| 00:00.04 | [7] | hm... is there a way to install a "clean" ubuntu, like on the pandaboard? |
| 00:00.42 | buZz | [7]: you can use a debian howto, and just swap the repositories for ubuntu |
| 00:00.48 | [7] | with the normal unity desktop or headless, the normal amount of kernel modules available, ... |
| 00:01.00 | buZz | unity is normal? HAHAHAHAHAHA |
| 00:01.17 | buZz | :) |
| 00:01.28 | buZz | well, its possible i guess |
| 00:01.29 | [7] | "normal" as in "stock configuration" |
| 00:01.40 | buZz | there is no stock ubuntu for Cubieboard |
| 00:01.44 | [7] | this linaro ubuntu doesn't have a driver for my wifi adapter, which even openelec has |
| 00:01.46 | buZz | :P |
| 00:01.53 | buZz | then compile it |
| 00:02.04 | buZz | get kernel source, turn on option, compile modules, install |
| 00:02.21 | [7] | sure, but then I have to do that after every update |
| 00:02.40 | [7] | would be much easier if there would just be a more complete kernel package somewhere |
| 00:04.15 | [7] | i really like how this works on OMAP, especially in contrast to the rpi |
| 00:05.07 | buZz | there are more complete kernel packages all over the place |
| 00:05.22 | buZz | rm <- has a nice kernel for people who are unable to compile one |
| 00:05.36 | buZz | http://romanrm.ru/en/a10/kernel <- there |
| 00:06.00 | buZz | most ARM boards dont need a new kernel every update |
| 00:06.12 | buZz | usually there are only 2 kernel trees available |
| 00:06.13 | [7] | what is the real difference (except for package selection) between linaro and canonical ubuntu? |
| 00:06.27 | buZz | no clue, i dont like either |
| 00:06.35 | buZz | trying to avoid them |
| 00:11.32 | pidpawel | what should i do to use mplayer to play on HP output? |
| 00:12.53 | pidpawel | (i mean on linaro ubuntu) |
| 00:13.10 | buZz | what is HP? |
| 00:13.27 | buZz | ah headphones? |
| 00:13.30 | pidpawel | yes |
| 00:13.43 | buZz | should output to headphones port by default? |
| 00:14.01 | buZz | 16:57:14 < buZz> mashku: syntax for mplayer to select second alsa output is : mplayer -ao alsa:device=hw=1.0 bla.avi |
| 00:14.02 | pidpawel | alsamixer says sun4i-sndhdmi |
| 00:14.09 | buZz | aplay -l |
| 00:14.38 | pidpawel | oh. -ao works |
| 00:14.40 | pidpawel | thanks ;) |
| 00:15.09 | buZz | yw |
| 00:15.30 | pidpawel | have anyone tried to get ir working? |
| 00:16.02 | buZz | not me |
| 00:27.18 | *** join/#cubieboard Sikki (4c152be9@gateway/web/freenode/ip.76.21.43.233) |
| 00:29.40 | *** join/#cubieboard RavenIIII (~RavenII@c-76-29-25-66.hsd1.il.comcast.net) |
| 00:30.09 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 00:35.37 | *** join/#cubieboard RavenII (~RavenII@c-76-29-25-66.hsd1.il.comcast.net) |
| 00:37.48 | *** join/#cubieboard RavenIII (~RavenII@c-76-29-25-66.hsd1.il.comcast.net) |
| 00:46.39 | pidpawel | irrecord: gap not found, can't continue |
| 00:46.40 | pidpawel | ;/ |
| 00:47.09 | Gumboot | Not having so much luck with the serial console after loading th kernel. Should be 'console=ttyS0', right? |
| 00:47.26 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 00:48.48 | pidpawel | root@harp:~# cat /proc/cmdline |
| 00:48.49 | pidpawel | console=ttyS0,115200 console=tty0 disp.screen0_output_mode=EDID:1920x1080p60 hdmi.audio=EDID:0 mmc0_led=1 bootmenutimeout=10 datadev=mmcblk0p2 |
| 00:49.02 | Gumboot | Two consoles! |
| 00:51.01 | Gumboot | Oh yes. Adding ",115200" makes it work. |
| 00:53.31 | *** join/#cubieboard RavenII (~RavenII@c-76-29-25-66.hsd1.il.comcast.net) |
| 00:55.06 | Gumboot | Is there an output mode which disables display altogether? |
| 01:31.41 | *** join/#cubieboard stefanro1 (~stefan@pD9FFBEA7.dip.t-dialin.net) |
| 01:37.26 | *** join/#cubieboard ghw (~ghw@74.63.212.29) |
| 01:38.46 | pidpawel | nice. |
| 01:38.59 | pidpawel | cubie as radio player with ir remote control ;) |
| 01:39.32 | [7] | hm, how does the cubie's serial port work? |
| 01:39.44 | [7] | RX/TX from which point of view? |
| 01:39.51 | [7] | for me both seem to be stuck at ~3.2V |
| 01:40.09 | pidpawel | isnt it a simple ttl port? |
| 01:57.01 | rvalles | sound is pretty good... |
| 02:09.49 | *** join/#cubieboard GeorgeIoak (4c143130@gateway/web/freenode/ip.76.20.49.48) |
| 02:10.44 | Gumboot | does the 'discard' mount option apply only to SSD? Not any other solid-state media? |
| 02:11.16 | Gumboot | rvalles: are you in a position to evaluate the line-in, as well? |
| 02:11.19 | [7] | erm, SSD == solid state media, because that's the definition :P |
| 02:11.47 | Gumboot | I'm just looking at the Debian instructions, and looking at some other random rant on the internet, and wondering if the mount options should be unified. |
| 02:11.58 | Gumboot | But I don't know if discard is relevant for mmc. |
| 02:12.43 | Gumboot | Also, I have no idea whose 'tx' is labelled 'tx' and whose 'rx' is labelled 'rx'. To me they're just bad names for the colours of the wires I stick into them. |
| 02:13.12 | [7] | well in theory an MMC card can support TRIM/discard |
| 02:13.27 | [7] | I know that some eMMC devices do (samsung superbrick bug was caused by a bug in that) |
| 02:13.47 | [7] | I guess that some cheaper SD cards will just ignore it though |
| 02:13.57 | [7] | and TX/RX is from the board's point of view |
| 02:14.15 | Gumboot | Then does this http://superuser.com/questions/228657/which-linux-filesystem-works-best-with-ssd need to be merged into here http://linux-sunxi.org/Debian#Configure_the_new_rootfs ? |
| 02:15.06 | [7] | I don't think adding discard there could hurt... |
| 02:15.15 | [7] | (unless you're talking to a buggy samsung eMMC) |
| 02:15.18 | Gumboot | I've spent way too long trying to get Debian installed on a Samsung SSD only to find that it's been failing under load. I switched it for a Mushkin and that seems to be fine. |
| 02:15.22 | Gumboot | So far. |
| 02:15.57 | [7] | is running ubuntu on a samsung 830 series 256GB ssd, and it's working perfectly |
| 02:16.09 | [7] | lvm+ext4 |
| 02:16.15 | Gumboot | On a cubieboard? |
| 02:16.20 | [7] | nope, my laptop |
| 02:16.31 | Gumboot | Yeah. I stuck it on a cubieboard and it went wrong. |
| 02:16.44 | [7] | ok, so the problem is specific to the board and not the SSD? |
| 02:16.56 | Gumboot | I'd left it on my desk for about three months after buynig it, and I thought I _really_ needed to test it so I knew whether or not the shop had sold me a faulty part. |
| 02:17.02 | Gumboot | (Very high risk in this country) |
| 02:17.22 | Gumboot | Well, if you feel like it you could try plugging your SSD into a cubieboard and doing a lot of writes to it. |
| 02:17.30 | Gumboot | Sounds like yours is identical to mine. |
| 02:17.40 | Gumboot | Could be the PSU. |
| 02:17.43 | [7] | well, I don't have a spare SSD around |
| 02:17.46 | Gumboot | That _claims_ to be 2A, but they never are. |
| 02:18.06 | [7] | did that SSD log unclean shutdowns? |
| 02:18.09 | Gumboot | I seem to have two spare SSDs around. |
| 02:18.18 | [7] | that samsung model has a SMART counter for those |
| 02:18.25 | Gumboot | dmesg started raving about restarts. The SSD just gave up altogether. |
| 02:18.34 | Gumboot | Oh, right. |
| 02:18.36 | Gumboot | Umm... didn't check. |
| 02:18.36 | AndrevS | Yep, I am having possible power issues with a dx 2A supply as well.... |
| 02:18.54 | *** join/#cubieboard emilepetrone (~emilepetr@c-24-22-29-165.hsd1.or.comcast.net) |
| 02:18.54 | Gumboot | I wouldn't have thought SSD would be such a burden. |
| 02:18.54 | [7] | dx... well, that's to be expected |
| 02:19.02 | Gumboot | But maybe they're more sensitive to noise. |
| 02:19.20 | [7] | that specific model can easily draw 1A |
| 02:19.21 | AndrevS | I'm going to hook it up to a 6A power supply, just to be sure the psu isn't the issue |
| 02:19.30 | Gumboot | [7]: The samsung one? |
| 02:19.33 | [7] | yes |
| 02:19.38 | Gumboot | I guess it's meant to be some high-performance thing. |
| 02:19.49 | [7] | it isn't a particularly efficient one, uses more power than an average HDD while writing |
| 02:19.53 | Gumboot | Not sure it's proper to plug it into a 1GHz ARM. |
| 02:19.58 | [7] | and probably rather bursty |
| 02:20.00 | Gumboot | From ... what .. 2006? |
| 02:20.55 | [7] | starts experimenting with nand an FEL |
| 02:22.03 | [7] | from what I gather boot0 and boot1 are stored somewhere outside of the FTL user area |
| 02:22.37 | *** join/#cubieboard ganbold__ (~Ganbold@202.21.108.34) |
| 02:28.49 | *** join/#cubieboard des2 (~noone@pool-71-190-44-156.nycmny.east.verizon.net) |
| 02:29.02 | [7] | wonders what all this interesting stuff inside the nand boot partition is |
| 02:30.42 | [7] | is there some documentation how the regular SD card boot process works? |
| 02:30.54 | [7] | are there any .axf files (which seem to contain ELF executables) involved? |
| 02:30.58 | rvalles | Gumboot: don't think so... it's just I had played some music on rpi's sound output and that is pretty nasty |
| 02:32.56 | Gumboot | It's PWM. |
| 02:33.14 | Gumboot | Although I thought I saw a changelog entry saying they'd improved it. |
| 02:33.30 | Gumboot | Might have gone and done some funky noise shaping or something. |
| 02:33.38 | AndrevS | I'm playing music on my pi all the time |
| 02:34.02 | Gumboot | In fact... you know what? If that isn't the case then I know exactly who I should chase up next Wednesday to get that to happen. |
| 02:34.15 | AndrevS | My main problem with it is the "tick" between the songs |
| 02:34.37 | Gumboot | That might be a specific PWM problem. |
| 02:34.58 | Gumboot | If it's not balanced then the only way to output the 0V value is by continuously oscillating, which would be a waste of power. |
| 02:35.20 | Gumboot | So you'd expect a click as it returns to Vss or Vdd, which would be +1V or -1V and not 0V. |
| 02:35.45 | Gumboot | (I mean, +1V or -1V after the necessary biasing) |
| 02:35.49 | [7] | well you could at least slowly fade it from 50% to 0% duty cycle to get rid of the pop |
| 02:36.15 | AndrevS | that's what I would think. |
| 02:36.35 | Gumboot | Might be better to drive a balanced pair. |
| 02:36.45 | Gumboot | And then just set them both to the same value for 0V. |
| 02:37.06 | Gumboot | I haven't really thought that through very hard, though. |
| 02:37.09 | Gumboot | And it'd be a PCB respin. |
| 02:38.49 | minkley | Just building cubieboard_opentv.tar.gz - had a bunch of compile errors - turns out not ready for gcc/g++ 4.7 - need to compile with gcc-4.4/g++-4.4 |
| 02:39.21 | AndrevS | ouch |
| 02:39.34 | Gumboot | errors? |
| 02:39.56 | Gumboot | I expect new warnings with 4.7, but new errors is more likely a configuration problem. |
| 02:41.24 | Gumboot | wonders if he could get some kind of shell extension to see how quickly he types 'sudo apt-get install ' |
| 02:41.26 | minkley | True, although I did not see a 'configure' |
| 02:42.03 | minkley | Here was the error -> make: *** [out/host/linux-x86/obj/EXECUTABLES/grxmlcompile_intermediates/grxmlcompile.o] Error 1 |
| 02:42.24 | Gumboot | That's the one from make. Was there one from GCC earlier on? |
| 02:42.59 | minkley | installed gcc / g++ 4.4 and it is now compiling |
| 02:43.01 | AndrevS | usually the real error is a few lines above it |
| 02:43.19 | [7] | Gumboot: you're thinking of building a bridge driver for the headphone output? |
| 02:43.36 | [7] | won't work well unless you have a differential amplifier following it... the output channels share a common ground |
| 02:44.23 | Gumboot | I was thinking of some fiddly arrangement of capacitors, but by "thinking" I don't mean anything quite so advanced as the word implies. |
| 02:44.39 | minkley | It bombs pretty quickly once you run make -j4 |
| 02:44.55 | minkley | about 30000 lines ago on my compile screen |
| 02:45.01 | Gumboot | minkley: You're not building for soft-float or something, are you? |
| 02:45.35 | Gumboot | These days I can't be bothered thinking about compile errors. I usually just issue make -j64 over and over until everything but that one file is built. |
| 02:45.49 | Gumboot | Because sometimes I get lucky and don't actually care about the binary that it's needed for. |
| 02:45.53 | minkley | No - this is my first build so i am pretty much just getting familiar with everything, following the instructions on the A1x wiki |
| 02:46.15 | [7] | Gumboot: make -k |
| 02:46.16 | Gumboot | The only gcc-4.4 binaries I've seen were things people kept around for their soft-float builds. |
| 02:46.52 | minkley | gumboot: so are you saying other folks are building with 4.7 fine? |
| 02:47.01 | minkley | gumboot: thats interesting |
| 02:47.05 | Gumboot | minkley: No. I mean in general. You have a hard-float gcc-4.4? |
| 02:47.15 | Gumboot | What's the link to the tarball? |
| 02:47.22 | minkley | hold on |
| 02:47.40 | minkley | http://dl.cubieboard.org/software/sources/cubieboard_opentv.tar.gz |
| 02:48.05 | minkley | gumboot: I was following http://linux-sunxi.org/Building_Android_TV_image |
| 02:48.52 | Gumboot | Does berryboot put you in a chroot? And does lsof fail to play the game? |
| 02:48.59 | minkley | gumboot: Built a debian 12.10 dev VM on virtualbox on my mac |
| 02:50.36 | minkley | Sorry not debian :) Ubuntu |
| 02:50.41 | Gumboot | Looks like I've just switched out all my stuff for a fresh Debian install. I may take some time to get my tools in order. |
| 02:51.01 | minkley | which is a debian linux line |
| 02:51.26 | Gumboot | Ooh.. this time it doesn't crash. |
| 02:51.49 | Gumboot | Is it useradd or adduser to add a new user with all the support scripts and everything for doing it right? |
| 02:52.06 | Gumboot | Maybe it's neither of those. |
| 02:54.02 | Gumboot | It is adduser |
| 02:58.34 | minkley | actually its both :) |
| 02:59.13 | Gumboot | useradd is the unfriendly version. If you run it it'll fail to do a bunch of automated stuff. |
| 03:00.16 | AndrevS | On Archlinux, adduser has been removed from shadow, but is a separate package in AUR |
| 03:01.03 | AndrevS | Gah, I was like.... wtf why doesn't adduser work anymore?!?! |
| 03:01.12 | Gumboot | Debian's openssl is unoptimised, same as Linaro. |
| 03:05.11 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 03:05.31 | Gumboot | I think I can actually tell the difference in log-in times with the downgrade. |
| 03:05.44 | Gumboot | (after fixing name resolution issues) |
| 03:06.09 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 03:07.22 | Gumboot | I think actually I won't look at that cubieboard_opentv build issue. It's something a whole lot bigger than I expected. |
| 03:07.28 | swivelgames | Hey guys... I'm trying to flash my A13 board's NAND. Using sunxi tools. Anyone here able to coach me a bit? |
| 03:12.25 | *** join/#cubieboard lerc (~quassel@121-74-230-95.telstraclear.net) |
| 03:25.47 | Gumboot | minkley: THat tarball is an android image? You're building Android? |
| 03:25.59 | Gumboot | Android is soft-float. |
| 03:26.22 | Gumboot | The compiler downgrade isn't strictly necessary, but it's a popular way of getting soft-float. |
| 03:26.49 | Gumboot | IF you can find Linaro toolchain binaries from about a year ago, those are soft-float too. |
| 03:28.28 | *** join/#cubieboard rz2k (~rz2k@95.27.13.50) |
| 04:01.40 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 04:05.35 | minkley | Gumboot: I got round the build issues by installing Gcc/G++ 4.4 - then pointed the gcc/g++ links in /usr/bin to the gcc-4.4 and g++-4.4 binaries which I figured was easier than touching all the makefiles |
| 04:10.24 | *** join/#cubieboard TheSeven (~quassel@rockbox/developer/TheSeven) |
| 04:17.15 | minkley | Gumboot: Not sure I understand why Android uses Soft-Float - I thought the A10 had FPU (vFPv3 or something) |
| 04:34.37 | Workboot | They do it because they always have. It's hard to change, now. |
| 04:35.19 | Workboot | Originally it was to support shitty old devices which don't matter anymore. |
| 04:37.18 | Gumboot | Oh, wrong window. |
| 04:37.39 | Gumboot | Wrong keyboard and screen, in fact. |
| 05:13.06 | *** join/#cubieboard KiSM (~KiSM@CPE-120-145-19-206.lnse2.wel.bigpond.net.au) |
| 05:15.40 | *** join/#cubieboard rvalles (~rvalles@unaffiliated/rvalles) |
| 05:20.34 | *** join/#cubieboard grog (~grog@S01060003ba6c9cbc.cg.shawcable.net) |
| 05:25.31 | *** join/#cubieboard grog (~grog@68.146.29.46) |
| 05:34.14 | *** join/#cubieboard titankiller (~titankill@117.214.20.72) |
| 05:43.21 | *** join/#cubieboard user__ (~user@93-35-240-165.ip57.fastwebnet.it) |
| 05:54.10 | *** join/#cubieboard imRance (~Rance@116.52.106.224) |
| 06:05.59 | *** join/#cubieboard titankiller (~titankill@117.201.246.44) |
| 06:26.12 | *** join/#cubieboard _bry4n (~nyrb@64-148-253-143.lightspeed.snantx.sbcglobal.net) |
| 06:50.06 | *** join/#cubieboard lkoranda (lkoranda@nat/redhat/x-lhscosfseyxslldl) |
| 07:07.53 | *** join/#cubieboard sv (~discopig@unaffiliated/discopig) |
| 07:41.40 | *** join/#cubieboard user__ (~user@93-36-14-247.ip57.fastwebnet.it) |
| 07:43.40 | *** join/#cubieboard rellla (~rellla@p4FE562C9.dip0.t-ipconnect.de) |
| 07:46.47 | swivelgames | Hey guys. I was about to compile `awutils` (`awflash`) to allow me to flash my A13 board's NAND with a new image. However, I don't know what partition to flash it to. Can anybody help me out here? Do I flash boot? misc? system? data? |
| 07:47.49 | *** join/#cubieboard titankiller (~titankill@117.201.246.44) |
| 07:53.52 | lawrence_ | nanda should be the boot |
| 07:53.58 | lawrence_ | nandb should be data |
| 07:54.04 | lawrence_ | depends on os installed at moment though |
| 07:54.19 | lawrence_ | you're android, assuming you're the one who just posted on cubieforums |
| 07:54.28 | swivelgames | Indeed. |
| 07:54.54 | lawrence_ | (i'm busy trying to setup kandan on there so its been up/down for the last few min) as i install passenger, then swear at the ruby gods |
| 07:56.02 | lawrence_ | i'm off to find a ruby channel so i can ask nicely why passenger no want to autostart things... |
| 07:56.20 | swivelgames | lawrence_, do you know what partition to flash? |
| 07:56.31 | swivelgames | Should I be flashing boot? |
| 07:56.46 | lawrence_ | i wouldn't touch anything yet till backing up |
| 07:56.53 | swivelgames | Indeed |
| 07:57.06 | lawrence_ | i'd probably use the nand_boot tools from sunxi-tools |
| 07:57.18 | lawrence_ | github://linux-sunix/sunxi-tools |
| 07:57.38 | lawrence_ | patrick updated those from the allwinner source |
| 07:57.54 | swivelgames | I have sunxi-tools. Was going to use fel to move the image over but... I still have no clue what partition to flash the image on to. |
| 07:58.05 | swivelgames | That's the question no one seems to be able to answer :/ |
| 07:58.10 | lawrence_ | i'm not so worried about the partitions |
| 07:58.26 | lawrence_ | i'm more worried about where exactly it expects the ipl / spl to be |
| 07:58.46 | lawrence_ | as thats not clear from a look at the code |
| 07:59.12 | lawrence_ | and a dump of the partition shows me a pseudo fat16 + preheader at the first 8k odd |
| 07:59.25 | lawrence_ | i need to do a proper dump though |
| 07:59.35 | lawrence_ | or download someones to compare |
| 07:59.53 | lawrence_ | did you read patricks notes on the miniand forum? |
| 07:59.59 | lawrence_ | https://www.miniand.com/forums/forums/development/topics/boot-ubuntu-lubuntu-from-internal-4g-flash?page=2#post-3046 |
| 08:00.15 | lawrence_ | mbr is the old name for the nand_tools |
| 08:00.51 | lawrence_ | see page 1 of that post also |
| 08:02.12 | lawrence_ | so nand_part - create partitions |
| 08:02.21 | lawrence_ | then add whatever as per normal in them |
| 08:03.09 | lawrence_ | looks like uboot goes at 32k in nand same as on sd |
| 08:16.30 | *** join/#cubieboard imRance1 (~Rance@116.52.33.220) |
| 08:22.18 | *** join/#cubieboard asiekierka (~asiekierk@fluttershy.pl) |
| 08:28.13 | *** join/#cubieboard emilepetrone (~emilepetr@c-24-22-29-165.hsd1.or.comcast.net) |
| 09:18.34 | *** join/#cubieboard c4nn1b4l (~cannibal@87.97.30.82.pool.invitel.hu) |
| 09:19.59 | lawrence_ | ruby is my new swearword for the day |
| 09:20.32 | swivelgames | hehehe |
| 09:20.59 | lawrence_ | the advise was - oh install a different version for that gem bundle |
| 09:21.32 | lawrence_ | i feel like mr mcsweary_very_bad_language today |
| 09:21.42 | swivelgames | The response, I'm sure, was similar to - stab stab stab |
| 09:22.01 | swivelgames | hahaha, mr mcsweary_very_bad_language |
| 09:22.06 | swivelgames | It has a ring to it... |
| 09:22.09 | swivelgames | Kind of... |
| 09:22.21 | swivelgames | :P |
| 09:23.11 | swivelgames | Anyway, I'm out for the night. olimex guys aren't giving me any guidance on this. Would be nice to get a straight answer out of the manufacturer of my board, but you know... trivial things. |
| 09:23.23 | swivelgames | 03:30 here... |
| 09:23.35 | swivelgames | Hope you have better luck, lawrence_ |
| 09:23.49 | lawrence_ | would like to get a chat app going on cubieforums |
| 09:23.59 | lawrence_ | as kandan looks nice |
| 09:24.09 | lawrence_ | albeit a @#$%$##%#$ pain to install |
| 09:24.37 | swivelgames | Most likely |
| 09:24.53 | swivelgames | Especially with luck the way it is lately... hehe |
| 09:25.13 | lawrence_ | hopefully github works today so i can pull kernel |
| 09:25.59 | lawrence_ | i think i can tell passenger to use development instead of production to fake out the kandan bits |
| 09:31.46 | *** join/#cubieboard rellla (~quassel@p4FE562C9.dip0.t-ipconnect.de) |
| 09:37.56 | *** join/#cubieboard KiSM (~KiSM@CPE-120-145-19-206.lnse2.wel.bigpond.net.au) |
| 09:46.28 | lawrence_ | and that worked finally |
| 09:47.37 | *** join/#cubieboard titankiller (~titankill@117.201.246.44) |
| 09:49.12 | lawrence_ | anyone want to try out chat.cubieforums.com ? |
| 09:49.22 | lawrence_ | just to see if it does work |
| 09:58.53 | *** join/#cubieboard qleelulu (~qleelulu@116.21.80.128) |
| 10:02.00 | rm | just put this link somewhere http://webchat.freenode.net/?randomnick=1&channels=cubieboard&uio=MT11bmRlZmluZWQb1 |
| 10:02.21 | *** join/#cubieboard qwebirc43245 (052db453@gateway/web/freenode/ip.5.45.180.83) |
| 10:02.28 | qwebirc43245 | it works :) |
| 10:09.27 | buZz | lawrence_: does that go to this irc channel? |
| 10:09.37 | lawrence_ | not yet |
| 10:09.59 | lawrence_ | i spent half the day finally getting it going, but it looks like its a crap solution, will look for something else |
| 10:10.08 | buZz | also, why does it have authentication for chatting? |
| 10:10.28 | buZz | mibbit.com also has a nice and easy webchat |
| 10:10.30 | lawrence_ | most definitely want irc integration so then people can just -> there for the #cubieboard love on irc |
| 10:10.54 | buZz | and rm's link also works fine |
| 10:16.19 | *** join/#cubieboard provel_ (~matte@host4-127-dynamic.10-87-r.retail.telecomitalia.it) |
| 10:34.19 | lawrence_ | webchat doesn't load for me |
| 10:34.25 | lawrence_ | i'd like a hosted version that works |
| 10:34.59 | *** join/#cubieboard titankiller (~titankill@117.201.246.44) |
| 10:35.10 | lawrence_ | mibbit looks interesting though |
| 10:44.40 | TheSeven | swivelgames: which partition to flash to depends on what you want to flash |
| 10:44.53 | TheSeven | but how do you even make it flash something? |
| 10:45.57 | TheSeven | lawrence: boot0/1 are outsude the /dev/nand accessible area |
| 10:46.24 | lawrence_ | ok, thats good |
| 10:46.47 | lawrence_ | so the /dev/nand is an abstraction not the actual nand |
| 10:47.01 | lawrence_ | its not provided by MTD driver? |
| 10:47.22 | lawrence_ | and for a "raw" nand read I just read off the mtd ? |
| 10:47.47 | lawrence_ | i haven't looked at it too closely yet |
| 10:48.24 | TheSeven | well i only have my cubie since yesterday, but i don't think you can access the raw nand easily through mtd |
| 10:48.48 | TheSeven | /DEV/NAND is an FTL with some proprietary partition table mechanism |
| 10:49.31 | TheSeven | so it is a highlevel block device that's already wear leveled and bad block remapped |
| 10:49.41 | TheSeven | so you can just throw ext4 into that |
| 10:49.42 | lawrence_ | looking at patricks code seems 32k it expects data (not clear on ipl or spl for that) but that seems logical, seeing as the SD is similar |
| 10:50.48 | lawrence_ | and its similar to what i see on other soc's |
| 10:50.54 | TheSeven | i have no idea what's outside of /dev/nand, but inside it there's just a fat16 partition |
| 10:50.59 | lawrence_ | they expect X @ location Y on medium |
| 10:51.20 | lawrence_ | https://github.com/linux-sunxi/sunxi-tools/blob/master/nand-part.h |
| 10:51.28 | lawrence_ | and the .c |
| 10:52.04 | TheSeven | that doesn't access the raw chip |
| 10:53.04 | lawrence_ | looks to me like its same as the SD basically |
| 10:53.10 | TheSeven | to check that i just dd'ed zeros into the first 32MB of /dev/nand yesterday and it still gets as far as boot1 failing to mount the (now overwritten) fat16 partition |
| 10:53.12 | lawrence_ | 8k IPl , then 32k uboot |
| 10:53.16 | lawrence_ | then off to partitions |
| 10:53.59 | TheSeven | during a regular nand boot uboot is loaded much later than during sd card boot |
| 10:54.11 | TheSeven | most likely because they used a different boot1 |
| 10:54.28 | lawrence_ | ipl? |
| 10:54.42 | lawrence_ | (although allwinner calls it fel-boot) |
| 10:55.11 | lawrence_ | going to relook at the uboot code again |
| 10:55.35 | TheSeven | well IPL is technically the BROM, spl is boot0/boot1 |
| 10:55.44 | lawrence_ | yup |
| 10:55.48 | lawrence_ | IPL - setup chip |
| 10:55.52 | lawrence_ | SPL - uboot |
| 10:56.02 | TheSeven | it isn't for nand boot |
| 10:56.03 | lawrence_ | then uboot hands off |
| 10:56.11 | lawrence_ | thats what i'm trying to see |
| 10:56.19 | TheSeven | uboot is the 3rd stage of spl for nand boot |
| 10:56.56 | lawrence_ | the doc's are so … not there. |
| 10:57.47 | TheSeven | first is boot0 (initialize sdram, other lowlevel stuff), then boot1 (setup voltage regulators and cpu clock, mount FTL, mount FAT16, load various .axf files), then that somehow hands off to uboot |
| 10:57.57 | lawrence_ | yup |
| 10:58.26 | lawrence_ | boot0 - would be what i would call ipl probably |
| 10:58.31 | lawrence_ | < 8k to fit into chip |
| 10:58.38 | lawrence_ | and inits defaults |
| 10:58.57 | TheSeven | i don't think boot0 is inside the BROM |
| 10:59.14 | lawrence_ | well, its via boot_sel |
| 10:59.25 | lawrence_ | so it just needs to sit where the cpu expects it |
| 10:59.33 | lawrence_ | and in whatever format it thinks |
| 11:00.02 | lawrence_ | (eg default nand block size vs actual nand block size) |
| 11:00.06 | TheSeven | sure? it's unusual for the BROM part to produce serial port chatter for obvious reasons |
| 11:00.18 | lawrence_ | not sure for this SoC |
| 11:00.23 | lawrence_ | but its typical for others i've used |
| 11:00.29 | lawrence_ | so assuming its fairly similar |
| 11:00.43 | lawrence_ | as Allwinner / BoxChip / ChinaChip all seem to be quite cushy |
| 11:01.00 | lawrence_ | and thats how other chinachip stuff i've seen works |
| 11:01.03 | TheSeven | and boot0 is talking quite a bit... i suspect it's actually the first SPL stage that's sitting at 8k on the nand |
| 11:01.10 | lawrence_ | so just thinking in that direction |
| 11:01.54 | TheSeven | which attempts to load boot1 from multiple copies at fixed locations to tolerate bad blocks |
| 11:02.19 | lawrence_ | http://rhombus-tech.net/allwinner_a10/a10_boot_process/ |
| 11:02.22 | TheSeven | boot1 then mounts the actual bad block remapping and wear leveling layer |
| 11:02.37 | lawrence_ | its bascially what i thougt |
| 11:04.22 | TheSeven | that page confirms that boot0 is SPL |
| 11:05.00 | TheSeven | being on the nand chip itself, not inside the brom |
| 11:05.20 | TheSeven | i wonder what boot1 is, that page doesn't mention it |
| 11:05.40 | TheSeven | boot.axf is probably run even later |
| 11:05.43 | *** join/#cubieboard eebrah (~chatzilla@212.49.88.107) |
| 11:06.51 | TheSeven | wonders what the code that's in the sd card spl is exactly |
| 11:07.38 | lawrence_ | looks like its identical to the chinachip stuff i'm used to anyway |
| 11:13.01 | buZz | boxchip == allwinner btw |
| 11:13.48 | lawrence_ | i'm not sure on that exactly. the chinese keep saying its taiwanese (boxchip) |
| 11:14.12 | lawrence_ | but its unclear, and irrelevant anyway |
| 11:14.17 | buZz | true |
| 11:15.10 | buZz | its just sun3i is Boxchip F20, sun4i is Allwinner A10 |
| 11:15.21 | buZz | maybe they just merged or something |
| 11:15.30 | lawrence_ | and Boxchip F20 is the older ChinaChip soemthing else |
| 11:16.25 | buZz | ChinaChip something then probably is sun2i and maybe sun1i |
| 11:16.48 | lawrence_ | F20 datasheet hre - http://wenku.baidu.com/view/aef3b509844769eae009ed23.html |
| 11:17.02 | buZz | and here http://dl.linux-sunxi.org/F20/F20%20Datasheet%20-%20v1.01%20%282010-12-12%29.pdf |
| 11:18.55 | lawrence_ | SC9800 |
| 11:18.57 | TheSeven | lawrence_: https://github.com/linux-sunxi/u-boot-sunxi/blob/1076d3bdd67db39f34bc91857c636525874441ae/nand_sunxi/src/format/nand_format.c#L44 |
| 11:19.22 | lawrence_ | F10 derives from the ChinaChip SC9800 |
| 11:20.01 | lawrence_ | sorry, then 索智SC9800 (SoChip) |
| 11:20.15 | lawrence_ | which derives from the chinachip CC1800 |
| 11:20.31 | lawrence_ | i think |
| 11:20.33 | buZz | those might be sun2i and sun1i then |
| 11:20.33 | buZz | :P |
| 11:20.35 | lawrence_ | anyway, |
| 11:20.55 | lawrence_ | 合智 = BoxChip |
| 11:21.08 | lawrence_ | 索智 = SoChip |
| 11:22.33 | lawrence_ | 全志 = Allwinner |
| 11:26.12 | lawrence_ | TMSC is making the chips |
| 11:26.35 | lawrence_ | A10 is on a 55nm process |
| 11:28.02 | lawrence_ | ok, back to the boot process |
| 11:28.35 | lawrence_ | i need to make a boot 0, stick it at sector 0 on the nand, and give it magic bits. |
| 11:28.58 | lawrence_ | boot0 is my fex config for cpu |
| 11:29.08 | lawrence_ | i can build the boot0 with https://github.com/hipboi/mksunxiboot/blob/master/mksunxiboot.c |
| 11:30.07 | lawrence_ | and i can make the fex config via the sunxi-tools |
| 11:34.04 | lawrence_ | http://linux-sunxi.org/LiveSuit_images |
| 11:34.14 | lawrence_ | thats actually clearer on what to grab |
| 11:35.15 | *** join/#cubieboard titankiller2 (~titankill@117.204.93.51) |
| 11:38.04 | lawrence_ | @TheSeven |
| 11:38.18 | lawrence_ | boot1 is the uboot |
| 11:39.13 | lawrence_ | so, looking at this |
| 11:39.25 | lawrence_ | i need to wrap the ipl with the mksunxiboot.c code |
| 11:39.48 | lawrence_ | and see if that ends up looking similar to the current data i have at sector zero on my nand |
| 11:40.08 | lawrence_ | off to test nuking my nand ;) |
| 11:41.29 | TheSeven | that's uboot? http://pastie.org/6032967 |
| 11:43.33 | lawrence_ | http://linux-sunxi.org/LiveSuit_images |
| 11:43.41 | lawrence_ | These are bootloaders to be put in nand. Boot0 is the SPL - Second Program Loader (sun4i-spl.bin when uboot is built from source) |
| 11:43.41 | lawrence_ | Boot1 is the U-boot! |
| 11:43.58 | lawrence_ | according to that page, boot0 is init, boot1 is uboot |
| 11:45.39 | lawrence_ | read this |
| 11:45.40 | lawrence_ | http://forum.doozan.com/read.php?6,8904 |
| 11:47.26 | lawrence_ | mkdir /tmp/nand |
| 11:47.27 | lawrence_ | <PROTECTED> |
| 11:47.37 | lawrence_ | its vfat, and can copy files straight off |
| 11:49.21 | *** join/#cubieboard AndrevS (~andre_bk@2001:980:55e0:1:20f:eaff:fe58:28f8) |
| 11:56.56 | TheSeven | yeah I have those files... but what is all this AXF stuff? |
| 11:57.07 | lawrence_ | thats android |
| 11:57.09 | TheSeven | and how would I go about replacing the boot1 on the nand? |
| 11:57.33 | lawrence_ | you need to pack the spl code so its formatted correctly |
| 11:57.47 | lawrence_ | i'm busy breaking mine so i see how to do it |
| 11:58.19 | lawrence_ | basically |
| 11:58.40 | lawrence_ | first page of flash contains a specially formatted boot0 |
| 11:58.51 | lawrence_ | which you make with - https://github.com/hipboi/mksunxiboot/blob/master/mksunxiboot.c |
| 11:59.34 | lawrence_ | and sun4i-spl.bin |
| 12:00.05 | lawrence_ | just trying to see where i need to stick uboot |
| 12:00.13 | lawrence_ | should be @32k i think |
| 12:00.31 | lawrence_ | then its 1st bootable partition for your kernel as normal sd.. |
| 12:01.47 | lawrence_ | as i care not for android |
| 12:01.59 | lawrence_ | and wish to vape it from space |
| 12:02.20 | *** join/#cubieboard int3nz0r (~int3nz0r@86.86.66.141) |
| 12:02.50 | TheSeven | https://github.com/linux-sunxi/u-boot-sunxi/blob/1076d3bdd67db39f34bc91857c636525874441ae/nand_sunxi/src/format/nand_format.c#L44 are the boot1 positions on the nand |
| 12:03.18 | lawrence_ | ok, cool. |
| 12:14.40 | *** join/#cubieboard titankiller (~titankill@117.204.93.51) |
| 12:30.54 | *** join/#cubieboard hipboi (~hipboi@27.23.148.225) |
| 12:55.30 | TheSeven | lawrence_: do you happen to know if there's anything special about that boot1 in the factory flash image? could it be used with normal linux instead of android as well? |
| 12:55.41 | TheSeven | and how would I go about accessing the raw flash? |
| 12:55.49 | lawrence_ | raw is /dev/nand |
| 12:55.55 | TheSeven | no, it isn't |
| 12:55.59 | TheSeven | that's on top of the FTL layer |
| 12:56.16 | TheSeven | so you can't access boot0/boot1 through that |
| 12:56.46 | lawrence_ | then maybe only accessible via fel |
| 12:56.54 | lawrence_ | or we write custom code |
| 12:57.21 | buZz | lawrence_: i had great fun replying to that 'C programming' question on the ML :P |
| 12:57.46 | lawrence_ | i was a bit, eh? really? |
| 12:57.51 | buZz | yeah :D |
| 12:58.02 | lawrence_ | at him, not you |
| 12:58.06 | buZz | about 50% of question on the ML are just 'how do i linux' |
| 12:58.07 | buZz | :) |
| 13:00.40 | TheSeven | lawrence_: how do I access it through FEL? from what I undestand FEL is basically a means of uploading code to RAM and executing it |
| 13:02.44 | lawrence_ | look at the livesuite page |
| 13:02.48 | lawrence_ | grab one of the images |
| 13:02.53 | lawrence_ | and start looking at the code |
| 13:03.03 | lawrence_ | its arm, so you need an arm decompiler also |
| 13:03.12 | lawrence_ | lemme know if you need one |
| 13:03.49 | TheSeven | have something better than ida? :) |
| 13:05.10 | TheSeven | and where is the ARM code in question? |
| 13:05.23 | lawrence_ | you have the ida arm decompiler? |
| 13:05.40 | lawrence_ | hex-rays one ? |
| 13:05.45 | lawrence_ | i've been looking for that... |
| 13:06.02 | TheSeven | well i have the IDA disassembler at least, hexrays not yet but that didn't work too well on x86 either |
| 13:06.16 | TheSeven | I'm used to looking at ARM assembly, so I don't really care much about hexrays |
| 13:06.40 | lawrence_ | i use this generally - http://www.openipcam.com/files/Reverse%20Engineering/arm2html.exe |
| 13:07.38 | TheSeven | hm, need to check that out at some point |
| 13:07.41 | TheSeven | is usually on linux |
| 13:07.47 | lawrence_ | ditto |
| 13:07.56 | lawrence_ | but i don't have a decent arm decompiler |
| 13:08.01 | TheSeven | anyway, I doubt it can beat ida, being that small and freely available |
| 13:08.37 | TheSeven | so does FEL basically upload and execute some kind of flasher stub, which then receives the data to be written to the flash? or how does all of this work? |
| 13:09.18 | TheSeven | the livesuit images wiki page seems to contain all kinds of documentation, but no pointer to an actual livesuit image to look at |
| 13:09.37 | TheSeven | is there a livesuit recovery image for the cubie? |
| 13:09.44 | buZz | many |
| 13:09.56 | buZz | TheSeven: just so you know, there is nothing specific about cubie |
| 13:10.01 | buZz | compared to all the other A10 devices |
| 13:10.04 | TheSeven | i mean a factory state restore image |
| 13:10.19 | TheSeven | some things will be specific in that |
| 13:10.21 | buZz | i imaged the stock nand (just /dev/nand) from inside debian if you want? |
| 13:10.58 | TheSeven | I have that... I'm more interested in the stuff outside of the FTL, and the livesuit mechanism itself |
| 13:11.27 | buZz | did you find http://linux-sunxi.org/ already? |
| 13:11.42 | buZz | it has a lot of information already .. and can always use moar |
| 13:12.26 | TheSeven | I followed quite a bunch of links on that wiki, but it seems to lack the really interesting bits |
| 13:12.27 | *** join/#cubieboard Gumboot_ (~sh1@rev.bovine.muck.net.nz) |
| 13:12.42 | buZz | feel free to add them ;) |
| 13:12.57 | TheSeven | I'll have to figure them out first... :/ |
| 13:13.10 | lawrence_ | this is where i'm at so far http://www.cubieforums.com/index.php/topic,63.0.html |
| 13:14.41 | lawrence_ | which is tie partition -> https://github.com/linux-sunxi/linux-sunxi/blob/sunxi-3.0/drivers/block/sunxi_nand/nfd/mbr.h |
| 13:14.44 | TheSeven | "nand-part is a tool to manage the NAND partitions, while fel is a tool for writing directly to the NAND (specifying blocks, start and size)." - how does the latter work? it seemed to be much rather a tool to write to RAM from what I understood |
| 13:15.27 | lawrence_ | the fel bits talk to the SoC internal BROM as far as i see |
| 13:16.04 | TheSeven | http://pastie.org/6033166 |
| 13:16.08 | TheSeven | nothing flash related to be seen |
| 13:16.13 | RaYmAn | fel is a part of the bootrom code, yes - it lets you read, write and execute memory (initially, only SRAM/IRAM, but once ram is enabled, also ram) |
| 13:17.06 | *** join/#cubieboard Gumboot_ (~sh1@rev.bovine.muck.net.nz) |
| 13:19.37 | TheSeven | "This matches beatifully with https://github.com/linux-sunxi/u-boot-sunxi/blob/1076d3bdd67db39f34bc91857c636525874441ae/nand_sunxi/src/format/nand_format.c#L44 |
| 13:19.37 | TheSeven | That has "partition table copy locations", which match where ours are." - I think the wrong link went in there... |
| 13:20.45 | lawrence_ | work in progress (tm) |
| 13:22.17 | lawrence_ | fel needs to be used while chip is in boot_sel high |
| 13:22.27 | *** join/#cubieboard Dave77 (~dave@host-2-96-51-239.as13285.net) |
| 13:22.44 | lawrence_ | so press boot_sel, plug in otg, then query over that from external device |
| 13:23.49 | lawrence_ | @theseven - no, its correct (link) |
| 13:24.15 | lawrence_ | as nand_format says it looks for data @ X,Y,Z locations, and the hexdump shows data in those locations |
| 13:24.21 | TheSeven | I think those are the boot1 locations on the raw nand, not the partition table locations |
| 13:24.21 | lawrence_ | which was what i was trying to get at |
| 13:24.39 | TheSeven | the partition table copies are just consecutive at the start of the highlevel nand |
| 13:24.49 | lawrence_ | going to take another look |
| 13:25.46 | lawrence_ | as i see paritition tables at 0x0 , 0x400, 0x800, 0xc00, and 0x1000 has the FFF value |
| 13:29.57 | lawrence_ | of course, you're right, and i fail math |
| 13:29.59 | lawrence_ | ;) |
| 13:34.00 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 13:35.04 | lawrence_ | theseven, - the phoenix and livesuite tools can make a sd-> nand flashable image |
| 13:35.10 | lawrence_ | so its a matter of exploring those |
| 13:35.21 | lawrence_ | which is pretty much the livesuit wiki page |
| 13:36.04 | *** join/#cubieboard hipboi (~hipboi@27.23.142.72) |
| 13:37.25 | lawrence_ | 欢迎光临 :) |
| 13:41.43 | *** join/#cubieboard Gumboot (~sh1@rev.bovine.muck.net.nz) |
| 13:51.50 | *** join/#cubieboard user__ (~user@82.84.103.158) |
| 13:52.05 | *** join/#cubieboard Gumboot (~sh1@rev.bovine.muck.net.nz) |
| 13:55.45 | *** join/#cubieboard leowt (~leowt@188.80.82.63) |
| 13:56.14 | leowt | is there any cad files available for cubieboard? |
| 13:56.33 | *** join/#cubieboard hipboi (~hipboi@27.23.142.72) |
| 13:56.46 | user__ | look on website |
| 13:57.48 | Dave77 | what do I need to make the green LED light on the cubieboard in C? |
| 13:58.01 | buZz | Dave77: you need to google how to do GPIO from C in linux |
| 13:58.09 | buZz | there are many tutorials |
| 13:58.19 | buZz | also, you could just use the leds class |
| 13:58.27 | buZz | see /sys/class/leds , iirc |
| 14:00.58 | Dave77 | add something to FAQ if missing.. can't find out how |
| 14:01.40 | buZz | http://linux-sunxi.org/Cubieboard/Programming/StatusLEDs |
| 14:01.44 | buZz | you suck at searching ;) |
| 14:03.18 | Dave77 | example don't work.. no such file etc |
| 14:03.32 | buZz | have you enabled the driver in the kernel?\ |
| 14:03.54 | Dave77 | don't know how? |
| 14:03.57 | buZz | and modified your FEX to list those options? |
| 14:04.14 | *** join/#cubieboard Gumboot (~sh1@rev.bovine.muck.net.nz) |
| 14:04.21 | buZz | maybe compiling a kernel with the options needed should be on your todo list ;) |
| 14:04.42 | Dave77 | can I recompile my existing kernel? |
| 14:06.50 | Dave77 | are there any cubie SD card images that can controls LEDs out of the box? |
| 14:07.57 | buZz | Dave77: yes and yes |
| 14:08.06 | buZz | config should be in /proc/config.gz |
| 14:08.28 | buZz | grab that, download kernel source, uncompress config.gz to .config in the root of kernel tree, make oldconfig |
| 14:10.06 | *** join/#cubieboard Avernos_ (~avernos@111.192.243.1) |
| 14:10.34 | *** join/#cubieboard hi (0250bc84@gateway/web/freenode/ip.2.80.188.132) |
| 14:10.44 | hi | hi |
| 14:10.49 | buZz | hihi |
| 14:10.59 | Guest43882 | how are you |
| 14:11.09 | buZz | very adequat |
| 14:11.35 | buZz | i am cutting my fingernails |
| 14:11.44 | *** part/#cubieboard Guest43882 (0250bc84@gateway/web/freenode/ip.2.80.188.132) |
| 14:11.55 | buZz | ehe |
| 14:12.02 | buZz | that scared him off! |
| 14:13.30 | *** join/#cubieboard Gumboot (~sh1@rev.bovine.muck.net.nz) |
| 14:15.21 | *** join/#cubieboard Cy-Gor (~Brian@cpe-70-124-70-140.austin.res.rr.com) |
| 14:20.19 | hipboi | anyone tried xbmc on android? |
| 14:20.35 | buZz | which android? |
| 14:20.37 | buZz | hi hipboi :D |
| 14:21.41 | hipboi | xbmc tv box version |
| 14:21.56 | hipboi | with external video player mx player |
| 14:22.18 | hipboi | it's usable on cubieboard |
| 14:22.22 | buZz | mx player doesnt support CedarX |
| 14:22.28 | *** join/#cubieboard Gumboot (~sh1@rev.bovine.muck.net.nz) |
| 14:22.52 | hipboi | i am playing 1080p smoothly with low cpu usage |
| 14:23.14 | buZz | 1080p x264? |
| 14:23.17 | buZz | or 1080p xvid |
| 14:23.26 | buZz | or 1080p mpeg2 :D |
| 14:23.35 | hipboi | i am not sure |
| 14:23.39 | hipboi | the big bunny video |
| 14:23.41 | buZz | big difference ;) |
| 14:24.16 | buZz | anyway, what did you want to ask about XBMC on Android? |
| 14:24.39 | buZz | you wanted to include it in rel3 of android tv image? |
| 14:24.45 | hipboi | yes |
| 14:24.52 | buZz | i would say, go for it |
| 14:25.21 | buZz | those images are just for livesuit, right? not for phoenixcard? |
| 14:26.16 | hipboi | i have no problem with another 1080p video, Girls' generation mv |
| 14:26.32 | hipboi | xvid mpeg-4 |
| 14:26.32 | hipboi | 24 fps |
| 14:26.36 | hipboi | livesuite image is phonex card image |
| 14:26.42 | buZz | is it? |
| 14:26.53 | buZz | i thought you needed to convert between the two |
| 14:28.12 | hipboi | and the big buck bunny is h.264/avc 24fps |
| 14:28.31 | buZz | nice |
| 14:29.01 | rellla | hipboi: if you include it, can you do it with xbmc-pvr-addons? |
| 14:29.16 | buZz | powervr? |
| 14:29.20 | buZz | how would that help? |
| 14:29.22 | buZz | oh |
| 14:29.26 | buZz | personal video recorder |
| 14:29.27 | buZz | i bet |
| 14:29.27 | buZz | :P |
| 14:29.34 | rellla | hipboi: if you include it, can you do it with xbmc-pvr-addons?:) |
| 14:29.39 | rellla | sry |
| 14:30.17 | hipboi | xbmc-pvr-addons? |
| 14:30.41 | *** join/#cubieboard Avernos (~avernos@unaffiliated/avernos) |
| 14:30.55 | *** join/#cubieboard Gumboot (~sh1@rev.bovine.muck.net.nz) |
| 14:31.27 | rellla | https://github.com/opdenkamp/xbmc-pvr-addons |
| 14:31.53 | rellla | and https://github.com/pipelka/xbmc-addon-xvdr |
| 14:32.19 | rellla | don't if it is possible. i use it in the linux version to stream tv |
| 14:32.30 | rellla | * don't know |
| 14:35.49 | buZz | i use VDR on linux for tv streaming and recording |
| 14:38.16 | rellla | me too |
| 14:38.34 | rellla | a10 server, a10 client |
| 14:39.22 | buZz | but was hipboi right? can i just use livesuit images _directly_ with phoenixcard? |
| 14:51.08 | *** join/#cubieboard hipboi (~hipboi@27.23.142.72) |
| 14:59.26 | Gumboot | What does chinese text look like right now, then? |
| 14:59.31 | Gumboot | has finally switched from screen to tmux. |
| 15:02.17 | hipboi | 中文？ |
| 15:02.38 | Gumboot | Hm. Three boxes. |
| 15:02.40 | *** join/#cubieboard GPSFan (~kenm@64.92.145.112) |
| 15:02.42 | Gumboot | Was that three characters? |
| 15:03.01 | hipboi | chinese in chinese |
| 15:03.11 | jelly-home | and a nice wide question mark |
| 15:03.12 | hipboi | you need some chinese fonts |
| 15:03.22 | TheSeven | <buZz> but was hipboi right? can i just use livesuit images _directly_ with phoenixcard? |
| 15:04.14 | hipboi | actually there is no livesuite image or phoenix card image |
| 15:04.20 | hipboi | there is allwinner image |
| 15:04.35 | hipboi | allwinner image is used by allwinner tools |
| 15:04.41 | hipboi | livesuite and phonex card |
| 15:04.59 | Gumboot | Oh, they look right on this terminal. |
| 15:05.08 | Gumboot | ©® |
| 15:05.14 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 15:06.15 | TheSeven | so does someone know how to access the raw nand from linux? |
| 15:06.43 | Gumboot | cat /dev/nand is as raw as I've managed. |
| 15:06.56 | TheSeven | yeah, I need even more raw (mtd level) |
| 15:08.54 | buZz | cool, thanks |
| 15:17.19 | TheSeven | thinks that there is a lot of confusion... |
| 15:17.38 | TheSeven | it seems like the boot1 that the cubie ships with is *not* uboot |
| 15:19.45 | jelly-home | hi, my cubieboard doesn't power up with two different 5V power bricks, one a HTC 5V/1A that works with my phone, one a generic 3/3.3/5/6/7V/2100mA thing; both seem to give at least 4.8 when connected to the DC jack. I haven't tried power over OTG yet due to lack of miniUSB cable. |
| 15:20.42 | TheSeven | jelly-home: are you using the cable that shipped with the cubie? |
| 15:21.03 | TheSeven | that cable has a pretty high resistance (like 1.2 ohms per direction) and will make boot fail even with the most stable power supply |
| 15:21.49 | jelly-home | TheSeven: only with one power brick; the other has its own cable. |
| 15:22.05 | jelly-home | TheSeven: do you mean 1.2kOhm? |
| 15:22.20 | TheSeven | no, I mean 1.2 ohms |
| 15:22.29 | TheSeven | which is way too high for a power supply cable |
| 15:22.40 | jelly-home | why would that matter with a decent supply |
| 15:22.56 | TheSeven | because the voltage at the board will be much lower than the voltage at the power supply |
| 15:23.13 | TheSeven | it means that at 500mA there will be 1.2V voltage drop => just 3.8V at the board which isn't enough |
| 15:23.15 | jelly-home | yeah... no, I measure 5.3V idle and 4.8V when connected |
| 15:23.55 | TheSeven | it probably dips way below the 4.8V briefly while trying to boot up and then comes back up after the board gave up |
| 15:24.48 | jelly-home | where on the 2012-09-09 should I take measurements to confirm this? |
| 15:32.06 | *** join/#cubieboard debianin (~user@dynamic-adsl-94-36-210-171.clienti.tiscali.it) |
| 15:33.14 | jelly-home | so the board has a very low internal resistance when powered up, like 3.8Ohm? |
| 15:45.16 | TheSeven | well it consumes a certain amount of current |
| 15:45.20 | TheSeven | up to about 1A I think |
| 15:45.29 | TheSeven | and doesn't tolerate low voltage very well |
| 15:55.48 | jelly-home | huh... blimey, found an old mp3 player supply with miniUSB and the board lights up |
| 15:57.50 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 15:58.34 | TheSeven | lawrence_: http://permalink.gmane.org/gmane.comp.hardware.netbook.arm/298 |
| 16:03.50 | jelly-home | what about too high voltage, ie. if I switch the generic power supply to say 6V? |
| 16:05.34 | TheSeven | jelly-home: you need a stable power supply, with an inner resistance (including cable) of less than 0.2 ohms if you want the board to work properly |
| 16:05.34 | TheSeven | the upper limit of what's tolerated is usually somewhere around 5.5V |
| 16:08.11 | jelly-home | well crap, this means I need yet another PS (or the replacement cable) |
| 16:08.39 | jelly-home | on the other hand, I don't have to return the board which seems to work after all :-) |
| 16:12.47 | *** join/#cubieboard DARLAS (~davidkarl@BSN-176-200-62.dial-up.dsl.siol.net) |
| 16:16.58 | *** join/#cubieboard rz2k (~rz2k@95-27-13-50.broadband.corbina.ru) |
| 16:17.11 | Turl | (12:06:56) TheSeven: yeah, I need even more raw (mtd level) <- our kernel doesn't use mtd |
| 16:17.34 | Turl | allwinner implemented their own bad blocks and stuff stack and exposes it as block device :< |
| 16:21.51 | *** join/#cubieboard eebrah (~chatzilla@212.49.88.103) |
| 16:25.22 | TheSeven | I know, but I want to get at boot0/boot1 which are outside of that FTL layer |
| 16:38.55 | *** join/#cubieboard debianin (~user@93.35.233.182) |
| 17:05.11 | *** join/#cubieboard bfree_ (~niall@109.255.245.157) |
| 17:05.41 | TheSeven | http://pastie.org/6033911 |
| 17:05.53 | TheSeven | getting closer... |
| 17:14.22 | *** join/#cubieboard bfree (~niall@109.255.245.157) |
| 17:42.45 | lawrence_ | lemme know when you see eGON.BT0 in your boot0 :) |
| 17:42.52 | lawrence_ | as thats the magic |
| 17:45.15 | *** join/#cubieboard eebrah (~chatzilla@212.49.88.103) |
| 17:46.08 | TheSeven | i know where that boot0 is, just not how to access that area |
| 17:46.24 | TheSeven | next up: build uboot with nand drivers |
| 17:46.28 | *** join/#cubieboard Davn (2e928376@gateway/web/freenode/ip.46.146.131.118) |
| 17:46.40 | lawrence_ | or just reuse the one from the android boot |
| 17:46.41 | *** join/#cubieboard txt_file (~txt@i59F6B1CD.versanet.de) |
| 17:46.44 | lawrence_ | as that already has |
| 17:46.55 | lawrence_ | then its |
| 17:46.55 | lawrence_ | setenv bootargs 'console=${console} root=${nand_root} init=${init} loglevel=${loglevel} mac_addr=${mac}' |
| 17:46.56 | lawrence_ | fatload nand 0:1 0x43000000 script.bin; fatload nand 0:1 0x48000000 linux/uImage;bootm 0x48000000 |
| 17:47.07 | TheSeven | the one from android is stoneage |
| 17:47.09 | TheSeven | from some time in 2011 |
| 17:47.15 | lawrence_ | yes, but it does nand |
| 17:47.22 | lawrence_ | and the current one i see doesn't |
| 17:47.22 | TheSeven | and it has some odd embedded commands |
| 17:47.28 | lawrence_ | although it does have zfs |
| 17:47.34 | lawrence_ | for some wacky reason |
| 17:48.00 | TheSeven | script.bin is already loaded by boot.axf |
| 17:48.08 | txt_file | does the cubieboard support accelerated video playback on linux? |
| 17:49.19 | lawrence_ | if you read from the uboot do you see the boot0 data? |
| 17:49.22 | lawrence_ | i haven't tried that yet |
| 17:49.24 | lawrence_ | you tried? |
| 17:49.34 | lawrence_ | eg uboot, then from there read nand |
| 17:51.50 | *** join/#cubieboard Cygor (~Brian@cpe-70-124-70-140.austin.res.rr.com) |
| 17:53.23 | TheSeven | lawrence_: the uboot that I currently have in there has no nand driver |
| 17:53.39 | lawrence_ | i'm just repartitioning mine now |
| 17:53.58 | lawrence_ | once i'm done will check via my uboot over ttl |
| 17:56.49 | TheSeven | swivelgames: you need a kernel patch to expose the /dev/nand device |
| 17:57.08 | TheSeven | which is included in the latest linaro ubuntu image it seems |
| 17:57.14 | swivelgames | TheSeven, hmmm |
| 17:57.34 | lawrence_ | theseven - we can't get at the boot0/1 |
| 17:57.35 | TheSeven | at least if installed through berryboot |
| 17:57.48 | TheSeven | lawrence_: I'm sure we can - just not that easily |
| 17:57.50 | lawrence_ | but if you want to clear out stuff, easy enough |
| 17:57.56 | lawrence_ | *i know we can |
| 17:58.05 | lawrence_ | i'm just trying to explain what he can do |
| 17:58.14 | swivelgames | I'm starting to think I should have bought a cubieboard, instead of a OLinuXino board... |
| 17:58.17 | swivelgames | lol :P |
| 17:58.27 | lawrence_ | so swivelgames - |
| 17:58.32 | lawrence_ | install sunxi-tools |
| 17:58.34 | lawrence_ | compile |
| 17:58.41 | swivelgames | Done and done |
| 17:58.42 | lawrence_ | then from that folder in the cubie |
| 17:58.50 | lawrence_ | ./ nand_part /dev/nand 'linux 8000000' |
| 17:58.56 | lawrence_ | then reboot |
| 17:59.11 | lawrence_ | that will wipe the nandb and onward partitions |
| 17:59.13 | Gumboot | swivelgames: I haven't followed this discussion. Is there something behind that statement which I could repeat to anybody asking me what kind of board they should buy? |
| 17:59.36 | lawrence_ | actually before you do anything i document first on mine |
| 17:59.40 | lawrence_ | then you can repeat if you wnat |
| 17:59.42 | lawrence_ | ;) |
| 18:01.26 | swivelgames | Gumboot, I purchased an A13 OLinuXino board... love the board, but there is almost no community or resources surrounding it. I can't find anything to help me flash this thing except for people saying "Use LiveSuit!"... It is very frustrating, especially since LiveSuit isn't available for my linux distro. |
| 18:01.37 | lawrence_ | swivel |
| 18:01.40 | lawrence_ | gimme 10 |
| 18:01.44 | lawrence_ | i have instructions for you |
| 18:01.48 | swivelgames | lawrence_ sure thing! Thanks! :D |
| 18:02.15 | lawrence_ | essentially both theseven and i want at lower level access than you need |
| 18:02.17 | swivelgames | Gumboot, I have found a plethora of resources and the cubieboard community has actually engaged with me and seems very knowledgable. |
| 18:02.19 | TheSeven | swivelgames: if you can boot from an SD card, I just wouldn't care about livesuit for now |
| 18:02.43 | lawrence_ | but you don't need that access, and can write to the nand without hassling |
| 18:02.52 | lawrence_ | i'm just doing that now (writeup) for you |
| 18:02.59 | swivelgames | lawrence_ perfect! :P |
| 18:03.13 | lawrence_ | also - the script.bin is per device |
| 18:03.13 | swivelgames | TheSeven, Yeah, just don't have an SD card atm, sadly. Not in the position to run out and get one for a while either :/ |
| 18:03.18 | lawrence_ | so keep yours and put safe. |
| 18:03.42 | swivelgames | TheSeven, besides that, I'd rather have it on NAND anyway. |
| 18:03.58 | swivelgames | lawrence_ gotcha. I saved my and tucked it away in a safe place :P |
| 18:04.08 | TheSeven | swivelgames: I meant that booting from an SD card is the easier way to flash the NAND than using livesuit |
| 18:04.10 | lawrence_ | in fact your whole /dev/nanda |
| 18:04.20 | lawrence_ | as thats your boot partition |
| 18:04.24 | lawrence_ | copy / backup |
| 18:05.40 | TheSeven | lawrence_: anything except for the mac address being device specific in there? |
| 18:05.50 | TheSeven | (i don't care about model specific stuff for now) |
| 18:06.07 | lawrence_ | well for him, i'll reuse the existing boot.axf BS |
| 18:06.16 | lawrence_ | and stick a uboot / uImage in the linux folder |
| 18:06.20 | TheSeven | ok, looks like i've built a working nand uboot: http://pastie.org/6034305 |
| 18:06.30 | lawrence_ | and then partition the nandb and on as a ext4 |
| 18:06.34 | TheSeven | yeah that's what I'm doing right now as well |
| 18:06.35 | lawrence_ | and have him copy over FS |
| 18:06.46 | swivelgames | lawrence_, I did a ./fel read, is that enough to backup? |
| 18:06.49 | lawrence_ | i'm just doc'ing that, as mine appears to not have farked up. |
| 18:06.51 | lawrence_ | :) |
| 18:07.04 | lawrence_ | i'm not 100% yet, but i did get mine in nand |
| 18:07.16 | lawrence_ | although again, still haven't touched existing boot0/1 yet |
| 18:07.33 | lawrence_ | needs to document, so won't talk for a few min |
| 18:08.29 | TheSeven | swivelgames: I don't think fel read does read from the flash |
| 18:09.22 | swivelgames | TheSeven, really? bummer... whats the best way to backup a block? |
| 18:09.43 | TheSeven | for inside-ftl stuff: do it from within linux |
| 18:09.48 | lawrence_ | 1st block on flash is guaranteed for 1000 writes |
| 18:09.58 | lawrence_ | so thats probably our real boot0 / boot1 area |
| 18:10.06 | lawrence_ | or at least boot0 area |
| 18:10.16 | lawrence_ | as that boot0 can't really do too much |
| 18:10.22 | lawrence_ | and cpu expects data at X |
| 18:10.30 | lawrence_ | so thats where i think it probably lies |
| 18:10.44 | buZz | 19:03:35 < lawrence_> so keep yours and put safe. |
| 18:10.46 | lawrence_ | underneath the abstraction layer |
| 18:10.52 | buZz | there is a public repo with script.fex |
| 18:10.56 | buZz | for numerous devices |
| 18:11.08 | lawrence_ | we can't touch it from the nand we get shown at the driver level |
| 18:11.12 | lawrence_ | we need mtd level really |
| 18:11.17 | lawrence_ | as its abstracted |
| 18:11.23 | buZz | https://github.com/linux-sunxi/sunxi-boards/tree/master/sys_config/a10 |
| 18:11.30 | buZz | just FYI |
| 18:11.38 | lawrence_ | not so useful |
| 18:11.52 | Gumboot | It was script.fex I was meant to tweak to disable HDMI output, right? |
| 18:11.58 | lawrence_ | yes |
| 18:12.10 | lawrence_ | bin2fex script.bin |
| 18:12.11 | lawrence_ | edit |
| 18:12.17 | lawrence_ | fex2bin script.fex |
| 18:12.18 | lawrence_ | replace |
| 18:12.34 | lawrence_ | and add the dest name which i forgot |
| 18:15.30 | Gumboot | XXX_used=0, is it? |
| 18:15.44 | lawrence_ | its in the wiki |
| 18:15.44 | Gumboot | dac_used=0, lcd1_used=0, etc. |
| 18:15.55 | lawrence_ | search for FEX |
| 18:18.16 | Gumboot | That's not ideal documentation. Am I better to enable the initialisation of the display so as to ensure that screen output types are set to 0? |
| 18:18.22 | *** join/#cubieboard swivelgames (~swivelgam@172-4-81-47.lightspeed.rcsntx.sbcglobal.net) |
| 18:18.37 | Gumboot | Or do I disable disp_init, so that nothing is initialised at all? |
| 18:19.00 | Gumboot | Kind of depends on what the power-on defaults are, and whether or not initialising the display inherently activates something which will consume unnecessary bandwidth regardless of the output types. |
| 18:19.47 | swivelgames | Not sure why I got kicked off |
| 18:19.52 | *** join/#cubieboard asiekierka (~asiekierk@fluttershy.pl) |
| 18:22.15 | Gumboot | swivelgames: Using a mac? |
| 18:24.47 | Gumboot | puts the new script.bin in place and leaves it there ready to confuse the hell out of himself next time he reboots. |
| 18:25.42 | Turl | swivelgames: "quit: Quit: This computer has gone to sleep" |
| 18:26.59 | swivelgames | Turl, thanks. I think my son closed my laptop and opened it again. Explains why :P |
| 18:29.02 | Gumboot | What you want to do, there, is to put something like a cubieboard in a cupboard, running irssi, and to simply attach to that. |
| 18:29.33 | Gumboot | That's the sort of thing I've been spending this whole weekend on. Killing off all those always-on PCs and replacing them with cubieboards. |
| 18:29.51 | Turl | nice |
| 18:30.10 | Gumboot | And having it in a cupboard means I want to kill off all the display-related logic so it doesn't draw power or memory bandwidth. |
| 18:30.55 | Turl | I've got a VPS for my network always-on stuff and a mele for my "I'd rather run in-house" always-on stuff |
| 18:31.18 | Turl | Gumboot: then disable disp and mali on kernel build and you should be good |
| 18:31.30 | Gumboot | Oh god... the kernel too? |
| 18:31.56 | Turl | rm has server kernels on his page too |
| 18:32.07 | Turl | well, you can use a hwpack and not load the modules too |
| 18:32.17 | Gumboot | How much of the hardawre is supported by mainline kernels? I pretty much just need serial and ethernet, here. |
| 18:32.21 | Gumboot | And SATA. |
| 18:32.30 | Gumboot | Serial, ethernet, and SATA. And I don't really need serial. |
| 18:32.34 | Turl | mainline 3.8? booting |
| 18:32.38 | Turl | and serial |
| 18:32.52 | Gumboot | "booting"? |
| 18:32.55 | lawrence_ | ack, my nanda didn't save changed |
| 18:33.03 | Turl | Gumboot: yeah |
| 18:33.11 | Turl | Gumboot: 3.9 will have gpio too, and some clock support |
| 18:33.13 | lawrence_ | gonna redo, then sync, then make sure it farking keeps it before rebooting |
| 18:33.16 | Turl | hopefully |
| 18:33.27 | Gumboot | So there's no ethernet or SATA in mainline? |
| 18:33.32 | Turl | nope |
| 18:33.35 | Gumboot | OUch. |
| 18:33.58 | Turl | stefanro1 was working on ethernet last I heard |
| 18:34.54 | Turl | Gumboot: http://linux-sunxi.org/Mainlining_Effort |
| 18:36.58 | Gumboot | Neither ethernet nor sata are actually mentioned on that page. Are they just "Device Tree"? |
| 18:37.20 | Gumboot | I think Device Tree might be something else, mightn't it. |
| 18:37.42 | Turl | device tree is most of the "booting" part |
| 18:37.51 | Turl | indeed, they're missing |
| 18:38.01 | Turl | I believe they should be under "DMA dependent drivers" |
| 18:38.55 | lawrence_ | ok, i got mine booting off nand |
| 18:39.44 | lawrence_ | http://www.cubieforums.com/index.php/topic,63.msg221.html#msg221 |
| 18:43.44 | lawrence_ | swivelgames |
| 18:43.57 | lawrence_ | http://www.cubieforums.com/index.php/topic,63.msg221.html#msg221 |
| 18:44.00 | lawrence_ | appears to work for me |
| 18:44.15 | lawrence_ | will clean that up more and make into a proper tips n tricks page |
| 18:45.41 | lawrence_ | i also need to change my nand uboot.bin to use a 1G one, as mine is not currently |
| 18:46.10 | lawrence_ | mr TheSeven - yours working yet (uboot) |
| 18:46.25 | lawrence_ | you can save me recompiling if yours is done |
| 18:46.29 | TheSeven | somewhat... not quite doing what I want yet |
| 18:46.36 | TheSeven | the problem is that lichee-dev is fairly old |
| 18:46.44 | lawrence | yeah |
| 18:46.48 | lawrence | i have mine booted up |
| 18:46.49 | TheSeven | I'm pondering to port the nand driver over to hno's latest sunxi uboot |
| 18:47.00 | lawrence | although the uboot is only 512m (the one off miniand) |
| 18:47.25 | lawrence | i'm writing it up better first, hopefully you'll have that done by the time i get to that part again ;) |
| 18:47.47 | lawrence | i'm not sure why they removed the nand drivers from the uboot |
| 18:47.50 | lawrence | seems a bit wierd |
| 18:48.02 | lawrence | we have zfs, but no nand drivers. wtf.. |
| 18:48.10 | lawrence | seems a strange choice to me! |
| 18:48.17 | lawrence | although i love having zfs |
| 18:54.22 | Turl | lawrence the nand driver was awful code, certainly not acceptable for mainlining |
| 18:55.05 | lawrence | would rather have nand than no-nand |
| 18:55.17 | Turl | if you want to use nand, there's lichee-dev branch |
| 18:55.18 | Turl | https://github.com/linux-sunxi/u-boot-sunxi/tree/lichee-dev |
| 18:56.38 | TheSeven | well, that's kinda stoneage... |
| 18:58.36 | TheSeven | ponders to throw it all out |
| 18:58.45 | Turl | hno and slapin were working on a new nand driver |
| 18:59.18 | TheSeven | might make most sense to drop the whole FTL and make it work with an MTD driver |
| 18:59.29 | lawrence | ++ on that |
| 18:59.46 | TheSeven | an FTL clearly has its advantages, but this one is a bit weird |
| 19:00.20 | Turl | yeah, they're doing it "the right way" :) |
| 19:01.33 | TheSeven | Turl: so another FTL approach... will the on-flash format be compatible, or something entirely new? |
| 19:01.37 | TheSeven | or some kind of industry standard? |
| 19:01.42 | *** part/#cubieboard MarkusBec (znc@80.69.39.131) |
| 19:02.15 | Turl | they're going for mtd support I think |
| 19:05.22 | TheSeven | hno: what's the difference between linux-sunxi's sunxi uboot branch, and your sunxi-patchqueue branch? is the latter an upstreaming effort? are all improvements/fixes from there also in the normal sunxi branch? |
| 19:07.26 | pidpawel | TheSeven: you were the one who wanted BROM dump? http://pliki.pidpawel.eu/random/brom1.dump |
| 19:07.54 | *** join/#cubieboard Dave77 (~dave@host-2-96-51-239.as13285.net) |
| 19:11.39 | swivelgames | lawrence, those instructions would only help if I was trying to flash the SD card though, right? |
| 19:11.50 | lawrence | no, works for nand |
| 19:11.52 | lawrence | just did it |
| 19:11.54 | swivelgames | oh ok |
| 19:11.57 | lawrence | redoing again just to double check |
| 19:12.24 | swivelgames | Ok... you're just using a flashed SD to move it over, right? |
| 19:12.24 | *** join/#cubieboard RavenII (~RavenII@c-76-29-25-66.hsd1.il.comcast.net) |
| 19:12.37 | lawrence | http://www.cubieforums.com/index.php/topic,72.0.html |
| 19:12.48 | lawrence | the main thing is to wait for the nand to finish writing |
| 19:13.02 | lawrence | if you don't sync, and reboot too soon it many not have |
| 19:13.06 | lawrence | thats a bit buggy imho |
| 19:13.15 | lawrence | the nand code should block till finished |
| 19:13.38 | swivelgames | Gotcha |
| 19:13.48 | lawrence | so, you can repeat - worked for me |
| 19:13.59 | *** join/#cubieboard titankiller (~titankill@117.204.93.51) |
| 19:14.00 | lawrence | However you need to make sure that the uboot.bin is compat for you |
| 19:14.02 | lawrence | that i used |
| 19:14.10 | swivelgames | The sun5i images are for the entire NAND, right? |
| 19:14.27 | lawrence | if not, you'll need to recompile one for the a13 via the lychee uboot code |
| 19:14.27 | lawrence | dunno |
| 19:14.34 | swivelgames | k |
| 19:14.45 | swivelgames | Can I boot from USB instead of SD/ |
| 19:14.45 | BigWhoop | should i download an irc client? :) |
| 19:14.46 | swivelgames | ? |
| 19:14.47 | lawrence | essentially we're keeping NANDa as thats our boot loader (we reuse the existing) |
| 19:14.48 | swivelgames | I assume not |
| 19:14.54 | lawrence | and wiping from NANDb onward |
| 19:15.00 | swivelgames | lawrence indeed |
| 19:15.05 | lawrence | then redoing NANDb as ext4 |
| 19:15.19 | lawrence | and copying kernel and a uboot and a script.bin into the NANDa |
| 19:15.29 | lawrence | and a FS into NANDb |
| 19:15.35 | lawrence | thats the long and short of it |
| 19:16.39 | lawrence | i've tested and retested on my cubie |
| 19:17.00 | lawrence | the only gotcha for me was mounting / remounting to double check it finished writing to the nand |
| 19:17.05 | swivelgames | oh ok |
| 19:17.10 | lawrence | for you, not sure the uboot i used will work for you |
| 19:17.32 | lawrence | suggest you don't nuke the nandb -> g |
| 19:17.37 | swivelgames | Indeed. Different device and all... might not work. But, I may still be able to manage. |
| 19:17.43 | lawrence | and first copy the uboot i used over |
| 19:17.45 | lawrence | and a kernel and check |
| 19:17.50 | lawrence | as thats non-destructive |
| 19:18.01 | lawrence | especially if you backed up and renamed like in my instructions |
| 19:18.03 | swivelgames | Indeed |
| 19:18.13 | lawrence | so do that, see if it works for you |
| 19:18.29 | lawrence | if so, then nuke --with-extreme-prejudice /nandb-> nandg |
| 19:18.32 | lawrence | ;) |
| 19:18.58 | lawrence | i have mine working off nand, it see's 1G and i'm good :) |
| 19:19.22 | BigWhoop | cubiboards in stock for germany :) yea |
| 19:19.23 | lawrence | now i'm going back into the uboot to read nand raw if possible ;) |
| 19:19.48 | swivelgames | BigWhoop, hoping to hear that about usa here soon! |
| 19:19.51 | lawrence | also - you'll need a ttl cable |
| 19:19.59 | lawrence | otherwise its hard to debug the uboot ;) |
| 19:20.05 | swivelgames | Yeah, true.. |
| 19:20.10 | BigWhoop | after new year swivel :) |
| 19:20.39 | swivelgames | BigWhoop, awesome! |
| 19:23.37 | swivelgames | BigWhoop, do you know if fullset will be available after the 18th? |
| 19:23.41 | swivelgames | full set* |
| 19:23.49 | BigWhoop | nope |
| 19:24.08 | swivelgames | Ah, that'll be a little later then? |
| 19:24.16 | swivelgames | Or you don't know :P |
| 19:24.26 | BigWhoop | any hints for a working ttl usb plug? |
| 19:24.32 | BigWhoop | dont know that. |
| 19:25.13 | swivelgames | I know Adafruit has one, BigWhoop |
| 19:25.20 | swivelgames | http://www.adafruit.com/products/954?gclid=CJGJ5_jumrUCFVSSPAod1kwAGQ |
| 19:25.26 | lawrence | any FTDI based is ok |
| 19:25.33 | lawrence | or PL2303 |
| 19:25.39 | lawrence | FTDI best |
| 19:25.43 | lawrence | PL2303 acceptable |
| 19:26.27 | swivelgames | AdaFruit's is PL2303HX.A, not FTDI. Might keep looking |
| 19:26.44 | swivelgames | http://www.adafruit.com/products/70 |
| 19:27.21 | Turl | I've got a CPsomething one and it works fine |
| 19:27.30 | Turl | CP2102 maybe |
| 19:27.32 | swivelgames | I'll be back, grabbing lunch :P |
| 19:28.49 | BigWhoop | me, too. |
| 19:30.34 | lawrence | i'm going to sleep |
| 19:30.36 | lawrence | 3:30am again |
| 19:34.54 | lawrence | theseven, if you want to ship me a uboot nand that does 1M, feel free. my name at computersolutions dot cn |
| 19:38.11 | *** join/#cubieboard emilepetrone (~emilepetr@c-24-22-29-165.hsd1.or.comcast.net) |
| 19:38.38 | *** join/#cubieboard eebrah (~chatzilla@105.230.158.247) |
| 19:42.01 | rvalles | just make sure it's at 3.3v |
| 19:42.31 | rvalles | I don't know if it's 5v tolerant but almost everything uses 3.3v and few are 5v tolerant so you want it to be 3.3v. |
| 19:44.05 | BigWhoop | this one? http://www.ebay.de/itm/FTDI-Basic-6-PIN-3-3-5V-For-Arduino-Free-USB-Cable-/230820610037?pt=LH_DefaultDomain_0&hash=item35bdfafff5 ? |
| 19:44.53 | BigWhoop | or this? http://www.ebay.de/itm/FTDI-Basic-Breakout-Arduino-5V-USB-TTL-6-PIN-3-3-for-MWC-MultiWii-Lite-/300776726903?pt=US_Character_Radio_Control_Toys&hash=item4607b0a177 |
| 19:51.09 | buZz | http://www.ebay.de/itm/FTUSB-USB-RS232-converter-3-3V-signal-level-3V3-FTDI-FT230X-from-EU-/170969733233?pt=LH_DefaultDomain_0&hash=item27ce971871#ht_1052wt_1190 |
| 19:51.18 | buZz | i would grab that |
| 19:54.50 | rvalles | either grab whatever cheapest (but ideally ftdi) or spend a little more and grab a http://dangerousprototypes.com/docs/Bus_Pirate |
| 19:55.45 | rvalles | also an option http://www.seeedstudio.com/wiki/Seeeduino_v3.0#Introduction |
| 19:55.49 | rvalles | when set to 3.3v |
| 19:55.57 | f1r3 | too pity that the ftdi cable does not fit 1:1 to the cb |
| 19:56.10 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 19:56.15 | rvalles | these are convenient only if you think you can put them to use in other ways |
| 19:56.26 | rvalles | (they're useful for stuff other than serial bridge) |
| 19:56.57 | rvalles | else grabbing a very cheap ft232r-based serial bridge is best |
| 19:57.06 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 19:57.52 | f1r3 | i use one of this kind with the plug rewired to fit cb |
| 19:58.58 | f1r3 | you can carefully pull out the cables of the plug, reoder them to fit the cb |
| 19:59.03 | f1r3 | http://www.ebay.de/itm/USB-to-TTL-Serial-Cable-FTDI-chipset-/300844503937?pt=UK_Computing_Parallel_Serial_PS_2&hash=item460bbad381 |
| 19:59.48 | Turl | 13 eur looks like expensive |
| 20:00.14 | f1r3 | it is |
| 20:00.40 | Turl | I cannot read german, but looks like cheaper http://www.ebay.de/itm/CP2102-Module-Modul-USB-to-TTL-Converter-Konverter-/280736990675?pt=DE_Computing_USB_Kabel_Hubs_Adapter&hash=item415d3a85d3 |
| 20:01.32 | Turl | I like those with pins, and then just use dupont cables to connect them to the devices |
| 20:03.46 | TheSeven | just used an ft232-based adapter and a simple 2-transistors discrete level shifter that he hacked up last night |
| 20:05.11 | TheSeven | and ftdi just IS expensive |
| 20:05.21 | TheSeven | but also much better than all this prolific crap |
| 20:08.51 | Gumboot | I don't think I've ever seen a non-prolific USB-232 adapter. |
| 20:09.18 | Gumboot | 'cept maybe once at work. A marvell one? |
| 20:10.52 | *** join/#cubieboard asiekierka (~asiekierk@fluttershy.pl) |
| 20:12.06 | *** join/#cubieboard Cy-Gor (~Brian@cpe-70-124-70-140.austin.res.rr.com) |
| 20:14.03 | lawrence_ | FTDI is best. |
| 20:14.11 | lawrence_ | spend the 3$ for one |
| 20:14.47 | buZz | just not all ftdi are configured for 3v3 levels |
| 20:14.53 | buZz | something to be aware of |
| 20:15.00 | buZz | neither are any of the others |
| 20:15.10 | buZz | you need to be aware of the levels your cable outputs :D |
| 20:15.23 | buZz | also, they should match the system you want to connect to ;) |
| 20:16.57 | lawrence_ | http://www.ftdichip.com/Products/ICs/FT232R.htm |
| 20:17.00 | lawrence_ | data sheet says |
| 20:17.15 | lawrence_ | 5v , 3.3v, 2.8v, 1.8v TTL input |
| 20:17.48 | lawrence_ | so what kind of FTDI chip are you using buZz |
| 20:17.55 | buZz | i am not using the chip |
| 20:17.55 | lawrence_ | some lame ass cheap one? ;) |
| 20:18.08 | buZz | i am using the cable that came with the cubieboard rich package |
| 20:18.23 | buZz | hmm do i have that here? |
| 20:18.26 | buZz | searches |
| 20:18.30 | lawrence_ | thats a PL2303 |
| 20:18.44 | lawrence_ | those are crap for flow control |
| 20:18.59 | lawrence_ | although really not needed for us as we no haz flow control |
| 20:19.12 | lawrence_ | but i have had issues with them before for that |
| 20:19.14 | buZz | Bus 007 Device 002: ID 067b:2303 Prolific Technology, Inc. PL2303 Serial Port |
| 20:19.15 | lawrence_ | buffer underruns |
| 20:19.17 | buZz | yes |
| 20:19.20 | lawrence_ | i know |
| 20:19.24 | lawrence_ | i'm all knowing |
| 20:19.26 | lawrence_ | ;) |
| 20:19.29 | buZz | i have issues with it aswell ;) |
| 20:19.35 | lawrence_ | and non sleeping compatible |
| 20:19.42 | buZz | well, i dont use it |
| 20:19.53 | buZz | my current debian rootfs doesnt even have serial terminal ;) |
| 20:19.57 | lawrence_ | essentially buy an adaptor with FT232R |
| 20:20.05 | lawrence_ | or go with the golden MAX232 |
| 20:20.08 | buZz | yeah i have some ftdis aswell |
| 20:20.09 | lawrence_ | which i never have issues with |
| 20:20.16 | buZz | arduino boards etc |
| 20:20.24 | lawrence_ | but then again, finding serial on modern laptops is nigh on impossilbe |
| 20:20.31 | lawrence_ | my thinkpads have |
| 20:20.38 | lawrence_ | but i build mine from scratch |
| 20:20.43 | lawrence_ | (thinkpads) |
| 20:20.45 | buZz | this one has none :( |
| 20:20.54 | buZz | lawrence_: wow nice |
| 20:21.06 | buZz | build me a 8" 1080P thinkpad with core i7 plz |
| 20:21.15 | lawrence_ | maybe |
| 20:21.21 | buZz | and 2*1tb in raid, and CVBS/Svideo |
| 20:21.22 | buZz | ;) |
| 20:21.24 | lawrence_ | i can only get 10" chassis though |
| 20:21.28 | lawrence_ | and eDP in that size |
| 20:21.34 | buZz | eDP is? |
| 20:21.42 | lawrence_ | so need to find a motherboard in that size+- with eDP out |
| 20:21.48 | lawrence_ | DP for small people |
| 20:21.49 | buZz | so you buy chassis and motherboard? |
| 20:21.50 | lawrence_ | ;) |
| 20:21.57 | buZz | ah displayport |
| 20:21.59 | buZz | no no no |
| 20:22.02 | buZz | _real_ analog video |
| 20:22.04 | buZz | needs it |
| 20:22.07 | lawrence_ | these days? |
| 20:22.09 | lawrence_ | unpossible |
| 20:22.09 | buZz | yes |
| 20:22.15 | lawrence_ | lvds or eDP |
| 20:22.20 | lawrence_ | or TTL |
| 20:22.21 | buZz | i use an analog video setup for VJ-ing |
| 20:22.30 | lawrence_ | i'm talking about raw panel |
| 20:22.31 | buZz | yeah so i'll stick to my core2duo dell ;) |
| 20:22.35 | buZz | oh |
| 20:22.39 | buZz | i am talking about ports ;) |
| 20:22.42 | lawrence_ | you stick a driver on it anyway (assuming lvds) |
| 20:22.58 | buZz | lvds -> driver -> panel? |
| 20:23.22 | buZz | oh, i need the analog output to be a second framebuffer (at least) |
| 20:23.22 | *** join/#cubieboard asiekierka (~asiekierk@fluttershy.pl) |
| 20:23.28 | lawrence_ | well motherboards generally do lvds |
| 20:23.30 | lawrence_ | in lappers |
| 20:23.33 | buZz | ya |
| 20:23.36 | lawrence_ | so lvds on mobo -> panel |
| 20:23.42 | lawrence_ | or eDP on mobo -> panel |
| 20:23.43 | buZz | but is there a pcb between lvds and panel? |
| 20:23.45 | buZz | yeah, ok |
| 20:23.53 | buZz | but |
| 20:23.55 | buZz | basically |
| 20:24.05 | buZz | you buy thinkpad chassis, thinkpad motherboard |
| 20:24.10 | buZz | and just add parts |
| 20:24.12 | lawrence_ | sure |
| 20:24.20 | buZz | no i mean, thats what you do? |
| 20:24.31 | lawrence_ | no, its not what i do |
| 20:24.34 | lawrence_ | but i have done |
| 20:24.42 | buZz | lol, pf, ok :) |
| 20:24.54 | lawrence_ | mostly as i has qxga panels |
| 20:24.57 | buZz | _thats_ what you ment with 21:20:55 < lawrence_> but i build mine from scratch |
| 20:25.30 | lawrence_ | i find a post referencing me on thindpads.com |
| 20:25.31 | lawrence_ | deng deng |
| 20:26.15 | buZz | :D |
| 20:26.35 | lawrence_ | http://forum.thinkpads.com/viewtopic.php?f=29&t=104956 |
| 20:26.43 | buZz | 2048×1536 |
| 20:26.47 | buZz | so nice |
| 20:27.47 | lawrence_ | so you see, i don't kid, when i say from scratch ;) |
| 20:28.48 | lawrence_ | i'll probably do a test arm based 10" using eDP display after CNY |
| 20:28.56 | lawrence_ | with thinkpad chassis |
| 20:29.04 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 20:29.12 | lawrence_ | was brainstorming that with mnemoc last week on #arm-laptop or whatever it is |
| 20:30.01 | Turl | -netbook |
| 20:30.03 | buZz | which size are those qxga screens? |
| 20:30.12 | buZz | arm-deskbook |
| 20:30.13 | buZz | :P |
| 20:33.09 | *** part/#cubieboard txt_file (~txt@i59F6B1CD.versanet.de) |
| 20:35.37 | *** join/#cubieboard eebrah (~chatzilla@105.230.126.63) |
| 20:38.41 | lawrence_ | qxga is 15" |
| 20:39.03 | lawrence_ | i can get eDP ones in 9" for 400rmb these days |
| 20:39.21 | lawrence_ | eDP QXGA 9" (iPad3) screens are 400rmb odd |
| 20:39.33 | lawrence_ | or 10" whatever that resolution is |
| 20:39.42 | lawrence_ | anyway, almost 5am, off to sleep |
| 20:39.49 | lawrence_ | have a meeting at 2pm tomorrow |
| 20:40.10 | Turl | lawrence_: your email still arrives as spam :/ |
| 20:40.35 | lawrence_ | meh. |
| 20:40.37 | Turl | although not all of it these days |
| 20:40.47 | lawrence_ | we do spf, dmarc, signing etc |
| 20:41.04 | Turl | DKIM? |
| 20:41.13 | lawrence_ | not dkim, as too painful |
| 20:41.19 | lawrence_ | anyway sleep time for loz |
| 20:41.21 | lawrence_ | night |
| 20:41.34 | Turl | it says it couldn't verify you sent it, it probably wants DKIM :/ |
| 20:41.39 | Turl | night |
| 20:44.30 | TheSeven | wow, now that was straightforward |
| 20:44.48 | TheSeven | porting over the nand driver from lichee to current uboot-sunxi worked first try |
| 20:45.02 | TheSeven | at least I can successfully get a directory listing of my boot partition |
| 20:52.29 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 21:17.30 | buZz | awesome :D |
| 21:21.36 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 21:23.40 | *** join/#cubieboard lerc (~quassel@121-74-230-95.telstraclear.net) |
| 21:39.33 | *** join/#cubieboard emilepetrone (~emilepetr@c-24-22-29-165.hsd1.or.comcast.net) |
| 21:41.55 | BigWhoop | thanks a lot guys. |
| 21:49.48 | *** join/#cubieboard GPSFan (~kenm@64.92.145.112) |
| 22:03.31 | rvalles | what's a bit silly and I'd like to fix is gpu reserves way too much ram. As I'm not really using accel, holding the framebuffer would be more than enough. |
| 22:04.35 | rvalles | It's probably some kernel parameter or some silly script.bin thing... but it's not on the fb article in the wiki... |
| 22:06.40 | Dave77 | http://cubieboard.org/buy/ <== Where is Hong Hong? |
| 22:09.09 | Turl | rvalles: disable *MALI* options on kernel config |
| 22:12.53 | rvalles | so by leaving just the fb it won't reserve as much ram anymore? :) |
| 22:13.05 | rvalles | I guess I'll try that when I build a kernel next |
| 22:13.18 | Turl | FB reserves 32M |
| 22:13.50 | Turl | mali 64M |
| 22:13.51 | rvalles | which is way more than necessary on 1920*1200*4 |
| 22:13.56 | rvalles | but better. |
| 22:14.21 | rvalles | ... wait, what? |
| 22:14.24 | rvalles | [root@alarm ~]# free -m |
| 22:14.26 | rvalles | <PROTECTED> |
| 22:14.28 | rvalles | Mem: 814 789 24 0 54 496 |
| 22:14.30 | rvalles | -/+ buffers/cache: 238 575 |
| 22:14.32 | rvalles | Swap: 2047 0 2047 |
| 22:14.54 | Turl | (1920 * 1200 * 32 * 2) * bits = 17.578125 megabytes |
| 22:14.54 | *** join/#cubieboard emilepetrone (~emilepetr@c-24-22-29-165.hsd1.or.comcast.net) |
| 22:15.02 | rvalles | here's what I'm seeing... that's probably some 192 + kernel + kernel's fixed-size caches |
| 22:15.15 | Turl | rvalles: run dmesg|less and check the top of it |
| 22:15.20 | Turl | mem reservations are printed there |
| 22:15.49 | rvalles | 1920*1200*4 (4*8=32) is 9216000 |
| 22:15.52 | *** join/#cubieboard emilepetrone (~emilepetr@c-24-22-29-165.hsd1.or.comcast.net) |
| 22:16.20 | rvalles | you reserve 2, but does the fb even do front&back buffers? |
| 22:16.53 | rvalles | [ 0.000000] Memory cut off: |
| 22:16.55 | rvalles | [ 0.000000] MALI : 0x5c000000 - 0x5fffffff ( 64 MB) |
| 22:17.00 | rvalles | [ 0.000000] Memory Reserved: |
| 22:17.03 | rvalles | [ 0.000000] SYS : 0x43000000 - 0x4300ffff ( 64 kB) |
| 22:17.05 | rvalles | [ 0.000000] VE : 0x44000000 - 0x48ffffff ( 80 MB) |
| 22:17.07 | rvalles | [ 0.000000] G2D : 0x58000000 - 0x58ffffff ( 16 MB) |
| 22:17.09 | rvalles | [ 0.000000] LCD : 0x5a000000 - 0x5bffffff ( 32 MB) |
| 22:17.11 | rvalles | that's way too much |
| 22:17.19 | rvalles | the hell is all that. |
| 22:17.21 | Turl | depends on what you use, but buffer swapping is common, android does it |
| 22:17.45 | Turl | VE is CedarX |
| 22:17.56 | Turl | G2D is 2D accelerator, unused for the most part |
| 22:18.00 | Turl | LCD is the framebuffer |
| 22:18.20 | rvalles | buffer swapping is common and useful, but as long as I'm using X with fbdev... no way it's benefiting from it. |
| 22:18.41 | rvalles | G2D's for the xf86-video-mali thing, then? |
| 22:18.53 | rvalles | I thought that was unaccelerated too |
| 22:18.58 | rvalles | so I guess it's not used at all |
| 22:19.43 | rvalles | well, next kernel I'm getting rid of everything but the fb and "hdmi" options. |
| 22:20.18 | rvalles | long term I'll probably play with the gpu... but that'd be with lima |
| 22:20.53 | Turl | G2D doesn't have anything to do with mali, I believe ssvb was going to add support for it on sunxifb X11 driver |
| 22:21.30 | Turl | but yeah, if you're using X, it's currently sitting idle |
| 22:22.43 | rvalles | text scrolling on fbcon is a slideshow |
| 22:22.48 | rvalles | so it probably isn't used there either |
| 22:23.38 | rvalles | (using xterm as terminal gives much better results... it is smarter on when to scroll and when to just draw a whole screen apparently) |
| 22:24.28 | rvalles | haven't heard about the lima & other arm gpu free driver projects talk in fosdem yet |
| 22:24.58 | rvalles | A basic drm/kms driver would be nice but that's dreaming on my end. |
| 22:26.20 | Turl | a kms driver would be unrelated with lima and such though, right? |
| 22:28.02 | rvalles | http://people.freedesktop.org/~libv/FOSDEM2013_arm_gpus_1-4_preliminary.pdf |
| 22:28.12 | rvalles | found this from #lima logs... |
| 22:28.24 | rvalles | apparently "actual driver work will start after fosdem" |
| 22:28.58 | rvalles | but they say in there that basically the reversing is pretty much done, they know how it works and they only need to get writing code. |
| 22:30.00 | Turl | reads pdf |
| 22:30.12 | rvalles | Turl: afaik kms implies a drm fb and some degree of knowledge by the kernel on how gpu memory is organised |
| 22:32.10 | *** join/#cubieboard RavenII (~RavenII@c-76-29-25-66.hsd1.il.comcast.net) |
| 22:33.56 | Turl | rvalles: but gpu is irrelevant to display modes |
| 22:34.46 | Turl | rvalles: it's not like on PC where you have a 'graphics card' that does both the output and 3D |
| 22:36.39 | rvalles | "gpu memory" was abuse of language |
| 23:29.52 | *** join/#cubieboard ssvb (~ssvb@212.16.98.80) |
| 23:37.06 | *** join/#cubieboard minkley (4068f926@gateway/web/freenode/ip.64.104.249.38) |
| 23:41.09 | *** join/#cubieboard marvinstuart (~marvinstu@cpe-184-57-132-151.cinci.res.rr.com) |
| 23:41.43 | *** join/#cubieboard pirea (~sabayonus@188.27.188.127) |
| 23:43.42 | TheSeven | arrrrrrrrrrrrrrgh |
| 23:43.56 | TheSeven | misaligned memory accesses in the nand driver! |
| 23:44.17 | pirea | who want a image for nandflash with archlinuxarm? :> |