**** BEGIN LOGGING AT Thu Oct 11 02:59:59 2018 Oct 11 08:45:10 hello. when are patches sent to the oe-devel mailing list applied, and by whom? i've been subscribed for about a week and haven't seen it happen yet. Oct 11 08:45:47 also, what's the deal with master and master-next? Oct 11 08:46:28 master-next isn't mentioned here https://www.openembedded.org/wiki/How_to_submit_a_patch_to_OpenEmbedded Oct 11 09:01:23 master is the current development branch, master-next is where patches from the list are queued for testing Oct 11 11:05:32 Hi guys! Question about OPKG and busybox: Oct 11 11:06:04 I upgraded busybox and some other packages. When doing this, the prerm of busybox removes all it's links. Oct 11 11:06:46 So the packages after busybox will fail because they can't use busybox commands any more. Oct 11 11:07:25 Is there a way around this, so for instance, make opkg upgrade packages sequentially? Oct 11 12:44:05 paulbarker, are you watching to FOSDEM stand deadlines? Oct 11 13:13:08 Crofton: Usually I don't think about FOSDEM until after ELCE, probably worth a quick check when the stand deadlines are though Oct 11 13:14:23 2nd Nov so will put it on my list for after the conference Oct 11 13:16:41 OK Oct 11 13:16:54 Just needed to make sure it is in your brain Oct 11 13:17:41 Thanks for reminding me! Oct 11 13:19:42 I figured you had it under control, but once upon a time the guy that normally submits the SDR romm forgot and was on vacatin at deadline, so I hacked something together at the very very last minute :) Oct 11 13:48:44 I'm about to request an account at https://www.openembedded.org/wiki/Special:RequestAccount but the "Terms of Service" link on that page is dead Oct 11 13:50:25 hmm, Oct 11 13:50:31 likely always been dead ... Oct 11 13:51:54 I was wondering if it was some kind of test to see who blindly accepts without checking :) Oct 11 13:52:00 hah Oct 11 13:52:08 likely boilerplate that was never filled in Oct 11 14:07:25 maelcum: you can track commits on: https://patchwork.openembedded.org/ Oct 11 16:35:08 maelcum: I maintain meta-openembedded, best way to look at patch flight is patchwork as vmeson said, if you are mailing list only person then subscribing to openembedded-commits ml might be helpful Oct 11 16:35:45 I do not send closure emails to patches like "applied" instead, I reply why I am not applying it Oct 11 16:36:09 thats less of email churn for me Oct 11 16:36:59 last update for meta-oe http://lists.openembedded.org/pipermail/openembedded-commits/2018-October/229892.html Oct 11 18:07:40 eabi/volk/1.5.0-r0/git/kernels/volk/asm/neon/volk_32fc_32f_dot_prod_32fc_a_neona Oct 11 18:07:40 smvmla.s:41: Error: selected processor does not support `pld [taps,#128]' in ARM Oct 11 18:07:40 mode Oct 11 18:08:04 but Oct 11 18:08:12 [balister@prague build]$ bitbake -e volk | grep ^ARM_INS Oct 11 18:08:13 ARM_INSTRUCTION_SET="thumb" Oct 11 19:13:54 thanks for the info Oct 11 19:14:40 (not mentioning everyone by name to avoid unnecessary distraction) Oct 11 19:35:08 does oe-core have an easy way to do a build with qemuarm setup for armv7? Oct 11 19:37:38 Crofton: standalone no, but I believe there are a few machine definitions floating around Oct 11 19:37:49 probably about time it got updated in oe-core Oct 11 19:38:34 yeah Oct 11 19:38:46 does anyone know if D.Reyna hangs out on irc? Oct 11 19:38:47 need to add for OEDEM Oct 11 19:38:56 sometimes Oct 11 19:39:16 looks like in #yocto Oct 11 19:39:34 ah, thanks Oct 11 19:40:19 silly I know Oct 11 19:41:01 bluelightning, I'm having som easm fails I've never seen before building meta-ti with master Oct 11 19:41:47 Crofton: hmm, have you pinged denix? Oct 11 19:45:06 he suggetsed meta-yocto-bsp Oct 11 19:45:20 then we balme oe-core :) Oct 11 19:47:34 Crofton, bluelightning: testing any other armv7 platform should be enough. too bad oe-core doesn't have armv7 qemu... Oct 11 19:50:14 denix: RP might say someone should send patches again (not sure) Oct 11 19:50:35 bit late for 2.6 though Oct 11 20:09:51 what happened to meta-yocto-basp? Oct 11 20:14:09 denix: meta-yocto-bsp does have beaglebone-yocto so armv7 is tested Oct 11 20:14:30 denix: and I did list changing qemuarm to v7 on the 2.7 plans Oct 11 20:14:42 (rename qemuarm -> qemuarmv5) Oct 11 20:15:36 bluelightning: off the top, would you recommend i stick with master for doing the devtool ypdd, or would sumo be good enough (i.e. any significant additions since sumo?) Oct 11 20:19:41 great Oct 11 20:20:10 I'm seeing something stupid with asm I do not understand Oct 11 20:22:56 RP: yes, that was my suggestion to Crofton as well to test against beaglebone-yocto. he has some ASM code that was tested on armv5 and armv8, so I suspect the issue might be specific to armv7 Oct 11 20:23:40 I'm pretty sure it has built for zynq at some point Oct 11 20:24:14 but only through rocko Oct 11 20:24:40 Crofton, use a modern version Oct 11 20:24:47 Crofton: have you tried meta-ti rocko as well? or you went straight to master? Oct 11 20:26:05 armpit: laverne should be all you need! Oct 11 20:26:27 bluelightning: hmm, doesn't look like it, i only count 4-ish commits to scripts/lib/devtool since Apr-2018 (do i have to look somewhere else?) and i don't think any of them are adding new features Oct 11 20:27:27 Crofton: all my machines are doing is including corresponding tune-cortexa8.inc and setting DEFAULTTUNE ?= "armv7athf-neon" Oct 11 20:28:25 needs some asm files Oct 11 20:29:06 Crofton: there were some thumb specific fixes between rocko and master... Oct 11 20:30:13 Crofton: and armv7athf/armv7vethf seem to be the default in or-core anyway Oct 11 20:30:21 oe-core Oct 11 20:42:19 | /home/balister/opensdr/ti-thud/build/tmp-glibc/work/cortexa8hf-neon-oe-linux-g Oct 11 20:42:19 nueabi/volk/1.5.0-r0/git/kernels/volk/asm/neon/volk_32f_x2_dot_prod_32f_a_neonas Oct 11 20:42:19 m_opts.s: Assembler messages: Oct 11 20:42:19 | /home/balister/opensdr/ti-thud/build/tmp-glibc/work/cortexa8hf-neon-oe-linux-g Oct 11 20:42:19 nueabi/volk/1.5.0-r0/git/kernels/volk/asm/neon/volk_32f_x2_dot_prod_32f_a_neonas Oct 11 20:42:20 m_opts.s:46: Error: selected processor does not support `sbfx r11,r1,#2,#1' in A Oct 11 20:42:22 RM mode Oct 11 20:42:51 MACHINE = "beaglebone-yocto" Oct 11 20:47:04 so same issue on non ti bsp Oct 11 20:47:18 Error: selected processor does not support `sbfx r11,r1,#2,#1' in ARM mode Oct 11 20:47:42 Crofton: like I suspected, something more generic to armv7 Oct 11 20:48:10 [balister@prague build]$ bitbake -e volk | grep ^ARM_INS Oct 11 20:48:10 ARM_INSTRUCTION_SET="thumb" **** ENDING LOGGING AT Fri Oct 12 03:00:00 2018