#
…

2026-07-08

02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-dev
02:03lacamar_ has joined #asahi-dev
02:04lacamar- has quit [Ping timeout: 480 seconds]
02:09truita_ has joined #asahi-dev
02:14truita has quit [Ping timeout: 480 seconds]
02:17truita_ is now known as truita
03:02tjemke773 has joined #asahi-dev
03:08tjemke77 has quit [Ping timeout: 480 seconds]
03:08tjemke773 is now known as tjemke77
03:18user has joined #asahi-dev
03:18user is now known as Guest12880
03:29Guest12880 has quit [Quit: WeeChat 4.7.1]
03:29user2 has joined #asahi-dev
03:31user2 has quit []
03:31user2 has joined #asahi-dev
03:54nora_ has joined #asahi-dev
03:56nora has quit [Ping timeout: 480 seconds]
04:20notwasm[m] has joined #asahi-dev
04:39mattia013 has joined #asahi-dev
04:56Retr0id has quit [Ping timeout: 480 seconds]
04:59Retr0id has joined #asahi-dev
05:12Retr0id has quit [Ping timeout: 480 seconds]
05:14Retr0id has joined #asahi-dev
05:19rosalyntg has quit [Read error: Connection reset by peer]
05:20rosalyntg has joined #asahi-dev
05:49mattia013 has quit [Remote host closed the connection]
06:01Pokemod977362084712818330 has quit []
06:08Pokemod977362084712818330 has joined #asahi-dev
06:22mripard has joined #asahi-dev
06:50lowspoly[m] has left #asahi-dev [User left]
06:57hdbngr has quit [Remote host closed the connection]
06:57hdbngr has joined #asahi-dev
07:07alessaba has quit [Ping timeout: 480 seconds]
07:08alessaba has joined #asahi-dev
07:16hdbngr has quit [Ping timeout: 480 seconds]
07:19elvishjerricco has quit [Ping timeout: 480 seconds]
07:24hdbngr has joined #asahi-dev
07:29elvishjerricco has joined #asahi-dev
07:31jannausven: do you already have an idea how to support it with another bootloader in-between? letting it proxy the calls? would have made sense to cc: u-boot on the RFC
07:31svenit "just" needs to pass along the runtime memory descriptors and that additional table
07:32hdbngr has quit [Ping timeout: 480 seconds]
07:36hdbngr has joined #asahi-dev
07:47yuka27.0b3 on t6050: SYS_IMP_APL_VM_TMR_FIQ_ENA_EL2 is unlocked, smp is broken
07:47yukaprogress, yay
07:53hardy94 has joined #asahi-dev
07:56StanfordAppliedCyber[m]smp is broken as in you do the cpustart write and nothing happens?
08:01hdbngr has quit [Ping timeout: 480 seconds]
08:06midou has joined #asahi-dev
08:13pb17 has quit [Ping timeout: 480 seconds]
08:13yukathe cores start, but never enter our code
08:14yuka"start" is judging from pmgr regs
08:14pb17 has joined #asahi-dev
08:15StanfordAppliedCyber[m]yeah same symptom as my macbook neo
08:15StanfordAppliedCyber[m]pmgr says they started but they don't execute our code
08:22picnoir has quit [Ping timeout: 480 seconds]
08:30picnoir has joined #asahi-dev
08:30yuka27.0b3 on t8132 is unchanged to 27.0b2: SYS_IMP_APL_VM_TMR_FIQ_ENA_EL2 is locked and iboot regularly panics
08:46StanfordAppliedCyber[m]with yuka's help i have successfully gotten debugusb to work on the neo
08:46StanfordAppliedCyber[m]i had to set nodisplay
08:47sveni get that they had to disable GXF/SPRR and whatever but did they really have to also break all kinds of other random crap and just not test this mode? :(
08:47StanfordAppliedCyber[m]apparently otherwise the dcp causes a lot of problems for reasons we don't quite understand, but also do not care about
08:47StanfordAppliedCyber[m]knock wood i can get a hypervisor PR out in a few days
08:48StanfordAppliedCyber[m]i'm about 70% of the way writing cleanroom docs on my emulator already
08:48StanfordAppliedCyber[m]the hypervsior pr will need a lot of massaging before it can be upstreamed, but that's a problem we can work through and come with with solutions that everyone can be happy about
08:48nickchanapple has not much incentive to test the mode unless they face legal pressure of "this is not really a computer" and had to be taxed differently
08:48StanfordAppliedCyber[m]anyways time to sleep at a reasonable hour for once
08:48JamesCalligeros[m]sven: why spend good money on engineers when ~80,000 non-revenue-generating mac owners will be forced to test it for you for free :3
08:49svenyeah :(
08:49JamesCalligeros[m]Something like that yeah
08:49svenif we could at least file bugs easily
08:49nickchan(computers that could run custom operation systems are taxed differently in some)
08:49svencitation required
08:54nickchansometimes i thought about making the fact that this mode is ho
08:54nickchanrribly broken more visible like on social media to attract scrutiny about these "computers" but then at the same time I feel like that could pick some fights that one does not want to deal with
08:55runxiyu_ is now known as runxiyu
08:56yuka> stats
08:56yukalol, the two users with a Mac Pro
08:56yukaone of them was jannau testing it
08:56yukaI'm sure of it
08:57nickchani wonder who is the other one
08:57nickchanabout the revenue now I still use macOS a lot but I think my way of using it doesn't generate much revenue for apple
08:58JamesCalligeros[m]I have an iphone and will not buy anything else but i dont pay for any apple services other than apple music
08:59JamesCalligeros[m]And because australian unis let you keep your email address basically forever, i can still verify my student status every year and get the student plan for that (and 10% off hardware at the apple store)
08:59nickchanfraud ! ! !
08:59JamesCalligeros[m]As long as it’s costing american companies money shrug
09:00JamesCalligeros[m]I’m doing my part!
09:00jannauit's possible that both reported mac pro installs are from me. I usually don't report test installs though
09:00nickchanthere is about 5 times more Linux on iPhone 8 users than mac pro asahi linux users then
09:01nickchanI can't tell for sure i don't have the stats
09:01jannauI doubt apple has less than 10% margin on anything
09:01JamesCalligeros[m]Opportunity cost then :p
09:01svenso 10 iphone 8 users then? :D
09:02nickchani say there's about 5
09:02nickchanif you include every a11 user there's maybe 10
09:03nickchani can already read from the nand from m1n1 python on a8x so a7/a8/a8x storage support would be coming eventually
09:04nickchaniboot traces makes it so much simpler
09:05hdbngr has joined #asahi-dev
10:15pb17 has quit [Ping timeout: 480 seconds]
10:18pb17 has joined #asahi-dev
10:19hdbngr has quit [Read error: Connection reset by peer]
10:38rosalyntg has quit [Remote host closed the connection]
10:39rosalyntg has joined #asahi-dev
11:13hdbngr has joined #asahi-dev
11:36hardy94 has quit [Read error: Connection reset by peer]
11:44kettenisjannau: yes, I was that in my mail
11:45kettenissven: the code in efi.rs is enough to make the Linux kernel EFI stub happy?
11:45svenyes, it's the bare minimum to make it happy
11:45svenpretty sure it violates the spec though
11:46kettenisno kidding ;)
11:49sveni expect that to be only useful for directly chainloading to the kernel fwiw, u-boot (or whatever) probably needs a different protocol to essentially pass along that new table i made up + the runtime mappings we need
11:50kettenisI'm amazed that you get away with not implenting any boottime services
11:50svenyup, same :D
11:50ketteniswould have expect you'd need ExitBootServices() at the very minimum
11:50sveni expected that to crash
11:51svenit works because i set runtime_services_supported to "nothing" and at least linux respects that and doesn't try to call into any efi services
11:52kettenisanyway, yes, you'll need to pass something in the device tree that we can parse in u-boot
11:53svenit's all a bit silly because we could also just add the memory map and a pointer to the psci handler in the FDT and just skip all the EFI stuff but oh well
11:57kettenisthe memory regions can probably be passed along as reserved-memory entries with some additional property to indicate that they're EFI runtime things
11:58svenyeah, i also plan to restrict the MMIO mappings to what we actually need instead of just mapping everything
11:59svenother than MMIO it's just m1n1 .data/.rodata/.bss and .text anyway
12:34lord_breadcrumb has joined #asahi-dev
12:35lord_breadcrumb has quit [Remote host closed the connection]
12:56runxiyu has quit [Remote host closed the connection]
13:01runxiyu has joined #asahi-dev
13:17hdbngr has quit [Ping timeout: 480 seconds]
13:26barusu has quit [Ping timeout: 480 seconds]
13:40barusu has joined #asahi-dev
13:42nickchannot really a fan of making the secondary stack a static allocation
13:43nickchanwith m5 ultra 36 cpus it would make the m1n1.bin over 2 MiB of just stacks
13:44Noahbds has joined #asahi-dev
13:44svenso? these machines have GiB or RAM
13:44nickchanyes but the static allocation is everywhere
13:44nickchanas the size of the .bin
13:45Sobek[m](And M5 Ultra have a higher minimum RAM compared to smaller chips ?)
13:45nickchanand im dealing with devices that may have only less than 512 MiB of usable dram
13:45Sobek[m](Would it make sense to have different m1n1 bin depending on the HW ?)
13:45Sobek[m]nickchan: ? Macs ?
13:46nickchanSobek[m]: doubt it it would go to the point of needing different bins since the m1n1 size is not optimized at all now
13:47nickchanbut then it's still okay, though with psci-via-efi branch the size does become *slightly* concerning but probably won't cause immediate issues
13:47nickchanwhat i am saying is there's tons of rom for optimization and splitting bin is not going to be required
13:49fl0_id_afkSobek[m] not macs, iphones
13:57Noahbds has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
13:59NickChan[m]maybe top_of_memory_alloc() can be used?
14:01nickchansince the memory needs to be reserved anyway
14:02nickchanand then the secondary stack size can still be 16K i think
14:06nickchanonly for the secondary though since this would leak memory if another m1n1 is chainloaded after secondary has been started
14:06svenmaybe we figure this out later
14:20nickchannp it wouldn't cause immediate problems so i can make it less messed up later
14:21nickchanone optimization im thinking is moving the main stack to just after top_of_kdata and using a seperate variable to tell the allocator about this fact
14:22nickchanthe boot-args makes it easy to do in asm
14:22hardy94 has joined #asahi-dev
14:25nickchan- has joined #asahi-dev
14:25nickchan has quit [Read error: Connection reset by peer]
14:27nickchan- has quit []
14:27nickchan has joined #asahi-dev
14:27svenyeah, i first want to get the psci-via-efi or whatever thing ready and then we can always optimize before the next m1n1 release
14:27svenright now my raw binary is about 3 MiB with that branch
14:27svenwhich is still fine for now imho
14:28nickchan though moving the main EL3 stack would require an appropriate reservation in FDT on the devices that need it
14:29nickchanthese two stacks accounts for 75% of the difference between the .macho and the .bin which is caused by whether .bss is included
14:30nickchanon the main branch
14:36nickchancould probably get rid of dummy_stack entirely as well, it should never be used and in practice it also doesn't work because it's not 16 byte aligned
14:40nickchanand reduce that main el3 stack to 16k since its main use is for the exception code which is the same code used by the secondaries that had 16K stack already in experiments like find_regs
14:51hardy94 has quit [Remote host closed the connection]
14:59pb17 has quit [Ping timeout: 480 seconds]
15:01hdbngr has joined #asahi-dev
15:12mattia013 has joined #asahi-dev
15:17pb17 has joined #asahi-dev
15:18hdbngr has quit [Ping timeout: 480 seconds]
15:40svenlooks like xnu doesn’t lie and the ovrd register are indeed per cluster
15:41sventhat makes shutting individual CPUs off a bit annoying
15:47nickchanCPU_OVRD/CYC_OVRD should be per core
15:47chaos_princessthe relevant bits are in ACC_OVRD
15:49nickchanright but wouldn't those be for shutting down the cluster
15:50nickchanI suppose it could be a problem when shutting down the last core in a cluster
15:51sventhen tell me how I can power off a cpu without acc_ovrd
15:51nickchancpu_sleep(0) in m1n1
15:52chaos_princessthat is not what we are talking about
15:53nickchan"going into deep sleep" in smp_stop_cpu() of m1n1 does sound like powering down the cluster
15:54nickchanand if deep_sleep is false there then the code does not touch acc_ovrd
16:13Guest12860 has quit [Remote host closed the connection]
16:19nickchanhonestly though cpu sleep is magic
16:19nickchana11 has been working by luck this whole time :)
16:20nickchanon a11 the pcore and ecore is in same cluster (?) and there's a ACC_EBLK_OVRD that is should be but never set in m1n1
16:37nickchan has quit [Quit: ZNC 1.9.1+deb2+b3 - https://znc.in]
16:37nickchan has joined #asahi-dev
17:18nickchani don't understand why the secondary stack must be accessible in the EFI PSCI context
17:19nickchanput it as reserved memory no map; then
17:19nickchanfor CPU_ON the secondary is outside the kernel with mmu off so from that core's perspective the stack is perfectly accessible
17:19iusearchbtw has joined #asahi-dev
17:20nickchanfor CPU_OFF the secondary stack would not even be used
17:20WindowPain5 has joined #asahi-dev
17:20nickchanand for CPU_SUSPEND either state is kept like with retention deep wfi so kernel stack and all, or lose state then it's like CPU_ON
17:21sven🤷‍♂️
17:21svenstill not gonna change anything for now
17:22svenI truly do not care about making anything more complicated right now to save 2MiB or so
17:23iusearchbtwjannau should know how to and its 14.8.3 not 16.8.3 because macos 16 did not exist its only 13 14 15 and 26
17:24chaos_princessjust run test.sh
17:25iusearchbtwchaos_princess i want to test it on j613
17:25iusearchbtwchaos_princess error[E0463]: can't find crate for `core` | = note: the `aarch64-unknown-none-softfloat` target may not be installed = help: consider downloading the target with `rustup target add aarch64-unknown-none-softfloat` = help: consider building the standard library from source with `cargo build -Zbuild-std` For more information about this error, try `rustc --explain E0463`. error: could not compile `bitflags` (lib) due to 1 previo
17:25nickchanno operating system image support these machines
17:26iusearchbtwnickchan who in the kaboom community has a m1
17:26chaos_princessiusearchbtw: i really like rustc, it is great at telling you what exactly went wrong, and how to fix it
17:26chaos_princessgreat piece of software
17:26WindowPain has quit [Ping timeout: 480 seconds]
17:26WindowPain5 is now known as WindowPain
17:27iusearchbtwchaos_princess A crate was declared but cannot be found. Erroneous code example: extern crate foo; // error: can't find crate You need to link your code to the relevant crate in order to be able to use it (through Cargo or the -L option of rustc, for example). Common causes * The crate is not present at all. If using Cargo, add it to [dependencies] in Cargo.toml. * The crate is present, but under a different name. If using Cargo, look for p
17:27nickchani am typing this from a macbook pro 2021 14 inch with m1 pro
17:27iusearchbtwnickchan fastfetch it
17:27chaos_princessiusearchbtw: did you just download the zip and clone the repo
17:27iusearchbtwclone the repo
17:27chaos_princessclone the submodules
17:28chaos_princessalso on a j314 here :P
17:29iusearchbtwchaos_princess compile the installer from source from https://github.com/jannau/asahi-installer/tree/m3-macos-16.8.3-ota-install and fix all of the errors and put it on a website
17:29iusearchbtwchaos_princess there is alot of errors when building it
17:30chaos_princesssorry, was the first part an order in my direction?
17:30iusearchbtwchaos_princess compile the installer from source from https://github.com/jannau/asahi-installer/tree/m3-macos-16.8.3-ota-install and fix all of the errors
17:33iusearchbtwchaos_princess what linux is on the j314
17:38iusearchbtw was kicked from #asahi-dev by chaos_princess [consider not ordering people around]
17:53chaos_princess has joined #asahi-dev
18:23XeR has quit [Remote host closed the connection]
18:24XeR has joined #asahi-dev
18:25XeR has quit [Remote host closed the connection]
18:28XeR has joined #asahi-dev
18:47pjv has joined #asahi-dev
18:55jannausigh, what happened here.
18:55jannauI make interesting typos though
18:57dyld has joined #asahi-dev
19:07Faelar[m]<insert y u no fix all errors meme here> 🙄
19:07Faelar[m]Quick support message here, thank you for all for what you do and sorry you have to deal with people like this.
19:23dyld has quit [Remote host closed the connection]
19:38hdbngr has joined #asahi-dev
19:38delinka has joined #asahi-dev
19:46yukaI decided I want to have a working dockchannel uart driver before trying to make it work with the mailbox
19:46yukagot tx mostly working including multiple channels, console support and earlycon
19:53FLHerne_ has joined #asahi-dev
19:54FLHerne has quit [Remote host closed the connection]
19:58pjv has quit [Remote host closed the connection]
19:58pjv has joined #asahi-dev
20:29mischa85 has quit [Quit: Lost terminal]
20:33pjv has quit [Ping timeout: 480 seconds]
20:39mattia013 has quit [Ping timeout: 480 seconds]
20:56cyberhead_ has joined #asahi-dev
20:59cyberhead has quit [Ping timeout: 480 seconds]
21:13hdbngr has quit [Ping timeout: 480 seconds]
21:13cy8aer(sorry wc)
21:39StanfordAppliedCyber[m]the standard hv dev process is chainload a new m1n1 image then run the default entrypoint and send the kernelcache over the wire right?
21:59hdbngr has joined #asahi-dev
22:02hardy94 has joined #asahi-dev
22:07hdbngr has quit [Ping timeout: 480 seconds]
22:32pb17 has quit [Ping timeout: 480 seconds]
22:32hardy94 has quit [Quit: hardy94]
22:51pb17 has joined #asahi-dev
22:59hdbngr has joined #asahi-dev
23:03nela has quit [Ping timeout: 480 seconds]
23:06nela has joined #asahi-dev
23:08hdbngr has quit [Ping timeout: 480 seconds]
23:42elvishjerricco8 has joined #asahi-dev
23:48elvishjerricco has quit [Ping timeout: 480 seconds]
23:48elvishjerricco8 is now known as elvishjerricco
23:49StanfordAppliedCyber[m]currently refactoring my entire HV (minus the sptm emulator) to be upstreamable with the standard utilities
23:52delinka has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
…