**** BEGIN LOGGING AT Thu May 08 02:59:58 2014
May 08 03:08:39 <mranostay> Abhishek_: not for the bindings
May 08 03:09:49 <Abhishek_> and then, the server that directly calls libsigrok and libsigrokdecode would have to be GPLed too?
May 08 03:15:55 <mranostay> yeah the problem is?
May 08 03:16:00 <mranostay> well the agent
May 08 03:16:27 <mranostay> i'm not a lawyers for the record
May 08 03:16:29 <mranostay> :)
May 08 03:16:59 <Abhishek_> I'm not a lawyer too, but ds2 would definitely have something to say in this matter
May 08 03:17:15 <Abhishek_> I did ping biot on #sigrok now for it
May 08 03:17:57 <Abhishek_> mranostay: The code snippet is 3-clause BSD
May 08 03:19:44 <Abhishek_> ds2: We need you here
May 08 04:26:00 <Abhishek_> ds2: k.
May 08 04:31:07 <mranostay> ds2: if it is a binary blob that libsigrok interfaces through ioctl or whatever there is no issue
May 08 04:32:47 <Abhishek_> I plan to write the asm code myself though. The purpose of decompilation was to understand a few bits, mainly data transfer.
May 08 05:16:38 <Abhishek_> ds2: Is your concern resolved?
May 08 05:18:17 <Abhishek_> ds2: I mean, has the concern that has been raised by you resolved?
May 08 08:09:38 <praveendath92> Morning :)
May 08 08:10:10 <praveendath92> I'm looking for a screen recording software for windows 8. Suggestions ?
May 08 08:10:46 * Abhishek_ is thinking of something similar for Ubuntu 14.04LTS
May 08 08:20:00 <Abhishek_> I found recordmydesktop (and gtk-recordmydesktop) on Linux
May 08 08:37:44 <Abhishek_> praveendath92: Does CamStudio work for you?
May 08 08:38:10 <praveendath92> I'm trying that.
May 08 08:38:23 <praveendath92> Getting issues.
May 08 08:41:23 <praveendath92> It's working now
May 08 08:42:45 <praveendath92> Issue with avi recording earlier
May 08 08:54:17 <vvu|Log> praveendath92: morning
May 08 08:54:36 <praveendath92> Hello
May 08 08:55:06 <praveendath92> Working on the video :)
May 08 08:55:31 <vvu|Log> oki
May 08 08:56:13 * vvu|Log hates morning classes
May 08 08:58:36 * Abhishek_ is almost used to going to 7:30 classes and catching up with lost sleep if he has a chance
May 08 08:59:31 <vvu|Log> as always
May 08 10:21:29 <praveendath92|PC> av500: vvu|Log http://praveenkumar.co.in/#!/Downloads
May 08 10:21:32 <praveendath92|PC> bard.pdf
May 08 10:21:41 <praveendath92|PC> How does it look ?
May 08 10:22:24 <av500> nice
May 08 10:22:46 <praveendath92|PC> Any changes or inclusions ?
May 08 10:23:02 <praveendath92|PC> Or shall I proceed towards the video with this ?
May 08 10:28:37 <av500> proceed
May 08 10:29:06 <praveendath92|PC> roger :)
May 08 15:06:29 <jkridner> praveendath92: so, AOA doesn't provide a "display class", but does provide a way for arbitrary data to be passed. Your Linux driver on BeagleBone will make the phone, which implements AOA, look like a display and use AOA to pass the display information to the phone. You'll then have an app that can receive the data and show it on the display. Did I understand from your presentation? av500?
May 08 15:07:14 * Abhishek_ is back after a 3 hour timezone skip sleep
May 08 15:08:19 <praveendath92> jkridner: That's correct.
May 08 15:08:47 <praveendath92> AOA is the protocol used.
May 08 15:09:35 <praveendath92> I saved detailed explanations for the video.
May 08 15:09:36 <DiegoTc> jkridner: is this size ok for the Zoom In Lightbox plugin card: http://jsbin.com/vunum/2
May 08 15:10:09 <DiegoTc> I'm still working on it, (it's not working fine)
May 08 15:12:42 <jkridner> DiegoTc: the functionality isn't right, but the general appearance is fine. I'm still not sold on the modal approach. I'm more inclined to have it like learn.adafruit.com
May 08 15:12:57 <jkridner> ie., links
May 08 15:18:29 <DiegoTc> jkridner: you want them to take you to a new page? and the page has a next back and a view all with the card on the centeR?
May 08 15:19:03 <DiegoTc> Similar to https://learn.adafruit.com/neopixel-punk-collar/prototype-circuit, but with a better view
May 08 15:19:10 <jkridner> I think so.
May 08 15:19:40 * DiegoTc that's much easiert than trying to make modals working the way I want :)
May 08 15:20:07 <DiegoTc> but doing it that way we loose the idea of flipping cards
May 08 15:24:38 <jkridner> how so?
May 08 15:25:27 <jkridner> It does make browsing lots of different tutorials slower....
May 08 15:26:11 <jkridner> and I can see value in having a lightbox view to flip through a larger *preview*, but I think you want a directly like to go through a tutorial.
May 08 15:26:29 <jkridner> I just wouldn't want to be executing my tutorial in some modal view.
May 08 15:29:03 * praveendath92 is following this discussion
May 08 15:30:01 <DiegoTc> I get you
May 08 15:30:05 <DiegoTc> but why?
May 08 15:30:07 <praveendath92> So a preview could contain, like 4 crucial slides of that section.
May 08 15:32:59 <DiegoTc> jkridner: for example the first tutorial here http://diegotc.github.io/bone101/Support/GSOC/tempboot3cards.html works
May 08 15:33:18 <DiegoTc> if you connect the beaglebone with the led, is running (HARD CODE)
May 08 15:33:33 <jkridner> DiegoTc: I don't find it acceptable that there isn't a URL for the tutorial.
May 08 15:33:35 <praveendath92> DiegoTc: You intend to show different sections in small cards and on choosing a section opens up in a light box and he can flip through that sections tutorial. Did i get it right ?
May 08 15:34:02 <Abhishek_> ds2: I hope you are here
May 08 15:34:03 * DiegoTc thinks jkridner is correct
May 08 15:34:16 <praveendath92> # change of urls can be used.
May 08 15:34:21 <jkridner> Maybe you can have both views.... click somewhere to open in "full page" mode or something.
May 08 15:34:33 <praveendath92> Hope you guys don't mind me coming in
May 08 15:34:37 <jkridner> the fast browsing is nice.
May 08 15:34:54 <DiegoTc> praveendath92: yes you're right
May 08 15:35:06 * jkridner wonders if http://stackoverflow.com/questions/824349/modify-the-url-without-reloading-the-page is usable
May 08 15:35:14 <praveendath92> Yep
May 08 15:35:19 <praveendath92> That is
May 08 15:35:32 <praveendath92> I used that infact
May 08 15:35:42 <jkridner> window.history.pushState({"html":response.html,"pageTitle":response.pageTitle},"", urlPath);
May 08 15:36:17 <jkridner> The address needs to update for cut-and-paste and link save purposes.
May 08 15:36:29 <jkridner> not just #asdfadsf or ?asdfasdf
May 08 15:36:37 <jkridner> need real urls
May 08 15:36:49 <praveendath92> http://praveenkumar.co.in/
May 08 15:37:06 <praveendath92> Please try that link once
May 08 15:37:07 <jkridner> #! is a bit odd.
May 08 15:37:18 <praveendath92> It is. True.
May 08 15:37:29 <praveendath92> Twitter use that !
May 08 15:37:36 <praveendath92> #something also works
May 08 15:38:46 <praveendath92> Open a section. Copy url and try in new tab with that url.
May 08 15:38:50 <DiegoTc> jkridner: I will try to do a page with a the card as center,and see how it works
May 08 15:39:00 <praveendath92> That is basically a thought.
May 08 15:39:30 <praveendath92> jkridner idea is also nice.
May 08 15:39:47 <DiegoTc> praveendath92: which one of all
May 08 15:39:59 <DiegoTc> getting quite confuse :p
May 08 15:40:19 <praveendath92> I guess i didnt follow all. The new page for a section.
May 08 15:40:33 <praveendath92> Sorry. I'm adding confusion.
May 08 15:41:10 <DiegoTc> ahh that's what I'm thinking don't worry praveendath92
May 08 15:41:37 <praveendath92> DiegoTc: Your initial idea was to do similar to Google keep ?
May 08 15:41:46 <praveendath92> With different card design ?
May 08 15:42:57 <DiegoTc> not really. It was something similar to: https://learn.adafruit.com/category/beaglebone
May 08 15:43:20 <DiegoTc> instead of going to new pages just next and back
May 08 15:46:02 <DiegoTc> with a modal
May 08 15:46:04 * praveendath92 hates his slow internet:/
May 08 15:47:45 <vvu|Log> praveendath92: u have not seen the wi-fi at my university
May 08 15:47:48 <praveendath92> Still loading. I will get back once it does.
May 08 15:48:22 <praveendath92> I'm at home. A 2g internet is all have or I can get.
May 08 15:49:12 <DiegoTc> 2mb?
May 08 15:49:21 <DiegoTc> sortty 2gb?
May 08 15:50:42 <praveendath92> 2G network mode I'm talking about. In 2g/3g/LTE context.
May 08 15:51:09 <DiegoTc> jkridner: just to confirm no scrollbar with this pages
May 08 15:51:11 <vvu|Log> no ethernet connection ?
May 08 15:51:26 <DiegoTc> I will try think about a cool, nice design then
May 08 15:52:43 <praveendath92> Nope.
May 08 15:52:51 <praveendath92> Finally. Loaded.
May 08 15:52:57 <DiegoTc> praveendath92: you are from India?
May 08 15:53:04 <praveendath92> That's an elegant design !
May 08 15:53:09 <praveendath92> Yep.
May 08 15:53:46 <DiegoTc> this was my first crazy idea: http://diegotc.github.io/bone101/Support/GSOC/tempboot3cards.html
May 08 15:53:50 <praveendath92> It's just my home. Back in college we get crazy speeda
May 08 15:55:22 <DiegoTc> praveendath92: I imagined internet connection is that way in certain areas of the country?
May 08 15:55:26 <praveendath92> The design used on that learn.adafruit  is nice.
May 08 15:55:44 <praveendath92> Yeah it is. In villages.
May 08 15:56:07 <praveendath92> Image in a card having no padding and all.
May 08 15:56:30 * praveendath92 is a big fan of Google card design
May 08 15:57:07 <praveendath92> That design with no scroll bars looks great !
May 08 15:57:22 <praveendath92> Any design with no scroll bars actually.
May 08 15:58:14 <DiegoTc> praveendath92: check this one: http://masonry.desandro.com/
May 08 15:59:53 <praveendath92> This is good. Actually i had a suggestion about fruit page related to this.
May 08 16:00:38 <praveendath92> Due to this layout, all cards in the second row won't be at same height
May 08 16:00:55 <DiegoTc> But right now, we change that: http://jsfiddle.net/DiegoTc/KVWFj/7/embedded/result/
May 08 16:01:05 <DiegoTc> I'm going to use pagination
May 08 16:01:44 <praveendath92>  I personal felt, though it makes good use of space, is difficult when looking for details in card
May 08 16:02:02 <praveendath92> Or while searching for a section
May 08 16:03:23 <praveendath92> This looks great. Each section is like a book.
May 08 16:04:20 <praveendath92> And each row at same height.
May 08 16:04:37 <praveendath92> There will be a flip animation ?
May 08 16:04:50 <DiegoTc> don't know
May 08 16:09:32 <praveendath92> I will stay in the loop.
May 08 16:09:55 <praveendath92> :)
May 08 16:43:11 <Abhishek_> mranosta1: what did you mean when you said: "Don't think of them as GPIO, they aren't"?
May 08 16:44:45 <mranostay> Abhishek_: they are GPO or GPI for one
May 08 16:50:34 <Abhishek_> okay, does that mean if I am using them for input, I cannot then use them as output or they are accessed using separate registers i.e. I write to R30 and read from R31?
May 08 16:51:28 <mranostay> one or the other
May 08 16:51:37 <mranostay> pinmux sets what the state will be
May 08 16:51:51 <Abhishek_> pinmux with the device tree, correct?
May 08 16:52:43 <Abhishek_> ah, pinmux with the PRU
May 08 16:52:57 <mranostay> no!!
May 08 16:53:15 <mranostay> you use device tree.. if i see you touch the pinmux registers from the PRU ... :)
May 08 16:55:58 * Abhishek_ wonders what mranostay will do when he says :)
May 08 16:58:09 <ds2> depending on which "pin", you can use it for in and out
May 08 16:58:20 <Abhishek_> hi ds2
May 08 16:58:40 <ds2> 16 I think is special in that it is routed to 2 physical pins
May 08 16:59:31 <ds2> hi
May 08 16:59:41 <mranostay> schematic explains all
May 08 17:00:10 <mranostay> Abhishek_: i highly doubt you'll have any outputs
May 08 17:01:00 <Abhishek_> indeed. I have all inputs, I was just clarifying what you said in the mail.
May 08 17:02:11 <Abhishek_> Another was the latencies Charles talked about in the mail, are they one-time latencies?
May 08 17:05:25 <ds2> latencies he was referring to is off PRU resources
May 08 17:05:47 <ds2> in PRU, if you access via the register is 1 cycle for all the signals
May 08 17:06:20 <ds2> and if you are clever, you do something like
May 08 17:06:58 <ds2> mov r0.b0, r30.b0; mv r0.b1, r30.b0...... for all the registers to get a very quick 100MHz burst
May 08 17:07:27 <ds2> next step is 50MHz if you want to keep things evenly timed
May 08 17:07:58 <ds2> i.e. mov .. nop mov... nop ... mov nop mov swap banks... mov... signal other PRU... mov...
May 08 17:08:19 <ds2> blah sorry, off by a factor of 2
May 08 17:08:30 <ds2> 200MHz burst is the first...100MHz is the 2nd
May 08 17:14:14 <mranostay> use the registers then pipe over to the PRU RAM
May 08 17:56:03 <Abhishek_> mranostay: You mean to say that I first read from R31 into the registers, and then use the other PRU to move that data into the RAM, correct?
May 08 18:20:57 <ds2> bah
May 08 18:22:54 <ds2> Abhishek_: still awake?
May 08 18:23:29 <Abhishek_> yeah, would be for quite sometime as I slept in the evening
May 08 18:29:33 <ds2> okay... quick overview of the PRUSS is needed based on your email.
May 08 18:30:02 <ds2> 1. The PRUSS is a block ("chunk of silicon") on the AM335x that the BeagleBone boards use (BBB and BBW).
May 08 18:31:00 <ds2> 2. The PRUSS internally has several resources: 2 processors, the PRU's, memory for instructions, memory for data, counter/timer, uart, and signals from and to the outside world.
May 08 18:31:39 <ds2> 3. The PRUSS subsystem is linked to the internal shared L3 bus on the AM335x that goes to most parts of the chip, including the ARM Cortex-A8.
May 08 18:32:34 <ds2> 4. On the AM335x, almost every signal that goes to the outside world (i.e. a "ball" on the BGA) goes through a multiplexer. aka the pinmux
May 08 18:33:24 <ds2> So given the above, if you want to take signal from the outside world and look at it on the PRU processor, you will first need to setup the pinmux to route that ball to the PRUSS subsystem's input signal.
May 08 18:34:09 <ds2> If you cannot do that routing (i.e. pinmux limitations due to the ball not being connected to the PRUSS's input signal), you need to get to it via the L3 bus.
May 08 18:34:37 <ds2> What this comes down to is - If you read a signal through the PRUSS's input signals, there is no additional latency. It is just another register read.
May 08 18:35:17 <ds2> However, if you read a signal through the L3 bus, you need to contend with L3 protocol and traffic which can cause significant latencies. This is done with the SBBO/LBBO instructions (vs mov and other instructions for the internal signals)>
May 08 18:35:20 <ds2> Clear?
May 08 18:35:57 <Abhishek_> ds2: I was a little confused with respect to the latencies when I saw Charles' code [I had seen it before too as it had popped up in the BeagleBoard group], because I could not differentiate between PRU R30 and R31 specialized GPIO subsystems and the AM335X own GPIO system which is on L4 .
May 08 18:38:31 <ds2> Abhiskek_: *nod* hence the attempt at making that clear... there are 2 different resources. the latency rules are pretty straight forward. If you leave the PRUSS, it gets slow.
May 08 18:39:54 <Abhishek_> So does that mean that I cannot write MOV R31.b0, I'll have to do LBBO and SBBO ?
May 08 18:41:02 <ds2> No
May 08 18:41:10 <ds2> there are 2 types of lines
May 08 18:41:31 <ds2> if you use a regular GPIO and muxed it as such, you have to use LBBO/SBBO which is cycle expensive.
May 08 18:42:10 <ds2> if you use a EGPIO (the signals on the PRUSS), you can do mov's with r31/r30
May 08 18:42:23 <ds2> the mov's are 1 cycle instructions. SBBO/LBBO is multi-cycle
May 08 18:42:47 <ds2> SBBO/LBBO accesses things outside of the PRUSS via the L3
May 08 18:43:13 <Abhishek_> Yeah, as you write in No. 3
May 08 18:44:42 <ds2> >
May 08 18:45:01 <ds2> anything leaving the PRUSS gets expensive cycle wise. it can also get nondeterministic :(
May 08 18:46:20 <ds2> with a few minor exceptions, most things within the PRUSS are 1 cycle..the exceptions mainly deal with writing to the data memory/shared memory... but keep in mind things can happen 32bits at a time. so if you are dealing with 8 bits, you can gain a factor of 4 timing
May 08 18:50:54 <mranostay> Abhishek_: scratchpad should be used for this too
May 08 19:06:30 <Abhishek_> ds2: When I use device tree for the pinmux and I connect a pin to pru1_r31_0, does that mean that I can use MOV R31... in the PRU to access it now, correct?
May 08 19:08:26 <ds2> Abhishek_: if you did it right, yes.
May 08 19:08:59 <ds2> that's probally mode 5 or mode 6, IIRC
May 08 19:09:03 <ds2> mode 4 is regular GPIO
May 08 19:24:33 <Abhishek_> Yeah, mode 6.
May 08 21:40:42 <ds2> l
May 08 21:41:32 <Abhishek_> !
May 08 23:13:28 <Abhishek_> G'night
May 08 23:51:24 <ds2> \
**** ENDING LOGGING AT Fri May 09 03:00:01 2014