**** BEGIN LOGGING AT Mon Jun 20 02:59:59 2016
Jun 20 12:53:27 <jic23> ZeekHuge: All going well?
Jun 20 12:54:00 <jic23> Not to worry if you are busy - I'm just eating lunch and bored ;)
Jun 20 12:56:35 <ZeekHuge> jic23: Hi ! ahh, struggling with PRU code
Jun 20 12:56:37 <ZeekHuge> and
Jun 20 12:57:08 <ZeekHuge> trying to understand the new ways that TI has introduced
Jun 20 12:57:31 <jic23> Sadly not an area I know anything much about...
Jun 20 12:58:10 <jic23> Out of curiosity has anyone done an effective bandwidth test on rpmesg?
Jun 20 12:58:11 <ZeekHuge> I have started to make posts about it, though I am slow at it (also, my father came here to meet me, and it was fathers day here yesterday, so was with him almost all the time )
Jun 20 12:58:39 <jic23> Sure, was fathers day here yesterday too.  Did my kind notice? Nope... Hmph.
Jun 20 12:58:41 <ZeekHuge> Ahh.. I was going to do that ... but Abhishek_ suggested me to use rpmsg anyway
Jun 20 12:58:54 <jic23> Definitely use it 'first'.
Jun 20 12:59:11 <ZeekHuge> and suggested that we would work to improve it later
Jun 20 12:59:40 <ZeekHuge> rpmsg as it would be the way with the standards
Jun 20 12:59:46 <jic23> If it doesn't work for bulk transfers then that is fine.  However, we should figure standard ways of setting up bulk channels as this isn't exactly an unusual problem.
Jun 20 13:00:13 <jic23> rpmsg could be used as configuration interface for say dma configuration.
Jun 20 13:01:24 <jic23> Guess see how far you get with rpmsg first...
Jun 20 13:02:26 <ZeekHuge> okay so for, I should go with what is available, right
Jun 20 13:02:27 <ZeekHuge> ?
Jun 20 13:02:42 <ZeekHuge> *for now
Jun 20 13:03:18 <jic23> absolutely, but when you hit the boundaries the trick is how to move the boundaries in a way useful to you and everyone else hitting the same issue.
Jun 20 13:03:19 <ZeekHuge> jic23: zeekhuge.github.io
Jun 20 13:03:58 <ZeekHuge> https://zeekhuge.github.io
Jun 20 13:05:02 <ZeekHuge> have to add some more commands and comments section there ..
Jun 20 13:06:14 <ZeekHuge> so, instead of starting with rpmsg itself, I have started from basics, in the posts. this way, It will help to organize all that in my mind too ..
Jun 20 13:06:24 <jic23> was about to say the font was hideous then noticed something changed all my default fonts.
Jun 20 13:06:26 <jic23> :)
Jun 20 13:06:35 <jic23> So the whole internet is hideous today.
Jun 20 13:06:58 <jic23> Makes a lot of sense.  Definitely a good way to clarify things in your mind.
Jun 20 13:06:58 <ZeekHuge> another browser ?
Jun 20 13:07:45 <ZeekHuge> And regarding your code review mail, I will first make the corrections and then reply back
Jun 20 13:08:37 <jic23> cool.
Jun 20 13:08:54 <jic23> Yeah, can't be bothered to figure what is wrong with fonts.  Will probably clear up on a restart of the browser.
Jun 20 13:10:03 <ZeekHuge> I will soon add a comment section there, but if you find a mistake please let me know .
Jun 20 13:11:01 <ZeekHuge> oh ! and I had a question ..
Jun 20 13:11:48 <ZeekHuge> jic23: you asked me to use a better approach than using bools for locks. so I guess you were suggesting to use mutex right ?
Jun 20 13:11:55 <jic23> yup
Jun 20 13:12:24 <ZeekHuge> Cool ! will do that . but this PRU code first ...
Jun 20 14:03:51 <chanakya_vc> Hey bradfa there?
Jun 20 14:03:59 <chanakya_vc> mdp ^^
Jun 20 14:19:47 <bradfa> chanakya_vc: hi!
Jun 20 14:20:09 <bradfa> chanakya_vc: ZeekHuge seems to also be wresling with rpmsg implementation, might be worth having you two discuss together some about what you're finding
Jun 20 14:20:24 <bradfa> could be good resources for each other
Jun 20 14:23:50 <chanakya_vc> bradfa, I pushed some code for the driver.But it was based on a character device implementation.Right now I am following this:http://invo-tronics.com/spi-driver-for-linux-based-embedded-system/
Jun 20 14:25:14 <chanakya_vc> bradfa, I will. At the moment I want to just get the basic framework of the SPI driver up.Get it to send data via spi_tansfer_message()
Jun 20 14:25:23 <bradfa> chanakya_vc: ok
Jun 20 14:29:18 <bradfa> chanakya_vc: I'm replying to your weekend email, again, I believe you are correct in your 2nd point.
Jun 20 14:31:23 <pmezydlo> bradfa: hi, now I'm testing slave driver in pio mode. Maximum transmit length is 32 bytes; the driver has no limit for receiving.
Jun 20 14:31:38 <bradfa> chanakya_vc: that link you just posted here, looks more like a tutorial on how to write a protocol driver
Jun 20 14:31:51 <bradfa> pmezydlo: awesome! :)
Jun 20 14:31:59 <pmezydlo> I think This is a good result
Jun 20 14:32:00 <bradfa> pmezydlo: are you using the fifo inside mcspi?
Jun 20 14:32:05 <pmezydlo> yeah
Jun 20 14:32:23 <bradfa> pmezydlo: can you send a link to your current code?
Jun 20 14:32:29 <chanakya_vc> bradfa, Okay. As it is I have realized it is not making much sense to me :P
Jun 20 14:32:33 <pmezydlo> the driver must load 32 bytes before transfer
Jun 20 14:32:43 <bradfa> pmezydlo: for MISO?
Jun 20 14:35:42 <pmezydlo> pls wait a few minutes, i'm uploading the code
Jun 20 14:35:48 <bradfa> pmezydlo: no worries :)
Jun 20 14:39:28 <pmezydlo> bradfa: https://github.com/pmezydlo/SPI_slave_driver_implementation
Jun 20 14:40:16 <bradfa> pmezydlo: awesome! :)
Jun 20 14:40:16 <pmezydlo> bradfa: My next goal is basic communiaction with user. After that DMA
Jun 20 14:40:29 <bradfa> pmezydlo: sounds like you're making very good progress :)
Jun 20 14:41:12 <pmezydlo> that's great work, I'm very glad.
Jun 20 14:42:17 <pmezydlo> bradfa: I hope that, DMA solve the problem of the maximum transfer
Jun 20 14:42:41 <pmezydlo> slave->master max 32 bytes
Jun 20 14:43:24 <bradfa> pmezydlo: it will likely help, although you still need to have a reasonable thing to do if the master clocks more than you expect, PIO or DMA
Jun 20 14:45:13 <chanakya_vc> bradfa, Okay received your mail.I am clear on how the PRU will access the shared memory.
Jun 20 14:45:58 <pmezydlo> bradfa: Thanks, I'm going back to work.
Jun 20 14:45:58 <bradfa> chanakya_vc: the PRU can also reach DDR, if you wanted to have one PRU core read/write DDR and copy data betwen the 12kiB shared memory while the other PRU core does the bitbanging
Jun 20 14:46:15 <bradfa> chanakya_vc: I'm not sure which is a better solution, performance wise, but just pick one to start with
Jun 20 14:46:38 <chanakya_vc> But what Abhishek_  pointed to me was the address over here:https://github.com/beagleboard/linux/blob/4.4/arch/arm/boot/dts/am33xx.dtsi#L946
Jun 20 14:46:42 <chanakya_vc> bradfa, ^^
Jun 20 14:46:43 <bradfa> chanakya_vc: probably the way you are going, with linux writing into the 12 kib Shared memory will be less complicated
Jun 20 14:47:35 <chanakya_vc> bradfa,  This is the address ARM would probably use to write to that shared mem.
Jun 20 14:47:38 <bradfa> chanakya_vc: that's a good example from Abhishek_
Jun 20 14:48:09 <chanakya_vc> Yes I think I get the whole idea about addresses I am going to be using.
Jun 20 14:48:14 <bradfa> chanakya_vc: :)
Jun 20 14:51:42 <chanakya_vc> bradfa, Now I am currently stuck with the driver bit.I know that a struct spi_master() along with spi_transfer_one message has to be declared.
Jun 20 14:52:24 <bradfa> chanakya_vc: if you take your char driver idea and just create a hack for now to make sure that your pru firmware does the right things, I think that'll be the easiest way to start
Jun 20 14:52:47 <bradfa> chanakya_vc: then after you have that working, start trying to make a driver which is a spi master
Jun 20 14:53:02 <chanakya_vc> bradfa,  Okay.So let me work on the char driver first.
Jun 20 14:53:48 <chanakya_vc> bradfa, There also I have a few doubts:Now I will have define a function that I will bind with .read in file operations structure right?
Jun 20 14:54:04 <chanakya_vc> To read data coming in from the userland?
Jun 20 14:55:24 <bradfa> chanakya_vc: yes
Jun 20 14:55:33 <bradfa> chanakya_vc: in lddv3 there should be a skull driver example, iirc
Jun 20 14:55:50 <bradfa> chanakya_vc: it should give you a good overview of read/write interface for a char driver
Jun 20 14:55:56 <chanakya_vc> bradfa, Something of /dev driver given here:http://www.linuxdevcenter.com/pub/a/linux/2007/07/05/devhelloworld-a-simple-introduction-to-device-drivers-under-linux.html?page=2
Jun 20 14:56:36 <chanakya_vc> http://www.linuxdevcenter.com/pub/a/linux/2007/07/05/devhelloworld-a-simple-introduction-to-device-drivers-under-linux.html?page=3
Jun 20 14:56:47 <chanakya_vc> The next pg bradfa ^^ Something like this?
Jun 20 14:58:00 <bradfa> chanakya_vc: like this https://lwn.net/images/pdf/LDD3/ch03.pdf
Jun 20 14:58:13 * bradfa misspelled "scull" before, sorry
Jun 20 15:01:58 <chanakya_vc> bradfa, Okay let me go through this.Are you free today?
Jun 20 15:03:33 <bradfa> chanakya_vc: I'll be around, yes
Jun 20 15:03:58 <chanakya_vc> Okay. I will read this and clear all my doubts. bradfa
Jun 20 15:04:37 <chanakya_vc> bradfa, I am sorry for my slow progress. But once I get done with SPI, I2C won't take so much time
Jun 20 15:06:27 <bradfa> chanakya_vc: don't worry, your progress is not slow, this is not simple
Jun 20 15:06:45 <bradfa> chanakya_vc: you are correct that i2c should go faster, though, you'll be able to reuse a ton of knowledge and code :)
Jun 20 15:07:03 <bradfa> chanakya_vc: I'm learning a lot working along with you on this
Jun 20 15:12:34 <chanakya_vc> bradfa, : )
Jun 20 15:13:34 <chanakya_vc> bradfa, I will get back to you after reading up on this.I am really motivated to finish this now : ) Thanks.
Jun 20 16:05:10 <ZeekHuge>  Haha ! I got PRU1 -> PRU0 -> ARM communication !
Jun 20 16:05:47 <m_w> ZeekHuge: excellent work
Jun 20 16:06:54 <ZeekHuge> :) Thank you. Well, there is just one part in the code, I am bit confused about .. so diving in more deep now /
Jun 20 16:06:55 <ZeekHuge> .
Jun 20 16:08:31 <ZeekHuge> and its using C, need to port PRU1 code to assembly.
Jun 20 16:09:17 <m_w> C not efficient enough?
Jun 20 16:16:38 <ZeekHuge> yeah, ds2 said that it should be in assembly, I was able to generate about 16MHz using C . But that was it, it was doing nothing else, so I think that it wont be efficient enough.
Jun 20 16:22:31 <bradfa> ZeekHuge: is inline assembly possible to just drop into your inefficient C code sections?
Jun 20 16:26:22 <ZeekHuge> Yeah, I was thinking about that too. And I think that it will work.
Jun 20 16:26:28 <ZeekHuge> bradfa: ^
Jun 20 16:34:56 <kiran4399> alexhiam: there?
Jun 20 17:26:19 <pmezydlo> m_w: hi, Should char driver and slave driver be together in the same file, as one driver?
Jun 20 17:28:05 <m_w> pmezydlo: if the intent is making a framework, I would use a separate driver so that other processors can use it
Jun 20 17:28:51 <m_w> if you want a standalone driver then the same file is fine
Jun 20 17:30:12 <pmezydlo> I think that standalone driver should be better.
Jun 20 17:30:45 <m_w> it would certainly be easier and we can look into restructuring for a framework later
Jun 20 17:37:04 <pmezydlo> m_w:I will make framework standalone char driver and communication with slave driver. so it should be faster and easier
Jun 20 17:40:58 <m_w> okay
Jun 20 17:43:40 <alexhiam> hi kiran4399
Jun 20 17:47:24 <kiran4399> alexhaim: so.. I am in the middle of completing the following stuff which I am ready to show in the mid-eval..
Jun 20 17:50:36 <kiran4399> 1. IMU, PWM, Baro, ADC APIs
Jun 20 17:51:03 <kiran4399> 2. Full documentation on the readme.md on my github repo
Jun 20 17:51:25 <kiran4399> 3. Patched MPU-9250 kernel driver
Jun 20 17:54:10 <kiran4399> 4. Details of each kernel driver on the github wiki page.
Jun 20 17:54:11 <alexhiam> kiran4399: cool. what's the status of the drivers? do you have both sensors working?
Jun 20 17:54:31 <kiran4399> alexhiam: both sensors?
Jun 20 17:54:40 <alexhiam> imu and pressure
Jun 20 17:55:04 <alexhiam> or are you still having trouble with the bmp driver?
Jun 20 17:55:19 <kiran4399> alexhiam: yeah.. thanks to your hacks!! ;)
Jun 20 17:55:45 <alexhiam> yeah they are both working?
Jun 20 17:56:29 <kiran4399> alexhiam: btw.. for the APIs I tested the kernel driver for mpu6050.. it is working perfectly.. I am assuming that mpu9250 also would work..
Jun 20 17:56:37 <kiran4399> alexhiam: though I did not test..
Jun 20 17:57:15 <alexhiam> yeah, I think the only difference will be the extra directory for the mag
Jun 20 17:57:38 <kiran4399> did you make a note of the additional sysfs entries for mag?? What are they?
Jun 20 17:57:42 <kiran4399> alexhiam: ^^^
Jun 20 17:57:43 <alexhiam> kiran4399: so you can get mpu9250.c switched over to using the sysfs entries then?
Jun 20 17:57:56 <alexhiam> no.. I thought you had patched the driver as well?
Jun 20 17:57:58 <kiran4399> alexhiam: yeah..
Jun 20 17:58:22 <alexhiam> plus separate out the bmp280 stuff to its own file and do the same
Jun 20 17:58:49 <kiran4399> alexhiam: oh.. no not yet.. I patched the kernel and compiled it.. but I don't understand how to load it into the debian 8 image?
Jun 20 17:59:12 <alexhiam> https://gist.github.com/alexanderhiam/29fcf814f731dd94f8850961f1a61ae3
Jun 20 17:59:18 <alexhiam> just build debs
Jun 20 17:59:36 <alexhiam> https://gist.github.com/alexanderhiam/29fcf814f731dd94f8850961f1a61ae3#file-beaglebone-kernel-dev-sh-L10-L17
Jun 20 18:03:25 <kiran4399> alexhiam: directly install the deb file on the board?
Jun 20 18:03:27 <kiran4399> so easy??
Jun 20 18:03:31 <alexhiam> yup
Jun 20 18:19:50 <jkridner> hi amragaey... haven't seen any BBUI patches lately. what's going on?
Jun 20 18:20:53 <alexhiam> kiran4399: and I'd like to see the travis-ci build passing again. If you make sure it's all building every time you make a change it'll save you debugging time later on
Jun 20 18:21:26 <kiran4399> alexhiam: alright.. I'll do that..
Jun 20 18:21:45 <kiran4399> alexhiam: anything else you want me to do?
Jun 20 18:22:58 <alexhiam> kiran4399: it would be ideal if you were also writing unit tests for each component of the APIs as you work on them
Jun 20 18:24:17 <kiran4399> alexhiam: yeah.. I wanted to ask you.. since I am building the apis differently.. should I make the unit tests?
Jun 20 18:24:24 <amragaey> hi jkridner, yes I haven't pushed anything yet
Jun 20 18:25:12 <alexhiam> kiran4399: well anything that they have test scripts for can be adapted, which should hopefully just be a matter of changing the include
Jun 20 18:25:41 <kiran4399> alexhiam: I am sorry.. I did'nt get that..
Jun 20 18:25:50 <amragaey> I'm went this week on Events, trying to get how it works, and I understood it.. currently I'm working in the events and UI for the action after dragging buttons into graph
Jun 20 18:26:01 <amragaey> I went*
Jun 20 18:27:20 <amragaey> I'll try to get some output and push today, jkridner
Jun 20 18:33:42 <ZeekHuge> jkridner av500 _av500_ : what should I do about the ADC EVM thing ?
Jun 20 18:35:37 <kiran4399> h
Jun 20 18:43:34 <alexhiam> kiran4399: I just mean if there's already strawson test programs you should be able to use them, since the APIs aren't changing, just how they're implemented
Jun 20 18:45:39 <amragaey> jkridner, isn't 192.168.7.2:3000 the url of bb c9 ?
Jun 20 18:46:44 <amragaey> (window.location.host) gives the localhost ip 127.0.0.1
Jun 20 18:48:26 <amragaey> jkridner, can you give example for another socket connection url ?
Jun 20 19:04:51 <Wormo> nerdboy: you didn't hear back from visaoni yet did you?
Jun 20 19:09:10 <nerdboy> nope
Jun 20 19:16:18 <chanakya_vc> bradfa, mdp I am writing the code now for the driver(have understood the ch bradfa told me read). Should be done by tomorrow.
Jun 20 19:17:13 <chanakya_vc> bradfa, Will push the code by tomorrow evening.
Jun 20 19:21:26 <Abhishek_> ZeekHuge: aren't they going to collect the taxes from you when they deliver the package to you?
Jun 20 19:22:37 <ZeekHuge> Abhishek_: Yes, that is what I said in the first mail.
Jun 20 19:22:46 <Abhishek_> Then?
Jun 20 19:23:09 <ZeekHuge> but, I will still need the invoice details to get fill in form 50
Jun 20 19:23:47 <Abhishek_> Did you speak to a customer care representative yet?
Jun 20 19:24:26 <ZeekHuge> 3 times ... to get the status ...
Jun 20 19:24:58 <Abhishek_> Call the Mumbai FedEx office where the shipment is held.
Jun 20 19:25:13 <Abhishek_> Customer care offices can only do so much. You need to escalate.
Jun 20 19:32:15 <Wormo> Abhishek_: Does he need to escalate? Sounded like he just needs some info from jkridner to complete the form, and then all will be well
Jun 20 19:33:32 <Abhishek_> Wormo: usually in India, shipments for individuals may not require as much processing a commercial shipment to a company
Jun 20 19:33:41 <Abhishek_> *as a commercial
Jun 20 19:34:55 <ZeekHuge> Abhishek_: okay, I will try that tomorrow morning as the number seems to be busy.
Jun 20 19:36:28 <Abhishek_> Call them during regular office hours, probably 9-5
Jun 20 19:42:54 <ZeekHuge> 9-9 as stated on their website. Yeah, I was just trying now ... was not expecting that they will attend the call.
Jun 20 19:44:01 <Wormo> Well good that you know something about it Abhishek_, all I know is that customs can cause long delays
Jun 20 19:46:53 <_av500_> jkridner: ping
Jun 20 19:47:31 <bradfa> chanakya_vc: ok, sounds good! :)
Jun 20 20:12:58 <Abhishek_> I don't think it's even customs that's causing the delay (they will collect it on delivery), it's just some state Sales Tax formalities.
Jun 20 20:16:05 <ds2> hmmmm
Jun 20 20:19:37 <jkridner> _av500_: pong
Jun 20 20:21:10 <Wormo> jkridner: it was about ZeekHuge's form, what to put on it, see email
Jun 20 20:21:19 <gcl-bot> [jadonk] k
Jun 20 20:25:58 <ZeekHuge>  I didnt understand the point that pru code needs a different compiler ... are they talking about clpru ? and clpru not being a gcc compiler ?
Jun 20 20:26:04 <ZeekHuge> https://groups.google.com/forum/#!topic/beagleboard/cYHCN3GWw_E
Jun 20 20:40:44 <Wormo> ZeekHuge: yes
**** ENDING LOGGING AT Tue Jun 21 02:59:58 2016