**** BEGIN LOGGING AT Mon Jun 27 02:59:58 2016 Jun 27 03:23:55 hi all Jun 27 03:24:11 any one worked with device drriver in BBB Jun 27 06:40:11 zmatt: thanks, that looks precisely like what i've been looking for. if everything works out as hoped, i might be able to contribute to the workflow (networkmanager should, after a recent patch, be suitable for serving all that is needed) Jun 27 07:15:31 ew, why would you want to use networkmanager Jun 27 07:17:28 zmatt: because it's already running on my system. works the same using plain dnsmasq, though. (i prefer to avoid recommending people to `git clone / make / execute` code when things can be done with what is shipped in distributions) Jun 27 07:18:22 hmm? oh you mean to use bbblfs? Jun 27 07:19:45 yeah I have a variant working now that only uses dnsmasq Jun 27 07:20:16 including a u-boot that's a bit better behaved than the one bbblfs uses Jun 27 07:20:44 still need to nice up the kernel and initramfs Jun 27 07:20:48 bbblfs' usb_flasher segfaults for me, unfortunately Jun 27 07:21:19 weird, but you can serve its spl, u-boot, and fit file using dnsmasq Jun 27 07:21:48 right now i'm using the following dnsmasq configuration, and both files are loaded, but i don't get a block device: Jun 27 07:21:52 http://paste.debian.net/762619/ Jun 27 07:22:09 eh Jun 27 07:22:21 no, SPL wants to load u-boot Jun 27 07:22:25 u-boot loads fit Jun 27 07:22:28 (that's the config i concluded from skimming through the main.c source code Jun 27 07:22:38 ah, i've missed a step -- that explains it, then Jun 27 07:22:42 moreover, u-boot annoyingly contains a hardcoded ip to contact as tftp server Jun 27 07:23:17 so spl is kind of a u-boot preloader? Jun 27 07:23:19 hold on I still have notes on how to get the bbblfs files working without building a new u-boot Jun 27 07:23:38 yes, spl's main job is initializing ram Jun 27 07:23:51 because they couldn't manage to fit a bootloader in 109 KB -.- Jun 27 07:24:14 (available internal SRAM you need to fit in to be loaded by the ROM bootloader) Jun 27 07:24:40 mhm, that seems to be a common pattern in bootloaders Jun 27 07:25:55 http://pastebin.com/isBDJKaa Jun 27 07:27:14 and you need to have ip 192.168.1.9/24 since that's what bbblfs' u-boot is hardcoded to contact to load the fit (which actually should actually be named /itb ) Jun 27 07:27:42 yeah, i've just configured that. funny thing is that uboot seems to reboot when nmap'd :-) Jun 27 07:27:52 (i was wondering whether it would offer anything like a network shell) Jun 27 07:27:58 lol no Jun 27 07:28:10 actually it has no networking except when explicitly needing it Jun 27 07:28:19 it has no multitasking or interrupts afaik Jun 27 07:28:31 so while it's doing network it can't do anything else Jun 27 07:28:49 reasonable enough for a bootloader -- i was just poking Jun 27 07:29:20 yeah, except in 109 KB you could fit a multitasking OS and complete networking stack just fine :P Jun 27 07:30:00 * chrysn nods (having played around with riot-os and bootloaders) Jun 27 07:30:58 what I don't understand is why it hardcoded an ip rather than using bootp/dhcp just like the first two stages Jun 27 07:31:19 then you can just use dhcp && bootm as bootcmd in u-boot Jun 27 07:32:28 that's better asked to the uboot ppl; but generally, if i get this working here for a first step, i'd be looking into more official files for spl, uboot and fit/itp, maybe there are better images that do things right Jun 27 07:32:57 no u-boot people would ask the same question, it's the bbblfs guy who hardcoded those parameters Jun 27 07:34:36 shouldn't then everything but the actual bbblfs image (fit/itb) be swappable? Jun 27 07:36:13 like I said I built a better-behaved u-boot, although I took a hacksaw-and-flamethrower approach to u-boot configuration since I wasn't in the mood to sort that mess out properly :P Jun 27 07:38:47 the official Kconfig mechanism is nearly useless for the am335x target anyway since it still has a metric ton of config being done in giant messy includes Jun 27 10:45:51 . Jun 27 11:44:19 Hello People Jun 27 14:21:01 zmatt: Hi, I'm using your tool py-uio to enable pruss and one output for now ;) Jun 27 14:21:59 Still some question can I use "exclusive" parameter in the dtsi file like dts ? Jun 27 14:41:16 zmatt: dtsi file with question http://pastebin.com/CTbQvdKs if you can take a look. Thanks Jun 27 17:12:25 HI Jun 27 17:12:38 i am new to beagle bone black Jun 27 17:15:39 hello Jun 27 17:17:15 kms: hi. do you have a question? if so, you should just ask it (to the channel, not to me) ;) hope you enjoy your BBB, I sure like mine. Jun 27 17:18:53 yes Jun 27 17:19:28 actuallyy i am planning to start learning device drivers concepts from basics like writting from scratch Jun 27 17:20:20 is it possible? Jun 27 18:34:52 dirkk: http://pastebin.com/GFCwSATX Jun 27 18:35:31 dirkk: I have no idea why people use such silly redundant node names for child nodes of &am33xx_pinmux Jun 27 18:35:48 especially since they only have to be locally unique (rather than globally unique like labels) Jun 27 18:36:51 metadata is not supported by my perl script currently... I could already add support for it I suppose but I'm not terribly motivated since I know the kernel ignores it anyway Jun 27 18:37:15 (at least the configfs mechanism does) Jun 28 02:04:41 Hello... Jun 28 02:04:56 Has the Bash Scripting changed since 2005? Jun 28 02:06:47 See here...I am typing up shell scripts on my BBB. Jun 28 02:07:20 My book from Blum is not collective. I am sure of it. What do you think? Jun 28 02:14:18 ... Jun 28 02:14:32 Sorry...I will get back to it. Jun 28 02:23:31 What is the difference between the $ and the # at the prompt? Jun 28 02:24:21 BBB uses # and Ubuntu uses $. I thought Ubuntu was a form of Debian...heh? Jun 28 02:29:09 Heh...forget it. I remembered. Jun 28 02:32:54 I am going to jump out on a limb and say people were ruining their BBBs with the available home $ Desktop directory shown. Jun 28 02:33:26 Okay... Jun 28 02:34:29 If I am a superUser, do I need to go into my bin directory in the BBB while bash scripting, i.e. #!/bin/bash? **** ENDING LOGGING AT Tue Jun 28 02:59:58 2016