IRC log of #cubox of Tue 05 Nov 2013. All times are in CET < Back to index

00:09 taz_ rabeeh: your home is near your office ;)
00:48 sam_nazarko rabeeh: are you around?
00:50 taz_ see you good night
02:20 tenedos can anyone help me with uboot. i am prob doing something that wont work but i have a arch linux build built for the imx6 and it has a file 6x_bootscript in the root of the filesystem but i have no idea how to get uboot to run it to try to boot the OS
09:48 DHowett (I forget: does the cubox pro do usb-otg?)
09:49 DHowett (I imagine probably not, but it couldn't hurt to ask)
09:52 DHowett aha; of the armada 510, Each USB 2.0 interface contains a single dual-role controller that can act as a host or a peripheral controller (aka USB controller).
12:15 rabeeh sam_nazarko: ping
12:19 tazou hello rabeeh :)
12:33 rabeeh tazou: hi
13:25 tazou hi how are you ? rested ? :)
13:46 jnettlet__ rabeeh, my naming suggestion for the contest is in.
13:47 jnettlet__ took some thinking over, but the nice autumn weather inspired me.
13:47 rabeeh it's not published yet
13:47 rabeeh what is it?
13:54 jnettlet__ The SolidRun Roadster, with appropriate badges for the various boards. S1, D1, Q2 etc. The name invokes a feeling that matches the hardware, free spirited, fun, fast, it is even open-air and the top comes off :-) As an automobile the roadster is historically a platform that has been tuned up and tricked out to make hotrods and this is what the SR Roadster computing platform is all about.
14:03 bencoh SR roadster ? hmm
14:03 bencoh sounds fun :)
14:08 jnettlet bencoh, doesn't it?
14:09 bencoh :)
14:11 jnettlet Come and take one for a spin. We can get you into an entry level model for a very reasonable price :-)
14:14 Bluerise Heh, sounds nice
14:29 tazou rabeeh, I try to compile a new kernel yesterday (i don't know if it's required) but it fail to boot : http://www.solid-run.com/phpbb/viewtopic.php?f=7&t=1566#p8365 can you look it please when you have time ? :) thanks
14:31 rabeeh jnettlet: i like it :)
14:35 rabeeh tazou: just replied to you on the forums
14:35 rabeeh the cubox installer is actually a single uImage that has both the kernel and initramfs.
14:36 tazou ho rabeeh thank you :)
14:37 rabeeh np
14:37 tazou rabeeh, must I compile this https://github.com/rabeeh/cubox-installer with my corss toolchain ?
14:37 rabeeh no
14:37 rabeeh buildroot already has it's own cross compilers
14:38 rabeeh http://www.solid-run.com/mw/index.php?title=CuBox_Installer_Developer_Page
14:38 rabeeh there should be a output/images/uImage file
14:38 tazou ok thanks I read it...
14:38 rabeeh just place it on top of the one that you have on the USB stick
14:41 tazou ok, i'm git clone the project on my laptop, to do "make cubox_installer_defconfig"
14:41 tazou download in progress...
14:44 tazou dependancies installed
14:44 tazou make in progress
15:13 tazou rabeeh, compilation in progress... Question : that i'm compiling , is already patched for the mtd "bug" ?
15:41 tazou rabeeh, the compilation failed :( do you know why please ? here is the end logs http://pastebin.com/tjz14g9M
16:15 mhoney What is the naming contest for? Is it a name for the board or the whole product with case and all? Is it going to be the same base name for the whole product line or is it for a particular version like the i4 pro?
16:19 Bluerise mhoney: For the board, C1
16:22 mhoney ok thx
16:40 _rmk_ jnettlet: have you managed to sort out the video timings with uboot yet?
16:40 _rmk_ jnettlet: or are we just going to declare uboot hdmi video output not worth the effort?
16:46 _rmk 16:46 * _rmk_ just got a bit of broken audio out of the tv... with less than 400 lines of code :)
16:46 _rmk_ compare that to 1437 lines in the FSL BSP ASoC driver
17:06 jnettlet _rmk_, I plan on sorting it out, but it is not a high priority on my list.
17:11 MikeSeth I am going to get my cube this month, am I?
17:59 tazou go to home, see you
18:54 jnettlet _rmk_, are you doing a straight alsa driver and skipping out on ASoC?
18:57 _rmk_ jnettlet: indeed; why use asoc when there's nothing there that needs asoc.
18:57 jnettlet _rmk_, is the audio shuffled around with the cpu or using a dmaengine?
18:58 _rmk_ its got another random non-standard buffer format - you have to fill the buffer with something like iec958 frames but not quite
18:58 jnettlet great
18:58 _rmk_ freescale decided to do the conversion in the kernel... in... interrupt... context.
18:58 jnettle 18:58 * jnettlet is glad you are working that out
18:59 jnettlet ewww
18:59 _rmk_ another problem is that the hardware is incapable of properly returning where in the dma buffer it is.
19:00 _rmk_ it tends to tell you that it's 512 bytes from the end address, or the previous "known" location (iow, 512 bytes from the end of a previous buffer)
19:00 _rmk_ I have startup and playback working, its just what happens at the end now.
19:01 _rmk_ alsa's silence fill produces noise.
19:08 taz_ hi
19:10 taz_ rabeeh: I answer on the forum : http://www.solid-run.com/phpbb/viewtopic.php?f=7&t=1566&sid=17ae3c58b5f627db5f76b7d4c66e2d0a&start=10#p8367 it's a compilation error
19:10 taz_ maybe you have a uImage already compiled on some website, available to be downloaded ?
20:13 _rmk_ jnettlet: damn it, looks like pulseaudio goes in at the hw:0,0 level in alsa, and even though it negotiates iec958 format (the only one the driver supports) it gives it S16LE or similar samples
20:18 jnettlet _rmk_, that sounds like a server configuration setting
20:20 jnettlet yep
20:20 jnettlet ; default-sample-format = s16le
20:20 jnettlet ; default-sample-rate = 44100
20:21 jnettlet but if the driver doesn't support that then it shouldn't advertise it as an available format. Obviously you know that, just trying to talk my way through why PA would try to use that.
20:23 dv_ can the pulseaudio resampler be disabled?
20:24 dv_ in case it also picks an unsupported sample rate and just resamples to whatever the hardware supports
20:25 jnettlet that is really all PA should be doing. If it detects a hardware supported sample rate it is supposed to just push that through.
20:28 dv_ the resampler does make sense for sw mixing input
20:28 dv_ but not for output
20:32 jnettlet which is one of the only meritorious things that PA accomplishes.
20:40 _rmk_ you know what film should be watched (on the cubox) tonight? especially if you're in the uk?
20:41 jnettle 20:41 * jnettlet awaits with bated breath
20:41 jnettle 20:41 * jnettlet wonders if there is anyone in the UK other than _rmk_ on this channel.
20:42 _rmk_ sam_nazarko is actually within 20 miles of me.
20:43 shesselba V
20:43 _rmk_ close enough :)
20:43 jnettlet those lurkers are always to close for comfort
20:50 jnettlet so _rmk_ you going to tell us what film we are supposed to be watching?
20:50 shesselba V
20:51 shesselba guy fawkes day, V refers to it
20:51 jnettlet The TV movie/series/or V for Vendetta
20:51 shesselba the latter
20:51 jnettlet great movie love it.
20:52 jnettlet regardless what most people say I think the Wachowski brothers did that graphic novel a solid.
20:53 jnettlet I am always a bit amazed at the emotion that Hugo Weaving can relay through that mask.
20:53 _rmk_ jnettlet: absolutely
20:57 jnettlet okay they have it on Netflix. I am off to finish making dinner and watch it.
20:57 _rmk_ heh :)
20:57 jnettle 20:57 * jnettlet is very susceptible to peer pressure
21:02 taz_ hi, someone have knowledge about building uImage from source ?
21:36 taz_ rabeeh: are you here this evening ?
22:24 jas-hacks_ rabeeh: Any plans for C1 to have imx6q som?
22:29 dv_ jas-hacks_: btw I pushed several significant bugfixes to gstreamer-imx
22:30 jas-hacks_ dv_: oh cool, I will test them out, great work your doing on gstreamer !
22:36 jas-hacks_ dv_: what board are you testing on?
22:48 dv_ jas-hacks_: carrier one
22:48 dv_ before I had that, I tested on a sabre SD duallite whenever I could
22:49 jas-hacks_ dv_: playback is fine on C1?
22:50 dv_ I will soon introduce a new parameter to adjust the number of allocated frames. unfortunately this parameter is necessary to choose between more frames (= higher DMA memory usage) and less frames (= potential for choppy playback due to necessary memcpy calls)
22:50 dv_ its a limitation in the VPU wrapper and apparently also in the VPU itself
22:50 dv_ jas-hacks_: yes
22:51 jas-hacks_ dv_: I was testing on custom imx6q board
22:51 dv_ encoding too
22:51 dv_ well the problems you saw could have been caused by the bugs I fixed
22:51 jas-hacks_ dv_: only decoding
22:51 dv_ no I mean, encoding also works well on the C1
22:51 jas-hacks_ ah, good
22:51 dv_ rabeeh: btw I think the transcoding bottleneck is the DMA transfer
22:52 dv_ high vs baseline profile of the source material makes very little difference, leading me to believe that its not computation that is the limiting factor here
22:53 dv_ perhaps the uncached/writecombine allocations with the CMA will help
22:54 jas-hacks_ I thought some of the limitation was because of the DDR bus rate
22:54 dv_ yeah thats probably what limits dma
22:55 dv_ I have a few ideas for potential speed improvements though . will follow up on these in the next few days.
22:56 jas-hacks_ dv_: Is sound working?
22:56 dv_ I did not test sound yet on the C1. but I do see an alsa device.
23:50 _rmk 23:50 * _rmk_ wonders if jnettlet enjoyed the film :)