ChanServ changed the topic of #asahi-dev to: Asahi Linux: porting Linux to Apple Silicon macs | Non-development talk: #asahi | General development | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Logs: https://alx.sh/l/asahi-dev
pb17 has quit [Ping timeout: 480 seconds]
chadmed_ has quit [Remote host closed the connection]
john-cabaj has quit [Read error: Connection reset by peer]
john-cabaj1 is now known as john-cabaj
john-cabaj has quit [Quit: john-cabaj]
mkurz has quit [Read error: Connection reset by peer]
<maz>
jannau: forget what I said. it reproduces the same way. I still had this silly dwc3 hack in my tree.
JayBeeFOSS has quit [Ping timeout: 480 seconds]
JayBeeFOSS has joined #asahi-dev
thunfisch is now known as Guest6444
thunfisch has joined #asahi-dev
Guest6444 has quit [Ping timeout: 480 seconds]
thunfisch is now known as Guest6447
thunfisch has joined #asahi-dev
pb17 has quit [Ping timeout: 480 seconds]
Guest6447 has quit [Ping timeout: 480 seconds]
<kettenis>
SME without SVE seems really weird
<maz>
kettenis: yeah. it breaks a lot of expectations in userspace that you can exit streaming mode and still use SVE. but it you look at the architecture, there is no actual dependency and you are allowed to build one without the other.
<maz>
for Apple, that's not really a problem: they already had AMX without SVE, so no SW expects SVE at all. but running non-Apple SW on that HW becomes more challenging.
<chadmed>
kwin does the right thing and if a cursor swap is reject will immediately fall back to software cursors
<chadmed>
i probably could have saved ~3 days if i had just tried that initially...
<chadmed>
s/reject/rejected
thunfisch is now known as Guest6455
thunfisch has joined #asahi-dev
<chadmed>
ugh firefox...
thunfisch is now known as Guest6456
<chadmed>
im gonna guess off the bat that mutter does not behave sensibly here....
thunfisch has joined #asahi-dev
Guest6455 has quit [Ping timeout: 480 seconds]
<jannau>
this was discussed in#dri-devel a while ago with the conclusion that not all compositors fall back to sw cursors
<chadmed>
yeah just read it, that discussion was not very helpful
<chadmed>
"let's do a new plane type but also tell all of userspace that plane types are now irrelevant" ???
<chadmed>
cmon....
<chadmed>
ugh yeah this is a mess, weston just crashes
<chadmed>
gtk hangs the entire commit chain
Guest6456 has quit [Ping timeout: 480 seconds]
thunfisch is now known as Guest6460
thunfisch has joined #asahi-dev
Guest6460 has quit [Ping timeout: 480 seconds]
pb17 has quit [Ping timeout: 480 seconds]
aradhya7 has joined #asahi-dev
nela has quit [Ping timeout: 480 seconds]
<leio>
chadmed: I think I really should get the upgrade in first too; my version doesn't do some of the fancy stuff yet
<leio>
chadmed: BUT on that note, it might be worth first testing how the older version acts.
<chadmed>
leio: no rush or anything, like i said judging by how gtk apps behave i have little hope that mutter does the right thing currently :p
<leio>
gtk is probably going to differ wildly on whether it's gtk3, gtk4 or gtk-4.14
<chadmed>
firefox and sysprof both misbehave
<chadmed>
gimp also misbehaves
thunfisch is now known as Guest6469
thunfisch has joined #asahi-dev
<leio>
firefox is not gtk; sysprof is gtk4, gimp is gtk2 unless you've keyworded the gimp3 beta
<chadmed>
firefox is gtk3...
<leio>
not really, it just uses it for theming and some low-level usage of things
Guest6469 has quit [Ping timeout: 480 seconds]
<leio>
it's so custom we can't consider it anywhere near a native gtk3 app in any way
<leio>
it does its own wayland subsurfaces too, etc
thunfisch is now known as Guest6470
<leio>
of course kind of an important app to have working good :)
thunfisch has joined #asahi-dev
<chadmed>
im of the opinion that we should just ship this as-is to provide impetus to fix this stuff
Guest6470 has quit [Ping timeout: 480 seconds]
<chadmed>
but breaking all gtk and firefox is probably too much breakage
pb17 has joined #asahi-dev
<leio>
how does it even impact those, thought we're talking about compositor compositing mouse cursor
<chadmed>
that was what i was asking myself too, but all my non-gtk non-firefox applications are fine
<leio>
I don't think gnome-shell people would be unwilling to have chats about it on their matrix
<leio>
they are all in a HDR hackfest right now though
<leio>
some gtk devs are not, main one that should know about this stuff is IRC only
<leio>
but I can try hopefully before your next early evening
<chadmed>
hmmm libreoffice and vlc also have the same behaviour...
<chadmed>
maybe worth testing on mutter
<leio>
and vlc is qt; libreoffice is another case of its own toolkit on top of lower level gtk3
<leio>
and I think they have a qt variant too these days
<leio>
eh, that toolkit is called VCL; so that and VLC :)
<chadmed>
yeah thats why i mentioned them, seems to not be gtk-specific so might be worth seeing if mutter by itself does the right thing and this is an application-level issue for some reason
<chadmed>
also yeah libreoffice has a kf5 backend too which is what im using obviously
<chadmed>
will see if this clears up when libreoffice gets qt6/kf6
nela has joined #asahi-dev
m42uko_ has joined #asahi-dev
m42uko has quit [Read error: Connection reset by peer]
* j`ey
can't find the kwin code that implements fallback
pjakobsson has quit [Remote host closed the connection]
pjakobsson has joined #asahi-dev
<jannau>
chadmed: I also don't understand how applications would be effected by offering cursor/overlay planes. that is either a compositor or driver bug. applications usually don't have access to the kms device and shouldn't try to use it (ignore special cases like mpv's kms vo)
<jannau>
I would assume so. also thunderbolt 4 for devices with internal display
<j`ey>
neat
<jannau>
cpu cores still have sawtooth/everest in their compatibility
<maz>
jannau: is that looking at the DT?
<jannau>
yes
<maz>
shame this thing is so locked down. would have made a hell of a "shelf computer"...
<j`ey>
maz: it'll come to a mini/studio eventually!
<maz>
j`ey: by then I will have missed the opportunity to get $work to pay for it! ;-)
<j`ey>
hehe
irth has quit [Quit: Connection closed for inactivity]
pb17 has quit [Ping timeout: 480 seconds]
jeisom has quit [Quit: Leaving]
pb17 has joined #asahi-dev
jeisom has joined #asahi-dev
ddxtanx has joined #asahi-dev
<maz>
jannau: another issue I forgot to mention is that on my M2 Air, I get a constant screaming of "[ 235.022683] power_supply macsmc-battery: driver reporting unsupported charge behaviour"
<maz>
removing macsmc_power "fixes" the behaviour.
<j`ey>
thats probably a fw issue?
<j`ey>
oh, nvm, them2 air should be on not 11.x fw
<maz>
j`ey: maybe. but it was OK with 6.8, and only the same patches with 6.9 showed the issue. so it has to be a change that is unrelated to the Asahi patches.
<jannau>
there was a patch flying around with charge behavior
<jannau>
"phy-apple-atc b03000000.phy: pipehandler lock not acked, this type-c port is probably dead until the next reboot." is still there and usb3 is not working