Brainium has quit [Remote host closed the connection]
Brainium has joined #asahi
possiblemeatball has joined #asahi
amarioguy has quit [Remote host closed the connection]
jamespmorgan has joined #asahi
jamespmorgan has quit [Ping timeout: 480 seconds]
Brainium has quit [Quit: Konversation terminated!]
SmileyKeith has quit [Read error: Connection reset by peer]
SmileyKeith has joined #asahi
SmileyKeith has quit [Ping timeout: 480 seconds]
possiblemeatball has quit [Quit: Leaving]
nicolas17 has quit [Quit: Konversation terminated!]
dmmcf has joined #asahi
dmmcf has quit []
jluthra has quit [Remote host closed the connection]
jluthra has joined #asahi
qyousef has quit [Read error: Connection reset by peer]
qyousef has joined #asahi
rvalue- has quit [Remote host closed the connection]
rvalue has joined #asahi
Zopolis4 has quit [Quit: Connection closed for inactivity]
SSJ_GZ has joined #asahi
Treibholz has joined #asahi
thevar1able has quit [Remote host closed the connection]
thevar1able has joined #asahi
thevar1able has quit [Remote host closed the connection]
thevar1able has joined #asahi
mxz has quit [Quit: cya]
mxz has joined #asahi
yuyichao_ has joined #asahi
iaguis has joined #asahi
yuyichao has quit [Ping timeout: 480 seconds]
marvin24 has joined #asahi
<iaguis>
I noticed systemd-{journald,logind,homed} getting killed with SIGABRT when resuming from s2idle. It seems the watchdog times out and kills them.
<iaguis>
I ran "python3 -c "import time; print(time.monotonic())"" before and after suspend and this is indeed what's happening on my M2 Air.
<iaguis>
I'm not sure if this is happening because something is kept awake while the laptop is asleep or this is normal kernel behavior on s2idle suspend.
<jannau>
iaguis: that's currently expected. the cpu keep running in s2idle
<iaguis>
also, my Thinkpad T14s doesn't show this behavor when doing s2idle: it seems CLOCK_MONOTONIC is stopped when suspended
<iaguis>
or do you mean it's currently expected but it should be fixed in the future?
hendry has joined #asahi
<jannau>
I was describing how the system behaves
Treibholz has quit [Quit: WeeChat 3.7.1]
Treibholz has joined #asahi
<iaguis>
I see, at least I know it's not something weird with my machine :)
<iaguis>
but I think it should be fixed
yuyichao_ has quit [Quit: Konversation terminated!]
yuyichao has joined #asahi
jacksonchen666 has quit [Ping timeout: 480 seconds]
jacksonchen666 has joined #asahi
possiblemeatball has joined #asahi
delsol has joined #asahi
Techcable has quit [Ping timeout: 480 seconds]
delsol has quit [Remote host closed the connection]
Techcable has joined #asahi
flying_sausages has joined #asahi
yumasi has joined #asahi
yumasi has quit [Remote host closed the connection]
<corion>
Went from 100% to 55% in 8h standby. Not too shabby. Not sure how asleep it was though. I thought I turned it off but I think it must've started so maybe it was just idle.
<corion>
I dont know if this is relevant to your thing.
yumasi has quit [Remote host closed the connection]
<karolherbst>
ahh yes, so DP-MST works with macos
<karolherbst>
so I guess that solves this mystery
<karolherbst>
so yeah... are there any DP-MST WIP patches?
<sven>
what you’re using is altmode and not MST. MST would be at least two separate DP streams
<tpw_rules>
i'm getting an occasional hang on boot from dcp failing to start ("wait for power timed out"), possibly in a racy way: https://pastebin.com/J5LkdH7F seems like it might be more likely to happen rebooting from macos (asahi is my default boot option) but i can't make it happen on command
<ChaosPrincess>
karolherbst: what you have is the standard chinese usb hub, and that one is altmode
<ChaosPrincess>
i have one from the same oem but with different logo
<sven>
to make thing more confusing: MST mirror/duplicate/however-it’s-called is backward compatible with normal DP single stream mode
<sven>
+s
<karolherbst>
ah yeah.. that might be that this dock supports both
<karolherbst>
and that macos only uses altmode there
<sven>
as I said, this hardware doesn’t support MST
<karolherbst>
fair enough.. but I was under the impression that all this is actually still MST even if it's called "alt mode"
<corion>
I dont know the first thing about this, but sven seems confident enough to indicate he knows what he's talking about. I'd say fold your hand, karolherbst :)
<sven>
i can’t route two DP streams to a single output in the display crossbar when using DP mode
<karolherbst>
which is totally fair
<sven>
what that hub is using is that DP altmode Pin assignment where two lanes are used for the USB3 signals and the two other lanes for DP
<karolherbst>
probably
<sven>
corion: I’ve read way too many usb c and display port specs in the past months to get this working :(
<corion>
sven: I figured :)
<corion>
(and I'm glad you did!)
<ChaosPrincess>
sven: wait, whats that about display crossbar?
<ChaosPrincess>
Isnt it dcp -> crossbar -> dp hardware?
<sven>
ChaosPrincess: the thing that handles dcpextN to atcphy-{dpout, usb4-dpin-0/1}
<karolherbst>
anyway... I guess with the altmode patches it might just work on my system then?
<sven>
so you can route a single dcpext stream to a atcphy port when using DP altmode or two streams when using usb4 tunneling
<ChaosPrincess>
Oh, ok
<sven>
i wouldn’t call the state they are in “work”, but sure
<karolherbst>
:D fair
<sven>
they still have a few issues like you can’t have the display connected during boot
<ChaosPrincess>
you said "routing dp to crossbar", so that was confusing
<sven>
and it just doesn’t like sone displays yet
<sven>
*some
<karolherbst>
apparently it also didn't like the dock to be connected while suspended
<karolherbst>
or at least I had to replug it so it identifies all the USB devices on it
<ChaosPrincess>
sven: btw, what machine are you doing this work/research on?
<corion>
I've noticed that I have to keep my headphones connected during boot for them to work (re: boot stuff).
<corion>
Which makes sense I guess.
<sven>
ChaosPrincess: m1 Mac Mini
<ChaosPrincess>
Oh, nvm then
<sven>
but other people have tested it on the pro/max and m2
<karolherbst>
I just wanted to check if the m2 can handle my external 4K display better than intel does...
<ChaosPrincess>
Im kinda curious about if there is extra weirdnes on m1 mbp
<ChaosPrincess>
For touchbar
<corion>
Isn't the touchbar watchOS?
<sven>
I’d be surprised tbh, touchbar should be mostly separate from the atcphy mess
<ChaosPrincess>
on t2, not m1
<corion>
ah
yumasi has joined #asahi
<ChaosPrincess>
There is digitizer, which is basically the same chip as in the iphone display, and something mipi related
cylm_ has joined #asahi
<corion>
If I wanted to make some asahi stickers, would I need some sort of license approval from whoever made that awesome logo? Or what license is it under?
<corion>
"Licensed under CC BY-SA 4.0", found it, I think.
cylm has quit [Ping timeout: 480 seconds]
<corion>
So it seems I can make stickers using the logo. Sweet!
<marcan>
karolherbst: DP-MST hardware will usually send a single stream to multiple outputs
<marcan>
so all those "MST" docks with two ports just end up spitting out the same thing to both when used with non-MST hosts
<marcan>
they don't *require* MST
<marcan>
as long as you don't need two discrete outputs
<karolherbst>
ahh, fair
<marcan>
re suspend that is known broken
<marcan>
(lots of things are broken in suspend, but USB especially)
<marcan>
USB is very cursed on these machines
<karolherbst>
heh
<tpw_rules>
s/ on these machines//
<marcan>
I need to do deeper testing, but from what I saw simply putting ports into suspend and then trying to take them out doesn't work
<karolherbst>
huh
<marcan>
I mean, it's a Synopsys core
<marcan>
so broken is the default state
<sven>
we also need to reset the entire usb controller and phy every time a new device is connected
<marcan>
at least it's not dwc2 broken, where you need CPU low latency assist to even make the USB ports work and perform properly
<karolherbst>
wait what?
<marcan>
this is why USB sucks on the raspberry pi
<marcan>
those use dwc2
<sven>
yeah, this thing is quite broken
<marcan>
M1s use dwc3
<karolherbst>
ahh
<marcan>
so dwc2 was *really* bad, dwc3 is just mostly bad
<corion>
Progress!
<tpw_rules>
but dwc2 is still usb3?
<marcan>
no, dwc2 is usb2
<karolherbst>
well.. as long as I can trash my x86 laptop in a year I don't mind if it takes a bit longer 🙃
<ChaosPrincess>
marcan: wtb story about dwc2 brokenness.
<ChaosPrincess>
Cause wtf, how do you even break it that badly
<marcan>
I need to log what macOS does to figure out whether ports breaking in suspend mode is expected (maybe macOS does full resets, that'd be cursed...) or whether there's a magic incantation to make it not explode
<marcan>
ChaosPrincess: Synopsys.
<karolherbst>
marcan: mhh.. just do a full reset on resume and move on or something?
<sven>
ohhhh… i think macOS might do a full reset. Remember that thing where I broke the ports when disconnecting the cable at just the right time in macOS?
<sven>
the thing that worked in Linux because I did that additional hard reset
<sven>
I could also restore those ports with a suspend cycle under macOS
<marcan>
karolherbst: no, that is cursed because it means either 1) your USB storage devices fall off the bus on suspend and your filesystems crash, or 2) you enable the stupid cheat mode that works around that, then someone unplugs a USB driver and plugs in a different but identical drive during sleep and on resume their filesystem gets corrupted to hell and back
<marcan>
I'm going to be very sad if 2) is the state of affairs on macOS
<karolherbst>
sure, but.. does it even work on x86?
<marcan>
yes it does
<marcan>
it also works on the Type A ports on M1s
<marcan>
because those use an xHCI controller not made by Synopsys
<karolherbst>
heh
<marcan>
as I said, Synopsys
<marcan>
sven: the only safe way to make it work with a full reset is if we can detect a disconnect during sleep, which is probably possible with tipd irq flags or something, but
<marcan>
that means even *more* cursed abstraction breaking bullshit in linux
<tpw_rules>
sven: is dcp startup still known to be racy? i'm getting hangs on boot and "apple-dcp 38bc00000.dcp: wait for power timed out". i have more logs and info if helpful, but i thought it was a semi-known problem
<marcan>
and I'm not even sure if it can be made to work with ports chained off of hubs, because if you reset everything wouldn't that lose the disconnect state on DFPs?
<marcan>
I hate USB.
<marcan>
I'm going to sleeep.
<sven>
urgh
icanflyit has joined #asahi
<sven>
yeah, we can probably detect it with some tipd hacking but ugh
<sven>
I’d rather not :D
<marcan>
if we can't make it work with downstream hubs it's a lost cause anyway
<marcan>
but then we should be able to cause epic filesystem corruption on macOS too
<marcan>
that would be a fun bug to file
<marcan>
I should test that...
<sven>
tpw_rules: i was only tricked into adding dcpext support to DCP, jannau does all the other DCP stuff :)
<icanflyit>
I have a question - what branch is the linux-asahi-edge kernel built from?
<marcan>
asahi tagged releases
<tpw_rules>
sven: ah okay, thank you
<sven>
maybe macOS only reset when it thinks nothings connected? That would also fix that broken port state I saw
Dementor has quit [Remote host closed the connection]
Dementor has joined #asahi
delsol has joined #asahi
delsol has quit [Remote host closed the connection]
delsol has joined #asahi
<jannau>
tpw_rules: I don't think that's an indication of a race. either the timeout is under some circumstances too low (unlikely for laptops since I never seen that even with tracing), there's a logic bug or something crashes dcp
jamespmorgan has joined #asahi
unicordion has quit [Quit: unicordion]
jamespmorgan has quit [Ping timeout: 480 seconds]
jamespmorgan has joined #asahi
possiblemeatball has joined #asahi
__nick__ has joined #asahi
___nick___ has quit [Ping timeout: 480 seconds]
yumasi has quit [Remote host closed the connection]
possiblemeatball has quit [Quit: Leaving]
hertz has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
__nick__ has quit [Ping timeout: 480 seconds]
jamespmo_ has joined #asahi
Brainium has joined #asahi
jamespmorgan has quit [Ping timeout: 480 seconds]
Guest345 has quit [Quit: G-line: User has been permanently banned from this network.]
vx has joined #asahi
vx is now known as Guest551
jamespmo_ has quit [Ping timeout: 480 seconds]
<icanflyit>
so I'm trying to modify the kernel so I can use the space in the notch (obviously wont be recognized by applications or WMs but I don't mind), so I've removed where notch_height is added in dcp_flush and then in parse_mode, I have it return the mode without the notch_height.
<icanflyit>
That all works fine in x, but my question is what do I need to do to get that working with the mesa driver
<icanflyit>
somewhere a resolution of 3456x2160 is defined, so when I try to run gpu accelerated with mesa (im testing with gnome) It attemps to use 3456x2160 and then I get invalid mode
<j`ey>
icanflyit: there's already a command line option to use the notch
<j`ey>
apple_dcp.show_notch=1 not officially supported yet
jamespmorgan has joined #asahi
<icanflyit>
That option works on latest tagged release?
<j`ey>
yeah
jamespmorgan has quit [Ping timeout: 480 seconds]
<icanflyit>
show_notch=1 means that the screen extends through the notch?
<jannau>
yes. thought process was that the notch becomes noticeable if the used display area extends left and right of it
pthariensflame has joined #asahi
delsol has quit [Remote host closed the connection]
<jannau>
that option will not be the officially supported option. There will be eventually display modes for the partially occluded display
jamespmorgan has joined #asahi
<cy8aer>
so selecting the resolution direct in the desktop, right? I tried the option and it is possible to let the notch invisible in gnome when modify the height to 48px and put the clock to the right hand side.
<cy8aer>
(panel height of course)
<cy8aer>
And because I can fully live with it it is my default now. jannau please warn when "apple_dcp.show_notch=1" disappears 😉
jamespmorgan has quit [Ping timeout: 480 seconds]
jamespmorgan has joined #asahi
jamespmorgan has quit [Ping timeout: 480 seconds]
rowin has joined #asahi
<rowin>
hello
<rowin>
anyone here?
<tpw_rules>
nope, sorry
<ChaosPrincess>
no, there is absolutely nobody here :P
<rowin>
when this chat usually achieve?
<rowin>
active?
<ChaosPrincess>
if you have a question, you should just ask it
<rowin>
my question is, how long will it take for asahi linux to have full support for M1/M2 macs