**** BEGIN LOGGING AT Thu Apr 08 02:59:57 2021 Apr 08 12:06:42 cool Apr 08 12:06:58 Abhishek: Please comment on my proposal if you are free, there's not much time left for submission deadline Apr 08 12:07:15 jkridner: It doesn't work :( Apr 08 12:07:17 When is it? Apr 08 12:07:22 13th Apr 08 12:07:25 5 days left Apr 08 12:07:52 Give me some time in the evening today Apr 08 12:08:33 Looks like we are operating on different scales of urgency. Apr 08 12:08:47 Submit whatever you have and keep updating it Apr 08 12:09:00 Sure, I am free anytime Apr 08 12:10:39 Note to all students: (i) you must submit your proposals on GSoC site using the formal process (make a pdf and upload). Merely having a eLinux proposal and not putting anything on GSoC site will disqualify you Apr 08 12:12:40 (ii) When considering proposals for selection, the GSoC copy of the proposal will be authoritative and used to judge. You can edit eLinux proposals but any new information added post app deadline will be only considered after proposal selection, if it happens. Apr 08 12:14:25 Okay, I already made the account a few days back, will add the pdf and descriptions too Apr 08 12:14:51 * vedant16[m] uploaded an image: (298KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/PRBbIjtFAiVgokLZqZrBxcXJ/image.png > Apr 08 12:14:59 You can download pdf of elnux from here Apr 08 12:15:13 Please make sure that you submit your proposals on GSoC website and keep updating regularly. GSoC has a zero exception policy and there are no appeals permitted under any circumstances Apr 08 13:08:45 mentors: I have shared my draft via the GSoC portal.. Please give me a go ahead if it looks good so then I will upload the PDF Apr 08 13:09:04 you can also view it here: https://elinux.org/BeagleBoard/GSoC/2021_Proposal/DhruvaGole Apr 08 13:46:01 * DhruvaGole[m] < https://matrix.org/_matrix/media/r0/download/matrix.org/mcEXTsqNTZlswEgiNsxWWhoU/message.txt > Apr 08 13:55:34 lorforlinux: ? Apr 08 14:27:30 Looks good to me ๐Ÿ‘ Apr 08 15:03:16 ds2: I read your mail. Tried to dig more into the TIDL stuff, but it seems that either we chop the network into many little pieces and try accelerating them 1 by 1 (probably a terrible and hacky idea) or not use TIDL at all. It would be a shame to lose the potential of EVEs but well, we cannot help it Apr 08 15:04:10 ds2: is there some other way of using EVEs for accelerating convolutions? As this is the only operation that is quite hardware specific in YOLO network (convolutions generally are quite computationally intensive) Apr 08 15:05:54 ds2: I am not sure about the future of EVEs on BB's but it would be nice to make most use of them Apr 08 15:23:46 I think the YOLO idea is finally borked Apr 08 15:25:25 the idea of using GPGPU seems to be good enough, however should it be a simple tutorial or more of a unified framework for doing computations on the GPGPU, if so then I would be interested :) ds2 Apr 08 15:26:00 anything that allows for reusing and for some flexibility with implementation is compelling for me Apr 08 15:27:41 I could do something like creating a framework (modifying your existing one) and writing a tutorial for GPGPU computations on BB's ds2 Apr 08 15:27:46 jduchniewicz: If you had to pick another proposal, which one would interest you? Apr 08 15:28:19 so in the future it would be a good building step for deploying DL models with GPU's Apr 08 15:29:23 since the YOLO one seem to be infeasible (at least for now), the GPGPU with GLES one. Micropython for BeagleConnect Freedom is also interesting, but I have no idea about the last one whatsoever Apr 08 15:29:52 For BeagleConnect, reach out to jkridner asap Apr 08 15:30:20 He might need to be bugged several times for a response if he doesnโ€™t see this right away Apr 08 15:31:07 yeah sure, I am wondering what is exact implementation detail for this jkridner Apr 08 15:31:09 Try to stick around when he appears; itโ€™s US eastern so he might Apr 08 15:31:16 oh he has so many identities Apr 08 15:31:36 anyway he will be pinged on the IRC for mentioning jkridner Apr 08 15:34:49 I can do either of these proposals as they seem quite interesting to me, don't know how much knowledge I need for the Micropython one Apr 08 15:35:14 Both of them have quite big impact (or at least seem to have :) Apr 08 15:35:37 ds2: I will summarize my thoughts in a mail Apr 08 15:37:08 jduchniewicz1: looking at BeagleConnect? Apr 08 15:37:37 So far quite casually (borked my YOLO proposal just seconds ago) Apr 08 15:37:59 either this as an alternative or the GPGPU one Apr 08 15:38:34 YOLO not working out? Apr 08 15:39:09 Have you had a chance to look at https://github.com/jadonk/beagleconnect ? Apr 08 15:42:42 I just skimmed it, will dig more into this while waiting for ds2 Apr 08 15:43:15 yes, it seems like TIDL does allow only for all or nothing approach with computing the network layers Apr 08 15:43:38 so we cannot do it the way we planned Apr 08 15:43:58 also considering creating acceleration framework on GPGPU Apr 08 15:58:06 satacker: How's the progress for your project proposal(s)? Apr 08 15:58:40 Do you mean about beagle-config? Apr 08 15:59:27 yeah Apr 08 16:00:59 I am making the changes suggested by Abhishek Apr 08 16:00:59 Like adding recipe outlines, comparing available text based guis and some more little but important things Apr 08 16:00:59 Do u have any suggestions too? Apr 08 16:02:21 Wait, looking at you proposal page Apr 08 16:04:15 There's one more doubt especially for you, regarding cape configuration. How do I best configure capes? Apr 08 16:05:10 Also for capes as most of them are not available on Indian sites, from where did you order? Apr 08 16:06:47 Not most , but i couldn't find any on robu.in Apr 08 16:06:58 Don't worry about the hardware, BeagleBoard.org will provide you with all the required hardware for you project Apr 08 16:08:08 Why we creating a GUI with qt? Apr 08 16:08:11 I'm working on getting stock of hardware needed at robu.in with help of vedant16[m] Apr 08 16:08:27 but... as mentioned, Beagle will buy hardware needed. Apr 08 16:09:17 A client or host side for configuring network sharing (wrapper around iptables and network forwarding) Apr 08 16:10:42 And firewall, to provide a seamless one click experience Apr 08 16:10:42 Ssh within the app Apr 08 16:11:04 Will there be a terminal based GUI for BeagleBone? Apr 08 16:11:14 Yes Apr 08 16:11:56 Are you aware of `config-pin` and `show-pins` utilities? Apr 08 16:12:00 Yes Apr 08 16:13:44 > <@satacker:matrix.org> And firewall, to provide a seamless one click experience Apr 08 16:13:44 > Ssh within the app Apr 08 16:13:44 How are we planning to do this? Apr 08 16:14:56 Omg. I didn't give a thought about ssh. Apr 08 16:16:26 No worries, let's figure it out right now! Apr 08 16:17:14 go on, present your thoughts Apr 08 16:19:12 thank you , it opened my eyes Apr 08 16:19:23 libssh seems the way Apr 08 16:19:45 Overlays are not particularly associated with `bb-cape-overlays` repo satacker Dhruva Gole Apr 08 16:20:24 That's just the tool, what's the way? Apr 08 16:20:54 I'll have to go through its docs๐Ÿ˜… Apr 08 16:21:18 Path for all new overlays -> https://github.com/beagleboard/BeagleBoard-DeviceTrees/tree/v4.19.x-ti-overlays/src/arm/overlays Apr 08 16:21:25 jduchniewicz1: did you get a chance to look at the beagleconnect stuff? sufficiently confused yet? Apr 08 16:22:31 I see you glanced at it while waiting to hear back from ds2. Apr 08 16:22:49 not sure if you go the point of wanting to ask some questions. Apr 08 16:23:54 satacker: The idea is pretty good albeit now polished. I wanted to deliver a similar terminal gui app with my project last year for configuring the overlays but, time didn't allowed that. vedant16's project inspired me for that :) Apr 08 16:25:10 Any student present here who want some help with the Project Proposal? Apr 08 16:25:46 I will polish it as much as I can , and with the suggestions both you as well as Abhishek have stated, I am sure it would be lot better than it's current state. Apr 08 16:26:06 Pardon? Didn't get yoh Apr 08 16:26:12 * Pardon? Didn't get you Apr 08 16:26:40 Time ๐Ÿ•Š Apr 08 16:27:21 jduchniewicz1: I know it is last minute, but can you tell me more about your interests to see if something better suits if YOLO isn't working out? Apr 08 16:27:22 you were looking for the I2C overlay(s) right? I have answered you comment on Github also. Apr 08 16:27:28 lots of AI interest? Apr 08 16:27:45 Yes yes Apr 08 16:27:56 Saw that, visited the link also Apr 08 16:28:01 htt"> Dhruva Gole: Apr 08 16:28:13 Actually I was under the wring idea that overlays were incomplete Apr 08 16:28:39 Feel free to look at my proposal and suggest improvements: https://elinux.org/BeagleBoard/GSoC/2021_Proposal/simpPRU_Improvements Apr 08 16:28:46 Just couple days back I realized that issue was also in the userspace codes Apr 08 16:29:19 Ie. It lacked many libraries for bbai Apr 08 16:29:30 In my opinion the statement `Incomplete` can only be used for the LCD capes overlays ;) Apr 08 16:30:09 like what? Apr 08 16:30:16 on it! Apr 08 16:32:44 satacker: If you are first making changes in your GDoc please share it with me also -> `deepaklorkhatri7@gmail.com`. Apr 08 16:33:31 lorforlinux[m]: did you do overlays for the roboticscape then? Apr 08 16:33:34 * jkridner slept since then. Apr 08 16:33:39 I first change the page and then ctrl c+ V but I would add you too Apr 08 16:33:41 no he did not Apr 08 16:33:59 jkridner : he was unable to recieve the robot cape Apr 08 16:34:15 hence those overlays will have to be done by me Apr 08 16:34:17 lorforlinux[m]: this channel is publicly logged. :-( Apr 08 16:34:38 nslu2-log: is logging us and is crawled. Apr 08 16:34:52 I guess it is kinda subtle in the topic. Apr 08 16:35:11 might want to use caution sharing e-mail addresses, though I suppose they all get spammed anyway. Apr 08 16:35:45 giuliomoro[m]: any luck drawing in students for Bela projects? Apr 08 16:36:22 Nope, I didn't had the robotics-cape in hand. Apr 08 16:37:04 * jkridner feels we aren't fishing in the right waters somehow to get audio, Zephyr and Linux kernel students. AI seems to be more the topic-du-jour. Apr 08 16:37:19 lorforlinux[m]: ok, didn't think so. Apr 08 16:37:57 yeah, It's already everywhere on the internet. Apr 08 16:38:26 :-D Apr 08 16:42:54 jkridner: did you see the Email from Leah? Apr 08 16:43:07 no. I'll look. Apr 08 16:43:16 jkridner: it is written ``TODO: Use /dev/uioX to isolate permissions as well as abstract the offset`` [here](https://github.com/beagleboard/librobotcontrol/pull/183/commits/89d227cff70f6b287edaaea0d484771f313df650) this is all be done by me right? Apr 08 16:43:37 * jkridner: it is written `TODO: Use /dev/uioX to isolate permissions as well as abstract the offset` [here](https://github.com/beagleboard/librobotcontrol/pull/183/commits/89d227cff70f6b287edaaea0d484771f313df650) this is all to be done by me right? Apr 08 16:44:27 Archisman Dey: Proposal looks good overall, you can describe a bit more in detail about how you will update the grammar for the 'only calling break/continue inside loops ' and 'return statements can be called from anywhere in the function' milestones. Apr 08 16:47:24 There is one student `Leo` on the forum who is interested in the project! Apr 08 16:47:27 * jkridner wishes he pulled introductory kernel patch tasks from http://bbb.io/t for all new students. Apr 08 16:47:35 laimonade: Hi, did you start building upon the idea and make a draft proposal? Apr 08 16:47:36 Okay, I will add how it will look after the project for the break/continue inside loops, and how it worked before in the return statements in functions. Apr 08 16:48:15 I'm not sure how bad we scare students away with complexity and/or simply being vague, but I imagine if there were some more clear introductory exercises it would really help with the on-boarding. Apr 08 16:49:16 lorforlinux[m]: I saw that, but I'm not sure how well giuliomoro[m] has been doing at getting engaged with a student. Apr 08 16:49:59 there is some nice back-and-forth on the list though, which is a start. Apr 08 16:51:26 jkridner: I am planning to create an introductory course on BBB/BBAI at `Upside Down Labs`. If you are interested in the idea let me know, I would love to talk on the details with you :) Apr 08 16:52:22 Totally agreed! Apr 08 16:53:30 hey jkridner I will look a little bit more into the beagleconnect later today or tommorow (going afk). Could you just give some reasoning for micropython there? Ease of programming? Apr 08 16:54:31 jkridner: as for my interest and the general notion of AI being tempting :) -> mostly interested in using Rust in embedded projects or accelerating stuff on HW (be it GPU or FPGA) Apr 08 16:54:53 Archisman Dey: If you can search up what grammar you will write to make those changes, it will be helpful as well since there are a few days until the deadline Apr 08 16:54:55 jkridner: got some experience with ALSA and audio with FPGA Apr 08 16:55:29 I am eager to hear from ds2 regarding the GPGPU project as this may be something useful and quite rewarding Apr 08 16:55:59 I believe that's the case ๐Ÿ‘๏ธ Apr 08 16:56:05 Agreed, a lot a students are intimidated by the ideas list Apr 08 16:57:16 * jkridner[m]1 would like to empower anyone to give it an overhaul, perhaps moving it to forum.beagleboard.org/c/gsoc. Apr 08 16:58:13 I will pick-up on the messages later, feel free to ask anything jkridner Apr 08 16:58:22 Okay, I will try to do this too, thanks for the suggestion! Apr 08 16:58:22 I know it is too late for this year, but fixing it now still must be of some good. Apr 08 16:58:45 lorforlinux: please can you allow me to ssh into your bbai? (Kindly check your email inbox) Apr 08 16:58:48 jkridner: Shall try to think of some ideas Apr 08 16:59:41 jkridner: I was keen on contributing to the Micropython for BeagleConnect Freedom project as well,so I am putting an application for this project as well.I am going through the documentation part and repos given on the ideas page.Anything else you would like me to go through regarding the project? Apr 08 16:59:53 jduchniewicz: motivation for Micropython support on BeagleConnect Freedom is really to mimic the ease-of-use seen on other plaforms, like Circuit Playground Express. I *love* the experience with that board. Apr 08 17:00:24 But, about as important as ease-of-use is simply getting the Zephyr support for Freedom into the Zephyr mainline. Apr 08 17:00:25 I will brainstorm various ideas and discuss them with you tommorrow itself and will start making the proposal as well Apr 08 17:00:46 Adding Micropython support is something I think can be reasonably mainlined. Apr 08 17:00:55 @Students There are a few days until the deadline but please don't wait until the last moment to submit your proposals, there are always internet outages/emergencies and you don't want your lose your effort Apr 08 17:01:05 The Greybus stuff I think will be stuck being our own for a long while. Apr 08 17:01:22 ...and the Greybus stuff is simply harder to understand and the actions are less clear. Apr 08 17:01:45 Ideally, I'd have someone dive in and solve Bluetooth on Zephyr in the Greybus app! Apr 08 17:02:07 As a student here, the idea of working on the Linux kernel seems interesting to me too, but I have no experience with it so I did not look into the kernel ideas in too much detail, it would help if there was some kind of kernel based pre-application task ig. Just looked into the TechLab cape, seems cool! Apr 08 17:03:11 yeah, we need to move from a cross-compile "hello world" to a Linux kernel "hello world". Apr 08 17:03:18 I'll setup my BBBWl for you as, BBAI will just continue to reboot itself due to overheating :( Apr 08 17:03:44 sure, just email me the connection details Apr 08 17:05:30 jkridner: Thanks for your inputs will surely go through all the points you told me to go through. Apr 08 17:07:10 Yadnik Bendale: cool. Check out the status of CircuitPython on Zephyr too.... Apr 08 17:07:29 I'm really not sure what the best starting point is or what the latest status of the projects are. Apr 08 17:08:13 as I'd said, it is more an exercise of getting good support for the hardware into upstream Zephyr and exposing the functionality in an easy to use way. Apr 08 17:08:51 sorry my hours on here can be a bit random. Apr 08 17:09:01 not sure if anyone else has any Zephyr experience. Apr 08 17:10:09 No worries i'll keep putting my findings here Apr 08 17:10:41 We can discuss later on a mutual time Apr 08 17:13:58 Nice work Archisman Dey looks pretty good to me, please make the changes pratimu has asked for. Apr 08 17:15:27 Thanks, Abhishek has also said he will comment on it today, so I will incorporate both of their ideas by tomorrow! Apr 08 17:18:15 Okie ๐Ÿ‘๏ธ Apr 08 17:21:21 Do we have `Arooj Shahid` here? Apr 08 20:10:22 hi lpillsbury[m] thanks for joining the chat. lorforlinux[m] is here often as well. Apr 08 20:10:48 cool thanks for having me. Apr 08 20:11:07 lorforlinux[m]: absolutely excited about any introductory materials! Apr 08 20:12:35 so I think there's two issues here given limited time: 1) what things need to be done that other students haven't taken already and 2) what would be good for my skillset? and then maybe I dive deeper into the intersection of those 2 things and can ask questions while putting together a proposal? Apr 08 20:25:04 lpillsbury you can check the bela section of the Ideas page Apr 08 20:25:25 I believe most of the topics there are yet to be taken Apr 08 20:29:02 lpillsbury[m]: even if someone else has "taken" the idea, it is OK to make a proposal for the same idea. I'd say your focus should be on how you can make a solid proposal, which includes: engaging with the community/mentors/other-students, fleshing out milestones that are realistic and productive and otherwise demonstrating your ability to understand how you can engage and benefit the community with good quality software. Apr 08 20:29:45 we interact here in a public form with all the students such that we can all benefit from the clarification and answers. Apr 08 20:30:24 if anything makes you feel uncomfortable, please feel free to reach out to cwicks[m] or myself privately (we are the admins), but, otherwise, we like to keep things very open. Apr 08 20:30:49 ...even to the point of having students put their drafts up such that other students could potentially "copy" them. Apr 08 20:31:24 OK great. Apr 08 20:31:35 if things go right, there will be plenty of mentor interactions along the way (hopefully even in this short time-frame).... and we'll know who actually went through the effort to understand how they can contribute. Apr 08 20:32:30 Here's the idea page: https://elinux.org/BeagleBoard/GSoC/Ideas-2021 and don't forget to do the cross-compilation task! Apr 08 20:32:42 Thanks ArchismanDey[m] ! Apr 08 20:33:22 you'll find fellow students are often as or sometimes even more useful than mentors. :-D Apr 08 20:33:46 many former students are now around as mentors as well. Apr 08 20:33:59 * jkridner realizes there's no easy way to recognize. Apr 08 20:34:45 lpillsbury: have you finished the pre req. Task? Apr 08 20:34:49 all: please remember the ideas page is just that.... ideas.... proposals don't have to be on existing ideas. Apr 08 20:35:05 it is hard to come up to speed with Beagle stuff. Apr 08 20:35:40 for PRU stuff, I like to recommend people read beagleboard.org/prucookbook Apr 08 20:35:44 > * <@freenode_jkridner:matrix.org> realizes there's no easy way to recognize. Apr 08 20:35:44 There's a list of mentors on the ideas page Apr 08 20:36:23 lpillsbury: Make sure you meet the following ASAP first : Apr 08 20:36:23 https://github.com/jadonk/gsoc-application/tree/master/ExampleEntryJasonKridner Apr 08 20:36:48 librobotcontrol has beagleboard.org/librobotcontrol, but much of the progress made by lorforlinux[m] last year to abstract the controller boards using overlays and udev rules haven't become part of that documentation, so the challenges are not obvious. Apr 08 20:37:11 also, prucookbook doesn't have any background on simpru. Apr 08 20:37:22 (I probably even spelled it wrong) Apr 08 20:37:36 Thanks DhruvaGole[m] ! Apr 08 20:37:47 simppru* Apr 08 20:38:13 showing you can cross-compile is a good step to demonstrate you can work with github and ARM toolchains. Apr 08 20:38:38 last year we said we'd make compiling a kernel module a requirement, but I never firmed-up appropriate instructions. Apr 08 20:39:35 all: don't hesitate to spill over to #beagle for more information. there are some great experts there that don't idle here. getting more input is always good. Apr 08 20:46:00 Dhruva Gole: thanks for letting me know, I see it in the instructions and will do it shortly Apr 08 20:47:25 jkridner can you please merge my PR? Just small changes they are.. Apr 08 20:47:33 Librobot control Apr 08 20:47:35 #beagle never shows up if I search for it on the Element desktop client, I had to join using the link given in beagleboard.org/chat btw Apr 08 20:47:55 * ArchismanDey[m] uploaded an image: (284KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/ejSyKFIHDGVaUylibzTPxjPs/image.png > Apr 08 20:49:24 Archisman Dey: search in freenode Apr 08 20:49:50 Archisman Dey: Also I went through your proposal, very well written. I have a small nitpick Apr 08 20:49:55 * vedant16[m] < https://matrix.org/_matrix/media/r0/download/matrix.org/NZVsteHMBcGwVQUVvwgKyPmG/message.txt > Apr 08 20:49:59 still does not show up, I have selected all rooms anyway, but ig it's probably private Apr 08 20:50:13 * vedant16[m] < https://matrix.org/_matrix/media/r0/download/matrix.org/uLxhniaiXCpIZEtNHQHzLHdB/message.txt > Apr 08 20:51:54 https://matrix.to/#/!eyeOxukenDejXCKcKX:matrix.org?via=matrix.org&via=mailus.de&via=selea.se Apr 08 20:51:56 join Apr 08 20:52:04 * no need for this, this will add unnecessary complexity Apr 08 20:52:32 Also, there's a lot work to do and I don't see the benefit of this feature. Apr 08 20:53:04 <@vedant16:matrix.org> ```"> It seems useful to me though, but I can remove it if you think it will be too hard to do Apr 08 20:53:27 Can you give me some usecases ? Apr 08 20:53:50 hard as in, increases complexity, then you'll have to extend this to arrays too Apr 08 20:54:08 also, there's no way to notify undefined behaviour. Apr 08 20:54:10 I joined from the link in the beagleboard site, but thanks! Apr 08 20:55:53 After thinking again, I can't really think of any, I just added it because most programming languages have it so it will be more familiar for users ig Apr 08 20:56:04 Okay, I will remove this Apr 08 20:56:13 see, ๐Ÿ˜‚ Apr 08 20:56:56 I meant undefined behaviour literally - the compiler does not consider this case at all Apr 08 20:56:57 what can really help is, caps of assigning hex to ints Apr 08 20:57:49 Umm, check what happens if you assign 1000 to uint8_t Apr 08 20:58:52 and adding left/right operators to bitwise operations Apr 08 20:58:56 as in writing, int a := #F will work and a will become 16? Apr 08 20:59:09 > <@vedant16:matrix.org> what can really help is, caps of assigning hex to ints Apr 08 20:59:09 * as in writing, int a := #F will work and a will become 15? Apr 08 20:59:13 yes right Apr 08 20:59:17 0xF Apr 08 20:59:48 int a := 0xFFFFFFFF Apr 08 20:59:59 max range of 32-bit int Apr 08 21:01:24 Why this will help, write drivers for i2c devices Apr 08 21:01:34 easy to handle it with hex support Apr 08 21:01:37 This actually is well defined in C, I checked this while writing the proposal too: https://stackoverflow.com/questions/18294350/how-assignment-from-int-to-char-works-in-c, but I did not want to add too much complexity to simppru, but I guess it is better to just remove this. Apr 08 21:02:19 There's also no % in arithmetic operators I think? Apr 08 21:02:38 Yes, it is. Apr 08 21:03:12 as I said, it will increase work, and no significant benefit/usage Apr 08 21:03:38 Okay, seems useful, I will add this Apr 08 21:04:04 Yes, there isn't you can add that too Apr 08 21:04:29 Agreed, I will remove it, thanks for the suggestion! Apr 08 21:05:20 Okay, I will add <<, >>, and % operators for int and char. Apr 08 21:05:23 idt how fast you can code, but better to underquote a bit than overquote and not complete the work. Apr 08 21:05:50 If you checked my proposal, I did underquote. and ended up doing a lot more. Apr 08 21:05:55 >, and % o"> Cool :) Apr 08 21:07:06 I kept the last two weeks light and the one before that a bit vague, shouldn't that be enough? Apr 08 21:07:33 I am advising in general, not wrt to your proposal :) Apr 08 21:08:54 Oh, okay, I think I can manage to do whatever I wrote in my proposal Apr 08 21:09:08 Cool, then :) Apr 08 21:09:38 I guess you'll end up doing more than proposed, which will be good Apr 08 21:10:59 Let's see, and I will add/remove the things you suggested by tomorrow Apr 08 21:11:22 I'll add UART support once I am a bit free. Apr 08 21:11:45 Cool, also don't forget to submit on the platform Apr 08 21:12:09 CC: Abhishek Please go through the above discussion when you see this. Apr 08 21:12:28 Yes, I have made the GSoC account and will submit at least a day or two before. Apr 08 21:12:39 fine Apr 08 21:14:13 jkridner: https://www.ti.com/tool/PRUCAPE this looks like a good cape for BBB, we should give this to students working on PRU projects ? Apr 08 21:50:04 * StevenSchrstedt4 < https://matrix.org/_matrix/media/r0/download/matrix.org/kLYivJsLxUSkXYelztPDsrIE/message.txt > Apr 08 23:18:55 vedant16[m]: makes sense, yes. Apr 08 23:19:10 FWIW I have been experimenting with isar recently, which already has some risc-v support. Might be a good blend of Debian and custom OS for the BeagleV. Apr 08 23:19:46 moto-timo: oh? Apr 08 23:20:06 https://github.com/siemens/isar/blob/master/meta-isar/conf/machine/sifive-fu540.conf Apr 08 23:21:03 TL;DR isar uses debootstrap to build a base image (qemuamd64-buster took 6m34s on my ~10 year old dual Xeon E5-2630 box) Apr 08 23:21:25 but you can still build a custom kernel and patched or completely custom applications Apr 08 23:22:37 using kas as well gives the magic "manifest plus configuration" of a one-liner build Apr 08 23:22:57 https://github.com/siemens/kas Apr 08 23:23:19 wihtout all that annoying repotool phone home to Google every single time you launch Apr 08 23:32:11 that *is* a bit annoying... Apr 08 23:33:23 * nerdboy was just building oe-core version of qemuriscv64 yesterday... **** ENDING LOGGING AT Fri Apr 09 03:00:09 2021