00:10 | i4pro | jnettlet: sorry i was away for a bit, did you get the same thing with sintel? |
00:10 | jnettlet | i4pro, I haven't sorry. I am caught up in work. |
00:11 | jnettlet | I will get back to you once I have tested. probably in the morning |
00:11 | i4pro | no worries |
00:17 | iio7 | I wanted to install Debian on cubox pro without using the cubox installer. Everything is running fine until during installation the Debian installer says "No installable kernel was found in the defined APT sources." |
00:27 | jnettlet | iio7, I doubt one is packaged. You will need to manually add the cubox kernel |
00:32 | iio7 | yeah, found the uImage-3.5.3-cubox, gonna use that rather than compiling |
00:32 | iio7 | unless it won't work on ext3 for some reason :) |
00:43 | jnettlet | it should work fine with ext3 |
00:45 | iio7 | yes! Finally.. new debian install with ext3 on external usb drive. |
02:12 | _rmk_ | rabeeh: doesn't the imx6 have vfp with 32 double registers, unlike the dove which only had 16... |
02:28 | _rmk_ | rabeeh: looking at the fpsid value, imx6 has a non-trapping fp implementation, which basically means all fp instructions are fully supported in hardware |
02:29 | _rmk_ | in other words, it'll only trap out to the support code when the unit is disabled (for context switching) |
10:04 | davorin | morning from .ch (o; |
10:26 | jnettlet | morning |
10:27 | jnettlet | davorin, why does switzerland use .ch? I could never understand that |
10:27 | davorin | ch = confederation helvetia |
10:28 | davorin | http://en.wikipedia.org/wiki/Helvetia |
10:29 | jnettlet | interesting |
10:44 | davorin | rabeeh already awake? (o; |
10:45 | davorin | have some mini-pcie connectors here to populate my c1 |
11:00 | davorin | someone doing android stuff? |
11:07 | jnettlet | _rmk_, I think I found out my card eject problem with D3CD. When doing the voltage switch DAT3 is used to verify that the switch was complete. Completion of voltage switch sequence is checked by high level of DAT[3:0]. Any bit of |
11:07 | jnettlet | DAT[3:0] can be checked depends on ability of the host. |
11:08 | jnettlet | which explains why Freescale recommends using gpio mode. |
11:10 | jnettlet | but the voltage switch drives DAT3 low |
11:10 | jnettlet | rabeeh, ^ can you also read that and verify my findings. |
11:12 | jnettlet | sorry, skip the last but. I think I was answering myself questioning myself :-) One of those mornings |
11:29 | jnettlet | so I guess the proper thing to do is if a gpio-cd property is not set that we automatically disable 1.8volt support |
11:40 | hste__ | jnettlet: have u seen those 4.1.0 vpu patches? https://github.com/wandboard-org/linux/commit/b65ef76276775ed1781f7a767610f52d7939cf86 https://github.com/wandboard-org/linux/commit/7cbd06b2904c1855109084ca6b0c84990bc69233 https://github.com/wandboard-org/linux/commit/1c86a3a256ebdac32d4ba5535b077a92358adaf8 https://github.com/wandboard-org/linux/commit/bad7e05d38fb78401143ed41f38688e7b6ea27e2 |
12:29 | bencoh | 3/5 |
12:29 | bencoh | woops |
12:31 | davorin | = 0.6 |
12:43 | jnettlet | hste__, those don't matter for 3.10 because the coherent memory is reserved through CMA. rabeeh may want to pull them into his 3.0.35 kernel though |
12:48 | _rmk_ | rabeeh: what does gpio5 bit 13 control on the oscillator - just an enable for it, or the power (and if it's power, what voltage please) ? |
12:49 | _rmk_ | rabeeh: same question for gpio3 bit 19 for the wifi too |
14:03 | jnettle | 14:03 * jnettlet is wondering if freescale has actually tested the UHS-i support at all. |
16:27 | jhonny | hi everyone. quick question i can't get either of my keyboards to work on my cubox-i2ultra. Keyboards I'm using are apple: USB wired A1243 and Bluetooth Wireless A1314 |
16:28 | jhonny | i already tried deleting one of they keyboard layout files trick, that didn't work and my bluetooth keyboard still doesn't work |
16:31 | jnettlet | jhonny, Apple's bluetooth keyboard probably requires a kernel module that isn't being built into the kernel |
16:32 | jnettlet | yeah a quick searchg and it looks like support wasn't added to the kernel until 3.4 |
16:32 | jhonny | i was reading about people having problems on getting their bluetooth apple keyboards not working on devices like nexus 7 after upgrading to android 4.3 |
16:32 | jhonny | upgrading to 4.4 fixed the issue |
16:33 | jhonny | or even removing one of the keyboard layout files ... because the issue I'm experiencing is .. |
16:33 | jhonny | my apple bluetooth keyboard connects and gets detected by bluetooth |
16:33 | jhonny | I can pair it to the cubox-i2ultra, however when i try to type something it doesn't work, nothing happens |
16:35 | jnettlet | oh android? I have no idea. haven't even looked at the kernel sources for that yet. |
16:39 | jhonny | yeah also the build seems very unstable, i get random lock ups and restarts and weird artefacts on the screen. And i know its not the cubox-i itself because geexbox runs like a charm on it |
16:47 | jnettlet | rabeeh, is this the proper pin config to select 1.8v's? MX6QDL_PAD_KEY_ROW1__SD2_VSELECT 0x17059 |
16:49 | _rmk_ | this is.. insane. |
16:50 | jnettlet | still working on the wifi support? |
16:50 | _rmk_ | I have my logitech receiver in the top USB, which is the usbotg port... and it works there. |
16:50 | _rmk_ | I've given it a fixed voltage regulator on gpio3 bit 22 |
16:51 | _rmk_ | according to /sys/kernel/debug/gpio... |
16:51 | _rmk_ | gpio-86 (usb_otg_vbus ) out lo |
16:51 | _rmk_ | but... it's active high. |
16:51 | _rmk_ | so either that gpio isn't used on cubox-i, or that file is lying, or..., something else |
16:55 | jnettlet | _rmk_, I am pretty sure the OTG port is the lower of the two usb ports |
16:55 | _rmk_ | the other annoying thing is that gpio number 0 doesn't seem to be claimed by the regulator code |
16:56 | _rmk_ | nope, the receiver is definitely plugged into the upper, and... |
16:56 | _rmk_ | 72: 0 0 0 0 GIC 72 ci_hdrc_imx |
16:56 | _rmk_ | 75: 49 0 0 0 GIC 75 ci_hdrc_imx |
16:56 | jnettlet | the top port has always worked and the lower port started working after rabeeh added the gpio 3 22 to the u-boot init. |
16:57 | _rmk_ | irq 75 corresponds with 43 in the dt file, which is the otg. |
16:57 | _rmk_ | irq 72 corresponds with 40 in the dt file, which is usbh1 |
16:57 | _rmk_ | and irq 75 increases when I hit keys... |
16:57 | _rmk_ | 75: 57 0 0 0 GIC 75 ci_hdrc_imx |
17:01 | jnettlet | I guess that is something rabeeh will have to answer |
17:17 | _rmk_ | I guess it's time to start throwing printk()s galore into the kernel to work out wtf is going on, why it's not parsing dt properly, why the regulator stuff seems to not work, etc |
17:20 | _rmk_ | if (config->ena_gpio && gpio_is_valid(config->ena_gpio)) { |
17:20 | _rmk_ | that's not helping |
17:20 | davorin | rabeeh awake now? (o; |
17:26 | _rmk_ | well... the conclusion I'm coming to is that modern kernels are total crap and we can't really support real hardware with them... and that the hummingboard and cubox-i work is just a pure fluke |
17:26 | _rmk | 17:26 * _rmk_ gives up with hummingboard/cubox-i hacking for the time being. |
17:42 | _rmk_ | and... if Rabeeh wants support to move forward in mainline kernels, then I need the schematics. I'm fed up of trying to work out what the design is from the 3.0.35 git repository and try to convert that to a DT representation. |
17:48 | davorin | general question: what are you using as host os and which toolchain for i.mx cross-development? |
17:51 | iio7 | I'm trying to install debian testing unto a new SD card from Kingston, but during boot there is a halt with the following error (info about the card is there too): http://paste.debian.net/76885/ |
17:51 | iio7 | Is that because the card is unsupported? |
17:52 | davorin | so you have also this in your log? |
17:52 | davorin | "Kernel panic - not syncing: No init found" |
17:52 | iio7 | yes |
17:52 | iio7 | on the console |
17:52 | davorin | have you tried another card, like sandisk? |
17:53 | davorin | i had extreme problems in the past using kingston on raspberry pi..with sandisk, never |
17:53 | iio7 | I had no problem with the default card that came with cubox. |
17:54 | davorin | i would definitively try a sandisk... |
18:02 | sickness | sandisk++ |
18:08 | iio7 | Using the cubox installer the card works fine. |
18:08 | iio7 | I just tried. |
18:08 | iio7 | It must be some boot arg settings or something. |
18:16 | davorin | when you add "rootfstype=ext4" ? |
18:18 | iio7 | No that didn't help. Gonna take a look at the boot.scr from cubox installer. |
18:31 | iio7 | Using identical bootargs from what the cubox installer uses didn't help. I guess this has to do with the initrd image from Debian then. |
18:32 | MarcusVinter | cubox installer working for i2's yet? |
18:32 | MarcusVinter | I'm only curious, I don't need it to. |
20:55 | rhollan | just wondering how responsive the UI in XBMC is on this device |
20:57 | rhollan | just wondering how responsive the UI in XBMC is on this device |
21:00 | xraxor | very |
21:01 | rhollan | I saw some youtube videos, and rendering was fine, but navigation seemed slow, like it is on the Raspberry Pi. Then again, it could simply have been network latency to content. |
21:02 | xraxor | i will post more videos on it soon |
21:02 | xraxor | but the software still meeds some optmisatiin |
21:03 | xraxor | its really fast, |
21:05 | rhollan | hmm. sounds promising then. |
21:06 | xraxor | it really is, once we get a few more distros and more devs get to work it it will get even better |
21:06 | xraxor | what xbmc distro do you use on the Pi? |
21:09 | rhollan | Oh, I don't... still looking to build an xbmc client and have been reviewiewing videos. |
21:09 | rhollan | Last client I built ran myth on a nanoITX |
21:10 | rhollan | That was at a time whene silent (fanless) 1080i (never mind p) rendering was impressive |
21:11 | rhollan | also cost $1k to put together with slim DVDROM, 500GB HD, and Silverstole LC-07 case. Case alone was about $200, and slim DVDROMS were expensive |
21:12 | rhollan | Silverstone |
21:12 | rhollan | So a handful of cuebox for living, and kids' rooms strike me as great. |
21:13 | rhollan | Got a Silicon Dust 3ch cable card network tuner that XBMC supports, so I could ditch the HD cable box that Comcast rapes me $16 a month for. |
21:35 | xraxor | my first media player was the Pi, for what it is its really good. |
21:35 | xraxor | but was slow to open stuff, like 1channel opening folders and so on |
21:35 | xraxor | now i have the cubox-i4pro |
21:35 | xraxor | havent used the pi for xbmc since |
21:37 | xraxor | that was a lot of money you spent on the pc |
21:37 | xraxor | i like the idea of energy saving with these arm devices |
21:43 | rhollan | Yeah, it was big $$$. But that was around 2007, when rendering even 1080i FANLESS was rather difficult. I took it as a challenge. Sadly, the nano-ITX platform proved flakey. |
21:44 | rhollan | How much is the i4Pro? |
22:12 | rhollan | does anyone know how much the i4Pro costs? |
22:13 | i4pro | its on their website |
22:13 | i4pro | i think its 124 now |
22:15 | rhollan | yeah, found it. Their site isn't the easiest to navigate. |
22:49 | _dab_ | rmk, unimpressed with the kernel? Does anyone have a stable kernel? I don't need an optimized video. Just stable basics. |
22:52 | _dab_ | biker, how is i4p kernel working for you? |
23:05 | iio7 | I got cubox because of the gigabit ethernet, but copying files using NFS unto external USB drive (drive has been tested elsewhere) seems really slow. I have tested using both XFS and EXT4. Is this like with Raspberry PI where the USB and Ethernet are sharing the same bus? |
23:09 | i4pro | iio7: you know the gigabit is limited to like 470 |
23:10 | i4pro | also you have jumbo frames on your network? |
23:10 | i4pro | jnettlet: you play sintel yet? |
23:13 | iio7 | i4pro, 470? I have gigabit lan with several boxes. Copying from one box to another gives me speeds at about 85 MB/s, but on the Cubox I only get about 16 MB/s. |
23:15 | i4pro | Micro-USB to RS-232 serial port connectivity joins the HDMI 1.4 and two powered USB 2.0 ports, and S/PDIF optical audio out. There's Gigabit Ethernet (although a maximum bandwidth of 470 Mbps is available due to limitations of the chip), a built-in Real Time Clock with battery backup, and an IR receiver and transmitter. 802.11n Wi-Fi and Bluetooth are included, though you'll still need to pay extra for a power adapter to be supplied with |
23:16 | i4pro | iio7: read this about jumbo frames, all your network equipment from pc to pc (and all switches in between need to support them) |
23:16 | i4pro | http://www.cyberciti.biz/faq/rhel-centos-debian-ubuntu-jumbo-frames-configuration/ |
23:17 | i4pro | if your network supports them, enable jumbo frames on your NFS box and the rest of the devices on your network |
23:18 | iio7 | Thanks, ill look into it! |
23:22 | i4pro | iio7: might do that for the rest of your network, but its not gunna work on the cubox, at least not with the dirsto i have i dont know about other ones.. ifconfig eth0 mtu 9000 SIOCSIFMTU: Invalid argument (cant even set mtu to 2000) |
23:23 | iio7 | I only got the cubox because of the gigabit, I already have 2 pi's :) Kindda a big bummer for me! |
23:36 | jhonny | Hey everyone, sorry to bother again, can someone help me try and troubleshoot why my apple USB keyboard doesn't work on Android 4.3 ? |