<[Pokey]>
Hey PaulFertser, quick update. It was going so well, but alas, the compile is failing and I do not know how to correctly interpret the results. Not sure I have even included enough details, but I have tried https://pastebin.com/JMRmc2PK
goliath has joined #openwrt-devel
<robimarko>
Ansuel and others, what should be done in regards with ath10k-ct?
<robimarko>
It seems not to be actively maintained or developed anymore
<robimarko>
I mean, its still stuck on 5.15 version
<rmilecki>
robimarko: if we don't get an update, i'm ok with dropping it
<rmilecki>
robimarko: we can't take all maintenance cost
<robimarko>
rmilecki: Hopefully he will continue/resume updating it
<robimarko>
Otherwise I agree its best to revert to in-tree ath10k
<robimarko>
rmilecki: While you are here, thoughs on disabling WEXT?
<rmilecki>
robimarko: does wl use it?
<rmilecki>
robimarko: i know we configure "wl" driver using "wl" tool
<rmilecki>
i'm not sure if WEXT is involved there at all?
<robimarko>
Question is whether wl even compiles
dangole has quit [Ping timeout: 480 seconds]
<robimarko>
Oh, seems like Ansuel fixed it couple of weeks ago
<robimarko>
And yeah, it depends on WEXT
<robimarko>
However, since its limited to bcm47xx and bcm63xx we can disable WEXT in generic configs
<rmilecki>
robimarko: a definitely good idea
<mrkiko>
off-topic question - In bash, I have variables named var1_name var2_name and so on; I need to iterate across all of them. How can I do something like $var$index_name to access the vars?
<mrkiko>
I guess I might be going after some eval trickery but...
<Ansuel>
on second tought branch pattern matching for main and openwrt- is problematic for everyone that wants to run test in his own fork
<Ansuel>
that effectively limit the CI run on any branch that doesn't match
<aparcar[m]>
looks good to me
<Ansuel>
a better solution is to just disable master... side effect user will have just another reason to migrate from master to main
<aparcar[m]>
I'll write a list of things that have to be done for that
<aparcar[m]>
we need a bit of grace period
<aparcar[m]>
blocktrron: speaking of that, could you tell your downstream project to switch to `main`?
<Ansuel>
btw a thing to do on packages repository... add a ci run that label each issue/pr with a tag related to the package
<Ansuel>
should be very easy to do
<aparcar[m]>
Ansuel: you want a label per package?
<Ansuel>
yes that is the idea
<Ansuel>
mainatiners would identify it quickly
Tapper has joined #openwrt-devel
<aparcar[m]>
ideally people start the title always with the package name
<Ansuel>
yep that is the pattern but i see tag more organized, or probably it's just me that see the packages pr section too ""empty""
<aparcar[m]>
issues can't really be tagged automatically. if it's based on the title someone could create arbitrary labels just by settings title to "my_fun_label: fake bug"
<aparcar[m]>
I wonder if codeowners is then a better solution
<blocktrron>
aparcar[m]: we can do in the future
<blocktrron>
But our tags reference branch x hash
<Ansuel>
my idea for pr was a mix of the action to check the changed files and regex to derive the changed packages... and then use api to assign the label
<blocktrron>
We do not want to get rid of master as a link, do we?
<tom->
for a short period on tuesday master was already missing on github
<blocktrron>
aparcar[m]: why do we have to migrate the branch name at all?
dgcampea has quit [Remote host closed the connection]
<blocktrron>
At the meeting it sounded like both can be used for cloning
dgcampea has joined #openwrt-devel
<aparcar[m]>
blocktrron: both can be used but people stated that they prefer only having one (main) and not an alias
<blocktrron>
aparcar[m]: where is this decision done
<blocktrron>
This is dumb. We break stuff from the past just because "We prefer havin only one"?
<blocktrron>
When the vote was casted, the premise was we do not break existing setups with this change.
<blocktrron>
And we should adhere to that.
<Ansuel>
the idea was to have a migration period (still to be decided the timeframe)
<Ansuel>
to permit downstream project to align to the change
<Ansuel>
and then do the switch
<blocktrron>
Ansuel: where can i read about that?
<Ansuel>
i remember it was written in the vote proposal
<ynezz>
"It's important to have a long transition period in order to avoid
<ynezz>
breaking downstream users' workflows."
<blocktrron>
in hindsight, making the switch like hostapd did it would have been the better way.
Lynx- has quit []
Lynx- has joined #openwrt-devel
<ynezz>
how they did it?
goliath has joined #openwrt-devel
<f00b4r0>
quick question for the wireless gurus: if I want to kick a client from an (open) SSID to force them to restart the whole association, is using WLAN_REASON_PREV_AUTH_NOT_VALID correct? My happy testing shows it works (whereas WLAN_REASON_UNSPECIFIED does not as the client immediately reconnects)
<robimarko>
Just a clean cut, cause this is just causing extra work
<f00b4r0>
aka "suck it up" ;P
<oliv3r[m]>
why would we want to go through the pain and switch from master to main? Other then to make some wokeness happy; it only causes pain, and doesn't yield anything?
<ynezz>
blocktrron: I think, that such proposal would be rejected by the majority
<f00b4r0>
oliv3r[m]: this has been going on ad nauseam. the wokeness won, time to FIDO ;P
<oliv3r[m]>
lol I know workness won; but it just creates so much busy-work; when there's already so much to do; and what will improve?
<f00b4r0>
apparently this was *important*
* f00b4r0
hides
<ynezz>
2 keystrokes saved for the start
<f00b4r0>
lol
* stintel
mumbles tab completion
<stintel>
now you have master and main and actually need 1 extra keystroke ;)
<f00b4r0>
meanwhile scratch what I said, above, WLAN_REASON_PREV_AUTH_NOT_VALID is no better than WLAN_REASON_UNSPECIFIED. I have a problem :P
<oliv3r[m]>
reading the vote-thread; and i'm sure this train can't be stopped anyway; hope the people who do the busy-work get payed handsomly :) btw, I doubt linus will migrate upstream master to anything else?
<ynezz>
stintel: my vim doesn't autocomplete branches in config files :P
<f00b4r0>
oliv3r[m]: I believe linus' position will be his usual "over my dead body" ;)
<Lynx->
so now openwrt will have to rework all the master/slave stuff, and more and more nonsense depending on the political climate
<oliv3r[m]>
as the wokeness continues, we'll soon have no words left to use; as someone will always be offended. I'm personally very offended by 'main'. There should be no main role, we are all equals :D
<stintel>
oliv3r[m]: and soon IQ tests will be forbidden because dumb people will be offended they have low scores
<stintel>
etc
* stintel
steps out
<Lynx->
as a small petty act of defiance, you can change all your github projects form main to master
<f00b4r0>
stintel: the MiniTru is working hard ;)
<f00b4r0>
s/stintel/oliv3r[m]/
<robimarko>
Its pointless to fight back now, it passed the vote, now we just live with the consequences
<Lynx->
with a divided and pissed development team
<f00b4r0>
part of which is now happier
* f00b4r0
ducks
<robimarko>
I am just waiting on all of the new devs that now feel inclusive and are not offended anymore
<f00b4r0>
lol
* ldir
quacks
<Lynx->
anyway, what fun now changing all the master/slave docs and references everywhere. Master/slave is a totally routine engineering term I had in my course.
<johnf>
ok, so I think I've figured out this strange problem I'm having
minimal has joined #openwrt-devel
<johnf>
I'm testing some minor changes to files baked into a firmware image
* f00b4r0
pulls more hair on captive portal (utter lack of) standardisation
<johnf>
and when I sysupgrade -n, I keep not getting a clean system
<Ansuel>
robimarko btw love the amount of comments you are getting with your simple series...
<johnf>
this time I've validated the inode numbers on the files in /etc, and they are changing
<robimarko>
f00b4r0: Huh, standardization, be lucky some of then even work
<Ansuel>
i'm impressed how for big change everyone is silent and for a god damn function that expose a hex id we have all sort of qcom guy LOL
<robimarko>
Ansuel: Well yeah, half of QCA and maintainers now jumped in
<f00b4r0>
robimarko: at least then you can shift the blame when something doesn't work and your implementation is standard-compliant ;)
<johnf>
so I think what's happening, is that a previous sysupgrade backup is being recovered when I boot the new firmware
<Ansuel>
time to spam CAN SOMEONE CHECK WHY REMOTEPROC IS BROKEN GUYS?
<johnf>
I think that it's stuck at the end of the filesystem or something, and it's not being erased because I'm not really writing anything to the filesystem
<johnf>
does anyone have any idea how I can address this problem?
<robimarko>
Ansuel: Well, Bjorn silently ignored my email
<johnf>
I guess I could fill the flash with 0s
<Lynx->
Ansuel you voted in this woke vote about changing master to main right? But then your avatar on the OpenWrt forum is mocking Jesus.
<Lynx->
wouldn't using a less offensive avatar be more inclusive?
<oliv3r[m]>
maybe we should fork openwrt to lede again :)
<Ansuel>
let not be too toxic about this guys... aside from downstream project... for a dev it's really a command to move from master to main...
<Ansuel>
codeaurora litterally changed the entire platform and nobody said a word...
<Lynx->
Well I wonder if you might consider changing your avatar, because it is genuinely offensive.
<robimarko>
Everything is offensive
<Ansuel>
we have tons of package broken that still point to the old link and are working only because we have it on mirrors...
<Ansuel>
for real guys lets not exagerate for a branch rename
<Lynx->
everything is offensive too, but it's really silly to change main to master and tolerate ridiculing Jesus...
<johnf>
there are more than enough real problems, while you may not agree with this decision it's made and done
<johnf>
not much left to deal but with the consequences
<johnf>
though whining on irc is always an option
<Lynx->
I think it's really a sign of our times that mocking Christianity is just fine, but pandering to dubious concepts like master/main.
<johnf>
s/left to/left to do/
<Ansuel>
this will be my last response. one may be a reference to slavery and something really bad... the other is fking put glasses on an image on some books... 2 very different topic... ALSO I EVEN VOTED NEUTRAL ON THAT... anyway back on actually working on what community deserve than wasting time on these futile discussion on a branch rename that were already discussed for month and a vote was already taken ;)
<johnf>
anyone got any ideas on my device that's resisting a version change?
<robimarko>
What is the device?
<oliv3r[m]>
johnf: ah, but it's easy to play the woke card, and say 'we must do this' and then let OTHERS do the majority of the work. "Dealing with the consequences" would mean, the yay-sayers do the work too? yeah, or just shove others work on their plate :) that's partially the crux
<Lynx->
It's not just a random picture. It's Jesus. And you care about it enough not to change it. I'll not say any more. Neutral says enough.
<Lynx->
Tolerance is fine, just so long as it's the right kind of tolerance right? Mocking Christianity is just fine.
<oliv3r[m]>
what was the outcome of the vote anyway? 51% in favor :p like brexit :D
<oliv3r[m]>
anyway, ansuel is right; best to continue with doing some work :)
<oliv3r[m]>
there's a good pun in there, but I'll refrain :)
<robimarko>
Lets just move on with life
<johnf>
robimarko: glinet mifi
<johnf>
though I've managed to get it to lose the configuration, so I'm happyish
<johnf>
I just don't understand what causes this behaviour and I've seen it before
<Slimey>
heh
<oliv3r[m]>
robimarko: have you run into the fact that dsa_bridge_join & co has change dinterfaces? any work as a result from that? (other then changing the API, I did that)
<johnf>
is there a message I'll see in logread that can tell me a sysupgrade configuration is being restored?
<robimarko>
oliv3r[mg: In 6.1 or ?
<oliv3r[m]>
6.1 indeed
<robimarko>
Well, last kernel I played with DSA around was 6.0
<robimarko>
And I dont remember making any changes
<PaulFertser>
[Pokey]: hi. I'm back to the keyboard now, hope to ping you later today, and if forget, feel free to remind me.
<Christophe[m]1>
Is anyone willing to give me quick advice how to migrate a redmi ax6000 to {stock,ubootmod}
<robimarko>
Ansuel: What exactly is broken?
<Ansuel>
mh?
<robimarko>
You mentioned packages being broken
<[Pokey]>
PaulFertser: Alrighty, thank you
<oliv3r[m]>
robimarko: ah ok, i kind of hoped your patchset for qca included port_bridge_join changes I could look at :)
<oliv3r[m]>
linux/master doesn't seem to show much, just the interface change, but no idea what the consequences of that change is
<PaulFertser>
[Pokey]: fwiw, the paste seems useless, as you're running not with -j1 you're getting normal outputs from some of the threads at the end, while the failing thread with its essential output is somewhere way up.
<[Pokey]>
I also tried immortalwrt's release via sysupgrade, package installed but the driver kept causing constant USB disconnects so that didn't work
<[Pokey]>
Tried it with Raspbian, seems the driver loaded and is working
<robimarko>
oliv3r[mg: Well, I can see they are just passing extack now to port_bridge_join
<robimarko>
And that is fully optional
<PaulFertser>
[Pokey]: I will try to spend some reasonable time later today to compile it on my own. But probably the result might not be what you like as realtek code is crap.
<[Pokey]>
PaulFertser: I appreciate that, thank you. I don't care too much so long as it works. I don't even need AP mode
<PaulFertser>
[Pokey]: if you're ready to test images from me. What tag exactly you want? Latest stable?
<tmn505>
johnf: I'm not sure if any message is printed, fstools are responsible for restoring backup from flash, so adding prints to it should do that.
<[Pokey]>
PaulFertser: Absolutely willing. Yes, just 22.03.5
<[Pokey]>
Can confirm I have network access using that driver under Raspbian
<PaulFertser>
[Pokey]: cool, I hope I'll be able to cook something for you before tomorrow begins but no promises, need some rest before.
<[Pokey]>
No rush! Thank you
<[Pokey]>
157MB/s, pretty stable too on Raspbian
<[Pokey]>
Mb/s*
<PaulFertser>
[Pokey]: are you building the tip of the branch for the repo we discussed? What is the kernel version?
<[Pokey]>
I am building the tagged v22.03.5
<PaulFertser>
[Pokey]: I mean for raspbian.
<[Pokey]>
As for the driver one, that pastebin error log is for a version in December (you can find the commit hash in the makefile from immortalwrt, and I tried latest but patches failed to apply not surprisingly)
<[Pokey]>
Uhh kernel on Raspbian... one sec
<[Pokey]>
6.1
<[Pokey]>
6.1.21+
<PaulFertser>
[Pokey]: and the driver version you're building for raspbian ?
<oliv3r[m]>
any idea why with 6.1 of_node_to_fwnode(np) returns me a null pointer? np is valid; i think in 5.15 this was 'just there'?
<robimarko>
Well, all that of_node_to_fwnode() does is return &np->fwnode if it exists
floof58 is now known as Guest1342
<robimarko>
Or if not returns NULL
floof58 has joined #openwrt-devel
Guest1342 has quit [Ping timeout: 480 seconds]
Borromini has joined #openwrt-devel
Borromini has quit [Quit: Lost terminal]
fakuivan has quit [Ping timeout: 480 seconds]
<oliv3r[m]>
Lol obviously ;) zo i should have asked that ;) howcome np->fwnode seems tot be null since 6.1
fakuivan has joined #openwrt-devel
<robimarko>
I guess this is somewhere in IRQ code?
<robimarko>
Cause, that node needs to come from somewhere
<oliv3r[m]>
Yeah, the header that defines it uses it a lot. Irq domains seem tot heavily telt on fwnodes. Zo i wonder now, what changed .. All i do, is use the DECLARE irq macro, zo i thought it would pass me a proper device_nodde ...
<robimarko>
Can you point me to the place that finds that node in the first place?
ptudor has joined #openwrt-devel
philipp64 has quit [synthon.oftc.net reflection.oftc.net]
dgcampea has quit [synthon.oftc.net reflection.oftc.net]
johnf has quit [synthon.oftc.net reflection.oftc.net]
Mangix has quit [synthon.oftc.net reflection.oftc.net]
Slimey has quit [synthon.oftc.net reflection.oftc.net]
jakllsch has quit [synthon.oftc.net reflection.oftc.net]
Snuupy has quit [synthon.oftc.net reflection.oftc.net]
yolo has quit [synthon.oftc.net reflection.oftc.net]
chder has quit [synthon.oftc.net reflection.oftc.net]
hurricos has quit [synthon.oftc.net reflection.oftc.net]
hgl has quit [synthon.oftc.net reflection.oftc.net]
johnf has joined #openwrt-devel
philipp64 has joined #openwrt-devel
yolo has joined #openwrt-devel
Slimey has joined #openwrt-devel
jakllsch has joined #openwrt-devel
Snuupy has joined #openwrt-devel
chder has joined #openwrt-devel
Tapper has quit [Quit: Tapper]
robimarko has quit [Quit: Leaving]
<[Pokey]>
PaulFertser: Apologies for not seeing sooner
<[Pokey]>
Just setting up to test
<[Pokey]>
PaulFertser: It appeared to be working, the driver loaded and I could scan for WiFi networks, however as soon as I added a Client (WDS) to the radio, it locked up. Rebooted, it boots, but networking appears to hang completely, and `ip a` yeilds an infinite pause waiting for a response, which cannot be CTRL + C'd