**** BEGIN LOGGING AT Tue Oct 20 10:59:57 2020 Oct 20 13:39:48 2 years ago, armpit and I were drinking in the Brewdog Oct 20 14:27:25 moto-timo: Yes. Make sure to get the 1.4 version, and fab it on 0.8 MM boards Oct 20 14:28:55 moto-timo: I usually fab through OSH park, but they can't do 0.8 MM 4 layer boards, so I had to go through pcbway with a minimum of 10... also solderpaste, a stencil, and a heat gun are pretty much mandatory to fabricate them by hand Oct 20 14:32:09 * JPEW really needs to finish my blog posts about that Oct 20 19:17:55 denix: I forgot to sent that gatesgarth patch for meta-ti which I did now please apply it to master its breaking otherwise Oct 20 19:22:00 its the last layer in my setup left to have this Oct 20 19:39:50 khem: thanks Oct 20 21:23:53 JPEW: thanks for the confirmation. I assumed as much. Did you have any trouble ordering the BOM? Oct 20 21:24:42 thank you everyone that patched the various layers for gatesgarth Oct 20 21:55:10 moto-timo: I manually ordered the parts on DigiKey... the BOM was only partially useful there. It needed fixing up. I had no trouble getting the boards fabbed with the gerber files they provide Oct 20 21:57:00 moto-timo: I think I made a BOM on digikey I can share if you want Oct 20 22:46:27 JPEW: that would be great Oct 20 22:47:02 Someone in Switzerland has boards available and they said the SD card holder wasn't available on Mouser/DigiKey Oct 20 22:47:11 I figured I'd verify that Oct 20 22:47:25 literally PCBs and the socket... not assembled boards Oct 20 22:48:00 and qty 10 is no problem, I usually do that for PCBs anyway Oct 20 22:48:47 for USB 3.0 sticks: https://www.sabrent.com/product/USB-SW30/usb-3-0-sharing-switch-for-multiple-computers-and-peripherals-led-device-indicators/ Oct 20 22:48:59 but I need to hack the switch with an external control Oct 20 22:49:08 or reverse engineer the circuit Oct 20 22:49:45 (not all my targets have sd card sockets) Oct 20 22:54:02 moto-timo: Oh, ya the SD slot is available from some other site.... can't recall ATM Oct 20 22:56:52 moto-timo: https://drive.google.com/file/d/1zlmzoKIqxFemOK4LzCELqG_glGgsmwWQ/view?usp=sharing Oct 20 22:57:18 Hmm, blue LEDs are out of stock... I've had a lot of trouble with them lately Oct 20 23:07:27 hmm, qemu-system-native for x86 fails for me for the x86_64-poky-linux target, with a bitops.h:20:41: error: missing binary operator before token "(" error, complaining about #define BITS_PER_LONG Oct 20 23:07:32 did anyone see something like that? Oct 20 23:07:36 I'm on dunfell# Oct 20 23:11:32 that i beleive is a kernel header change infecting glibc and breaking qemu Oct 20 23:12:58 should find details of fix on qemu list Oct 20 23:13:22 oh ok, so its a known thing that's not fixed on yocto yet Oct 20 23:13:26 will try to find it Oct 20 23:13:34 thanks! Oct 20 23:22:40 JPEW: thank you Oct 20 23:23:03 JPEW: I'm going to ping adafruit and see if they might consider building them as an actual product Oct 20 23:23:49 JPEW: although I kind of want the SD male part of that to be on a flex board Oct 20 23:36:40 XorA: one thing I don't quite get - the problem is with the host kernel, right? i.e. qemu-native builds for host kernel headers, or does yocto install some sanitized native kernel headers version? Oct 20 23:45:16 yeah i think so Oct 20 23:45:23 the first part i mean Oct 20 23:45:55 ok.. so now I know what to patch.. i.e. it has to be outside of yocto Oct 20 23:45:56 sorry the issue looks familiar but ive never personally hit it Oct 20 23:46:29 i guess your on fedora or arch or debian sid? Oct 20 23:46:54 hmm, I am not sure what they installed on that server for me, I usually ask for centos but I think this is some ubuntu LTS box with a 4.19 kernel Oct 20 23:47:06 so it needs the backport... Oct 20 23:47:27 maybe ive sent you on wrong path then Oct 20 23:47:31 would be nicer if I could use some kernel-native stuff to make sure the build does not depend on the host, but I guess that's not easily possible? Oct 20 23:47:50 I am not sure you did, I found a number of reporits about this error, not only qemu related Oct 20 23:47:58 and it sounded like it was a missing backport in older kernels Oct 20 23:48:08 so your hint wasn't off Oct 20 23:49:44 I think I need something like that https://lore.kernel.org/lkml/20200213142147.17604-1-borntraeger@de.ibm.com/ Oct 21 00:02:45 can I forece a native recipe to use the non host kernel includes? Oct 21 00:03:12 or would that be "bad" in the sense of what the compiled result is, since it may not match the system? Oct 21 00:13:31 armpit: meta-odroid changes to switch bbappends to linux-yocto are causing it to override COMAPTIBLE_MACHINE unconditionally see https://github.com/akuster/meta-odroid/pull/62 Oct 21 00:14:03 armpit: Please review and merge this as my CI is currently broken https://github.com/YoeDistro/yoe-distro/runs/1283791808 Oct 21 00:21:23 khem: btw, who is maintaining the h2? Oct 21 00:22:52 Jin^eLD: its armin Oct 21 00:23:04 I only care for c2 and c4 Oct 21 00:24:54 armpit: did you update the BIOS recently and did they drop legacy non-uefi mode on the H2 as well? :> Oct 21 00:35:40 will check back in tomorrow, nite Oct 21 02:01:57 moto-timo: Hmm, ya. I'd be a little leery of the flex connector given the history of the device. One of the revs (1.4 maybe)? went to a 4 layer PCB b/c they where having trouble with the impedance matching on the SD lines Oct 21 02:02:42 moto-timo: Ya adafruit making them would be great! I would pay some money to save the hours it takes to fab them :) Oct 21 03:24:01 Hello...I get this error when trying to start toaster using "source toaster start" Oct 21 03:24:10 Failure while trying to setup toaster: Problem installing fixtures: no such table: orm_build__old Oct 21 03:24:21 More details here --> https://pastebin.com/qfMusX7V Any ideas? Oct 21 03:24:31 All toaster dependencies are installed inside python virtual env ie. I am using python virtualenv for this **** ENDING LOGGING AT Wed Oct 21 10:59:57 2020