marcan changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | General project discussion | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
<marcan>
(I had this on my TODO list but he beat me to it)
<rwhitby>
I have one on order, so I'll be following how to update the firmware :-). I like how the streaming time is just after I finish work for the day :-)
<rwhitby>
ah, the github page is pretty self-explanatory
<dhewg>
that's not the macrosilicon thing, is it?
<JTL>
Not what you want, but I have a PCIe capture card that can do lossless captures with a DVI input up to 2048x204. I beleive it uses a Xilnx FPGA internally.
<JTL>
er 2048x2048 iirc
<marcan>
dhewg: no, these are the VXIS ones
<marcan>
the ones that do real USB3 and uncompressed YUV 1080p30
<JTL>
not bad
<marcan>
but unfortunately have a hardcoded sharpening filter, and somewhat poor color conversion; this patch fixes the former
<marcan>
I want to do a deep dive into the video scaler registers to see if I can fix the latter
<marcan>
I suspect poorly generated gamma tables
<dhewg>
nice, but do i read that correctly, and *that* only works in one usb plug orientation?
<marcan>
yes, because it's type C but they were too lazy to put a mux in there
<dhewg>
lol?
<marcan>
type C requires muxes on every connector for SuperSpeed
<JTL>
standards noncompliant USB-C?
<marcan>
since you don't know which side the lines end up on
<JTL>
Funny
pakl has quit [Ping timeout: 260 seconds]
pakl has joined #asahi
aimileus[m] is now known as aimileus
pakl has quit [Ping timeout: 272 seconds]
pakl has joined #asahi
VinDuv has quit [Quit: Leaving.]
acelogic has quit [Read error: Connection reset by peer]
acelogic has quit [Remote host closed the connection]
acelogic has joined #asahi
volpe has joined #asahi
bostjan has joined #asahi
acelogic has quit [Remote host closed the connection]
acelogic has joined #asahi
volpe_ has joined #asahi
florianscholz has joined #asahi
florianscholz is now known as fsc
mogery has joined #asahi
wiloon has joined #asahi
mogeryy has joined #asahi
fsc_ has joined #asahi
fsc has quit [Ping timeout: 256 seconds]
mogery has quit [Ping timeout: 256 seconds]
Esmil has quit [Ping timeout: 256 seconds]
Esmil has joined #asahi
Axenntio has joined #asahi
minicom7 has joined #asahi
minicom has quit [Ping timeout: 240 seconds]
xarkes has quit [Ping timeout: 240 seconds]
DrWhax has quit [Ping timeout: 240 seconds]
Hexagon has quit [Ping timeout: 240 seconds]
Foxboron has quit [Ping timeout: 240 seconds]
vijfhoek has joined #asahi
<j`ey>
marcan: weird, I didnt get a notification for the stream
DrWhax has joined #asahi
Hexagon has joined #asahi
Foxboron has joined #asahi
<jix>
I got one, but I think randomly dropping notifications is a youtube thing? at least it's a common complaint AFAICT
<vijfhoek>
I believe most streamers have a secondary notification on Discord (or IRC in our case)
<vijfhoek>
because YouTube's (and also Twitch's) notifications just don't work well
<j`ey>
so it looks like he has a irqchip-aic.c
<j`ey>
anything else "important" happen?
<jix>
I'm still catching up, but where I am, he just fixed some samsung uart settings so FIFO mode works
<jix>
love youtube's keyboard shortcuts for this, pressing L on youtube skips 10 secs fwd (and pressing ? shows all shortcuts)
leonlg has joined #asahi
xarkes has joined #asahi
acelogic has quit [Ping timeout: 264 seconds]
Axenntio has quit [Ping timeout: 272 seconds]
r1fl has joined #asahi
jaXvi has quit [Quit: Leaving]
thresh has joined #asahi
taziden has quit [Ping timeout: 240 seconds]
jaXvi has joined #asahi
mogeryy has quit [Read error: Connection reset by peer]
taziden has joined #asahi
volpe has quit [Remote host closed the connection]
volpe_ is now known as volpe
raster has joined #asahi
mogery has joined #asahi
mogery is now known as Guest99946
Guest99946 has quit [Client Quit]
mogeryy has joined #asahi
maor26 has joined #asahi
wiloon has quit [Quit: Connection closed for inactivity]
mogeryy has quit [Client Quit]
mogeryy has joined #asahi
mogeryy has quit [Remote host closed the connection]
mogeryy has joined #asahi
mogeryy has quit [Remote host closed the connection]
mogeryy has joined #asahi
maor26 has quit [Remote host closed the connection]
maor26 has joined #asahi
lucid_0x80 has joined #asahi
<Necrosporus>
What is chickens.py doing?
snalty has joined #asahi
<j`ey>
setting magic values in CPU registers
<Necrosporus>
How does he have assembler code in py file?
<j`ey>
he copied and pasted it :P
parallax has left #asahi [#asahi]
mogeryy has quit [Remote host closed the connection]
fewfewwe has joined #asahi
amw2 has joined #asahi
amw2 has quit [Ping timeout: 246 seconds]
vimal has joined #asahi
vimal has quit [Remote host closed the connection]
vimal has joined #asahi
amw2 has joined #asahi
amw2 has quit [Ping timeout: 265 seconds]
<yrlf>
Necrosporus: the python code talks over uart to m1n1, which is effectively an RPC interface for experimenting
<yrlf>
so it can say "yo, m1n1, can you execute that instruction real quick?"
<Necrosporus>
yrlf, weren't mar|can going to make mini talk with the host via USB directly without UART?
minicom7 is now known as minicom
<jix>
that's still planned eventually AFAIK
amw2 has joined #asahi
amw2 has quit [Ping timeout: 256 seconds]
amw2 has joined #asahi
PdimH has joined #asahi
<luca020400>
marcan: it's guarded by CONFIG_ARM, not a thing on arm64
amw2 has quit [Ping timeout: 265 seconds]
ephe_meral has joined #asahi
maor has joined #asahi
lucid_0x80 has quit [Read error: Connection reset by peer]
maor26 has quit [Ping timeout: 256 seconds]
egghead has joined #asahi
egghead has quit [Excess Flood]
egghead has joined #asahi
jamadazi has joined #asahi
egghead has quit [Excess Flood]
egghead has joined #asahi
fewfewwe has quit [Quit: Connection closed]
<luca020400>
There's fixed-clock
<luca020400>
marcan: ^
aknt has joined #asahi
plainbits has joined #asahi
ephe_meral has quit [Ping timeout: 256 seconds]
ephe_meral has joined #asahi
TomJepp has joined #asahi
PdimH has quit [Quit: WeeChat 2.9]
modwizcode has joined #asahi
djb has quit [Read error: Connection reset by peer]
djb_ has joined #asahi
aknt has quit [Quit: Leaving]
Janhouse has joined #asahi
<Janhouse>
I'm late to the party, so I was wondering if Asahi has anything to do with guys from Corellium or is that completely unrelated?
<Janhouse>
I guess they just booted from external drive, while you are trying to get it going from the mac itself?
<yrlf>
marcan has a stricter copyright policy, since he absolutely wants to make sure that the asahi work is upstreamable
<yrlf>
Janhouse: both are booting from the mac itself, iBoot doesn't handle external devices
<yrlf>
(AFAIK)
<Janhouse>
yrlf, thanks for the info!
<Janhouse>
What are correlium guys doing differently when it comes to copyright? :)
<jix>
Janhouse: their project they're basing their efforts on didn't follow the guidelines for upstream linux, so now they are refactoring that to clean things up
<jix>
while the asahi project starts from scratch following those guidelines and some additional project guidelines to make following upstream guidelines easier
<jix>
but for stuff that is ready for upstream linux, I expect there to be collaboration... for example m arcan is already using one of the patches corellium sent to the LKML and they added him on CC for their patches
<davidrysk[m]>
hmm, one of their changes is -- Apple processors do not keep register state across WFI. As such, put a mechanism in cpu_ops to put a custom sleep function instead.
<Shiz>
yeah, their WFI invokes a WFE state instead
<modwizcode>
I feel like the corellium people are doing this because we annoyed them somehow
<volpe>
so they are going to upstream first so that it would make mar|can life easier? So toughtful of them
raster has quit [Quit: Gettin' stinky!]
Tokamak has joined #asahi
skg has quit [Quit: 〜バイバイ〜！]
raster has joined #asahi
skg has joined #asahi
<Shiz>
personally if they're properly upstreaming stuff and it all checks out copyright-wise i'm all for it
<sven>
if they go through the upstream process for their project it's a win-win for everyone. m arcan won't have to redo that work and will still be able to contribute to the upstream process (by independently reviewing the work). and on the other hand m arcan's work also gets a review by the Corellium guys (hopefully) who have a lot of knowledge on apple
<sven>
silicon.
<sven>
the whole process should go much faster that way and result in higher quality code in the kernel :)
<Shiz>
aye
<sven>
and m arcan also gets to focus on all the tricky little things earlier that will come up after most hardware works for the first time
<sven>
and there will be a *lot* of those things
sharpiro has joined #asahi
<volpe>
Yeah on the topic of arm porting (and in particular apple hardware) they probably have more experience so if everything goes to upstream good for them and good for asahi. To benefit is really Linux in the end.In my opinion the going forward there's only some risk of duplicated efforts but I hope they could coordinate in some way to avoid that.
<modwizcode>
It will help I think even if efforts are duplicated because then two people have tried it and we have greater confirmation of how stuff works]
<vijfhoek>
Yeah, effort duplication isn't a bad thing per se
<jix>
especially if different methods for reversing were used... poking the HW as black box here, and (I assume) observing what xnu does for corellium (given that virtualizing iOS is what they do as a company)
jamadazi has joined #asahi
jamadazi has quit [Client Quit]
<volpe>
Yeah did't thought of that... I'm absolutely no expert in this stuff ahah
VinDuv has joined #asahi
hungryhugo has quit [Quit: Connection closed]
macc24 has joined #asahi
macc24 has left #asahi ["WeeChat 3.0"]
Axenntio has joined #asahi
alicela1n[m] has joined #asahi
anarsoul has joined #asahi
<alicela1n[m]>
hi!
alicela1n[m] is now known as Alice[m]
<jn__>
hello Alice[m]
Axenntio has quit [Ping timeout: 272 seconds]
uberushaximus has quit [Ping timeout: 256 seconds]
plainbits has quit [Quit: Bye!]
leonlg has quit [Ping timeout: 260 seconds]
<jix>
marcan: from yt: ob1cannoli: rx and tx are backwards
<j`ey>
yeah, not very helpful without context :|
<j`ey>
marcan: "zadintuvas1
<j`ey>
Corellium seems to have 1 << 9 for TX and 1 << 8 for RX."
<emily>
wait he's still streaming? I like, slept and everything!
TheLink has quit [Ping timeout: 240 seconds]
TehLink is now known as TheLink
<emily>
tfw the cause of Linux-on-Apple-Silicon is being massively accelerated by the power of mutual petty spite
<modwizcode>
lmao I thought the same thing
<volpe>
Maybe he confused the clocks and he's going to go to sleep at 23:00 UTC :P
<modwizcode>
he forgot to set his internal RTC registers
HotSwap has joined #asahi
Bublik has quit [Ping timeout: 240 seconds]
jamadazi has joined #asahi
Bublik has joined #asahi
<jamadazi>
uh i really hope marcan doesnt burn himself out due to internal pressure or something, i hope he isn't making himself work 12+ hrs a day just because of said petty spite or some other silly reason like that
<dottedmag>
Come on, it's just a thrill of making things work.
<jamadazi>
yeah, hope so :)
<modwizcode>
I feel like he might always be like this :p
<jix>
jamadazi: not the first time he pulls debugging sessions like this
<jn__>
late night debugging can be pretty tempting when stuff *almost* works. it's like in a CTF when you almost have the flag
<jamadazi>
oh yes i know the feeling very well :) been there
<jamadazi>
as long as it's healthy and sustainable, all is good
<jamadazi>
it's awesome to see all the enthusiasm to get this thing working
<volpe>
Did he even have dinner at some point? ahah
<jamadazi>
idk, but i can totally relate to the feeling of getting totally carried away coding, that's why i was sharing these friendly concerns about not forgetting to take care of your health and not burning out
citruscitrus[m] has joined #asahi
<vijfhoek>
as long as this doesn't happen too often, it's fiiine
<Fanfwe>
He probably could feel he was so close, he couldn't go to bed without finishing this
<Fanfwe>
And it paid off, he now has a working userspace
<j`ey>
hopefully..
<j`ey>
he had some issus before that printks "fixed"
VinDuv has quit [Quit: Leaving.]
QuinnDTE has joined #asahi
QuinnDTE has left #asahi [#asahi]
bostjan has quit [Quit: Leaving]
<tbodt>
didn't he say earlier that he was trying to fix his sleep schedule
<modwizcode>
I think he did lol
<modwizcode>
maybe this is fixing it
<modwizcode>
if he just keeps going he'll overflow the sleep schedule
<volpe>
Oh shit, hes going to end the stream!
<tbodt>
marcan: sleep
<volpe>
marcan: sudo bonk
<brinly>
wow, 12 hour stream
<tbodt>
14 hour
<brinly>
Oh wow
<tbodt>
youtube might have cut off the first 2 hours though
<brinly>
I slept and did work this morning
<brinly>
Already
<brinly>
And he’s still streaming from last night
<tbodt>
almost 15 hours (14:51)
<brinly>
I missed most of it, what got achieved? and how close is it to a stage that external contributors can meaningfully contribute
<jn__>
brinly: FIQ+timer works now, framebuffer works now (it was relatively little work), and earlycon UART output works (but the full UART driver has problems)
<eta>
did he not solve the UART
<jn__>
eta: correct
<eta>
I swear the last 6 hours were UART debugging, feels bad man :(
<brinly>
it’s patches to a existing kernel driver or a brand new one?
<j`ey>
former
<jn__>
wait, now i'm misreading stuff
<brinly>
on mobile -> is the source tree up?
<jn__>
sorry, i'm bad at negative questions
<j`ey>
brinly: dont think so
<tbodt>
he did not solve the UART
<davidrysk[m]>
AIC issues
<eta>
huh YT haven't archived the stream yet
<davidrysk[m]>
it's >12h
<davidrysk[m]>
and youtube doesn't guarantee archival of >12h streams