**** BEGIN LOGGING AT Tue Mar 18 02:59:59 2014 Mar 18 07:13:18 1JTAALA8Q ds2 kevinh__ LorenzMeier panto tridge: i've uploaded my gsoc proposal, pls check it out and provide me feedback... Mar 18 07:13:33 ok, will do Mar 18 10:15:50 panto: ping Mar 18 12:20:48 sidbh_, pong Mar 18 14:07:34 sidbh: ping Mar 18 14:31:33 anujdeshpande: pong Mar 18 15:19:16 @jkridner I sent u the mail regarding the proposal for beaglepilot..will you please give a look to it and tell me the corrections or work to be done Mar 18 15:29:55 Umm... Has the ros integration been already done with the pilot code ? Mar 18 15:30:54 Because recently I have been tweaking around with the urg lidar that I have and would like to do some 3d mapping on d board itself Mar 18 16:43:27 kevinh__:ping Mar 18 16:46:00 kevinh__: we already have created beaglepilot mailing list for discussions: https://groups.google.com/forum/#!forum/beaglepilot Mar 18 16:47:03 kevinh__: it is open, but i'll send the membership request to all the mentors asap Mar 18 16:54:02 anujdeshpande: received my mail? Mar 18 16:56:28 sidbh: yepp. sounds like a good plan Mar 18 17:23:52 sidbh: please do not use that address for me or risk having things perm-spam listed... ping me for an address to use Mar 18 17:23:56 thanks. Mar 18 17:25:06 sidbh: sorry!, pls share ur email id Mar 18 17:26:17 sidbh: due to spam problems, each class of usage gets a new address. usage outside of the class will be tagged semiautomatically as spam Mar 18 17:29:36 * sidbh scratches his head... Mar 18 17:31:02 mailling lists get unique addresses. mail going to that address not coming from the mailling list is assumed to be spam Mar 18 17:31:32 same thing with company contacts... they sell the address, address gets blacklisted as spam infested Mar 18 17:32:53 ok! now i understand :) Mar 18 17:33:26 it also identifies dishonest businesses who tell people they do not sell addresses Mar 18 17:34:30 :D Mar 18 17:50:25 panto: ping Mar 18 17:50:33 yeah yeah I know Mar 18 17:50:38 it's on my to do list Mar 18 17:51:43 panto: no no, i wanted to discuss another topic, if you are free!! :) Mar 18 17:52:03 ok, go on but I don't have much time Mar 18 17:54:07 panto: i was studying remotproc, and was stuck at "how is the *rproc created for PRU that is to be passed to the functions everywhere?" Mar 18 17:54:36 the driver seems to pull off that trick Mar 18 17:56:21 i studied the pru_rproc.c, but was unable to found anything , can you point me to a part of code or something Mar 18 17:59:29 struct pruproc *pp = platform_get_drvdata(pdev); -> in struct pruproc struct rproc *rproc; Mar 18 18:01:06 panto: yeah i found that but it requires device structure/ platform_device structure Mar 18 18:01:23 and? Mar 18 18:01:33 how do you suppose you'll do it? Mar 18 18:01:43 and please don't tell me you plan to rewrite it Mar 18 18:02:15 i'm unaware how to do it :( Mar 18 18:03:48 you won't get far without having a device Mar 18 18:08:03 sidbh, ok, how about you start slow Mar 18 18:08:16 most of the code in there is generic Mar 18 18:08:18 split it up Mar 18 18:08:29 into one generic PRU part (loader+low level functions) Mar 18 18:08:47 and other files pertaining to a special function Mar 18 18:08:48 i.e. pwm Mar 18 18:08:53 or beaglepilot one Mar 18 18:09:03 you do not want to start from scratch Mar 18 18:09:07 you will run out of time Mar 18 18:10:12 panto: ok, seems like a nice plan! Mar 18 18:11:16 one thing about open source is use other people's stuff Mar 18 18:11:32 you starting from scratch, would require really extensive kernel knowleget Mar 18 18:11:37 *knowledge Mar 18 18:12:44 panto: ok, i'll keep that in my mind Mar 18 18:15:50 sidbh: panto do you think it's convenient to divide the librpru part in 2 ? I was thinking on working on PWM, PPM and PPM-SUM. Mar 18 18:16:20 ok, let's take it from the top Mar 18 18:16:40 the pru stuff can be split in one major part which is the generic loader + comms Mar 18 18:17:00 and an arbitrary number of modules dealing with specific peripheral instances Mar 18 18:17:07 so depending on the firmware you load Mar 18 18:17:20 you could have 1 PWM, 2 PPM, one PPM-SUM, one SPI, etc Mar 18 18:17:50 all of the different peripherals might need different 'modules' (no relation with kernel modules) Mar 18 18:18:07 panto: i see. so one person should tackle the loading+comms and one should work on the smaller modules ? Mar 18 18:18:17 ideally Mar 18 18:19:12 panto: but how will the smaller modules will be tested before loading+comms is prepared? Mar 18 18:19:29 panto: works well for me then. I can work on the smaller modules plus some other features like a dts , a ground station control and then the smaller modules. Mar 18 18:19:39 loading+comms are there Mar 18 18:19:48 you only need to expose them to user-space Mar 18 18:20:14 no-one expects you guys to build this stuff from scratch Mar 18 18:20:48 you need to figure things enough so that mold what's there for your task Mar 18 18:35:04 panto: so loading+comms will include making an auto tools type library which can be a standard install. using remoteproc inside of that for the actual loading. plus setting up the comm. all of this will compromise one project. sidbh is in a better position to do this i think. Mar 18 18:36:01 loading+comms will take place in the kernel Mar 18 18:36:16 you will extend what's there with a user-space interface Mar 18 18:36:28 and then you'll wrap it up the interface with a library Mar 18 18:37:04 panto: understood. so kernel -> interface -> library -> modules Mar 18 18:37:21 yes Mar 18 18:37:23 -> flying beagles Mar 18 18:37:28 hopefully :) Mar 18 18:37:34 panto: :D Mar 18 18:37:38 I need one as a toy for my daughter Mar 18 18:37:51 she should be 9 months old by the time you're done Mar 18 18:38:39 panto: vmayoral is doing a very neat project. I wonder if you've seen it. http://erlerobot.com that'll be running beagle pilot. so one of those for all present here :-p Mar 18 18:39:13 heh Mar 18 20:47:41 hi mikehhawley. Note, there are logs of the chats and some of the folks here should be able to catch you up on existing project proposal thoughts. Mar 18 20:47:45 * jkridner goes afk Mar 18 21:37:08 sidbh: ping **** ENDING LOGGING AT Wed Mar 19 02:59:58 2014