**** BEGIN LOGGING AT Mon Jun 13 02:59:58 2016 Jun 13 04:39:42 alexhiam: there? Jun 13 04:40:29 Has anyone got a mail from TI regarding TI One stop shop order? Jun 13 05:00:40 hi m_w ! Jun 13 05:00:58 just woke up ... Jun 13 05:02:37 hey ZeekHuge Jun 13 05:02:55 Hi ! Jun 13 05:02:59 what were you asking earlier, I was on the phone with my mom Jun 13 05:03:06 was distracted Jun 13 05:03:21 oh .. ahh .. just a sec Jun 13 05:03:36 m_w: https://github.com/ZeekHuge/BeagleScope/tree/master/examples/kernel_examples/n-blinky Jun 13 05:03:52 I have completed this example module Jun 13 05:04:05 and I was asking to review it . Jun 13 05:04:36 I have tried to take care of every single thing. coding style, function names, use of generic methods and everything that i have known till now, about the kernel module writing Jun 13 05:04:57 I have to add comments and some mutex_locks Jun 13 05:05:11 but will do them along with the next steps Jun 13 05:05:31 okay it looked very clean and orderly Jun 13 05:05:55 :) Jun 13 05:06:46 and was like .. ahh .... okay for upstreaming kind if things ? Jun 13 05:07:08 m_w: I mean the writing style and all. Jun 13 05:08:40 And yeah you can try it :) but for me, its not loading when being compiled on BBB. So I had to keep it on kernel source and then compile Jun 13 05:10:19 also, It was correct versioning number on BBB earlier and I dont know what changed that it stopped getting versioning number afterwards. Jun 13 05:11:18 it looks like it is following coding conventions Jun 13 05:11:43 it is a bit late for me to try it now Jun 13 05:12:24 okay ! no problem:) Jun 13 05:12:43 good night :) Jun 13 05:13:58 I would like to give it a try before the meeting so I can give some feedback Jun 13 05:14:24 I have to prep my BBB first Jun 13 05:14:35 needs new kernel right? Jun 13 05:20:58 4.4.11-ti-r29 Jun 13 05:23:37 m_w: I got a mail from ti regarding ti one stop shop order.. what was that? Jun 13 05:24:31 I have no idea Jun 13 05:25:13 are you expecting more hardware? Jun 13 05:29:36 m_w: yes, its here https://github.com/RobertCNelson/linux-stable-rcn-ee/releases/tag/4.4.11-ti-rt-r29 Jun 13 05:30:19 did you load it manually or can you upgrade it through the debian package manager? Jun 13 05:30:46 I did it manually. Jun 13 05:32:41 did you compile on the board or on your host? Jun 13 05:35:58 host. Actually I was compiling it on my board first. It was working. Then suddenly, It began to complain about un availability of version number while loading (that is when I asked that question yesterday) and then I had to crosscompile whole kernel and the module. Jun 13 05:49:20 kiran4399: I too have the mail Jun 13 05:49:39 and the total cost is $0 :D Jun 13 05:50:15 sunglasses and USB bracelets and graph pads and all :D Jun 13 05:50:32 and a beach ball too ! hahaha ! Jun 13 05:51:51 ZeekHuge.. what is that about? Jun 13 05:51:56 is that really for us? Jun 13 05:53:16 I believe so. Its on the exact same address that i mentioned in my GSoC profile and I receive all GSoC things there. Jun 13 05:54:09 also, IIRC it had a note saying something like : "this will be shared with other companies to send you promotional material" Jun 13 05:54:42 strange Jun 13 09:36:09 Abhishek_: there ? Jun 13 12:18:18 hey bradfa !There? Jun 13 12:19:44 hi chanakya_vc Jun 13 12:20:10 I'm here for the next hour or so, I'm going to take a look at your work and emails from the weekend, sorry was very busy this weekend Jun 13 12:21:21 Np bradfa Take a look at this:http://pasteboard.co/1DsNjN8E.png Jun 13 12:22:06 Will you be free later today bradfa? Jun 13 12:22:22 Maybe in 4 to 5 hours? Jun 13 12:23:14 chanakya_vc: yes, I have meetings until about my lunch-time but then "free" after that. I should be at my computer and on IRC Jun 13 12:23:26 chanakya_vc: I like your logic analyzer plot! :) Jun 13 12:24:25 This was given to me by alexhiam.I am unable to reproduce this on my computer.I would need your help with that bradfa. Jun 13 12:25:22 chanakya_vc: ok, can sigrok find your saleae clone yet? Jun 13 12:25:38 bradfa, Yes. I downloaded pulseview too Jun 13 12:27:25 So there when I click on devices that require fxlawfw ,it shows saleae logic Jun 13 12:27:37 And is connected to it too. Jun 13 12:28:44 Also bradfa,I have come up with a way to let the ARM talk to the PRU via the PRU-ICSS shared RAM. I have some doubts and would like to clear them up with you tonight. Jun 13 12:30:19 chanakya_vc: can you do a simple capture with no data with pulseview? Jun 13 12:30:50 Okay simply run it? Jun 13 12:31:01 bradfa, right?Wait Jun 13 12:32:58 bradfa, http://pasteboard.co/1DsNjN8E.png Jun 13 12:39:07 chanakya_vc: that's the same link you sent before Jun 13 12:39:21 and that doesn't look like pulseview to me.. Jun 13 12:45:55 bradfa, Ohh sorry...http://pasteboard.co/1Eu3g2uC.png Jun 13 12:46:00 My mistake :P Jun 13 12:58:41 chanakya_vc: ok, so it can do captures, that's the first step Jun 13 12:59:25 chanakya_vc: (sorry haven't read your emails yet, got interrupted by hardware guys here) and you have your pin muxing setup to turn both the SPI clock and MOSI lines from the PRU intout GPIO outputs? Jun 13 13:00:15 I have bradfa . Jun 13 13:01:08 chanakya_vc: ok, and if you put your logic analyzer on the SPI output and set a capture on the rising edge of CLK with a sample rate of something like 24 MHz, you don't see anything? Jun 13 13:06:19 bradfa,I have got to run. Maybe we can discuss this in about 4 hours when you are done with your meetings? Jun 13 13:09:09 chanakya_vc: hopefully... although now I have a meeting from 12:30 till 2 EDT... but I will connect with you between all my meetings fun :) Jun 13 13:09:48 chanakya_vc: but just ask any questions you have and when I'm at my computer I'll write back Jun 13 13:09:54 or others can possibly help, too Jun 13 13:24:11 hi jic23 you there ? Jun 13 13:26:25 on phone Jun 13 13:37:54 back but doing a bit of assembly as can't find anyone else to do it today... Jun 13 13:51:55 jic23:, alexhiam: there? Jun 13 13:53:43 Kiran4399 - in and out. Looking for some screws... Jun 13 13:54:32 jic23: so.. what I have been doing is I developing mpu6050 driver myself.. can't wait for someone to get the patchset. :-( Jun 13 13:54:57 so.. I've seen from the android driver about some more sysfs entries relating to quatronians, dmp and so on.. Jun 13 13:55:10 sort of adding some code in the mpu_core.c Jun 13 13:55:20 some more attributes. Jun 13 13:55:30 jic23:^^ Jun 13 13:59:43 kiran - don't waste your time. Run the set I posted.. Will already work as Leonard has been working with precisely that part to test the driver.... Jun 13 13:59:52 There are so much more useful things to be doing.... Jun 13 14:00:03 Though obviously it might be interesting + you might learn stuff... Jun 13 14:00:43 Ok, the quarternion output might be useful... Is the hardware spitting that out with the normal firmware? Jun 13 14:00:58 okay a question on what you said, ans when you are free :) . So you said that i have "half" supported the multiple instances of the device, that "half" is because of the /dev/ file name ? as it will be same for the 2nd instance and thus would not be successfully created ? Jun 13 14:01:11 jic23: ^ Jun 13 14:01:50 jic32: yeah.. I guess so.. Jun 13 14:01:52 ZeekHuge, no half supported because the driver explicitly instantiates precisely one device. Jun 13 14:02:40 jic23: https://git.kernel.org/cgit/linux/kernel/git/jic23/iio.git/commit/?h=experimental-beaglefun Jun 13 14:02:40 kiran4399, hmm. datasheet is 'limited' on these... Jun 13 14:02:46 is this the link you posted? Jun 13 14:02:53 how do I get the patch? Jun 13 14:03:45 easiest will be to fetch that whole tree into your kernel git tree. Then do gitk -n 10 remotename/experimental-beaglefun. Jun 13 14:03:54 Then you can use gitk to cherry pick the relevant patches. Jun 13 14:04:14 Which is all the new stuff on top of rcn's tree other than that last patch which will stop the beagle running (probably) Jun 13 14:04:35 jic23: what are the improvements? Jun 13 14:04:44 support for the magnetometer. Jun 13 14:05:20 or technically support for any slave device that we have a driver for, but unless your breakout routes some fairly unusual pins out to the edge, you probably won't get any other slaves on the bus,,,# Jun 13 14:05:57 any extra attributes for mpu6050 added? like dmp and so on? Jun 13 14:06:15 nope. Jun 13 14:06:37 dmp stuff is to put it lightly basically proprietary and undocumented last time I looked anyway. Jun 13 14:06:50 plus all stuff better done in software on a linux board. Jun 13 14:07:23 Their fusion algorithms are likely to be simplistic for starters... Though who knows since they don't tell anyone what they are doing ;) Jun 13 14:08:06 If you can find the doc referenced : “Programming Sequence for DMP Hardware Functions.” Jun 13 14:08:23 that would be cool, but google hasn't found it for me... Jun 13 14:08:44 jic23: I am sorry.. but can you tell me how to clone a git tree? Jun 13 14:08:47 + last time IIRC anyone asking about this, they got a 'No way are we releasing that'. Jun 13 14:09:02 Do you have a current copy of the source for RCNs tree? Jun 13 14:09:16 If not you can grab the whole lot from my repo but it'll take longer :( Jun 13 14:09:17 umm.. 4.4 yeah.. Jun 13 14:09:19 jic23: Jun 13 14:09:36 Git checkout or a tar.gz type download? Jun 13 14:11:38 jic23: or can I simply copy your /drivers/iio/imu/inv_mpu6050 onto mine? Jun 13 14:12:04 kiran4399 - nope, do it right! Jun 13 14:12:17 So how did you get rcn's source? Jun 13 14:12:38 I cloned from beagleboard/linux Jun 13 14:12:47 cool - easy then. Jun 13 14:12:59 Open a terminal and change to the root of that tree. Jun 13 14:13:41 then type: git remote add iio git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git Jun 13 14:14:15 Which will add a git remote which is kind of instructions to git on where to get code from. beagleboard/linux will be your origin remote already. Jun 13 14:14:27 Then type git fetch iio Jun 13 14:14:57 and finally, gitk -n 10 iio/experimental-beaglefun Jun 13 14:15:10 (assuming you have a graphical setup?) Jun 13 14:15:28 graphical setup? Jun 13 14:15:33 jic23: ^^ Jun 13 14:15:33 then you can right click on the patch you want (starting at the oldest) and select cherry pick. Jun 13 14:15:47 As in can display a GUI program... Jun 13 14:15:50 (rather than terminal only) Jun 13 14:15:56 (seems likely ;) Jun 13 14:16:07 but you never know! Jun 13 14:16:44 jic23: do we have that gui for linux? Jun 13 14:16:50 The fetch will pull in a fair amount of objects to git as my tree contains a few additional versions of linux on the normal rcn one, but won't do any harm. Jun 13 14:16:59 yeah, gitk is part of most git installs on whatever OS. Jun 13 14:17:33 Other useful tool is git gui which is used to easily break up your patches into nice clean commits. Jun 13 14:18:53 back in a few mins - need to stick a pci card in pc. Jun 13 14:38:23 kiran4399: this page looks kinda familiar... https://github.com/kiran4399/bb_blue_api/wiki/PRU-Kernel-driver Jun 13 14:50:11 kiran4399: I mainly bring it up because you missed the most important line in my gist, which was crediting where that info originally came from Jun 13 14:50:18 i.e. https://groups.google.com/d/msg/beagleboard/od6h9yTKUD4/jzGE6KaxAQAJ Jun 13 15:25:19 alexhiam: sorry.. won't do that next time.. Jun 13 15:25:44 alexhiam: btw. there is no module named pru_rproc?? Jun 13 15:26:10 alexhiam: and currently working on applying the patchset for mpu devices.. Jun 13 15:26:18 kiran4399: there should be Jun 13 15:27:10 https://github.com/beagleboard/linux/blob/4.4/drivers/remoteproc/pru_rproc.h Jun 13 15:27:34 builds as a module in the bb.org_defconfig Jun 13 15:29:39 alexhiam: ok.. will try.. btw.. did you test mpu 9250? Jun 13 15:34:34 kiran439: hasn't arrived yet Jun 13 15:35:17 but I got the kernel built with those patches, so I'm ready to test when it does Jun 13 15:35:45 you mean the patches which jic23 suggested? Jun 13 15:35:54 yeah Jun 13 15:35:54 alexhiam: ^^ Jun 13 15:36:23 btw... bmp280 is connected to i2c2 right? Jun 13 15:36:33 yeah Jun 13 15:36:38 alexhiam: but even mpu9250 is connected to i2c2? Jun 13 15:36:46 yeah Jun 13 15:37:01 how? Jun 13 15:37:19 glc-bot: .w i2c Jun 13 15:37:25 :( Jun 13 15:37:29 https://en.wikipedia.org/wiki/I%C2%B2C Jun 13 15:37:30 [WIKIPEDIA] I²C | "I²C (Inter-Integrated Circuit), pronounced I-squared-C, is a multi-master, multi-slave, single-ended, serial computer bus invented by Philips Semiconductor (now NXP Semiconductors). It is typically used for attaching lower-speed peripheral ICs to processors and microcontrollers in short-distance, intra..." Jun 13 15:44:41 alexhiam: do miniature stuff like buttons, leds, dc motors work by kernel drivers? Jun 13 15:45:01 they can Jun 13 15:45:24 alexhiam: I mean in this project how do you want me to implement them? Jun 13 15:47:50 test Jun 13 15:48:42 kiran4399: the motors are all on PWM outputs, so that can stay as it is Jun 13 15:48:47 it's already using the PWM driver Jun 13 15:48:53 you mean ehrpwm? Jun 13 15:48:56 same or the GPIO stuff Jun 13 15:48:57 yeah Jun 13 15:57:44 alexhiam: for setting the led high or low.. strawson apis use mmap. Jun 13 15:57:57 alexhiam: do you want me to use the same thing? Jun 13 15:58:07 oops. That'll have to be changed to the gpio driver Jun 13 15:58:25 alexhiam: there is a gpio kernel driver ???? Jun 13 15:58:27 no mmap for any userspace hardware control Jun 13 15:58:31 yea.. Jun 13 15:58:47 google is your friend ;) Jun 13 16:04:01 alexhiam: can I use this? Jun 13 16:04:02 https://github.com/derekmolloy/exploringBB/blob/master/extras/kernel/button/button.c Jun 13 16:07:29 I don't see any reason not to just use the gpio driver Jun 13 16:08:17 alexhiam: I mean I don't know what do you mean by gpio driver.. Jun 13 16:08:29 there is no am335x-gpio.c in /drivers/gpio Jun 13 16:10:02 https://www.kernel.org/doc/Documentation/gpio/gpio.txt Jun 13 16:10:38 ok................I got it.. Jun 13 16:10:47 it's gpio-omap.c Jun 13 16:10:57 I remember doing it in my 2nd semester.. Jun 13 16:11:14 the OMAP processors have the same GPIO subsystem Jun 13 16:11:24 or at least close enough to all use that driver Jun 13 16:11:32 http://derekmolloy.ie/gpios-on-the-beaglebone-black-using-device-tree-overlays/ Jun 13 16:11:58 there's plenty of tutorials out there, just do some googling Jun 13 16:12:40 alexhiam: ya.. I knew these things when I was in my 1st year.. forgot now.. :-P Jun 13 16:13:58 alexhiam: I think I'll finish off the apis for led and switch by today.. Jun 13 16:14:54 sorry button.. Jun 13 16:17:00 kiran4399: before you change the apis, can you get it travis-ci setup with the repo? Jun 13 16:17:10 I set it up.. Jun 13 16:17:15 But I got a build failed.. Jun 13 16:17:28 may be because I did not set it properly.. XD Jun 13 16:17:39 Has anyone else also received the mail from TI? Jun 13 16:17:45 me.. Jun 13 16:17:53 Ohh.. Jun 13 16:17:55 chanakya_vc: Jun 13 16:17:58 do you know why? Jun 13 16:18:04 TI caring for us kiran4399 Jun 13 16:18:09 ZeekHuge, ^^? Jun 13 16:18:12 chanakya_vc:haha.. Jun 13 16:18:20 yeah.. even ZeekHuge got it.. Jun 13 16:18:24 yep ! me too .. Jun 13 16:18:27 kiran4399: your .travis.yml doesn't have any build instructions Jun 13 16:18:31 Although why they are sending us a beach ball Jun 13 16:18:34 Abhishek_: you there ? Jun 13 16:18:42 I don't get that ZeekHuge kiran4399 :P Jun 13 16:18:44 chanakya_vc: Yeah ! thats funny ! Jun 13 16:18:51 yeah.. we can play together if you all are here!! :-P Jun 13 16:19:08 Goa probably ! :P after project :P Jun 13 16:19:27 ZeekHuge, kiran4399 Lets all meet up in Goa? Jun 13 16:19:33 What do you say? Jun 13 16:19:36 when? Jun 13 16:19:37 nope ! just kidding ! Jun 13 16:19:45 ZeekHuge, Ohh :P Jun 13 16:19:51 are ! i was not serious bhai ! Jun 13 16:20:10 Got it ZeekHuge bro! Jun 13 16:20:31 ZeekHuge: khabhi tho phadayi chodo :-P Jun 13 16:20:38 Anyway,I was pleasantly surprised with the mail. Jun 13 16:20:51 Looks like TI is following our project Jun 13 16:20:59 projects Jun 13 16:21:39 Actually .. TI is benefited a lot with BBB projects and GSoC projects on it . So they are just trying to take care :) Jun 13 16:21:54 I think so. Jun 13 16:22:32 ds2: you there ? Jun 13 16:22:49 And I thought they didn't care.You know ZeekHuge Cathy Wicks,who has ordered it on our behalf is the head of TI's University program Jun 13 16:22:52 kiran4399, ^^ Jun 13 16:23:04 oh.. is it.. Jun 13 16:23:14 She visited our lab once. Jun 13 16:24:35 I think some mentor in BeagleBoard.org must have told TI. alexhiam ,_av500_ av500 any idea about this? Jun 13 16:25:19 Cathy has always been involved with / followed bb.org gsoc Jun 13 16:25:21 The links to the scope screen images that i clicked today : Jun 13 16:25:27 http://postimg.org/image/lhb0s3et3/ Jun 13 16:25:29 and Jun 13 16:25:38 http://postimg.org/image/6eysheu9j/ Jun 13 16:25:44 Abhishek_: ds2 ^^ Jun 13 16:26:27 So the voltage is under 3 V at freq of about 17MHz Jun 13 16:26:52 ZeekHuge: that's from the PRU? Jun 13 16:26:53 and that freq, was produced by using __delay_cycles(1) Jun 13 16:26:56 yes Jun 13 16:28:03 m_w: ^^ Jun 13 16:28:44 FYI you can get a more accurate frequency measurement by measuring the time between all those pulses then dividing by the number of pulses Jun 13 16:28:58 Ohh alexhiam That's great. Jun 13 16:29:23 TI should care.They do benefit from all these projects Jun 13 16:29:30 ZeekHuge: what's the bandwidth of the scope? Jun 13 16:30:18 oh and i have put the example driver to communicate to the pru up here :https://github.com/ZeekHuge/BeagleScope/tree/master/examples/kernel_examples/n-blinky Jun 13 16:30:38 alexhiam: its 100MHz. Jun 13 16:30:50 ZeekHuge, kiran4399 Lets hope they do remember us when we apply to TI for an internship/job :P Jun 13 16:30:55 ZeekHuge: what about the probes? Jun 13 16:31:45 you mean 10x and 1x ? Jun 13 16:31:52 alexhiam: ^ Jun 13 16:31:59 ZeekHuge: no, the bandwidth Jun 13 16:34:13 ZeekHuge: what I'm getting at is that it's likely 0-3.3V at the pin (assuming good routing), and you're seeing less on the scope screen because of some combination of bandwidth limit/probing (inductance)/out-of-cal scope vertical amp Jun 13 16:35:30 actually I dont have the scope here and i was trying to find it in images that i have , I wasnt able to find there ... Jun 13 16:36:33 that's fine. I'm just saying you should take that voltage measurement with a grain of salt Jun 13 16:56:50 alexhiam: What I could just see is that the inductance of input is 1Mohms||15pF Jun 13 16:57:50 alexhiam: http://postimg.org/image/sg9jrithd/ Jun 13 17:00:26 that's not the inductance, that's the scope's input characteristics. The inductance I'm talking about is a parasitic property of things like the wires in the probe Jun 13 17:00:50 in a perfect world wire's wouldn't have any inductance, but it ain't a perfect world Jun 13 17:02:55 yeah I meant characteristics :| Jun 13 17:07:12 * nerdboy has a green with dangle-y stuff attached Jun 13 17:08:30 a little disappointed they got rid of the barrel connector Jun 13 17:08:56 nerdboy: don't you know? you're supposed to power everything from usb these days Jun 13 17:11:57 uh huh, because it specs out at 500 mA Jun 13 17:13:53 alexhiam: tell me one thing.. Jun 13 17:14:30 alexhiam: the sysfs entries for gpio are already loaded in /sys/class/gpio right.. Jun 13 17:15:13 kiran4399: yeah Jun 13 17:15:34 alexhiam: so we can directly cat or echo the pin right? Jun 13 17:15:44 kiran4399: though the pins aren't necessarily muxed for gpio Jun 13 17:15:48 right Jun 13 17:15:57 alexhiam: so for this case we need not use any overlay right/ Jun 13 17:16:49 kiran4399: the base am335x-boneblue.dts will want to mux those pins for gpio, but you can just use config-pin for now if you want Jun 13 17:17:25 alexhiam: config-pin? Jun 13 17:17:40 alexhiam: I think that should be good.. Jun 13 17:17:54 becuase I am getting lot of conflicts in dmesg while loading the overlay.. Jun 13 17:17:54 config-pin -h Jun 13 17:18:21 kiran4399: are those conflicts documented on a wiki page? Jun 13 17:18:44 no... not yet.. Jun 13 17:19:44 alexhiam: http://pastebin.com/nRf8wnQx Jun 13 17:20:59 "pin 44e109ac.0 already requested by 48038000.mcasp" Jun 13 17:21:14 you'll need to disable HDMI video & audio Jun 13 17:23:24 alexhiam: I think it is already disabled.. Jun 13 17:23:25 http://pastebin.com/VApeWbE7 Jun 13 17:24:02 line 11 Jun 13 17:24:14 needs to be uncommented Jun 13 17:27:21 alexhiam: yeah.. will do it.. Jun 13 17:28:11 alexhiam: so what should i do to get correct measurements ? get new probes ? Jun 13 17:29:59 ZeekHuge: well the voltage doesn't really matter, as you know the GPIO is running at 3.3V. But if you do want to try and get a more accurate measurement, first check the bandwidth on your probes. If it's high enough, you can reduce the inductance by keeping all your wires as short as possible Jun 13 17:31:14 ZeekHuge: if you want to know more about that stuff, w2aew on youtube has some great videos Jun 13 17:31:17 https://www.youtube.com/watch?v=zodpCuxwn_o Jun 13 17:31:22 https://www.youtube.com/watch?v=-4q8geE5ef8 Jun 13 17:33:50 bradfa, I will be mailing you the ideas about miso tonight or tomorrow Jun 13 17:37:25 chanakya_vc: it might be worth adding a CS signal now, it'll make it much easier to debug the data if you have a CS to trigger on, and many SPI decoders require a CS (not sure about sigrok, might support a timeout also, but CS will make it easier regardless) Jun 13 17:37:43 it could just a hard coded pin for now Jun 13 17:39:09 Okay simply a gpio pin going low for now will do right? Jun 13 17:39:11 alexhiam, ^^ Jun 13 17:39:26 chanakya_vc: you'll want to use a PRU GPO Jun 13 17:39:41 but yeah, just go low before sending then high after Jun 13 17:39:41 Yes PRU GPO alexhiam Jun 13 17:40:11 eventually the pin and the active state will want to both be configurable, but active low is fine for now Jun 13 17:40:54 so you'd just need to add 2 lines Jun 13 17:41:17 Okay got it. alexhiam. Jun 13 17:43:05 alexhiam, I had a doubt yesterday. Regarding the addresses of the shared data ram.Table 4 in the ref guide shows that the address of instruction RAM is 0x0000.While this is also the address of the the internal data ram of 8kb(in table 5) Jun 13 17:43:25 of the PRU? How is this possible? Jun 13 17:45:31 alexhiam, Also,wanted to ask that I would want to use the shared RAM of 12kb or these internal data RAM's?What would you suggest? Jun 13 17:45:37 chanakya_vc: those are difference memory maps. And 'data access' will use the data memory map Jun 13 17:46:57 alexhiam, Okay so 0X0000_0000 of the instruction RAM is not the same as that of the data RAM? Jun 13 17:47:06 right Jun 13 17:47:07 hmmnn got it. Jun 13 17:48:44 I'm not actually sure if the stack is put in that 8kB instruction ram or in the data ram.... Jun 13 17:49:13 that's up to the compiler Jun 13 17:50:24 But variable memory will be in the data RAM only right?alexhiam? Jun 13 17:51:07 depends Jun 13 17:51:21 http://www.ti.com/lit/pdf/spruhv7 should have the answers... Jun 13 17:51:39 Instruction RAM would contain the firmware code I guess because there is a limit to the code size that we can upload I guess Jun 13 17:52:05 alexhiam, ^^ Should come because of the limited size of the IRAM? Jun 13 17:52:47 right, the firmware goes in that IRAM Jun 13 17:53:44 Pg 99 contains stuff about memory Jun 13 17:53:45 alexhiam, Jun 13 17:56:07 I'm guessing the stack is at the end of the IRAM Jun 13 17:58:05 anyone else here making a robot and can tolerate the 96borads? Jun 13 18:01:13 alexhiam, If that's the case then I put at stuff at random mem addresses,wouldn't that cause a problem? Jun 13 18:01:55 chanakya_vc: nvm, the compiler is using the DRAM: http://git.ti.com/pru-software-support-package/pru-software-support-package/blobs/master/examples/am335x/PRU_access_const_table/AM335x_PRU.cmd#line73 Jun 13 18:02:23 chanakya_vc: so only the 12kB shared ram is safe to use Jun 13 18:02:29 e.g. https://e2e.ti.com/support/arm/sitara_arm/f/791/p/485199/1752049#1752049 Jun 13 18:04:43 So data Ram it is..alexhiam just clear a fundamental doubt of mine. So any variable I declare or dynamically allocate is stored in this stack?.But if I declare a pointer to a certain mem,would it also come in this stack? Jun 13 18:06:18 I mean I am asking that addresses are allocated to a variable are in a stack,or does the stack store the variable name with the random address allotted to it? Jun 13 18:07:18 chanakya_vc: the 12kB shared data RAM, that is. If you declare a pointer locally in a function, that pointer (4 bytes because it's a 32-bit architecture) will be stored in the stack Jun 13 18:07:47 https://mcuoneclipse.com/2013/04/14/text-data-and-bss-code-and-data-size-explained/ Jun 13 18:13:54 Okay alexhiam .But I donot understand this variable stack.Does it simply contain all the initialized variables along with the addresses to where there values are stored? Jun 13 18:14:36 Or this stack is the place where the values of the variables are stored? Jun 13 18:14:52 along with its name? Jun 13 18:16:39 chanakya_vc: variable names aren't stored, they just let the compiler know what chunk of memory you're talking about. The stack stores anything that is dynamically allocated. e.g. any time a function is called, the memory required for that function (loval variables, arguments, return values, etc.) is allocated on the stack Jun 13 18:17:04 unless malloc is used, in which case it is put in the heap (but don't use malloc on the PRU) Jun 13 18:18:26 'dynamically allocated' wasn't really the right term, I mean allocated at run time, the compiler knows how much memory to use for each function Jun 13 18:18:42 alexhiam, Okay so if I said int a=6 and then int *b=&a,b would point to a mem location on the stack? Jun 13 18:18:54 yup Jun 13 18:20:36 Okay so all memory allocated at run time is from this stack?And this stack exists in the Data RAM in case of PRU right? Jun 13 18:22:47 if you declare a variable at the global level it gets put in the .data section, e.g.: http://git.ti.com/pru-software-support-package/pru-software-support-package/blobs/master/examples/am335x/PRU_RPMsg_Echo_Interrupt0/main.c#line69 Jun 13 18:23:30 alexhiam, In the above link,the .text part is stored in Flash mem and the initialized variable bit goes in RAM Jun 13 18:23:47 So the satck there is in the RAM right? Jun 13 18:24:05 yeah, flash is way to slow for a stack Jun 13 18:24:20 Although it says the that the initialization value goes into FLASH Jun 13 18:24:44 but it has the benefit of being non-volatile, so you can store your instructions there and they won't disappear when you remove the power Jun 13 18:25:50 right, so when the program runs it can grab that initialization value and copy it to ram Jun 13 18:27:19 Ohh.Got it alexhiam : ) Jun 13 18:29:54 Just one more doubt regarding this,when our binary is just sitting there i.e not being executed,its still stored the same way? Jun 13 18:31:07 alexhiam, Like we see one file at a single loc but its actually stored between RAM and Flash.I imagine it would contain directions to find this variable stack? Jun 13 18:31:40 that .cmd file is the instructions Jun 13 18:33:02 But we don't load that to the PRU? We load the .out right? Jun 13 18:33:27 Or I have been trying to load the wrong file so far :P Jun 13 18:33:29 alexhiam, ^^ Jun 13 18:35:26 right, the cmd file is used by the linker. The finished binary is essentially the entire PRU memory map Jun 13 18:35:34 IRAM and DRAM Jun 13 18:35:46 er, data RAM that is Jun 13 18:38:14 can we disable hdmi cape on green? Jun 13 18:38:26 i only see 3 .dtb files... Jun 13 18:38:48 Okay alexhiam .Got it.Next I have to figure out a way to let the arm talk to the shared mem. Jun 13 18:39:56 anybody know what am335x-bonegreen-wireless.dtb does? Jun 13 18:40:17 nerdboy: that's for the new green wireless Jun 13 18:40:24 has a wifi/bt4 chipset Jun 13 18:42:41 nerdboy: hmm, seems like the green might not use hdmi by default? Jun 13 18:43:05 oh right, which makes sense because it doesn't have an hdmi connecter Jun 13 18:43:15 connectOr Jun 13 18:46:08 alexhiam, Here in the am335x guide there is the concept of Global addresses.On page 20 there is table 6 which gives offset addresses(same as local address for PRU-ICSS).So all the ARM has to do is call Address of PRU-ICSS+Offset address? Jun 13 18:47:10 you'll want to use dma Jun 13 18:47:20 alexhiam: okay so you were suggesting that the probe does not has high enough bandwidth and it was attenuating higher harmonics of the signal . An that is what resulting in that shape . Jun 13 18:47:24 right ? Jun 13 18:47:35 *and Jun 13 18:48:41 I'll check the probes tomorrow. Jun 13 18:50:10 apparently they left that off too Jun 13 18:50:19 didn't even notice... Jun 13 18:50:20 alexhiam, Okay I will to research on that a bit.I thought all I had to do was to use offset addresses. Jun 13 18:50:37 ZeekHuge: well, if the probe has too low of a bandwidth that would lead to some attenuation. The ringing (where it bounces above/below on the rising/falling edge and tapers off at a higher freq) is likely caused by the length of the ground lead (longer lead means more inductance) Jun 13 18:51:21 ZeekHuge: you also want to make sure your probes are compensated for the scope inputs Jun 13 18:51:33 yep saw that video and read some text. Jun 13 18:52:08 yeah . And that will need a sig generator. Jun 13 18:52:31 ZeekHuge: the scope will have a probe compensation output Jun 13 18:52:56 square wave, often @ 1kHz Jun 13 18:52:58 do i need to mess with capemgr on green? Jun 13 18:53:22 chanakya_vc: on the Linux side, your program/driver will use virtual memory, so you can't just create a pointer like that Jun 13 18:53:56 nerdboy: if you just want to mux pins, it should have the universal-io overlay loaded already and you can use config-pin Jun 13 18:53:57 alexhiam: can you see that in the image ? its an old model, I dont think so it has one. Jun 13 18:55:06 ZeekHuge: I don't see it, but it should have one. Maybe under the screen or on the side Jun 13 18:55:38 oh yes ! I remember it has some ports there .. Jun 13 18:55:41 Okay alexhiam .I got it.I will read up on DMA a bit. Thanks for your help! Jun 13 18:55:50 np Jun 13 18:56:04 hey bradfa are you free ? Jun 13 18:56:41 kiran4399: you got the driver working with the new patches? Jun 13 18:57:48 alexhiam: a question, it may sound funny .. how do you know so much ? Jun 13 18:58:06 okay . dont ans that if you dont want to :) Jun 13 18:59:40 ZeekHuge, Haha,Even I wonder that sometimes.I feel sometimes that I learn more from alexhiam than I do from my profs Jun 13 19:00:21 I just get fascinated by so much of knowledge everyone here has .. And we students are always like "what is this ? " & "what is that ?" .. Jun 13 19:00:24