**** BEGIN LOGGING AT Tue Aug 12 02:59:58 2014 Aug 12 13:26:13 jkridner: are you here? Yesterday we did tests with Steve Aug 12 13:26:44 the issues talked yesterday were fix Aug 12 13:28:23 I'm working on the documentation Aug 12 13:28:34 jkridner: is there any special format for doing it? Aug 12 13:28:39 or we are free to do it? Aug 12 13:29:16 please let me know Aug 12 14:20:54 DiegoTc: please move ahead with Steve's recommendations and wrap up coding ASAP Aug 12 14:23:06 vvu: I'm looking at trying your bootloader. which tag should I use? Aug 12 14:23:14 vvu: I'm going to use a BeagleBone Black as the host. Aug 12 14:38:11 go latest commit Aug 12 14:38:22 that tag is old Aug 12 14:38:28 need to make a new one Aug 12 14:38:44 jkridner: ^ Aug 12 15:00:02 k Aug 12 15:16:42 jkridner: I'm doing documentation this way: https://github.com/DiegoTc/bone101/wiki/githubOauth.js Aug 12 15:17:13 Explaining the functions of each javascript file I wrote, so future developers can get reference of it Aug 12 15:17:16 k, using the wiki is good Aug 12 15:19:26 jkridner: I won't be able to make it to the irc meeting tomorrow, anything I should know/do besides making sure things are finished/cleaned up/documented? Aug 12 15:20:14 don't think so... just try to chat with students to make sure their coding is wrapped up and documentation is being done. Aug 12 15:20:30 ok Aug 12 15:33:36 <_av500_> jkridner: vvu: i am on family vacations Aug 12 15:33:43 <_av500_> but I should be there tomorrow for the meeting Aug 12 15:47:56 _av500_: no worries, i'm always lurking here Aug 12 15:48:15 takin advantage of the extebded weekend? Aug 12 15:48:54 <_av500_> vvu: 3 of them Aug 12 15:48:58 <_av500_> all connected together Aug 12 15:49:41 yeah, thu i'm finishing too here my work at TI and shipping myself to Romania Aug 12 15:50:35 <_av500_> ic Aug 12 15:51:39 praveen was here lately? Aug 12 17:37:31 vvu: can't help trying Mac again... Aug 12 17:37:47 vvu: libusb_open_device_with_vid_pid never returns non-null. Aug 12 17:38:13 vvu: I do see the device in the system info though. Aug 12 17:38:43 jkridner: mac does not work Aug 12 17:38:46 needs a codeless kext Aug 12 17:38:59 the mac kernel takes the ROM as an ethernet interface Aug 12 17:39:05 with the RNDIS kext is has Aug 12 17:40:23 if you have patience 3 more weeks i'm gonna get my hands on a mac mini and write that kext :) Aug 12 17:40:25 k, but why doesn't libusb free it up? Aug 12 17:40:34 they do not have support for that Aug 12 17:40:56 https://github.com/libusb/libusb/wiki/FAQ#How_can_I_run_libusb_applications_under_Mac_OS_X_if_there_is_already_a_kernel_extension_installed_for_the_device Aug 12 17:40:57 I see logs of you asking: http://sourceforge.net/p/libusb/mailman/message/31517196/ Aug 12 17:41:22 yes, for the windows thing....my hardware on the Dell laptop I have is crap.. Aug 12 17:41:36 Abhishek_ tested on his windows machine and it works Aug 12 17:41:53 in a couple of days I will reach home and i have there a couple of windows desktops to continue Aug 12 17:42:16 btw, I think splitting the bootloader from the flasher in name would be good... Aug 12 17:42:26 simply being able to boot is helpful. Aug 12 17:42:35 ah, so just booting you sau Aug 12 17:42:36 say Aug 12 17:43:04 we can do that...with a kernel the user supplies and a basic busybox ? Aug 12 17:43:10 or how would you say ? Aug 12 17:44:58 when i reach home i will look into this...now i have a 19" screen, cannot do anything on this :) Aug 12 17:49:13 vvu: just needs to be split. it if frightening to run usb_flasher to merely do a bootload. Aug 12 17:49:47 I'm not saying the user should need to provide anything, but everything should be able of being replaced at every step. it is the unix way. Aug 12 17:50:29 I can't seem to find what to unload. Aug 12 17:51:42 horndis Aug 12 17:51:46 that is the kext as I remember Aug 12 17:52:18 jkridner: do a kextstat i think Aug 12 17:52:25 and you can see what is has loaded Aug 12 17:52:34 https://developer.apple.com/library/Mac/technotes/tn2315/_index.html#//apple_ref/doc/uid/DTS40014014-CH1-TNTAG4 seems to be missing the downloadable. :( Aug 12 17:52:49 vvu: yeah, did that.... tried to remove anything that could look like it. Aug 12 17:53:11 jkridner: i have an example from the guys from OpenOCD Aug 12 17:53:17 for their STM probes Aug 12 17:53:19 or smth like that Aug 12 17:53:23 lemme search for it Aug 12 17:53:25 unloaded 2 that looked like they might be it, but com.apple.driver.AppleUSBCDCACMData wouldn't unload Aug 12 17:54:37 jkridner: https://github.com/texane/stlink/tree/master/stlinkv1_macosx_driver Aug 12 17:55:02 currently i do not have Apple platform :( Aug 12 17:59:08 jkridner: if you can test it that would be nice :) Aug 12 17:59:14 k Aug 12 17:59:39 atm I have in qt this: https://github.com/ungureanuvladvictor/BeagleBoneBlack-USBFlasher Aug 12 18:00:04 just bootloading Aug 12 18:02:31 vvu: that seems like a bat starting point.... not codeless Aug 12 18:03:09 the one from stlink ? Aug 12 18:04:01 vvu: https://github.com/ungureanuvladvictor/BBBlfs/pull/11 Aug 12 18:04:02 yes Aug 12 18:04:46 https://developer.apple.com/library/Mac/technotes/tn2315/_index.html#//apple_ref/doc/uid/DTS40014014-CH1-TNTAG4 is supposed to have a codeless example, but doesn't seem to. Aug 12 18:05:42 https://github.com/piha/k8055-mac-codeless-kext Aug 12 18:05:49 it is a matter just to replace the vid/pid i think so Aug 12 18:05:53 no binary or anything Aug 12 18:05:56 just in the plist files Aug 12 18:10:19 hmmm... replugging enabled me to discover the board with libusb Aug 12 18:10:48 last status message: libusb: info [darwin_claim_interface] no interface found; setting configuration: 1 Aug 12 18:14:50 can you pinpoint the #LOC ? Aug 12 18:15:47 it has 2 interfaces, and i use only #1 which is the bulk interface Aug 12 18:22:16 jkridner: https://gist.github.com/ungureanuvladvictor/aa6e3751cb3818b2f11c can you try this, just replace the vid/pid Aug 12 18:35:21 fixed typo: https://gist.github.com/jadonk/34d2161376853dc446a2 Aug 12 18:36:10 vvu: segfaults on call to get_device Aug 12 18:36:17 jkridner: change also the vid/pid Aug 12 18:36:43 #49 Aug 12 18:38:10 vid = 0x0451 pid = 0x6141 Aug 12 18:38:44 still segfaulting Aug 12 18:39:12 it cannot open the device if so Aug 12 18:40:44 usbhdl is 0. Aug 12 18:41:16 doing sudo avoids segfault Aug 12 18:42:15 usbhdl stll 0 Aug 12 18:42:55 reset target to better luck Aug 12 18:43:16 jkridner: i will reach a mac device in a couple of weeks and check everything, if you have time to work on it fine, if not fine again :) Aug 12 18:43:16 last line: num endpoints: 1 -- 131 Aug 12 18:43:35 resetting device enable discovery Aug 12 19:07:45 vvu: did you see the last line I got? Aug 12 19:07:56 I got distracted and haven't yet figured out its meaning Aug 12 19:08:32 num endpoints ? Aug 12 19:17:30 jkridner: https://gist.github.com/ungureanuvladvictor/839f4195724219d97de0 Aug 12 19:17:41 this is what is correct, remove the reset device. dunno why the example had it there Aug 12 19:18:04 https://gist.github.com/ungureanuvladvictor/aa6e3751cb3818b2f11c Aug 12 19:18:06 try with this Aug 12 19:28:31 doesn't print anything. Aug 12 19:31:18 k, now I seem to be consistently getting cur config = 0 Aug 12 19:33:14 does the ROM bail out at any point ? the rom should reset itself after some timeout Aug 12 19:34:20 Shouldn't I claim the interface, like https://gist.github.com/jadonk/d9a65609beb66cbbd45b Aug 12 19:35:26 does not matter now, you should get the info about the device even without claiming Aug 12 19:36:01 k, but it was saying that there was no active config. Aug 12 19:36:07 I was getting the device. Aug 12 19:36:13 jkridner: libusb_set_configuration(usbhdl, 1); Aug 12 19:36:15 call this Aug 12 19:36:19 try and set config 1 Aug 12 19:36:31 * vvu does not know the difference between config 1 and 0 Aug 12 19:37:10 well, you see from the printout in the comments of the gist I sent that it already things it is using config 1 Aug 12 19:37:25 what are you seeing that doesn't look right? Aug 12 19:37:58 i think looks right, never user alt_settings in my code Aug 12 19:38:04 neither I think it counts Aug 12 19:38:16 can you try now and claim interface 1 ? Aug 12 19:38:41 HI jkridner Aug 12 19:39:13 hi Abhishek_ Aug 12 19:39:47 jkridner: were you able to try out the web interface with a Bone? Aug 12 19:40:11 vvu: inside of usb_flasher? Aug 12 19:40:34 vvu: I tried 'sudo ./usb_flasher' again. reminder, this is where it sits: libusb: info [darwin_claim_interface] no interface found; setting configuration: 1 Aug 12 19:41:31 is it possible that the osx kernel took just one interface under control / Aug 12 19:41:35 Abhishek_: no... I need to try out new kernels and I'm trying to use vvu's flasher to do so... really need to improve my ability to iterate on new kernels. Aug 12 19:42:02 vvu: can you explain the log output? Aug 12 19:42:41 can you change libusb_set_debug(ctx, 3); to libusb_set_debug(ctx, 4); Aug 12 19:42:45 and give a full gist ? Aug 12 19:43:27 jkridner: so do you want to be able to boot the kernel via USB and then have it load the rootfs from the eMMC / SD card ? Aug 12 19:47:54 Abhishek_: among other things, yes. Aug 12 19:49:34 vvu: http://pastebin.com/JtkLVc0p Aug 12 19:53:27 vvu: seems to be hanging in the bulk_transfer call. Aug 12 20:02:14 vvu: It would be interesting to see if the IN tokens are going out on the bus. Aug 12 20:02:36 seems odd that the ROM would need to SEND data first. Aug 12 20:03:39 vvu: are you just prepping the buffer and expecting this to be asynchronous? Aug 12 20:17:03 jkridner: sorry I was afk Aug 12 20:17:15 jkridner: the rom sends a bootp request Aug 12 20:17:17 vvu: I see it is a synchronous call. Aug 12 20:17:20 k. Aug 12 20:17:23 packet into a RNDIS packet Aug 12 20:17:31 so plug in the BBB and run the flasher Aug 12 20:17:39 becuase after some time the ROM resets itself Aug 12 20:17:43 not sure if the IN tokens are going out, but clearly the stack isn't seeing the data. Aug 12 20:17:47 if it does not receive the bootp answer Aug 12 20:18:09 jkridner: i need to get a mac to see what is happening there...like this is super blind :) Aug 12 20:18:30 I have a bus analyzer I can hookup at some point here. Aug 12 20:18:58 check if the BBB is sending stuff Aug 12 20:19:06 i think it hangs somewhere maybe in some undefined state Aug 12 20:19:48 the same happens in windows if the usb controller on the Host side is not working nice with libusb Aug 12 20:19:55 just hangs the BBB in an undefined state Aug 12 20:19:58 and does not reset itself Aug 12 21:08:28 Hi to all! Could someone please amke me the favor to run this script with your BBB http://beagleboard.org/Support/BoneScript/analogWrite/ jkridner Aug 12 21:08:34 Abhishek_: vvu Aug 12 21:08:36 thanks :) Aug 12 21:11:27 we're testing with Steve and we can't run that example neither on the beagleboard.org website, neither the bone101 Aug 12 21:11:51 DiegoTc: you might need to be patient.... Aug 12 21:12:04 it builds a device tree overlay and loads it... Aug 12 21:12:24 I checked it out yesterday to see if it was working and it did work for me.... Aug 12 21:12:34 doesn't mean it isn't quite buggy. Aug 12 21:12:56 thanks jkridner ! Steve was asking me and I don't have a BBB with me right now Aug 12 21:13:01 to test Aug 12 21:16:00 DiegoTc: run the browser on the BBB and load the page ? Aug 12 21:16:05 or can i do it from a linux host ? Aug 12 21:17:09 jkridner: the last resort will be to make native apps for Windows and OS X Aug 12 21:17:20 and use native USB support not libusb.. Aug 12 21:17:23 but that can be super pain Aug 13 01:35:27 jkridner: do we have to post updates on our respective threads for this meeting as well? **** ENDING LOGGING AT Wed Aug 13 02:59:58 2014