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
ahuar88_ has joined #asahi-dev
ahuar88 has quit [Read error: Connection reset by peer]
ahuar88_ has quit [Read error: Connection reset by peer]
ahuar88 has joined #asahi-dev
ahuar88_ has joined #asahi-dev
ahuar88 has quit [Read error: Connection reset by peer]
ahuar88 has joined #asahi-dev
ahuar88_ has quit [Read error: Connection reset by peer]
ahuar88 has quit [Read error: Connection reset by peer]
ahuar88 has joined #asahi-dev
user982492 has joined #asahi-dev
ntal has joined #asahi-dev
darkapex1 has joined #asahi-dev
rpirea has quit [Ping timeout: 480 seconds]
rpirea has joined #asahi-dev
darkapex has quit [Ping timeout: 480 seconds]
deteg1337 has joined #asahi-dev
deteg1337 has quit [Ping timeout: 480 seconds]
nsklaus has quit [Quit: ZZZzzz…]
pthariensflame has joined #asahi-dev
pthariensflame has quit []
ntal has quit [Remote host closed the connection]
<marcan> :/
deteg1337 has joined #asahi-dev
deteg1337 has quit [Ping timeout: 480 seconds]
user982492 has quit [Quit: Textual IRC Client: www.textualapp.com]
A_L_I_C_E has joined #asahi-dev
nicolas17 has quit [Quit: Konversation terminated!]
nicolas17 has joined #asahi-dev
deteg1337 has joined #asahi-dev
A_L_I_C_E has quit [Read error: Connection reset by peer]
A_L_I_C_E has joined #asahi-dev
deteg1337 has quit [Ping timeout: 480 seconds]
nicolas17 has quit [Remote host closed the connection]
deteg1337 has joined #asahi-dev
deteg1337 has quit [Ping timeout: 480 seconds]
A_L_I_C_E has quit [Quit: Quit]
A_L_I_C_E has joined #asahi-dev
richyliu2 has joined #asahi-dev
<richyliu2> what's the status on the microphone? I want to help with the reverse engineering if that hasn't been done already
A_L_I_C_E has quit [Ping timeout: 480 seconds]
A_L_I_C_E has joined #asahi-dev
deteg1337 has joined #asahi-dev
deteg1337 has quit [Ping timeout: 480 seconds]
bisko has joined #asahi-dev
bisko has quit []
ahuar88 has quit [Remote host closed the connection]
ahuar88 has joined #asahi-dev
ahuar88 has quit [Remote host closed the connection]
ahuar88 has joined #asahi-dev
deteg1337 has joined #asahi-dev
<marcan> ChaosPrincess: can you send the t8112 PWM DT changes too? (on top of the asahi-soc/dt branch)
deteg1337 has quit [Ping timeout: 480 seconds]
<marcan> I guess it's about time I bring up the M2 Max...
<marcan> might not bother to stream this one since honestly it's going to be boring unless something really unexpected comes up
<marcan> just chickens and fuses everywhere
nsklaus has joined #asahi-dev
A_L_I_C_E has quit [Read error: Connection reset by peer]
A_L_I_C_E has joined #asahi-dev
Ry_Darcy has joined #asahi-dev
<Ry_Darcy> Any sign of the M2 Mini being ready for testing soon. I have been looking at this expensive paperweight for three months now. I appreciate you are all doing your best, but you know how it is...
<marcan> nope.
<marcan> buying unsupported devices is on you, not us :)
<marcan> this might end up requiring a complete rewrite of the DCP code in m1n1 to be much more complex, TBD
A_L_I_C_E has quit [Ping timeout: 480 seconds]
A_L_I_C_E has joined #asahi-dev
<ChaosPrincess> marcan: sent.
<marcan> thanks!
<ChaosPrincess> i personally havent tested this one (no m2 machines), but iirc someone else did :P
<marcan> assuming it's what we have in the downstream kernel it works for me :p
<ChaosPrincess> and stream the chickens anyway, at least its somewhat educational, and i might come and shitpost in chat :P
<ChaosPrincess> its almost what we have in downstream, yes
<marcan> actually someone else already did the chickens, I just need to deal with fuses I guess
<Ry_Darcy> @marcan True, but has been noise about the M2 Mini since last year with corresponding code activity. Not unreasonable to acquire a device and hope for the best. Maybe in 2024...
<Ry_Darcy> but there has been..
___nick___ has joined #asahi-dev
<marcan> if you want it faster, you can help out ;)
<Ry_Darcy> @marcan an old Mainframer like me? Not a chance! I will leave that to the young ones.
<Ry_Darcy> Maybe I will sell the "*ç&&/ thing.
<ChaosPrincess> well, arent those machines kinda mainframe-y just due to the amount of io coprocessors?
<psykose> every machine is a mainframe if you squint really hard
<Ry_Darcy> No idea. Don't get me start on I/O heavy lifing and PRSM. Greetings from White Plains and Poughkeepsie. Enough said.
___nick___ has quit []
<Ry_Darcy> start = started
<psykose> PRecision Strike Missile?
<Ry_Darcy> lifing = lifting. All thumbs this morning.
<ChaosPrincess> processor resource/system manager
<ChaosPrincess> ibm really likes their weird names for normal stuff
<ChaosPrincess> this is basically a hypervizor, somewhat like xen
<Ry_Darcy> Ahem a Type 1 Hypervisor.
___nick___ has joined #asahi-dev
<ChaosPrincess> xen is a type 1
<ChaosPrincess> iirc
<jannau> the device is public for 2 1/2 months, anything before that is useless to use. we can't add support based on rumors
<jannau> just because you don't see code changes means there are none
<Ry_Darcy> @jannau I am not partial to rumours. Sad nevertheless.
<psykose> i'll give you a tenner for it
<jannau> the devicetree for the m2 mini is already submitted for upstream and except for display, wlan and bt everything works
<Ry_Darcy> Fair enough, don't forget the house, life savings and the dog too.
<Ry_Darcy> @jannau I could live without BT and WLAN, but the display is the killer obviously. I have read the discussions...
<ChaosPrincess> they got a new wifi/bt chip due to 6e, didnt they?
<jannau> ChaosPrincess: yes, hopefully not much work to support (at least without 6GHz support)
deteg1337 has joined #asahi-dev
ahuar88_ has joined #asahi-dev
ahuar88_ has quit []
ahuar88_ has joined #asahi-dev
ahuar88_ has quit [Remote host closed the connection]
A_L_I_C_E has quit [Remote host closed the connection]
A_L_I_C_E has joined #asahi-dev
ahuar88_ has joined #asahi-dev
ahuar88_ has quit []
deteg1337 has quit [Remote host closed the connection]
<jannau> marcan: are you looking into dcp? I have hackish ipc changes for 12.3 / 13.2 FW coexisting
deteg1337 has joined #asahi-dev
deteg1337 has quit [Remote host closed the connection]
Mary6 has quit []
raveling has joined #asahi-dev
Mary has joined #asahi-dev
<jannau> also are the annotations for ActiveRegion correct? "max_x max_y min_w max_w"? any idea how they work?
<povik> richyliu2: re microphones. it's pretty much figured out. there's been one puzzling thing about the LEAP processor that i finally cracked yesterday
<povik> other than that there are holes in understanding of registers for the LEAP<->PDM glue but that does not block finishing the linux driver at this point
i509vcb has quit [Quit: Connection closed for inactivity]
Thsf has joined #asahi-dev
deteg1337 has joined #asahi-dev
Thsf has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
gabuscus_ has joined #asahi-dev
<chadmed> fwiw t6020 only required DTs and secondary cpu cluster info in m1n1 to boot (mostly) correctly
<chadmed> the missing pieces being fuses for pcie and pmgr stuff for the DTs
<marcan> interestingly, it seems there are no fuses for pcie
deteg1337 has quit [Ping timeout: 480 seconds]
c10l has quit [Ping timeout: 480 seconds]
gabuscus__ has joined #asahi-dev
gabuscus has quit [Ping timeout: 480 seconds]
<chadmed> :o
<marcan> I mean the reg entry is there but I don't see it ever getting hit in relation to pcie except one weird sequence where it's writing ECAM stuff and the fuse read is 0 anyway?
<marcan> maybe iBoot initializes this now
<marcan> PCIe has definitely changed to some extent, ADT reg layout is slightly different but things do somewhat line up with the old one. will probably need some tweaking.
<marcan> jannau: once everything else boots yeah
gabuscus_ has quit [Ping timeout: 480 seconds]
<jannau> ok, I'll prepare everything I have. for m1n1 work it's mostly the 13.2 tracing and the dptx implementation for m1n1
c10l has joined #asahi-dev
<marcan> ack
<marcan> jannau: fyi force pushed main for a fixup
<marcan> there's also an apcie-ge thing in the driver, not instantiated in the DT
<marcan> wild guess: Mac Pro PCIe GEneral purpose PCIe slots.
cylm has joined #asahi-dev
<marcan> AppleInterruptControllerV3 is a thing
<marcan> so this all points towards another chip for the Mac Pro, with AICv3 and more PCIe
<jannau> looking macos 13.3? I was wondering if we want use to 13.3.1 for the 2023 devices. I'm sure the dcp firmware interface is very close if not identical
<marcan> I was wondering the same
<jannau> there is t6500 without known use
<marcan> if you want to go for 13.3, sure
<marcan> I'm on 13.2 right now, but I can switch once this is working
<marcan> also given the 2-die stuff in AIC and obviously cropped die shot (lol), I think I'm going to do the die split preemptively even if we don't have the 2-die chip yet
<marcan> might as well save us a big DT shuffle
<jannau> one reason for 13.2 would be that the 2023 devices will probably ship for a while with 13.2. the installer would work without update. otoh most would probably update on first run
<marcan> actually, mine came with 13.0-special
<marcan> so nah
<marcan> had to update it
<marcan> I assume later ones come with 13.2
<jannau> ah, yes, mine came with 13.0-special as well. I forgot that
<marcan> (well, I say mine, this is lina's laptop but I need to bring up the base stuff first :p)
<marcan> anyway, time to write device trees I guess
<marcan> does anyone have any of that done already?
<jannau> the m2 or m2 pro mac mini (or both) came with 13.0-special
<marcan> this m2max 16" also did
<jannau> I think kettenis_ or chadmed have started with devicetrees
<chadmed> kettenis got further than i did, we both gave up on pmgr though
<marcan> heh, ok
<chadmed> one second
<chadmed> these mostly worked
<marcan> apple,t8103-cluster-cpufreq, really?
<marcan> I expect t8112-cluster-cpufreq
<marcan> (though the difference is minimal, affects feedback)
bps has joined #asahi-dev
<marcan> anyway, will base off of that
<jannau> probably not the first priority fix after copying the t600x devicetrees
<jannau> +to
<chadmed> the only ones we werent sure of were dieX and nvme
<chadmed> everything else should be pretty much good
<chadmed> cant remember at what stage i uploaded this gist though and i lost my local branch when i detonated / the other day
<chadmed> and ofc pmgr
<jannau> dieX is just stuff which is on both m1 ultra dies, behaves like the die0 include if included with DIE 0 defined
<chadmed> yeah but it was all the cpufreq/pmgr stuff that we gave up on :P
<chadmed> values in the opp table come from m1n1 experiments with fixed clusters etc
<chadmed> https://github.com/chadmed/m1n1/tree/t6020 some stuff is done here, depends on the t602x chickens PR
<marcan> ooh the pcores have 2 extra pstates over the baseline m2
<marcan> apple really cranking it up
<marcan> (useless until we do the deepsleep thing but hey)
<marcan> and the ecores lose the lowest pstate
<chadmed> im still hoping they end up using noticeably less power than t600x, those two extra ecores should the pcores stay quiet for longer
<marcan> oh interesting, but state 0 is a dummy
deteg1337 has joined #asahi-dev
bps has quit [Ping timeout: 480 seconds]
deteg1337 has quit [Ping timeout: 480 seconds]
<jannau> the 16" m2 max has a high performance mode with a higher max frequency
<jannau> 3.66 GHz vs. 3.5 GHz
raveling has quit [Ping timeout: 480 seconds]
Thsf has joined #asahi-dev
<marcan> interesting, we'll have to drop those in the other DTs then
Thsf has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ChaosPrincess> got the full init sequence for dfr display in m1n1. Now to figure out which of the fifo writes we _actually_ need
ahuar88 has quit [Quit: Leaving...]
deteg1337 has joined #asahi-dev
<marcan> povik: is it possible the mca IRQs are off by one on t600x?
<marcan> I see mca0 on 1111 but we have it as 1112
deteg1337 has quit [Ping timeout: 480 seconds]
<kettenis_> that has my attempt
<kettenis_> and changes to make u-boot work on the m2 pro mini
<kettenis_> basically just a slightly tweaked memory map
<chadmed> did the pmgr and smc end up working? i cant remember
<kettenis_> some of it at least; can't remember either
<kettenis_> but I can check when I'm back home in a couple of hours
<ChaosPrincess> hah, so, there is even a dumb framebuffer mode on dfr display
<chadmed> iirc i got hard resets when trying to poke the smc with the m1n1 experiments but could do memory reads on its addresses without blowing anything up
<kettenis_> I had smc working
<kettenis_> but the rtc stuff wasn't working
<kettenis_> I believe I had nvme working as well, but not 100% sure
<chadmed> i probably should have saved all that stuff before wiping this machine...
<chadmed> https://gist.github.com/chadmed/2101227831c0af272bcfcdb8326a57ea oh i kept a m1n1 log and the pstate latencies at least :P
<marcan> and the pmgr_mini base for t600x is also wrong O_o
<marcan> I wonder how that ever worked, maybe it aliased?
<povik> marcan: yes. we don't use them
<marcan> povik: oh we don't use the IRQs?]
<povik> no, we don't
<marcan> OK, that explains how it works :p
<povik> :p
Thsf has joined #asahi-dev
deteg1337 has joined #asahi-dev
Thsf has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
deteg1337 has quit [Ping timeout: 480 seconds]
Ry_Darcy has quit [Remote host closed the connection]
<marcan> huh, and the ATC fuses also went away
<marcan> also on newer firmware on t8112 I think, at least
<marcan> did they just move all this to iBoot?
<marcan> (great for us!)
Thsf has joined #asahi-dev
Thsf has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Thsf has joined #asahi-dev
raveling has joined #asahi-dev
Thsf has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
deteg1337 has joined #asahi-dev
Thsf has joined #asahi-dev
deteg1337 has quit [Ping timeout: 480 seconds]
Thsf has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Thsf has joined #asahi-dev
<marcan> [ 7.482000] usb 2-1: new SuperSpeed USB device number 2 using xhci-hcd
<marcan> it really does work without fuses
<marcan> sven: rejoice!
Thsf has quit []
<sven> nice :D
ChaosPrincess has quit [Ping timeout: 480 seconds]
i509vcb has joined #asahi-dev
deteg1337 has joined #asahi-dev
<marcan> pushed what I have so far to t602x/bringup and m1n1:main. it's pretty broken still, but at least the pmgr stuff isn't causing instant explosions, SMC and such works. device tree should have ~everything except MTP and the GPU power stuff and spi0 (need to work out pins).
<marcan> m1n1 pcie init works as far as I can tell (ew magic pokes, though some are probably unnecessary)
<marcan> linux however does not
<marcan> but at least it looks like some important-ish registers might have moved moved so linux probably needs changes
<marcan> nvme is also broken, haven't looked at it yet
<marcan> going to sleep now :p
<marcan> also dwc3 superspeed works so there's that positive note :p
<kettenis_> IIRC I had to fiddle with the power domains a bit to get nvme working
<marcan> yeah, I have a feeling
<kettenis_> and by no means i'm implying that I know what i'm doing in that area
deteg1337 has quit [Ping timeout: 480 seconds]
raveling has quit [Ping timeout: 480 seconds]
<marcan> ok, linux is poking the ANE_SYS power domain and this makes *no* sense. either I screwed up something major (but I can't find it) or there is something very wrong in linux genpd
<marcan> [ 7.798483] apple-pmgr-pwrstate 28e080000.power-management:power-controller@260: PS ane_sys: pwrstate = 0xf: 0xf00000f
<marcan> [ 7.802556] apple-pmgr-pwrstate 28e080000.power-management:power-controller@260: PS ane_sys: pwrstate = 0x0: 0xf0000f0
<marcan> even linux thinks it's ane_sys, wat
<kettenis_> I also didn't intend to keep you from getting some sleep ;)
<marcan> yeah but this is weird
<marcan> inb4 core linux genpd bug...
<marcan> ah ok, I screwed up the pmgr ranges and it was child PDs of that trying and failing to probe repeatedly
Thsf has joined #asahi-dev
Thsf has quit []
ChaosPrincess[w] has joined #asahi-dev
<ChaosPrincess[w]> Can someone with a m1 mbp run this from m1n1 python shell and say if colors look correct https://github.com/WhatAmISupposedToPutHere/m1n1/blob/f4517bdaf72cb36ac8999903a5d85aecddd0ed28/proxyclient/experiments/touchbar_rainbow.py
raveling has joined #asahi-dev
<marcan> interestingly it works if m1n1 initializes nvme first
Thsf has joined #asahi-dev
ChaosPrincess has joined #asahi-dev
<marcan> ok, it's just that iBoot leaves it powered and Linux doesn't have the wakeup fallback thing m1n1 does.
<marcan> adding some stuff to m1n1 to shut it down if it's running
<marcan> though we probably should make linux just do it the robust way anyway...
ChaosPrincess[w] has quit [Quit: Page closed]
ntal has joined #asahi-dev
<marcan> ah, but since linux tries to reset it anyway that doesn't work
<marcan> meh, let's just have m1n1 clean up
<marcan> ah, wasn't that hard, fixed it in linux
ntal has quit [Remote host closed the connection]
<marcan> pushed the fix, nvme works
<marcan> pcie is kinda broken in m1n1 too, it SErrors for some reason (but not with the tracer?)
Thsf has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<marcan> fixed the SError stuff, so now just need to fix pcie and add MTP and we're mostly there
<marcan> good night :)
<marcan> jannau: if you have a chance, mind reviewing the DCP stuff? I think m1n1 needs changes too etc.
<marcan> re ActiveRegion I think I pulled that out of binary RE but I have no idea how it works
Thsf has joined #asahi-dev
<marcan> cc axboe_ I think you were waiting for this bringup :p
<marcan> will try to do MTP tomorrow and see if I can work out PCIe, that should get us to non-edge-kernel feature parity
<marcan> then just DCP and GPU
deteg1337 has joined #asahi-dev
<kettenis_> if you get pcie working, I can work on the asmedia usb stuff in u-boot ;)
Thsf has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
deteg1337 has quit [Ping timeout: 480 seconds]
cylm_ has joined #asahi-dev
cylm has quit [Ping timeout: 480 seconds]
deteg1337 has joined #asahi-dev
deteg1337 has quit [Remote host closed the connection]
Thsf has joined #asahi-dev
Thsf has quit []
Thsf has joined #asahi-dev
Thsf has quit [Read error: Connection reset by peer]
___nick___ has quit [Ping timeout: 480 seconds]
deteg1337 has joined #asahi-dev
deteg1337 has quit [Ping timeout: 480 seconds]
A_L_I_C_E has quit [Ping timeout: 480 seconds]
A_L_I_C_E has joined #asahi-dev
<jannau> why apple? "/arm-io/dart-dcp/mapper-dcp" vs. "/arm-io/dart-dcp0/mapper-dcp0" on the m2 pro. will there be a dcp1 on the VR headset, or on the m2 ultra for 10 displays
drubrkletern has joined #asahi-dev
yrlf has quit [Quit: The Lounge - https://thelounge.chat]
yrlf has joined #asahi-dev
A_L_I_C_E has quit [Ping timeout: 480 seconds]
cylm_ has quit [Ping timeout: 480 seconds]
c10l has quit [Quit: Bye o/]
drubrkletern has quit [Remote host closed the connection]
c10l has joined #asahi-dev
deteg1337 has joined #asahi-dev
deteg1337 has quit [Ping timeout: 480 seconds]
deteg1337 has joined #asahi-dev
deteg1337 has quit [Ping timeout: 480 seconds]