side only, and test. |
21:39 | _rmk_ | yea, at the cubox end I think |
21:40 | _rmk_ | it was the cable provided with the icybox enclosure so probably wasn't the best quality cable |
21:41 | frilled | can't say I ever had probems with those, although I only have usb2 and usb3 ones |
21:41 | frilled | I'm using a raid box on the sata port now |