03:35 | ichthys | Does anyone have any advice for getting the cubox to use the entire SD. mine is using 2GBs of a 8GB SD |
03:35 | cbxbiker61 | how did you install it? |
03:36 | ichthys | debian img file. then i used windows diskimager |
03:37 | cbxbiker61 | i'm not a debian expert, but most distros have a way of resizing partitions (if the fs is supported) |
03:38 | ichthys | gotcha, i tried gparted it just throws errors |
03:38 | ichthys | is there a distro you would recommend using with the cubox? |
03:39 | cbxbiker61 | i can't really make that recomendation, since i do my own distro |
03:39 | cbxbiker61 | i'm still building up packages for the cubox-i, i have about 450 out of 2000 done |
03:40 | ichthys | almost there now! |
03:40 | cbxbiker61 | it'll be pretty easy from this point forward...the first ones are always the most time consuming |
03:42 | cbxbiker61 | we really need an sd based installer that could create the filesystems, that'd make it easier for everyone, in particular users without a regular linux box |
03:43 | ichthys | i have linux vms. but i try to mount the micro sd (in a usb reader) and it says its busy :/ |
03:43 | cbxbiker61 | is that with virtualbox? |
03:44 | ichthys | vmware player |
04:16 | _dab_ | jnettlet: rebuilt my 3.15 kernel. ondemand 146 Mbits/sec, performance 207 Mbits/sec |
04:26 | ichthys | poweroff |
09:06 | _dab_ | jnettlet: rebuilt my 3.15 kernel. ondemand 146 Mbits/sec, performance 207 Mbits/sec |
09:06 | jnettlet | _dab_, which is slower right? |
09:07 | jnettlet | well upstream is the same but downstream is slower |
09:07 | _dab_ | 207 is the fastest I have had |
09:07 | _dab_ | the kernel craps itself performance governer |
09:08 | _dab_ | now and then |
09:08 | jnettlet | interesting. |
09:08 | _dab_ | so 50mbit extra for performance gov |
09:08 | jnettlet | can you post the output of your /sys/kernel/debug/clk/clk_summary? |
09:09 | jnettlet | _dab_, which is correct. I have a patch for the ondemand governor to fix this. The clock speeds are tied to the cpufreq. By default the ondemand governor doesn't ramp cpu on iowait, which causes slow IO speeds on our platform. |
09:10 | jnettlet | freescale skirts the issue by shipping the interactive governor by default. |
09:10 | jnettlet | so that gives you the powersaving, but doesn't cripple IO performance |
09:11 | _dab_ | slow down for me. 1 see http://pastebin.com/index/fk7vdHwM and yes I will post the output you asked for |
09:11 | jnettlet | that is with the performance governor? |
09:12 | _dab_ | note that changing the governer did not alter tx speeds, so even more curious |
09:13 | _dab_ | unknown directory, i dont have it |
09:13 | jnettlet | oh you need to enable it in your kernel config |
09:14 | _dab_ | can do, will take a while |
09:14 | jnettlet | CONFIG_DEBUG_FS=y |
09:14 | _dab_ | yes to the gov question |
09:14 | _dab_ | no failures with other governers |
09:16 | _dab_ | i already have CONFIG_DEBUG_FS=y. the sys/kernel/debug/ directory is empty |
09:17 | jnettlet | oh it might not be getting mounted automagically |
09:17 | jnettlet | mount -t debugfs nodev /sys/kernel/debug |
09:18 | _dab_ | that did it |
09:22 | _dab_ | http://pastebin.com/rkZgYG2a |
09:23 | _dab_ | BTW 3.15-rc1 wifi is still v. slow |
09:24 | _dab_ | but it works |
09:29 | jnettlet | _dab_, clocks look fine |
09:29 | jnettlet | can you post dmesg for the latest kernel |
09:30 | _dab_ | ok |
09:32 | _dab_ | http://pastebin.com/fnXUXWDs |
09:33 | _dab_ | AND echo 'heartbeat' > /sys/class/leds/imx6\:red\:front/trigger sends the kernel into an infinite loop |
09:35 | jnettlet | _dab_, oh interesting you are negotiating flow control off, I am negotiating flow control rx/txt |
09:35 | jnettlet | rx/tx |
09:36 | jnettlet | _dab_, what switch are you plugged into? |
09:37 | jnettlet | can you try just plugging the cubox-i directly into another gigabit card? |
09:37 | _dab_ | netgear fvs318n |
09:38 | _dab_ | and yes, I can do that. I did it ages ago (on rabeehs kernel) with no real effect. plugged directly to an intel ethernet card on a HP pc |
09:42 | _dab_ | back in a little while |
10:01 | _dab_ | http://pastebin.com/bPru1GuW |
10:01 | _dab_ | that was connected to a dell intel ethernet |
10:01 | _dab_ | shame it cant do that all the time400/550 is pretty good |
10:05 | _dab_ | jnettlet: you still there? |
10:10 | jnettlet | _dab_, yep |
10:15 | _dab_ | you got that ok? |
10:16 | _dab_ | eating... |
10:18 | jnettlet | _dab_, okay that looks more normal. Can you check the logs. I bet that is with flow control on. |
10:18 | jnettlet | that is pretty important for gigabit ethernet |
10:53 | _dab_ | flow control rx/tx |
11:06 | _dab_ | tx off on the PC |
12:45 | jenkins101 | hi! |
12:45 | jenkins101 | I have noticed that if you read a lot for USB the ir reciver stops responding |
12:45 | jenkins101 | using latest 3.10.30 kernel |
12:45 | jenkins101 | o cuboxi |
12:45 | jenkins101 | anyone else noticed this? |
12:51 | jenkins101 | rabeeh: ? |
12:51 | jenkins101 | have you tested this? |
13:08 | jnettlet | jenkins101, please provide a test case of exactly what you are doing. |
13:17 | rabeeh | jenkins101: what do you mean? you have a hard drive on the USB? and then IR stops working? |
13:17 | rabeeh | please provide details; and this is first report about this |
13:23 | jenkins101 | hehe :-) |
13:23 | jenkins101 | ok |
13:24 | jenkins101 | I am actully running a dvb adapter |
13:24 | rabeeh | hey |
13:24 | jenkins101 | tuned to a channel |
13:24 | rabeeh | oh; nice !!! |
13:24 | jenkins101 | then IR is goon |
13:24 | rabeeh | which? t/s/c? |
13:24 | jenkins101 | c |
13:24 | rabeeh | fta content? |
13:24 | jenkins101 | cinergy htc usb xs |
13:24 | jenkins101 | yes |
13:24 | rabeeh | which country? |
13:24 | jenkins101 | sweden |
13:24 | rabeeh | which codec? |
13:24 | jenkins101 | codec? |
13:24 | rabeeh | xbmc livetv? |
13:24 | jenkins101 | aa yes |
13:24 | rabeeh | codec - mpeg2/h.264? |
13:25 | jenkins101 | an if I just tunne to the channel... |
13:25 | jenkins101 | aa mpeg2 |
13:25 | jenkins101 | or all actully :-\ |
13:25 | jenkins101 | both |
13:26 | rabeeh | maybe there is some sort of transmitter on the dongle that is blocking? |
13:26 | rabeeh | if you do 'cat /dev/lirc' you should see some output when hitting any key on the remote |
13:27 | rabeeh | try with the dongle and without and lets see what happens |
13:27 | rabeeh | btw - is de-interlacing enabled? this is a major feature in the i.mx6 vs. most other xbmc boxes that it has built-in hardware deinterlacer ! |
13:28 | jenkins101 | ok |
13:28 | jenkins101 | well if I tunne to a channel |
13:28 | jenkins101 | dvbdump... |
13:28 | jenkins101 | ir stops working directly |
13:29 | jenkins101 | I dont get anything from /dev/lirc |
13:29 | jenkins101 | not even when it is working? |
13:31 | rabeeh | which distro are you using? |
13:31 | jenkins101 | OE :-) |
13:31 | rabeeh | GB/OE/XB? |
13:32 | rabeeh | oh - nice :) |
13:32 | rabeeh | which build? latest? |
13:32 | jenkins101 | yes... |
13:32 | jenkins101 | I whas testing the CEC just now... |
13:32 | rabeeh | that fast? |
13:32 | jenkins101 | the mails you know... ;-) |
13:32 | rabeeh | i think only yesterday wolfgar provided the patches |
13:34 | jenkins101 | yes |
13:34 | jenkins101 | I reported |
13:38 | jenkins101 | it works good now by the way |
13:38 | jenkins101 | :-) |
13:38 | jenkins101 | but now it is this IR thing.. |
13:57 | dxt | 13:57 * dxtr just set up a private GitLab thing |
13:59 | jenkins101 | rabeeh: yes the de-interlacing seems to work finne! :-) |
14:19 | jenkins101 | rabeeh jnettlet I tested with a sundtek card and IR works when I use that card?!? |
14:20 | rabeeh | dxtr: GitHub? |
14:20 | rabeeh | jenkins101: can u judge from quality point of view? (the deinterlacing) |
14:20 | jenkins101 | rabeeh: how do you mean |
14:21 | rabeeh | the deinterlacing function; do you see artifacts? |
14:21 | rabeeh | it's pretty new feature added to xbmc |
14:22 | jenkins101 | no... |
14:22 | jenkins101 | it works good I think |
14:23 | jenkins101 | linnes get a little blure and distorted when sweeping in sports... |
14:24 | jenkins101 | but I only have a old ION to compare and thats more or less the same... |
14:32 | rabeeh | i wonder which algorithm they are using for deinterlacing; there are two modes in the ipu - slow and fast motion |
14:47 | jenkins101 | rabeeh: you know anythin about HD audio pass-trough? |
15:14 | rabeeh | jenkins101: i need to get back; but i think GB has it up and running now |
15:14 | dxtr | rabeeh: No. Gitlab |
15:15 | dxtr | It's like Github but you can set it up yourself if you wanna |
15:15 | dxtr | https://www.gitlab.com/ |
15:15 | dxtr | http://gl.dxtr.im |
15:15 | dxtr | http://gl.dxtr.im/u/dxtr :) |
15:15 | dxtr | (Also available over https with my own self-signed certificate) |
15:20 | rabeeh | nice ! |
15:22 | dxtr | I've been spending a little while importing my projects from Github |
15:22 | dxtr | rabeeh: If you wanna try it I could give you an account :P |
15:23 | dxtr | It isn't very far from github (It's actually really, really close. Like.. If you're familiar with github you'll learn gitlab in five minutes or less) |
15:24 | dz0ny | https://github.com/gogits/gogs |
15:24 | dz0ny | another github clone |
15:26 | dxtr | dz0ny: I'll have to try it |
15:27 | dxtr | Go and everything |
15:27 | dxtr | That's cool |
16:00 | jenkins101 | rabeeh: no... it is not working in GB... at least not for me... :-) |
16:01 | jenkins101 | this was 20140426 |
16:01 | jenkins101 | HD audio that is |
16:02 | jenkins101 | DTS and AC3 is finne. |