**** BEGIN LOGGING AT Sat Jul 15 03:00:05 2017 Jul 15 03:08:02 another satisfied customer Jul 15 03:08:50 you know if its a connection from gateway/web/freenode/etc that its a waste-of-time. Jul 15 03:08:56 I dunno why that gateway even exists. Jul 15 03:09:20 *ahem* Jul 15 03:09:45 that's freenode's web chat client. Jul 15 03:09:59 Some people use it when they're behind firewalls. Jul 15 03:10:11 Some TI people have to do that sometimes Jul 15 03:10:19 hrm thats .. typical .. lol Jul 15 03:10:19 or don't know how to set up a proper irc client. Jul 15 03:10:26 ^ Jul 15 03:10:36 some of each, but you asked why it exists :) Jul 15 03:10:36 yes That. Jul 15 03:10:48 don't shit on them for that. Jul 15 03:11:15 So this has always been the problem here. The instructions aren't good enough. There's too much bad information, and the process is too complicated Jul 15 03:11:40 I think if you can't configure an IRC client .. WTF you playing with embedded computing devices for .. Jul 15 03:11:42 alan_o: +1 Jul 15 03:11:48 The idea that anyone should be using this without a console connection is insane. Jul 15 03:12:20 But there's all this automated behind-the-scenes stuff that's supposed to make it work without a proper console, right up until it doesn't Jul 15 03:12:28 or right up until it's _perceived_ not to Jul 15 03:12:30 yup Jul 15 03:12:57 and at that point the facts don't matter much Jul 15 04:09:24 Do you people think that because you have exorbitant time to research the insane amoutn of errors and dead end paths related to BBB and embedded linux that you are somehow special or deserving of respect? Jul 15 04:10:11 open source is bad Jul 15 04:29:16 kill your selves Jul 15 04:35:34 w/e lol Jul 15 04:37:24 wastes of space Jul 15 04:51:33 I think someone needs a ban ... :) Jul 15 04:52:01 lil emo bitch :) Jul 15 07:54:02 Anyone know what the status of the images at rcn-ee.com is? It seems there's preview images for stretch available there, but I'm not sure if this the staging ground for official images, or just some random site that offers images of unknown quality, so to speak :-) Jul 15 07:57:46 Also, is there any guidance on what images are used for what hardware? It seems there's "BBB", "bbx15", and "bone" flavours, but I can't find any reference about what image is for what hardware. In particular, I'm looking for the image to use on a BB green Jul 15 08:04:23 If the screenshots at http://wiki.seeed.cc/BeagleBone_Green/ are any hint, I should be using the "bone" version it seems Jul 15 08:49:04 blathijs: the images are somewhat more described here -> http://elinux.org/Beagleboard:BeagleBoneBlack_Debian Jul 15 08:51:53 zmatt: did you move to hawaii? Jul 15 08:52:02 zmatt: or do you have just a weird sleep rythm? Jul 15 08:52:34 sleep rhythm? what sleep rhythm? Jul 15 08:53:43 ^^; Jul 15 08:54:02 and what the hell is going on with those flashers? I've heard people saying it turns off after flashing, now nate_ had one that rebooted after flashing... which is really not good Jul 15 08:57:51 * KotH shurgs Jul 15 08:57:54 no idea Jul 15 08:58:06 also, i am not 100% sure nate understood what he was doing Jul 15 08:58:12 could be equally well PEBCAK Jul 15 09:01:26 perhaps, although flashers shutting down instead of ending with all leds on like tons of docs say is also not good Jul 15 09:03:57 all these random changes and the resulting wrong docs are really frustrating, especially for new users Jul 15 09:08:54 especially users new to linux in generall Jul 15 18:43:00 is it possible to stop the blinking leds in BBB and use them as GPIO ? Jul 15 18:49:22 yes Jul 15 20:01:41 using them as gpio would obviously require some solderwork to get access to the signals Jul 15 20:02:10 what do you mean? Jul 15 20:02:33 or do you mean control the leds manually? Jul 15 20:02:42 yes, i want to control them manually Jul 15 20:03:12 i want to change the GPIO using the sysfs driver , and get the led on/off Jul 15 20:03:20 you can control them via /sys/class/leds Jul 15 20:04:06 Okay. thanks. Jul 15 20:06:00 just write 'none' to the trigger attribute to change them to manual control, and then write 0 or 1 to brightness Jul 15 20:06:06 another thing , is the dtb of BBB is only a single file ? am335x-boneblack.dtb ? why there are like 20 files at /boot/dtbs/ Jul 15 20:06:35 different hardware uses a different dtb Jul 15 20:07:00 so can i keep only am335x-boneblack.dtb ? (or when i change the dt, can i copy only this file?) Jul 15 20:07:46 when customizing the dt I'd actually recommend picking a different name to avoid confusion with the standard ones Jul 15 20:08:10 (you can select which dtb to use with the "dtb" variable set in /boot/uEnv.txt ) Jul 15 20:14:47 zmatt: how would you suggest to debug ethernet intergration with external switch? which tools do you suggest to trace the ethernet issues? Jul 15 20:16:09 question too vague and underspecified to give a meaningful answer Jul 15 20:18:24 Okay , here my scenario , i'm connecting the RGMII1 bus into external Gbit ethernet switch port. (which have internal PHY). after i configured it with the BBB DT. which tools i should use in order to show the network status? Jul 15 20:20:19 you're not getting link up? Jul 15 20:20:52 what kernel messages are you getting related to ethernet/cpsw ? Jul 15 20:21:07 double-check that the pins have been configured right Jul 15 20:21:40 check status registers in the switch Jul 15 20:22:22 well it says that the link is up Jul 15 20:22:34 ok, so that's pretty substantial Jul 15 20:23:04 then you probably don't have to check pinmux Jul 15 20:23:39 this is what i'm geeting at linux console: Jul 15 20:23:48 [115630.066481] Generic PHY fixed-0:00: attached PHY driver [Generic PHY] (mii_bus:phy_addr=fixed-0:00, irq=-1) Jul 15 20:23:53 please use pastebin Jul 15 20:23:57 debian@arm:~$ [115631.104495] cpsw 4a100000.ethernet eth0: Link is Up - 10Mbps/Full - flow control off Jul 15 20:24:31 ok so that sounds like misconfiguration Jul 15 20:24:42 sorry what is pastbein Jul 15 20:24:47 pastebin.com Jul 15 20:25:02 or one of many similar sites for sharing chunks of text Jul 15 20:25:05 (like, say, a log) Jul 15 20:25:16 crap, it is blocked by my company Jul 15 20:25:30 hastebin.com ? paste.debian.net ? Jul 15 20:26:17 also, why the *hell* does your company block pastebin? Jul 15 20:26:52 http://paste.debian.net/976676/ Jul 15 20:27:03 got the pate.debian works Jul 15 20:27:21 my company blocks pastebin too, the IT dept are losers Jul 15 20:28:10 oh you forced it to 10Mbps Jul 15 20:28:43 so... it's probably not doing link detection and you're getting a bullshit "link up" Jul 15 20:29:16 the switch doesn't offer a phy-like mdio interface? Jul 15 20:29:38 oh never mind, you actually got rx packets even Jul 15 20:30:29 even if it has MDIO , the cpsw driver can work with it? i don't need to change the cpsw code? Jul 15 20:31:00 how can i verify that these RX/TX are valid packets? Jul 15 20:32:00 if the rx packets were invalid then they would be listed as RX errors Jul 15 20:32:12 so they pass the CRC check Jul 15 20:32:34 it is still possible something is going wrong with transmitted packets Jul 15 20:32:51 or it might be an issue with your switch Jul 15 20:32:55 so it's a probably problem with the TX? Jul 15 20:33:28 I'd suggest debugging this as a networking issue first rather than a hardware issue.. i.e. use a packet sniffer to see if anything transmitted from your board shows up on the other switch ports Jul 15 20:33:52 I don't suppose you have activity leds on the ports? Jul 15 20:34:58 also check the registers of your switch. presumably it has statistics counters where you can see if it's receiving any packets on the port that connects to the AM335x, or whether it's receiving corrupted packets Jul 15 20:34:59 mmm , the other ports works fine. Jul 15 20:35:10 a-ha! great idea Jul 15 20:35:31 okay i'll check this... thanks.. Jul 15 20:35:35 it might also have settings for applying internal delay on the clock Jul 15 20:35:46 even with 10M ? Jul 15 20:36:12 oh right, I'm not really sure how that works with rgmii Jul 15 20:39:37 ah right, at 10/100 it doesn't use DDR Jul 15 20:41:12 wow the rgmii spec sucks Jul 15 20:41:26 really not clear at all Jul 15 20:49:04 fancy, I just stumbled over some gigabit phy that has registers where you can add delay to each rgmii pin (in or out) in steps of 0.06 ns Jul 15 20:49:27 hehe Jul 15 20:49:37 that's pretty crazy Jul 15 20:50:02 another thing, in DT when i add group to am33xx_pinmux will it always assign this pinmuxing ? (even if it's not related to any other module at DT ?) Jul 15 20:50:21 no, it needs to be requested by some driver Jul 15 20:51:10 or, well, the default pin config of a device is automatically requested by default I think Jul 15 20:51:25 but merely adding a child node to &am33xx_pinmux does not suffice Jul 15 20:51:54 i see , so i need to find the default group ? Jul 15 20:52:00 ? Jul 15 20:52:39 https://github.com/RobertCNelson/dtb-rebuilder/blob/4.9-ti/src/arm/am335x-bone-common.dtsi#L380-L382 Jul 15 20:52:39 i want to add a pinmuxing to something to debug , and i don't have any driver to attach to it Jul 15 20:52:47 ah Jul 15 20:54:05 you can probably just add a silly device that gets automatically configured (e.g. use compatible = "simple-bus"; ) and attach pinmux to that Jul 15 20:55:11 zmatt : Okay I see this. thanks. you are very helpful guy ! :) Jul 15 20:56:16 http://paste.debian.net/plain/976682 Jul 15 20:56:19 something like that Jul 15 20:56:36 haven't tested it but I'm pretty sure that should work Jul 15 20:58:22 excellent! Jul 15 21:00:48 internal delay btw does sound like a good candidate for trouble Jul 15 21:01:59 the am335x does not support internal delay and requires the necessary clock delay to be implemented on pcb or in the phy Jul 15 21:02:35 so assuming none is implemented using pcb trace delay, you'll need to ensure internal delay is enabled in the switch for both rx and tx clock Jul 15 21:04:08 probably the switch will by default use internal delay for the rx clock it transmits but not the tx clock it receives Jul 15 21:04:26 which fits your symptoms Jul 15 21:07:31 great, that sounds promosing. i'm on it... i'll try to turn the internal delay on and check the packets error report Jul 15 22:46:14 jsnuhskzsi Jul 15 22:46:21 jjsnjdjis Jul 15 22:46:37 hi frns Jul 15 23:08:46 ... Jul 16 01:16:58 Hi there, I've been fighting with my beagle logic prudaq for quite some time now, and I have it running under /dev/beaglelogic, and the overlay seems to load correctly. I tried to sample some signales by connecting P8.46 to an external signal generator (I can see the signal on the wire with my scope). I also tried to connect the reference daq's reference voltage. On the webinterface, I can that the measurement, but Jul 16 01:16:58 everything comes out 0. Same when I try something like "hexdump -d /dev/beaglelogic | head". My kernel is 4.9.36-ti-r46. Any ideas how to further debug this? How can I figure out what element in the whole chain causes this? https://www.irccloud.com/pastebin/U8I4OZRI/ Jul 16 01:19:13 And here my current uEnv https://www.irccloud.com/pastebin/cBE58TTB/ Jul 16 01:19:57 do I have to choose a different pru overlay for the prudaq? Is that why? Jul 16 02:45:06 +- **** ENDING LOGGING AT Sun Jul 16 03:00:00 2017