<ynezz>
cbeznea: yeah, we need to restart buildbot in order to add new subtarget
<ynezz>
cbeznea: while looking into that, I've just noticed, that for at91/sama7 subtarget you've introduced new cortex-a7-vfpv4 architecture, which is going to increase build resources usage for phase2 builders by additional ~8h hours
danitool has joined #openwrt-devel
Kali_ has joined #openwrt-devel
<ynezz>
cbeznea: so ideally it should simply use currently available arm_cortex-a7 or arm_cortex-a7_neon-vfpv4 builders, AFAIK there is no public documentation available about that product, so unsure how to proceed
<ynezz>
cbeznea: in other words I don't see currently much of the value to have our infra build "vfpv4 optimized" packages for single eval board sama7g5-ek
robimarko has joined #openwrt-devel
slh has quit [Quit: leaving]
slh64 has quit [Quit: gone]
floof58 has quit [Ping timeout: 480 seconds]
kopijahe has quit [Read error: Connection reset by peer]
ekathva has joined #openwrt-devel
slh has joined #openwrt-devel
slh64 has joined #openwrt-devel
guidosarducci has quit [Remote host closed the connection]
guidosarducci has joined #openwrt-devel
AtomiclyCursed has quit [Ping timeout: 480 seconds]
fda- has joined #openwrt-devel
fda has quit [Ping timeout: 480 seconds]
clayface has quit [Ping timeout: 480 seconds]
clayface has joined #openwrt-devel
ekathva_ has joined #openwrt-devel
<aiyion_>
Does somebody know whether Henrik Ginstmark has a handle on this server?
<T-Bone>
arg. I just realized ImageBuilder downloads base-files pkg, and that overrides any and all changes made to target/**/base-files.
T-Bone is now known as f00b4r0
<f00b4r0>
wonder why we even bother shipping base-files within IB
valku has joined #openwrt-devel
srslypascal has joined #openwrt-devel
goliath has joined #openwrt-devel
danitool has quit [Ping timeout: 480 seconds]
<ynezz>
f00b4r0: I assume it's for /etc/rc.common and services handling
danitool has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
<f00b4r0>
ynezz: I'm not sure I follow? target/**/base-files/ relates to board.d and upgrade stuff. And it's overwritten by the contents of the base-files packages. AFAICT none of the files in the target folder are used?
<f00b4r0>
luckily for me it seems I can get away with cp'ing them to files/ and using FILES=files
<f00b4r0>
to override the override :)
<ynezz>
you need that for DISABLED_SERVICES
<ynezz>
check prepare_rootfs target in include/rootfs.mk
<ynezz>
it will use rc.common on the host
<f00b4r0>
ynezz: I think we're talking about different things? target base-files don't have anything to do with services?
<f00b4r0>
or I'm confused?
kopijahe has quit [Ping timeout: 480 seconds]
<f00b4r0>
prepare_rootfs doesn't touch target/**/base-files. In fact I can't find anything that uses them
<f00b4r0>
ynezz: I'm sorry, I must be really dense but I don't see how that relates to the contents listed here: https://pastebin.com/R5nuGVHG
kopijahe has joined #openwrt-devel
rua has joined #openwrt-devel
kopijahe has quit [Read error: Connection reset by peer]
<ynezz>
ok, misunderstanding on my side
<f00b4r0>
maybe I didn't explain myself clearly either :) Basically I added a patch to enable support for a specific device in a 21.02 checkout. Then I built an ImageBuilder archive from the modified tree, expecting to be able to build images for that device from this IB. It only half works, because as IB fetches the upstream base-files ipk, the content of that packages are used instead of the contents of the IB's target base-files folder.
<f00b4r0>
so while the image build was successful (the DTS and makefile changes were in the IB), the base-files glue was missing from the image, resulting in a soft-bricked device.
<f00b4r0>
then again I suppose nobody does that :)
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<ynezz>
you shouldn't modify the files in place, use FILES=""
<f00b4r0>
ok, you didn't understand me
<f00b4r0>
I'm not modifying anything in place
<f00b4r0>
from a patched git checkout of 21.02, I built an Image Builder
<f00b4r0>
from that unmodified IB, I built an image
<ynezz>
ah, you forget to modify the feeds.conf ?
<f00b4r0>
the image was incomplete because the base-files glue needed, and present in the IB, were wiped by the upstream base-files packages
<f00b4r0>
it turns out that in the Image Builder target/ directory, it would seem that the various base-files subfolder are simply unused. So I wonder if that's by design, and if so, why we bother packaging them with the Image Builder
<f00b4r0>
do I make more sense now? :)
<ynezz>
well, IB uses packages specified in repositories.conf, so if you want to use yours you should add yours in there
minimal has joined #openwrt-devel
<f00b4r0>
:'
<f00b4r0>
:(
<ynezz>
you can try to delete that files
<ynezz>
and see what breaks
<f00b4r0>
I don't want to use any out of tree package
<f00b4r0>
I was expecting that a patched git tree would lead to a useable "patched" ImageBuilder. It doesn't.
<f00b4r0>
specifically because the files that are shipped with the imagebuilder are *not used*
<ynezz>
echo 'src custom file:///from a patched git checkout of 21.02/bin/packages' >> repositories.conf
<ynezz>
then you can use it
<f00b4r0>
...
<ynezz>
you just don't get how IB works, it uses binary packages specified in repositories.conf
<f00b4r0>
I understand that
<f00b4r0>
and that's fine.
<f00b4r0>
what I do not understand is: what is the use of shipping in IB under "target/*/*/base-files", folders and files which are completely unused?
<ynezz>
delete it and find out
<ynezz>
if it works without it, then it can be probably removed?
<f00b4r0>
well I wish they *were* used instead :)
<ynezz>
lol
<f00b4r0>
essentially it means you cannot build an IB from a modified tree and expect it to work. At the very least maybe *that* should be documented :)
<f00b4r0>
the point is from my perspective I'm not using a *custom package*
<f00b4r0>
I'm only touching target-dependent files
<f00b4r0>
and these files are shipped with the image builder.
<f00b4r0>
put differently: cherry pick any device support patch from master into a checkout of 21.02, build an image builder, try to use it to build a 21.02 image for said backported device, watch it fail.
<f00b4r0>
maybe that's by design, and that's ok. I just learnt the hard way :)
<ynezz>
you need to point IB at your binary packages in repositories.conf
<f00b4r0>
I don't have any binary packages
<tmn505>
f00b4r0: IB builds images mostly from ipk, which base-files is also one (base-files are merged with target ones), if You don't specify Your custom packages repo, then it sucks from upstream one
<f00b4r0>
yeah I get that now. Which suggests we can probably get rid of shipping target/**/base-files. I'll try to delete that and see what happens :)
<tmn505>
yes
<tmn505>
or mostly yes, I haven't tried that
<ynezz>
I did, it builds, didn't run test it
Kali_ has quit [Quit: leaving]
<f00b4r0>
meanwhile copying the contents to files/ does work. I'll use that as a lazy workaround :)
Tapper has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
guidosarducci has quit [Remote host closed the connection]
torv has quit [Read error: Connection reset by peer]
cmonroe_ has joined #openwrt-devel
guidosarducci has joined #openwrt-devel
torv has joined #openwrt-devel
Slimey has quit [Read error: Connection reset by peer]
Slimey has joined #openwrt-devel
arifre has joined #openwrt-devel
kopijahe has joined #openwrt-devel
kopijahe has quit [Read error: Connection reset by peer]
Borromini has joined #openwrt-devel
Misanthropos has joined #openwrt-devel
Borromini has quit [Ping timeout: 480 seconds]
Tapper has quit [Ping timeout: 480 seconds]
Misanthropos has quit [Ping timeout: 480 seconds]
shibboleth has joined #openwrt-devel
Tapper has joined #openwrt-devel
Misanthropos has joined #openwrt-devel
<mangix>
nasty. I bricked my desktop. No OpenWrt development for a while.
<minimal>
rsalvaterra: yeah Alpine recently found changing kernel builds to use mawk rather than Busybox awk made a major different for build time on x86_64
<mangix>
minimal: alpine still has various issues with building openwrt. biggest issue is perl.
<minimal>
mangix: wasn't clean enough, I meant Alpine building the *Alpine* kernel package
<minimal>
s/clean/clear/
<mangix>
got it.
<minimal>
I think gawk was a little faster than Busybox awk but mawk was faster still
<mangix>
Alright, looks like I have to disassemble my desktop and flash the BIOS with a serial programmer. This will be fun.
Borromini has quit [Quit: Lost terminal]
PaulFertser has quit [Ping timeout: 480 seconds]
srslypascal is now known as Guest2431
srslypascal has joined #openwrt-devel
GNUmoon has quit [Ping timeout: 480 seconds]
<aiyion_>
Grommish: I just received an EdgeRouter 10x and was looking into adding support for it; only to find you already did that :)
<Grommish>
aiyion_: Not quite, the LEDs dont' work, I can't test PoE, and Port 5 refuses to work, although 0-4 and 6-9 do
<Grommish>
aiyion_: I've not been able to touch it for a while because of the other stuff I've got going, but it'll at least boot
<Grommish>
aiyion_: The mixing of the MT7621 + RTL8367RB mixes DSA and swconfig, so I'm surprised it worked at all :D
f00b4r0 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
f00b4r0 has joined #openwrt-devel
Guest2431 has quit [Ping timeout: 480 seconds]
<aiyion_>
Any chance you'd introduce me to you findings next week? Like obviously I have only little experience with device support, but at at least the LTE-router and the omega worked out well :)
aiyion_ is now known as aiyion
<aiyion>
Maybe we're lucky and fix another issue :)
<aiyion>
for starters, do you have a public staging branch that contains your progress?
<Grommish>
aiyion: of course, hit me up whenever.. I still have the device, although it's running EdgeOS right now. I did all my testing via tftp initramfs loading to make it eaaier
<aiyion>
I have not even opened up this one, but after the weekend I'd like to get started reading up on the previous work. I thought the thread in the forum looked promising.
<Grommish>
I suspect it's a DTS issue still but...
<aiyion>
thanks
<aiyion>
Im far from suspecting anything yet ^^
<Grommish>
At the time, I had zero DTS exp and I'm not much further along these days, so anyone who can peep it is always welcome
<rsalvaterra>
mangix: Oh, dear…! What have you done? o_O
GNUmoon has joined #openwrt-devel
PaulFertser has joined #openwrt-devel
robimarko has quit [Quit: Leaving]
PaulFertser has quit [Ping timeout: 480 seconds]
<mangix>
rsalvaterra: turns out I just needed to reset it. I was hoping I could use an SPI programmer with an SOIC8 clip. Nope. Would need to desolder.
PaulFertser has joined #openwrt-devel
gladiac has quit [Quit: Ping timeout (120 seconds)]
gladiac has joined #openwrt-devel
GNUmoon has quit [Remote host closed the connection]
valku has quit [Remote host closed the connection]
GNUmoon has joined #openwrt-devel
valku has joined #openwrt-devel
Misanthropos has quit [Ping timeout: 480 seconds]
shibboleth has quit [Quit: shibboleth]
Tapper has quit [Ping timeout: 480 seconds]
Misanthropos has joined #openwrt-devel
Znullptr has quit [Quit: Connection closed for inactivity]