IRC log for #oe on 20200529

01:34.47*** join/#oe yann|work (~yann@91-170-159-152.subs.proxad.net)
04:15.14*** join/#oe davidlt (~davidlt@78-63-27-146.static.zebra.lt)
05:26.52*** join/#oe AndersD (~AndersD@h83-209-96-136.cust.a3fiber.se)
05:57.51*** join/#oe kroon (~kroon@213.185.29.22)
06:05.14*** join/#oe rob_w (~rob@unaffiliated/rob-w/x-1112029)
06:07.11*** join/#oe AndersD_ (~AndersD@h83-209-96-136.cust.a3fiber.se)
06:11.06*** join/#oe AndersD__ (~AndersD@h-17-226.A137.corp.bahnhof.se)
06:13.36*** join/#oe leon-anavi (~Leon@78.130.197.211)
06:34.58*** join/#oe adelcast (~adelcast@2605:6000:101c:3b5:bf3:a7d4:c2b5:cbb1)
06:38.12*** join/#oe diego_r (~diego@mob-2-41-210-190.net.vodafone.it)
06:42.18*** join/#oe kroon (~kroon@213.185.29.22)
07:04.55*** join/#oe ao2 (~ao2@87.19.23.254)
07:13.15*** join/#oe sstiller (~sstiller@p200300f07f12ef018d67a76084e9e25b.dip0.t-ipconnect.de)
07:17.22*** join/#oe jkridner (~jkridner@pdpc/supporter/active/jkridner)
07:22.08*** join/#oe Leatherface (~leatherfa@helvetet.eu)
07:24.32*** join/#oe cpriouzeau (~cpriouzea@164.129.115.76)
08:02.56*** join/#oe infobot (ibot@c-174-52-60-165.hsd1.ut.comcast.net)
08:02.56*** topic/#oe is OpenEmbedded Developer Lounge | Web: http://openembedded.org | Repositories: http://git.openembedded.org/ | Primary Repo Mirrors: https://github.com/openembedded | This is not a distro or machine support channel
08:04.39*** join/#oe AndersD_ (~AndersD@h83-209-96-136.cust.a3fiber.se)
08:07.47*** join/#oe AndersD (~AndersD@h-17-226.A137.corp.bahnhof.se)
08:18.19jaeckelkhem: thanks! I wanted to avoid modifying src_uri, but I also don't see another option (protocol=http is already enabled)
08:24.43*** join/#oe florian (~florian_k@Maemo/community/contributor/florian)
08:24.48*** join/#oe clopez (~tau@neutrino.es)
08:52.08*** join/#oe BlastuR (~magnus@unaffiliated/blastur)
09:17.16*** join/#oe kroon (~kroon@213.185.29.22)
09:27.47*** join/#oe dlan (~dennis@gentoo/developer/dlan)
09:37.18*** join/#oe vquicksilver (~nobody@gentoo/contributor/vquicksilver)
09:45.16*** join/#oe mihai (~mihai@unaffiliated/mihai)
10:29.17*** join/#oe hpsy (~hpsy@85.203.15.116)
11:42.33*** join/#oe berton (~berton@181.220.84.90)
11:49.48*** join/#oe leon-anavi (~Leon@78.130.197.211)
13:07.26*** join/#oe ubuntu (~ubuntu@unaffiliated/letothe2nd)
13:08.34*** join/#oe ericch (~ericch@pool-108-34-251-214.prvdri.fios.verizon.net)
13:12.11*** join/#oe troulouliou_div2 (~troulouli@unaffiliated/troulouliou-div2/x-0271439)
13:22.26*** join/#oe rcw (~rcw@45.72.195.246)
13:26.00JPEWJin^eLD: There are examples of baremetal toolchains out there... I don't recall where ATM
13:27.57*** join/#oe tprrt (~tprrt@upc31-1-78-208-110-13.fbx.proxad.net)
13:28.33Jin^eLDJPEW: I only found the RTOS layer example with a toolchain for qemuarm, but I had a hard time to come with the right machine.conf values for a new architecture
13:29.41Jin^eLDin the end I got to a point which allowed me to compile gcc-cross-avr but the resulting toolchain was unusable, I guess I messed up the settings after all
13:37.44Jin^eLDI guess my problem is anyway not multiconfig as such, but rather my lack of hardware knowledge and knowing what options to use for building the cross compielr
13:59.30*** join/#oe sakoman (~steve@rrcs-66-91-142-162.west.biz.rr.com)
14:39.09*** join/#oe khem (~khem@unaffiliated/khem)
14:39.32*** join/#oe hipr_C (~Thunderbi@89.187.178.5)
15:04.50JPEWJin^eLD: I know there are examples out there of baremetal machine confs :)
15:05.27Jin^eLDwell, if you run across a non-arm one, please let me know ;)
15:06.16JPEWRP: ^^
15:06.52JPEWRP: I though I recalled you saying you had done an AVR machine conf at one point
15:07.19Jin^eLDI saw an avr32 in siteinfo.bbclass
15:07.41Jin^eLDbut I am not sure if thats compatible to those 8 and 16 bit avr mcus?
15:10.20Jin^eLDand then I was also not sure about SITEINFO_BITS
15:29.57*** join/#oe florian_kc (~florian_k@Maemo/community/contributor/florian)
16:39.04RPJPEW: I have, yes
16:39.41RPneeds to clean that up so its sharable
16:40.20Jin^eLDRP: which avr did you add?
16:41.44RPJin^eLD: I was using an 8 bit one with custom code, ATmega8535 I think
16:42.28Jin^eLDI think that comes very close to what I am interested in, I am trying to bundle the mcu part of the klipper 3d printing firmware in a linux image
16:43.34Jin^eLDso I need to compile it for a bunch of mcus, like atmega1284p, atmega2560, 328.. long lit
16:43.36Jin^eLD*list
16:44.01Jin^eLDbut I think all of those can be handled by the same compiler configuration
16:44.30Jin^eLDand then another bunch would be for cortex m0/m3/m4
16:45.05Jin^eLDif you have anything to show (does not even need to be fully working) - I think this would already help a lot
16:45.26RPJin^eLD: I have a couple of systems with ARM/Linux with an MCU alongside so I had recipes for a build system cross compiler and an on target one
16:45.38RPso you could ssh into the arm system and recompile the AVR code
16:46.00Jin^eLDthat's what I wanted to avoid... my goal was to prebuild everything and just ship the binaries
16:46.14RPJin^eLD: right, this was just for hacking during development
16:46.32RPI made it do build time too
16:46.38Jin^eLDcurrently klipper guys suggest to install some debian spin and then manually add klipper, build the mcu fw on the target (their setup is aimed at a raspi) and stuff like that
16:46.56Jin^eLDand I was not too happy with that and also not with the raspi focus, so started to hack a minimalistic distro that would work out of the box and have everything preinstall
16:47.09Jin^eLDnowhere nearly finished https://github.com/jin-eld/klipper-linux
16:47.29Jin^eLDand currently I am a bit stuck with the mcu part
16:47.36RPJin^eLD: let me see if I can quickly put the code somewhere
16:47.43Jin^eLDcool, thank you
16:49.53*** join/#oe bluelightning (~paul@pdpc/supporter/professional/bluelightning)
16:55.04RPJin^eLD: http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/t222&id=d344684a3fc161a77f10edf2cc636f54d52a5d20 - its pyro based so old but you can see the idea
16:57.06Jin^eLDwhere exactly is the avr machine conf, not seeing it so far
16:57.55Jin^eLD?
16:58.00RPJin^eLD: oh, I did something a little different, I just hacked it alongside my normal machine
16:58.22Jin^eLDah, you did not use multiconfig?
16:58.26RPJin^eLD: so I had two different toolchains and some special recipes which didn't build for MACHINE but for my avr part
16:58.36RPJin^eLD: multiconfig did not exist when I wrote this!
16:58.57Jin^eLDoh right, its pyro..
16:59.19RPThere is nothing which says you can't have two toolchains in a given build or that they have to target the same thing ;-)
16:59.33Jin^eLDhmm damn! :) then its not quite what I am looking for; I already did some hacks with pulling in m4 baremetal toolchains and hacking variables to compile for the mcu within the linux build
16:59.34*** join/#oe ao2 (~ao2@87.19.23.254)
16:59.44Jin^eLDand I think ther eis a meta-microcontroller layer specifically for AVR to do that
17:00.02RPWell, those recipes could be used standalone too
17:00.22Jin^eLDmy main issue is coming up with a meaningful machine conf for the mcu
17:00.25RPi.e. if they work in this config, as standalone machine config should be possible
17:00.50RPI'm not sure it matters too much to start with, just copy an ARM one
17:01.00Jin^eLDin theory if the machine.conf is fine, the default gcc-cross recipe would spit me out the right toolchain, rihgt?
17:01.01RPYou'll figure out the pieces over time
17:01.23RPNo, AVR is interesting enough that you'll need the config in those recipes I shared
17:01.42RPWhichever way you do it you need that config, that is the magic which gives you an AVR compiler
17:01.55Jin^eLDaah, that's the trick
17:02.24RPremember its not a glibc target for example
17:02.28Jin^eLDI got to the stage where gcc-cross-avr would build "something" but I was missing what you mentioned just now, so the resulting toolchain did not work
17:02.41Jin^eLDyes, right, but there is the baremetal TCLIBC option now
17:02.42Jin^eLDwhich I used
17:02.55RPyou probably want the avr-libc I have there
17:03.00Jin^eLDso I optimistically assumed it would do the trick :)
17:03.03Jin^eLDyes, true
17:03.27Jin^eLDthat I found out yesterday night, but I think I need the cross compiler first? or does it already need avr-libc to compile?
17:03.40Jin^eLDI'll go through your recipes after work, thanks for sharing!
17:03.44RPdon't remember. See what I put in the recipes? :)
17:04.02Jin^eLDright:)
17:04.19RPI know that compiler works. Sadly I didn't have the time to make this into something nicer on a more modern version :(
17:04.28Jin^eLDthis is another "parallel toolchain" approach: https://github.com/schnitzeltony/meta-microcontroller
17:05.08Jin^eLDwell, Alejandro published some stuff on how to get this all set up with multiconfig
17:05.32RPright, that looks like similar ideas with the recipes
17:05.33Jin^eLDso I think that would be the "modern" way, but the exampls are for microblaze or qemuarmv5
17:06.00Jin^eLDand I was failiing to get multiconfig going due to difficulties coming up with a meaningful machine.conf
17:06.12Jin^eLDit was complaining about tunes and features and what not, basically the sanity checker did not like me :)
17:06.29RPright, at that point its a full new arch and it will get a bit grumpy
17:07.01Jin^eLDyes, that's what I saw then, I ended up in siteinfo.bbclass and was then wondering if adding SITEINFO_BITS=8 was a good idea :)
17:07.26RPI'd just use a dummy arm machine to start with ;-)
17:07.43Jin^eLDand then I was also not quite sure how to deal with the "variants" i.e. atmega1280, atmega2560 and so on, all of them are technically compilable with an "avr2" gcc
17:07.59RPthat is where the tune file would come in
17:08.08Jin^eLDah, makes sense
17:08.15RPif dome properly anyway
17:08.44Jin^eLDdummy arm.. so for starters I should not bother with SITEINFO_BITS but make sure I have the endiannes right and then work out the magic from your gcc recipes?
17:11.39Jin^eLDgenerally it'd be awesome if we had mcu layers for most common stuff, like the m4 (a lot of boards with two processors are out there) and multiconfig sounds like a clean way to do it
17:11.57Jin^eLDespecially since you can then pull in the package from the mcu build into your "main/parent" image
17:12.33*** join/#oe florian_kc (~florian_k@Maemo/community/contributor/florian)
17:12.38Jin^eLDbasically thats what Alejandro was showing with the Xilinx board and FreeRTOS setup
17:14.18Jin^eLDlet's see how far I get with the avr and anything worth submitting will come out of it :>
17:22.02RPJin^eLD: I'd love to do it, just never had the time or the people asking for it
17:23.49Jin^eLDI struggle a bit because I am actually a userspace guy, so arch and hardware specific details is something where lack a lot of knowledge, but if you'd have time for a few hints from time to time and maybe a review - that'd be great; I'll simply go on hacking the avr multiconfig thing in that github klipper repo
17:24.22Jin^eLDand if you steer me into a direction where it would look generic enough (and an avr machine conf should be generic enough), I could then submit something vs oe/yocto
18:14.11*** join/#oe troulouliou_div2 (~troulouli@unaffiliated/troulouliou-div2/x-0271439)
19:09.40*** join/#oe ao2 (~ao2@87.19.23.254)
20:01.15*** join/#oe hipr_C (~Thunderbi@89.187.178.5)
20:33.07*** join/#oe mattsm (~mattsm@76-205-175-243.lightspeed.austtx.sbcglobal.net)
20:40.16*** join/#oe jkridner (~jkridner@pdpc/supporter/active/jkridner)
20:40.17*** join/#oe florian_kc (~florian_k@Maemo/community/contributor/florian)
20:40.52*** join/#oe mattsm (~mattsm@76-205-175-243.lightspeed.austtx.sbcglobal.net)
21:42.00*** join/#oe jkridner (~jkridner@pdpc/supporter/active/jkridner)
21:48.03*** join/#oe ant__ (~ant__@95.250.164.184)
21:52.45*** join/#oe jkridner (~jkridner@pdpc/supporter/active/jkridner)
23:23.31*** join/#oe creich_ (~creich@p200300f6af28d810ab89d0976f0525ed.dip0.t-ipconnect.de)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.