**** BEGIN LOGGING AT Wed Mar 24 02:59:57 2021 Mar 24 04:45:41 using that rpmsg-lite for python for pru statement Mar 24 04:46:15 Okay, let me take a look Mar 24 04:46:46 Okay so PRU side Mar 24 04:47:04 Yeah, it might be just port and recompile Mar 24 04:47:11 Best case Mar 24 04:47:57 But obviously a lot more details Mar 24 04:48:12 porting itself might be a complete gsoc project if Mar 24 04:48:53 also, python transpiled makes more sense than setting up a interpreter, what say? Mar 24 04:49:08 like what simpru was doing Mar 24 04:49:45 Python transpiled to C? Mar 24 04:50:26 yes Mar 24 04:52:11 Cythonize ? Mar 24 04:52:22 It’s going to be tricky Mar 24 04:52:52 Python has libraries too - how do you bring libs in? Mar 24 04:53:49 more like micropython port Mar 24 04:54:06 PRU doesn’t have enough resources Mar 24 04:54:13 Last I checked Mar 24 04:54:27 if there's a python frontend to consume python file, we could make something in the backend Mar 24 04:54:40 yes, transpiled micropython Mar 24 04:54:51 It’d be really cool Mar 24 04:55:22 we could drop library support for now. only make the core functionality working as gsoc project Mar 24 04:56:28 That would be awesome to work on Mar 24 04:59:01 can you scout around micropython docs to see if it is possible Mar 24 04:59:29 specifically look if they provide a python frontend to convert code into AST graph Mar 24 04:59:56 or else you will have to roll your own parser and lexer Mar 24 05:01:31 I am on it Mar 24 05:18:22 one more idea I have, you can page PRU's RAM using dma Mar 24 05:18:34 and then port micropython to run on it Mar 24 05:19:03 pru can access memory of arm core Mar 24 05:19:21 so, write a paging handler on pru 0 Mar 24 05:23:39 Abhishek jkridner is this feasible or even technically possible? Mar 24 05:25:57 https://summerofcode.withgoogle.com/archive/2020/projects/6579687536459776/ Mar 24 05:26:32 I asked him, he said reverse will be very tough. Mar 24 05:26:52 you will have to make everything yourself. Mar 24 05:27:19 In that case Clang gave AST graph, not sure if micropython can do same Mar 24 05:28:09 http://docs.micropython.org/en/latest/develop/compiler.html#parsing Mar 24 05:28:09 Yes there's no frontend Mar 24 05:30:02 http://docs.micropython.org/en/latest/develop/compiler.html#emitting-bytecode Mar 24 05:30:21 see in this part, instead of emitting bytecode, make it emit C code Mar 24 05:30:26 it is doable ig Mar 24 05:46:06 First we need pru_rp_msg_lite Mar 24 05:47:27 Porting it will take time right ? Mar 24 05:47:58 * satacker[m] uploaded an image: (40KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/XMRHHHdsddSfpUzBXyvPnLXk/image.png > Mar 24 05:48:06 ---from its website Mar 24 05:59:58 I'm not sure how much this will benefit us Mar 24 06:00:30 You need to touch yellow files only Mar 24 06:22:29 Alrighty Mar 24 15:58:39 Hello Students! I am facilitating as an admin for this year's GSoC. Please let me know if you have any questions but also if you have not found a mentor on this chat space to discuss your proposal. Mar 24 16:01:21 Our mentors are here to assist you in developing your proposal. So before you lock in the official form on the GSoC site, work with mentors by putting together your proposal onto the git so you can collaborate, edit and get the best proposal together. Recommended format is here https://elinux.org/BeagleBoard/GSoC/2021ProposalTemplate Mar 24 16:28:36 You are awesome to be on at this time Abhishek and at India friendly times also - thank you! Mar 24 16:31:07 Hi Abhishek Mar 24 16:35:30 Hi all Mar 24 16:36:25 Hi Mar 24 16:36:50 Hi all ! Mar 24 16:38:41 Hello everyone! Mar 24 16:39:11 Hi pratimu Hi satacker Hi lorforlinux Hi vedant16 good to "see" everyone today - I hope all are well from the pandemic Mar 24 16:40:13 hi all Mar 24 16:41:47 Any updates on simppru statement Archisman Dey ? Mar 24 16:42:13 Hi cwicks everything is fine here, hopefully its same for you too. Mar 24 16:42:14 Regarding usb configfs in device tree : not really sure how we can use libusbg with device trees, never found an example Mar 24 16:42:39 thanks for asking :) hope everything is good on your end Mar 24 16:45:52 Hi everyone, joining in ~15 minutes after having my dinner Mar 24 16:47:16 Usb can be configured in device tree , but in the references ; a link for libusbg is mentioned. The ppt in the link describes some features of libusbg and it's gadget schemes . Mar 24 16:48:53 Hi, bumping the message. Opened a discussion on the mailing list Mar 24 16:54:46 I mostly understood what your code (lexer.l, parser.y, ast.c, symbol_table.c, code_printer.c) does, these are the files that need to be changed for improving the error handling in bison, updating the grammar to add support for return statements anywhere in the function, and keeping control statements inside if-elseif-else, right? Mar 24 16:55:57 for return statment handling you don't need to touch code_printer, it needs to be handled in AST graph generation step Mar 24 16:56:22 other than that you are right Mar 24 16:56:29 Also, I noticed that range in for loops only support range1 : range2 with range1 < range2 right now, I thought it would be useful to add more general range declaration like range1: range2: step_count where these are any integers Mar 24 16:57:07 Yes, it can be done, sure. synatx changes are welcome Mar 24 16:57:39 also you can add functionality to call C code inline Mar 24 16:57:49 this should also maintain backwards compatibility Mar 24 16:59:24 yes, it won't matter much if it doesn't. We can always warn users during a release. Mar 24 16:59:59 this would have to have limits right? Mar 24 17:00:22 limits as in ? Mar 24 17:00:43 I was thinking we could add support for calling c functions, this could be useful for someone wanting to use functions written by someone else Mar 24 17:00:58 cwicks: I would like to contact Hunyue Yau as they replied to my submission on the google pages thread and I would like to continue developing the idea for a proposal Mar 24 17:01:35 sounds good, it would cover lot of things. Mar 24 17:01:43 I am not sure how adding support for writing arbitrary C code will look like, can you provide an example? Mar 24 17:01:57 jduchniewicz: his tag is ds2 Mar 24 17:02:21 what I had in mind was, that code won't interact with simppru, but simply be copied to the transpiled C code Mar 24 17:03:22 because variables and other things need to be synced from C to simpPRU, and we have to check that this does not have any contradictions, so we basically need to write part of a C compiler if we need to add arbitrary C support right? Mar 24 17:04:06 Oh, then my reply is irrelevant now Mar 24 17:04:35 not necessarily, you can let the C compiler take care of it. add this as a beta feature only. Mar 24 17:05:37 I have to think about how exactly I could do this :thin Mar 24 17:05:52 🤔* Mar 24 17:06:00 :thin was a typo Mar 24 17:06:05 This can be a secondary goal, if you have time left Mar 24 17:06:17 you edit on matrix Mar 24 17:06:50 > <@satacker:matrix.org> Usb can be configured in device tree , but in the references ; a link for libusbg is mentioned. The ppt in the link describes some features of libusbg and it's gadget schemes . Mar 24 17:06:51 > Just wondering why the reference of libusbg for device trees Mar 24 17:06:51 lorforlinux also what do we mean by configfs in device tree Mar 24 17:07:00 * you can edit on matrix Mar 24 17:07:11 Okay, this a good idea Mar 24 17:07:51 jduchniewicz: His chat name is ds2 Mar 24 17:08:25 vedant16: You also mentioned adding support for the beaglebone blue, and I also noticed there is an issue about this on github, but I'm not sure how to go about this Mar 24 17:08:56 you need to generate pin mappings Mar 24 17:09:38 https://github.com/VedantParanjape/simpPRU/blob/master/generate_pinout_table_json.py Mar 24 17:10:24 then convert the json to a string and add here: https://github.com/VedantParanjape/simpPRU/blob/master/include/pin_config.h Mar 24 17:10:24 Okay, I will look into this and ask you if there are any further doubts Mar 24 17:11:59 https://github.com/VedantParanjape/simpPRU/blob/master/src/pin_config.c also look here Mar 24 17:12:46 Also, another part of the project would be to add unit tests for every feature right? Mar 24 17:13:05 for every existing and new feature Mar 24 17:13:13 ig you have gone through last years mailing lis Mar 24 17:13:16 list* Mar 24 17:13:28 yes, that's something we need Mar 24 17:13:36 yes Mar 24 17:14:17 it was also in https://github.com/VedantParanjape/simpPRU/wiki#things-to-be-implemented-in-future Mar 24 17:15:39 This also mentions "Add support for using PWM using eCAP module", I have not looked into this at all Mar 24 17:16:29 This is very tough thing to make it work 😆 Mar 24 17:16:36 Reminder to students - start your proposal on the git with this template to guide you https://elinux.org/BeagleBoard/GSoC/2021ProposalTemplate Mar 24 17:16:49 I think at this point, it's a bug in the processor Mar 24 17:17:06 This way mentors can see your ideas summarized help edit them and encourage your proposal for GSoC to take form Mar 24 17:17:21 I have no idea about this tbh Mar 24 17:17:53 Then I am not looking into this 😅 Mar 24 17:18:30 Thanks, did not know about this! Mar 24 17:18:33 <@satacker:matrix.org> Usb can"> Not sure if I mentioned configfs with DT anywhere 🤔 Mar 24 17:19:28 Hi @lorforlinux, I’m going to do a proposal for the BeagleBone AI / Robotics Cape librobotcontrol support. I’ve worked extensively with robotics control systems in the past so this should be a nice project Mar 24 17:20:30 vedant16: Also, are comments in simpPRU supported yet? The lexer has support I saw, but it is written TBD on the doc. Mar 24 17:21:44 yes they are Mar 24 17:22:28 Oh, okay, ig the doc needs to be updated then Mar 24 17:22:34 ohh, i forgot to write docs for it ig, if you have time send a PR Mar 24 17:22:42 Hey Aedan Cullen Mar 24 17:22:42 That's awesome, Let us know if you need any help regarding the proposal. Also, you have to complete the hello world task, do that asap :) Mar 24 17:22:42 Good luck Mar 24 17:22:49 Sure, I will Mar 24 17:23:16 For the project OpenGLES acceleration for DL can some mentor please suggest a few good resources with respect to the BBAI for hardware acceleration.I have already run OPENGL ES on RPI 4 and was planning to accelerate atleast 1 layer using OPENGL ES in the coming week. Mar 24 17:24:32 I am almost done with the hello world task anything specific i could contribute to? Mar 24 17:25:11 Thanks for the help, I will write a draft proposal before the next meeting then Mar 24 17:25:38 sounds good Mar 24 17:27:47 Its mentioned as goal *USB configfs in device tree such that ...* Mar 24 17:30:27 The example in bindings shows almost everything Mar 24 17:30:49 Yadnik Bendale I can help in validating things. Mar 24 17:31:28 ds2 is the primary mentor for your project, he can answer your questions better. Mar 24 17:36:05 usb_alias:usb@blah{ Mar 24 17:36:46 * satacker[m] < https://matrix.org/_matrix/media/r0/download/matrix.org/qqxCrJcRPXslTRTMFxJOJQdw/message.txt > Mar 24 17:38:02 Right now i am not sure how kernel looks for UDC through this Mar 24 17:40:20 * lorforlinux[m] < https://matrix.org/_matrix/media/r0/download/matrix.org/hPbzfkQFfRdSdXPWLlwfXnnl/message.txt > Mar 24 17:40:55 Try one thing Mar 24 17:41:31 Create a device tree file for thia, compiler it and load. Mar 24 17:42:25 Then search for the reference from / Mar 24 17:42:29 Then you'll understand this better. Mar 24 17:44:25 Yeah sure, thanks for the heads up Mar 24 17:51:22 <@satacker:matrix.org> usb_ali"> I looked it up and the "compatible" field allows for the kernel to identify UDC Mar 24 17:51:42 * I looked it up and the "compatible" property allows for the kernel to identify UDC Mar 24 17:51:42 Thank you Mar 24 18:08:54 DhruvaGole[m]: unfortunately, the images don't show up in irccloud... I'll try to look on matrix. Mar 24 18:10:01 vedant16[m]: I think paging is technically feasible to enable micropython, but it requires an understanding of compilers such that the compiler can be modified to perform the page loads. Mar 24 18:11:26 Ohh, can it be done, like paging on pru0 and program on pru1 ? Mar 24 18:18:03 Sure Thank you Mar 24 22:54:40 you need the compiler to create pages out of the code segments and perform page-swaps on calls. Mar 24 22:56:10 there would be a maximum function size. **** ENDING LOGGING AT Thu Mar 25 02:59:56 2021