**** BEGIN LOGGING AT Thu Feb 06 02:59:57 2020 Feb 06 03:01:13 (14 channels when running from SD card, reduced to 12 channels when using eMMC) Feb 06 03:05:41 hmm, maybe better to just limit it to 12 channels anyway... those last two channels are eMMC cmd and clk, so if you happen to toggle in a valid command by accident the eMMC will drive a response onto the cmd line. unless you hold the eMMC in reset, which requires enabling its reset pin in its OTP configuration Feb 06 03:08:03 That doesn't just sound messy... Feb 06 03:09:39 I'd prefer to boot from eMMC anyway Feb 06 03:16:56 Hey, should I need internet access to get my drivers installed on the BBBW or are they already programmed to install when plugged in? Feb 06 03:17:08 what drivers? Feb 06 03:17:20 The USB driver Feb 06 03:17:35 host driver Feb 06 03:17:55 ah that's not what you said (those are not "installed on the BBBW") Feb 06 03:17:59 anyway, you don't need any Feb 06 03:18:04 Hmm. Feb 06 03:18:06 Okay. Feb 06 03:18:25 That is what I thought. The driver is already implemented into the BBBW. Feb 06 03:18:27 just run Linux Feb 06 03:18:43 The people had Win OS systems at the "makerspace." Feb 06 03:18:52 set_: no, appropriate drivers are part of each relevant host OS Feb 06 03:19:04 Oh. Okay. Feb 06 03:19:30 So, it used to be just for Win OS and Linux was covered. Got it. Now, both should be covered. Feb 06 03:19:40 B/c...you guys made your own driver. Feb 06 03:20:41 I plugged the BBBW in and it said on Win 10, makerspace set up, drivers not installed. Feb 06 03:20:56 USB networking uses either RNDIS or CDC-ECM (the BBB supports both). CDC-ECM is the usb standard protocol for it, and is implemented by Linux and Mac. Microsoft doesn't give a fuck about standards and instead uses their custom RNDIS protocol, which is also supported by Linux. Feb 06 03:20:58 I did not know what that meant. Feb 06 03:21:09 was it running an uptodate image? Feb 06 03:21:16 Yes. Feb 06 03:21:25 I did not have internet access, though. Feb 06 03:21:59 it has traditionally been notoriously difficult for the BBB to convince Windows to use their own fucking drivers for their own fucking proprietary RNDIS protocol Feb 06 03:22:11 Oh. Okay. Feb 06 03:22:26 and Microsoft periodically breaks it somehow, requiring the USB descriptors to be tweaked Feb 06 03:22:34 which is why it's important to be uptodate Feb 06 03:22:35 I was wondering. I thought those people were messing w/ me. But, get this... Feb 06 03:22:44 when it is, it *should* work plug and play Feb 06 03:23:21 I opened, I know you do not use it, Cloud9 IDE w/ 192.168.7.2 into the address bar. Feb 06 03:23:48 did it work? Feb 06 03:23:52 I could get into my system and use it but things were odd. Odd equals my source changing in front of my eyes. Feb 06 03:24:11 if you can get into your system, then that means you have all the drivers you need Feb 06 03:24:19 I mean it. I would type the number two, i.e. 2. And, then it would erase specific lines. Feb 06 03:24:24 I got home to test it. Feb 06 03:24:35 I have no idea what you could possibly be talking about Feb 06 03:24:37 To make sure it was not the image or me. Feb 06 03:24:45 It works fine on my computer. Feb 06 03:24:54 it was probably you Feb 06 03:24:55 But not at the makerspace i went to today. Feb 06 03:25:14 OH well. All is well that ends well. Feb 06 03:28:32 if drivers were missing, then going to 192.168.7.2 in the web browser would just result in "This site can't be reached" Feb 06 03:33:52 Okay. Feb 06 03:34:17 I do not know why the machine I was using stated on a pop up, "Drivers not installed." Feb 06 03:34:26 And... Feb 06 03:34:46 I have no clue why the machine allowed me to get to the online server. Feb 06 03:35:23 The only thing I can think of, since it is a BBBW, my system was not updated/upgraded and w/out wifi. Feb 06 03:36:39 I mean this too. Do not get me wrong. I am not complaining. I just thought you or the bbb.io peoples should know. Feb 06 03:37:12 In case others were upset over this idea, it is most likely firewall/password issues w/ wifi/internet access. Feb 06 03:37:45 ? Feb 06 03:38:11 the driver warning might have been about the CDC-ECM interface I guess. regardless, it was of no importance since usb networking was working for you Feb 06 03:38:34 so the only driver you care about was clearly functional Feb 06 03:39:00 But...@zmatt: You should have seen the issues that arose while I tried to run source and edit the source. Feb 06 03:39:10 hmmm need to check old BBB's moron fun for the mentally impaired. Feb 06 03:39:17 set_: the fact you could run or edit source *at all* means the driver was working Feb 06 03:39:32 Lines and characters were disappearing and okay. Forget it. Feb 06 03:39:56 I have no idea what you were experiencing or why, but it had nothing to do with drivers Feb 06 03:40:04 Okay. Feb 06 03:40:13 As long as you can admit it, fine. Feb 06 03:41:30 well I cannot magically look over your shoulder across time and space, and your problem description is so incredibly vague that it's completely impossible to form any sort of opinion about whether it was a real problem or just you Feb 06 03:42:10 Trust me, finally, just trust me. It was not me this time. Oh and i know. It is not your issue. It is not even an issue. I just thought you should know. Feb 06 03:42:39 Forget I brought it up. Sheesh. Feb 06 03:47:06 hey set_ if you get frustrated take a breather, that saved me destroying a lot of things. Feb 06 03:47:27 GenTooMan: I am clam. Clam! Feb 06 15:55:41 okay, tried re-breadboarding the working PB, and I can get low-speed USB devices to register, but not high-speed ones, either no response or similar power-related errors as before Feb 06 15:55:54 so signal integrity is also an issue when USB1 is working Feb 06 15:56:08 I'll post my notes to the forum, thanks again Feb 06 16:13:51 for high-speed usb, signal integrity can easily be an issue with a pcb design, let alone a breadboard Feb 06 16:14:44 keep in mind that high-speed usb is nearly half a gigabit/s Feb 06 16:30:48 thus it acts more like a controlled transmission line than "digital"? Feb 06 16:33:26 I mean, all "digital" signals are analog, and all traces are transmission lines... but the extent to which you can ignore this reality in favor of a pleasant fantasy depends on trace length and signal rise/fall time Feb 06 17:04:33 well yes... a lot of people love pleasant fantasy except when they aren't. LOL Feb 06 20:44:38 Has anyone successfully muxed the pins on the BeagleBone AI? AI to enable Feb 06 20:55:14 Hmm. Feb 06 20:55:31 Me, no but others may have set up muxing so far. Feb 06 20:55:47 JoePo: How far along the task are you? Feb 06 20:57:33 They have this fellow. He has muxing info. for specific pins. he might show up soon. Feb 06 20:58:37 https://docs.google.com/spreadsheets/d/1mSqEpV_BAUHfeNApytxHcGhgTZwypy564GyOr66Nphs/edit#gid=854495639 Feb 06 20:58:51 That is from @zmatt (I am pretty sure). Feb 06 21:03:25 JoePo: I just plugged in. If you want to do some source and hardware stuff, let a brother know. Feb 06 21:04:13 Is the google groups page giving anyone else trouble these days? I cannot get to that page for some reason. Feb 06 21:08:01 JoePo: Go to /var/lib/cloud9/BeagleBone/AI/ Feb 06 21:08:13 There is some source I found on the board. Feb 06 21:10:45 blinkLED.js has a header pin for use. P9_15! Feb 06 21:14:25 Check out analogInOut.js. It is a nifty set up and has two muxed pins for use. Feb 06 21:15:31 Thanks, ill have a look at it Feb 06 21:16:22 You are welcome. Feb 06 21:17:24 It may take some testing to see what is available w/out config-pin thus far. I know some people love .dts files. They may come in handy w/ this board so far. Feb 06 21:19:28 My AI dir is empty besides the pru dir. Do i need to update my examples? Feb 06 21:20:16 I've tried to create a .dts but how would i know which Label SPI has? Feb 06 21:23:38 Yes to your first question. Feb 06 21:23:59 Let me go and see how to update it w/out the ole update/upgrade w/ apt. Feb 06 21:25:01 Oh. SPI, I see it is referenced in that docs.google link I posted from @zmatt. Feb 06 21:25:58 https://beagleboard.org/upgrade is the upgrade page for the AI. Feb 06 21:27:34 JoePo: That question you asked about the SPI and how to label it in the .dts file is a good question. I might be able to hash something together. I think BeagleBoard.org persons may have put something together on their github page. Feb 06 21:30:01 I might need to check the TRM. This will be an adventure. Do you have about three weeks? Feb 06 21:30:13 Just kidding and not really kidding. Feb 06 21:35:56 p. 5930 is the start of it all. It may take longer than three weeks w/ me. Feb 06 21:36:09 JoePo: Do you have time? Feb 06 21:37:18 Right now or the three weeks? Feb 06 21:39:27 Right now and "forever," would be nice but now is better than ever. Feb 06 21:40:08 Okay. So, I found some McSPIm modules. m in McSPIm stands for 1 through 4. There are four of them on a single am57xx. Feb 06 21:42:43 Not really as its already late and I have to get up early. Feb 06 21:42:49 So, let us say we use McSPI1 (for instance). We will need >>> spi1_sclk, spi1_d[0], spi1_d[1] (Where 0 and 1 depend on configuration), and... Feb 06 21:42:51 Oh. Feb 06 21:42:52 Okay. Feb 06 21:43:04 No issue. Feb 06 21:43:12 When you come back, we can start again. Feb 06 21:43:36 Thanks for the help! Feb 06 21:43:37 <<< this is me (most of the time). Feb 06 21:43:39 Yep. Feb 06 21:43:59 Really appreciate it and wish you a nice evening Feb 06 21:44:12 Have fun and geaux Cajuns! Feb 06 22:24:35 Hey, what LCD are you guys planning on implementing into the AI construction? Feb 06 22:30:12 construction, cough. I meant development. Feb 06 22:57:31 A little late set_ Feb 06 23:01:20 I know. I goof up in my speech. Feb 06 23:02:21 My words are from '89 and my education is only so old. Mix that w/ this and blah...construction comes out. Feb 06 23:02:53 GenTooMan: Do you know about the LCD the bbb.io folks are putting together. Feb 06 23:04:16 Last but not least, the "makerspace" lady said I cannot use their computers any longer w/ my BBBW. Is that odd or what? Feb 06 23:04:58 "Blame it on the higher-ups." That usually works. Feb 06 23:08:09 MCSPI_xxxx[x] is a bunch of complicated there and back. There are some rules, notes, regulation, and order. It is not easy. Feb 06 23:08:33 I am glad I did not tell JoePo about the ease of it all. That would have been a lie. Feb 06 23:14:02 It is like UART, GPIO, and PWM are all added into one MCSPI_xxxx[x] pin. Feb 06 23:14:59 What device module is on the AI for SPI? Feb 06 23:38:18 I'm not sure as I have not looked at the AM57xx Feb 06 23:39:24 unfortunately lots of things may be lagging in documentation Feb 06 23:39:46 Okay. No issue. I just thought from the TRM, it would tell someone instead of just reference it casually. Feb 06 23:40:26 I think the bbai has three spi ports available on its expansion headers Feb 06 23:40:42 But. That chip has four accessible, right? Feb 06 23:40:53 The AI has only three, got it. Feb 06 23:41:27 In my book, that is two too many. W/ one, you can do all sorts of slave/master/slave configs. Feb 06 23:41:36 yeah, 1 out of 4 has no usable pinout on the AI (2 out of 5 if you include qspi) Feb 06 23:42:07 Oh. Is that normal b/c of all the other available ports? Feb 06 23:42:38 I have not read about the qspi at all yet. Feb 06 23:43:09 I got sidetracked in the source code for irq somehow. Feb 06 23:43:10 I mean, it's like that on nearly every embedded systems... there are more peripherals than you can use simultaneously, everything is heavily pinmuxed, and not all processor pins are available on the expansion headers (a whole bunch are used for on-board stuff obviously) Feb 06 23:43:49 there's more stuff on the BBAI than on the BBB, and more mux modes per pin, but conceptually it's the same Feb 06 23:44:08 the part that's really annoying is that glitch-free I/O configuration can only be done in u-boot Feb 06 23:44:19 Right. I understood and understand. I just figured w/ the amount of processors on this SoC, there would have been more pins available. Feb 06 23:44:19 Oh! Feb 06 23:44:29 @zmatt: That is nice to know. Feb 06 23:44:38 I know nothing about u-boot. Lucky me. Feb 06 23:44:49 the number of cores and the amount of I/O have no reason to be related to each other Feb 06 23:44:58 but the AM572x *does* have a ton more pins Feb 06 23:45:23 Right, I saw that idea to be true. It is a larger chip w/ more balls. Feb 06 23:45:50 Options galore! Feb 06 23:47:27 does it have the same annoying power options as the AM335x? Feb 06 23:47:49 The TPSxxxxx? No, it is an updated chip. Feb 06 23:49:00 Or...do you mean for the pins and the power options on those pins? Feb 06 23:50:43 So, those u-boot settings that should be handled need to be set up before boot, in u-boot, w/ what is available in the TRM? Feb 06 23:51:10 GenTooMan: the BBAI does not resemble the BBB in any way, every chip is different Feb 06 23:51:34 it's just the same form factor and in theory a somewhat compatible expansion header Feb 06 23:51:42 except physically.... :D Feb 06 23:51:49 That's probably good. There were things that made the BBB design... rather annoying. Feb 06 23:51:50 (if you manage to get your pins configured) Feb 06 23:51:58 Right, w/ a ton more configurable pins. Feb 06 23:51:59 Ha. Feb 06 23:52:53 I know two GPIO pins available so far. I really would like to be a person that blows the lid off this bugger. But, u-boot. Feb 06 23:53:29 I think most pins are configured as gpio by default Feb 06 23:54:18 Is bootlin the people still mfg. u-boot? Feb 06 23:56:01 or is u-boot being, for the BBB/BBAI, taken over by rcn? Feb 06 23:57:34 Oh. Denx? Feb 07 00:00:54 Anyway, I better figure this out before my head explodes (GenTooMan). Feb 07 00:05:28 Boomo! Feb 07 00:25:22 * GenTooMan gets the mop out. Feb 07 00:28:13 I am okay. They just released one, a newer version compared to what I was reading. Feb 07 00:34:15 GenTooMan: Have you ever made a patch for BBBs for u-boot? Feb 07 00:34:48 I just looked and this is why I am asking. It seems it would take more than one GenTooMan for that task. Feb 07 00:50:32 I'm uncertain what you mean, but O-tay. Feb 07 00:51:37 patching it is pretty straight forward Feb 07 00:52:10 esp if all you are changing is the pinmux... IIRC - it is a .h file. edit, rebuild, install, and enjoy the working results :D Feb 07 00:54:42 Oh. It seemed long and long. I guess I could see it another way. Feb 07 01:03:37 it isn't much more difficult then the kernel Feb 07 01:04:39 Oh. So, I am still reading. The McSPI module stuff is not familiar and it is fun, double-o, fun. Feb 07 01:05:20 So, to program, I would need to initialize global modules somehow. Still learning here. Feb 07 01:06:51 Then, off to operational modes which is configured by MCSPI_CHxCONF. Feb 07 01:09:23 So, after that, I go to Common Transfer Sequence to pick my module mode(s). Feb 07 01:10:22 Oh and ds2: I saw the kernel .h files and .asm files. In one small batch of source, there are like 25 plus preprocessor directives. Feb 07 01:11:06 for instance: https://github.com/torvalds/linux/blob/master/arch/arm/kernel/irq.c. Feb 07 01:12:13 There is no way to learn all that source and use those directives in one single file. I think this is why outsourcing is so gained. Feb 07 01:12:33 nothing too complex there Feb 07 01:15:09 Yea right, ds2. I guess in time, 10 plus years, one could fathom all that info. I mean, all those people have been coming and going on the same idea since '92. There is no way to catch up. Feb 07 01:18:02 @zmatt: What are "OCP accesses?" Feb 07 01:23:00 Are they just ports? Feb 07 01:28:04 how can i find the hw address of PCRM_CM_PER_SPI1_CLKCTRL? Feb 07 01:28:23 I guess I should google more. Dang it. Feb 07 01:29:09 TRM might help a lot Feb 07 01:29:52 also depends on how the headers are constructed for the processor specific defines. Feb 07 01:30:00 Aw. Feb 07 01:30:06 Thank you, mighty sir. Feb 07 01:30:30 I need to find the pru_cfg.h and "resource_table_empty.h" files. Sheesh. Off to look. Feb 07 01:31:25 Are people working on making these file updates or are we supposed to use older models of the files? Feb 07 01:35:06 depends on who maintains them. Feb 07 01:35:49 for example TI likely maintains the AM572X specific header files but all else is probably someone else. Feb 07 01:37:03 If you grab Code Composer Studio you can get the processor specific headers but not easily. You have to browse to where TI squirrelled them away. Feb 07 01:37:39 Oh. Feb 07 01:37:40 Okay. Feb 07 01:38:18 TI it is. Feb 07 01:43:16 hint You need to grab the file specific to the series of processor. Don't be surprised if you are inundated with a huge pile of headers. Feb 07 01:45:01 Okay. I just grabbed pru_cfg.h and "resource_table_empty.h" but I think they are for the am335x. So, off to CCS. Awesome! Feb 07 01:45:58 And yes. I typed that out in chat. Feb 07 01:49:41 I updated the CCS by accident. Now, time to wait. Hey, GenTooMan, do I go to the application store on CCS or do I do it another way? Feb 07 01:57:02 Okay. I found a way to get info. again. am5792, right? Feb 07 01:58:56 5729, sorry. Feb 07 02:02:03 are you using windows or linux? Feb 07 02:07:24 Windows. Feb 07 02:08:10 I just found a dude on Element14 would seemingly used these two files in his/her PRU usage. I just cannot figure out where in the world he/she got them. Feb 07 02:08:31 ... Feb 07 02:08:36 That almost made sense. Sorry. Feb 07 02:08:36 ...! Feb 07 02:08:43 ? Feb 07 02:08:51 would = who Feb 07 02:10:02 Yea. I use this chat on the Win OS computer. Feb 07 02:10:08 anyhow for TI the path is nominally \TI\ Feb 07 02:10:14 Okay. Feb 07 02:10:50 I have an arm compiler and a pru compiler available. Feb 07 02:10:58 That is about it, though. Feb 07 02:11:19 are you looking for PRU or ARM? the PRU might be different per processor. Feb 07 02:12:29 Oh. I have v9 of CCS. Does that matter? Feb 07 02:12:41 I am looking for PRU. Feb 07 02:12:44 That should be the latest... Feb 07 02:12:56 Yea. I did not even have to sign in like regular. Feb 07 02:13:07 They are just throwing it out like garbage. Feb 07 02:13:21 no calls, no emails, no nothing. Just CCS! Feb 07 02:14:21 I have the Linux tools installed and the PRU compiler tools installed. Feb 07 02:15:08 It cannot locate my board for some reason and they have no chip info. for am5729. Feb 07 02:16:48 ewwwwwwww CCS Feb 07 02:17:21 Hey GenTooMan and ds2: Do you guys know where I can find the pru_cfg.h file for the am5729? Feb 07 02:17:29 Or... Feb 07 02:17:33 Do I need to write it up? Feb 07 02:17:45 what's the pru_cfg? Feb 07 02:17:56 Header file. Feb 07 02:18:21 It is a config file for the PRU chips. Feb 07 02:25:34 is it labeled pru_math.h? Feb 07 02:25:52 No sir. Feb 07 02:25:56 pru_cfg.h Feb 07 02:26:22 it might be generated. Feb 07 02:26:58 headerfile for what? Feb 07 02:27:02 I can show you the file for the am335x. Maybe I should delve into pru stuff on the am5729 and change the am335x into the am5729 for this specific .h file. Feb 07 02:27:04 PRU. Feb 07 02:27:05 I doubt it is a kernel header Feb 07 02:27:10 I think it is. Feb 07 02:27:12 waht software package!? Feb 07 02:27:17 pru Feb 07 02:27:24 Please hold. I will get it. Feb 07 02:27:45 if its bare metal, you don't need them to provide you with one... that should be in the docs Feb 07 02:28:36 https://github.com/jadonk/pru-software-support-package/blob/master/include/pru_cfg.h Feb 07 02:29:10 I suspect that one will work Feb 07 02:29:36 the PRU is just a block... IIRC - that doesn't have hard coded absolute addresses so... Feb 07 02:29:49 Okay. Feb 07 02:30:24 I keep getting the same error: cannot find pru_cfg.h. I guess b/c it is one of my preprocessor directives. Feb 07 02:32:58 It actually says in the error: No such file or directory. Feb 07 02:34:52 did it say where it was looking for it? Also ... what are you trying to compile? Feb 07 02:35:41 I am cp the files to /usr/local/include/ now. Feb 07 02:36:11 Oops. Feb 07 02:37:00 now rsc_types.h. Geesh. I do not even use this as a directive. Feb 07 02:43:39 Oh well. There are some files that cannot be found thus far. Feb 07 02:45:26 I got the file to compile and run but it went into "Segmentation fault" status. Feb 07 02:47:05 My "master" port, the OCP port, has some source but I think it is incorrect for the am5729. It is for reading external memories, supposedly. Feb 07 02:52:17 Okay so. I found the register summary. This is food-good but my pru_cfg.h and "resource_table_empty.h" is jacked up. I need to alter them to suit my needs to test for the McSPI1 module. Feb 07 02:53:34 and I need to actually order my scope / logic analyser like I need to. So I have one, available sooner than never. :D Feb 07 02:57:21 Scopes and PRUs/McSPI did it! **** ENDING LOGGING AT Fri Feb 07 02:59:57 2020