**** BEGIN LOGGING AT Wed Dec 20 03:00:02 2017
Dec 20 14:27:07 <ant_work> greguu, btw
Dec 20 14:27:09 <ant_work> [PATCH, RFT] ARM: use --fix-v4bx to allow building ARMv4 with future gcc
Dec 20 14:27:14 <ant_work> today
Dec 20 14:32:36 <Varti> hi
Dec 20 14:32:58 <ant_work> hello
Dec 20 14:33:04 <ant_work> greguu, ah, and  bootsplash: Add animation support
Dec 20 14:33:29 <Varti> ARMv4 would be useful for Collie only IMHO
Dec 20 14:33:45 <ant_work> yes, someone up there still cares
Dec 20 14:34:38 <Varti> that's nice :)
Dec 20 14:34:58 <Varti> I see now, those are patches for /drivers/video/fbdev/core/bootsplash.c, isn't it?
Dec 20 14:37:23 <ant_work> yes
Dec 20 14:38:27 <Varti> and, if I understood correctly, gcc 7.x does not support -march=armv4 anymore, but it still supports -march=armv4t
Dec 20 14:38:59 <ant_work> yes, gcc and binutils are patched
Dec 20 14:39:43 <Varti> with normal armv4 CPUs requiring now --fix-v4bx as a workaround for the linker
Dec 20 14:40:04 <ant_work> yep
Dec 20 14:40:45 * Varti wonders if greguu will ever try to compile a kernel for the Collie
Dec 20 14:41:28 <Varti> everything else will have to be anyway recompiled too, there are no binary packages for ARMv4
Dec 20 14:42:03 <ant_work> note that you could use those binaries on pxa zaurus without much loss
Dec 20 14:43:00 <ant_work> what is important is the ABI. armv4 historically was OABI
Dec 20 14:43:42 <ant_work> followed by the actual EABI spec
Dec 20 14:44:25 <ant_work> this means that the older armv4 files for collie from elsi etc. will not run :/
Dec 20 14:45:03 <ant_work> afais only OpenEmbedded (and OpenZaurus before) used EABI for collie/armv4
Dec 20 14:46:43 <Varti> aha
Dec 20 14:48:27 <Varti> I thought that ARMv5 had instructions not supported on ARMv4
Dec 20 14:48:41 <ant_work> if the kernel is compiled with OABI compatibility then things are ok...I think I disabled that ;)
Dec 20 14:49:10 <ant_work> Varti, right, you can run armv4 code on armv5te but not the opposite
Dec 20 14:49:10 <Varti> heh, I don't think that would feature be useful in kexec :)
Dec 20 14:49:18 <Varti> ok
Dec 20 14:49:34 <ant_work> actually you could use a singl ekexec for all Z
Dec 20 14:50:13 <ant_work> ..but tere are some little configuration changes (delays) so we have different versions
Dec 20 14:50:27 <ant_work> this should be cleaned out...
Dec 20 14:58:20 <Varti> I seem to recall you talked in the past of having a single kexec binary for all the models
**** ENDING LOGGING AT Thu Dec 21 03:00:00 2017