**** BEGIN LOGGING AT Sat Jul 01 03:00:02 2017 Jul 01 09:58:49 Hello! Anyone who saw this already? https://paste.kde.org/p6pebyrzz Jul 01 14:25:32 RP: fixed the sysklogd issue that was failing with gcc hardening patches on AB, now my local builds are clear Jul 01 14:26:43 RP: I was wondering if we should replace sysklogd with rsyslog on poky-lsb, all major distros have dropped sysklogd over 5 years ago and moved over to either rsyslog or systemd Jul 01 14:27:59 I just sent out a v3 of my pull request which should address the issue and I also bundled glibc 2.26(rc) upgrade patches in same pull Jul 01 17:38:12 Hello, is it possible to exclude QEMU from SDK/extSDK build? Jul 01 20:52:56 Hey - is there a way I can build a custom initramfs? Jul 01 20:54:32 use an initramfs recipe with the cpio / cpio.gz image fstypes. find oe-core -name \*initramfs\*.bb. now what you *do* with that, whether it's alongside the kernel and you configure your bootloader to use it, or bundle it with the kernel, that's a different thing Jul 01 20:57:05 Thanks kergoth Jul 01 20:57:10 np Jul 01 20:57:48 i have no idea how the bootloader configuration piece is done with yocto, that'd depend on whether you build your bootloader at all with it, which bootloader it is, whether you use wic or not, etc Jul 01 20:57:55 I'm just having a bear of a time booting up the kernel on my board based off the NXP Sabre iMX7d Jul 01 20:57:59 iirc there's a variable to bundle it into the kernel, see kernel.bbclass, but i've never tested it Jul 01 20:58:15 It seems to ... well it doesn't "hang", but it waits at UTP: waiting for device to appear Jul 01 20:58:46 i assume you've tried making the kernel more verbose? iirc there's a commandline option to change the console log level for kernel log messages Jul 01 20:58:46 I've snipped out everything regarding drivers/usb/gadgets (which is the only place that UTP appears in the kernel) Jul 01 20:59:05 that'd be myr ecommendation, try to get more messages to figure out exactly where in the boot process it stops Jul 01 20:59:06 but it's still doing that. I've checked my rootfs and it doesn't have anything related to UTP in it Jul 01 20:59:18 so am I looking in the wrong place by checking the initramfs? Jul 01 20:59:50 Yes - I set the log level to 7, and enabled every debug option I know of Jul 01 20:59:57 huh, that's odd, i'd start by making sure it's actually using your new kernel at all, if it's still waiting for UTP when you disabled anything utp-related. Jul 01 20:59:58 * kergoth shrugs Jul 01 21:00:07 I also put my own custom printk's in the kernel source to try to narrow it down Jul 01 21:00:15 no idea what utp even is, maybe you could check with other folks using your board? the manufacturer, or layer maintainer? Jul 01 21:00:43 Good call - for a while I was accidentally booting from the emmc image, but the printks helped me figure that out and ensure I'm using my kernel Jul 01 21:00:49 * kergoth nods Jul 01 21:01:01 UTP = micro transfer protocol Jul 01 21:01:10 that's usually m y first suspect.. if somehting seems inexplicable, check my assumptions :) Jul 01 21:01:10 (I didn't know about it till I encountered this) Jul 01 21:01:20 Yep :) Jul 01 21:01:32 Problem is nobody in the world other than me is using this board atm Jul 01 21:01:40 that's unfortunate Jul 01 21:01:48 It's custom in-house design based loosely off the NXP Sabre Jul 01 21:01:50 Yeah it is :( Jul 01 21:02:03 do you have serial connectivity? Jul 01 21:02:07 The hardware designer, in his infinite wisdom, re-routed a bunch of stuff unnecessarily Jul 01 21:02:19 Like... the i2c0 used to go to the power management unit Jul 01 21:02:21 i'd consider using the kernel debugger, connect to it over gdb, worst case step through it Jul 01 21:02:28 ah Jul 01 21:02:42 but now it goes to our additional realtime clock, and then he *added* an i2c bus to go to the pmu Jul 01 21:02:53 why not just leave i2c0 on the pmu and add a bus for the rtc? Jul 01 21:03:02 ... and then he left the company. Before I joined. Jul 01 21:03:27 kergoth: I would love to do that, but I don't know how. Do you have a link explaining how to use the kernel debugger over serial? Jul 01 21:04:05 Before this job (which I've had for an entire month) I was strictly software, working on video games and web apps and desktop apps. Jul 01 21:04:23 Now I'm diving into schematics, multimeters, 4000 page datasheets, and the kernel :-p Jul 01 21:04:31 I'm picking it up ok, but there's a TON to learn Jul 01 21:04:56 The device tree alone is quite a bear Jul 01 21:07:56 Not complaining though - it's fun tackling totally new challenges. Thus me voluntarily working on a Saturday Jul 01 21:12:03 it's pretty easy over serial, iirc you enable it in the kernel config, add 'kgdbwait' or something to the commandline, and then connect to the serial port with a command in gdb. you can probably google for the exact steps Jul 01 21:12:06 i haven't done it in a while Jul 01 21:13:09 Ok. I'll look that up. Thanks for the suggestion. I wasn't sure what debugging options were actually available at the kernel level **** ENDING LOGGING AT Sun Jul 02 03:00:01 2017