**** BEGIN LOGGING AT Wed Mar 21 03:00:04 2018 Mar 21 03:23:12 m_w_: I looked at the expect-send functionality of c-kermit. That would be easy enough to implement. I could even extend it with regex pattern matching for the expect string Mar 21 03:24:21 Something like "SCRIPT UART1 " Mar 21 03:25:21 and "SCRIPT UART1 list", "SCRIPT UART1 del " to add and remove scripts Mar 21 03:29:21 m_w_: Also, I have some experience with PCB design and fabrication. I personally designed and routed half the boards we used for Hyperloop India's pod electronics. If you'd like, designing the cape can be another goal. What do you think? Mar 21 03:37:35 jkridner[m]: I believe i can finish those goals(blockly-bonescript) in my thread within the schedule as I have been working on blockly for a while , I have some doubts regarding solving the issues in bonescript: what should I do if some of the errors are not reproducible? Mar 21 03:42:18 jkridner[m]: yes, I have been in the field of web development(both dynamic sites and PWAs) for over two years now, I am comfortable working with JS, blockly and UI customization . Mar 21 04:01:56 * anujdeshpande[m] tips his hat and wishes everyone a wonderful Wednesday Mar 21 04:43:06 spiritus: do you already have a draft of the proposal ? Link ? Mar 21 05:35:08 muneeb17: Hey ! Read your proposal. Do you really think all this will take 12 weeks ? Mar 21 05:35:21 RPMsg support is already there .. Mar 21 05:35:30 and rproc is stable too Mar 21 05:35:51 the python files will ultimately need to write to the sysfs entries. thats it Mar 21 05:36:01 so .. why 12 weeks for that ? Mar 21 05:36:51 and what do you mean by `mmap() for some memory management` ? Mar 21 05:37:26 :s/python files/python scripts/ Mar 21 05:41:29 zeekhuge: I actually have 4-5 weeks planned for remoteproc/rpmsg. The other major phase will be to move examples to pru-as and documentation is also there. Also I plan on implementing userspace interrupts with RPMsg. Mar 21 05:43:16 what do you mean by "userspace interrupts" ? RPMsg framework can easily send messages to PRUs and the PRUs will be waiting for them to arrive, and can do programmed tasks for different msgs. Is that what you mean by "userspace interrupts " ? Mar 21 05:43:44 Also, you have included RPMsg in "strech goals" Mar 21 05:44:07 Yeah, I'll include that in the main timeline now Mar 21 05:44:11 do you have experience in python ? can you do simple file read and write ? If yes, it should not take more than 2 weeks. Mar 21 05:45:10 or .. better question is . why should it take more than 2 weeks ? Mar 21 05:46:07 memory functions are also there.... the existing API uses prussdrv mem functions. Mar 21 05:47:12 yeah. but why porting it to remoteproc and RPMsg will take 12 weeks ? Mar 21 05:47:41 which would mostly involve changing python scripts to just read/write to sysfs entries ? Mar 21 05:48:11 No, I'll use the rproc and rpmsg drivers to implement the API Mar 21 05:48:14 sysfs entries = corresponding to the remoteproc and RPMsg framework. Mar 21 05:48:26 and how will you "use" them ? Mar 21 05:49:34 those are kernel drivers, right? Can't we use those functions in the userspace Mar 21 05:50:31 that is what sysfs entries are. And they are almost like simple files, which you read from, or write to. And all that already exists for RPMsg, as well as Remoteproc Mar 21 05:51:03 I thought sysfs entries will be slower Mar 21 05:51:23 porting PyPruss will only be writting to these sysfs "files". Mar 21 05:51:50 what other way do you have ? Mar 21 05:54:08 What do you suggest now? I planned on using the drivers directly from the beginning. I dont have any other way. Mar 21 05:54:49 As far as i know, that is probably the only way to interact with the kernelspace. And sysfs entries will be as fast as the cpu can handle them to be. Mar 21 05:58:40 zeekhuge: Thanks for your input. What would you suggest I do now? I am seriously confused Mar 21 06:06:01 abhishek_: Can you give your input on this discussion ^ Mar 21 06:10:41 illustris: I think the timeline need to be more detailed. Tell us about what all things (maybe different frameworks ) you will use to implement the week's task, or just how about you will go implementing it .. etc. Mar 21 06:11:16 Okay. I'll elaborate on the points Mar 21 06:11:27 Timeline is arguably the most imp part of the proposal. Should give a clear idea if you really know what you have to do. Mar 21 06:12:02 illustris: Also, is the 2nd week's task not already done ? Mar 21 06:12:18 I mean .. there should be some examples related to that . Mar 21 06:12:57 Its preferred if we are able to use already-existing code. Mar 21 06:13:53 Sure, there's code to read from serial and write to file. But What I want to implement is to do that and cat the logs without any buffering. In every example I found, the log wouldn't update till a newline or the buffer got filled Mar 21 06:14:55 so if you press a backspace in the interactive shell, the corresponding escape sequence gets written to log, and mirrored on any other non-interactive sessions live Mar 21 06:16:09 But you're right, it shouldn't take me a week. I'll adjust the timeline Mar 21 06:17:07 And mention that too (the details) .. in the timeline .. Mar 21 06:17:28 will do. Thanks! Mar 21 06:22:55 hjhee_: Hey ! The familarizing the PRUs part is a prerequisite. their is enough examples out there that you can try and questions you can ask here, to know the working of rproc and PRUs. Mar 21 06:23:24 And you are dedicating 2nd-6th week for it Mar 21 06:24:20 And I think PTP implementation and testing is what will require more time .. Mar 21 06:26:52 Also, if you are planning to use DMAs (not-RPMsg, which might not suffice the bandwidth beaglelogic needs) I think the deliverables to be more generic and not beaglelogic-specific. So that others are able to utilize it as well. Mar 21 06:28:07 you may then add the task of attaching this "generic-synchronous data" project to beaglelogic, in your timeline. Mar 21 06:35:02 Pac23: There was atleast some work already done on this project last year. right ? Mar 21 06:41:33 hjhee_: Also, in the "suggestions" section of your proposal, you are saying "Additionally I think the project should also consider utilizing the EDMA controller for data transmission", but your timeline already includes to use the DMA. You realize that both of them are same actually ? Mar 21 06:48:37 Pac23: recommend digging into the previous year repo in this time - building it, looking for pieces you can run now lacking the full set of hardware, be able to discuss in concrete terms which parts you expect to need extended Mar 21 06:49:24 When continuing on a project, getting your head into the existing codebase is really crucial for success Mar 21 06:49:46 otherwise you can end up rewriting most of it, and there goes the schedule for adding new features Mar 21 06:51:14 sdabhi23: The timeline needs work Mar 21 06:52:27 From all the other links (github, your blogs) it looks like you have experience in android and web dev. there is no indication that you have experience in embedded systems. And then ... the timeline is so short, that .. it tells us (as mentors) nothing to decide on. Mar 21 06:53:35 So, timeline and project implementation part needs work. They should convince us that you can pull it off. Mar 21 07:35:52 zeekhuge: Wormo: I made the timeline more detailed. Please take a look Mar 21 08:55:45 spiritus: should probably see channle-logs. Mar 21 09:09:20 illustris: getting kind of late here, will try to get you feedback tomorrow Mar 21 09:09:44 Wormo: Cool. Thanks! Mar 21 09:11:58 zeekhuge: Thanks! Mar 21 10:42:57 hjhee: should probably see channel logs Mar 21 10:57:04 Hey, is the primary objective of cubesat project integration with beaglepilot/mavlink? Mar 21 14:46:32 @zeekhuge yes in 2016 and 2017 there was work done on the project,although as of now only the 2D axis anemometer is ready everything else is still to be worked on Mar 21 14:47:20 @wormo yes i have seen the previous years repo,i will write more about it in my proposal and update you then,thank you for the feedback Mar 21 15:16:06 jkridner: please review my proposal https://elinux.org/BeagleBoard/GSoC/etcher-beagleboot Mar 21 15:17:29 jkridner: in your elc presentation, you mentioned about crash while tftp when NETCONSOLE is on, I tried tftp from node console (netconsole), I didn't get any crash Mar 21 15:19:20 what did you mean by u-boot crash actually? Mar 21 15:46:08 meeting in 15 minutes Mar 21 15:58:37 hi cwicks! Mar 21 15:58:47 so, https://bbb.io/gsocmeet still doesn't work? Mar 21 16:00:10 seems like it only gives the time... and gives the wrong date, ie., not today, ie., now! Mar 21 16:00:12 :-( Mar 21 16:00:38 ravikp7: I have some questions about Etcher and usbboot, but I'd like to do this meeting first. Mar 21 16:00:55 pmezydlo: thank you for joining! Mar 21 16:01:19 hi Jason, hi all Mar 21 16:01:42 so, who all is here in a mentoring capacity? Mar 21 16:01:43 ds2? Mar 21 16:01:51 hello Mar 21 16:02:06 here Mar 21 16:02:17 I can take one student Mar 21 16:03:37 _av500_: ? Mar 21 16:03:44 _av500_, you going to be the ring leader again this year? Mar 21 16:03:54 thanks m_w_ and ds2 for being here. Mar 21 16:04:05 Anyone seen Wormo? Mar 21 16:04:09 nerdboy? Mar 21 16:04:39 Abhishek__: ? Mar 21 16:04:40 that google calender thing did not work well Mar 21 16:04:40 o12 hello everyone - happy to help admin this year Mar 21 16:05:10 I can be a mentor, not sure if main mentor but co-mentor for sure Mar 21 16:05:15 an ICS attachment would probally be more useful Mar 21 16:05:59 i can help, no time for primary mentoring right now... Mar 21 16:06:00 if you can just make a .ICS, I can just point the URL to it. I'm struggling even making one... Mar 21 16:06:02 I guess I can figure it out. Mar 21 16:06:20 Reminder student applications need to be in by 3/27 Mar 21 16:06:26 'k Mar 21 16:06:28 ping stephanie harder... Mar 21 16:06:57 hi nerdboy ! Mar 21 16:07:02 what I can take on depends on the project Mar 21 16:07:22 zeekhuge, you want to mentor again this year? Mar 21 16:07:31 nerdboy: Stephanie is Wormo, yes? Mar 21 16:08:05 * jkridner[m] kicks summerofcode.withgoogle.com Mar 21 16:08:37 think wormo == stephanie == person at e-ale (at SCaLe at least) Mar 21 16:08:47 Hey ! Mar 21 16:08:52 jkridner: Hi I'm here Mar 21 16:09:04 mw: I think yes. Mar 21 16:09:32 abhishek_[m], you have time to mentor this year? Mar 21 16:09:42 zeekhuge, awesome Mar 21 16:09:46 I'll be interested in taking up PyPRUSS and related PRU projects Mar 21 16:10:05 moar mentors! Mar 21 16:10:49 Can also help with shaping up of node-beagle-boot if it's taken up this year. Mar 21 16:11:45 I have these mentors signed up: ds2 , m_w_ , zeekhuge , abhishek_ , rma Mar 21 16:11:50 anyone seen rma? Mar 21 16:12:00 nope Mar 21 16:12:16 any other mentors that aren't yet invited? Mar 21 16:12:21 pmezydlo: have you seen rma? Mar 21 16:12:38 no Mar 21 16:13:04 but I can be a mentor Mar 21 16:13:11 nerdboy: I think my invite is still outstanding for you. Mar 21 16:13:58 was hoping to see anujdeshpande too. Mar 21 16:14:26 pmezydlo: can you add yourself to the bottom of bbb.io/gsocides ? Mar 21 16:14:31 bbb.io/gsocideas even. Mar 21 16:15:04 also, PM me the e-mail address where you want the invite. Mar 21 16:15:11 are you looking to do more audio things, PRU, ??? Mar 21 16:15:32 seems we have some real critical mass on PRU mentorship and not so many PRU proposals. Mar 21 16:15:58 as seen on the proposals list or elsewhere? Mar 21 16:16:09 No good one at least. Mar 21 16:16:14 As of now. Mar 21 16:16:39 ah...well, as of now, there isn't anything in my area Mar 21 16:16:41 just based on interactions in this channel and draft proposals up on eLinux. Mar 21 16:17:05 I hope students know that eLinux is the best way to get feedback friendly to most mentors ahead of the proposal deadline next week. Mar 21 16:17:59 ds2: weren't you looking for PRU-to-kernel projects, ie., serial, I2C, SPI, etc. examples where PRU is acting as the host controller and providing a host controller interface to the kernel? Mar 21 16:18:19 jkridner[m]: yes... didn't see much of that Mar 21 16:18:43 did I miss something on the elinux side? see passing talk of it Mar 21 16:18:49 bbb.io/gsocides is it works? I have a problem with joining Mar 21 16:18:52 hopfully some prospective students are monitoring our discussion and seeing where there are supporting mentors without students making attractive proposals. Mar 21 16:19:25 or the GPGPU or a convincing offline smart speaker Mar 21 16:19:27 https://elinux.org/Category:GSoCProposal <-- good way to find proposals. Mar 21 16:19:34 i don't really do JS :) Mar 21 16:19:50 if people haven't tagged with that category, they might not be discovered as easy. Mar 21 16:20:16 unfortunately, the category has been used in other years, so you have to pay attention. Mar 21 16:20:35 i was hoping for folks to have started something on sommerofcode Mar 21 16:20:46 * jkridner[m] is happy to see when people use https://elinux.org/BeagleBoard/GSoC/ProposalTemplate Mar 21 16:21:04 however, someone seems to have defaced it a bit. :-) Mar 21 16:22:06 Any useful edits in https://elinux.org/index.php?title=BeagleBoard%2FGSoC%2FProposalTemplate&type=revision&diff=462851&oldid=438026 ? Mar 21 16:23:38 * jkridner[m] reverted. Mar 21 16:23:46 ds2: easier for you to review there? Mar 21 16:24:26 jkridner[m}: problem is finding 2018 stuff Mar 21 16:24:30 is it good to encourage students to do both? Mar 21 16:24:55 yes.... esp. giving the time line...google doesn't recogonize anything else anyways Mar 21 16:25:11 i see 3 relevant proposals in the weeds Mar 21 16:26:03 k. I'll try to give some individual pings. Mar 21 16:26:06 are we suppose to use Discussions on elinux to comment? Mar 21 16:26:11 any mentors un-pinged? Mar 21 16:26:20 can we get all potential mentors to review https://bbb.io/gsocml ? Mar 21 16:27:22 even if it is just to tell the students, "hey, this doesn't seem useful" Mar 21 16:27:28 Jic23? Mar 21 16:27:31 that is the same as beagle-gsoc@ right? Mar 21 16:28:27 you currently have 9 proposals in the app - 1 in "final" 8 in "draft" Mar 21 16:29:26 And one of them is a g-doc. Mar 21 16:30:16 jic23 doesn't seem to be very active... probably could just consult on stuck issues. Mar 21 16:30:35 ds2: yes, just one of my crazy short-links. Mar 21 16:31:07 zeekhuge: :-( not sure if everyone feels the same as me that opening all those G-Docs is a pain. Mar 21 16:31:34 having 30 eLinux tabs doesn't seem to kill my machine as much. Mar 21 16:31:40 jkridner: Especially on Chrome? Mar 21 16:32:05 a bigger problem with G-Docs is having to check your credentials before hand Mar 21 16:34:05 so, my main agenda items walking in was to make sure any mentors were invited and see where we were at in a mentor interest level... Mar 21 16:34:41 any other mentors want to speak up about what interests them? I may post some logs to try to encourage potential students to read. Mar 21 16:34:59 abhishek_: +1, yup, esp on Chrome/Chromium. Mar 21 16:35:28 I added myself to mentors list Mar 21 16:35:42 pmezydlo: thanks. Mar 21 16:37:28 pmezydlo: I sent a mentoring invite to your gmail address Mar 21 16:37:52 I see it thx Mar 21 16:40:20 pmezydlo: m_w_ abhishek_ zeekhuge any comments on the project types that interest you? Mar 21 16:40:43 nerdboy: ^^^ same Mar 21 16:40:59 nerdboy: trying to do more with the anemometer? Mar 21 16:41:53 beng-nl: you have your own GSoC this year? Mar 21 16:42:55 I'd be interested in PRU-related projects - PyPRUSS and the like. Also would be interested to help with node-beagle-boot future prospects. Mar 21 16:44:04 nerdboy: you looking at these terminal server things? Mar 21 16:44:17 Yeah .. PRU and kernel maybe. Mar 21 16:44:17 abhishek_: thanks for reminding us. Mar 21 16:44:33 zeekhuge: are there any notable kernel project ideas or proposals? Mar 21 16:44:54 Not really. Mar 21 16:45:23 wow... we are very lopesided this year Mar 21 16:45:34 m_w_: you looking at this terminal server stuff? right direction to encourage prospective students? Mar 21 16:46:05 just didn't put in any good kernel ideas into our ideas list. :-( Mar 21 16:46:27 jkridner[m]: unfortunately not; neither student nor org Mar 21 16:46:27 and our description probably wasn't kernel-heavy enough. Mar 21 16:46:43 having e-ale materials would really be helpful for new kernel students too. Mar 21 16:47:14 beng-nl: oh no! well, you are always welcome to engage here! :-) Mar 21 16:47:50 jkridner[m], I would mind doing the serial terminal stuff if the right proposal gets sent Mar 21 16:47:53 some resources in other areas? Mar 21 16:48:00 illustris: thanks for doing both an elinux draft and a summerofcode.withgoogle.com draft Mar 21 16:48:39 FPGAs, mainlining PocketBeagle kernel Mar 21 16:50:16 Pac23xyz: I see your elinux entry. do you have a summerofcode.withgoogle.com draft as well? Mar 21 16:51:24 m_w_: can you float a description of a mainlining project to the top of bbb.io/gsocideas ? Mar 21 16:51:46 seems it would get the most attention from mentors... Mar 21 16:51:54 just getting basic PocketBeagle support in mainline seems to basic... Mar 21 16:52:11 PRU remoteproc might still be a gap.... Mar 21 16:52:18 not sure what else is a gap in mainline. Perhaps eCAP/eQEP? Mar 21 16:53:22 let's feel very free to redirect interested students who seem easy to engage. Mar 21 16:53:40 mainlining is an ongoing process Mar 21 16:53:44 m_w_: i thought you liked the terminal server idea... Mar 21 16:53:51 I do Mar 21 16:53:54 kernel patches can seem intimidating, but don't have to be so hard. Mar 21 16:54:23 are they any good? if i you wanna take one i can "help" Mar 21 16:54:25 jkridner: what gap in the remoteproc ? Mar 21 16:54:28 for me Linux kernel driver scope, fpga project or maybe audio project, will be interesting, but I dont have any ready proposition Mar 21 16:55:18 any mentor (potential) or otherwise, that is strong in OpenGL? Mar 21 16:55:36 more beagle-wire stuff is good in my eyes. something with some kernel drivers to go with peripherals. Mar 21 16:55:45 i think eQEP is in mainline now? Mar 21 16:56:12 or is that different than the patches from last year? Mar 21 16:57:33 * jkridner[m] would love more https://github.com/beagleboard/bb.org-overlays for mikroE click boards... Mar 21 16:58:10 that's a bit of a kernel project, at least for devices that don't have drivers already... Mar 21 16:58:40 for example, ESP8266 Click support is easy if you are just doing serial commands, but no good solution exists for PPP to Linux so we can use a whole Linux TCP/IP stack. Mar 21 16:58:57 nerdboy: is eQEP mainline now? I don't keep track. I need to track to closure better. :-( Mar 21 16:59:00 sorry for being intermittent; on a red hot PCB project right now Mar 21 16:59:25 hmm? Mar 21 16:59:31 * nerdboy stuck in a startup-y push to first demo release... Mar 21 16:59:37 you mean no ESP8266 PPP firmware? Mar 21 17:00:13 yeah. Mar 21 17:00:15 maybe it is there. Mar 21 17:00:26 no idea where stephanie is today... Mar 21 17:00:52 there should be no new kernel stuff for PPP. the modem stuff with pppd should be all that's needed on Linux Mar 21 17:01:56 if someone wants to support these semismart Wifi modules, i have started the linux side (SOCKS based) Mar 21 17:02:15 it mostly works against the SiLabs one already Mar 21 17:02:42 mentors... ok, that's the end of our hour. Can you all be here next week? Proposals should be in. Summary of actions: mentors to make sure the top few entries in the ideas page are actually viable for them to mentor, mentors to comment on the mailing list to direct students to those projects, I'm to contact students with eLinux draft proposals and encourage them to also post to summerofcode.withgoogle.com, any Mar 21 17:02:43 actions missing? Mar 21 17:02:57 I might be on the road next week Mar 21 17:04:16 ds2: yeah, bad kernel example since it really is just firmware on the ESP8266. but there could be many software gaps for various Clicks. Here's how I fixed the OLED C: https://gist.github.com/jadonk/837bca668111a8f39bf06e1a2c274158 Mar 21 17:04:22 I still mean to make a blog post out of the above. Mar 21 17:05:34 ds2: k about being out next week. thanks if you can take on the other mentor actions (ideas list, encourage students to gravitate towards kernel/etc. projects) Mar 21 17:06:25 ds2: (aside) are their mikroE Click boards for those WiFi modules? Mar 21 17:07:05 jkridner[m]: not from SiLabs but don't see why not...they can be USB/SPI/UART interfaced Mar 21 17:07:22 anujdeshpande: m_w_ nerdboy pmezydlo ds2 cwicks zeekhuge anything else before I close the official meeting? Mar 21 17:07:47 jkridner[m]: who are the admins this year? i know you and cwicks. anyone else? Mar 21 17:08:00 abhishek_: ^^ Mar 21 17:08:30 Gary Beaver... need to work on getting him here. Vlad... also not engaged. Mar 21 17:08:39 Nothing that I can remember right now. Mar 21 17:08:39 I might still be able to add one if someone wants to help with those duties. Mar 21 17:08:51 I am looking to hire an intern for the summer to help with some such tasks. Mar 21 17:09:05 (local intern) Mar 21 17:09:22 face-to-face interview tomorrow, so fingers crossed Mar 21 17:10:09 _av500_: we miss you! Mar 21 17:15:24 jkridner: I'd like to be considered for attending this year's GSoC mentor summit. Mar 21 17:15:37 starting to lobby early? :-D Mar 21 17:16:16 help us get this year's project proposals to be awesome! :-D Mar 21 17:16:30 more mentor-to-student engagements through more attractive proposals. Mar 21 17:45:54 jkridner: I don't completely understand the usbboot framework in etcher, I'd have to have a chat with etcher developers to finalise what fits best for our use case. If you have some questions, I'll try to answer that to the best of my knowledge Mar 21 17:52:04 meanwhile this can be helpful https://github.com/resin-io/etcher/issues/1541 Mar 21 17:52:11 https://github.com/resin-io/etcher/pull/1686 Mar 21 19:42:48 Hmmm... does https://beagleboard.org/gsoc/meeting.ics work for anyone? Mar 21 19:43:15 using mimetype of text/calendar, but still not loading as calendar. Mar 21 19:55:03 muneeb17_: How many different languages do you know ? How about creating c/c++ libs, python mods and RUST and also nodejs bindings. Mar 21 19:56:15 Also, understand that being able to work with PRUs using Rproc and RPMsg is a pre-requisite. There is enough out there to learn it. Mar 21 20:06:23 Once you understand it, using them from different languages is just about reading from and writing to the correct sysfs. **** ENDING LOGGING AT Thu Mar 22 03:00:02 2018