| 15:24 | ralix | morning |
| 15:24 | ralix | echo ======== Setting bootargs ======== |
| 15:24 | ralix | setenv bootargs 'console=ttyS0,115200n8 console=tty1,115200n8 vmalloc=384M root=/dev/sda1 rw video=dovefb:lcd0:1280x720-32@60-edid clcd.lcd0_enable=1 clcd.lcd1_enable=0 usb0Mode=host usb1Mode=host' |
| 15:24 | ralix | echo ======== Loading kernel ======== |
| 15:24 | ralix | ext2load ide 0:1 0x00200000 /boot/uImage |
| 15:24 | ralix | echo ======== Booting kernel ======== |
| 15:24 | ralix | bootm |
| 15:24 | ralix | ups sorry |
| 15:25 | ralix | I'm just trying archlinux only on a sata drive to install. However, I am having difficulty with the boot.src. On the console, he remains in stand. " |
| 15:25 | ralix | ** Unable to read "/ boot.scr" from ide 0:1 ** |
| 15:25 | ralix | ===> Executing ext4load ide 0:1 0x02000000 / boot / boot.scr |
| 15:25 | ralix | Loading file "/ boot / boot.scr" from ide device 0:1 hda1 |
| 15:25 | ralix | 434 bytesRead |
| 15:25 | ralix | # # Executing script at 02000000 |
| 15:25 | ralix | Setting bootargs ======== ======== |
| 15:25 | ralix | Loading kernel ======== ======== |
| 15:25 | ralix | Loading file "/ boot / uImage" from ide device 0:1 (hda1) " |
| 16:11 | rabeeh | ralix: why do you get a space between '/' and 'boot'? |
| 16:12 | ralix | only from copy |
| 16:17 | ralix | ok i boot from sd and my root ist sda1 (esata), its ok for me |
| 18:46 | lioka | hi |
| 18:47 | lioka | cubox, self-compiled xbmc. black box, logs are full of 'ERROR: GL: Error compiling vertex shader' -- rings any bell ? |
| 18:47 | lioka | and ERROR: GUI Shader [guishader_frag_default.glsl] - Initialise failed |
| 18:49 | rabeeh | lioka: please look at the wiki for more information how to build xbmc |
| 18:49 | rabeeh | i think your main issue is that you building it with GL graphic acceleration hardware instead of GLES. |
| 18:51 | lioka | rabeeh: no, with GLES/EGL, resulting binary is linked with those. |
| 18:51 | rabeeh | which distro are you using? |
| 18:52 | rabeeh | http://www.blackbox.com/ ? |
| 18:52 | lioka | rabeeh: altlinux. infact, i'm the xbmc maintainer :] probably not very good tho |
| 18:53 | rabeeh | hehe. xbmc is a tough nut to break |
| 18:53 | rabeeh | lots of tweaks. |
| 18:53 | rabeeh | i think for today that most advanced xbmc port for CuBox comes for Arch, GeexBox and XilkaX (in alphabetic order) |
| 18:53 | rabeeh | have you looked at any of those? |
| 18:53 | rabeeh | are you running Eden or Frodo? |
| 18:54 | lioka | rabeeh: xbmc is configured with --enable-gles and --with-platform=marvell-dove |
| 18:54 | lioka | rabeeh: eden, yes |
| 18:57 | rabeeh | lioka: do you have X and other stuff up and running? do you have some sort of rootfs.tar with a kernel? |
| 18:57 | lioka | rabeeh: i'm using patches from github.com/huceke/xbmc, if that matters |
| 18:57 | rabeeh | i'm asking since there is the CuBox installer that is a centralized place where distros goes in for easy installation on top of CuBox. |
| 18:57 | lioka | rabeeh: well, we are hardfp and self-hosting |
| 18:58 | rabeeh | lioka: those patches are ok to start with; they should give you a working xbmc but not best performance |
| 18:59 | lioka | rabeeh: so yes, we sorta have rootfs, about 2000+ packages as for now |
| 18:59 | rabeeh | lioka: all i need is a link then a small script like the following - |
| 18:59 | rabeeh | http://research.microsoft.com/apps/pubs/default.aspx?id=102069 |
| 18:59 | rabeeh | oh; sorry. |
| 18:59 | rabeeh | wrong link |
| 18:59 | rabeeh | http://dl.dropbox.com/u/72661517/CuBox/installer/XilkaX/install-XilkaX |
| 18:59 | lioka | indeed |
| 19:00 | rabeeh | :) |
| 19:00 | rabeeh | that it is a small add-on script for the installer that gets XilkaX specific stuff installed. |
| 19:00 | rabeeh | the installer already takes care of the main UI (will be HDMI based soon); and already takes care of partitioing and getting distros list from the web. |
| 19:01 | rabeeh | lioka: are you using the hardfp GPU drivers? |
| 19:02 | rabeeh | http://download.solid-run.com/pub/solidrun/cubox/packages/marvell-opengl/ |
| 19:02 | lioka | rabeeh: as of X and other stuff -- yes, xorg dovefb, hardfp drivers, 3.5.7 kernel from your git and and and |
| 19:03 | lioka | at least i can use xfce |
| 19:04 | lioka | and xbmc log contains 'NOTICE: GL_VENDOR = Marvell Technology Group Ltd |
| 19:04 | lioka | and so on |
| 19:18 | rabeeh | try vDorst kernel |
| 19:18 | rabeeh | mine is outdated |
| 19:19 | lioka | OH SH*T. found what was wrong |
| 19:19 | rabeeh | there is a kernel patch for getting the r3184 driver |
| 19:19 | rabeeh | GPU driver |
| 19:19 | rabeeh | what? |
| 19:19 | lioka | libGLESv2SC.so was missing |
| 19:19 | rabeeh | oh |
| 19:19 | rabeeh | and? |
| 19:19 | lioka | and at least it shows menus |
| 19:19 | rabeeh | are you using pulseaudio? |
| 19:20 | rabeeh | good :) |
| 19:20 | lioka | dunno yet :] it is first time i'm seeing something from xbmc |
| 19:20 | rabeeh | for hardfp vmeta, you need libmiscgen.so - |
| 19:21 | rabeeh | http://download.solid-run.com/pub/solidrun/cubox/packages/marvell-vmeta/ |
| 19:21 | lioka | rabeeh: yeah, already have, thanks |
| 19:21 | rabeeh | i suggest you build it from scratch |
| 19:21 | rabeeh | oh; one notice for you (and all - jnettlet) Russel King had first attempt to upstream the vmeta uio driver and was rejected because of ioctl |
| 19:22 | rabeeh | he is rewriting that tiny kernel driver to be char based driver; so probably in some point of time that will be upstreamed and libvmeta.so needs to be ported to support the char device ioctls instead of the uio |
| 19:23 | rabeeh | probably will be mainly porting the pollin function and get the vmeta interrupt indication via ioctl instead. |
| 19:23 | jnettlet | rabeeh, oh I rewrote the driver to use sysfs nodes |
| 19:24 | jnettlet | then I updated the libvmeta to support it |
| 19:24 | jnettlet | that was upstreams recommendation when Marvell first tried to upstream it. |
| 19:26 | jnettlet | But ultimately I am holding off upstreaming it as I want to convert it to use dmabuf instead of pmem. Marvell is working on switching to ION (I already wrote an adaption of that as well). |
| 19:27 | jnettlet | I was at the Linaro Connect conference today and was discussing dma-buff and UMM in terms of Marvell with the Linaro guys doing the omap and whatever is driving the new chromebook drivers. |
| 19:27 | rabeeh | aargh |
| 19:27 | jnettlet | I am hoping to get some hacking time in with them later this week to push my work along where I am having some problems. |
| 19:27 | lioka | is ~15 fps at fullhd is normal for xbmc gui (not video yet) or i'm screwed again somewhere ? |
| 19:28 | rabeeh | lioka: yes |
| 19:28 | rabeeh | jnettlet: i was under impression that rmk was in sync with you |
| 19:29 | jnettlet | rabeeh, nah I haven't heard from him in weeks. We sent him an XO-4 but don't know what he is doing with it |
| 19:31 | jnettlet | rabeeh, you guys are still using bmm right? |
| 19:36 | N30N | jnettlet: yep, we are. |
| 19:38 | jnettlet | N30N, that is what I thought thanks. |
| 19:38 | rabeeh | jnettlet: so the strategy should be moving bmm to dma-buf? |
| 19:40 | jnettlet | rabeeh, that is what I am planning to do for out implementation. |
| 19:41 | jnettlet | it doesn't matter for you guys, but for us that will allow camera buffer to be captured and sent to graphics engine to display, and vmeta for encoding all without using the CPU |
| 19:42 | jnettlet | but ultimately this is needed to switch to using the Unified Memory Manager which is the holy grail of graphics memory on ARM |