**** BEGIN LOGGING AT Sun Apr 02 03:00:02 2017 Apr 02 08:40:17 Hello jkridner> Apr 02 08:40:19 ds2> Apr 02 08:47:40 join Apr 02 08:48:26 hello Apr 02 09:23:33 Can anyone plz explain what this means "the project will involve developing drivers to reserver the device on behalf of the PRUs" in the project "PRUs to offload processing of raw data from on - board peripherals" Apr 02 09:40:42 help Apr 02 10:09:47 Akki: Hi ! please see how others have created their proposal on the elinux page. GSoC requires the ability to figure out things yourself. Apr 02 10:12:52 ok i will figure that out but would you plz answer my previous question Apr 02 10:41:27 jkridner: Hi! I've changed the binaries as perhaps I shared the wrong binaries. Please test the updated binaries from the same drive folder that I shared earlier. I've also included its link in the "experience and approach" section of the proposal. I'll be submitting the final proposal at gsoc site later today. So, please give your final thoughts on the Apr 02 10:41:28 updated proposal, how's the timeline? Apr 02 11:17:58 Akki: so, you realize the PRUs are remote-processors and not the main processing unit. Apr 02 11:18:15 Hmm Apr 02 11:18:24 and all the peripherals are controlled by drivers on the main processor Apr 02 11:18:43 so, when the PRU wants to access a peripheral Apr 02 11:19:05 It might be off, since the MPU (main processing unit) do not knows about it. Apr 02 11:19:33 the MPU do not know that the PRU wants to access that peripheral. Apr 02 11:20:20 but then, a peripheral can be switched on by loading the concerned DTO file. or adding the entry into the dts anyway .. Apr 02 11:20:54 but again, if for some reason, the MPU too wants to access that peripheral at the same time the PRU is using it . then ? Apr 02 11:21:13 that is what you have to find a solution for ... Apr 02 11:21:42 there are many other cases where resources are shared between different proceesors. Apr 02 11:22:12 maybe you can ask on the device-tree channel on how to add the hwmods in such a situation .. Apr 02 11:22:42 also, maybe you could see into the remoteproc_core.c Apr 02 11:23:28 http://lxr.free-electrons.com/source/drivers/remoteproc/remoteproc_core.c#L511 Apr 02 11:24:28 getting something.. Apr 02 11:24:51 any quick queries ? Apr 02 11:25:01 yes Apr 02 11:25:19 please go ahead .... Apr 02 11:25:26 lets say i want to take the data of say eCAP module which is not in the local memory of PRU Apr 02 11:25:42 so i have to access global memory via L3 connect Apr 02 11:26:27 so it will take more time so is it advisable to use this technique to fetch data of peripherals Apr 02 11:26:37 ? Apr 02 11:27:24 and also i am trying to create a page on elinux but seems like account verification is taking too long Apr 02 11:27:48 any other idea how you might be able to do that then ? Apr 02 11:28:29 L3 connect definitely makes it loose its real-time nature. Apr 02 11:28:33 not till now this is what i m able to understand as of now .. Apr 02 11:29:36 then thats the only way .. Apr 02 11:30:26 peace ! Apr 02 11:30:55 ok .. if am getting this right i will have to make C library to access the correct registers from memory and do some calculations over it (if required) Apr 02 11:39:16 first goal is to be able to do the library thing for max number of peripherals. Apr 02 11:39:30 then if time allows, examples on how to use them. Apr 02 11:39:50 and when to schedule drivers? Apr 02 11:40:12 actually, atleast one example for each should be there ... that should be in the schedule. Apr 02 11:40:26 that too, along the project. Apr 02 11:40:38 ok Apr 02 11:41:02 but you need to get the get and idea of how much time that will take .. Apr 02 11:41:17 *get the idea of Apr 02 11:41:57 Hi! Apr 02 11:41:57 ok i will try my best to do what i can in the time remaining but do i need to create an account for making an elinux page Apr 02 11:42:48 How can i find quotes here about my project Apr 02 11:43:09 Akki: I think so .. Apr 02 11:54:30 How can i find quotes here about my project Apr 02 12:04:40 pointhex: present your project and proposal to the mentors through IRC or mailing list and request them to provide quotes. Apr 02 12:35:23 hi everybody Apr 02 12:36:23 nerdboy: I have created the final proposal, and also created a workflow of the data acquistion Apr 02 12:37:30 in that workflow I have tried to explain process till data transferred to user space Apr 02 12:38:18 http://elinux.org/GSoC2017_sonic_anemometer_proposal#Requirement_Analysis Apr 02 12:38:32 http://elinux.org/GSoC2017_sonic_anemometer_proposal#WorkFlow_for_Data_Acquistion Apr 02 12:42:06 thetransformer : can you help me regarding creating an elinux page. I have requested an account but it's been an hour or 2 that i haven't received a mail with a password. Any suggestions? Apr 02 12:42:19 hi zeekhuge:,ds2: would you please check out my proposal and express your suggestions Apr 02 12:42:49 Akki:I think for account, you may have to wait few more hours Apr 02 12:44:28 and once you get an account, to create a elinux page, we have to search a topic that isn't available on elinux wiki, so for ex you can search a term like"GSoc-2017 project-proposal " Apr 02 12:44:59 and then if that is not available as a page on wiki, you should get a link to create a new page Apr 02 12:51:45 thetransformerr: how are you going to transfer data to the MPU ? Apr 02 12:52:39 what is the frequency of I/O you need ? Apr 02 12:55:26 for PRU to MPU, I still have to figure out a concrete solution, although I have thought of using low level libraries like zeromq Apr 02 12:55:52 you can suggest me a possible solution Apr 02 12:56:33 will there be a GUI or something ? to get the analyzed data ? Apr 02 12:56:50 that is what the 2nd PRU used for Apr 02 12:57:05 in most of the applications like beaglelogic or beaglescope Apr 02 12:57:16 frequency for I/O to MPU is going to very low If we can detect the transmission or receiving of wave in PRU only, is it possible??? Apr 02 12:57:42 ofc that is possible .. Apr 02 12:58:04 but the analyzed data should be available to the MPU Apr 02 12:58:20 GUI is planned during later stage, currently we have planned to output readings in text file sort of arranegment Apr 02 13:00:01 analysis in PRu is just limited to detect receiving of wave, I am thinking of sending out the time of flight from PRU to MPU and other calculation is intended to be done in user sapce Apr 02 13:00:57 whats the frequency of I/O on the PRU side ? like to your transducer ? Apr 02 13:01:32 If you have enough time to do calculations on the PRU itself, you can Apr 02 13:02:22 I have decided 400 kHz sampling rate for 40kHz wave Apr 02 13:03:24 that gives you how many nano seconds of gap ? Apr 02 13:03:44 I mean just try to figure out .. Apr 02 13:03:53 anyway .. maybe a task for later time .. Apr 02 13:04:10 but in any case you will have to transfer data to the MPU Apr 02 13:04:23 I think we have ample time because with a target of 20Hz result output, we have got 50 milliseconds for calculation of one result set and that means 50000 microseconds, and out of it sound wave just takes 800 microseconds for one side, Apr 02 13:05:18 "but in any case you will have to transfer data to the MPU" would you please specify what does term "data" encompasses here in itself Apr 02 13:06:26 if you do calculation on the PRU, then the data= final result. if you wish to do calculation on the MPU then data = raw time of flight . Apr 02 13:06:55 otherwise how would a user be able to see the wind speed or any other analyses ? Apr 02 13:07:13 "that gives you how many nano seconds of gap ?" I think I have calculated this with 40kHz we get a sample in time duration of 2 microseconds and PRU have one time cycle of 5ns Apr 02 13:08:49 yeah thats true, this are actually finer details, about which I can't claim anything surely at this point without a practical experience, but I think if we do calculation in user space, we can involve other parameters also like humidity et. Apr 02 13:09:55 and also this project is originally intended to be 3D axis, so I think loading PRU with too much of calculations might hamper our scalability factor Apr 02 13:11:34 ohkay ! so are you going to use the PRUDAQ with beaglelogic image directly ? Apr 02 13:12:18 that also I haven't decided for sure, but that's what I am planning to do in first attempt Apr 02 13:12:33 then maybe you wont have to worry about data transfer to MPU then. Apr 02 13:14:38 are you talking about these transducers - http://www.amazon.in/HC-SR04-Ultrasonic-Distance-Measuring-Arduino/dp/B01LXFUAFV?tag=googinhydr18418-21&tag=googinkenshoo-21&ascsubtag=0c4bf88e-4544-4bbe-934b-030e68cad056 ? Apr 02 13:17:44 yeah they could be a good choice, but they are integrated on same board, so either we have to change structure of anemometer from figure that I have shown in proposal in timeline section or we can't use them Apr 02 13:18:45 I haven't zeroed on the exact model I am going to use, but thanks for pointing me out this Apr 02 13:21:01 that can eliminate the need of using PRUDAQ or and ADC module and will give you the time-of-flight .. not sure about accuracy .. Apr 02 13:37:43 you are actually doing a gross injustice with wind speed by comparing it to distance measurement 😱 , wind is hell lot of dynamic than distance and add up to that the outdoor environment conditions like snow et. Apr 02 13:38:57 I have thought about something like these http://uk.farnell.com/multicomp-ultrasonic-transducers Apr 02 13:48:52 oh..kay .. I am really sorry for the injustice that I did :P ;) Apr 02 13:50:39 anyway .. I have no experience of such application but since those sensors are based on the same principle, I thought maybe you could use it. Apr 02 13:53:23 😀 thanks for reviewing proposal,I guess I would require your help with PRU Apr 02 14:02:12 ds2? Apr 02 14:29:13 ds2, jkridner, nerdboy: please review my proposal http://elinux.org/BeagleBoard/beaglecv_stereo Apr 02 15:24:57 thanks for your responses .. Apr 02 15:25:15 here is my proposal .. please review it Apr 02 15:25:21 https://docs.google.com/document/d/1HpZhU7EGvouX0ep-oEJzWszGwZrbBoKuQHaAg6MKY3o/edit?usp=sharing Apr 02 15:38:24 hey kiran4399 Apr 02 16:28:19 Hi NP95 Apr 02 16:31:28 thanks for your responses . here is my proposal .. please review it Apr 02 16:31:45 https://docs.google.com/document/d/1HpZhU7EGvouX0ep-oEJzWszGwZrbBoKuQHaAg6MKY3o/edit?usp=sharing Apr 02 17:24:26 hey m_w I added my proposal to the GSoC site. Is it ok to include the head section like on the eLinux proposal? Apr 02 17:29:12 kartik__: i thought that was supposed to be on the wiki for submission? Apr 02 17:29:46 looks pretty good, maybe a bit more detail in the timeline Apr 02 17:37:39 kiran4399: also looks good, do you have mentor(s) lined up? Apr 02 17:37:52 nerdboy: ds2, m_w Apr 02 17:44:41 ohkae .. ill expand the timeline .. apart from that is everything okay ? Apr 02 17:44:48 Apr 02 17:45:09 nerdboy: do you have anything to say on this? Apr 02 17:45:16 m_w: there: Apr 02 17:45:26 m_w: can you please look at my proposal? Apr 02 17:47:14 i think breaking out some task detail is the main thing Apr 02 17:47:25 wiki submission? Apr 02 17:47:36 oh ohkay.. ill do that now Apr 02 17:48:11 nerdboy: can you elaborate more? Apr 02 17:50:53 bullet 6. on the gsoc page Apr 02 17:51:22 http://elinux.org/BeagleBoard/GSoC/Ideas ^^ under general reqs Apr 02 17:55:28 nerdboy: can you please tell me which task detail you want me to break? Apr 02 17:57:15 i mean, eg, make a bullet list of things to do in those 2-week periods Apr 02 17:58:08 and that was for kartik__ (the task details) Apr 02 18:02:02 kiran4399: i can't think of much else for you, except maybe relate it more to the "GPU offload" idea? Apr 02 18:02:30 nerdboy: yeah.. thanks.. I will do it. Apr 02 18:03:03 as long as you have a clean interface for the opengl part Apr 02 18:04:27 then you could have a piece to "provide sample programs to show using GLES2 for computations" for stereo vision? Apr 02 18:05:00 nerdboy: yeah.. that's exactly what I was thinking.. Apr 02 18:09:07 nerdboy: please checkout this section in my proposal , it is sort of flowchart you asked for,http://elinux.org/GSoC2017_sonic_anemometer_proposal#WorkFlow_for_Data_Acquistion Apr 02 18:10:11 though I have still not defined concretely the way to transfer data from PRU , currently I am hoping for beaglelogic Apr 02 18:23:26 i think edma interface is the "official" TI way... Apr 02 18:24:31 *between PRU and main memory Apr 02 18:25:03 https://docs.google.com/document/d/1HpZhU7EGvouX0ep-oEJzWszGwZrbBoKuQHaAg6MKY3o/edit# Apr 02 18:25:24 did you look at prudaq branch from last year? including the wiki? Apr 02 18:25:38 ^^ thetransformerr Apr 02 18:26:10 I had a look at both, last year prudaq branch isn't much developed, Apr 02 18:26:32 please see the changes .. is it ok? Apr 02 18:26:42 I think it just had basic c file for calculation from top, Apr 02 18:27:02 and on PRUDAQ wiki I saw round robin example Apr 02 18:29:59 bullet 6 on gsoc page? i dont understand completely Apr 02 18:31:54 i mean the wiki on the beagle-sonic github Apr 02 18:34:39 yeah, that diagram is helpful Apr 02 18:35:30 "Essentially copied straight from the prudaq repo." this is the commit message by him for prudaq branch Apr 02 18:35:58 and I hope there are no other branch than the one created by last year GSoC student Apr 02 18:39:03 yeah wiki is helpful, but I hadn't given it much priority considering that most of the content is related THS1206 and also he don't specify much clearly about the transducers he is using Apr 02 18:41:41 nerdboy: do you have time to check out my proposal too? m_w was the only one to have a look at it until now I guess Apr 02 18:49:26 link? Apr 02 18:52:32 sure Apr 02 18:52:59 nerdboy: http://elinux.org/BeagleBoard/GSoC/BeagleBone-based_serial_terminal_server Apr 02 19:18:00 ordsen: that looks pretty good Apr 02 19:18:39 plan B is always good for the big ticket questions Apr 02 19:20:42 thanks! :) Apr 02 19:25:38 mentors : plz give your views about my proposal http://elinux.org/BeagleBoard/GSoC/beagle_offloader Apr 02 19:40:10 nerdboy: do you have time to check out my proposal ? Apr 02 19:47:04 zeekhuge : plz give your comments.. Apr 02 19:49:39 Akki: you know the cuurent git repop is https://git.ti.com/pru-software-support-package right? Apr 02 19:50:06 v5 was the last release plus it looks like new stuff Apr 02 19:50:13 *since v5 Apr 02 19:50:54 look at the last item on the ideas page Apr 02 19:52:32 aren't these both are same? Apr 02 19:52:45 not necessarily Apr 02 19:52:55 the beagleboard one is old style Apr 02 19:53:11 that's why it's a project idea Apr 02 19:53:17 http://processors.wiki.ti.com/index.php/PRU_Training:_Hands-on_Labs Apr 02 19:54:38 what's the difference between these two? Apr 02 19:54:40 the labs should match v5 in the ti git repo but they use the PRU cape so the labs need to be adjusted for just a beaglebone plus other stuff (without the cape) Apr 02 19:55:26 the point is the interface has changed over the life of beaglebone and several kernels from 3.8 until now Apr 02 19:56:13 v5 is curreent as far as the proper PRUSS interface using rproc/rpmsg Apr 02 19:56:49 ^^ works with mainline-ish bbkerenl and ti-dev kernel Apr 02 19:57:36 i havent made a wiki page .. Apr 02 19:57:57 just have the google docs Apr 02 19:58:15 export text and paset it in the wiki? Apr 02 19:58:22 *paste even Apr 02 19:59:12 https://eewiki.net/display/linuxonarm/BeagleBone+Black <= these are the latest kernels Apr 02 19:59:15 okay .. btw should i add something more to my proposal Apr 02 19:59:38 they both apply patches to mainline or ti-dev Apr 02 20:00:27 it looked okay, paste it in the wiki and clean it up Apr 02 20:01:07 they are also prebuilt/packages in the debian/ubuntu rootfs in that page Apr 02 20:01:30 the default debian on bb uses those repos Apr 02 20:01:43 *debian package repos Apr 02 20:02:26 anyone who hasn't compiled a kernel should do it now Apr 02 20:03:06 using those build repos ^^ you can drop patches in the patches/ dir and it will apply them Apr 02 20:05:04 hmm...seems like things will take time to get clear. I will read about these. Apr 02 20:05:35 we need to compile just a simple c code ? Apr 02 20:06:29 Students will demonstrate their ability to cross-compile and utilize version control software by creating a "Hello World" application and generating a pull request to https://github.com/jadonk/gsoc-application/tree/master/ExampleEntryJasonKridner. Apr 02 20:06:34 4. Apr 02 20:07:03 building a kernel is just a good idea... Apr 02 20:08:27 using the above repo, run build_kernel.sh to start a fresh build tree, or us tools/rebuild.sh to use the current kernel source checkout Apr 02 20:08:41 s/us/use/ Apr 02 20:10:56 hopefully that helps? unless jkridner told you something different, those are the general reqs. Apr 02 20:11:32 any suggestions regarding the proposal? Apr 02 20:12:57 just make it clear you're going to use the current interfaces not the old uio stuff Apr 02 20:15:33 Hello , this is vary silly question but how to make wiki page , in proposal they said fill this fill out this template but when I tried to do so I found only edit potion which will make change in original one. I have account on elinux Apr 02 20:18:03 make a link to your new page in parent page Apr 02 20:18:24 then you can "create" it from there Apr 02 20:19:45 ok Apr 02 20:24:25 i thought there as a parent page... Apr 02 20:24:31 *was even Apr 02 20:26:41 hmm, i see one that started from the user page Apr 02 20:26:52 so at ;east that works Apr 02 20:55:22 Akki : you deleted offload data processing from idea page and inserted your proposal on that place, please correct it . Apr 02 20:56:54 Oops ... Thanks for pointing out Apr 02 21:02:31 I have removed mine ... Any ideas how to restore back the original Apr 02 21:04:12 Akki : go to view history menu on that page and select undo against changes you made. Apr 02 21:05:18 don't remove by edit option Apr 02 21:11:28 Akki ??? Apr 02 21:12:40 No option for undo here Apr 02 21:14:56 Go to idea's page on top there are options like Read,edit, view history Apr 02 21:16:08 in view history it is showing number of changes you made so just undo it. Apr 02 21:16:32 okk? Apr 02 21:17:08 See this is happening by undoing Apr 02 21:19:49 Akki: Any reason for the priority of the peripherals ? Apr 02 21:20:23 I would expect a bit more details about how you are going to achieve the goal .. in terms of code and tools you will use ... Apr 02 21:20:56 a bit about if you understan what RemoteProc and RPMsg framework is .. Apr 02 21:21:41 and see what peripherals are already inside the PRU-ICSS, no need to access such peripherals outside PRUSS Apr 02 21:23:03 and I wont recommed using statements like "(if something left)" Apr 02 21:23:33 rishabh_bb: sorry for inconvenience but it is sorted Apr 02 21:24:04 ok, no problem Apr 02 21:24:33 You need to be clear about "I will do at least these things .. " and then comes "If time allows I'll extend move to ..." Apr 02 21:25:04 *I'll move to .. Apr 02 21:26:02 zeekhuge: i just make some priorities because at some point of time i have to decide which one to first Apr 02 21:26:32 And moreover I mentioned that i can be altered according to situation or need Apr 02 21:27:25 * priorities can be altered Apr 02 21:27:39 *decide which one to do first Apr 02 21:31:29 zeekhuge : "see what peripherals are already inside the PRU-ICSS, no need to access such peripherals outside PRUSS" by this do you mean to access only those peripherals which has shared local memory only not global? Apr 02 21:31:59 zeekhuge: UART is inside PRU-ICSS so that think we don't need, ok ? but what about Ethernet ? PRU-ICSS does have Ethernet for profinet . Apr 02 21:44:05 thats industrial Ethernet if I remember correctly .. Apr 02 21:44:18 on the PRUSS Apr 02 21:44:25 yes Apr 02 22:04:24 zeekhug : sorry ,anything I missed ? my internet connection lost Apr 02 22:07:33 rishabh_bb : make an account on IRC Cloud, it's a good IRC client along with an android app. Apr 02 22:08:42 :ok , thanks Apr 02 22:09:10 and u haven't missed anything Apr 02 22:09:57 In offload processing do we need to write user space application to check developed API's ? or APIs will use only in PRU firmware ? Apr 03 01:33:01 ds2? **** ENDING LOGGING AT Mon Apr 03 03:00:01 2017