marcan changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | "Does XXX work yet?": https://alx.sh/fs | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-alt #asahi-stream #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
moby1 has quit [Ping timeout: 480 seconds]
moby has quit [Ping timeout: 480 seconds]
moby has joined #asahi
moby1 has joined #asahi
<sven>
if you can run things as root you can also just load a kernel module that manually goes ahead and does evil things
<sven>
and if you protect against that with some custom config you can also just protect the speaker daemon the same way *shrug*
moby1 has quit [Ping timeout: 480 seconds]
moby has quit [Ping timeout: 480 seconds]
Zopolis4 has joined #asahi
moby has joined #asahi
moby1 has joined #asahi
<psykose>
why did speakersafetyd not stop me from blasting my speakers with a rifle?
stipa has quit [Read error: Connection reset by peer]
moby1 has quit [Ping timeout: 480 seconds]
moby has quit [Ping timeout: 480 seconds]
stipa has joined #asahi
moby has joined #asahi
moby1 has joined #asahi
<ayke_>
chadmed: understood, I'm sure you have a much better idea of what is going to be acceptable upstream
moby has quit [Ping timeout: 480 seconds]
moby1 has quit [Ping timeout: 480 seconds]
moby has joined #asahi
moby1 has joined #asahi
blazra has quit [Ping timeout: 480 seconds]
blazra has joined #asahi
moby has quit [Ping timeout: 480 seconds]
moby1 has quit [Ping timeout: 480 seconds]
<bluetail>
I honestly wouldn't have thought rsync would put such a load on my m1 mac mini, copying a dozen of files over: https://ibb.co/HnDk5Yw
<bluetail>
( 56 % )
moby has joined #asahi
moby1 has joined #asahi
<chadmed>
the kernel reports load utilisation as a function of the core's capacity, frequency and the task's consumption of time so that number depends on the pstate and whether its running on an icestorm or firestorm core
<bluetail>
ic
<chadmed>
e.g. with EAS enabled pipewire's resampling "uses" 70% of an icestorm core at the lowest pstate, but if i peg it to a pcore at 3ghz its like 1% utilisation
<bluetail>
recently switched to kde plasma because I seen myself mostly fiddling with the config all day... like no end in sight. What do you guys use? marcan and lina use kde-plasma too it seems... I mean, alone adding a new monitor config takes time in sway...
<chadmed>
i use a lightly customised plasma
<bluetail>
what annoys me in plasma is I haven't found a toggle for maximize/restore yet!
<bluetail>
in sway I just used fullscreen to meta+f and press it twice restored previous state
<chadmed>
super+pgup
<bluetail>
TY!!!
<chadmed>
(pgup on the macbooks being fn+up arrow key)
<chadmed>
so super+fn+up arrow
<bluetail>
I'm searching right know to what its mapped to so that I can change it
<chadmed>
Maximise Window
<chadmed>
system settings -> shortcuts -> kwin
<bluetail>
do you use bismuth? focus change isnt possible it seems... so I use alt+tab
<bluetail>
just a small change
<chadmed>
never even heard of bismuth
<bluetail>
enables tiling in kde plasma
<bluetail>
and its kinda broken
<chadmed>
kwin has a tiling mode built in now
<bluetail>
often when I launch terminal its floating instead
moby has quit [Ping timeout: 480 seconds]
<bluetail>
ooooh
moby1 has quit [Ping timeout: 480 seconds]
<chadmed>
not sure how exactly to enable it because i stopped caring about twms a long time ago
<chadmed>
but its there
<bluetail>
I just want to use the mouse less often
<bluetail>
is that possible without tiling?
<chadmed>
yeah have a rifle through the kwin key binds in the system settings
<chadmed>
there are plenty that are unbound and you can set up how you like
<bluetail>
a ... rifle?
<chadmed>
oh um
<chadmed>
rifling through something means like
<chadmed>
going through it quickly
<chadmed>
old language from when youd use your index and middle finger to flick through manila folders in a cabinet
<bluetail>
I see. Thank you. That was so helpful. Love your support!
<Retr0id>
I see "SPI NOR" is listed as a supported feature, on the feature support wiki page. Purely out of curiosity, what is this used for under linux, and is there a way for me to dump the NOR contents?
zef has quit [Ping timeout: 480 seconds]
moby1 has quit [Ping timeout: 480 seconds]
moby has quit [Ping timeout: 480 seconds]
<Retr0id>
seems like it's not in the asahi device tree, which answers my first question ("it isn't"), and perhaps also answers my second question ("add it!")
bluetail9 has joined #asahi
moby has joined #asahi
moby1 has joined #asahi
<chadmed>
someone is working on it since it's going to be needed for linux-macos bluetooth continuity and such
<chadmed>
but its another safety thing you can easily clobber your machine and require a trip to apple or an afternoon with apple configurator if you screw with it
bluetail has quit [Ping timeout: 480 seconds]
moby1 has quit [Ping timeout: 480 seconds]
moby has quit [Ping timeout: 480 seconds]
<Retr0id>
I can imagine
pbsds3 has quit []
moby has joined #asahi
pbsds has joined #asahi
moby1 has joined #asahi
D-Spirits has quit [Ping timeout: 480 seconds]
myon98 has joined #asahi
chadmed_ has joined #asahi
moby1 has quit [Ping timeout: 480 seconds]
moby has quit [Ping timeout: 480 seconds]
chadmed has quit [Ping timeout: 480 seconds]
moby has joined #asahi
moby1 has joined #asahi
chadmed_ has quit [Read error: No route to host]
chadmed_ has joined #asahi
avocicltb^ has quit [Remote host closed the connection]
moby1 has quit [Ping timeout: 480 seconds]
moby has quit [Ping timeout: 480 seconds]
Mary6 has quit []
Mary6 has joined #asahi
moby has joined #asahi
moby1 has joined #asahi
moby has quit [Ping timeout: 480 seconds]
moby1 has quit [Ping timeout: 480 seconds]
<ChaosPrincess>
Retr0id: spi nor can be used to read/write the nvram. If you want to dump the entire nor, just delete the partition map from device tree, and it will give you access to the whole chip
<Retr0id>
which part of the device tree is that in?
chadmed_ has quit [Remote host closed the connection]
moby has joined #asahi
moby1 has joined #asahi
moby has quit [Ping timeout: 480 seconds]
moby1 has quit [Ping timeout: 480 seconds]
Zopolis4 has quit []
moby has joined #asahi
moby1 has joined #asahi
moby has quit [Ping timeout: 480 seconds]
moby1 has quit [Ping timeout: 480 seconds]
moby has joined #asahi
moby1 has joined #asahi
<Retr0id>
ah ty
moby has quit [Ping timeout: 480 seconds]
julio7359 has quit [Ping timeout: 480 seconds]
moby1 has quit [Ping timeout: 480 seconds]
myon98 has joined #asahi
robinp_ has joined #asahi
chadmed has joined #asahi
robinp has quit [Ping timeout: 480 seconds]
julio7359 has joined #asahi
ptudor has quit [Ping timeout: 480 seconds]
MichaelLong has quit [Remote host closed the connection]
MichaelLong has joined #asahi
axt has joined #asahi
chadmed_ has joined #asahi
robinp_ has quit [Remote host closed the connection]
robinp has joined #asahi
chadmed has quit [Ping timeout: 480 seconds]
ptudor has joined #asahi
K900_ has joined #asahi
<K900_>
Hey guys, not quite Asahi-related question, but all the relevant changes I could find were Asahi-related
<K900_>
Kernel 6.2.3 broke wireless for me on a Raspberry Pi 4
<K900_>
(bcm4345)
<K900_>
Any ideas before I try to bisect it?
<marcan>
K900_: I see mostly non-asahi changes from 6.1 to 6.2, or are you saying it broke in stable 6.2.3?
<K900_>
Stable 6.2.3, yes
<K900_>
I'm bisecting now but it's not exactly fast hardware
<K900_>
6.2.2 works
<K900_>
The really weird thing is that there's just nothing in the logs
<K900_>
iwd just says "operation failed" (extremely helpful)
<K900_>
And nothing in dmesg or iwd's own log
<marcan>
does the device probe?
<K900_>
Yes
<K900_>
Actually let me see if I can reproduce it on a cross-built kernel
<marcan>
it kind of sounds like you're looking in the wrong place, I don't think any of the brcmfmac changes in 6.2.3 make any sense as breaking anything
<marcan>
maybe something else broke it? sdio controller, something in the 80211 core...
<K900_>
That's what I'm feeling too
<K900_>
But I guess there's only one way to find out now
hightower2 has quit [Remote host closed the connection]
<K900_>
OK at least it reproduces on cross
<jannau>
K900_: what broke exactly? authentication? can you still join open networks? a rebase onto 6.3-rc1 broke authentication
<K900_>
Uhh, I haven't actually tried
<marcan>
I went through all the commits in 6.2.3 and didn't see anything notable though
<K900_>
Let me see
<marcan>
well, all the driver commits
<K900_>
Yes
<K900_>
Can join open networks
<K900_>
Just tested with my phone
<marcan>
ah wait, I do see cfg80211
<marcan>
changes
<marcan>
outside any drivers
<marcan>
K900_: yeah I bet on a cfg80211 change doing it
<K900_>
And now to figure out which one
<K900_>
Fun
chadmed_ has quit [Remote host closed the connection]
chadmed has joined #asahi
<marcan>
I bet on f4b6a138efb8 / 015b8cc5e7c
<K900_>
OK, more fun
<K900_>
I dug up my laptop to see if it's even brcmfmac specific
<K900_>
And mt7921 now just locks up
<K900_>
Is it the same regression? Is it a different one? Fuck if I know
<K900_>
OK scratch that that was just a one-off glitch
<K900_>
Rebooting fixed it
<K900_>
So at least it's not consistently broken on 6.2.3
<K900_>
Let's see if reverting f4b6a138efb8 helps
<jannau>
marcan: you won. reverting 015b8cc5e7c in 6.3-rc1 fixes authentication
<jannau>
maz: ^^^
<marcan>
I'm not sure if the patch is broken or the fix is exposing another bug, but there you go
<K900_>
Welp
<K900_>
So where does one find cfg80211 people these days
<marcan>
linux-wireless mailing list I guess
<K900_>
jannau: just to confirm, you're also on brcmfmac, right?
<marcan>
yes
Bey0ndB1nary has joined #asahi
<jannau>
yes, all apple silicon devices use brcmfmac
<K900_>
So I guess we know brcmfmac breaks but mt76 doesn't
<marcan>
pretty sure that patch is broken
<marcan>
hm no, wait
<K900_>
I'm still rebuilding
<jannau>
clearing connect->* in cfg80211_connect breaks it here, authentication is still fixed if I revert only the last hunk of 015b8cc5e7c
chadmed has quit [Remote host closed the connection]
<marcan>
ok so that patch is replacing a schedule that calls cfg80211_conn_work() with only a direct call to cfg80211_conn_do_work()
<marcan>
jannau: oh so it's that one
<marcan>
I wonder if that's a userspace usage problem then
<K900_>
Are you using iwd or wpa_supplicant jannau?
<K900_>
iwd here
<K900_>
But I can try wpa_supplicant
<K900_>
(but also, iwd works fine with mt7921e so)
<jannau>
iwd
chadmed has joined #asahi
<marcan>
note that iwd uses a different codepath on brcmfmac and we *specifically* chose it because it works with wpa3 there
<marcan>
(wpa_supplicant doesn't or didn't when I first tested)
<marcan>
has to do with offload etc, might have been SME
<K900_>
Doesn't work with WPA3 for me on RPi4 but I believe that's a firmware issue
<K900_>
(and definitely not related to this directly)
<marcan>
we have a patch for WPA3 support
<marcan>
like another one
<marcan>
(which is not upstream yet)
<K900_>
Oh, OK
<marcan>
brcmfmac is hilariously undermaintained, it's missing some major features
<K900_>
I am still stuck with a bunch of devices that won't connect to WPA3 anyway
<K900_>
OK, rebooting with 015b8cc5e7c reverted now
<marcan>
that cached_keys stuff looks for WEP only, but then when it's not set they go and wipe all the crypto info?
<marcan>
yeah that's broken
<marcan>
completely broken
<marcan>
(unless you use WEP)
<K900_>
Yep, the revert worked
<K900_>
Should I test wpa_supplicant?
<marcan>
try removing just the last two lines
<marcan>
the ones that screw around with connect->crypto
<marcan>
I'm pretty sure that just outright breaks crypto offload for everything
<K900_>
Trying...
<jannau>
just removing those two lines fixes authentication as well
<marcan>
I'll send an angry email
<jannau>
thanks
<K900_>
So those two lines just aren't correct at all?
<K900_>
I guess let me rephrase
<K900_>
Can there be a better fix than just removing those?
<K900_>
I kinda want to get the fix into NixOS
<marcan>
K900_: how should I credit your report?
<K900_>
What do you need, an email?
<K900_>
me@0upti.me
<marcan>
K900_: there may be a better fix but this sounds like a wext problem where the fix is being shoehorned into the core wrongly
<marcan>
and a name (any name)
<K900_>
K900 or Ilya or w/e
<K900_>
I guess I have kernel commits signed with that email so anyone that wants to dox me can already
<K900_>
lol
<marcan>
:p
<marcan>
at least they got rid of the silly real name policy now
<K900_>
Also FWIW I don't really care about getting credited
<K900_>
But your call
<K900_>
I was going to ask you to CC me on the email but then I realized I probably don't want half of linux-wireless in my inbox ^^"
<marcan>
I mean I was going to CC you by default :p
<marcan>
if you don't want it I'll drop you
<marcan>
I doubt you'll get more than a few emails out of this though
<K900_>
Actually, yeah, do it
<K900_>
If someone finds a good fix I can get into NixOS quickly that will be worth it
<marcan>
sent
donk has joined #asahi
<K900_>
Thanks
<j`ey_>
marcan: that was hardly angry!
<mps>
idk if it is related but mrvl chip works with 6.2.3, yesterday upgraded kernel and didn't noticed any problem with wifi on it
<marcan>
my standards for angry are very different from the kernel average :p
<marcan>
mps: this breaks offload, likely your chip is too dumb for offload
<K900_>
I can test iwlwifi too
<K900_>
Just to get more data
<marcan>
I'm pretty sure software WPA still works or whatever
<mps>
marcan: that explains why I don't have problem
<marcan>
the broadcom guy still didn't reply to me about whether it works on broadcom firmware (sigh)
<K900_>
Uhh
<K900_>
I can do that
<K900_>
But I'm going to need a few hours probably
<marcan>
if you can test that it works I'll use that data point when I send a v2 (I also didn't hear from the cypress people about that so now I need to add a conditional under the hypothesis that they diverged, sigh)
<marcan>
why are broadcom/cypress/infineon people so unresponsive :(
<K900_>
Actually
<K900_>
Maybe not
<marcan>
make sure you use iwd (that is known to work)
<K900_>
I was thinking I'll have to reconfigure the AP
<K900_>
But maybe I can set up a hotspot on my phone
<K900_>
Let me see if that does WPA3
<K900_>
It does
<K900_>
Thank you daddy google
<K900_>
Do I still need [PATCH 06/10] brcmfmac: cfg80211: Pass the PMK in binary instead of hex?
<marcan>
just to make the patch apply, yeah
<marcan>
(and also maybe outright)
<marcan>
that one is already merged but for 6.4 I think
<K900_>
OK building now
<K900_>
Also rebuilt with just the last two lines removed
<K900_>
And that worked
<K900_>
Might be worth mentioning on the list too
<marcan>
I did mention that the last 2 lines are the problem
<K900_>
Oh nvm I can't scroll
roxfan2 has joined #asahi
tertu has quit [Ping timeout: 480 seconds]
roxfan has quit [Ping timeout: 480 seconds]
<K900_>
marcan: does not work
<marcan>
ok, could be a firmware issue then, or we need yet another mechanism for non-apple chips
<marcan>
thanks for testing
<K900_>
Nothing interesting in dmesg or iwd logs
<K900_>
iwd just says "operation failed"
<K900_>
Anywhere else I should look?
Zopolis4 has joined #asahi
<K900_>
Anyway I think I'll just go back to 6.1 for now instead of trying to run weird cross-compiled kernels
<K900_>
Thanks for helping debug this folks :)
<K900_>
I'm not on IRC much but feel free to ping me on email/matrix/fedi (@k900:toot.0upti.me) if you need more RPi testing, I have a spare one and a decent amount of interest in getting the wireless to be less terrible
<K900_>
...oh fuck.
<K900_>
That commit got backported to 6.1
<psykose>
:D
<marcan>
it probably got backported to everything, because stable
<K900_>
Yep
<mps>
K900_: you could ask on #iwd OFTC channel, maybe there someone could help
<K900_>
With what/
<K900_>
*?
<mps>
ask if iwd people could help to fix issue
<K900_>
The auth issue or the WPA3 issue?
<mps>
both
<K900_>
The auth issue marcan has tracked down and posted on linux-wireless
<K900_>
So probably up to them now
<K900_>
The WPA3 issue is that the firmware on the RPi is from like 2020 so it probably doesn't know what WPA3 even is
<mps>
ah, ok
K900_ has quit [Quit: Page closed]
chadmed has quit [Remote host closed the connection]
Mary6 has quit []
Mary6 has joined #asahi
chadmed has joined #asahi
joshtaylor has quit [Quit: Connection closed for inactivity]
chadmed_ has quit [Remote host closed the connection]
chadmed has quit [Read error: No route to host]
chadmed has joined #asahi
roxfan has joined #asahi
robotrip has joined #asahi
roxfan2 has quit [Ping timeout: 480 seconds]
Hibyehello_ has joined #asahi
roxfan2 has joined #asahi
c10l has quit [Quit: Bye o/]
roxfan has quit [Ping timeout: 480 seconds]
Hibyehello has quit [Ping timeout: 480 seconds]
<robotrip>
Hello all. Apologies for being a noob but my friend recently bought a mac mini off ebay for school and its account locked. Is it possible to boot asahi on an account locked mac?
c10l has joined #asahi
veeyee has quit [Quit: Leaving.]
veeyee has joined #asahi
<marcan>
robotrip: no
<marcan>
iCloud locked macs are as good as a brick
<marcan>
this is by design
<marcan>
there is no bypass
<robotrip>
ahh ok. i figured as much based on what i've read. thanks for confirming.
c10l5 has joined #asahi
c10l has quit [Read error: Connection reset by peer]
robotrip has quit [Quit: Page closed]
<psykose>
scammed :-)
veeyee has quit [Ping timeout: 480 seconds]
Bey0ndB1nary has joined #asahi
Bey0ndB1nary has quit [Max SendQ exceeded]
jamespmorgan has joined #asahi
Bey0ndB1nary has joined #asahi
possiblemeatball has joined #asahi
<mort_>
for some reason, bluetooth works in kde but not using bluetoothctl in sway: https://p.mort.coffee/pWO
jamespmo_ has quit [Ping timeout: 480 seconds]
Raito_Bezarius has quit [Remote host closed the connection]
Raito_Bezarius has joined #asahi
Raito_Bezarius has quit [Read error: Connection reset by peer]
Raito_Bezarius has joined #asahi
Raito_Bezarius has quit [Remote host closed the connection]
Raito_Bezarius has joined #asahi
Hibyehello_ has quit [Remote host closed the connection]
<blazra>
I just got a freeze of the system while running asahi-dev. I had to restart by holding the power button. Kernel log says: "Unable to handle kernel paging request...". Kernel log and asahi-diagnose (freeze happened at 22:48): https://gist.github.com/blazra/d6e12a942b1a6ef90c28616c75831c9c
<blazra>
marcan: ^
rayanamukami3 has joined #asahi
jordy_ has quit [Remote host closed the connection]
chadmed has quit [Remote host closed the connection]