02:31 | Coburn|2 | so I guess I need to use a 3.5/3.6 kernel if I want video output? |
02:31 | Coburn|2 | LAME. |
02:31 | Coburn|2 | unless dv_ can shine some light onto it |
02:32 | dv_ | ? |
04:39 | Coburn|2 | dv_: Where you doing a opensource GPU driver for Cubox? |
07:46 | wumpus | Coburn|2: https://github.com/laanwj/etna_viv |
07:53 | Coburn|2 | wumpus: is it usable for 3.10.x kernels? |
07:53 | Coburn|2 | Also, does it 2D Xorg FB HWA ? |
07:53 | Coburn|2 | does it do* |
08:17 | wumpus | if the GPL kernel driver has been ported to that, sure |
08:17 | wumpus | no one wrote an xorg driver yet |
08:18 | Coburn|2 | hmm |
08:18 | wumpus | I only wrote the gallium/mesa 3d driver, the 2d part has been reverse engineered though |
08:18 | Coburn|2 | ok |
08:19 | Coburn|2 | Just thinking the Xorg HW Acceleration may not be needed, because one of my contacts said the CPU's iwmmxt chip makes dragging things less hardware intensive |
08:19 | Coburn|2 | like on the R-Pi, usage will go from 10% idle with Xorg, to like 80% dragging a window |
08:21 | wumpus | also robclark was planning to work on a kernel driver that fits into the Linux drm interface instead of the totally alien Vivante GPL driver |
08:21 | wumpus | ok, yes it may well be that for imx6 there's a better way to do the blitting than the 2D engine |
08:22 | Coburn|2 | i'm talking about the calssic CuBox here |
08:22 | Coburn|2 | not sold on iSeries really |
08:22 | wumpus | but from what I heard, most window managers on X use 3D operations internally these days for composition and such |
08:23 | Coburn|2 | well, I'm planning to use openbox |
08:24 | wumpus | I don't know whether that uses a 3D (like compiz) or xrender (like metacity) |
08:26 | Coburn|2 | well, it is very lightweight |
09:28 | jnettlet | openbox uses standard X11 2d drawing. It doesn't even support compositing. |
09:47 | jnettlet | rabeeh, got your pull request merged. Sorting through SPL changes. My wife had invited guests over for dinner last night so didn't get a chance to sit in front of the computer. Trying to fit that in today. |
10:16 | hste_ | jnettlet:: whats the url to the latest git? |
10:16 | MikeSeth | rabeeh: you guys really should post shipping update on the forums, people are guessing from order numbers etc |
10:55 | dv_ | robclark, wumpus: whats the state of the drm driver? |
10:56 | dv_ | MikeSeth: they probably dont know exact dates |
10:57 | MikeSeth | dv_: nobody does, which is why I am going to the post office every morning in vain hope |
10:59 | wumpus | dv_: hasn't progress any |
11:08 | _rmk_ | dv_: which drm driver? |
11:08 | dv_ | _rmk_: robclark's one |
11:08 | dv_ | for etnaviv |
13:55 | robclark | dv_, state is mostly still vaporware.. I got busy w/ a few other boards (plus "real" work) before I managed to get the up and running w/ carrier1. I do have right uart cable now, etc.. now all I need is time ;-) |
13:56 | robclark | hope to finally get back to it in a couple weeks |
14:23 | rabeeh | robclark: i can help with questions; but can't help with time |
14:23 | rabeeh | time is still my biggest enemy |
14:29 | hste_ | rabeeh: its not easy to win a battle against time.. :) |
14:35 | _rmk_ | rabeeh: you need a tardis |
14:43 | robclar | 14:43 * robclark needs a tardis :-P |
14:43 | robclark | rabeeh, sure, no prob.. |
14:44 | robclark | I guess at least on the upside, imx drm hdmi support should be in better shape now than when I first got the board ;-) |
15:07 | _rmk_ | robclark: yea, when I can kick the patches out and get people to ack them :) |
15:46 | neofob | so i got cryptodev-linux works with cubox (marvell version), openssl needs to recompile with CRYPTODEV flags and all |
15:50 | rabeeh | neofob: have performance number with and without? |
15:50 | rabeeh | as i recall on big buffer (8k) you should be reaching 350Mbps on AES128 cipher |
15:51 | neofob | rabeeh: openssl speed -evp md5 gives over 800MB/s |
15:51 | neofob | vs. 65-73MB/s before |
15:52 | neofob | i need to apply the patch from debian's openssl tree to avoid the message "no version info found..." |
15:53 | rabeeh | MByte or Mbit? |
15:53 | rabeeh | either sounds too high |
15:53 | neofob | MByte |
15:55 | neofob | i'll try with AES128 later today (i dont have cubox with me now) |
15:55 | robclark | _rmk_, speaking of kicking out patches, I don't suppose you care to send the 8 byte get_user patch? I think by now it is more your patch than mine (and, well, I hope someday to get the atomic modeset stuff finished, which will depend on it) |
15:55 | neofob | is `openssl speed -evp md5` reliable to test out the machine speed? |
15:56 | Coolgeek | hi, is it possible that my cubox have some heating issue ? |
19:08 | MarcusVinter | Hello everyone. Just thought I'd come and introduce myself here. I'm just joining the community and working with the Carrier-One right now but I'm going to be fairly active in here from now on. Would be nice to know if anyone else here is working on it right now and your progress. Cheers. |
19:12 | rabeeh | MarcusVinter: welcome to the channel. |
19:12 | MarcusVinter | Thanks Rabeeh. I actually work for i-Virtuals so I thought this would be a nice direct line of communication. |
19:13 | rabeeh | oh; great to see you here |
19:14 | MarcusVinter | Thanks. I should be fairly active over the next few weeks anyway as I work on the C-1. Looking forward to getting stuck in properly. |
19:15 | MarcusVinter | I'm going to head off anyway for the day. Take care and I'll speak to you all soon. Thanks! |
21:21 | jas-hacks_ | dv_: Still having problem using playbin, is that related to your upcoming fixes? |
21:21 | dv_ | hi |
21:21 | dv_ | what problems exactly? |
21:22 | jas-hacks_ | I can play video using "gst-launch-1.0 -fe filesrc location=$1 ! qtdemux name=dmx dmx. ! h264parse ! imxvpudec ! eglvivsink dmx. ! queue ! fakesink" |
21:24 | jas-hacks_ | But fails using " gst-launch-1.0 playbin uri=" |
21:25 | jas-hacks_ | osting on bus error message from element 'qtdemux0': GstMessageError, gerror=(GError)NULL, debug=(string)"qtdemux.c\(3948\):\ gst_qtdemux_loop\ \(\):\ /GstPlayBin:playbin0/GstURIDecodeBin:uridecodebin0/GstDecodeBin:decodebin0/GstQTDemux:qtdemux0:\012streaming\ stopped\,\ reason\ error"; |
21:25 | dv_ | I need exact details. run it with GST_DEBUG=2 and GST_DEBUG_DUMP_DOT_DIR=. grab the generated .dot file with "PAUSED_PLAYING" in its filename, convert it to PNG using imagemagick convert, and upload the image somewhere |
21:39 | jas-hacks_ | dv_: http://stende.no-ip.info/jas/c1/PAUSED_PLAYING.png |
21:39 | jas-hacks_ | dv_: Thats pretty cool |
21:39 | dv_ | and do you see any output with "ERROR" written on the log? |
21:42 | dv_ | DISPLAY=:0 GST_DEBUG=2,*vpu*:9 gst-launch-1.0 -fe playbin uri=file:///media/mmcblk0p3/ironman3-tlr1-m4mb0_h720p.mov video-sink="eglvivsink" <- this plays fine for me |
21:44 | jas-hacks_ | Only error was this : 0:00:04.073867051 3368 0x2c501d50 ERROR imxvpudec ../src/vpu/decoder/decoder.c:847:gst_imx_vpu_dec_handle_frame: failed to decode frame: failure |
21:45 | dv_ | ok re-run it with my GST_DEBUG settings |
21:45 | dv_ | then copy the log output to codepad.org |
21:45 | dv_ | (remember to set it as "text") |
21:54 | jas-hacks_ | dv_: http://codepad.org/iNJRE6BP |
21:57 | dv_ | what gstreamer version are you using? |
22:03 | jas-hacks_ | http://packages.ubuntu.com/raring/gstreamer1.0-plugins-base |
22:04 | dv_ | hmm I am using 1.2.0 |
22:04 | dv_ | 1.0.6 is quite old now |
22:05 | dv_ | if possible, try to build and install 1.2.0 , and see if it works with this version |
22:05 | dv_ | I can imagine h264parse has some bugs in 1.0.6 that were fixed since then |
22:05 | jas-hacks_ | I'll try compiling 1.2 sources |
22:08 | dv_ | 1.0.6 was released in march . now see this: http://cgit.freedesktop.org/gstreamer/gst-plugins-bad/log/gst/videoparsers/gsth264parse.c |
22:08 | dv | 22:08 * dv_ wonders if 1.2.0 should be a minimum requirement |
22:09 | jas-hacks_ | could try patching it up or not worth the effort? |
22:10 | dv_ | first lets check if your issues are gone in 1.2.0 |
22:39 | Bluerise | he came after that was said :p |
22:39 | Bluerise | oops |
22:40 | Bluerise | without context, this sounds bad. |