2026-07-31
00:00linuxgemini has quit [Quit: Ping timeout (120 seconds)]
00:00linuxgemini has joined #asahi-dev
00:00hdbngr has quit [Ping timeout: 480 seconds]
00:07Calandracas_ has joined #asahi-dev
00:14Calandracas__ has quit [Ping timeout: 480 seconds]
00:27nepeat has quit [Ping timeout: 480 seconds]
00:36nepeat has joined #asahi-dev
00:45picnoir has quit [Ping timeout: 480 seconds]
00:53picnoir has joined #asahi-dev
01:50hdbngr has joined #asahi-dev
01:59hdbngr has quit [Ping timeout: 480 seconds]
02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-dev
02:00ciara has quit [Remote host closed the connection]
02:12ciara has joined #asahi-dev
02:27nnym has quit [Ping timeout: 480 seconds]
02:30nnym has joined #asahi-dev
02:43voidah has joined #asahi-dev
02:43hdbngr has joined #asahi-dev
02:51hdbngr has quit [Ping timeout: 480 seconds]
03:11nora has joined #asahi-dev
03:13nora_ has quit [Ping timeout: 480 seconds]
03:17hdbngr has joined #asahi-dev
03:26hdbngr has quit [Ping timeout: 480 seconds]
03:30hdbngr has joined #asahi-dev
03:35voidah has quit [Ping timeout: 480 seconds]
03:39hdbngr has quit [Ping timeout: 480 seconds]
03:39NekoOvO[m] has joined #asahi-dev
03:45voidah has joined #asahi-dev
03:53voidah has quit [Ping timeout: 480 seconds]
04:22hdbngr has joined #asahi-dev
04:32Ariadne has quit [Ping timeout: 480 seconds]
04:34hdbngr has quit [Ping timeout: 480 seconds]
04:44nnym has quit [Remote host closed the connection]
04:47Yushi has joined #asahi-dev
04:52nnym has joined #asahi-dev
04:54hdbngr has joined #asahi-dev
05:08hdbngr has quit [Ping timeout: 480 seconds]
05:31hdbngr has joined #asahi-dev
05:40hdbngr has quit [Ping timeout: 480 seconds]
05:43hdbngr has joined #asahi-dev
05:59hdbngr has quit [Ping timeout: 480 seconds]
06:07Lyude has quit [Quit: Bouncer restarting]
06:08Lyude has joined #asahi-dev
06:24tbodt has quit []
06:24tbodt has joined #asahi-dev
06:32hdbngr has joined #asahi-dev
06:35hardy94 has quit [Remote host closed the connection]
06:39hdbngr has quit [Remote host closed the connection]
06:39hdbngr has joined #asahi-dev
06:48hdbngr has quit [Ping timeout: 480 seconds]
06:50vup has quit []
07:05jackhumphries has joined #asahi-dev
07:05jackhumphries has quit []
07:05hdbngr has joined #asahi-dev
07:14hdbngr has quit [Ping timeout: 480 seconds]
07:24hdbngr has joined #asahi-dev
07:30Ariadne has joined #asahi-dev
07:44hdbngr has quit [Ping timeout: 480 seconds]
07:51sven[cpu0] [SPTM] "VIOLATION_ILLEGAL_DISPATCH_ENTRY_POINT": "validate_dispatch_entry_addr(sptm_validation.h:1406) - caller_domain(0x1), entry_addr(0xfffffe0018ab0000)
07:56hdbngr has joined #asahi-dev
08:08hdbngr has quit [Ping timeout: 480 seconds]
08:13hdbngr has joined #asahi-dev
08:14vup has joined #asahi-dev
08:28mmediouni[m]sven: a hack in the QEMU emulation that is still there: not injecting any interrupts while the guest is running in GL
08:28mmediouni[m]maybe you want to try that one (which I didn't bother tracking down...)
08:42hdbngr has quit [Ping timeout: 480 seconds]
08:43hdbngr has joined #asahi-dev
08:57hdbngr has quit [Ping timeout: 480 seconds]
09:05hdbngr has joined #asahi-dev
10:05pb17 has quit [Ping timeout: 480 seconds]
10:08pb17 has joined #asahi-dev
10:16flokliSon_Goku: following up on https://oftc.catirclogs.org/asahi-dev/2026-07-06#1783361776-1783363807; , do you have that firmware patch somewhere?
10:17flokliI also spoke with yuka a bit, and after reading on https://docs.kernel.org/driver-api/firmware/fw_search_path.html as well we thought maybe just making that take a list would be another option.
10:18chaos_princessflokli: https://lore.kernel.org/asahi/2026073008-crane-jolly-2204@gregkh/T/#mb073990c1131b43a3f45fde8efc247f28ed92970
10:23chaos_princessis this a nixos thingy
10:23jannau"$FIRST_PATH\0$SECOND_PATH" might work but replying that nixos has already an alternative use for this would be appreciated
10:25flokliThis is a nixos thingie for us, but it might also become an issue for some image-based distros.
10:27flokliWe have some mechanism to merge dirs at runtime, and maybe can use that mechanism, but other distros might not have the same option.
10:27flokliI can reply, yes. I'd probably not want to rely on nul bytes accidentially working as a separator.
10:35hdbngr has quit [Ping timeout: 480 seconds]
10:37hdbngr has joined #asahi-dev
10:42chaos_princesssven: is this intentional? https://github.com/AsahiLinux/m1n1/blob/main/src/kboot_atc.c#L170 . All other instances of AUS_VREG_TRIM have length at 5
10:43chaos_princessalso, how exactly is tb failing for you?
10:43chaos_princessi get [ 50.505540] thunderbolt-apple-acio 23b788000.nhi: RTKit: syslog message: lane_init.c:936: Timeout reading AT register spam instead
10:46hdbngr has quit [Ping timeout: 480 seconds]
10:49hdbngr has joined #asahi-dev
11:05flokliapparently phosh now has a semi-standardized format to describe cutouts: https://phosh.mobi/posts/notch-support/ . opened an issue to maybe move these parts to a separate repo, so this data can be consumed by other compositors (?): https://gitlab.gnome.org/World/Phosh/gmobile/-/work_items/16
11:06floklii recall there were also some proposals to add this to the kernel, but it got rejected as too complicated.
11:06JamesCalligeros[m]Theyre working on a wayland protocol for it
11:08floklithat would be sweet! then I assume we'd at some point default apple_dcp.show_notch=1, so compositors have the whole screen to juggle with?
11:08hdbngr has quit [Ping timeout: 480 seconds]
11:09JamesCalligeros[m]Yeah my preference would be to default it to on (i.e. change the param to hide_notch)
11:09JamesCalligeros[m]The kernel parts got bikeshedded into oblivion mostly because people couldn’t agree on the best way to describe panel occlusions
11:09JamesCalligeros[m]There were proposals for high precision and others for “meh close enough”
11:13hdbngr has joined #asahi-dev
11:18elisaado has joined #asahi-dev
11:18elisaadodo note that some composities /really/ don't like the extra pixels added by enabling the notch when fractional scaling is enabled
11:19flokliI'm happy with the format phosh came up with. and definitely better than just not moving at all. If people ever intend to carry this along in device trees rather than another catalogue repository they can still do that, and then there's already a format to use, and the bikeshed alredy happened in userspace.
11:26hardy94 has joined #asahi-dev
11:27hdbngr has quit [Ping timeout: 480 seconds]
11:28JamesCalligeros[m]We dont/cant know we dont have telemetry
11:29JamesCalligeros[m]Also its kind of immaterial. Our job is to make the hardware do what it’s supposed to, not fix every userspace bug that could possibly theoretically affect the platform
11:30JamesCalligeros[m]Compositors not handling fractional scaling properly is kind of out of scope
11:31JamesCalligeros[m]I would wager the split is probably pretty even but in favour of not having the notch
11:31JamesCalligeros[m]Purely because it requires work to enable and not everyone is going to do that
11:31hdbngr has joined #asahi-dev
11:38jannauI plan to add separate modes with and without notch to decp to get rid of the kernel param
11:40flokliJamesCalligeros[m]: jannau: it might make sense to talk with Phosh/gmobile folks on a way to pass this down from the ADT, where we have this info
11:40flokliMaybe this tooling could not just look at the compatible string but actually read some properties that m1n1 could add
11:41chadmedi think if plasma learns about occlusion zones it would be better to just have the notchful modes by default and have a hide_notch parameter rather than expose both sets at all times
11:41svenhttps://f.svenpeter.dev/e5f7c7239a6d6b7c2b6faca003b2ca6c1ed13df76c4748626e0694a5d0be6bd6_sptm_txm_emu.txt and with that i'm really out of excuses to not install macos27 for real
11:42chadmedif userspace plays nicely with the notch and rounded corners at the tops of the screens then theres little reason left to hide them unless someone is using a compositor that doesnt handle the notch properly. once plasma which is our flagship experience does then that becomes less important
11:42sventhe issue was that sptm check's VBAR_EL1 before returning to xnu and due to my banking hack it read what would be VBAR_GL1 and then panicked
11:50hdbngr has quit [Remote host closed the connection]
11:50hdbngr has joined #asahi-dev
12:07floklijannau: I'm not sure what you're proposing. But the current cutouts might not get accidentally applied in our show_notch=0 case, as the "resolution" doesn't match.
12:08hdbngr has quit [Ping timeout: 480 seconds]
12:08flokliI opened another issue about format stabilization, maybe we can carry this info via a device tree property. Would mean we don't need to populate that repository with some static set of cutout data for all models, but can do this in m1n1 possibly
12:10chadmedthe proposal is to simply get rid of the parameter entirely and just have 16:10 and notchful modes all exposed at once and selectable from userspace
12:11hdbngr has joined #asahi-dev
12:11flokliAh, as a mode. But then userspace would still know how to interpret these resolutions (one with, one without notch)
12:13chadmeduserspace doesnt care, we just push the framebuffer down on the screen away from the notch
12:14chaos_princessuserspace needs to care in notchful mocr
12:14chaos_princess*mode
12:15chadmedyes, my point is that once it can/does care then the notchless ones become largely superfluous
12:36hdbngr has quit [Remote host closed the connection]
12:36hdbngr has joined #asahi-dev
13:18hdbngr has quit [Ping timeout: 480 seconds]
13:20hdbngr has joined #asahi-dev
13:21jannauchadmed: I don't like the surface mask in the current form but I heven't spent time/thought how to make it nicer
13:29somewhere-else has joined #asahi-dev
13:30nela has joined #asahi-dev
13:33hdbngr has quit [Remote host closed the connection]
13:33hdbngr has joined #asahi-dev
14:00hdbngr has quit [Ping timeout: 480 seconds]
14:04somewhere-else has quit [Ping timeout: 480 seconds]
14:05hdbngr has joined #asahi-dev
14:11somewhere-else has joined #asahi-dev
14:19somewhere-else has quit [Ping timeout: 480 seconds]
14:23hdbngr has quit [Ping timeout: 480 seconds]
14:27somewhere-else has joined #asahi-dev
14:30nano has quit [Quit: User left the chat]
14:32nano has joined #asahi-dev
14:34Ariadne has quit [Read error: Connection reset by peer]
14:35hdbngr has joined #asahi-dev
14:35somewhere-else has quit [Ping timeout: 480 seconds]
14:42somewhere-else has joined #asahi-dev
14:43Ariadne has joined #asahi-dev
14:47hdbngr has quit [Ping timeout: 480 seconds]
14:50somewhere-else has quit [Ping timeout: 480 seconds]
14:52hdbngr has joined #asahi-dev
14:54ddxtanx has quit [Quit: Konversation terminated!]
14:58somewhere-else has joined #asahi-dev
15:00lynndotpy72 has quit [Quit: bye bye]
15:01hdbngr has quit [Ping timeout: 480 seconds]
15:02lynndotpy72 has joined #asahi-dev
15:03hdbngr has joined #asahi-dev
15:07somewhere-else has quit [Ping timeout: 480 seconds]
15:14somewhere-else has joined #asahi-dev
15:22somewhere-else has quit [Quit: somewhere-else]
15:57mmediouni[m]sven: so far on macOS 27: everything has SPTM, only A18 Pro and M5 get exclaves
15:57mmediouni[m](including M5 Pro/Max)
15:58hdbngr has quit [Ping timeout: 480 seconds]
15:58mmediouni[m]well yeah you'll have to virtualise EL2 too... or just patch a tiny bit incl a few defaults write and then it stops looking
15:59svenright now i'm shifting both XNU and SPTM by just patching a few sysreg accesses from EL2/GL2 to EL1/GL1
16:00svenboth of them already try to access stuff that's EL2/GL2-only while other paths seems to have a if (CurrentEL == 2) before it tries to access that stuff
16:00mmediouni[m] for now, defaults write "com.apple.coreanimation" CAEnableMedina -bool false disables the exclaves code path on M5+
16:01chaos_princesswhat do exclaves do on m5
16:01mmediouni[m]the MacBook Neo has no hardware camera light
16:02chaos_princessi know, interested in m5
16:02chaos_princesssince that uses the same mbp chassis that has it
16:02mmediouni[m]seems to do pretty much nothing on M5 rn
16:05hdbngr has joined #asahi-dev
16:08mmediouni[m]microphone indicator and always on stuff (hey Siri? - alwaysonexclavesd) apparently 🫠
16:08chaos_princess... okay
16:12mmediouni[m]on non exclaves, micactivityd - which is also a conclave host - uses com.apple.private.audio.suppress-mic-indicator and I presume that on exclaves (am on an M4 Max so can't test this) that's not an option anymore
16:16svenis there any information how exaclave core manages pagetables, etc? i guess it also calls into SPTM? does it share XNU's address space or is it actually separate?
16:33mmediouni[m]sven: look at /Library/Developer/CommandLineTools/SDKs/MacOSX27.sdk/System/Library/Frameworks/Kernel.framework/Versions/A/Headers/platform/sptm
16:34mmediouni[m]it's its own SPTM client (like XNU is)
16:35mmediouni[m]but instead of using genter #0 to call SPTM, it does... hvc #0
16:43NightShiftDev has joined #asahi-dev
16:48NightShiftDev[m] has joined #asahi-dev
16:49svenalright.. i'm very dumb because i just remembered that macos ofc supports thunderbolt and i can just connect that mac mini to the dock that's right next to it :D
16:51NightShiftDev has quit [Quit: WeeChat 4.9.5]
16:54enick_272 has quit [Quit: Bridge terminating on SIGTERM]
16:54rhysmdnz has quit [Quit: Bridge terminating on SIGTERM]
17:17elvishjerricco5 has joined #asahi-dev
17:24elvishjerricco has quit [Ping timeout: 480 seconds]
17:24elvishjerricco5 is now known as elvishjerricco
17:25heavenlydemonicprogrammer[m] has joined #asahi-dev
17:29heavenlydemonicprogrammer[m]Hi, I hope this is the right place to ask, I have an m3 MacBook Pro and want to test running asahi on it (I know it’s still unsupported but I want to get a feel of how far it is in development) and I want to know if there’s anything I can help with as a dev that could be fitting as an entry to contribute to the project, are there any guides on how to manually get asahi running without the automated script that would
17:29heavenlydemonicprogrammer[m]walk me through the work to do manually?
17:33chaos_princessbase m3 or pro/max
17:34heavenlydemonicprogrammer[m]I have the base m3 pro model
17:34* heavenlydemonicprogrammer[m]uploaded an image: (124KiB) < https://matrix.org/oftc/media/v1/media/download/ATah5nhnhQuKshgXv-5w6bKcDHhzKwP6mmK6SM2FaXinFu3zr6BS-0Z1w2nW5g9Gp00BfCbye-_CYVIso1VdLKBCegAVdghgAG1hdHJpeC5vcmcvdWJzTmZmUW9IY1FXTGJHc0dJT0dJbnVz >
17:35chaos_princessalso https://github.com/AsahiLinux/m1n1/blob/main/src/kboot_atc.c#L74 - this is most likely incorrect, as it should be only applied in usb case, and there is CIO_ACIOPHY_TOP that is for tb case (they are different)
17:35chaos_princessheavenlydemonicprogrammer[m]: check pull requests under installer repo, go from there
17:36heavenlydemonicprogrammer[m]Okay, thank you
17:47chaos_princessyes, i am on a j274
17:48chaos_princessit also happens that CIO_ACIOPHY_TOP is empt
17:48chaos_princessbut it skips applying USB_ACIOPHY_TOP
18:01heavenlydemonicprogrammer[m]chaos_princess: is it okay to ask questions here while trying to get a system running or there’s a better place for that?
18:01chaos_princesssure, why not
18:02heavenlydemonicprogrammer[m]okay thank you
18:06heavenlydemonicprogrammer[m]what I understood so far is that the script from the pr will get me to m1n1 stage 1 then I have to do the rest on my own and I have to build stage 2, the script won’t do that for me. am I correct?
18:06heavenlydemonicprogrammer[m]also I want to note that I’m reading https://asahilinux.org/docs/sw/m1n1-user-guide/ and using the help of an LLM
18:08heavenlydemonicprogrammer[m]my goal from this is to try to experiment with this and learn more about how asahi works and find a way to help
18:32chaos_princessinstaller has m1n1 as submodule, so it will install latest stage 1
18:32chaos_princessalso, please do not use llms, 1. we do not allow using it for code. 2. they will lead you into stupid directions
18:42heavenlydemonicprogrammer[m]chaos_princess: Okay
18:43chaos_princessthis is probably not helpful, but my m3 install was done by adding all the m3 support patches to an existing m1 install, and then dd-ing it over
18:43heavenlydemonicprogrammer[m]chaos_princess: Yeah I've seen them saying nonsense a lot before I just used it as a secondary search engine while reading the docs but I only use it if I get stuck with search
18:44heavenlydemonicprogrammer[m]chaos_princess: That actually sounds interesting what did you patch exactly and where are the patches from
18:44chaos_princessi don't care how you use them, just saying that if you ask someone "i have used a llm and ...", the most likely response is "have you considered not doing that"
18:44chaos_princessjust look at m1n1 and linux repos under the same org
18:44chaos_princessfor linux - branch asahi-wip + PRs
18:46heavenlydemonicprogrammer[m]chaos_princess: Okay then I will stop using it
18:46heavenlydemonicprogrammer[m]chaos_princess: Okay I'll go check them
18:46heavenlydemonicprogrammer[m]Thank you
18:46chaos_princess(you don't need to ping me in every msg :P)
18:47heavenlydemonicprogrammer[m]I just replied to the messages to make it explicit what each message is responding to, I'll stop doing it
18:49chaos_princessso, this channel is actually an irc channel, with matrix side being bridged to it. since i am on irc side, i don't even see those
18:49chaos_princessand yes, irc is the "primary" side
18:49heavenlydemonicprogrammer[m]Would it be better if I get on the irc side?
18:49chaos_princessup to you
18:50midou has joined #asahi-dev
19:00heavenlydemonicprogrammer[m]Okay, I'll give it a try later
19:00heavenlydemonicprogrammer[m]Thank your for your time and help
19:40roxfan2 has joined #asahi-dev
19:45roxfan has quit [Ping timeout: 480 seconds]
19:57chaos_princesswhich tb flavor worked back then, just "true" tb, or usb4 too?
20:04chaos_princessand the phy driver also did the "usb4 not implemented, fallback to usb2"?
20:05jannauprobably not since it has been rewritten since then but I'm not sure if that message is critical
20:08chaos_princessi am just wondering if that causes problems, as i don't have any "true tb" devices, only usb4 ones
20:11chaos_princessi changed a few other macos differences, didn't help either
20:11chaos_princessthe only major thing i see is that now atcphy_configure runs synchronously instead of in a mux_set_work
20:32amarioguy has joined #asahi-dev
20:42svenhttps://f.svpe.de/27f599ba4181b71a4630947fd6876cd7123f7e9bfb926c85d8e4fd7ccedb61ce_macos27.png yay, it boots. most weird bug though: it refuses to display anything on my display connected via thunderbolt :D
20:53svenhttps://f.svpe.de/dea4c79f127e3fed14e5dfe9f0556ac2de46be37309cc02be7d69b57c146c364_xnu_emulated.txt that's with emulated, panics somewhere probably because the emulation is inaccurate
21:11svenhttps://f.svpe.de/69f80a3ce9193fb3791cf27b1751d8c205699693b778ba07dbe6cd5dfcdb5d99_serror.txt buuuuuuut https://f.svpe.de/1a378cbdcc04d9fcb35732ec367e1409b42ed4767154c8add4dac08de55429a4_macos27cursed.jpeg
21:11chaos_princesswhy would launchd of all things need it
21:12chaos_princessmaybe it is not launchd but dyld?
21:13mmediouni[m]That's a nice result
21:14mmediouni[m]And yeah it's very slow I was quite happy when with a qemu impl it was possible to get a "mount" command execution below a second
21:15mmediouni[m]And even then not consistently because all the transitions caused the scheduler to put the whole thing on the ecores...
21:20mmediouni[m]It's good enough to cover the tracing part tbh so it does the job
21:22nnym has quit [Remote host closed the connection]
21:23nnym has joined #asahi-dev
21:24mmediouni[m]now that even A13 has SPTM, wonder if a custom kernel build able to boot macOS without SPTM will be possible once there's the xnu source release...
21:28svenstill serrors while it's in wfi apparently though: L2C_ERR_STS: 0x11000ffc00000000 L2C_ERR_ADR: 0x0 L2C_ERR_INF: 0x0
21:34Ariadne_ has joined #asahi-dev
21:35Ariadne has quit [Read error: Connection reset by peer]
21:37nnym has quit [Ping timeout: 480 seconds]
21:38DD65 has quit [Remote host closed the connection]
21:38DD65 has joined #asahi-dev
21:45nnym has joined #asahi-dev
21:51svenah, one thing that would be helpful: if someone can get me a live ADT from apple entitled boot mode on a machine with SPTM and TXM that would be great
21:53nnym has quit [Ping timeout: 480 seconds]
21:58svenand a msc student i've been advising did the SPTM bringup so i'll have to coordinate when i can publish that part with him
22:02amarioguythat commit is 5 years ago tho, so our fork should definitely be carrying that change...
22:03nnym has joined #asahi-dev
22:11nnym has quit [Ping timeout: 480 seconds]
22:18nnym has joined #asahi-dev
22:45hdbngr has quit [Ping timeout: 480 seconds]
22:55r0ni has quit [Ping timeout: 480 seconds]
23:05hdbngr has joined #asahi-dev
23:13r0ni has joined #asahi-dev
23:14hdbngr has quit [Ping timeout: 480 seconds]
23:24hdbngr has joined #asahi-dev
23:32hdbngr has quit [Ping timeout: 480 seconds]
23:34ciara has quit [Ping timeout: 480 seconds]
23:36ciara has joined #asahi-dev
23:40hdbngr has joined #asahi-dev
23:49hdbngr has quit [Ping timeout: 480 seconds]
23:51hdbngr has joined #asahi-dev