| 05:21 | mk01 | guten morgen |
| 05:27 | jnettlet | god morgen |
| 10:07 | matoking | Hey, how can I force the Cubox-i to boot into shell instead of Xfce desktop environment I have installed? |
| 10:09 | mk01 | motoking: init=/bin/bash |
| 10:10 | mk01 | as kernel command line parameter |
| 10:11 | matoking | Yeah, into uEnv.txt |
| 10:11 | mk01 | yes |
| 10:12 | mk01 | you should get single user mode - so uid=0 and no password check |
| 10:12 | matoking | Let's hope this works |
| 10:13 | mk01 | 5$ it will |
| 10:13 | matoking | The kernel I'm currently trying out doesn't have the wireless driver installed, and in order to enable it I actually have to see what's on the screen :) |
| 10:14 | matoking | Yeah, it works |
| 10:14 | mk01 | matoking: then maybe better solution would be to just disable Xfce, then reboot without init=/bin/bash |
| 10:14 | matoking | I didn't expect it to boot into bash this early in the boot process |
| 10:15 | matoking | But this should do |
| 10:15 | mk01 | because wlan0 will need to load some processes (wpa_supplicant) |
| 10:15 | mk01 | which can have troubles running on S |
| 10:15 | mk01 | you should get bash in 1-2 s |
| 10:17 | mk01 | as the bash is loaded instead standard init daemon |
| 12:49 | matoking | Okay, I need to prevent lightdm from starting up somehow |
| 12:50 | matoking | But I can't access the internet and some commands don't work probably because the bash shell is before most of the necessary utilities and such are started |
| 12:50 | matoking | dpkg for example complains about PATH |
| 12:50 | mk01 | I told you] |
| 12:50 | mk01 | what system it is |
| 12:50 | rabeeh | maybe run in different runlevel? |
| 12:51 | mk01 | yep. rabeeh is right |
| 12:51 | mk01 | just issue |
| 12:51 | mk01 | runlevel 1 |
| 12:51 | mk01 | :) |
| 12:51 | obinou | Hello, did someone tried tvheadend recently on cubox-i ? |
| 12:52 | obinou | I use the package xbian-package-tvheadend-6q |
| 12:52 | obinou | but I have issues |
| 12:52 | mk01 | obinou: which one ? |
| 12:52 | obinou | (tvheadend v3.9.431~g3f3fdc8 ) |
| 12:52 | obinou | Well, the program seems to hang |
| 12:52 | obinou | for no reason |
| 12:53 | obinou | tried with 2 different USB-DVB |
| 12:53 | mk01 | ok |
| 12:54 | obinou | For now , It seems to have completed the initial scan (found the correct number of channels) , but hang when it creates the channel mapping |
| 12:54 | obinou | I can kill it , but then it redo the whole scan at stratup |
| 12:54 | obinou | the other USB-DVB-t key works better |
| 12:55 | obinou | but trying to put any channel on fails with out of order frames |
| 12:55 | mk01 | there is new branch at tvheadend, do you know anything about ? |
| 12:55 | mk01 | origin/feature/network-scan |
| 12:56 | obinou | mhhh non but I was thinking about compiling it myself |
| 12:56 | obinou | (i've tested both the usb keys on my main linux PC & it works |
| 12:58 | mk01 | obinou: I set tv=headend for refresh & recompile |
| 12:58 | mk01 | in 20-30 minutes try new |
| 12:58 | mk01 | what kernel you use ? |
| 12:59 | obinou | root@xbian-cubox:~# uname -a |
| 12:59 | obinou | Linux xbian-cubox 3.10.30-g63976aa-dirty #449 SMP Mon May 26 06:05:19 CEST 2014 armv7l GNU/Linux |
| 12:59 | mk01 | ok |
| 12:59 | obinou | I wonder if I shouldn't rather recompile the whole xbian-6q |
| 13:00 | obinou | To have better control on everything |
| 13:00 | mk01 | there is not so much to recompile as system Debian |
| 13:00 | mk01 | in that case kernel |
| 13:00 | mk01 | + tvheadend |
| 13:01 | mk01 | tvheadend from git with no mods |
| 13:01 | mk01 | kernel from repo |
| 13:01 | mk01 | config = default + drivers media |
| 13:01 | obinou | ok |
| 13:02 | mk01 | even its on github -> xbian-patches including kernel .config |
| 13:03 | mk01 | I use the same source (tvheadend) without problems |
| 13:03 | mk01 | on x64 3.14 |
| 13:03 | mk01 | with Jessie as well |
| 13:03 | mk01 | so no diff on libs, |
| 13:03 | mk01 | only kernel |
| 13:04 | mk01 | and the same source compiled for RPI as RPI TVheadend package |
| 13:04 | mk01 | and there was no one freeze report |
| 13:05 | mk01 | that means wait 20m |
| 13:05 | mk01 | test |
| 13:05 | mk01 | then I can put my DVBT to cubox and check |
| 13:05 | mk01 | then do whatever you want |
| 13:06 | mk01 | just one question |
| 13:07 | mk01 | can you post "ldd /usr/local/bin/tvheadend" ? |
| 13:12 | matoking | Okay I finally managed to stop the lightdm from starting up by modifying its init script itself |
| 13:12 | matoking | Not an elegant solution, but it worked |
| 13:13 | mk01 | if you have initscript (/etc/init.d) then "update-rc.d lightdm -f remove" |
| 13:13 | jnettlet | if you have systemd then change the default target to multi-user.target |
| 13:14 | matoking | Oh right |
| 13:16 | matoking | Alright, now I'm getting close! |
| 13:17 | mk01 | obinou: http://xbian.brantje.com/pool/staging/main/x/xbian-package-tvheadend-6q/xbian-package-tvheadend-6q_1.0.1_armhf.deb |
| 13:20 | obinou | mk01: thanks i'll try that today if I can |
| 13:21 | mk01 | obinou: check if by accident there is no missing lib dependency |
| 13:22 | obinou | mk01: my current one (/usr/bin/tvheadend) misses no libraries |
| 13:38 | mk01 | obinou: fine |
| 13:40 | matoking | I need to run a command before the "Configuring network interfaces..." part in the bootup process |
| 13:41 | matoking | Or that could be done instantly after the kernel module (rtl8192cu) has loaded |
| 13:42 | mk01 | there is /etc/network/if-pre-up.d |
| 13:42 | matoking | Oh sweet, I think that's perfect |
| 13:42 | mk01 | any script (file with +x) being there will be started before "if up XXX is called) |
| 13:44 | mk01 | variables from environment are : IFACE = INTERFACE, $METHOD, $LOGICAL, $ADDRFAM |
| 13:46 | matoking | Gah, that didn't help |
| 13:47 | matoking | Okay, I have an ID for an USB WiFi dongle |
| 13:47 | matoking | It uses rtl8192cu chipset which is supported by Linux, but because the device identifier isn't known by the kernel I add it using a command |
| 13:47 | gabriel_ | Any news concerning delievery of CuBox-i? The delay is still 6 weeks? |
| 13:48 | matoking | The dongle only starts working when the login prompt pops up |
| 13:48 | matoking | I need it to be detected and started before the network interfaces are configured |
| 13:49 | matoking | I think running lsusb should force the USB devices to be detected |
| 13:50 | mk01 | motoking: is only SYSV available ? |
| 13:50 | mk01 | so no upstart/systemd ? |
| 13:50 | mk01 | then create rcS init script |
| 13:51 | matoking | Or maybe the script is not being executed because it's missing the shebang |
| 13:53 | ido___ | hi everyone, does anyone know if the newly released Gotham for android runs smoothly on Cubox-i and manages to run HD movies? |
| 13:54 | mk01 | but if-pre-up.d has to work |
| 13:54 | matoking | mk01: I created a script called 8192custart.sh inside it |
| 13:55 | mk01 | i'm not sure of the "." |
| 13:55 | mk01 | in name |
| 13:55 | matoking | It runs "echo "2001 330D" | sudo tee /sys/bus/usb/drivers/rtl8192cu/new_id" and after it runs "lsusb" |
| 13:55 | matoking | Well, I'll take it out and give it a try again |
| 13:55 | mk01 | check "man run-parts" |
| 13:55 | mk01 | as through it are "directories" started/processed |
| 13:56 | mk01 | you can test with |
| 13:56 | mk01 | run-parts --list /etc/network/if-pre-up.d |
| 13:57 | mk01 | if it gets listed, will be started |
| 13:57 | matoking | mk01: Thanks, removing .sh from the end made it work! |
| 13:58 | mk01 | perfect |
| 13:58 | mk01 | this is "clean" solution which will last over upgrades |
| 13:59 | matoking | Now, let's see if I can ssh to the Cubox |
| 13:59 | mk01 | as other hacks could be overwriten by new system files |
| 13:59 | mk01 | matoking: haha, I though it's already working |
| 13:59 | mk01 | :) |
| 14:00 | matoking | Yeah, I started seeing DHCPDISCOVER messages so I automatically assumed it's working now |
| 14:01 | matoking | YES, I JACKED IN |
| 14:01 | matoking | The local IP address was changed |
| 14:02 | matoking | That gave me a good scare for a minute there :) |
| 14:03 | matoking | Now, if my stupid router would allow me to change the assigned IP address :/ |
| 14:05 | matoking | I can't edit the assigned IP if the device is online |
| 14:05 | matoking | And if I'm reading this right, the device is considered to be online if it has connected at least once during the last decade |
| 14:06 | mk01 | why and what IP doesn't match ? |
| 14:06 | matoking | Well the local IP for the device is 192.168.1.40 |
| 14:06 | matoking | I'd rather not go through the trouble of changing it to 192.168.1.37 everywhere |
| 14:07 | matoking | And this ZyXEL router I have is filled with bizarre bugs |
| 14:07 | matoking | The maximum limit of port forwarding rules is 10 and if you reach that, you suddenly can't delete the older rules! |
| 14:07 | matoking | Needless to say I've already bought a new router, just have to wait till it gets here :P |
| 14:08 | mk01 | matoking: create dhcp reservation for the HW addr of the adapter |
| 14:12 | matoking | Okay I restarted the router and managed to change the IP |
| 14:12 | matoking | Let's try out the download speed performance |
| 14:13 | matoking | Woah |
| 14:13 | matoking | 2 MB/s |
| 14:14 | matoking | With the integrated chip it would crawl down to as low as 100 KB/s |
| 14:22 | matoking | But yeah this just solved pretty much the only problem I had with Cubox-i :) |
| 14:23 | jnettlet | mine doesn't make me coffee yet. that is my main problem with it |
| 14:25 | obinou | mk01: I tried your version 1.0.1 , seems to be _much_ better |
| 15:45 | mk01 | obinou: but _much_better_ means still not absolutely 100% |
| 17:08 | obinou | mk01: well, I felt back to 720p on xbmc, and now my kid is watching his cartoon without complains |
| 17:09 | obinou | mk01: so yes, it's quite usable :-) Now the issue I have are much more related to the video driver (**** AXIS ERROR ***) at boot sometime |
| 17:15 | mk01_ | obinou: with the DVI patches before you could use non CEA the same way as CEA ? (so with fine audio etc)? this is what you mean? |
| 17:15 | obinou | CEA == ? |
| 17:16 | mk01_ | obinou CEA = HDMI. 1920x1080, 1280x720, ... |
| 17:16 | mk01_ | not 1366x768, 1440x1050 etc :) |
| 17:17 | obinou | ok |
| 17:17 | obinou | no I just switch xbmc from 1920x1080 to 1280x720 , |
| 17:18 | obinou | since in France , DVB-T signal are mainly in "SD" , it works well |
| 17:18 | obinou | For the audio I still have issues , but I guess my audio system is crappy |
| 17:19 | obinou | The numeric input (optic TOS-link / AC3) has some sync problems |
| 17:19 | obinou | but that's also the case with other input devices |
| 18:36 | mk01_ | obinou: and what is problem with SD content on 1920x1080 screen ? |
| 18:37 | obinou | mk01_: it works but it looks very bad |
| 18:37 | mk01_ | (assuming deinterlacing OFF as there is really bug somewhere) |
| 18:37 | obinou | lots of lag |
| 18:37 | mk01_ | yes, deinterlacing problem |
| 18:38 | mk01_ | it looks "public" is quiet about this, but I'm sure everyone has this |
| 18:38 | obinou | ok i'll try to disable interlace |
| 18:39 | obinou | tonight |
| 18:39 | mk01_ | and even the issue is tracked somewhere |
| 18:39 | mk01_ | very very long time ago |
| 18:40 | mk01_ | let's ask Rabeeh if he re-appear here |
| 18:41 | rabeeh | mk01_: here |
| 18:42 | mk01_ | rabeeh: is the de-interlacing issue tracked / worked on / or put on hold? |
| 18:42 | rabeeh | mk01_: wolfgar / koying were working on that on imx6-xbmc |
| 18:42 | bencoh | hmm, SD content should never be scaled to 1080p without proper deinterlace :> |
| 18:42 | rabeeh | they had performance issues on 1080p; unclear what exactly |
| 18:43 | bencoh | it's bound to look ugly this way ;) |
| 18:43 | mk01_ | obinou: so you have your answer. |
| 18:43 | rabeeh | i'm not sure if the de-interlacing function itself was laggy or the upscaling part (they discussed either using the gpu or pxp) |
| 18:43 | rabeeh | mpeg2 SD content? |
| 18:43 | rabeeh | it should be deinterlaced first and then upscaled :) |
| 18:44 | bencoh | yup |
| 18:44 | mk01_ | rabeeh: MPEG2 |
| 18:44 | mk01_ | interesting is my SD content upscaled is perfectly ok |
| 18:44 | mk01_ | bud HD is fail |
| 18:44 | rabeeh | mk01_: i'm guessing here; most of the SD interlaced stuff is either satellite or cables; which is mostly mpeg2 |
| 18:44 | mk01_ | 13-20 fps |
| 18:44 | bencoh | rabeeh: mostly, but can be h.264 as well |
| 18:45 | bencoh | (depending on where you are and who's your provider) |
| 18:45 | rabeeh | bencoh: sure; can be both |
| 18:45 | bencoh | I mean, in real life :) |
| 18:47 | mk01_ | life is funny cause people responsible for this standards and decisions are ... let's say not on the right place |
| 18:47 | mk01_ | at least in SVK and other countries around |
| 18:48 | bencoh | :-) |
| 18:49 | bencoh | has anyone seen "internal compiler error: in push_minipool_fix" when building for cubox btw ? |
| 18:49 | mk01_ | so finally we have MPEG2 at SD resolutions, taking 3mbps / channel. impossible to put HD content to multiplexes etc etc etc. then I rather switch to internet stream (of the same TV) with 400kbps bandwidth, double resolution and x264 |
| 18:49 | rabeeh | bencoh: native? |
| 18:49 | rabeeh | native build? |
| 18:49 | bencoh | nope, cross |
| 18:50 | rabeeh | mk01: but then you missed all the 10000 other channels that are broadcasted :) |
| 18:50 | bencoh | rabeeh: arm-oe-linux-gnueabi-gcc, v4.8.2 (openembedded) |
| 18:54 | rabeeh | bencoh: are you building for yocto? |
| 18:55 | rabeeh | i just did that and it built core-image-x11 without any build errors |
| 18:56 | bencoh | not for yocto ... oe-core+meta-oe+meta-fsl-arm{,-extra} this got triggered when building biTSream from git |
| 18:56 | bencoh | I had no major issue when building core-image-minimal |
| 18:56 | bencoh | (apart from usual OE issues ;) |
| 18:57 | bencoh | it looks like an old bug in gcc that as never been taken care of |
| 18:57 | bencoh | has* |
| 19:20 | obinou | mk01_: rabeeh : Is there a specific github repo to recompile xxbian from scratch for imx6q ? I found the main xbian one, but there is no branch for imx6 specifics |
| 19:21 | obinou | I need to add a few features (for exemple using the wlan as an AP in a bridge, things like that |
| 19:23 | obinou | (for what I could try, /etc/network/interfaces is changed at boot to force eth0 as DHCP , where I would like to make a bridge and include wlan2 & eth0 inside |
| 19:23 | mk01_ | obinou: I will make you sad - all imx6 specific repos are on my build machine only. ad1) we are only now creating autobuild & apt platform and ad2) nobody from team have cubox till now, so nobody pushed on me to push to git |
| 19:23 | obinou | ok |
| 19:23 | mk01_ | obinou: 90% xbian packages is common for RPI and IMX6 |
| 19:24 | obinou | and the 10% come from patches here and there ? |
| 19:24 | mk01_ | from rest 10% - half is sharing source code, only binaries are different (RPI being wheezy and IMX being Jessie - so different libs to link to) |
| 19:25 | obinou | mhhh ok |
| 19:25 | mk01_ | and the second half from this 10% are binaries which are bundled exactly like manufacturer is providing it |
| 19:25 | mk01_ | (RPI firmware, IMX firmware) |
| 19:26 | obinou | If I grep dpkg -l|grep 6q , I can see all the 6q thinks are xbmc . I guess the kernel is specific too, of cource |
| 19:26 | mk01_ | and the two completely different are on my git pushed (libcec & kernel) |
| 19:26 | mk01_ | as soon as we finish build system I will move to xbianonpi too |
| 19:26 | mk01_ | so |
| 19:27 | mk01_ | that means |
| 19:27 | mk01_ | just push to repositories which are there |
| 19:27 | mk01_ | the same is deployed onto IMX6 |
| 19:27 | obinou | the binaries are common for raspi & imx6, or you did recompiled everything for armv7-neon ? |
| 19:27 | mk01_ | ]in your case system hooks & init scripts |
| 19:28 | mk01_ | obinou: most things needs to be recompiled due to different base distro (as I told before) |
| 19:28 | obinou | ok |
| 19:29 | mk01_ | once RPI will be upgraded to Jessie, recompiled will be only the ones where it make sense |
| 19:29 | obinou | (But I assume jessie could work on raspi too, if recompiled) |
| 19:29 | obinou | but ok |
| 19:29 | mk01_ | YES |
| 19:30 | mk01_ | but we speak about 1 or 2 packages there |
| 19:31 | mk01_ | XBian is providing 15-20 packages to install on top of debian, from them libcec, kernel, xbmc ... is binary |
| 19:31 | obinou | well, thank you for all your work |
| 19:31 | mk01_ | vnc server |
| 19:31 | mk01_ | and that;s maybe all |
| 19:32 | mk01_ | all other functions / custoizations / tools are dash code, or python . |
| 19:33 | mk01_ | (of course by using other tools like dialog, or xbmc libs/classes for xbmc code) |
| 19:34 | mk01_ | if you want to extend interfaces configuration / auto generation of "interfaces" file, clone xbian-config-shell and xbian-config-xbmc |
| 19:35 | mk01_ | the first is dialog + bash, the latter python |
| 19:35 | obinou | i'm doing just that right now |
| 19:35 | obinou | thanks your this great piece of help |
| 19:35 | mk01_ | you are welcome |
| 19:36 | mk01_ | but are you sure APd works and bridging too ? |
| 19:36 | mk01_ | I'm not so sure. |
| 19:36 | mk01_ | specially for wlan |
| 19:36 | mk01_ | (on cubox) |
| 19:38 | mk01_ | last info -> init scripts & base system custom serices or changes to defaults are contained in xbian-update package |
| 19:38 | mk01_ | it contains almost no real code / scripts. just mods to plain debian system & configs |
| 19:41 | obinou | I'll look into the repo xbian-update / xbian-config-shell / xbien-config-xbmc as soon as they are cloned (I'm on a satellite link so it takes a while) . For the AP stuff, I can use the "iw" command |
| 19:41 | obinou | which looks ok |
| 19:41 | obinou | but yes, it's possible that the firmware forbid it |
| 19:41 | obinou | One thing I have to check |
| 19:42 | mk01_ | it should work |
| 19:42 | obinou | also one of my other goal is to make the distro work on GK802 & Sabre Light, which I own too - shall be only a kernel DTS file to load |
| 19:42 | mk01_ | but current driver in 3.10.30 is failing to set promics mode |
| 19:43 | mk01_ | the kernel masters are saying it is some kind of stupid bug, but wlan is currently not a priority |
| 19:43 | mk01_ | what about gfx driver libs ? |
| 19:44 | mk01_ | it's IMX6 too ? |
| 19:44 | obinou | yes |
| 19:44 | mk01_ | ah ok |
| 19:44 | obinou | all are imx6q |
| 19:44 | mk01_ | then yes |
| 19:45 | obinou | the only pb is on the GK802 there are some thermal issues that forces the usage of accelerated video |
| 19:45 | obinou | I fried one when using software-decoding |
| 19:45 | mk01_ | obinou: you get oriented a bit in the topic you want & xbian packages & code |
| 19:45 | mk01_ | in between |
| 19:46 | obinou | mk01_: sure |
| 19:46 | obinou | thanks again for this |
| 19:46 | rabeeh | mk01_: i'll PM about xbian team machines |
| 19:46 | mk01_ | we are finishing today with the concept for auto management & build & deploy |
| 19:46 | obinou | i'm pulling these repo now I'll try to sort all of this to make my own distro |
| 19:47 | mk01_ | and until you have something, there will be all prepared also considering different build targets for relevant repos |
| 19:47 | obinou | (not my own - rather my "flavor") |
| 19:49 | obinou | mk01_: Then i'll dive into all of this code - I'm familiar with the kernel itself, not so much with xbmc & all the ecosystem other related tools |
| 19:49 | obinou | will use your stuff whenever it gets ready - no rush |
| 19:50 | obinou | there is tons of forum posts I need to read too |
| 20:07 | mk01_ | you can use xbian.orf forum as well |
| 20:07 | mk01_ | as again the code is shared |
| 20:07 | mk01_ | what is on RPI, that works now on cubox too |