**** BEGIN LOGGING AT Sun Jan 08 03:00:01 2017 Jan 08 03:00:34 use a button? Jan 08 03:00:40 pfft Jan 08 03:01:11 :-) Jan 08 03:01:13 * rah -> zzz Jan 08 03:01:16 well you can just go straight to wiping u-boot in eMMC, then it'll automatically fall back on the next device in the boot order: uSD Jan 08 03:01:43 but then you won't have a bootable system until whatever you put in the card is bootable Jan 08 04:27:49 Spidler: dm-verity might be interesting to you... although it was designed for verified boot, it also looks quite useful to me to catch an unexpectedly modified data block as early as possible Jan 08 04:42:11 I'm having a hell of a time with the BBGW - i'm trying to use pin 8.12 (pin #44 - GPIO2_4) as a digital output. But when I load a dto that has that pin, dmesg tells me: pin 44e10830.0 already requested by 47810000.mmc; cannot claim for ocp:blahblah . The BBB/BBG don't use this pin for mmc, and the seeed wiki shows that the pin should be available. What can I do? Jan 08 04:44:30 clockman_: can you try to see what my show-pins utility shows? https://github.com/mvduin/bbb-pin-utils Jan 08 04:44:50 hmm, 47810000 ... I think that's mmc2, used by the wifi Jan 08 04:45:35 the bbgw really sucks, instead of reusing the pins of the sacrificed ethernet for wifi (like the BBB wireless does) they left those pins unconnected and sacrificed even more pins of the expansion headers Jan 08 04:46:28 hey... now THAT's a handy tool Jan 08 04:46:34 thanks :) Jan 08 04:46:50 yea, it shows mmc2 d0 Jan 08 04:47:09 should i scrap this and get a bbb-wireless Jan 08 04:47:45 I have like 6 pins that I use for my digital IO, and 4 of them are used for mmc Jan 08 04:48:21 if you can't relocate to different pins then I'd suggest one of the other two wireless BBB-derivatives yes Jan 08 04:48:50 I already didn't get a good impression from the bbg, but the bbgw made seeed sink even further in my esteem Jan 08 04:49:15 yea, I already have about 15 circuit board made. working just fine with BBB and BBG, but as soon as i wanted built-in wifi.... seeed hates me. Jan 08 04:49:58 it's really dumb... the set of pins used by ethernet includes all pins needed for the wl18 chipset Jan 08 04:50:34 moreover that set of pins has a separate I/O supply, so seeed could have powered that on 1.8V and thereby avoided the need for level-shifters (since the wl18 chipset requires 1.8V I/O) Jan 08 04:51:27 (the bbbw didn't have that option since they chose to use the OSD335x, which ties all I/O supplies together) Jan 08 04:51:40 maybe their engineers are trying to sabotage the company Jan 08 04:52:28 the bbe (beaglebone enhanced) used an usb-wifi chipset instead, and didn't sacrifice ethernet but in fact upgraded it to gigabit Jan 08 04:54:33 so if i'm trying to keep as much stuff off usb as possible, i should try the BBBW? Jan 08 04:54:59 if you don't need ethernet, then that seems a good option Jan 08 04:55:11 I don't have any personal experience with it, I consider ethernet to be a hard requirement Jan 08 04:55:40 sounds good. I'll order one and see how it goes. Thanks for the suggestion (and showpins) Jan 08 04:55:47 you're welcome Jan 08 04:56:12 keep in mind the pin descriptions (leftmost column) are somewhat specific to the BBB of course Jan 08 04:57:03 https://goo.gl/Jkcg0w this is my pins spreadsheet, which the show-pins utility is based on Jan 08 04:57:25 it contains both generic info on the am335x, and BBB-specific tabs (marked in orange) Jan 08 04:57:43 I personally use the P9/P8 tabs (or printouts thereof) often as reference cards Jan 08 14:15:06 I'm following the getting started instructions for the BeagleBone Black, and I can't connect to the server on my BeagleBone. I am running Arch Linux on my laptop and I have set the wired interface to "up." Can someone suggest what I might be doing wrong? Jan 08 15:09:36 matt1: did it get an ip ? Jan 08 15:09:40 (dhcp) Jan 08 15:11:38 note that if you have a lan with dhcp server you can use ethernet instead of the usb networking; in that case you should be able to locate the bbb by name Jan 08 15:12:38 I'm not sure how to see if it has an IP. In the ip link output I see "link/ether b0:d5:cc:fa:b6:12 brd ff:ff:ff:ff:ff:ff" at the end of the entry for the wired connection; is this first hex string an ip address? Jan 08 15:13:38 no Jan 08 15:13:42 Would I run "# dhcpcd enp0s4f1u1"? (enp0s41u1 is the name of my wired interface) Jan 08 15:13:44 ip link doesn't show IPs, ip addr does Jan 08 15:13:55 no Jan 08 15:14:28 sorry, since you're using arch I kinda assumed you'd be familiar with how IP networking works :) Jan 08 15:17:21 oh wait dhcpcd, that's a client... if with "wired" you mean the usb networking interface then yes maybe Jan 08 15:17:32 I'm not used to having to do such a thing manually, on any distro Jan 08 15:18:25 also here it doesn't appear as enp* but just usb0, but that depends on udev rules hence may very well differ per distro Jan 08 15:18:43 Thanks for your help and apologies for being a bit slow with the IP stuff. ip addr shows a inet6 entry but not an inet entry, which makes me think it's only getting an IPv6 address. Jan 08 15:19:26 And I'm not sure how to navigate to an IPv6 address on FF or Chrome or any other browser. Jan 08 15:19:39 yeah IPv6 autoconfigures by default Jan 08 15:21:15 if you have avahi installed (specifically libnss-mdns) then you might actually be able to locate the beaglebone by name, but link-local addressing isn't always well supported by applications :) Jan 08 15:21:19 Just got it working! I ran "# dhcpcd [interface]" from my laptop and it gave it an IPv4 address. Jan 08 15:21:39 welcome to Arch, where nothing works out of the box ;-) Jan 08 15:21:54 Haha, yes it has its pros and cons. Jan 08 15:22:56 Thanks for your help and the "ip addr" reminder. Jan 08 15:23:08 you're welcome Jan 08 18:06:26 hello! I have a question. Where the idea for name "Beagle" come from? Jan 08 18:11:00 hello! I have a question. Where the idea for name "Beagle" come from? Jan 08 18:12:47 the logo does have the beagle dog in ;) Jan 08 18:13:59 good question, but no idea Jan 08 18:14:27 (same could be asked w.r.t. the pandaboard) Jan 08 18:53:18 The panda board is likely because it was done in the far east and it was another animal. Chinese like Panda's but they aren't particularly cute or cuddly in reality. I guess panda was a better animal and bigger? (than a beagle). I can't give a story for the name of the beagle board however. Jan 08 19:07:55 well, the pandaboard went extinct anyway ;P Jan 08 22:05:16 True after I bought 2 it exploded because TI discontinued the entire line. That's what I get. Jan 08 22:20:26 you bought two pandas and they exploded? Jan 08 22:21:19 I have a pandaboard5 ;-) at least I found some references that seemed to indicate the omap5-uevm was intended to be the "pandaboard 5" Jan 08 22:27:53 Yeah, my plan was to make a linux box that could "hang" on a DMZ line and be somewhat difficult to mess with by miscreant websites. The panda boards at least work well I even got them with boxes that had built in reset buttons so I could see what "troubled" sites did and hit reset when all else failed. Jan 08 22:29:36 It boots relatively quickly with the "limited" Image I found but I didn't have time to rebuild a debian distro on it. Maybe I could make it work with Gentoo since that is mostly source distributed. Jan 08 22:30:12 I don't see why the distro would have much impact, as long as the kernel works it doesn't hugely matter Jan 08 22:32:58 for my omap5-uevm I actually just took the base image I use for beaglebones and just replaced the kernel and bootloader Jan 08 22:33:01 works fine Jan 08 22:45:56 bbl Jan 09 01:56:34 hmm, I had expected the dTLB thrashing at 31+ pages, but I wonder what phenomenon is causing some slowdown for 12+ pages.... https://hastebin.com/raw/exelurohoz Jan 09 01:59:27 but ok, real-world TLB miss penalty is 36 cycles, not the minimum of 24 cycles mentioned in the cortex-a8 TRM Jan 09 02:16:30 ok, fixing an issue in the test code reduced time per inner iteration to 3 cycles (I already thought 4 was a bit much) but the strange jump at 12 pages remains... https://hastebin.com/raw/evucumequs **** ENDING LOGGING AT Mon Jan 09 03:00:00 2017