2026-01-20
00:18pb17 has quit [Ping timeout: 480 seconds]
00:30hdbngr has joined #asahi-dev
00:34fl0_idjannau 16 core. I thought it was referred to as t6031 as well... but yeah I'll try to check that
00:39hdbngr has quit [Ping timeout: 480 seconds]
00:52hdbngr has joined #asahi-dev
00:56mixi has quit [Ping timeout: 480 seconds]
00:57pb17 has joined #asahi-dev
01:06mixi has joined #asahi-dev
01:08hdbngr has quit [Ping timeout: 480 seconds]
01:11hdbngr has joined #asahi-dev
01:26hdbngr has quit [Ping timeout: 480 seconds]
01:52hdbngr has joined #asahi-dev
01:57midou has quit [Read error: No route to host]
01:57midou has joined #asahi-dev
02:01hdbngr has quit [Ping timeout: 480 seconds]
02:05midou has quit [Ping timeout: 480 seconds]
02:11pb17 has quit [Ping timeout: 480 seconds]
02:19midou has joined #asahi-dev
02:21pb17 has joined #asahi-dev
02:27k70_ has joined #asahi-dev
02:29k70 has quit [Ping timeout: 480 seconds]
02:29k70_ is now known as k70
02:31midou has quit [Ping timeout: 480 seconds]
03:12hexdump0815 has joined #asahi-dev
03:14hexdump01 has quit [Ping timeout: 480 seconds]
03:14hdbngr has joined #asahi-dev
03:15tobhe_ has joined #asahi-dev
03:22midou has joined #asahi-dev
03:22tobhe has quit [Ping timeout: 480 seconds]
03:23hdbngr has quit [Ping timeout: 480 seconds]
03:26midou has quit [Read error: Connection reset by peer]
03:32midou has joined #asahi-dev
04:04midou has quit [Read error: No route to host]
04:14pb17 has quit [Ping timeout: 480 seconds]
04:21midou has joined #asahi-dev
04:32pb17 has joined #asahi-dev
04:56nora has joined #asahi-dev
04:56hdbngr has joined #asahi-dev
04:58nora_ has quit [Ping timeout: 480 seconds]
05:05hdbngr has quit [Ping timeout: 480 seconds]
05:09hdbngr has joined #asahi-dev
05:15midou has quit [Ping timeout: 480 seconds]
05:18hdbngr has quit [Ping timeout: 480 seconds]
05:22midou has joined #asahi-dev
06:00hdbngr has joined #asahi-dev
06:04chadmed has quit [Quit: Konversation terminated!]
06:04chadmed has joined #asahi-dev
06:07midou has quit [Ping timeout: 480 seconds]
06:08hdbngr has quit [Ping timeout: 480 seconds]
06:15midou has joined #asahi-dev
06:35midou has quit [Ping timeout: 480 seconds]
06:54hdbngr has joined #asahi-dev
06:57jannaufl0_id: the 16 core variant is t6031. the possible issues I saw are with the 14-core variant (t6034)
07:02midou has joined #asahi-dev
07:03hdbngr has quit [Ping timeout: 480 seconds]
07:10midou has quit [Read error: Connection reset by peer]
07:10hdbngr has joined #asahi-dev
07:13midou has joined #asahi-dev
07:19hdbngr has quit [Ping timeout: 480 seconds]
07:21midou has quit [Ping timeout: 480 seconds]
07:38noopwafelmy linux boot on the m3 air is still pretty broken, I had to sabotage smt and chicken bits and the cpuidle driver
07:41noopwafelmy tree is also full of hacks for other things, so this _entirely_ might be my fault, but if anyone has tips to avoid me reinventing the wheel, let me know
07:45midou has joined #asahi-dev
07:47jannaunoopwafel: https://github.com/jannau/m1n1/tree/t8122_bringup + https://github.com/jannau/linux/tree/t8122-j504 does work for the limited hw enabled / present in the device tree
07:52jannaumoved the usb-c ports (not enabled) to the device .dts since j504 should be the only m3 device with i2c USB-PD port controllers
07:59midou has quit [Ping timeout: 480 seconds]
08:01hdbngr has joined #asahi-dev
08:09noopwafeljannau: thanks, that gives me working SMT after just adding a minimal j613 dts which is a big improvement at least :)
08:09hdbngr has quit [Ping timeout: 480 seconds]
08:11hdbngr has joined #asahi-dev
08:12chaos_princesssmt or smp? :P
08:12noopwafeldoes anyone have some idea what might be a reasonable firmware target for m3? I am running with 14.8.3 and wondering if I am going to waste my time porting to that
08:13chaos_princess14.latest seems fine?
08:20hdbngr has quit [Ping timeout: 480 seconds]
08:29ravikant_ has joined #asahi-dev
08:41nickchanShiz: setting bit[1] sets the corresponding EL's timer to 24 MHz subject to VHE caveats
08:43nickchanShiz: i mean we know that there's a whole set of custom timer registers that's 1:1 to the arch ones likely with just with a different freq
08:45nickchanhttps://github.com/apple-oss-distributions/xnu/blob/43a90889846e00bfb5cf1d255cdc0a701a1e05a4/osfmk/arm64/machine_routines.c#L2324
08:45hdbngr has joined #asahi-dev
08:45nickchanthis is not all need to play with hw to know what the rest of the arch register clones are
08:49midou has joined #asahi-dev
08:56nickchanive seen the same register first referenced as acntrdir_el1 but later agtcntrdir_el1 and the encoding is the same
09:10midou has quit [Remote host closed the connection]
09:17nickchannoopwafel: maybe find a register that reads out 0x16e3600 (24 MHz) or 0x3b9aca00 (1 GHz) other than CNTFRQ_EL0
09:19midou has joined #asahi-dev
09:29kitlithjannau: https://gitlab.freedesktop.org/kitlith/mesa/-/commit/279e9caf1090017001c235ed3ef9a5dc0a2d6f19 -- i'm not entirely sure that filtering out virtio at this stage is strictly necessary. It does show up as platform bus, I'm pretty sure it'll fail anyway when trying to find the connector, as drm_info fails to get any useful information when running under muvm. But I guess there's also little reason not to filter it out at this stage, aside f
09:55hdbngr has quit [Ping timeout: 480 seconds]
10:00hdbngr has joined #asahi-dev
10:08hdbngr has quit [Ping timeout: 480 seconds]
10:18hdbngr has joined #asahi-dev
10:27hdbngr has quit [Ping timeout: 480 seconds]
10:27___nick___ has joined #asahi-dev
10:29hdbngr has joined #asahi-dev
10:49hdbngr has quit [Ping timeout: 480 seconds]
10:51___nick___ has quit []
10:53___nick___ has joined #asahi-dev
10:53___nick___ has quit []
10:55___nick___ has joined #asahi-dev
10:57hdbngr has joined #asahi-dev
11:02maor26 has joined #asahi-dev
11:06culman has joined #asahi-dev
11:12chadmed_ has joined #asahi-dev
11:12chadmed has quit [Read error: Connection reset by peer]
11:16culman has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
11:27chadmed_ has quit [Quit: Konversation terminated!]
11:27chadmed_ has joined #asahi-dev
11:33jmax_ has quit [Remote host closed the connection]
11:37culman has joined #asahi-dev
11:38crabbedhaloablut has quit []
11:38crabbedhaloablut has joined #asahi-dev
11:53culman has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
11:53culman has joined #asahi-dev
11:53culman has quit []
12:25ravikant_ has quit [Ping timeout: 480 seconds]
12:30hdbngr has quit [Ping timeout: 480 seconds]
12:37hdbngr has joined #asahi-dev
12:46hdbngr has quit [Ping timeout: 480 seconds]
12:48hdbngr has joined #asahi-dev
12:49culman has joined #asahi-dev
13:05culman has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
13:05ravikant_ has joined #asahi-dev
13:22___nick___ has quit [Ping timeout: 480 seconds]
13:23culman has joined #asahi-dev
13:24___nick___ has joined #asahi-dev
13:26hdbngr has quit [Ping timeout: 480 seconds]
13:34hdbngr has joined #asahi-dev
13:42hdbngr has quit [Ping timeout: 480 seconds]
13:51hdbngr has joined #asahi-dev
14:00hdbngr has quit [Ping timeout: 480 seconds]
14:03hdbngr has joined #asahi-dev
14:07fl0_id___ has joined #asahi-dev
14:23midou has quit [Read error: No route to host]
14:31midou has joined #asahi-dev
14:40hdbngr has quit [Ping timeout: 480 seconds]
14:47hdbngr has joined #asahi-dev
14:54yukaI was able to map the uart mmio ranges in early boot and use debug_putc after enable_mmu, traced up to aic init
15:02culman has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
15:08fl0_id___yuka, do you have this up somewhere, just in case it helps me? I will probably also test some mcc.c changes, as tz_regs seem different, and i don't know what device was used for those.
15:11john-cabaj has joined #asahi-dev
15:14culman has joined #asahi-dev
15:14hightower3 has joined #asahi-dev
15:22hightower4 has quit [Ping timeout: 480 seconds]
15:35culman has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
15:39culman has joined #asahi-dev
15:43culman has quit []
15:50midou has quit [Ping timeout: 480 seconds]
15:57midou has joined #asahi-dev
16:52pb17 has quit [Ping timeout: 480 seconds]
17:01touch has joined #asahi-dev
17:02touchi wanna look into touch id... i see sep is already implemented so i guess it won't be very hard.
17:08touchi have no experience with asahi development, but i do have some exp with reverse engineering
17:11svenand this kinda needs a dedicated machine because if you mess that xART locker file up you might need to DFU revive the mac
17:14sventhen try to build some python script that talks to SEP. for touchid you need at least user management and the communication with that touchid peripheral
17:14sventhen once that works we can think about the linux driver and userspace daemon should look like
17:15svenhttps://github.com/AsahiLinux/m1n1/blob/main/proxyclient/m1n1/hw/sep.py no idea if it still works :-)
17:16nnym_ has joined #asahi-dev
17:17sventhere's certainly stuff happening on boot, but no one knows what exactly and what's actually required
17:18touchyeah and you need a password first to unlock so its interesting if it's something manadatory for the sep
17:18svenyeah, but i'm sure there's also lots of unrelated stuff happening. no one really looked into this so far :-)
17:18nnym has quit [Ping timeout: 480 seconds]
17:20svenunfortunately the sensor won't talk to us at all and e.g. give us raw fingerprint data so we can't just bypass SEP
17:21touchi'm gonna guess not a lot of debugging info will be there so there will be a lot of figuring out what each log means
17:23chaos_princessget xars/xarm endpoints first, nothing relevant will talk to you w/o it
17:23chaos_princessand those are the annoying ones with a high chance of getting a trip to dfu city
17:23sveni have some notes on the gigalocker format itself somewhere, but it's just a simple key/value store
17:24chaos_princessliterally the endpoint names
17:25chaos_princessendpoint 0xFD will tell you the endpoints that the sep firmware exposes and the buffer sizes
17:25chaos_princessthere is a stub driver in linux already that just loads the firmware and ignores every message received on 0xfd
17:25svenhttps://f.svpe.de/262f757f4eef4825164bd703903e9f5ef2f76bd6ca410511a101a361b62084ed_gl_parse.py.txt i guess those were my notes before i gave up due to the risk of dfu trips
17:26chaos_princessgoing straight to ghidra is spicy
17:26svenplease read https://asahilinux.org/copyright/ and especially the reverse engineering policy
17:27chaos_princessanti-rollback tokens inside of sep, if sep data inside gl file is lost, we suspect it will invalidate all boot signatures and destroy nvme encryption keys
17:28svenSEP will ask you to update that file via the xars/xarm endpoints and you kinda have to do that
17:28chaos_princessonce you start talking to sep and/or interfering with xnu trying to talk to it, you may get things wrong
17:28nickchanabout the rtkit version 10 endpoint question i have reached a point where im pretty sure i cannot answer the question properly without tracing xnu so time to bite that bullet i guess
17:29chaos_princessis it not z2?
17:30chaos_princesstouch: yes, but dfu restore wipes the disk
17:31chaos_princessit is the most low level recovery mode that is "everything wrong, we are setting up the machine as new"
17:33pb17 has joined #asahi-dev
17:33chaos_princessnot much, it is one of the few copros that need to be backwards compatible
17:35fl0_id___ has quit [Ping timeout: 480 seconds]
17:36__Tetris__4 has joined #asahi-dev
17:39__Tetris__4 has quit []
17:39__Tetris__0 has joined #asahi-dev
17:39ravikant_ has quit [Ping timeout: 480 seconds]
17:40nickchani will have pongoOS do some work before m1n1 for xnu tracing cause im not about to put sep firmware load race into m1n1
17:42__Tetris__ has quit [Ping timeout: 480 seconds]
17:42__Tetris__0 is now known as __Tetris__
17:44Bjarne[m]I was also poking around with macvdmtool. I managed to get serial debug output of the target on my macOS Host (DFU to DFU).
17:44nickchanconveniently there's *always* 0x800 contineous bytes of empty space + at least 0x400 bytes of empty space elsewhere inside xnu i could use
17:45Bjarne[m]But while SerialShim was running which should have hooked to the /dev/tty.debug-console right?
17:46nickchanRight i have to disable spectre mitigation cause otherwise almost all kernel memory gets unmapped in el0
17:49nickchanbecause of some ktrr the pagetables can't be freely modified but it is ok to change T1SZ
17:59fl0_idwould linux:asahi vs linux:asahi-wip have meaninful changes that could impact m3 bringup things? so far I always used asahi
18:00jannauBjarne[m]: sigh, I wasn't watching the repo. Did you see https://github.com/AsahiLinux/tuxvdmtool/pull/3
18:09jannauBjarne[m]: did you run `cargo fmt` on your changes? `const TPS_REG_MODE : u8 = 0x03;` look a little weird
18:12yukafl0_id my t8132 debug_putc work is here: https://cyberchaos.dev/yuka/linux/-/commit/8c86bd8c76bbe02dbf20f3e0f8fb0c311be30faf
18:24fl0_id___ has joined #asahi-dev
18:26yukathis is all with cpuhp_setup_state() call in aic init commented out btw, since it hangs there otherwise
18:35fl0_id___ has quit [Ping timeout: 480 seconds]
18:39Bjarne[m]I also have built a setup for merging macvdm and tuxvdm, I’m rather confident it will work but I’m trying to use as little C Code as possible.
18:45hdbngr has quit [Ping timeout: 480 seconds]
18:52hdbngr has joined #asahi-dev
19:01hdbngr has quit [Ping timeout: 480 seconds]
19:04Bjarne[m][jannau](https://matrix.to/#/@_oftc_jannau:matrix.org): I ran cargo fmt and added a commit to PR#4
19:05hdbngr has joined #asahi-dev
19:09chadmed_ has quit [Quit: Konversation terminated!]
19:09chadmed_ has joined #asahi-dev
19:14jannauBjarne[m]: manually merged. I already ran cargo fmt on each format. also fixed all clippy issues. thanks for this, I can now drop the horrible code from the downstream tipd driver
19:19jannauBjarne[m]: maybe wait for #3 or at least ignore the multiple transports aspects of it for now
19:36pb17 has quit [Ping timeout: 480 seconds]
19:49pb17 has joined #asahi-dev
19:49fatmalama has joined #asahi-dev
20:23john-cabaj1 has joined #asahi-dev
20:24john-cabaj has quit [Ping timeout: 480 seconds]
20:24john-cabaj1 is now known as john-cabaj
20:51arcade_droid has quit [Ping timeout: 480 seconds]
20:55JoshuaPeisach[mds] has joined #asahi-dev
20:55JoshuaPeisach[mds]kernel newbie here, question about the TODO in the “asahi” drm driver. I see it wants DMA address/range validation, is it basically just ensure address + range > kernel VA range global?
21:02chaos_princessmind linking the line?
21:04jannauI'm not sure but isn't the kernel VA range check more complicated as user space is free to place that where it likes
21:06___nick___ has quit [Remote host closed the connection]
21:07JoshuaPeisach[mds]chaos_princess: yeah, line 227, sorry https://github.com/AsahiLinux/linux/blob/f1879b3d78dcf69a8ff9be3b504c1bd6d5fa2f47/drivers/gpu/drm/asahi/mmu.rs#L207
21:33ravikant_ has joined #asahi-dev
21:33ravikant_ has quit []
22:11hightower3 has quit [Ping timeout: 480 seconds]
22:21fl0_id___ has joined #asahi-dev
22:51hdbngr has quit [Ping timeout: 480 seconds]
22:54hdbngr has joined #asahi-dev
22:57pb17 has quit [Ping timeout: 480 seconds]
22:59pb17 has joined #asahi-dev
23:19hdbngr has quit [Remote host closed the connection]
23:19hdbngr has joined #asahi-dev
23:22fl0_id___ has quit [Remote host closed the connection]
23:22fl0_id___ has joined #asahi-dev
23:36hdbngr has quit [Ping timeout: 480 seconds]
23:44hdbngr has joined #asahi-dev
23:52fl0_id___ has quit [Ping timeout: 480 seconds]