**** BEGIN LOGGING AT Mon Mar 16 02:59:58 2015 Mar 16 03:04:25 alexanderhiam_: from what I see in the library digitalRead returns a PyObject which needs to be freed at some point in time...that never happens from what I understand Mar 16 03:26:40 * nerdboy adds himself to an empty mentor thingie Mar 16 04:11:11 vvu: when you create PyObjects with Py_BuildValue (and pretty much any PyObject afaik) they'll get freed by the garbage collector once they no longer have any references Mar 16 04:13:56 vvu: can you post your code? Mar 16 04:26:51 mainly i have a while 1 loop where i take decisions based on digitalRead Mar 16 04:27:23 i checked out the tag before going to the new API and ram stays constant Mar 16 04:29:34 i need to catch some sleep, catch you later Mar 16 04:32:08 ok, I'll look into it, gn Mar 16 06:53:05 @ds2 is the automated testing aimed at testing the IO through external hardware or through software only? Mar 16 07:53:19 * karki is back :D Mar 16 07:57:15 karki: hey. I added gpio module support.it was discussed as a feature addition to pruspeak on irc a few days back. Sent you a merge request. Can you please evaluate it. Thanks :) Mar 16 07:57:43 DIO 8 to 16 works now Mar 16 08:01:26 Thanks for the pull request :) Mar 16 08:01:31 Really good work :D Mar 16 08:01:41 but, there are somethings you have to work on Mar 16 08:01:47 I'll talk about it later Mar 16 08:01:58 but *brilliant* first effort ! Mar 16 08:02:04 karki: okay :) Mar 16 08:02:34 karki: thanks. Can be easily expanded to other modules. Currently only gpio 0 Mar 16 08:03:42 few corrections here and there . (I'll talk about it later, have some tiny bit of work now) Mar 16 08:03:49 got back from singapore today Mar 16 08:04:02 was talking about BBB at a conference Mar 16 08:04:11 karki: okay. Ttyl. Have a few doubts. Mar 16 08:04:28 Awesome Mar 16 08:22:15 sidbh_: ping:) Mar 16 08:22:35 kiran4399, pong Mar 16 08:24:01 sidbh_: so when you asked me to remove PRU_SYSCGF stuff.. did you mean to remove both standby mode and enable mode lines and keep only enable before timer_init() or something else ? Mar 16 08:24:39 yes the first one Mar 16 08:29:48 sidbh_: Hmm.. did that .. but the beaglebone is ending up in a sort of deadlock.. Mar 16 08:31:15 kiran4399, ok, is everything alright if you don't do external access, just the older functionalities but after this fix? Mar 16 08:33:04 sidbh_: you mean to say if access by r30 ? Mar 16 08:33:11 yes Mar 16 08:34:36 anyway, you need some way to debug the code now... Mar 16 08:34:44 karki, you there? Mar 16 08:35:31 yeah Mar 16 08:35:33 kindof Mar 16 08:35:42 * karki is multi tasking :p Mar 16 08:37:04 gr8, can you help kiran4399 with some tools and ways to debug the code, I utilised the shared memory but If I'm not wrong there is even better way available with pruspeak using upcalls Mar 16 08:37:26 which code? Mar 16 08:37:32 I guess panto implemented printf Mar 16 08:37:39 yes he did Mar 16 08:37:44 the pruspeak code Mar 16 08:37:54 oh , ok :) Mar 16 08:38:06 kiran what do yu have a problem with Mar 16 08:38:09 the timer ;) Mar 16 08:38:42 * karki feels that the pru timer is the most evil thing created :p Mar 16 08:39:46 I guess he is actually having problems with accessing the global registers, e.g. GPIO on AM335x Mar 16 08:39:58 karki: hi... I don't know what my problem is called .. :-P but it is like when I mess up with the pru_syscfg enabling's and disabling's my beaglebone is kind of freezing.. Mar 16 08:40:30 karki: I am actually trying to access other gpio's using starterware.. Mar 16 08:40:45 karki: the library is working fine.. Mar 16 08:41:14 karki: do you have any debugging stuff with you ? Mar 16 08:41:18 karki, I guess most of his problem would be resolved if he had a nice way to debug it :) Mar 16 08:41:47 sidbh_: I am not able to access even r30 pins.. :-( Mar 16 08:42:06 hmmm... Mar 16 08:42:27 wait, are you using gpio as is Mar 16 08:42:33 on linux Mar 16 08:42:39 I mean starterware Mar 16 08:42:41 as is Mar 16 08:42:43 on linux Mar 16 08:43:17 sidbh_ : the best method panto had was writing into shared memory ;) Mar 16 08:43:28 I really don't have anything better to offer Mar 16 08:43:41 well.. I compiled it for pru.. and got a static library ... installed it.. included those functions in pru0_firmware.c... modified the make file.. Mar 16 08:44:28 you just don't combine a non linux thing with a linux thing Mar 16 08:44:40 I'll look into what you are doing a bit later Mar 16 08:44:41 karki, same as what I did :) Mar 16 08:44:50 kinda busy :/ Mar 16 08:45:03 * karki has backlog tasks for the week :'( Mar 16 08:45:14 kiran4399, let me tell you the way I debugged my pru codes :) Mar 16 08:45:27 what I am using in starterware are the api's for accessing the registers.. Mar 16 08:45:41 sidbh_: sure .. go ahead :-D Mar 16 08:47:13 kiran4399, check this line https://github.com/bugobliterator/ardupilot/blob/master/Tools/Linux_HAL_Essentials/pwmpru/prucomm.h#L73 Mar 16 08:47:53 you see I'm using this structure to communicate with the linux userspace with settings Mar 16 08:48:03 ok.. I see it is the start pointer for shared dram Mar 16 08:48:49 now I can have a similar structure for say a debug structure which contains all the info I need to pass on Mar 16 08:48:58 ok.. Mar 16 08:49:19 Now what I de is create a very simple app e.g. https://github.com/bugobliterator/ardupilot/blob/master/Tools/Linux_HAL_Essentials/test_codes/rcin_test.c Mar 16 08:50:00 ok.. Mar 16 08:50:05 where I just mention the shared DRAMs location from the linux userspace point of view, and just declare the structure here too Mar 16 08:50:46 in the rcin_test.c ... Mar 16 08:51:43 now I compile this code and run continuously, It will print away all my data, use ring_buffer if you want it robust or, just raising magic numbers to access will do too. Mar 16 08:52:37 so rcin_test runs on my linux polling for any debug data Mar 16 08:53:10 having similar debug setup in place will save you a lot of hassle Mar 16 08:53:28 so you want me to do this when my beaglebone hangs ? Mar 16 08:53:48 no just run the software the whole time Mar 16 08:54:20 you do the testing, it will run in say different ssh term/window/instance Mar 16 08:54:40 Oh... ok got it.. Mar 16 08:55:01 in one terminal I will do this pruspeak.. and another debugging right ? Mar 16 08:55:12 exactly Mar 16 08:56:32 I'm not sure how pruspeak uses shared memory, so make sure you don't clash debug and application memories Mar 16 08:56:54 keep them separated, debug stuff is there to help not to break Mar 16 08:57:32 sidbh_: OK.. Mar 16 09:04:00 sidbh_: err... can you tell me what structure I need to define in rcin_test.c.... the one you gave me... line 73 is a macro.. Mar 16 09:05:08 https://github.com/bugobliterator/ardupilot/blob/master/Tools/Linux_HAL_Essentials/rcinpru/prucomm.h#L15 Mar 16 09:05:14 they both should be same Mar 16 09:05:33 https://github.com/bugobliterator/ardupilot/blob/master/Tools/Linux_HAL_Essentials/test_codes/rcin_test.c#L17 Mar 16 09:05:58 the 1st one is from pro firmware, second one from userspace app Mar 16 09:06:40 s/pro/pru/ Mar 16 09:24:07 alexanderhiam_: if u need some debug info just ping Mar 16 09:30:28 sidbh_: can I check PRUCFG_SYSCFG register with the debug code ? Mar 16 09:31:13 yes, just pass it on to the structure and your userspace debugger will read it:) Mar 16 09:38:55 sidbh_: should I keep those ring_head and ring_tail ?? Mar 16 09:41:18 as i said its your choice how you implement, if you want to do a ring buffer , you keep them, if you don't use a three state flag or something to specify if PRU is using it or userspace or its idle and access accordingly Mar 16 09:41:56 Oh.. OK.. Mar 16 09:43:23 with ring buffer you would have very low chance of losing messages/changes, and its implementation is easy enough Mar 16 09:44:04 Its mostly done you just need to copy wisely :) Mar 16 10:13:20 sidbh_ : are you mentoring this time? :) Mar 16 10:19:22 karki, nope, I won't be much available when GSoC actually kicks off Mar 16 10:33:20 sidbh_: I am adding dpram starting pointer to this.. is it ok ? #define DPRAM_SHARED 0x00012000 Mar 16 10:35:04 its fine as long as its away from other utilised shared memory regions make sure of that Mar 16 10:36:45 no.. wait.. looks like I have another #define DPRAM_SHARED 0x00010000 Mar 16 10:37:56 sidbh_: shall I do it before 10000 ? Mar 16 10:39:39 no, read the reference!! manual 0x10000 - 0x20000 is shared ram 0x0000 - 0x10000 is PRU RAM Mar 16 10:40:27 sidbh_ : Oh.. my bad !!! so how do I know how much sharedram pruspeak is using ? Mar 16 10:40:56 check how much space that pointer with that address is attached with and then use a lot far from that but within 0x20000 Mar 16 10:41:17 e.g. for starters just use 0x16000 Mar 16 10:42:40 how much does ringbuffer take ? Mar 16 10:43:08 I mean .. how much of the dpram does it take ? (maximum) Mar 16 10:54:14 kiran4399, sorry I made a mistake with the range it will be 0x10000 - 0x13000 for shared ram (its 12 KB in size) Mar 16 10:55:54 kiran4399, to calculate the size of space acquired by any structure is fairly basic C question, answer to which you will find in detail easily using google :) Mar 16 10:57:43 keep referring pru reference and am335x user manual as frequently as possible, you will find something new and enlightening every time Mar 16 11:28:23 sidbh_: I have to tell you some of my observations... Mar 16 11:29:22 sidbh_: I though of first testing the code by removing some lines from the dio_handler... and seeing if it is hanging or not .. Mar 16 11:29:53 kiran4399, yes tell me Mar 16 11:30:40 sidbh_: So when I removed the init() function which consisted of gpio module enable, reset and pin_dir_set.... the beaglebone is not hanging anymore... but at the same time the desired pin is not being set.. Mar 16 11:31:02 sidbh_: so what do you think ?? Mar 16 11:31:28 I think the hanging problem came because of the init function... Mar 16 11:31:53 hmmm... interesting, does waiting after init() brings up same issue Mar 16 11:32:14 sidbh_: If we cannot come to a conclusion now.. I will go by the ring_buffer test.. Mar 16 11:32:59 you should setup the debug code anyway, you can't go blind overtime you hit an issue, you need to know what's going on Mar 16 11:33:14 sidbh_: OK.. Mar 16 11:33:18 actually that should be your first goal Mar 16 11:35:28 can you tell me some data you would want me to put in the structure ? Mar 16 11:35:37 sidbh_: ^^ Mar 16 11:36:20 just have a string and a array of 10 int32s should suffice in your case Mar 16 11:39:55 print this data continuously in userspace Mar 16 11:41:17 sidbh_: umm... in the add_to_ring_buffer function in the main source code.... with what should I add to the ring_buffer ? Mar 16 11:42:14 ohhh... you want to go ring buffer way Mar 16 11:43:41 https://github.com/bugobliterator/ardupilot/blob/master/Tools/Linux_HAL_Essentials/rcinpru/rcinpru0.c#L18 Mar 16 11:44:25 define this in your firmware code and pass char array and int array to it Mar 16 11:45:39 hello,what is the 3D graphic library used for the beagleboard UI embedded in the card Mar 16 11:45:56 keep the size of ring buffer so that (DRAM+sizeof(debug_struct)*ringbuffer_size) doesn't go beyond 0x13000 Mar 16 11:46:26 s/DRAM/SHARED_RAM/ Mar 16 11:47:31 sidbh__: ya sure... should I declare the same pin_read and delta_time in the buffer of the ring structure ? Mar 16 12:05:55 sidbh_: ya sure... should I declare the same pin_read and delta_time in the buffer of the ring structure ? Mar 16 12:08:45 kiran4399, I think you are aware that is not a debug structure, its the one I'm using to pass the settings to pwm firmware, I just gave you that as an example/reference code. most of it you can use as is but you need to make some heavy changes. for example pin_read and delta_time doesn't matter in your case, because they application specific Mar 16 12:10:51 sidbh_: Yes.. gotcha !! :-D Mar 16 12:11:37 kiran4399, are you sure ? don't hesitate, ask away if you still have any doubt Mar 16 12:14:28 sidbh_: no..no... I got it.. I was only under this confusion ... Mar 16 12:14:46 ok cool! Mar 16 12:32:25 kiran4399: hello,please can you help clearify this phrase "" Mar 16 12:34:05 kiran4399: " find the best way to maintained all information on gist.github.com Taking in consideration big images can't be uploaded right now." Mar 16 12:34:44 found on the " BoneScript web pages with live-running examples and documentation" project Mar 16 13:11:16 sidbh_: one more thing.. where will the debug information be printed ? Mar 16 13:58:24 kiran4399, check pseudo code I made for userspace to be compiled and run in userspace: https://gist.github.com/bugobliterator/c043686347d5feafc961 Mar 16 13:59:00 pseudo code for firmware side, to be merged with your firmware code: https://gist.github.com/bugobliterator/c043686347d5feafc961 Mar 16 13:59:30 hope this resolves your doubt, let me know if you face any problem Mar 16 14:12:42 your organisation is doing very good work and every year to present a really challenging problem ,and i am very interested in your organisation so will you tell me this year what is going to be your project Mar 16 14:14:05 sidbh_: Thanks a lot :-D But I think those 2 links are same ?? Mar 16 14:16:47 which two links?? Mar 16 14:17:27 14:58 < sidbh_> kiran4399, check pseudo code I made for userspace to be compiled and run in userspace: https://gist.github.com/bugobliterator/c043686347d5feafc961 Mar 16 14:17:30 14:59 < sidbh_> pseudo code for firmware side, to be merged with your firmware code: https://gist.github.com/bugobliterator/c043686347d5feafc961 Mar 16 14:22:14 sidbh__: Thanks a lot :-D But I think those 2 links are same ?? Mar 16 14:23:08 are they! sorry, there you go : https://gist.github.com/bugobliterator/4392feb74e88d3416472 Mar 16 14:23:50 this ones the pseudo userspace code Mar 16 14:24:03 hope this helps :) Mar 16 14:59:07 azizulhakim: How is the proposal coming? Mar 16 15:48:33 sidbh_: sid, why is some garbage string printing ?? Mar 16 15:49:15 point me to your code Mar 16 15:53:59 sidbh_: https://gist.github.com/kiran4399/f0a93d9f7136ba8c8087 Mar 16 15:55:59 and userspace side too Mar 16 15:57:26 kiran4399, you need to initialise ring_head and ring_tail at startup in firmware Mar 16 15:57:52 to 0 Mar 16 15:58:54 sidbh_: Oh... I got it.. Mar 16 15:59:53 also initialise the strings and integer array members of debug struct to "initial" and 0 Mar 16 16:06:30 how can i start cross compiling on mac os? Mar 16 16:06:58 Themacho: Easy way is install a Linux VM Mar 16 16:07:33 Then refer https://github.com/jadonk/gsoc-application/tree/master/ExampleEntryJasonKridner Mar 16 16:07:42 or just intall Linux Mar 16 16:07:54 That's better. Mar 16 16:08:17 rd_: 1) no PMs Mar 16 16:08:19 i just need to compile helloworld . c and make a pull request? Mar 16 16:08:24 2) no shouting Mar 16 16:09:42 ? Mar 16 16:10:01 3) sign up for the mailing list and propose your project there Mar 16 16:10:01 jkridner: Hello jkridner. Mar 16 16:10:07 howdy! Mar 16 16:10:12 applications start today!! Mar 16 16:10:47 jkridner: nice to hear from you !! :-D Mar 16 16:12:14 what happened to rahuldewangan? Mar 16 16:13:26 * jkridner has no idea Mar 16 16:13:32 should i know how to write device drivers and kernel patching? Mar 16 16:15:11 Themacho: Depends on the project you are applying for. Mar 16 16:15:42 what are the basic skills required for BB projects? Mar 16 16:16:36 Again, project specific. There no such thing as you MUST know this skill to apply. Mar 16 16:16:50 jkridner: i also stuck my name on RT kernel project Mar 16 16:16:58 But you should be able to learn the required skills as per demands. Mar 16 16:17:04 nerdboy: thanks!! Mar 16 16:17:05 since it was empty and looked lonely... Mar 16 16:17:12 :-D Mar 16 16:17:24 Hi jkridner Mar 16 16:19:08 maybe i will finally get the answer to the burning question of rt_group_sched vs rt_user_sched patch... Mar 16 16:21:40 howdy Abhishek_ Mar 16 16:21:50 okay, so are the top few 2014 ml threads the right place to look? Mar 16 16:22:25 * nerdboy noticing there is no gsoc repo as pointed at last year Mar 16 16:22:47 I'm interested in this project "Common bootloader for different all the BeagleBone/BeagleBoards" ,any help in where to start? Mar 16 16:22:52 do i smell a mentor meeting of some kind? Mar 16 16:24:03 googly hangout maybe? stepper motor driven morse code session? Mar 16 16:24:57 naval flags rendered on led panels? Mar 16 16:25:31 * nerdboy running out of stupid ideas on only one espresso... Mar 16 16:25:56 I'm interested in this project "Common bootloader for different all the BeagleBone/BeagleBoards" ,any help from where to start? Mar 16 16:26:58 did you post on the ML? Mar 16 16:27:56 yes Mar 16 16:28:14 i don't know the mentor on that one... Mar 16 16:28:34 Hunyue Yau Mar 16 16:29:14 ^= ds2 Mar 16 16:29:22 i can read with my old-guy glasses... Mar 16 16:30:13 * nerdboy remembers ds2 complaining about that one... hey, wait a minute... Mar 16 16:30:53 * jkridner wonders how av500 is feeling now. Mar 16 16:31:04 me too Mar 16 16:32:03 anyone know Class Ziemke? name sounds familiar to me. Mar 16 16:32:27 is it a person, place, or thing? Mar 16 16:32:38 jkridner: How's BB-X15 coming along? Mar 16 16:33:26 jkridner: Hello! Mar 16 16:33:49 jkridner: Hey could you please check my reply to our conversation about the nw.js app http://bit.ly/1Evqpf9 Mar 16 16:34:03 webglider: pulling it up. Mar 16 16:34:27 Abhishek_: seems several small revisions are needed. Should be available in May for additional pre-release testers. Mar 16 16:34:30 * nerdboy thinks maybe _av500_ needs more fluids and what-not... Mar 16 16:34:40 nerdboy: +1 Mar 16 16:35:47 jkridner: looking forward to it Mar 16 16:37:14 how can i contact mentor Hunyue Yau ? Mar 16 16:37:24 * nerdboy plays hammer-of-god on the invading ant army Mar 16 16:37:26 ds2: ^ Mar 16 16:39:47 how can i contact mentor Hunyue Yau ? Mar 16 16:40:10 Themacho, ds2 ^ Mar 16 16:41:18 sorry but i dont understand :D Mar 16 16:41:21 * nerdboy hands out 1st homework assignment for anyone who has not read Lucifer's Hammer... Mar 16 16:41:42 Themacho, ds2=Hunyue Yau Mar 16 16:42:03 thanks Mar 16 16:43:46 oh yeah, almost forgot Mar 16 16:43:59 * nerdboy starts the philly count-down clock Mar 16 16:44:39 a week and an hour. ;) Mar 16 16:45:00 jkridner, you making it elc next week? Mar 16 16:46:48 2 hrs to go to applications open Mar 16 16:47:39 rahuldewangan: ask out loud please Mar 16 16:47:49 * nerdboy looks for new guy hat Mar 16 16:49:38 rahuldewangan: either here or the mailing list... Mar 16 16:49:56 anyone here ever compiled qt5 statically for Linux ? Mar 16 16:50:54 rahuldewangan: if you write it up in a little more detail you can submit Mar 16 16:51:35 with a diagram even Mar 16 16:51:52 * nerdboy recommends operational concept diagram Mar 16 16:52:43 take off the /msg part Mar 16 16:58:00 jkridner: I got my proposal reviewed by alexanderhiam. Mar 16 16:58:29 jkridner: he suggested me some changes.. Mar 16 17:00:40 great! Mar 16 17:01:16 <_av500_> jkridner: HI Mar 16 17:01:24 Howdy _av500_! Mar 16 17:01:26 <_av500_> jkridner: I will be there for wednesday meeting Mar 16 17:01:32 * jkridner is headed to lunch. Mar 16 17:01:36 <_av500_> also, we should think about a mentors meeting Mar 16 17:01:40 _av500_: great! I expect it to be busy. Mar 16 17:01:44 <_av500_> right Mar 16 17:01:51 <_av500_> but that can be next week Mar 16 17:01:55 _av500_: do we want to do the mentors meeting 30 minutes earlier/later? Mar 16 17:02:01 _av500_: ok. Mar 16 17:02:21 i will be gone this week... Mar 16 17:02:22 <_av500_> jkridner: it looks to me like we have good engagement with mentors Mar 16 17:02:27 <_av500_> vvu: no prob Mar 16 17:02:40 <_av500_> jkridner: so maybe its ok to start next week to sort stuff out Mar 16 17:03:12 jkridner: so.. one of which is he told me it would be great if we use c9 ide to link up with pruduino.. Mar 16 17:03:19 proposals are due in 10 days from now, right ? Mar 16 17:03:52 <_av500_> vvu: 27th Mar 16 17:05:16 i think we will have a couple of good ones :) Mar 16 17:05:51 jkridner: what do you think ?? Mar 16 17:06:40 jkridner: One thing I saw in noduino is that it provides a nice startup for users... like it is very user-friendly.. Mar 16 17:07:52 rahuldewangan: since the application period opens today... Mar 16 17:08:21 write it up, and don't forget the OV-1 diagram Mar 16 17:09:34 most of the "project" ideas seem like developer itches so an interesting application with some good hacks seems fine to me Mar 16 17:10:41 rahuldewangan: are you familiar with https://www.understandrisk.org/ ? Mar 16 17:10:47 they love stuff like that... Mar 16 17:11:14 or crisiscommons maybe Mar 16 17:19:40 jkridner: I'd like to discuss about the nw.js project Mar 16 17:20:01 jkridner: Over the weekend I built this prototype: http://bit.ly/1bc8DBx Mar 16 17:20:21 jkridner: It's a nw.js app which can write images to sd cards on linux systems Mar 16 17:20:44 jkridner: It also keeps track of the progress during the process Mar 16 17:21:46 rahuldewangan: i was just pointing them out Mar 16 17:21:59 as in yes, it's a valid application Mar 16 17:22:44 and that it's not an "advance the development infrastructure" thing Mar 16 17:23:40 OV-1 is usually analogous to ops concept diagram Mar 16 17:24:17 just one that captures your operational env, users, core architecture, etc Mar 16 17:25:13 jkridner: After doing quite some search I figured out that the best way to write images to micro SD would be to first transfer the downloaded img to the BB and then over SSH the app can tell the board to write the img (since its linux) Mar 16 17:25:57 http://tinyurl.com/k2cdvv9 <= googly examples Mar 16 17:26:13 jkridner: This will keep it uniform across all platforms. For talking to the board via SSH we could use https://github.com/mscdex/ssh2 Mar 16 17:26:17 mostly DoD and NASA plus other govt stuff Mar 16 17:26:49 http://en.wikipedia.org/wiki/Operational_View Mar 16 17:27:17 jkridner: that's not too much homework, is it? Mar 16 17:28:07 rahuldewangan: you might want to include some rationale from UR or crisiscommons stuff Mar 16 17:31:06 rahuldewangan: doesn't need to be as fancy as some of those examples Mar 16 17:31:19 vvu: how long does it run before you see the memory grow? Are you just watching top? Mar 16 17:31:45 it imediately starts gradually Mar 16 17:32:19 i tested my code and removed parts from it, when I removed digitalRead call mem stayed at ~40mb constant Mar 16 17:32:52 sidbh_: can you tell me .. while debugging what I should get ?? Mar 16 17:33:17 what are you getting? Mar 16 17:34:49 rahuldewangan: *if* you're already familiar with eclipse you can try the polarsys/capella thing for creating your system architecture Mar 16 17:35:22 but i wouldn't spend much time on that if it doesn't make sense Mar 16 17:35:41 * nerdboy usually wings the diagrams when time is very short Mar 16 17:37:43 what you are asking me is to make a detailed flowchart of how system will work if i am right Mar 16 17:38:56 karki : i'll indent the code properly :) Other things might be messy too. It was all done in a couple of days. What improvements and further work would you suggest ? Mar 16 17:38:58 sidbh_: I am still getting some garbage string.. Mar 16 17:39:23 sidbh_: I initializing in both the codes.. Mar 16 17:39:28 shubhangi : the gpio control is ok. I'm fine with the code Mar 16 17:39:40 s/initializing/initialized Mar 16 17:39:41 what I'm not fine with is the pull req made Mar 16 17:39:49 you have not cleaned up as such Mar 16 17:39:58 you have added auto generated files etc Mar 16 17:40:05 oh Mar 16 17:40:13 will clean it up asap Mar 16 17:40:27 I'm pretty sure you don't need to touch 35 files and write 2.5K lines of code for extending GPIO Mar 16 17:40:40 rahuldewangan: look at the ops concept/ov-1 examples and wikipedia Mar 16 17:40:42 more like touch 3-5 files and 100 loc (max) Mar 16 17:41:04 shubhangi : I'd suggest to forget it for now Mar 16 17:41:12 focus on the proposal this week Mar 16 17:41:18 it can have some of the details but only where needed Mar 16 17:41:22 issuing the pull request is the main thing Mar 16 17:41:47 not getting it accepted , atleast thats not the priority now Mar 16 17:41:59 the technical stuff is (mostly) in the system views Mar 16 17:42:07 vvu: ok, I'm seeing it now too... though I really don't get it... Mar 16 17:42:14 okay. will complete proposal and submit. Shall i look into the decimal delay issue after it Mar 16 17:42:20 apaar , hows your proposal coming along Mar 16 17:42:28 cool Mar 16 17:42:36 or have you already worked on it ? Mar 16 17:42:53 focus on driver stuff, very important. pruspeak/pruduino is very much secondary Mar 16 17:43:00 I haven't worked on decimal Mar 16 17:43:08 not a spec, just the ov-1(ish) part Mar 16 17:43:15 driver stuff => your proposal Mar 16 17:43:17 karki : yes, i wrote an LEDtoggle demo driver Mar 16 17:43:26 I saw that Mar 16 17:43:39 heavily borrowed from beaglelogic ? Mar 16 17:43:42 helps you and the reviewers get a feel for the intended environment, users, data flow, etc Mar 16 17:43:46 karki: works fine. uses beaglelogic bindings Mar 16 17:43:50 yes, heavily Mar 16 17:43:54 hey karki :)..the proposal is coming along great should have it ready be tom Mar 16 17:43:55 I saw ;) Mar 16 17:44:04 apaar : cool Mar 16 17:44:10 i know the code inside out though Mar 16 17:44:23 alexanderhiam_: same here, looked through the code...don't really get from where that all RAM is getting consumed Mar 16 17:44:23 shubhangi : Now that you know the code - work on how to write it Mar 16 17:44:27 okay Mar 16 17:44:33 (y) Mar 16 17:44:37 beaglelogic style is not cool Mar 16 17:44:46 it's as it says = "glue" Mar 16 17:44:57 we need cement :D Mar 16 17:45:01 yes :) Mar 16 17:45:11 * karki logs out Mar 16 17:45:21 vvu: I just tried manually decrementing the references to the PyObject before returning it from digitalRead and it segfaults, so there's definitely only the one reference Mar 16 17:45:56 might be room for another one with software stack details, but the overview is probably best if's just text and the one graphic Mar 16 17:46:46 alexanderhiam_: same here, tried that too :) Mar 16 17:47:07 but if not from pybbio means something is from python internals :) Mar 16 17:47:13 since it makes you think about it, and also easier to explain... win-win Mar 16 17:47:40 vvu: oh, I wonder if it's in the state file descriptor memoizing, perhaps it's actually continually opening the files Mar 16 17:48:11 * nerdboy probably sounds like the instructor-dad... Mar 16 17:48:14 i gave a look but it is not...think so Mar 16 17:48:21 you put the lock when you first open the file Mar 16 17:48:25 and then you just check if it is locked Mar 16 17:48:43 i was reading the documentation to integrate ros with bonescript Mar 16 17:48:59 vvu: oh I think I know what it is! I'm pulling items out of the GPIO config dict and not decrementing the references when I'm done with them Mar 16 17:49:09 if PyGPIO_getStateFile Mar 16 17:49:11 in* Mar 16 17:49:30 hmm, maybe that's why they don't listen to me at home either... Mar 16 17:50:01 vvu: yeah, that's 4 PyObjects per digitalRead call! Mar 16 17:50:29 well i was digging the documentation http://wiki.ros.org/roslibjs to integrate bonescript with ROS Mar 16 17:50:53 but i have no idea where I should edit the library Mar 16 17:53:30 alexanderhiam_: :) well i can take a look and submit a PR if u are busy with it Mar 16 17:53:35 kick it in the junk Mar 16 17:54:23 vvu: doing it now. I'm wondering if the other C extensions could have siilar leaks... Mar 16 17:55:32 wait... what python interface is this? Mar 16 17:56:16 the normal BBB.GPIO thing? Mar 16 17:56:32 alexanderhiam_: ok, let me know when u push it Mar 16 17:56:40 kiran4399, print the value of ring head and ring tail too Mar 16 17:56:50 let me know what are there values Mar 16 17:56:55 sidbh_: OK.. Mar 16 17:58:04 * nerdboy looking at adafruit_bbio on github Mar 16 17:58:37 hmmmm Mar 16 17:58:43 themacho: yes? Mar 16 17:59:13 themacho: ahh... PM's... no PM's please. Mar 16 17:59:16 https://beagleros.wordpress.com/ in this documentation it is provided that the interfacing of BBB with dc motors altimeter etc in ROS platform Mar 16 17:59:56 themacho: Bootloader - goal there is to have a single unified bootloader between the different beagle platforms. Currently, we have the BBC/BBX sharing one and the BBW/BBB sharing another. Mar 16 18:00:05 but come tp bonescrript integration part i found that little bit complex Mar 16 18:00:32 https://github.com/adafruit/adafruit-beaglebone-io-python/ even Mar 16 18:00:33 themacho: ADC - look at the ML. there are limitations on the driver...goal there is to identify the limitations and improve the driver Mar 16 18:00:35 so i need some help for where to start Mar 16 18:00:53 kiran4399, also try to using simple for loops for data copying from array to shared ram structure array inside add_to_ringbuffer() function , as I'm not sure if memcpy works fine with data transfer between data RAM and shared RAM Mar 16 18:01:21 where can i start in the bootloader project ,and excuse me what's the ML :D Mar 16 18:01:58 sidbh_: I am getting this strange warning Mar 16 18:02:18 kiran4399, gist it Mar 16 18:02:21 :) Mar 16 18:02:33 while I try to print uint32 integer... Mar 16 18:03:28 looks like manual Mar 16 18:03:47 sidbh_: https://gist.github.com/kiran4399/c48aca0ece3a05117e12 Mar 16 18:04:24 sidbh_: that is why I removed integers in the printf for time being... Mar 16 18:04:25 http://wiki.ros.org/roslibjs does anybody have idea about integration of ROS and bonscript Mar 16 18:11:29 kiran4399, I think __packed__ should take care of that, but let's not take a chance here, make all the data types in the structure as char Mar 16 18:11:43 kiran4399, see what I did here, https://gist.github.com/bugobliterator/4392feb74e88d3416472#file-pseudo_userspace-cpp-L4 Mar 16 18:12:00 imitate that everywhere Mar 16 18:12:13 it should resolve the warning Mar 16 18:12:52 let me know if issue persists Mar 16 18:16:19 kiran4399, also change your firmware struct to this: https://gist.github.com/bugobliterator/c043686347d5feafc961#file-pseudo_firmware-cpp-L7 Mar 16 18:17:12 which part of bonescript I should start to edit Mar 16 18:17:24 for integrating ros Mar 16 18:21:42 sidbh_: you want me to remove __attribute__((__packed__)) in the struct in firmware side ? Mar 16 18:23:30 sidbh_: ooohh... It is printing the whole ring..... int[0] is always 10 !!! Mar 16 18:23:51 sidbh_: but the string is always garbage !! :-( Mar 16 18:24:02 we can solve that Mar 16 18:24:04 :) Mar 16 18:24:25 how are you initialising, let me see the current state of your code Mar 16 18:26:55 sidbh_: I am excited to know what it should print ?? Mar 16 18:27:06 what should it print ?? Mar 16 18:29:05 it will print what you write in the buffer inside firmware Mar 16 18:30:17 Oh.. Mar 16 18:30:20 for example as you call add_to_ring_buffer, with string and integer as its parameter, it will save them into the ring buffer and increment the head Mar 16 18:30:31 sidbh_: https://gist.github.com/kiran4399/f0a93d9f7136ba8c8087 this for the firmware side.. Mar 16 18:31:04 now the userspace app will fetch the data and increment the tail and print it until tail == head Mar 16 18:32:24 kiran4399, https://gist.github.com/kiran4399/f0a93d9f7136ba8c8087#file-pru0_firmware-c-L30 uint8_t integers[] not uint32_t Mar 16 18:32:32 sidbh_: https://gist.github.com/kiran4399/09c1b5bc7e531986b437 this is the userspace.. Mar 16 18:34:06 kiran4399, https://gist.github.com/kiran4399/f0a93d9f7136ba8c8087#file-pru0_firmware-c-L842 write "Initial" there instead of "" Mar 16 18:36:37 kiran4399, compile, run and show me exactly what output you are getting after this Mar 16 18:37:49 https://gist.github.com/kiran4399/f0a93d9f7136ba8c8087#file-pru0_firmware-c-L841 uint8_t not uint32_t Mar 16 18:38:04 and size 4 not 3 Mar 16 18:38:43 for the time being comment these https://gist.github.com/kiran4399/f0a93d9f7136ba8c8087#file-pru0_firmware-c-L854-L856 Mar 16 18:40:08 add add_to_ring_buffer(deb_str,deb_ints); RBUFF->ring_tail++; to https://gist.github.com/kiran4399/f0a93d9f7136ba8c8087#file-pru0_firmware-c-L845 Mar 16 18:40:19 s/to/before/ Mar 16 18:41:43 you should get "Initial 0 0" on your terminal after this change Mar 16 18:42:21 let me know if you got all the points I mentioned Mar 16 18:43:45 vvu: looks like that's not the problem after all, PyDict_GetItemString returns a borrowed reference. Decrementing the references causes a segfault Mar 16 18:49:01 praveendath92: not done with the proposal yet. working on it. Mar 16 18:51:08 alexanderhiam: oh well, more digging :) Mar 16 19:04:51 alexanderhiam: i checked mainly all calls you do there...all things return borrowed refs and all the string must not be deallocated by you. Mar 16 19:05:15 maybe a dumb question...if i do a printf in the library code where does the output go ? :) Mar 16 19:05:29 sidbh_: updated one.. https://gist.github.com/kiran4399/f0a93d9f7136ba8c8087 the output is not changing.. initial 0 0 is not printing.. Mar 16 19:05:38 it just goes to stdout Mar 16 19:05:56 join #beagle Mar 16 19:05:57 stdout when i run my python script ? Mar 16 19:06:06 yeah Mar 16 19:06:29 kiran4399, can I have a look at the output, gist it Mar 16 19:08:00 kiran4399, remove this line, its wrong, ring tail is incremented inside the function Mar 16 19:08:23 sidbh_: I was wondering about that !! :-) Mar 16 19:08:24 ohhh, I told you to that, my bad! Mar 16 19:08:27 hehe proposal time :) Mar 16 19:08:45 https://gist.github.com/kiran4399/f0a93d9f7136ba8c8087#file-pru0_firmware-c-L845 Mar 16 19:12:52 alexanderhiam: https://github.com/graycatlabs/PyBBIO/blob/master/bbio/platform/beaglebone/gpio.c#L144 Mar 16 19:12:57 don't you need to free there ? Mar 16 19:13:45 the string? Mar 16 19:13:52 sidbh_: why is my userspace program giving out an infinite loop when there is only 1 entry to print ? Mar 16 19:13:54 https://gist.github.com/ungureanuvladvictor/87409acd583b8852bb0d Mar 16 19:13:59 from https://docs.python.org/2/c-api/arg.html#parsing-arguments-and-building-values Mar 16 19:14:14 yes, the string Mar 16 19:14:22 yeah it shouldn't what is it printing? Mar 16 19:14:24 ah! Mar 16 19:15:09 just free it in PyGPIO_digitalRead Mar 16 19:15:13 sidbh_: garbage.. Mar 16 19:15:40 what's the value of ring->head Mar 16 19:16:11 s/ring->head/ring_head/ Mar 16 19:16:17 vvu: where's that doc from? Mar 16 19:16:24 sidbh_: tell me one thing.. how is the userspace getting pru_firmare's ring->tail ? Mar 16 19:16:24 oh Mar 16 19:16:31 you just told me :P Mar 16 19:16:38 the python api :) Mar 16 19:17:04 kiran4399, https://gist.github.com/kiran4399/09c1b5bc7e531986b437#file-test-cpp-L33 using this we get the pointer to the structure Mar 16 19:17:13 and we are reading that Mar 16 19:17:16 sidbh_: it is going from 0...19 0...19... Mar 16 19:17:37 also print ring_head Mar 16 19:17:54 sorry ring_tail Mar 16 19:18:45 vvu: that seems to only apply to the 'es' format Mar 16 19:19:11 * nerdboy recommends restoring pin state completely Mar 16 19:19:32 both clock and data Mar 16 19:20:31 ah...missed that one Mar 16 19:21:51 vvu: oh, must be the PyString_FromString Mar 16 19:21:57 especially if we're talking to i2c device Mar 16 19:22:13 vvu: https://github.com/graycatlabs/PyBBIO/blob/master/bbio/platform/beaglebone/gpio.c#L148 Mar 16 19:22:35 yes...new reference there Mar 16 19:22:41 yup Mar 16 19:22:59 do you have time to test it ? Mar 16 19:23:27 you have a lot of PyString_FromString in the code :) Mar 16 19:23:48 I sure do! Mar 16 19:23:49 sidbh_: ring->head and ring->tail are different all the time.. Mar 16 19:23:55 have a lot that is Mar 16 19:24:05 ring-> tail is some random value all the time.. Mar 16 19:27:42 vvu: ugh, I have various PyX_FromYs all over the place! Mar 16 19:28:09 alexanderhiam: oh well at least we found it Mar 16 19:28:22 yeah, thanks for noticing that! Mar 16 19:28:35 anytime Mar 16 19:39:30 vvu: looks like that did it, just pushed Mar 16 19:39:44 now on to the other extensions... Mar 16 19:41:24 kiran4399, checked the very simple tweaks it will start working: https://gist.github.com/bugobliterator/5ed636751b318eb85fe0 Mar 16 19:41:38 the user side code is perfect Mar 16 19:42:03 kiran4399, the code is working just tested on my BBB Mar 16 19:42:13 let me know if you get issue Mar 16 19:53:49 hendersa: i want to use my bacon cape potentiometer for analog input Mar 16 19:54:15 how can i configure it for analog input using device tree Mar 16 19:57:07 ankur: You can do that, but you must be very careful not to exceed 1.8v on the ADC pins. Mar 16 19:57:29 You have some pre-made bacon "capes" available to you already in the kernel. Mar 16 19:57:49 You can enable them via the capemgr kernel command line option, as you would other capes. Mar 16 19:58:39 Or, you can cut and put the necessary pieces out of those capes for your own custom device tree overlay. Mar 16 19:59:05 Check out the firmware/capes/BB-BONE-BACON*.dts files in the 3.8 kernel source. Mar 16 20:00:23 hendersa: i learnt a lot about device tree overlays and concluded that it is necessary to configure peripheral hardware without rebuilding the kernel Mar 16 20:00:38 is it all or i am missing out something Mar 16 20:02:24 No, you are correct. It isn't possible to enable every type of subsystem of the BBB at once (because some interfaces share pins with other ones on the P8/P9 headers), so you enable or disable overlays as appropriate to configure what you need. Mar 16 20:02:56 How can I get the open sources please? Mar 16 20:03:10 Several pre-made overlays are built into the kernel and disabled by default. These you can enable via the capemgr kernel command line options. Mar 16 20:03:38 karki: should be "Fevi Kwik" instead :P Mar 16 20:04:08 For custom device tree overlays not in the kernel, you can load them after kernel boot by placing the overlays in the "firmware" directory and loading them dynamically. Mar 16 20:04:22 yes, i found out that "echo BB­ADC > /sys/devices/bone_capemgr.*/slots " can be used to enable analog inputs using device tree Mar 16 20:05:45 Right. There is an overlay with that name implemented in the BB-ADC-00A0.dts file and built into the kernel. It is disabled by default. Mar 16 20:06:15 echo'ing the name of the fragment into the "slots" file triggers the dynamic enabling of overlay. Mar 16 20:06:23 How can I get the open sources please? Mar 16 20:06:54 Omar_: what do you mean by that? Mar 16 20:07:30 The open source cod Mar 16 20:07:33 code * Mar 16 20:07:57 Omar_: what open source code?? Mar 16 20:08:00 for? Mar 16 20:08:44 Didn't you join google summer of code ? Mar 16 20:10:05 Omar_: there are some mentors and potential gsoc students here... but I still don't know what you're asking... Mar 16 20:11:35 hendersa: will it be a good gsoc project if i make a android application for accessing all the functionality of bacon cape? Mar 16 20:11:53 Omar_: you can check the ideas page,there are link to github repositories for some of the projects. Mar 16 20:12:02 bacon? someboday say bacon?? Mar 16 20:12:11 "there are links to" Mar 16 20:12:27 I need to see the open sources to see if I can join with your company or not Mar 16 20:13:00 Omar_: here's where to start: http://elinux.org/BeagleBoard/GSoC/Ideas Mar 16 20:13:16 Thank you :) Mar 16 20:13:21 Omar_: and the mailing list: https://groups.google.com/forum/#!forum/beagleboard-gsoc Mar 16 20:13:21 nerdboy: It's a "Bacon" cape :P Mar 16 20:13:31 ankur: You're on the right path, but you might need several apps to demonstrate all of the various pieces that you'll need. Targeting the bacon cape is probably a good first target. Mar 16 20:14:06 is capemgr on the android image? Mar 16 20:14:32 It is on mine. I made the Android image with the 3.8 kernel for this kind of work. Mar 16 20:15:24 if that thing actually makes bacon, i will back the project right now Mar 16 20:15:45 You'll chip a tooth if you try to eat this bacon: http://elinux.org/images/thumb/8/80/Black-bacon1.jpg/448px-Black-bacon1.jpg Mar 16 20:16:05 hendersa: any solutions for runtime hardware config on 3.14 android images? Mar 16 20:16:51 Does any body know how I can get to the mentor Diego Turcios (DiegoTc) ,say email Mar 16 20:17:15 * nerdboy complains loudly about no truth in advertising any more... Mar 16 20:17:27 tanero: this is the best place to find mentors Mar 16 20:17:43 alexanderhiam: No such luck. Not yet, anyway. I have it on my to-do list to backport the new dynamic overlay support in 3.19 to 3.14 and move the I2C bits for capemgr forward to 3.14. Mar 16 20:17:44 or a post on the ml Mar 16 20:18:03 I actually had a port of the capemgr working on 3.13 a while back. Mar 16 20:18:34 http://i.imgur.com/Pf8Ffek.png Mar 16 20:18:47 oh, I know nothing of this dynamic overlay support in 3.19 Mar 16 20:18:56 can we go back to the beagle-powered bacon machine? Mar 16 20:19:04 panto's working on that Mar 16 20:19:08 * nerdboy votes for that one Mar 16 20:19:08 alexanderhiam: I want to work on a project he started last year's Gsoc.It looks like he is not on this IRC channel Mar 16 20:19:13 oh that's panto's stuff Mar 16 20:19:24 alexanderhiam: ok, will check when i reach my room Mar 16 20:19:25 yup Mar 16 20:20:00 tanero: he often is here. You can post to the mailing list with a summary of what you want to do and he'll see it Mar 16 20:20:08 Yeah, that overlay support made it into the 3.19 mainline. Mar 16 20:20:24 I didn't realize it made it in! that's great Mar 16 20:20:30 alexanderhiam: Ok,gracias Mar 16 20:20:36 hendersa: i think apps for controlling GPIO, reading I2C module, SPI module, PWM and ADC can be easily demonstrated using help of bacon cape Mar 16 20:20:43 rcn 3.19 works on my rev a6a Mar 16 20:21:12 nerdboy: Does the HDMI audio work in that version, though? Mar 16 20:21:24 didn't try it Mar 16 20:21:37 * hendersa notes that it is always a juggling act to find the right kernel with the right features. Mar 16 20:21:48 video/usb/ethernet Mar 16 20:21:57 I'm not sure if OpenGL ES is working with 3.19 right now, either. Mar 16 20:22:28 sgx builds and omaplfb loads/works Mar 16 20:22:41 is the next shipped image still going to be 3.14 then? Mar 16 20:22:50 x seems a little edgy on top of that, but... Mar 16 20:23:22 ankur: You can tailor your proposal to demonstration of all of the bacon cape features, if you would like. There are other interfaces to target beyond those, too. Mar 16 20:23:37 I would expect the next one to be the TI BSP 3.14. Mar 16 20:23:52 * nerdboy lucky he got bacon sandwich Mar 16 20:24:02 But it really comes down to what is available and working and what we all feel comfortable with. Mar 16 20:24:55 I just want a kernel that supports GLES, HDMI audio/video, and has the cape manager. Mar 16 20:25:06 * hendersa falls to his knees and shakes his hands at the sky. Mar 16 20:25:12 and wrapped in bacon... Mar 16 20:25:39 seems a bit odd to lose dynamic overlays with capemgr with 3.14 if it's supported in 3.19 Mar 16 20:25:46 We'll get there. All these moving bits to nail down. Mar 16 20:25:56 hendersa: did you enable that hdmi audio codec thing? Mar 16 20:26:20 The audio codec is for an audio cape, not the HDMI audio on the TDA framer. Mar 16 20:26:23 *in the HD audio config... Mar 16 20:26:41 there's a generic one i thought Mar 16 20:27:02 * alexanderhiam is speaking as a maintainer of a userspace library that relies on dynamic overlays Mar 16 20:27:10 hendersa: There are other interfaces to target beyond those. Can, you provide me just one example so that i can think in that direction. Mar 16 20:27:12 shouldn't matter if it's a cape or not Mar 16 20:27:50 alexanderhiam: There was pretty intrusive changes to 3.8 to get the capemgr in there. Aside from dynamic overlay support, there was also code to hit the I2C EEPROMs on the cape to get their info. Mar 16 20:28:10 The dynamic overlay support is now mainline in 3.19, but not the I2C bits. Mar 16 20:28:31 ankur: Well, the CAN bus is an example of one. Mar 16 20:28:36 right Mar 16 20:28:47 of course, there's all this stuff too CONFIG_OMAP4_DSS_HDMI_AUDIO=y and what-not Mar 16 20:29:17 hendersa: you enabled all that stuff? Mar 16 20:29:22 * Abhishek_ 's next planned HW design is an audio codec cape Mar 16 20:30:51 nerdboy: Yup. I have all that enabled in 3.18.5, I believe. Mar 16 20:32:27 no boop noises at all? Mar 16 20:32:40 * hendersa is looking through the device tree in arch/arm/boot/dts. Mar 16 20:32:42 Abhishek_: I've heard many people asking for that Mar 16 20:33:16 let me know if there's any specific feature I could incorporate Mar 16 20:34:01 hendersa: i see two SND_AM33XX_SOC modules Mar 16 20:34:48 Ahishek_: stereo in and out would be ideal of course Mar 16 20:35:21 last one is CONFIG_SND_SOC_HDMI_CODEC=m under CODEC drivers Mar 16 20:35:31 nerdboy: It might just be because of the device tree overlay settings. I see the am335x-boneblack-nxp-hdmi-no-audio.dtsi being included in am335x-boneblack-base.dts. Mar 16 20:36:21 Abhishek_: would you use the PRU? I seem to remember there not being a full I2S bus on the headers Mar 16 20:36:29 * nerdboy fresh out of ascii faces Mar 16 20:37:51 Is it? I don't see using the PRUs for audio though Mar 16 20:38:57 Abhishek_: I looked into it once, and I remember there being something like some missing pins Mar 16 20:40:22 you might be able to do an i2s or pcm stream through the pru Mar 16 20:42:00 nerdboy: Looks like the include for am335x-bone-bbb-exp-c.dtsi is commented out in am335x-bone.dts. Mar 16 20:42:17 * nerdboy goes to the cowboy barbershop Mar 16 20:42:24 bbiaf Mar 16 20:43:35 nerdboy : hey Mar 16 20:43:46 be back in a fortnight? Mar 16 20:43:57 Nope. Wrong one. Its the boneblack one. Geez. Mar 16 20:44:08 * hendersa looks at device tree files and grumbles. Mar 16 20:44:23 nerdboy : you are mentoring the beagleRT project . Mar 16 20:44:24 I need to write this down and follow the trail of .dts/.dtsi includes. Mar 16 20:44:51 nerdboy : ping Mar 16 20:45:47 praveendath92: finished DB exam? Mar 16 20:46:08 No. Still a lot of time to go. Mar 16 20:46:19 And a lot of pages to read. Mar 16 20:46:37 when do u have it / Mar 16 20:48:03 31st March Mar 16 20:48:30 hendersa, it's setup as: -.dtb Mar 16 20:49:17 rcn-ee: So the 3.18 tree from your repo had HDMI audio out of the box? Mar 16 20:49:33 rcn-ee: Or am I missing a spot where I need to comment it in? Mar 16 20:50:17 I don't think hdmi audio even worked, last i checked, there was 3 different patchset's trying to do the same thing. i just ended up droppign it till they figure it out on their own.. Mar 16 20:51:05 OK. That is what I figured. It looked like it should work with 3.18, but it wasn't and I didn't have the time to sink into it to sort it out. Mar 16 20:56:38 I am Wadinga Leonard from the University of Buea, Cameroon. I would like to work for beagleboard this year during the gsoc2015. In particular on the common bootloader for the different beagleboards/beaglebone. Mar 16 20:58:39 Hello, I am very interested in system programing, what is the inscription procedure for GSOC? Mar 16 20:59:42 Talking to some mentors on irc yesterday, i was informed there is already a common bootloader for the various devices: uboot. So i was wondering what is the objective of this project or are there some bugs to fix on this bootloader Mar 16 21:00:28 wadleo, the goal is allow us to have one "image" for BeagleBone/BeagleboneBlack + BeagleBoard x15... by tweaking the SPL "MLO" binary... Mar 16 21:00:54 It is still possible to looking for a mentor, now? Mar 16 21:02:33 rcn-ee: thanks for the tip. I have checked this devices u mentioned but can u give me a link to the SPL "MLO" binary or some documentation Mar 16 21:03:28 saruta: Its just starting,here you will always meet mantors,to get the IRC nick of mentors check at the bottom of http://elinux.org/BeagleBoard/GSoC/Ideas Mar 16 21:04:00 tanero: thank you so much Mar 16 21:04:03 wadleo, http://git.denx.de/?p=u-boot.git;a=blob;f=doc/README.SPL;hb=HEAD You'll also need to download the full datasheets for both the am335x & am57xx familes from TI.. Mar 16 21:04:36 wadleo, more here: http://git.denx.de/?p=u-boot.git;a=tree;f=doc/SPL;hb=HEAD Mar 16 21:09:29 rcn-ee: When is the final Jessie image expected? Mar 16 21:10:07 Abhishek_, the bug count is down to 85: https://udd.debian.org/cgi-bin/rcblog.cgi so april-ish.. Mar 16 21:10:43 the weekly autobuilder now is only doing jessie... (exept for the machinekit guys who aren't ready for jessie yet) http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Debian_Image_Testing_Snapshots Mar 16 21:20:34 jkridner: [reminder] Kindly check this http://bit.ly/1Evqpf9. I really need some feedback before I can start working on my proposal Mar 16 21:21:05 * jkridner was just going through some of the posts. Mar 16 21:22:37 msg wadleo please discuss this in public. Mar 16 21:22:54 wadleo: no PM's. please discuss this on the channel. Mar 16 21:23:23 okay, I am looking for a mentor for the ADC driver. Mar 16 21:41:34 webglider: k, I made some kind of response. Mar 16 21:43:28 jkridner: https://cdn.hackaday.io/images/3171841426081548319.jpg Mar 16 21:46:45 would be assembling it soon Mar 16 21:55:35 jkridner: Thanks for taking the time. Could you also check my previous reply on the same post: http://bit.ly/1Evqpf9 Mar 16 22:04:11 we have to do the ADC driver from scratch or we have already a little driver? Mar 16 22:12:23 please check the tree Mar 16 22:12:26 and mailling list Mar 16 22:12:45 the first thing needed for that project is to diagnose and figure out what the pain points are Mar 16 22:15:22 Abhishek_: is this a music quality CODEC or comm quality codec? Mar 16 22:15:54 music quality, more multimedia-oriented Mar 16 22:16:30 ah Mar 16 22:16:48 Abhishek_: so in otherwords, you have no respect for the current HDMI stuff? :D Mar 16 22:18:00 more as in augmenting my newly designed LCD cape. I pointed jkridner to the link a few lines above. Mar 16 22:18:42 Abhishek_: is the LCD cape using the display controller on the am335x? Mar 16 22:18:48 yeah Mar 16 22:19:44 then you can't run BeagleLogic on it :D Mar 16 22:20:12 yep, this is kinda orthogonal to BeagleLogic Mar 16 22:20:19 Oh :P Mar 16 22:20:33 Abhishek_: why yet another LCD interface? Mar 16 22:21:24 I found some tablet LCDs really cheap and there wasn't any such prior design for the BBB so I made one Mar 16 22:21:41 http://theembeddedkitchen.net/coming-soon-yet-another-beaglebone-displaycaptouch-cape/391 Mar 16 22:22:08 ah heh Mar 16 22:22:30 with the way the BeagleBone processor is - is that even worthwhile? Mar 16 22:23:50 the BBC/BBX is more meaningful for a proper display Mar 16 22:25:10 With the way SGX support is, it might not be good for 3D graphics and so but still, for some embedded control panel applications, it might be worthwhile Mar 16 22:25:41 it is more the SGX Mar 16 22:26:19 compared to the DM37x, the AM33x's display stuff is like comparing a bicycle to a BMW luxury car Mar 16 22:26:51 SGX can be used for other fun stuff like parallel processing Mar 16 22:27:09 I see Mar 16 22:27:19 even with out OpenCL Mar 16 22:27:47 the DM37x can do fun stuff like overlays, rescale, color conversions, etc all in HW Mar 16 22:28:16 does the Linux kernel use all those features optimally? Mar 16 22:28:19 I personally would think the LCD controller would make a nicer DDS Mar 16 22:28:24 with the right kernel, yes ;) Mar 16 22:28:50 you mean the signal generator DDS? Mar 16 22:28:55 yes Mar 16 22:29:04 feed the LCD lines into a DAC Mar 16 22:29:11 put the waveform into the framebuffer memory Mar 16 22:29:53 How does it compare with DMA2D on the STM32F429? Mar 16 22:30:18 donno, haven't spend much time on the Fn series where n > 1 Mar 16 22:34:02 so, for the adc stuff, we need to improve the drivers of /drivers/iio/adc ? Mar 16 22:36:19 saruta: yes... after triaging what the problems are and identifying what will fit in the GSoC timeframe Mar 16 22:43:30 maybe we can just unified the drivers... I mean simplify the access of the hardware's functions Mar 16 22:48:24 there is a driver Mar 16 22:48:32 the hw supports a lot of features not in the driver. Mar 16 22:48:42 look at the ML archives Mar 16 22:50:41 it is about hardware sampling? Mar 16 22:51:04 there is a lot of settings not exposed. Mar 16 22:51:10 hw sampling is one Mar 16 22:51:22 filtering is another Mar 16 22:51:34 selection of references is another Mar 16 22:51:36 ... Mar 16 22:51:39 read the TRM Mar 16 22:51:59 okay, okay Mar 16 22:52:09 this is not a trivial project Mar 16 22:54:12 but, what the skills needed? Mar 16 23:32:06 webglider: ping Mar 16 23:57:02 * nerdboy considers another espresso Mar 16 23:59:20 don't mess up your tastebuds for next week.. Mar 17 00:00:28 tastebuts? Mar 17 00:00:31 buds Mar 17 00:00:45 rcn-ee: where/when is the meetup? Mar 17 00:01:23 ds2, elc next week..San Jose Ca... i figure we can borrow a meeting room, or take over a food joint. ;) Mar 17 00:02:06 rcn-ee: "good luck" :) Mar 17 00:02:19 rcn-ee: how are you organizing it? Mar 17 00:03:31 * nerdboy still considering... Mar 17 00:03:41 ds2, at somepoint, the "presentations" end, i'll just walk to the front and say the bbb meeting will now start.. Mar 17 00:03:42 since i have get off my ass and make it Mar 17 00:03:49 rcn-ee: or should I just find you in person? ;) Mar 17 00:04:18 rcn-ee: oh okay... I'd be interseted in attending but I donno how integrated is ABS and ELC (officially, I am attending ABS) Mar 17 00:04:25 rcn-ee: i can bring chromebooks for googly video hangout Mar 17 00:04:34 with pirate headgear even Mar 17 00:04:56 * nerdboy was pirate for last night's mtg Mar 17 00:05:03 ds2, last year they were at the same hotel... if i remember right, all meeting ended around 5/6... so we just steal a room. :) Mar 17 00:05:17 or just make everyone hit "random" Mar 17 00:05:19 nerdboy, then i'll bring my chromebook too.. Mar 17 00:05:55 if you call it a BoF it can be official Mar 17 00:06:13 maybe with free refreshments Mar 17 00:06:16 rcn-ee: 'k - wasn't there last year Mar 17 00:06:43 * nerdboy looks sideways denix Mar 17 00:06:47 *at Mar 17 00:06:48 ds2, yeah it was pretty unified... the opening slide was "ELC & ABS".. Mar 17 00:07:18 nerdboy, we just got to sneak a couple cases of beer in, i think it was a couple blocks away.. Mar 17 00:07:39 no BoF? Mar 17 00:09:52 haven't read the rules. ;) Mar 17 00:10:50 * nerdboy poking upstream Mar 17 00:11:32 do we need ot pitch in and hire jkridner a fresh babysitter? Mar 17 00:11:57 babysitter? to sit the list or me? Mar 17 00:11:58 jkridner: you get half for nap time, half for mtg time Mar 17 00:12:26 * nerdboy assumed you have "responsibilities" Mar 17 00:13:28 well, that's a bit of a pass as beagleboard.org is my priority. Mar 17 00:13:37 still.... I do have distractions. Mar 17 00:13:54 could probably pull more of the community in to help with some of those distractions. Mar 17 00:14:16 did you make a list already? Mar 17 00:14:46 i was actually just wondering if there were any special instructions for the application period Mar 17 00:14:49 list of distractions? no. seems to change faster than I can list. Mar 17 00:15:01 engage with students first and foremost.... Mar 17 00:15:16 wouldn't it be nice if everything worked like the PRU - single cycle for everything? :D Mar 17 00:15:20 just help them write proposals that you think they can actually accomplish... Mar 17 00:15:27 how do we tell who has done the mechanical stuff? ie, crossbuild, etc? Mar 17 00:15:37 and then then give them some kind of little tasks that can test them. Mar 17 00:15:59 don't trust them because it "sounds" like they know what they are talking about... Mar 17 00:16:07 make them show you somehow. Mar 17 00:16:22 notches on gunbelt? Mar 17 00:16:32 small tasks, but something that shows they have the right skills to build an environment, etc. Mar 17 00:17:05 previous contributions are a great measure, but rare for a program that is meant to bring students INTO the open source way. Mar 17 00:17:30 for some projects it requires more then that Mar 17 00:19:20 * nerdboy should just put that on the wiki Mar 17 00:20:32 jkridner: the sticky group threads from 2014? are they still relevant now? Mar 17 00:20:49 * nerdboy notes repo does not exist Mar 17 00:21:56 jkridner: would everyone scream if we added that little bit of process? Mar 17 00:22:16 oh? Guess I will replace it. Mar 17 00:22:24 like a wiki page with what people have done so far? Mar 17 00:22:47 checkoff toolchain, u-boot, kernel, something... Mar 17 00:23:26 *if* they have hardware, can build that stuff and make/boot sdcard Mar 17 00:30:48 nerdboy: just updated the post. thanks for the heads-up Mar 17 00:31:11 nerdboy: process away!! Mar 17 00:31:39 nerdboy: just be sure to pull in av500 and myself for a bit of review, but heard away! Mar 17 00:31:43 herd away? Mar 17 00:32:01 * jkridner is feeling grammatically challenged lately for unknown reasons (distractions) Mar 17 00:32:29 nerdboy: note we have the private channel and github that we need to reinvigorate. Mar 17 00:32:47 collaboration tools = github, irc, e-mail. Mar 17 00:32:55 + wiki (elinux.org) Mar 17 01:00:56 PRU isn't exactly single cycle, those accessing memory are 2+ cycles Mar 17 01:00:57 rcn-ee: BoFs look full, fallback to Plan B(eer) Mar 17 01:11:01 * nerdboy also gets 2+ cycles Mar 17 01:11:08 one per cheek... Mar 17 01:21:01 I can't remember, do we have a template for proposals? Mar 17 01:21:28 * nerdboy surrounded by inlaws... fires up both cheeks... Mar 17 01:31:40 is the one in jason's gihub repo current? Mar 17 01:32:49 jkridner: is this still correct, or no? https://github.com/jadonk/gsoc-application Mar 17 01:33:55 I was thinking the melange app Mar 17 01:34:54 Mar 17 01:35:47 I feel like gsoc had their own template last year... I'll keep digging Mar 17 01:37:18 oh, what do you know, it's just in the faq: http://www.google-melange.com/gsoc/document/show/gsoc_program/google/gsoc2015/help_page#5._What_should_a_student_proposal_look Mar 17 02:18:23 alexanderhiam: we had one in my year :) Mar 17 02:18:29 but it got lost somehow in the process Mar 17 02:57:55 jkridner: i can't anything about the sauna... Mar 17 02:58:05 this club does have a suana, right? Mar 17 02:59:01 * nerdboy prepping with homemade black-bean-with-leftover-pork chili **** ENDING LOGGING AT Tue Mar 17 02:59:58 2015