**** BEGIN LOGGING AT Wed Mar 08 03:00:03 2017
Mar 08 08:49:43 <flappyshabeer> hi
Mar 08 08:49:47 <flappyshabeer> anyone there?
Mar 08 09:00:43 <Wormo> hi flappyshabeer
Mar 08 09:31:10 <ravikp7> tokencolour: Did you succeed  in flashing the kernel with BBBlfs ?
Mar 08 09:32:32 <tokencolour> ravikp7: Uhmm...I'm currently in the middle of an alternative. Will let you know.
Mar 08 09:34:28 <ravikp7> tokencolour: I've fixed some bugs in BBBlfs code, which could be the reason for flasher getting stuck. If you wanna try the fixed one, let me know
Mar 08 09:35:01 <tokencolour> ravikp7: okay. I'll let you know.
Mar 08 13:11:03 <markroxor> hello all, can someone direct me to some place where I can past gsoc proposals?
Mar 08 13:33:42 <uka_in> markroxor: Hi, you need to first post your proposal on mailing list(http://bbb.io/gsocml) for reviews and suggestions.
Mar 08 13:53:56 <prav> hey
Mar 08 13:54:13 <prav> anyone der?
Mar 08 13:55:50 <prav> hey
Mar 08 16:07:00 <ravikp7> jkridner: Official meeting ?
Mar 08 16:52:17 <ayushb> help
Mar 08 17:11:51 <jkridner> gm all... my power (and thus Internet) are going out frequently right now (wind storm)... but I'm trying to be here.
Mar 08 17:11:53 <jkridner> I think this is our meeting time.
Mar 08 17:12:34 <ds2> europe getting blown?
Mar 08 17:14:56 <pmezydlo> hi all
Mar 08 17:15:25 <maciejjo> hello :)
Mar 08 17:19:23 <kiran4399> jkridner: HI
Mar 08 17:19:39 <jkridner> not in Europe until Monday.
Mar 08 17:20:17 <kiran4399> jkridner: You want to move my proposal to wiki?
Mar 08 17:28:57 <ravikp7> jkridner: There is hard coded hardware address for BBB and host in the BBBlfs code ? are they arbitary addresses ?
Mar 08 17:30:54 <_av500_> jkridner: hi jkridner
Mar 08 17:31:20 <jkridner> ravikp7: I'd expect it to be arbitrary if assigned by the board.
Mar 08 17:31:32 <jkridner> _av500_: will you be at Embedded World next week?
Mar 08 17:31:37 <_av500_> jkridner: yes
Mar 08 17:31:41 <_av500_> u too?
Mar 08 17:31:58 <_av500_> jkridner: any goodies I can grab from ti booth?
Mar 08 17:34:31 <ravikp7> jkridner: In the BBBlfs code, before claiming the interface 1 of usb device(BBB), it checks if kernel driver is active for interface 0 and detaches it if so.
Mar 08 17:35:31 <ravikp7> jkridner: In tutorial of js that you shared, you were checking kernel driver active for interface 1, for claiming interface 1
Mar 08 17:37:23 <ravikp7> Which one is correct ? If BBBlfs one is right then, that's the reason of your device claim fail error probably
Mar 08 17:37:25 <jkridner> _av500_: yup. will be at bb.org booth,not ti
Mar 08 17:37:40 <jkridner> ravikp7: wasn't so much a tutorial as my own exploration
Mar 08 17:38:52 <_av500_> jkridner: im there tuesday!
Mar 08 17:39:10 <jkridner> _av500_: cool, it'll be great to catch up
Mar 08 17:39:23 <jkridner> BBBlfs works, my try hangs.
Mar 08 17:39:39 <_av500_> jkridner: yes!
Mar 08 17:39:56 <kiran4399> jkridner: there?
Mar 08 17:40:21 <ravikp7> jkridner: The packet node library  that you used, is not working for me,  either the  library is outdated or its documentation. Did it work for you ?
Mar 08 17:40:48 <kiran4399> jkridner: thinking of running libraries in kernel space along with porting ardupilot on beaglebone blue
Mar 08 17:42:34 <jkridner> ravikp7: yes, with node7
Mar 08 17:42:40 <jkridner> node v7, that is.
Mar 08 17:42:53 <jkridner> oh, wait...
Mar 08 17:43:07 <jkridner> no, the node-usb worked for me... I never tried the packet library really.
Mar 08 17:43:18 <ravikp7> v 7.6 here, packet not working
Mar 08 17:44:05 <jkridner> k. don't know if it is necessary, just being lazy if I can find an existing way of handling the binary data.
Mar 08 17:44:17 <uka_in_> jkridner:i cross compiled uboot, however there's some error while testing it on qemu
Mar 08 17:44:23 <ravikp7> jkridner: we need to explore some other binary parser
Mar 08 17:44:27 <ravikp7> ?
Mar 08 17:44:33 <jkridner> ravikp7: don't I need to claim the interface somehow?
Mar 08 17:44:56 <jkridner> ravikp7: k, any ideas? seems lots of people just do it manually. :(
Mar 08 17:45:25 <ravikp7> currently exploring it
Mar 08 17:46:36 <ravikp7> jkridner: why BBBlfs checks for kernel driver active on interface 0 when it has to claim interface 1
Mar 08 17:47:21 <maciejjo> ds2: hi, are you here? I've send you an email regarding PRU DMA project, did you have a chance to look at it?
Mar 08 17:48:03 <ds2> maciejjo: yes.... as long as you are using the EDMA driver, it should be fine
Mar 08 17:48:17 <maciejjo> ok, cool
Mar 08 17:48:26 <jkridner> hi vvu_
Mar 08 17:48:36 <kiran4399> jkridner: ardupilot?
Mar 08 17:48:44 <kiran4399> jkridner: kernel interfaces?
Mar 08 17:49:15 <jkridner> vvu_: do you have any thoughts on the process of claiming the interface?
Mar 08 17:49:38 <jkridner> kiran4399: ?
Mar 08 17:49:52 <kiran4399> jkridner: thinking of running libraries in kernel space along with porting ardupilot on beaglebone blue
Mar 08 17:50:06 <kiran4399> jkridner: will that suffice for a project this time?
Mar 08 17:50:13 <kiran4399> jkridner: along with ros wrappers..
Mar 08 17:50:20 <jkridner> kiran4399: ardupilot is already implemented on BeagleBone Blue and merged into its mainline.
Mar 08 17:50:22 <ds2> maciejjo: do you know much about the remoteproc/uio issues?
Mar 08 17:50:39 <kiran4399> jkridner: can you give me the link?
Mar 08 17:50:50 <jkridner> kiran4399: you could propose to either BeagleBoard.org or ArduPilot (who is mentoring their first GSoC) to improve the Blue support, but it is already there.
Mar 08 17:51:02 <jkridner> kiran4399: https://github.com/ardupilot/ardupilot
Mar 08 17:51:05 <maciejjo> I've used uio in the past, this week I started using remoteproc
Mar 08 17:51:32 <jkridner> kiran4399: https://github.com/ArduPilot/ardupilot/commit/61a88a9b5e0326acca6ee11e14219bef65b36e62
Mar 08 17:51:47 <maciejjo> should I support both interfaces for this project or can I focus on remoteproc?
Mar 08 17:52:09 <kiran4399> jkridner: what about kernel drivers?
Mar 08 17:52:17 <jkridner> kiran4399: https://github.com/ArduPilot/ardupilot/search?q=blue&type=Commits&utf8=%E2%9C%93
Mar 08 17:52:35 <jkridner> kiran4399: adding kernel drivers would be a great project in my opinion
Mar 08 17:52:48 <kiran4399> jkridner: for this years gsoc?
Mar 08 17:52:51 <jkridner> kiran4399: not sure of all the blockers last year.
Mar 08 17:52:58 <kiran4399> hmm.
Mar 08 17:53:12 <ds2> maciejjo: I think supporting one is good. both would be a stretch goal if you get bored and have extra weeks :D
Mar 08 17:53:14 <kiran4399> jkridner: what else can be done with blue? cloud9 support?
Mar 08 17:53:26 <jkridner> kiran4399: oooo.... sooooo much to do with Blue.
Mar 08 17:53:36 <ds2> maciejjo: you do need to have DMA working both ways (A8 -> PRU) and (PRU -> A8)
Mar 08 17:53:42 <jkridner> open source graphical programming tools like Blockly, ScratchX, etc.
Mar 08 17:54:03 <tokencolour> Hello ds2! I've been trying to run my own kernel onto the BBB as you suggested.
Mar 08 17:54:05 <kiran4399> jkridner: can we do something with robotics?
Mar 08 17:54:11 <maciejjo> ok, thats how I imagined it
Mar 08 17:54:19 <ravikp7> jkridner: I made a PR for BBBlfs https://github.com/ungureanuvladvictor/BBBlfs/pull/31 can you have a look
Mar 08 17:54:29 <tokencolour> I tried this http://www.armhf.com/boards/beaglebone-black/bbb-sd-install/
Mar 08 17:54:30 <ds2> maciejjo: the user space part can even be optional
Mar 08 17:54:32 <jkridner> an HTML "cockpit" with mjpg-streamer plus phone-based-browser accelerometer controls and just general controls/feedback.
Mar 08 17:54:42 <ds2> tokencolour: did you get anywhere?
Mar 08 17:55:06 <jkridner> ravikp7: k.
Mar 08 17:55:18 <ds2> blah... PR is such a poor acronym
Mar 08 17:55:35 <ds2> (PR == Purchase Request? or is it Push Request)
Mar 08 17:55:58 <maciejjo> ok, so it is sufficient if I manage to copy memory using DMA both ways between kmod and PRU, correct?
Mar 08 17:56:13 <nerdboy> Problem Report
Mar 08 17:56:34 <kiran4399> jkridner: are there any robotics frameworks which can be linked up with the beaglebone blue?
Mar 08 17:56:40 <tokencolour> I was able to compile and get an image and all the other necessary files(MLO and u-boot.img) but have had a problem placing them at the right place.
Mar 08 17:56:50 <ravikp7> ds2: from now only Pull Request for sure 😉
Mar 08 17:57:03 <ds2> maciejjo: I think so
Mar 08 17:57:11 <nerdboy> MLO has to be thr first thing on the boot partition
Mar 08 17:57:19 <ds2> tokencolour: new style or old style?
Mar 08 17:57:25 <nerdboy> kinda finicky
Mar 08 17:57:28 <tokencolour> But that link I shared right now had a pre-compiled image and it worked.
Mar 08 17:57:34 <jkridner> kiran4399: cylon.js, ROS, gobot.io, artoo.io
Mar 08 17:57:36 <tokencolour> MLO first.
Mar 08 17:57:43 <tokencolour> u-boot.img second
Mar 08 17:57:51 <ds2> IOW, are you using a FATfs or not
Mar 08 17:58:02 <tokencolour> and then we can place the rest of the files
Mar 08 17:58:14 <tokencolour> FAT16 for boot partition
Mar 08 17:58:19 <tokencolour> ext4 for rfs.
Mar 08 17:58:31 <ds2> maciejjo: I take it you are familiar with the DMA memory allocation stuff?
Mar 08 17:58:43 <tokencolour> Didn't get very far with tftp.
Mar 08 17:59:22 <ds2> tokencolour: make sure the partition is setup right - type 'C' and activated... and make sure you have an acceptable "geometry"
Mar 08 17:59:44 <maciejjo> I've done DMA support for drviers but on another platform
Mar 08 18:00:18 <jkridner> kiran4399: I also think hooking node-red and finishing node-roboticscape are useful.
Mar 08 18:00:29 <ds2> maciejjo: do you know how to get a kernel going on the BeagleBone/AM335x platform?
Mar 08 18:00:36 <jkridner> kiran4399: https://en.wikipedia.org/wiki/Open-source_robotics
Mar 08 18:00:37 <maciejjo> yeah
Mar 08 18:00:37 <gcl-bot> [WIKIPEDIA] Open-source robotics | "Open-source robotics (OSR) is a branch of robotics where the physical artifacts of the subject are offered by the open design movement. This open design movement applied to the field of robotics makes use of open-source hardware and free and open source software providing blueprints, schematics, and..."
Mar 08 18:00:39 <kiran4399> jkridner: but whatever software we would like to port.. I think we need to get the kernel drivers ready right?
Mar 08 18:00:48 <ds2> maciejjo: do you have one already?
Mar 08 18:00:54 <maciejjo> yes I do
Mar 08 18:01:23 <jkridner> kiran4399: I've been trying to get someone to to Paparazzi (https://en.wikipedia.org/wiki/Paparazzi_Project)
Mar 08 18:01:24 <gcl-bot> KeyError: u'extract' (file "/usr/local/lib/python2.7/dist-packages/sopel/modules/wikipedia.py", line 87, in mw_snippet)
Mar 08 18:02:15 <jkridner> kiran4399: I think getting the drivers into the kernel (with same or better latency) would have a big impact.
Mar 08 18:02:33 <ds2> maciejjo: have you done the basic - compile and push test (general exercise asked for by this org)?
Mar 08 18:02:33 <tokencolour> ds2: What is meant by the new and old one?
Mar 08 18:02:46 <jkridner> kiran4399: I sent a note to lkml regarding using BPF on IIO to filter/control.
Mar 08 18:02:56 <maciejjo> yes
Mar 08 18:03:05 <ds2> tokencolour: old way is to use the FATFS. new way puts it in an absolute location near the boot sector so no FATfs is needed
Mar 08 18:03:06 <kiran4399> jkridner: bpf?
Mar 08 18:03:44 <maciejjo> ds2: is the new way documented somewhere? I would like to try it as well, always used FAT partition before
Mar 08 18:03:46 <jkridner> ravikp7: PR seems helpful. not sure about getting rid of the infinite wait. might be nice to know it continues trying, but being able to have a known-sequence of starting the app and then plugging the board is important.
Mar 08 18:03:51 <tokencolour> tokencolour: okay so I did it the old way.
Mar 08 18:03:53 <ds2> maciejjo: not sure what more to add at this point :D
Mar 08 18:03:55 <jkridner> ravikp7: want to avoid possible race conditions.
Mar 08 18:04:24 <tokencolour> oops!
Mar 08 18:04:24 <ravikp7> tokencolour ran into the issue of infinite wait I guess
Mar 08 18:04:31 <ds2> maciejjo: yes... you can either do it the hard way (read the TRM) or look for the magic dd incantations on where to write it to...it is in the ML archives...some folks here may even know it off the top of my head
Mar 08 18:04:51 <jkridner> kiran4399: BPF is essentially jit-compiled x86 code used generally for packet filtering, but I think could be applied to IIO event filtering (and putting control loops into the IIO subsystem).
Mar 08 18:05:11 <tokencolour> ds2: I did it the old way.
Mar 08 18:05:31 <tokencolour> ravikp7: Yes I did.
Mar 08 18:05:52 <ds2> tokencolour: that is fine...check those things I mentioneded earlier (part type, active, and geo)
Mar 08 18:06:33 <tokencolour> ds2: I'm actually confused with what to do with zImage and the dts files.
Mar 08 18:06:55 <jkridner> kiran4399: I think it is a reasonable question as to where the guts of rc_balance should live... kernel or userspace.
Mar 08 18:07:11 <tokencolour> ds2: Have you documented somewhere what to do exactly after having these files ready.
Mar 08 18:07:26 <kiran4399> jkridner: why can't it stay in kernel space?
Mar 08 18:07:46 <kiran4399> jkridner: can you please outline the challanges of moving it into kernel space?
Mar 08 18:07:53 <ds2> tokencolour: not really... I just look at what Uboot is setup for. DTB is loaded by uboot unless you have a combined image. same with kernel
Mar 08 18:07:56 <jkridner> kiran4399: today it is in userspace.
Mar 08 18:08:09 <ravikp7> jkridner: Infinite wait helps nothing for debugging when it runs into that issue
Mar 08 18:08:43 <ds2> tokencolour: someone else can probally comment more... don't think I can explain uboot in a reasonable amount time
Mar 08 18:08:46 <jkridner> ravikp7: but what about the case where it isn't plugged in yet?
Mar 08 18:09:55 <tokencolour> The dtbs files are fetched when mentioned in the uEnv.txt, right?
Mar 08 18:10:03 <tokencolour> ds2: ^
Mar 08 18:10:33 <ravikp7> jkridner: it should be plugged in before using tool, if not user will run the script again
Mar 08 18:11:37 <ds2> tokencolour: sort of...there is a load command that gets run but the uboot configurers for the BB have decided to make things complicated
Mar 08 18:11:52 <ds2> I will defer that to a more appropriate person :D
Mar 08 18:11:56 <maciejjo> tokencolour: I did a small writeup on how I create image for BBB, maybe it will help you: https://maciejjo.pl/bbb_custom_image.html
Mar 08 18:12:13 <kiran4399> jkridner: what are your expectations from the kernel space? what all kernel drivers for peripherals are already in the mainstream and what are yet to built?
Mar 08 18:14:45 <tokencolour> ds2: I get confused when there are so many articles that say that I should place zimage into boot/ and that doesn't really help me boot it up.
Mar 08 18:15:03 <tokencolour> maciejjo: I'm reading your writeup. Thanks.
Mar 08 18:15:29 <ravikp7> jkridner: thoughts on other commits such as releasing interface everytime before claiming again and removing some unreachable code
Mar 08 18:16:23 <tokencolour> ds2: How should I continue, once I'm able to boot the kernel?
Mar 08 18:16:58 <ds2> tokencolour: things have changed over time... I think the ltest trend is to stick it in /boot on the rootfs partition
Mar 08 18:17:09 <uka_in_> I was trying to run uboot on qemu but ran into this error https://gist.github.com/madaari/ed965e11c2f6b54350c221a3bebd3f14 can any one please help??
Mar 08 18:17:19 <ds2> tokencolour: get familiar with it cuz do to the work, you will be rebuilding things often
Mar 08 18:17:35 <maciejjo> tokencolour: probably depends what u-boot you are using, but by default if you don't provide uEnv.txt it will look for zImage and DTB on rootfs partition at /boot
Mar 08 18:17:43 <ds2> tokencolour: know how to get info from kernel output (i.e. printk), read sysfs and proc
Mar 08 18:18:09 <tokencolour> ds2: I also tried doing the same with BBBlfs but to no avail. :(
Mar 08 18:18:18 <ds2> maciejjo: depends on how uboot was compiled... the default environment is patched
Mar 08 18:18:51 <ravikp7> uka_in_: I successfully cross compiled u boot for BBB, see if it helps https://gist.github.com/ravikp7/770ff7df85f0e0588d3b112b32403697
Mar 08 18:18:55 <maciejjo> ok - this is how unchanged u-boot from TI behaves
Mar 08 18:19:01 <tokencolour> maciejjo: I didn't know providing uEnv.txt is an option.
Mar 08 18:19:04 <jkridner> kiran4399: just look at libroboticscape (https://github.com/StrawsonDesign/Robotics_Cape_Installer.... almost none of is in the kernel.
Mar 08 18:19:17 <jkridner> kiran4399: he mmap()s just about everything...
Mar 08 18:19:30 <jkridner> kiran4399: I think some type of profiling would be most helpful.
Mar 08 18:19:56 <jkridner> ravikp7: looked mostly good to me.
Mar 08 18:20:03 <uka_in_> <ravikp7>I did that too.. but there was an error while loading it in uboot
Mar 08 18:20:19 <jkridner> ravikp7: do you have any fixes for the node.js experimentation?
Mar 08 18:20:32 <uka_in_> <ravikp7>* i mean in qemu
Mar 08 18:20:33 <jkridner> ravikp7: stuck until you get some hardware?
Mar 08 18:20:46 <ravikp7> jkridner: yeah
Mar 08 18:21:27 <ravikp7> jkridner: I changed your code https://gist.github.com/ravikp7/fbaa73ebd00297dfb3fe9d951268d199
Mar 08 18:21:43 <ravikp7> jkridner: test it
Mar 08 18:21:48 <uka_in_> jkridner:i do have hardware ,should i start with node-usb?
Mar 08 18:23:32 <ravikp7> jkridner: please confirm if it claims the interface with changed code or still an error
Mar 08 18:25:01 <tokencolour> ds2: we'll be using <linux/dma-mapping.h> and <linux/mm.h> when allocating DMA mappings, won't we?
Mar 08 18:25:14 <tokencolour> ds2: Sorry for the dumb questions.
Mar 08 18:25:32 <tokencolour> I'm a newbie here.
Mar 08 18:26:20 <ds2> tokencolour: I think so
Mar 08 18:26:33 <ds2> tokencolour: look at other drivers... i.e the ethernet one
Mar 08 18:28:08 <tokencolour> ds2: Thanks for pointers. I'll look into them.
Mar 08 18:29:26 <tokencolour> ds2: How will the ethernet drivers be relevant here?
Mar 08 18:30:04 <tokencolour> ds2: Also what questions should I answer once preparing the proposal?
Mar 08 18:30:32 <jkridner> ravikp7: k, I'll try it.
Mar 08 18:30:41 <jkridner> uka_in_: yeah, I'd think so.
Mar 08 18:30:51 <jic23> Jkridner, kiran4399: agreed, first step definitely a spot of profiling.  Need to know where bottle necks are!
Mar 08 18:31:16 <jic23> I can guess, but don't have the numbers
Mar 08 18:32:41 <jic23> Not looked at the code, but if doing control loop in kernel is it fixed point, or integer?
Mar 08 18:33:48 <jic23> Certainly a good project idea!
Mar 08 18:34:20 <jic23> Lots of fun stretch goals.
Mar 08 18:35:38 <ds2> tokencolour: it uses DMA
Mar 08 18:36:08 <tokencolour> oh okay.
Mar 08 18:39:52 <ravikp7> jkridner: one more question about BBBlfs, how do we know that the size of buffer is  450 bytes,  is it somewhere in datasheet or something else ?
Mar 08 18:39:59 <jic23> Jkridner, might be worth talking to Gregor Borrie at parrot. He has had one of their UAVs controlled off standard kernel drivers. Might be willing to share some experiences.
Mar 08 18:43:46 <jkridner> ravikp7: I added a comment to your gist.
Mar 08 18:43:57 <jkridner> ravikp7: it still had an error at the claim.
Mar 08 18:44:43 <jkridner> jic23: that'd be great! can you make an intro?
Mar 08 18:45:30 <jic23> Sure, will do when on pc.
Mar 08 18:47:27 <ravikp7> jkridner: what do you think now, what could be the problem ? problem with node usb ?
Mar 08 18:49:08 <jkridner> not sure... that's where I got stuck.
Mar 08 18:49:33 * jkridner has to go offline to record a video.
Mar 08 18:49:43 <jkridner> I'll be back in a couple hours.
Mar 08 19:13:00 <ravikp7> jkridner: If you can try it on some other systems, do try it.
Mar 08 19:18:39 <ravikp7> jkridner: as per your error logs, this.device.__claimInterface error occurs at node_modules/usb/usb.js line 168, but as I can see at https://github.com/tessel/node-usb , this.device.__claimInterface is at  node_modules/usb/usb.js line 180. Are you running latest version of node usb?
Mar 08 19:19:23 <ravikp7> uka_in_ can you test the node.js code as you have the hardware?
Mar 08 19:19:45 <uka_in_> yup i am working on it
Mar 08 19:20:11 <ravikp7> uka_in_ : got node installed ?
Mar 08 19:20:23 <uka_in_> not yet
Mar 08 19:20:29 <uka_in_> but soon will
Mar 08 19:20:37 <ravikp7> ok
**** ENDING LOGGING AT Thu Mar 09 03:00:02 2017