07:51 | tm512 | Hi, will the mainline linux kernel (3.16) work with the cubox/hummingboard? |
07:52 | tm512 | I see in the source tree that there is support for i.MX6 but I know that doesn't mean that full support is in |
08:17 | jnettlet | tm512, it has basic support but you will need some additional patches from rmk to get it fully functional. |
08:17 | jnettlet | what functionality are you looking for? |
08:18 | tm512 | I'd definitely want graphics support, and I want to use the etna_viv driver for 3D accel |
08:19 | tm512 | audio support through the 3.5mm jack would be important as well |
08:19 | jnettlet | I am not sure if audio is supported upstream. |
08:19 | jnettlet | also the etnaviv driver will only provide 3d acceleration on the framebuffer currently. |
08:20 | jnettlet | Xorg and DRM drivers are still a work in progress. We are just getting close to an initial alpha release. |
08:21 | tm512 | and that applies to all of the hummingboards and cuboxes? |
08:22 | jnettlet | what applies to them? |
08:23 | tm512 | only having 3D acceleration in the fb |
08:23 | jnettlet | for etnaviv yes. |
08:23 | jnettlet | if you use the binary drivers then you can have either accelerated Xorg or Wayland as well |
08:24 | tm512 | yea, not gonna do that though |
08:24 | tm512 | that was the case with the cubieboard I just sold today |
08:25 | jnettlet | well we are working hard on them, but graphics driver development takes time. |
08:26 | tm512 | I understand |
08:26 | tm512 | what about 2D acceleration inside of X? |
08:27 | jnettlet | same, we are just getting a very early alpha driver for 2d support with etnaviv |
08:28 | jnettlet | really I would recommend using the binary drivers until we make some more progress |
08:28 | jnettlet | if you want any sort of stability the OSS drivers are just not going to be there yet. |
08:28 | tm512 | but if the application got an EGL context, it could render accelerated into the framebuffer? |
08:29 | tm512 | i.e. fullscreen |
08:30 | jnettlet | yes, that would work for etnaviv on the framebuffer device. |
08:30 | tm512 | so SDL applications could be accelerated? |
08:31 | jnettlet | in theory yes, but again we are still talking alpha level stuff. FB might be closer to beta but still very much not something to be considered stable |
08:33 | tm512 | yea, I know |
08:34 | tm512 | I'm still debating whether to buy one or not |
08:34 | tm512 | the cubieboard ended up sitting in a closet for all but two weeks that I owned it |
08:35 | jnettlet | well it is a great little SOC and has some of the best community support you are going to find. |
08:35 | jnettlet | I guess it depends what you want to use it for. |
08:35 | tm512 | what I did on the cubieboard was set up Linux From Scratch |
08:35 | tm512 | using the musl libc, and busybox for the userland |
08:36 | jnettlet | well that isn't recommended. musl has some serious problems on ARM |
08:36 | tm512 | not in my experience |
08:36 | jnettlet | unless you are a core developer I would recommend against that. |
08:37 | tm512 | it just became really unweildy, gettext wouldn't generate .po files or something like that |
08:38 | tm512 | and pkgsrc failed to install the packages that involved gettext, I'd have to patch the file list, etc |
08:39 | tm512 | http://i.imgur.com/knCcrdF.png it worked though |
08:41 | jnettlet | well we have all sorts of XBMC distros, ARCH, Debian, Fedora, Android running the CBi/HB hardware |
08:42 | tm512 | my idea is to set up something similar with a hummingboard, using LFS/musl |
08:42 | tm512 | with the etna_viv driver though |
08:44 | tm512 | I guess there aren't any similarly priced SBCs with a Qualcomm SoC, are there? |
08:45 | tm512 | dunno if freedreno has an X driver |
08:51 | jnettlet | yes, Rob Clark has written one. |
08:54 | tm512 | I did find this, but that's pretty ridiculous for something that might just end up being a paperweight after I'm done playing with it http://www.inforcecomputing.com/product/moreinfo/ifc6410.html |
15:20 | jmontleon | jnettlet, any idea why I wouldn't get working usb on the hummingboard with your kernel? |
15:22 | jmontleon | i suspect these are part of the problem, but not sure: |
15:22 | jmontleon | [ 2.705318] usb_otg_vbus: disabling |
15:22 | jmontleon | [ 2.708958] usb_h1_vbus: disabling |
15:25 | jnettlet | jmontleon, what combo? |
15:25 | jmontleon | quad and hummingboard pro |
15:25 | jnettlet | let me take a quick look. I actually haven't had time to try my quad on an HBp |
15:26 | iCEBrkr | I'm hoping to replace my abandonded Sony Google TV with this cubox |
15:26 | iCEBrkr | Pfft. Honeycomb. :( |
15:27 | metrem | hi all, I just received my HummingBoard-i2eX, and I'd like to know if you can install debian wheezy armhf with video drivers that work? |
15:28 | jmontleon | jnettlet, thanks. fwiw, config I am (currently) using is: https://github.com/jmontleon/fedora-20-cubox-i_hb/blob/3.14.14/.config not much different than the defconfig you pointed me at. GZIP instead of LZO compression, cmment at top for fedora build, jump_labels off (was causing oops'es in conjunction with CONFIG_DEBUG_SET_MODULE_RONX when adding other stuff - guess this is a known thing, can point to a Fedora bug with more details) and CONFIG_LSM_ |
15:28 | jmontleon | MMAP_MIN_ADDR=32768 so selinux doesn't go batty |
15:28 | jmontleon | think that's about the sum of the differences |
15:29 | jnettlet | oh man lzo decompresses so much faster :-) |
15:30 | jmontleon | jnettlet, ya - I just need to add the lzop package buildrequire to the rpm spec and I can use it |
15:30 | jmontleon | was just me being lazy, but I had it in my head to fix that later |
15:31 | jmontleo | 15:31 * jmontleon saw some comparisons on size/speed and LZO looked to win the speed contest |
15:31 | jnettlet | jmontleon, sure. could you enable debug on the kernel commandline and then send the entire dmesg output? |
15:31 | jmontleon | sure |
15:31 | jnettlet | thanks. |
15:36 | jmontleo | 15:36 * jmontleon wonders if this is something else |
15:38 | metrem | some people have install http://gwolf.org/node/3896 image ? i need some help for install etna_viv (open source video driver) |
15:47 | jmontleon | jnettlet, http://paste.fedoraproject.org/126689/40845605/ |
15:52 | jnettlet | jmontleon, yeah those two vbus's getting disabled are your problem. Now did you boot up with a usb device plugged in? |
15:53 | jmontleon | just a keyboard, but it seems no better booting without it |
15:58 | jnettlet | have you tried both ports? |
16:00 | jnettlet | jmontleon, oh I see the problem. You are missing the USB drivers. CONFIG_USB_CHIPIDEA=y |
16:00 | jnettlet | CONFIG_USB_CHIPIDEA_UDC=y |
16:00 | jnettlet | CONFIG_USB_CHIPIDEA_HOST=y |
16:01 | jmontleon | uggh |
16:01 | jmontleon | i think I removed those because I thought they were not building right |
16:01 | jmontleon | let me try again |
16:02 | jmontleon | i probably only thought they weren't; there were a couple others making a mess and I probably wrongly accused it |
16:06 | cbxbiker61 | how is that sdma-imx6q.bin supposed to load with 3.14? |
16:07 | cbxbiker61 | my system log says it's not loading |
16:09 | jnettlet | cbxbiker61, ah I think the dts location is being used. I will need to switch that. |
16:10 | jnettlet | edit arch/arm/boot/dts/imx6qdl.dtsi and change line 764 |
16:11 | cbxbiker61 | so maybe put lib/modules in front? |
16:12 | cbxbiker61 | or do we use absolute path |
16:14 | cbxbiker61 | i guess the fsl-imx-sdma.txt seems to indicate full path |
16:16 | jnettlet | cbxbiker61, that path is relative to whatever is specified as your firmware directory |
16:16 | jnettlet | so if you used /firmware then the path would be /firmware/imx/sdma/ |
16:17 | cbxbiker61 | that's what i would think |
16:50 | jnettlet | cbxbiker61, is that the only problem you have had so far? |
16:50 | cbxbiker61 | hdmi audio is not working, i think it's related |
16:51 | cbxbiker61 | i'm going to use config_extra_firmware |
16:51 | jnettlet | cbxbiker61, yes. The fsl audio driver requires the sdma firmware |
17:06 | jmontleon | it's amazing what happens when you have the correct drivers |
18:25 | brantje | Anyone knows how to force 720p output? |
18:25 | brantje | On xbian |
18:27 | brantje | mk01? |
18:27 | curlymo | check the xbian wiki |
18:27 | brantje | Same as rpi? |
18:39 | brantje | I see the bootscreen for abit |
18:39 | brantje | then the screen is black |
18:41 | curlymo | have you checked the xbian wiki? |
18:41 | brantje | Yea |
18:41 | brantje | I did |
18:41 | curlymo | what did you change? |
18:42 | brantje | Changed it to setenv baseconfig ahci_imx.hotplug=1 raid=noautodetect telnet root=/dev/mmcblk0p2 video=mxcfb0:dev=hdmi,1280x720@60,if=RGB565 rootwait consoleblank=0 selinux=0 nohdparm splash |
18:42 | curlymo | but didn't work? |
18:42 | brantje | I see i see the cubox bootscreen for a second |
18:42 | brantje | - isee |
18:43 | curlymo | are you using HDMI video or a HDMI --> XXX adapter? |
18:44 | brantje | HDMI -> VGA |
18:45 | curlymo | then thinker a little with the resolution and RGB value |
18:45 | brantje | What are the possible values for RGB? |
18:45 | curlymo | 1280x960 maybe? |
18:46 | rdvn | I use HDMI -> DVI and have same problem |
18:52 | brantje | looks like it doesn't use my settings |
18:52 | curlymo | did you update the boot.src? |
18:55 | brantje | how? |
18:55 | brantje | i changed the .txt |
18:55 | brantje | and then rebooted |
18:55 | brantje | oh i gfound |
18:55 | brantje | sorry |
18:56 | rabeeh | brantje: i'm using an alternative way now with xbian |
18:56 | rabeeh | i removed the boot.scr and using a text file called uEnv.txt |
18:56 | rabeeh | http://pastebin.com/X3ZpQJ1m |
18:56 | rabeeh | as an example |
18:57 | rabeeh | it's not complete as xbian boot.scr but you get the idea |
18:57 | brantje | I try this first |
18:57 | kgp | Rabeeh! It's late :) |
18:57 | rabeeh | the difference is that you don't need to mkimage the boot.scr.txt |
18:57 | brantje | http://boundarydevices.com/configuring-i-mx6-machines-different-screens-nitrogen6x-sabre-lite/ |
18:57 | brantje | https://community.freescale.com/thread/304822 |
18:57 | kgp | Did you commit your script? |
18:57 | brantje | whee |
18:57 | brantje | it works |
18:58 | rabeeh | kgp: :) |
18:58 | rabeeh | damn xorg |
18:58 | rabeeh | mk01 might be up to something though |
18:58 | rabeeh | jnettlet: curlymo reported a bug on the gpio's |
18:59 | rabeeh | i have a fix and will be sending that to you (stupid me) |
19:00 | jnettlet | okay |
19:00 | brantje | lol... ERROR: The certificate of 'github.com' is not trusted. |
19:00 | brantje | ERROR: The certificate of 'github.com' hasn't got a known issuer. |
19:00 | rabeeh | jnettlet: any issue having the ahci controller hotplug on by default? |
19:01 | rabeeh | brantje: update your date/time :) |
19:01 | brantje | xbian@cubox /tmp $ date |
19:01 | brantje | Tue Aug 19 19:01:43 CEST 2014 |
19:01 | brantje | Looks good |
19:01 | rabeeh | which xorg? |
19:02 | rabeeh | 1.16? |
19:02 | brantje | xbian |
19:02 | jnettlet | rabeeh, doesn't bother me. It is mostly about saving power. It is off by default because that is what FSL committed upstream. |
19:09 | brantje | rabeeh, my time is correct, but i still get the error |
19:10 | curlymo | apt-get install ca-certificates? |
19:18 | brantje | ty |
19:20 | brantje | O_o |
19:20 | brantje | xbian@cubox /dev $ ls | grep lirc |
19:20 | brantje | xbian@cubox /dev $ |
19:20 | brantje | No lirc? |
19:24 | rabeeh | modprobe ir_lirc_codec |
19:24 | rabeeh | ? |
19:24 | rabeeh | brantje: i think you are not using jnettlet's defconfig as-is |
19:24 | jnettlet | or just the base_config |
19:25 | brantje | Well i have an MCE remote |
19:25 | brantje | xbian@cubox /etc/lirc $ sudo modprobe ir_lirc_codec |
19:25 | brantje | > no results |
19:25 | brantje | xbian@cubox /dev $ ls | grep lirc |
19:25 | brantje | lirc0 |
19:25 | brantje | lircd |
19:25 | curlymo | uname -r? |
19:27 | brantje | xbian@cubox ~ $ uname -r |
19:27 | brantje | 3.10.30-g63976aa-dirty |
19:27 | curlymo | gpio doesn't work on that version |
19:27 | brantje | Now doing an apt-get upgrade |
19:28 | brantje | http://xbian.brantje.com/images/xbian-image-imx6-20140818.img.xz -> this one does? |
19:28 | curlymo | just don't use XBian on the HB yet. Won't work well if you don't CEC/HDMI |
19:28 | curlymo | no |
19:28 | brantje | Hummingburd? |
19:28 | brantje | I have cubox-i4pro |
19:28 | curlymo | probably the same |
19:29 | curlymo | I only have hummingboards |
19:29 | curlymo | but IR is dependent on a working GPIO configuration and i'm sure xbian doesn't provide a kernel with working GPIO support |
19:29 | brantje | InfraRed Remote Controls tested with CuBox integrated IrDA reciever |
19:29 | brantje | This only lists infra-red remote controls that are compatible with CuBox integrated IrDA reciever. |
19:29 | brantje | Syabas Popcorn Hour A-100 Remote Control works with LIRC 0.9.0 and Linux Kernel 2.6.32 on Ubuntu 10.04 |
19:29 | brantje | Microsoft MCE Remote Control works with LIRC 0.9.0 and Linux Kernel 3.5.7 on Debian 7.0 |
19:29 | brantje | So it works on debian 7.0 |
19:30 | curlymo | Be patient |
19:30 | curlymo | we're working on it |
19:30 | brantje | Ok thanks =) |
19:35 | quitte | is there a target that creates an image of ubi volumes containing kernel/squashfs/someoverlay ? |
19:44 | quitte | oops wrong channel |
19:58 | jmontleon | think I should have a good and working 3.14.14 kernel for fedora coming up in a few minutes |
19:59 | jnettlet | jmontleon, great pass me a link when it is ready |
19:59 | jmontleon | sure thing |
20:28 | sam_nazarko | brantje curlymo you guys should just use osmc ;) |
20:36 | malte | @sam_nazarko: Open Source Monitoring Conference ? |
20:44 | anaconda | except it doesn't exist yet |
20:52 | sam_nazarko | it does, and test builds are available |
20:54 | brantje | load average: 1.30, 1.42, 0.86 > is this normal on xbian? |
21:08 | brantje | load average: 1.16, 1.34, 1.16 -> without xbmc open |
21:09 | brantje | I think there is something wrong? |
21:13 | jmontleon | sound is not working.. firmware is loading but when it goes to play things just start to hang |
21:13 | jmontleo | 21:13 * jmontleon guesses maybe pulseaudio needs some beatings |
21:19 | jnettlet | jmontleon, we were afraid that would be the case. |
21:20 | jnettlet | PA runs the buffers at near empty and the FSL drivers expects them to be almost full |
21:26 | iCEBrkr | This is the most productive IRC channel I've witnessed. |
22:43 | jmontleon | jnettlet, http://people.redhat.com/jmontleo/cubox-i_hb/rpms/test/ - there's a repo file ready to go at http://people.redhat.com/jmontleo/cubox-i_hb/cubox-i_hb.repo - just enable test and should be able to install |
22:44 | jmontleon | no image with it yet, but if they're at least booting should be easy enough to install |