07:41 <kroon> dv_, ping

07:51 <dv_> .

07:54 <dv_> kroon: the ipu gpu combination should be possible, though I never tried it

07:54 <dv_> note however that you can set the eglvivsink to borderless mode. it is then possible to exactly position it. I was able of doing picture-in-picture like combinations with this

07:55 <dv_> one eglvivsink showing a 720p video in the upper left corner, one 1080p video fullscreen in the background

07:56 <dv_> according to jnettlet, the gpu wont break a sweat over this

07:58 <jnettlet> yeah, two 30fps videos one not even fullscreen is nothing

08:01 <dv_> I would actually not recommend to use the ipu for this

08:01 <dv_> the ipu sink is much slower than the eglvivsink

08:02 <dv_> in fact, I was contemplating to remove the ipusink at some point. but some people need it.

08:14 <kroon> dv_, what I would like to do is render a static background image in one framebuffer, and have Qt5 render a gui in a second framebuffer. The reason I am interested in trying this out is that I suspect that I will get better performance if Qt5 doesnt have to render the background for each frame, but can just do basically a glClear(). You have any thoughts on this ?

08:15 <dv_> so qt5 should be the overlay?

08:15 <kroon> dv_, yes

08:15 <dv_> that should work

08:16 <dv_> then gstreamer is using only one sink (the ipusink) I take it

08:17 <kroon> I dont think I have to involve gstreamer, right ? I could just write the background image in /dev/fb0, and make sure Qt5 uses /dev/fb1 correctly, right ?

08:19 <dv_> oh so when you mentioned full hd it was just for one image

08:19 <dv_> yes I guess that would work

08:20 <kroon> dv_, very cool, thanks. I have experiementet with unblanking fb1, and resizing it to fullhd, same as fb0, and it does seem to do what I expect so far

08:35 <atiti> nice

08:35 <atiti> was thinking about doing something similar

11:35 <hste_> cnxsoft: internet is still open in thailand?

12:14 <cnxsoft> Test test

12:14 <cnxsoft> hste: yes :)

12:18 <hste_> cnxsoft: :)

12:35 <kroon> dv_, I'm not sure wether this will be an improvment to letting the gpu render the background anyway ...

17:03 <Er0l> hste: did try to boot old kernel 3.0.35 with this new uboot for 3.10.17?

17:07 <hste_> no

17:08 <hste_> Er0l: did u get sound over hdmi on 3.10.17?

17:08 <Er0l> I have issues with 3.10.17, it's unstable so I wanted to revert back to 3.0.35.

17:10 <hste_> Er0l: what problems?

17:10 <Er0l> kernel oops, it freezes. spent two days debugging still no progress

17:11 <hste_> mine is stable as long as I dont use the rtl8192 module

17:11 <Er0l> I have disabled rtl8192 and am using usb ethernet adapter.

17:12 <hste_> Do u use a powered hub?

17:13 <Er0l> no, shoud I ? it's connected directly to gk802 usb port.

17:13 <hste_> I had problems when not using a powered hub on usb, not enough power

17:14 <Er0l> I will test that right now.

17:17 <Er0l> with usb powered hub, my usb ethernet adapter goes from USB 2.0 to USB 1.1

17:22 <hste_> is it a usb 2.0 hub?

17:22 <Er0l> yes

17:23 <hste_> can u pastebin your dmesg?

17:25 <Er0l> http://pastebin.com/hhFixqPT

17:30 <Er0l> seems this hub is fake one so I can forget about this

17:32 <hste_> but it should be more stable with powered hub

17:37 <Er0l> I will try to revert to 3.0.35 since it was stable without powered hub.

17:44 <hste_> Er0l: you could try this http://stende.no-ip.info/imx6/kernel-3.10.31-mix.tar.xz

17:52 <Er0l> hste from where did you get the source for this kernel ?

17:54 <hste_> Er0l: its frescales new alpha with 3.10.17 vivantedrivers and rtl8192cu module

17:55 <hste_> Er0l: not rtl8192cu but the 8192cu module :)

17:55 <Er0l> hste: thanks, I am running it now. I will test and see if it is stable

17:56 <Er0l> hste: I also found rtl8192cu-fixes on github that worked , but I blacklisted the modules when I was trying to locate the cause of kernel panic.

17:57 <hste_> I guess u got 8192cu module loaded now :)

17:58 <Er0l> yes it's loaded. I will test it in a few moments :)

17:58 <hste_> I don't know if they work my wifi is caput. no antenna....

17:59 <Er0l> you pulled out the cable from the board ?

17:59 <hste_> yes it fall of long time ago

18:00 <Er0l> mine is still holding up, but I went for usb ethernet adapter since I am using NAS share I needed a little more troughput.

18:04 <Er0l> hste: 73.3 Mbits/sec

18:04 <Er0l> hste this is speed tested with iperf on my usb ethernet adapter

18:07 <hste_> Er0l: not bad for an usb dongle

18:07 <Er0l> hste: but wifi is really slow with this driver with 192.168.254.6 port 46879

18:07 <Er0l> sorry :) wrong paste just 3.81Mbits

18:10 <Er0l> hste: do you know where I can download this alpha kernel source 3.10.17 ?

18:10 <Er0l> hste: sorry I meant 3.10.31

18:25 <hste_> Erol: I packed mine here http://stende.no-ip.info/imx6/3.10.31_1.1.0_alpha_mix_oldvivante.tar.xz

18:27 <Er0l> hste: tnx a lot. I want to recompile and include cifs module.

18:31 <hste_> Erol:: I guess u could turn off some debug I had in the config

18:34 <Er0l> oops again, when unpacking your kernel. http://pastebin.com/dznhdDEN

18:41 <hste_> Er0l:http://www.linuxfoundation.org/collaborate/workgroups/networking/sk_buff

18:44 <hste_> Er0l: what usb netdongle do u have?

18:46 <hste_> Er0l: try blacklist 8192cu

18:46 <Er0l> hste: it's using asix driver

18:50 <hste_> Er0l: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=35773dac5f862cb1c82ea151eba3e2f6de51ec3e

19:04 <Er0l> hste: I'll test this, but mine is usb 2.0. we'll see. it worked without issues with 3.0.35, but it was different asix driver version.

19:11 <hste_> if you google you can find thing like this: http://www.spinics.net/lists/netdev/msg191801.html

23:28 <rz2k> hste_: I've setup yocto with chromium, qt5.2 and gcc rootfs ready at my home dir on your server

23:28 <rz2k> chromium only took a couple hours to build with your machine!

23:29 <rz2k> I've built it for wandboard-quad, but you can grab the rootfs and replace kernel modules with needed ones

23:30 <rz2k> and run it where you want