#
…

2026-06-03

00:03damas has quit [Read error: Connection timed out]
00:16NeOne has left #asahi-dev [#asahi-dev]
00:22damas has joined #asahi-dev
00:32damas has quit [Ping timeout: 480 seconds]
00:40voidah has joined #asahi-dev
00:50elvishjerricco has quit [Quit: The Lounge - https://thelounge.chat]
00:50elvishjerricco has joined #asahi-dev
00:50voidah has quit [Ping timeout: 480 seconds]
00:53averne_ has quit [Ping timeout: 480 seconds]
00:56averne has joined #asahi-dev
01:15damas has joined #asahi-dev
01:34millefy14 has quit [Ping timeout: 480 seconds]
01:43millefy14 has joined #asahi-dev
01:55vimproved has quit [Remote host closed the connection]
01:57vimproved has joined #asahi-dev
01:59voidah has joined #asahi-dev
02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-dev
02:09voidah has quit [Ping timeout: 480 seconds]
02:13gargantu1 has quit [Ping timeout: 480 seconds]
02:21damas has quit [Ping timeout: 480 seconds]
02:24Pig[m] has joined #asahi-dev
02:31damas has joined #asahi-dev
02:39damas has quit [Ping timeout: 480 seconds]
02:40jnn has joined #asahi-dev
02:44jn has quit [Ping timeout: 480 seconds]
02:49tobhe has joined #asahi-dev
02:56tobhe_ has quit [Ping timeout: 480 seconds]
03:02tjemke779 has joined #asahi-dev
03:08tjemke77 has quit [Ping timeout: 480 seconds]
03:08tjemke779 is now known as tjemke77
03:08marvin24 has joined #asahi-dev
03:10anon_ has quit [Quit: anon_]
03:18voidah has joined #asahi-dev
03:28voidah has quit [Ping timeout: 480 seconds]
03:36anon has joined #asahi-dev
03:41damas has joined #asahi-dev
03:43damas has quit [Remote host closed the connection]
03:50nora_ has joined #asahi-dev
03:52nora has quit [Ping timeout: 480 seconds]
04:00habiby has quit [Ping timeout: 480 seconds]
04:00habiby has joined #asahi-dev
04:25liangqi has joined #asahi-dev
04:26croissant` has quit []
04:27croissant has joined #asahi-dev
05:10TitouanT[m] has joined #asahi-dev
05:23AnirbanBose[m] has joined #asahi-dev
05:28AnirbanBose[m]You: hello I have a m4 air lying around I could help with the m1n1 bootloader stuff i believe
05:28AnirbanBose[m]You: can any provide me the docs and repo where the dev work is going on
05:30AnirbanBose[m]this the right docs i believe
05:48Yushi has quit [Remote host closed the connection]
05:49Yushi has joined #asahi-dev
06:39mildsunrise has quit [Quit: disconnect]
06:39mildsunrise has joined #asahi-dev
06:41nickchani am confused about print_regs
06:42nickchanwhy u64 sp = ((u64)(regs)) + 256;
06:42nickchantrying to find where that come from
06:45Sobek[m]What's the type of regs ? Is it possible, that regs is a pointer to somewhere on the stack, and that this is computing an offset from the start of this buffer to compute the stack pointer ?
06:47Sobek[m](If ARM stack grow down (I'm sadly more of an x86 person), this would be the top of the stack before pushing 32 64-bit registers in a buffer whose address is in regs ?)
06:48nickchanSobek[m]: got it now
06:48nickchanstruct exc_info is 256 bytes long
06:49nickchanso after allocating that space and pushing all the registers then the original sp must be location of exc_info + 256
06:49JamesCalligeros[m]Should probably s/256/sizeof(*regs->exc_info)/ or something
06:49Sobek[m]I agree, I have the same reaction to magic numbers.
06:51Sobek[m]For function calls, ARM uses a link register, and you push it on the stack as part of the function preamble, if your function will be overwriting it (callee saved), right ?
06:51Sobek[m](print_regs on x86 would have to account for the return address pushed on the stack, but not ARM ?)
07:01jannauit's not required to create frame pointers for tail calls. as long as the link register is not disturbed you can still return
07:46voidah has joined #asahi-dev
07:54voidah has quit [Ping timeout: 480 seconds]
08:02asahi has joined #asahi-dev
08:51voidah has joined #asahi-dev
08:58liangqi has quit [Remote host closed the connection]
08:58liangqi has joined #asahi-dev
08:59voidah has quit [Ping timeout: 480 seconds]
09:00chadmed has quit [Read error: Connection reset by peer]
09:00chadmed has joined #asahi-dev
09:05asahi has quit [Quit: asahi]
09:06asahi has joined #asahi-dev
09:06liangqi has quit [Ping timeout: 480 seconds]
09:09liangqi has joined #asahi-dev
09:17asahi1 has joined #asahi-dev
09:17liangqi has quit [Ping timeout: 480 seconds]
09:18liangqi has joined #asahi-dev
09:22asahi has quit [Ping timeout: 480 seconds]
09:22asahi1 is now known as asahi
09:23ralim has quit [Read error: Connection reset by peer]
09:23ralim has joined #asahi-dev
09:27liangqi has quit [Ping timeout: 480 seconds]
09:33liangqi has joined #asahi-dev
09:38voidah has joined #asahi-dev
09:43liangqi has quit [Ping timeout: 480 seconds]
09:46voidah has quit [Ping timeout: 480 seconds]
09:48liangqi has joined #asahi-dev
09:56catred has joined #asahi-dev
09:57liangqi has quit [Ping timeout: 480 seconds]
09:59ddxtanx_ has quit [Remote host closed the connection]
10:05liangqi has joined #asahi-dev
10:07tryhard[m]chaos_princess: any idea why i'm getting this crash? pmp is fully initialized.
10:07liangqi has quit [Remote host closed the connection]
10:08liangqi has joined #asahi-dev
10:11JamesCalligeros[m]What is that blocking work thread doing
10:16tryhard[m]i'm not sure. every probe i was doing already finished before.
10:41gargantua has joined #asahi-dev
10:43voidah has joined #asahi-dev
10:45chadmedis there anything else in the logs? it looks like it got stuck dealing with a power domain
10:50tryhard[m]nope, just another retry with the same crash
10:50tryhard[m]until i just rebooted
10:51voidah has quit [Ping timeout: 480 seconds]
10:54StanfordAppliedCyber[m]i'm at the point where i really need to think about how to upstream the sptm emulator because i'm 100% certain this is going to work and i think is the best option, looking for some maintainer input here
10:55StanfordAppliedCyber[m]basically, my current plan is the following things:
10:55tryhard[m]i also have this PM: genpd: Disabling unused power domains but i doubt its related
10:55StanfordAppliedCyber[m]1. upstream, myself, all the hypervisor stuff that does NOT involve sptm. this is my own code, where i own the copyright, there's some timer stuff that comes to mind
10:56StanfordAppliedCyber[m]2. clean room docs on how sptm works that i'm very careful to ensure don't include tainted info
10:56StanfordAppliedCyber[m]3. a runtime trace from my emulator just describing inputs and outputs
10:58Faelar[m] has joined #asahi-dev
10:59StanfordAppliedCyber[m]yuka has volunteered to do the clean room RE work
11:00StanfordAppliedCyber[m]is this a viable path forward?
11:01StanfordAppliedCyber[m]i say this is the most viable path forward btw because if we take the "just run sptm in a VM, proxy requests" option i think we're going to hit a LOT of issues, none of which are fundamentally unsolvable, but would delay progress by months if not years. it would also make exclaves way harder because sptm is responsible for organizing the shared memory between exclaves and xnu
11:02StanfordAppliedCyber[m]by contrast, on our own custom emulator, we can inspect state and then this becomes a lot easier, and i think supporting exclaves is absolutely something we want to do eventually because apple is going to keep putting more stuff in them and it'll likely be required to fully trace the hardware
11:03StanfordAppliedCyber[m]s/say/think/, s/viable/promising/, s/forward//
11:03StanfordAppliedCyber[m]not to mention that sptm is a very ornery piece of software that is pretty much always unhappy in my experience
11:09StanfordAppliedCyber[m]also in the meantime i'm very interested in the gpu and i assume i'm free to use the hypervsior to start REing the firmware interface? i've never disassembled any of the gpu kexts or firmware and am totally untainted
11:13yukaIs the firmware interface even that different from previous ones?
11:14StanfordAppliedCyber[m]but that's what i want to check :)
11:14StanfordAppliedCyber[m]i'm hoping it's not obviously, but at minimum initdata will need some fixes
11:15chaos_princess1. yes. 2. reach out to noopwafel/shiz
11:16StanfordAppliedCyber[m]this is in reply to my 3 part plan? or to yuka's question
11:17StanfordAppliedCyber[m]oh it's quite different? unfortunate, i was hoping it'd be similar, but this is apple so i knew better than to expect
11:17chaos_princessat least the initdata is materially different from 13.5 to 14.8.3
11:18StanfordAppliedCyber[m]yeah i expected that much, but i'm hoping it's just added/removed some fields, not oh hey we just totally changed the way the whole interface works
11:20chaos_princessadded/removed fields, restructured blobs on which field and which struct goes where, things of that nature.
11:21chadmedthe gpu and display controller have not stayed the same across a single major os release, it is a bit annoying :P
11:22chadmed"we moved the fb swap submit callback from d408 to d409 because ummm well ahhh you see mmmm"
11:23amarioguy has joined #asahi-dev
11:30J has joined #asahi-dev
11:31J is now known as Guest10713
11:32Jijdejiiedjjiedijede has joined #asahi-dev
11:41Guest10713 has quit [Remote host closed the connection]
11:43pb17 has quit [Ping timeout: 480 seconds]
11:49Sobek[m]The rust macro magic that Lina developped for this sort of versionning can still be used tp help with that issue, right ?
11:50chadmedyeah, it just has to be reversed and actually expressed
11:51chadmedsimilarly the dcp driver has a bunch of insane^H^H cool macros and such to deal with the divergent fw abis
11:52Jijdejiiedjjiedijede has quit [Remote host closed the connection]
12:03chadmed has quit [Quit: Konversation terminated!]
12:03chadmed has joined #asahi-dev
12:05pb17 has joined #asahi-dev
12:10voidah has joined #asahi-dev
12:17chadmedchaos_princess: if/when you rebase silence on 7.1 make sure to use the new macaudio and i2s bus keeper settings. the amps will ignore the current ones and you will get interesting results.
12:20voidah has quit [Ping timeout: 480 seconds]
12:22voidah has joined #asahi-dev
12:31voidah has quit [Ping timeout: 480 seconds]
12:50xdbob has quit [Remote host closed the connection]
12:52xdbob has joined #asahi-dev
13:22pb17 has quit [Ping timeout: 480 seconds]
13:24fl0_id has quit [Ping timeout: 480 seconds]
13:26pb17 has joined #asahi-dev
13:32fl0_id has joined #asahi-dev
13:41svenStanfordAppliedCyber[m]: we know how all the iommus work anyway, no need for traces there
13:42svenjust describe that domain N function 12 is dart_tWHATEVER_map page and we can just use the dart code inside m1n1 already
13:42svensame for the NVMMU and SART
13:42Marvin[m]1 has left #asahi-dev [User left]
13:43svensomeone also got sptm on m2 to run using nested virt fwiw, but that obviously doesn't translate 1:1 to m4+ because that one relies on GXF and SPRR
13:43svenotherwise that approach sounds good
13:46nela8 has joined #asahi-dev
13:48nela has quit [Ping timeout: 480 seconds]
13:48nela8 is now known as nela
14:29gargantua has quit [Ping timeout: 480 seconds]
14:33zkrx has quit []
14:35zkrx has joined #asahi-dev
14:57gargantua has joined #asahi-dev
15:19voidah has joined #asahi-dev
15:58asahi1 has joined #asahi-dev
16:04asahi has quit [Ping timeout: 480 seconds]
16:04asahi has joined #asahi-dev
16:09asahi1 has quit [Ping timeout: 480 seconds]
16:23asahi has quit [Read error: Connection reset by peer]
16:44voidah has quit [Ping timeout: 480 seconds]
16:50chaos_princessaop init seems to just work(tm), same for als/las, but there seem to be material changes in how the pdm decimator is setup
17:04Amid[m] has joined #asahi-dev
17:08voidah has joined #asahi-dev
17:16voidah has quit [Ping timeout: 480 seconds]
17:19michaelros444 has joined #asahi-dev
17:25jmax has quit [Quit: Q]
17:26michaelros444 has quit [Quit: The Lounge - https://thelounge.chat]
17:27voidah has joined #asahi-dev
17:27liangqi has quit [Remote host closed the connection]
17:27michaelros444 has joined #asahi-dev
17:27liangqi has joined #asahi-dev
17:29michaelros444 has quit [Read error: Connection reset by peer]
17:29michaelros444 has joined #asahi-dev
17:36liangqi has quit [Ping timeout: 480 seconds]
17:41liangqi has joined #asahi-dev
17:47tryhard[m]i suspect something wrong maybe with dt?
18:00tryhard[m]ok i think i fixed it
18:03ddxtanx has joined #asahi-dev
18:09gargantua has quit [Ping timeout: 480 seconds]
18:11liangqi has quit [Remote host closed the connection]
18:11liangqi has joined #asahi-dev
18:11michaelros444 has quit [Quit: The Lounge - https://thelounge.chat]
18:12michaelros444 has joined #asahi-dev
18:18jannauyuka: something went wrong in the cpufreq PR: "yuyuyureka wants to merge 3290 commits into AsahiLinux:bits/001-devicetree-m3 from yuyuyureka:t6031-cpufreq"
18:19liangqi has quit [Ping timeout: 480 seconds]
18:20yukaAh well
18:21yukanow better?
18:21jannaudoes the benchmark gives you the expected frequency for higher p-states starting around 3.7 GHz
18:22pjv has joined #asahi-dev
18:22jannaulooks good now
18:27yukajannau: only if I switch the boot core to a core in another cluster, see the other commits in https://github.com/yuyuyureka/m1n1/commits/pstate-latencies-boot-cpu/
18:33voidah has quit [Ping timeout: 480 seconds]
18:39jannauon the max you could test a core on the second perf cluster. I suspect switching to an efficiency core at the beginning will be sufficient. that probably explains why we did not see it on m1* and m2*
18:40yukaI'm also seeing it on A18 Pro and M4 (even with switching to an e-core), probably because they don't sleep properly
18:41voidah has joined #asahi-dev
18:50michaelros4444 has joined #asahi-dev
18:51michaelros444 has quit [Ping timeout: 480 seconds]
18:51michaelros4444 has quit [Read error: Connection reset by peer]
18:51michaelros444 has joined #asahi-dev
18:57michaelros4444 has joined #asahi-dev
19:00pjv has quit [Quit: pjv]
19:02michaelros444 has quit [Read error: Connection reset by peer]
19:03michaelros44445 has joined #asahi-dev
19:04tryhard[m]is there direct way to make linux change devices power state only after pmp is fully configured? or i need to do something manually
19:08gargantua has joined #asahi-dev
19:10michaelros4444 has quit [Ping timeout: 480 seconds]
19:10michaelros44445 is now known as michaelros4444
19:10liangqi has joined #asahi-dev
19:13pjv has joined #asahi-dev
19:17hardy94 has joined #asahi-dev
19:18michaelros4444 has quit [Ping timeout: 480 seconds]
19:19liangqi has quit [Ping timeout: 480 seconds]
19:21chaos_princessnot how pmp works, the report pm domains should ack power state changes immediately while pmp is booting, and only after it finishes - you start waiting for it to accet the change
19:22chaos_princesslook at what v2 is doing, and do same, but over rtkit
19:25commandoline has quit [Quit: Bye!]
19:26commandoline has joined #asahi-dev
19:27liangqi has joined #asahi-dev
19:29lukeleppan911 has joined #asahi-dev
19:30tryhard[m]how do they know pmp is booting?
19:31tryhard[m]i thought that's what causes my new crash executing from non-executable memory after gpu init
19:32commandoline has quit [Quit: Bye!]
19:32lukeleppan911 has quit []
19:33commandoline has joined #asahi-dev
19:34lukeleppan91 has quit [Ping timeout: 480 seconds]
19:36liangqi has quit [Ping timeout: 480 seconds]
19:38liangqi has joined #asahi-dev
19:40michaelros44445 has joined #asahi-dev
19:41michaelros444454 has joined #asahi-dev
19:42yukaBecause of the cpufreq things, I started looking into sleep behavior on M4 again
19:43yukaReading xnu code, it should configure sleep through ACC_OVRD like on other devices
19:43yukabut we are not allowed to do that :(
19:46liangqi has quit [Ping timeout: 480 seconds]
19:47lukeleppan91 has joined #asahi-dev
19:48michaelros444454 has quit [Remote host closed the connection]
19:48michaelros44445 has quit [Ping timeout: 480 seconds]
19:55liangqi has joined #asahi-dev
19:57regis has joined #asahi-dev
20:00regis has quit []
20:00regis has joined #asahi-dev
20:00regis has quit []
20:04lukeleppan91 has quit [Quit: Ping timeout (120 seconds)]
20:07lukeleppan91 has joined #asahi-dev
20:09liangqi has quit [Ping timeout: 480 seconds]
20:13chaos_princesskinda regretting putting pdm coefficients in driver, m3 has 2046 instead of 120
20:18liangqi has joined #asahi-dev
20:20jannaunot too late to change that
20:26liangqi has quit [Ping timeout: 480 seconds]
20:31liangqi has joined #asahi-dev
20:35pjv has quit [Quit: pjv]
20:35alessaba has quit [Ping timeout: 480 seconds]
20:35alessaba has joined #asahi-dev
20:40liangqi has quit [Ping timeout: 480 seconds]
20:44liangqi has joined #asahi-dev
20:56cy8aerHi there, I somehow have problems with the bluetooth stack and asahi-7.0.11-1. I connect some Linkbuds S via Bluetooth but no sounds. The only thing I can see there are some wireplumber "firmware bug" messages. I disconnected the phones and tried to reconnect but this was not possible anymore.
20:56cy8aerI do not see anything in dmesg, but I can offer a journalctl of the session: https://nc.g3la.de/index.php/s/6r4Le3kFzCdLxgT
20:58cy8aerDisabling bluetooth and enabling again does crash the system sometimes (!). I do not have any of these efects with asahi-7.0.10-2 or before.
20:59cy8aerThese wireplumber messages are around 22:43:14
21:01StanfordAppliedCyber[m]sven: thanks for the advice here!
21:01StanfordAppliedCyber[m]re: traces, the thing that concerns me is that there's a lot of edge case behavior that in some cases took me days/weeks to get to the bottom of because the failure symptoms were so far removed from the actual root cause
21:02StanfordAppliedCyber[m]there's a reason i'm not using the m1n1 code as it is currently
21:02jannaucy8aer: I switched brcmext patch from v2 to v3. can you try to revert 30fcc498ff7c and apply f5cd0acdc21d
21:03jannaubluetooth audio issues confirmed on 7.0.11, bt trackball works though
21:03StanfordAppliedCyber[m]this may be a theoretical barrier, maybe yuka or others are just better devs than me and can get it working just with the high level overview, but my expectation is that with a runtime trace resolving some of these edge cases will be extremely difficult if not impossible
21:06StanfordAppliedCyber[m]not to mention i figured the risk is extremely low-- even if the traces end up just confirming what we know about dart, it's functionally just the same as looking at an m1n1 hv trace and doesn't taint the people working on it
21:10cy8aerjannau: Compiling, stay tuned!
21:10voidah has quit [Ping timeout: 480 seconds]
21:15liangqi has quit [Ping timeout: 480 seconds]
21:15StanfordAppliedCyber[m]also wanted to ask about a specific edge case-- we have to pass a big struct to xnu containing information about sptm, memory mappings, a bunch of other stuff
21:16StanfordAppliedCyber[m]this struct is contained in the KDK and APSL2 licensed
21:16StanfordAppliedCyber[m]the fields of the struct are all in the public xnu source but the offsets need to be found either by REing apple's compiled kernel or from the kdk (how i found it)
21:17nicenickname6161 has joined #asahi-dev
21:17liangqi has joined #asahi-dev
21:19StanfordAppliedCyber[m]from looking at the code, I think this is safe to include in clean room docs, but i wanted to ask your thoughts here to confirm, and if you think it's overly risky then to ask for alternative paths forward
21:19chaos_princessyou write your own struct based on the apsl source
21:21cy8aerjannau: reverting 30fcc498ff7c and applying f5cd0acdc21d works for me. Complete different behavior: the phones connect by themselves and sound.
21:22StanfordAppliedCyber[m]"write your own struct based on apsl source", so there's a few things there, there's the data types which i can just use primitive types for instead of the apple named ones, but the names themselves mostly describe hardware facts
21:23StanfordAppliedCyber[m]i'll move over to the re channel as a precaution here though i think it's unnecessary
21:26liangqi has quit [Ping timeout: 480 seconds]
21:31nicenickname6161 has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
21:38liangqi has joined #asahi-dev
21:39yukaI can measure the first pcore (boot core) of the Neo at the highest expected clock
21:39yukaBut not the second pcore
21:40yukapresumably because the boot core is configure not to go to deep sleep
21:41yukaI think smp call is somehow broken after switch_boot_core
21:42yukaSo I can not measure latencies of the boot core after switching away from it
21:42StanfordAppliedCyber[m]what do you mean "measure the first pcore"?
21:42StanfordAppliedCyber[m]like check the clockspeed?
21:43jannauit's possible that there is still code checking cpuX.state == "running" to determine the boot cpu
21:43nicenickname6161 has joined #asahi-dev
21:44jannauit wasn't intended to switch the cpu core and keep using m1n1 but only for kboot
21:45voidah has joined #asahi-dev
21:45yukaOhhh right
21:47liangqi has quit [Ping timeout: 480 seconds]
21:50StanfordAppliedCyber[m]also hows chainload on the neo working? does it work yet?
21:50StanfordAppliedCyber[m]or is it still messed up
21:53hardy94 has quit [Remote host closed the connection]
21:53voidah has quit [Ping timeout: 480 seconds]
21:56nicenickname6161 has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
22:01yukaWorks with kisd
22:03jannauand the recently "fixed" chainload.py
22:11liangqi has joined #asahi-dev
22:14Lyude has quit [Read error: Connection reset by peer]
22:14Lyude has joined #asahi-dev
22:30liangqi has quit [Remote host closed the connection]
22:35ddxtanx has quit [Remote host closed the connection]
22:37voidah has joined #asahi-dev
22:37ddxtanx has joined #asahi-dev
22:38liangqi has joined #asahi-dev
22:46liangqi has quit [Ping timeout: 480 seconds]
22:47voidah has quit [Ping timeout: 480 seconds]
22:56liangqi has joined #asahi-dev
23:03voidah has joined #asahi-dev
23:04liangqi has quit [Ping timeout: 480 seconds]
23:12voidah has quit [Ping timeout: 480 seconds]
23:12hardy94 has joined #asahi-dev
23:13liangqi has joined #asahi-dev
23:18tuxq has joined #asahi-dev
23:19tuxq has left #asahi-dev [Leaving]
23:28liangqi has quit [Ping timeout: 480 seconds]
23:37hardy94 has quit [Quit: hardy94]
23:38liangqi has joined #asahi-dev
23:46liangqi has quit [Ping timeout: 480 seconds]
23:54voidah has joined #asahi-dev
23:55liangqi has joined #asahi-dev
…