**** BEGIN LOGGING AT Sun May 22 02:59:59 2016 May 22 05:34:36 Hello nerdboy ! May 22 05:42:30 hi ZeekHuge, you have access to system clock at msec/microsec resolution May 22 05:42:48 *with rtc May 22 05:43:10 probably with ntp May 22 05:44:03 ahh .. okay. So how to use it for the ADC ? May 22 05:45:07 system time libc? May 22 05:45:26 or python May 22 05:46:20 start with ntp and use at least 3 independent sources May 22 05:46:21 clock shouldn't be freerunning independent of the datalines though, how would he know when to sample May 22 05:46:48 when he asked about clock, he meant as sampling input not a time-source May 22 05:47:45 don't you get unix epoch, which correlates to external "time" May 22 05:48:04 in the PRU? no May 22 05:48:08 ntp gives valid ref clock May 22 05:48:19 no, outside the pru May 22 05:48:40 pru has fixed clock rate May 22 05:48:42 but it's the PRU that is doing the sampling and needs to have a handle on the timing May 22 05:49:24 sample on each cycle of clock, on the correct edge May 22 05:49:34 it has to come through an interface then May 22 05:50:08 also depends on how much latency and jitter, etc May 22 05:50:48 you could do some test code to count clock cycles May 22 05:51:07 the PRU is supposed to be good at consistent timing via instruction counting, it should be straight forward to implement a decent clock signal May 22 05:52:00 someone else will need to advise on the electrical connections for the clock pin though... May 22 05:52:06 you just want a clock pulse? May 22 05:52:10 yep May 22 05:53:32 if it's a good source it should correlate pretty well with external reference clock May 22 05:54:23 What i am confused about it, how to use that simple pulse train given out by the pru, to drive the input clock of this EVM http://cds.linear.com/docs/en/demo-board-manual/dc782A.pdf May 22 05:55:05 yeah we'll have to wait for hw mentor to chime in on the electrical connections May 22 05:55:41 Since, the max input limit at that clock input it 2V and we also need to terminate the transmission line. May 22 05:58:06 so, a voltage divider as a Double termination. May 22 05:59:16 like this http://www.marvintest.com/downloads/knowledgebase/Q200196%5CFig7b.JPG May 22 05:59:28 any suggestion for that ? May 22 06:00:43 sorry I do zero hw design, all I can do is occasional soldering based on someone else's design May 22 06:01:07 :) May 22 06:01:51 I kind of liked hw design in school, but it has been too long to be of use May 22 06:02:45 maybe you should ask about FFT's instead, at least I did that in grad school ;) May 22 06:04:28 Okay so lets discuss about software. Wormo : The coding period starts in about 20 hours .. so.. May 22 06:04:37 How to go then ? May 22 06:05:08 oh wow, finally the real thing :) May 22 06:05:50 I havent received the hardware yet, so no hardware things for now. I will first try to get on the remoteproc kernel and find out a way to communicate to it ? May 22 06:06:57 yes that part doesn't need the ADC, perfectly good place to start May 22 06:08:30 Did you get the earlier Beaglelogic working with the current kernel? May 22 06:08:49 okay,has it been shipped ? May 22 06:09:16 not yet. have to try it . With the latest kernel. May 22 06:09:58 I am not sure if it shipped yet, hopefully yes -- Digikey showed EVM as in stock May 22 06:11:22 nerdboy: do you know about the shipping of the ADC ? May 22 06:12:18 according to this post : https://groups.google.com/forum/#!topic/beagleboard-gsoc/lns9QnPQKnk, i will be using the iot kernel image. May 22 06:14:39 Have you dl that image yet? May 22 06:14:51 Wormo: Are these timing okay for you ? like, I was here earlier this day and now ? May 22 06:15:24 This is a pretty good time, too bad most of the other mentors are not here though May 22 06:17:13 just started downloading, the image at that location keeps on changing very frequently, so I was just waiting for the latter time to get the latest image. May 22 06:19:16 I think, ds2 is busy, he was not here earlier today either. And I think that he is usually busy during weekends. I am not sure. But i think so. May 22 06:21:58 you're probably right May 22 06:22:50 also, since we know that the pruss and reproc module has some "unloading" bugs, for which ds2 suggested to use them as in-kernel and not as loadable module. May 22 06:23:21 but, there is no difference btw a loaded module and a in-kernel thing ? right ? May 22 06:23:37 so what we can do is, just not un-load it, ie not lsmod it. May 22 06:23:49 right ? and everything will be fine . May 22 06:23:57 yes May 22 06:24:01 Cool ! May 22 06:24:21 well assuming it loads properly to begin with, then there's no difference May 22 06:24:36 So , we can use this very compiled image, and allow the module to load automatically at boot. May 22 06:24:54 yes, it does. at-least with the current version that i have, May 22 06:25:47 yes I think this one will too, since standardizing on a base that is stable was the point May 22 06:26:40 Cool then, now looking into the rproc and pruss source . May 22 06:27:12 okay, is there a way to know the kernel version inside a pre-compiled image ? May 22 06:27:49 okay got it. its 4.4.8-ti-r22 May 22 06:28:21 that's interesting, I was going to guess it was the 4.1 May 22 06:28:39 saw both 4.1 and 4.4 configs https://github.com/beagleboard/image-builder/tree/master/configs May 22 06:28:46 for IOT image May 22 06:39:41 Wormo: but there is no non-rt release tagged as 4.4.8 ... https://github.com/beagleboard/linux/releases May 22 06:41:50 okay so rt images are basically obtained by applying PREEMPT patch, that, i guess changes something in core like scheduler or something, so as far as pruss and rproc are concerned, i can see use their source in the 4.4.8-ti-rt-r22 May 22 06:41:54 isnt it ? May 22 06:44:03 you'll need the matching source tree to make sure modules are happy, I'll take a look May 22 06:45:06 you don't see the -rt in your kernel reported by uname? May 22 06:46:03 i dont see it here : http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Jessie_Snapshot_lxqt May 22 06:46:04 From the image builder scripts, it looks like -rt kernel does go into those debian IOT images... May 22 06:46:51 it says 4.4.8-ti-r22 there on that page. May 22 06:46:53 https://github.com/beagleboard/image-builder/blob/e829334ae836a44ab2831141dccc37046f8bcf51/configs/seeed-debian-jessie-iot-v4.4.conf May 22 06:46:56 yes May 22 06:47:08 so that's what I expect to be inside your image May 22 06:50:36 that = 4.4.8-ti-r22 ? May 22 06:51:16 You got the image all the way downloaded yet? May 22 06:52:04 not yet. May 22 06:52:23 low on bandwidth for now. will take some time. May 22 06:52:32 sure I know how that goes May 22 06:53:16 btw, what line in the conf file shows that its a rt image ? it just shows the installed packages na ? May 22 06:54:24 okay is that the linux header file's name ? May 22 06:54:37 or linux image line ? May 22 06:54:38 whoa you're right, I must be getting tired May 22 06:54:48 I thought I saw -rt in there somewhere May 22 06:55:11 so it agrees with the tag you found May 22 06:55:59 linux-image-4.4.8-ti-r22 May 22 06:56:22 yes, but the source tree has no such release. May 22 06:57:04 so where am i suppose to find the source ? May 22 06:57:51 okay first, we need to find the source .. right ? as it is the only way to know whats inside the pruss and rproc modules. May 22 06:58:18 it could be an RCN repo, he maintains kernel tree May 22 07:00:09 http://rcn-ee.net/deb/jessie-armhf/v4.4.8-ti-r22/ May 22 07:01:39 yeha ! May 22 07:01:41 yes ! May 22 07:01:44 finally ! May 22 07:01:53 I was there first ;) May 22 07:02:00 but yes, your idea ! May 22 07:02:03 there's git tree too, looking in eewiki to find it again May 22 07:04:42 source repos are on github May 22 07:04:48 ugh "WARNING, due to the setup of this repo, all "tag's" are un-reproducible." May 22 07:04:51 https://github.com/RobertCNelson/ti-linux-kernel-dev May 22 07:05:05 patch trees for mainline/dev May 22 07:05:36 on separate branches like the manifest thingie May 22 07:05:51 I don't see the exact tag though May 22 07:06:11 whereas the patch file at rcn-ee.net does have exact match May 22 07:06:42 nerdboy, I only see 'ti-linux-4.4.y' in the git May 22 07:07:21 as opposed to 4.4.8-ti-r22 in the image May 22 07:10:21 there are a few different arm repos with branches for the different patch sets May 22 07:11:20 normally it's either N.N.y branch checkout or tag May 22 07:11:25 I was looking on BBB eewiki page for the git repo that looked most relevant, you think there's a better one? May 22 07:13:30 Ok it is here in this repo May 22 07:13:32 https://github.com/RobertCNelson?tab=repositories pick one May 22 07:13:49 not the kernel source repo, rather a patch repo May 22 07:14:33 https://github.com/RobertCNelson/ti-linux-kernel-dev/tree/4.4.8-ti-rt-r22 May 22 07:14:53 it should be equivalent to the single patch in http://rcn-ee.net/deb/jessie-armhf/v4.4.8-ti-r22/ right? May 22 07:15:45 armv7-multiplatform, bb-kernel, armv7-lpae-multiplatform, armv7_devel, ti-linux-kernel-dev May 22 07:17:12 https://github.com/RobertCNelson/ti-linux-kernel-dev/tree/ti-linux-rt-4.4.y/patches <= somewhere in there? May 22 07:17:15 ZeekHuge: looks like you clone https://github.com/RobertCNelson/ti-linux-kernel-dev and it has the tag, you can run build_kernel.sh to apply the patches to upstream May 22 07:18:51 https://github.com/RobertCNelson/ti-linux-kernel-dev/tree/4.4.8-ti-r22 May 22 07:18:58 ask robert where the r22 comes from, builder script patch rev or some such... May 22 07:19:39 Wormo: yes , even the https://eewiki.net/display/linuxonarm/BeagleBone+Black#BeagleBoneBlack-LinuxKernel suggests so, under the heading TI BSP May 22 07:20:40 That part just said to checkout the branch to get latest May 22 07:21:28 but it turns out you can check out the exact tag (there's a giant list of tags and -rt ones come first so I didn't notice the matching one at first) May 22 07:23:15 heading for bed soon (saturday nights should be a little less late, don't want to miss church in the morning) May 22 07:23:31 gnight ZeekHuge, nerdboy May 22 07:23:38 just a second May 22 07:23:40 plz May 22 07:23:42 yes? May 22 07:23:56 to get source from that build_sh May 22 07:24:04 I'll have to comment out those May 22 07:24:12 https://github.com/RobertCNelson/ti-linux-kernel-dev/blob/4.4.8-ti-r22/build_kernel.sh#L250 May 22 07:24:24 make kernel and all those lines, right ? May 22 07:25:18 no it doesn't hurt to let it build kernel as well, you've already got the right compiler eh? May 22 07:25:32 https://releases.linaro.org/components/toolchain/binaries/5.3-2016.02/arm-linux-gnueabihf/gcc-linaro-5.3-2016.02-x86_64_arm-linux-gnueabihf.tar.xz May 22 07:25:39 comment out make_kernel, make_module_pkg, and make_firmware_pkg. May 22 07:25:43 ohkay. May 22 07:25:51 got it . May 22 07:25:54 Thank you :) May 22 07:25:58 Good night. May 22 07:26:10 future builds of modules will be quick then, kernel is good at partial rebuilds May 22 07:26:27 yeah . May 22 07:27:43 Good night :) May 23 00:23:20 Hello ! May 23 00:23:31 hi ZeekHuge, how's it going? May 23 00:24:04 * ZeekHuge just woke up, and hasnt even came out of bed, will be back after getting a cup of tea/coffee May 23 00:24:09 Hey Wormo ! May 23 00:24:13 Not that good .. May 23 00:24:45 The new image contains nothing like the pruss or rproc driver, May 23 00:25:41 just disabled in the config right? so you'll rebuild with those as modules? May 23 00:26:38 root@beaglebone:~# ls -l /lib/modules/4.4.8-ti-r22/kernel/drivers/remoteproc/ May 23 00:26:38 total 28 May 23 00:26:39 -rw-r--r-- 1 root root 25288 Apr 27 22:35 omap_remoteproc.ko May 23 00:29:04 so, if i wish to use rproc, I will have to use its source, from a different version of kernel, and compile it, which will make it a rproc porting project, and not the beaglescope project. May 23 00:30:52 and, I remember, I tried some older images too (ofc not all) and the only one in which the rproc was there was the last one which i was using. ( I dont remember but probably it was 4.1.17-ti-rSomething ) May 23 00:33:49 though the 4.4 branch has pruss and pru_rproc May 23 00:33:52 https://github.com/beagleboard/linux/tree/4.4/drivers/remoteproc May 23 00:34:03 * ZeekHuge will be right back. May 23 00:35:02 so you're sure it's not just turned off in the config? I need to download the matching kernel now, didn't get around to it last night May 23 01:09:40 back .. May 23 01:10:39 Wormo: Nope, i just saw the conf files at /etc/moprobe.d/ May 23 01:10:39 *modprobe May 23 01:10:41 but, the dmesg has something related to rproc May 23 01:11:17 http://paste.debian.net/695236/ May 23 01:12:18 btw, I got my welcome package last night :) It beautiful, a GSoC pen, a GCoC note making diary and a beautiful ( though small ) laptop sticker. May 23 01:12:24 yay May 23 01:12:44 post a pic on G+? May 23 01:13:09 Sure . its on fb. I'll put one on G+ too. May 23 01:15:31 any idea about that remoteproc thing in the dmesg ? May 23 01:17:31 Wormo: ? May 23 01:18:00 that's a different user of rproc, wkup would be a separate processor than pruss, so I'm not too concerned about that complaint May 23 01:18:16 still getting my kernel source set up here... May 23 01:18:33 home DSL also not too swift May 23 01:18:37 okay. May 23 01:22:16 sigh it's insisting in downloading toolchain again that I already had installeed May 23 01:22:17 definitely understand that :) no problem. May 23 01:22:36 place it in the tools/ folder ? May 23 01:22:41 build_kernel.sh that is, now I'm the one considering hacking it ;) May 23 01:23:35 I tried packing up my installed toolchain but I guess it looked at checksum & rejected the tarball May 23 01:24:08 what time is it there, getting late? May 23 01:24:20 oh ! not in the tools/ but in dl/ May 23 01:24:30 right that's where I put it May 23 01:24:32 nope. it 6:54 AM May 23 01:24:56 oh yeah that's right, you said you were just getting up May 23 01:25:20 yep :) May 23 01:26:10 that's still really early in my book, feel awful if I get up early like that May 23 01:26:22 so the coding period has started I guess ? May 23 01:26:50 I think you're right! May 23 01:28:11 I kept the source kernel tree at download, so was waiting for that. And decided to sleep and get up early. though the download got interrupted in the middle . I have made this simple bash while loop script now, that repeats any command given to it. May 23 01:28:52 UTC time : 1:25 AM, 23rd May 2016. May 23 01:30:21 and i found this cool way to get the kernel source on unreliable connection . https://cdn.kernel.org/category/site-news.html May 23 01:30:33 so just doing what it says. May 23 01:30:49 I will upgrade my data plan soon though .. May 23 01:32:29 ok now kerel downloading, this will take a while too May 23 01:33:36 also, I think there is a typo/bug in the build_kernel.sh ( or other .sh files that are being called ). It clones the kernel source from http://... , which on my system atleast, didnt work, I had to so I manually tries using git://.... May 23 01:33:42 and that did work . May 23 01:34:03 cloning kernel from git:// or http:// ? on your system ? May 23 01:34:15 https:// actually May 23 01:34:18 which does work May 23 01:34:51 oh. okay. May 23 01:34:54 2% done now... May 23 01:35:08 eta ? May 23 01:35:36 doesn't give one May 23 01:35:46 okay, yes . its cloning. May 23 01:35:59 right as opposed to the wget May 23 01:41:26 also, out of curiosity. Is the linux source at github exactly the same as at kernel.org ? ie the one we are downloading. May 23 01:41:58 and what does that warning means ? https://eewiki.net/display/linuxonarm/BeagleBone#BeagleBone-TIBSP May 23 01:42:03 *mean May 23 01:42:15 the ti one is not the same as linus tree May 23 01:42:56 ahh .. where are you cloning from ? May 23 01:44:10 This build system takes Linus' "vanilla" tree and applies patches to it May 23 01:44:49 vanilla, mainline, official, whatever you want to call it May 23 01:45:30 okay so this will work na ? https://cdn.kernel.org/category/site-news.html May 23 01:46:09 yes May 23 01:47:06 sorry for those 'na' in the conversation, that might confuse sometimes, but its just an Indian habit, means nothing imp. May 23 01:47:09 okay. May 23 01:48:38 Is it similar to how some people end questions in "no?" May 23 01:48:53 I can use 'git bundle' for this, no? May 23 01:49:00 Exactly ! May 23 01:49:21 No problem then, msg received ;) May 23 01:50:05 17% and at least I don't tend to get disconnects May 23 01:51:44 okay so, I just want to confirm my steps to build the exact kernel, I forked ti-linux-kernel-dev, cloned the 4.4.y branch only, added a new branch from the 4.4.8-ti-r22 tag, entered that branch and executed build_kernel.sh May 23 01:51:56 yes good May 23 01:52:03 cool ! May 23 01:52:59 Wormo, what other project are you in ? May 23 01:54:40 I run a local LUG, do some miscellaneous commits to projects on github May 23 01:55:09 some ansible-related repos lately, which has been fun to learn May 23 01:55:23 ah ... GSoC projects :) ? I will see the list. May 23 01:55:38 this year. May 23 01:55:39 No not GSoC projects, you're only one May 23 01:55:49 okay :) May 23 01:56:50 btw, i tried vim, like using it regularly for editing files, like i do use nano, for simple edits. but I that makes me use only some very basic commands. nothing else. May 23 01:57:58 so i know what ESC, ':q', ':qw' and 'i,j,k,l,h' does. May 23 01:58:10 and 'd' or 'x'. May 23 01:58:13 Hey have you seen 'tmate' or even tmux yet? May 23 01:58:23 those are all good basics May 23 01:58:47 I was thinking we should install tmate & do terminal sharing May 23 01:59:21 tmate , never used . May 23 01:59:25 https://tmate.io/ May 23 01:59:43 yep. No apt install package May 23 01:59:56 okay it has May 23 01:59:57 I think we must install it from git May 23 02:00:05 will have to add ppa repo May 23 02:00:11 ok that's fine too May 23 02:00:43 what ubuntu 16.04 has it in main repo May 23 02:01:05 that's convenient May 23 02:02:15 tmate needs to be on the computer running the terminal being shared May 23 02:02:24 rest can use normal ssh clients May 23 02:02:53 okay so should install on the bbb ? or my system ? May 23 02:03:05 your system May 23 02:03:41 then there would be different panes that have your BBB serial console and ssh sessions May 23 02:04:05 and then you can show your mentors exactly what you're seeing May 23 02:04:39 okay. May 23 02:04:41 sounds pretty awesome right? May 23 02:04:52 yep :) May 23 02:05:02 I've not used it yet, but have done similar with GNU screen + ssh May 23 02:05:03 updating. May 23 02:05:32 Have been thinking about using tmate for an openembedded tutorial session at our local hackerspace May 23 02:05:57 let people get into a build machine with the layers already set up May 23 02:06:53 41% ... think the clone will be done by the time I get back from dinner May 23 02:07:18 Sure, happy dining . :) May 23 02:07:40 I'm leaving in about 15 min, back a couple hours after that May 23 02:07:53 dinner + visit with family who live close by May 23 02:08:18 oh, okay. I had a discussion about that encode pin, with zmatt on #beagle. May 23 02:09:02 So, he helped a lot, and suggested to go through some basics of transmission line. May 23 02:09:17 av500: Hey ! Any idea about the ADC EVM ? May 23 02:09:34 related to its shipping ? May 23 02:09:42 _av500_: ^^ May 23 02:09:58 ds2: ^^ May 23 02:12:22 Wormo: just one question, are my previous mails okay ? May 23 02:12:30 yes May 23 02:12:43 I mean, is there something wrong in them ? May 23 02:13:30 okay. May 23 02:13:38 I don't know of anything, either the question ones or the pre-coding-plan one I requested May 23 02:14:12 okay. May 23 02:19:08 So I am assuming the BBB and the ADC will arrive not before the next week, even if it does, I'll start with the hardware only after this first week. For this week, goal is to get started with beagleLogic, and get the kernel part ready. Also to get more understanding and experimentation with the rproc drivers. May 23 02:19:38 yes that's a safe way to reorder, I was thinking that too May 23 02:20:02 keep doing the parts that don't need ADC until it actually arrives May 23 02:20:11 heading out now, ttyl May 23 02:20:16 sure ! May 23 02:20:19 :) **** ENDING LOGGING AT Mon May 23 02:59:58 2016