floof58 has quit [Remote host closed the connection]
Tapper has joined #openwrt-devel
Mangix has joined #openwrt-devel
floof58 has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
gladiac has quit [Quit: k thx bye]
floof58 has joined #openwrt-devel
floof58 is now known as Guest1560
floof58 has joined #openwrt-devel
Guest1560 has quit [Read error: Connection reset by peer]
<Christophe[m]1>
Can someone tell me what "WARNING: Applying padding in xxx/bin/packages/aarch64_cortex-a53/base/Packages to workaround usign SHA-512 bug" is about?
<ynezz>
Christophe[m]1: git log e1f588e446
floof58 has quit [Remote host closed the connection]
floof58 has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
floof58 has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
floof58 has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
floof58 has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
floof58 has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
floof58 has joined #openwrt-devel
floof58 is now known as Guest1566
floof58 has joined #openwrt-devel
Guest1566 has quit [Read error: No route to host]
floof58 has quit [Remote host closed the connection]
<aparcar[m]>
Ansuel: ynezz shouldn't we add some shellcheck for our internal scripts? I keep seeing people doing typos and things slip through
sorinello has quit [Ping timeout: 480 seconds]
floof58 has quit [Remote host closed the connection]
floof58 has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
floof58 has joined #openwrt-devel
danitool has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
<ynezz>
aparcar[m]: the more QA, the better
floof58 has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
floof58 has joined #openwrt-devel
floof58 has quit [Remote host closed the connection]
<aparcar[m]>
can I somehow tell within a Package Makefile the process to execute in a subfolder of the source? The sources contains multiple files, I only want to enter the ./rust folder
floof58 has joined #openwrt-devel
<Tapper>
floof58 Hi can you try and not keep leveing and coming back in to the room pleas? It's a bit spammy. Thanks.
<Ansuel>
it's ok... if you check how cache work, we would have a download at the start and an upload at the end (that should overwrite the directory, or just delete first and download)
<Ansuel>
to prevent any kind of bloat we can just set a single directory with something like ccache-kernel-ipq806x-generic
<ynezz>
it has mirror/rsync feature
<ynezz>
it = minio
<ynezz>
so should probably help to save some time
<ynezz>
in similar builds, thats it
<Ansuel>
the thing is that we do tons of test on kernel but very little time we do kernel bumps and with a 200mb of kernel ccache, the time decreated from 30 to 1 min
<ynezz>
I would probably consider adding branch name into that cache key, it might increase cache hits
<ynezz>
but as well might increase storage size
<Ansuel>
i have no idea about packages workflow (where we build everything) but for kernel it's 200mb for target
<ynezz>
just not sure if we can trust ccache that much, if the result is going to be reproducible on buildbots without ccache
<ynezz>
and I've been bitten few times in the past by some compilation failure which turned out to be caused by ccache
<ynezz>
so stopped using it, but it was many years ago, so maybe the situation is better now
<ynezz>
the time saving is very nice, so really worth trying it out
<Ansuel>
we had a compilation failure recently but it was us changing host tools config and ccache not detecting that... a wipe of the ccache dir fixed the problem so in case we have that kind of problem we can just wipe everything... they are scratch file after all
<Ansuel>
anyway i checked price for cloud storage it would cost 0,50€ of credit for 20gb
<ynezz>
yes, storage is usually peanuts, in/eg-ress transfers is what bothers me
<Ansuel>
0,02/GB mh
<ynezz>
but if we can operate mostly in mirror/rsync mode then it should be fine - here I'm unsure about number API calls, they're billed as well
gladiac has quit [Quit: k thx bye]
<Ansuel>
if we want to reduce call we can consider making a tar of everything and sync that
<Ansuel>
but yes looking at the traffic that might be a problem... if we want to implement that we should check the usage in a window of a week to see if the thing doesn't explode with cost
<ynezz>
honestly I don't know the internal ccache structure, so unsure how well it could be mirrored/rsynced either, but I would assume, that tarball would kill all of that as it would always change
<ynezz>
IMO GCS is good for the start, you can gather some data and we can adjust later, we wont run out of $10k in a week hopefully :P
<ynezz>
IIRC then aparcar[m] has access to the GCP and can probably provide you with access
<Ansuel>
nice!
<Ansuel>
plan is to add it on top of current system so if we have problems cache support is still there
<f00b4r0>
Ansuel: fwiw on phase1 we only ccache host tools; for the reasons highlighted by ynezz and you, and despite some local naysayers claiming ccache is rock solid ;)
<robimarko>
\x: Thermal management is a must as well
<robimarko>
CPR is missing as well, there is work to do
gladiac has joined #openwrt-devel
<\x>
what is cpr
<robimarko>
Core Power Reduction
<robimarko>
Its a HW peripheral that tweaks the OPP point voltage dynamically based on the peripherals being used and load factor
rmilecki has quit [Quit: Konversation terminated!]
rmilecki has joined #openwrt-devel
rmilecki has quit []
rmilecki has joined #openwrt-devel
rmilecki has quit []
rmilecki has joined #openwrt-devel
rmilecki has quit []
rmilecki has joined #openwrt-devel
<\x>
oh so something like a vrm controller
patrykk has quit [Read error: No route to host]
<Znevna>
can it make it HCF? (halt and catch fire) you know, like the recent AMD/Asus combo? :P
<robimarko>
Nope
<robimarko>
As there are min/max limits for each OPP
<robimarko>
It just provides the optimal voltage to OPP-s and then regulator gets adjusted
<Znevna>
hmm
<Znevna>
did something change in github? only devs with access can approve changes?:P
<robimarko>
Yes, it was intentional change from couple of days ago
<Znevna>
ok
_mfjlo54ys has joined #openwrt-devel
<_mfjlo54ys>
hey, i'm trying to compile but getting the error:openwrt/include/toplevel.mk:231: world] Error 2 , any ideas on what to do? Am i missing spe package i can install with som quick command?
<_mfjlo54ys>
missing some package?
<_mfjlo54ys>
i'm new to this, but when compiling i should compilr from root # and not $ right?
<_mfjlo54ys>
ok i missed that one! gonna do it over, you know about error 2 and some command?
<_mfjlo54ys>
i did get that earlier, the someone here wrote som clever command that installed packages i believe and it went away...
<_mfjlo54ys>
Gonna try without sudo now again though
<_mfjlo54ys>
thanks
<_mfjlo54ys>
i often just skip through text by the way so i missed that
<_mfjlo54ys>
obviously, but still. The world eeeor what that about?
<_mfjlo54ys>
world build error
<Znevna>
everything required for a build from source should be on that page
<Znevna>
assuming you don't skip too much
<Znevna>
why do you wanna build from source?
goliath has joined #openwrt-devel
rua has joined #openwrt-devel
Borromini has joined #openwrt-devel
<_mfjlo54ys>
ok, because i want to learn, and make it custom to my needs, plus theres less potential viruses or whatever in it that i don't trust. that's why
<_mfjlo54ys>
I would build my entire os if i could. i'm just for open source all the way
<Znevna>
ok
<_mfjlo54ys>
I'm no coder, i just trying to learn to compile and it's awesome how flexible openwrt seems, but libreboot seems more libre because they use a libre kernel
<_mfjlo54ys>
yeah
<_mfjlo54ys>
librecmc i meant
<_mfjlo54ys>
how do i keep the size around 7 mb? I can't compile many packages with the M option right? I can't check the size of it before compiling right..
<_mfjlo54ys>
Can one shrink an already built file easily?
<Znevna>
no
<Znevna>
you can practice with firmware selector, adding whatever packages you want, see what you end up with
<Znevna>
then build locally
<Znevna>
but it's the same thing in the end
<Znevna>
so wasted time
<_mfjlo54ys>
ok thanks. It's not wasted time if i'm learning howto compile
<_mfjlo54ys>
locally
<_mfjlo54ys>
It's good knowing a little bit about
<_mfjlo54ys>
plus, it's interestin to see how flexible openwrt seems to be!
<_mfjlo54ys>
interesting
<_mfjlo54ys>
I even learnt you can use / and search in menuconfig
<_mfjlo54ys>
:)
<_mfjlo54ys>
just now
<_mfjlo54ys>
but there are some commandthat checks if everything isset up correct right? Like the error 2 last time, the world issue builderror
<_mfjlo54ys>
but there are some command that checks if everything is set up correct right?
<_mfjlo54ys>
like downloading missing libraries maybe, or building tools, etc
<_mfjlo54ys>
it did work to build last time.. i just deleted the old openwrt, and the new one gave me that issue
<_mfjlo54ys>
forgot what i did do right lsst time i fixed it
<_mfjlo54ys>
last time
_mfjlo54ys has quit [Remote host closed the connection]
<Christophe[m]1>
Thanks wigyori for the risc-v sifive unleashed support. <3
Borromini has quit [Quit: Lost terminal]
SlimeyX has quit [Remote host closed the connection]
rmilecki has quit [Quit: Konversation terminated!]
rmilecki has joined #openwrt-devel
rmilecki has joined #openwrt-devel
minimal has joined #openwrt-devel
djfe is now known as Guest1591
djfe has joined #openwrt-devel
Guest1591 has quit [Ping timeout: 480 seconds]
rua has quit [Quit: Leaving.]
SlimeyX has joined #openwrt-devel
rua has joined #openwrt-devel
dangole has quit [Remote host closed the connection]
rmb has quit [Ping timeout: 480 seconds]
pkgadd_ has quit [Ping timeout: 480 seconds]
<mrkiko>
wigyori: thanks from my side as well for all your help, and for this port as well. Don't have the hw, but it's very nice to see riscv starting to get more popular as well
<robimarko>
Honestly, its only decent HW missing that is slowing RISCV down
<[Pokey]>
General question: If I go ahead and add support for an 8/64 device, well, is that even possible with 22 and upcoming 23, and would it get rejected? I have read and accept the warning of limitations
<[Pokey]>
I would just wanna replace the dodgy Chinese management interface with a decent and maintained one for an AP
<slh>
[Pokey]: https://downloads.openwrt.org/releases/23.05-SNAPSHOT/targets/ath79/generic/ suggests around 6 MB, so possible and presumably even for 24.xy. extrapolating from history with 4/32 and 8/32, device support is likely to be accepted, as long as you don't really need complex new (out-of-tree) drivers or a full new (sub-)target
<[Pokey]>
I might win a pair in eBay and their management interface fills me with dread and sadness
<[Pokey]>
Apparently they have native telnet and serial console with uboot accessible according to a friendly Amazon reviewer
<[Pokey]>
I really hope the bootloader will be easy to handle, and I might be able to add support
<slh>
well, I certainly wouldn't buy that stuff today - but if it is a mostly bog standard mt76x8 device, there wouldn't be any reason to reject device support patches
<[Pokey]>
It's cheap and does what I need
<[Pokey]>
I just wanted to neaten it up a little with a proper management interface, that's all
<[Pokey]>
These are 1KM directional PTP bridge APs
<[Pokey]>
WiFi in a marketing trenchcoat
<slh>
ah, o.k., didn't look at that aspect
<[Pokey]>
I neat to beam network from my house to a garage block across the street
<[Pokey]>
s/neat/need
<[Pokey]>
Nice cheap device you pointed out though, and I'm a brit for future reference :P
<slh>
makes a nice wired AP
<slh>
750-800 MBit/s
<[Pokey]>
Thank you for your advice, I will hopefully win the auction tomorrow, and I will do my best based on \x's previous assistance and your advice today. I might have to pop in and ask about bootloadery stuff though
<[Pokey]>
Thank you
<[Pokey]>
800mbps is very nice at that price
<[Pokey]>
They look like they would be good for just plopping around the house and pointing them at each others BSSIDs to build a mesh
<slh>
also possible, yes
<slh>
a good friend uses two of those dap-x1860 as WDS/4addr extenders
<[Pokey]>
inb4 PaulFertser notices WDS
<[Pokey]>
slh: I think that's basically how I'll set them up when/if I get OpenWRT on them
<[Pokey]>
AP (WDS) on one side, Client (WDS) and standard convenience AP on the other
ptudor has quit [Ping timeout: 480 seconds]
<djfe>
mt7628 is not the best choice wifi wise, but it will run and have enough flash space (probably) for a while
<djfe>
omg.de advertises these as "up to 8km" range
<slh>
very personally, I wouldn't really consider the old RaLink- or Mediatek <mt7615 wifi myself, mt7615 should be fine - and mt7915+ certainly is (very-) fine.
<\x>
if it was me itll be like cpe210/cpe710 hehe
<\x>
for beaming wifi, theyre cheap
<\x>
then you just do vxlan on both ends to pass your vlans across a wds or an 80211s link
danitool has quit [Remote host closed the connection]
<djfe>
oh wait actually he doesn't have the mentioned device, but the ldf, sorry
<\x>
no idea though if 11s issues are fixed
<slh>
a few hours ago, yes
<\x>
I do wonder if we can now have frame mode 2 on ath10k ct
<djfe>
what is that?
<djfe>
what 11s issues? a-msdu frames?
<[Pokey]>
djfe: well that's depressing
<[Pokey]>
I'm glad you pointed it out though, thank you
<[Pokey]>
I already have a bid placed on these units and mikrotik gear is way out of my price range
<[Pokey]>
I might end up retaining stock fw if the WiFi really is an issue unless it can get resolved
<djfe>
it's not like mt7628 is unusable but if you have to restart the device at some point to reestablish the connection.. (the original os probably won't have the issue, but since it has existed for years, it's unclear if this will ever be fixed or when)
<\x>
tbh as long as you can establish a link itll be easy