#
…

2026-06-18

00:00Extern has quit [Quit: WeeChat 4.9.0]
00:09KaisarMaytham[m] has quit [Ping timeout: 480 seconds]
00:09KaisarMaytham[m] has joined #asahi-dev
00:09norn[m] has quit [Ping timeout: 480 seconds]
00:09person_who_loves_stealing[m] has quit [Ping timeout: 480 seconds]
00:09BitBanger[m] has quit [Ping timeout: 480 seconds]
00:09dcavalca has quit [Ping timeout: 480 seconds]
00:09kirakun[m] has quit [Ping timeout: 480 seconds]
00:09goffrie[m] has quit [Ping timeout: 480 seconds]
00:09robp[m] has quit [Ping timeout: 480 seconds]
00:09Moderation[m] has quit [Ping timeout: 480 seconds]
00:09mmmm[m] has quit [Ping timeout: 480 seconds]
00:10person_who_loves_stealing[m] has joined #asahi-dev
00:12dcavalca has joined #asahi-dev
00:12norn[m] has joined #asahi-dev
00:14BitBanger[m] has joined #asahi-dev
00:18kaleryeblueberryoboe has joined #asahi-dev
00:22mmmm[m] has joined #asahi-dev
00:24goffrie[m] has joined #asahi-dev
00:24kirakun[m] has joined #asahi-dev
00:24Moderation[m] has joined #asahi-dev
00:25robp[m] has joined #asahi-dev
00:32LarstiQ has quit [Quit: ZNC - http://znc.in]
00:33LarstiQ has joined #asahi-dev
00:36steinsgate has quit [Quit: Connection closed for inactivity]
00:51pb17 has quit [Ping timeout: 480 seconds]
00:53kaleryeblueberryoboe has quit [Read error: Connection timed out]
00:53pb17 has joined #asahi-dev
01:16jsk_ has joined #asahi-dev
01:19jsk has quit [Ping timeout: 480 seconds]
01:20jsk_ is now known as jsk
01:39kaleryeblueberryoboe has joined #asahi-dev
01:50fr4gm3nt3dm3m0r1e5[m] has left #asahi-dev [Kicked by @draupnir-fedora:rory.gay : spam]
02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-dev
02:31tobhe_ has joined #asahi-dev
02:35kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
02:38tobhe has quit [Ping timeout: 480 seconds]
03:16kaleryeblueberryoboe has joined #asahi-dev
03:24marvin24 has quit [Ping timeout: 480 seconds]
04:26kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
04:32mattia013 has joined #asahi-dev
04:34kaleryeblueberryoboe has joined #asahi-dev
04:48kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
04:50kaleryeblueberryoboe has joined #asahi-dev
04:59kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
05:06kaleryeblueberryoboe has joined #asahi-dev
05:15ipatch has quit [Quit: WeeChat 4.9.0]
05:16ipatch has joined #asahi-dev
05:23kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
05:31kaleryeblueberryoboe has joined #asahi-dev
05:32weirdtreething has quit [Read error: Connection reset by peer]
05:32weirdtreething has joined #asahi-dev
05:39kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
05:39vimproved has quit [Ping timeout: 480 seconds]
05:41kaleryeblueberryoboe has joined #asahi-dev
05:55kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
06:02kaleryeblueberryoboe has joined #asahi-dev
06:05mattia013 has quit [Remote host closed the connection]
06:07pb17 has quit [Ping timeout: 480 seconds]
06:23vimproved has joined #asahi-dev
06:28yurix has quit [Read error: Connection reset by peer]
06:29yurix has joined #asahi-dev
06:31pb17 has joined #asahi-dev
06:51sofus has joined #asahi-dev
06:51hardy94 has joined #asahi-dev
07:34hardy94 has quit [Remote host closed the connection]
08:03kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
08:26sofuswould it be silly to let the cm3 handle the tunatables?
08:27JamesCalligeros[m]I mean youve already written your own firmware :p
08:28chaos_princessarguably better in some way, you can shove all of model-specific stuff into the firmware blob
08:28sofusokay great, thanks
08:28JamesCalligeros[m]My issue was less the doing them part and more the magic numbers. We have a ~decent idea of the registers via R and even if you just #define UNK_TUNABLE0 0x… itll look a bit nicer
08:30JamesCalligeros[m]But yes also stuffing all the model specific stuff in a fw blob is a good idea
08:30kaleryeblueberryoboe has joined #asahi-dev
08:30JamesCalligeros[m]If we take this approach going forward, would you be amenable to transferring the fw repo to the github org?
08:31sofusi have gotten the tunatables names for all devices now, some of the newer ones have 346 different registers
08:31sofusJamesCalligeros[m]: yes ofc
08:34sofusfor the tunatable registers, R basicly copied names from the kc, not sure if i can do the same?
08:34nickchanthis should be fine
08:35nickchanit's not any different from the register names copied from proprietary datasheet in all places in linux
08:36JamesCalligeros[m]NAL but yes “facts about the hardware” are not copyrightable
08:36JamesCalligeros[m]Register names that tell you what the register does are facts about the hardware and not novel/original code
08:38yuka has quit [Remote host closed the connection]
08:44kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
08:44chaos_princess... debatable
08:44sofusthanks, the names are super inconsistent, some tunatables equal, but one or two names are different
08:45chaos_princessyou can make an argument that the name is someone's interpretation of how to write the hw description in a concise way
08:46JamesCalligeros[m]Sure but that only applies if you directly copy the name exactly. If the register is called AVC_MACROBLOCK_SIZE in the kext and I call it MBLK_SZ_H264 then who cares
08:47chaos_princesswell, AVC_MACROBLOCK_SIZE is also probably fine
08:47chaos_princessmy take is "this is probably fine in this case"
08:50sofusits actually all dma config, any clue what RD and WR could mean?
08:51sofusits one of the things they are inconsistent with
08:52yuka has joined #asahi-dev
08:56sofuswhat about BFR? There is always one DMA_BFR_CONFIG with the same name after an DMA_CONFIG
08:56JamesCalligeros[m]Would expect buf and not bfr but who knows
08:57sofusthat would kinda make sense
09:01chaos_princessdisplay still can't survive pmgr_misc even with min-state
09:18pb17 has quit [Ping timeout: 480 seconds]
09:24pb17 has joined #asahi-dev
09:30gargantua has joined #asahi-dev
10:55fl0id_oftc is now known as fl0_id_afk
11:02fl0_id_afk is now known as fl0id_oftc
11:48chadmed has quit [Quit: Konversation terminated!]
11:50chadmed has joined #asahi-dev
12:10gargantua has quit [Ping timeout: 480 seconds]
12:47Josh[m] has joined #asahi-dev
13:02sofus has quit [Ping timeout: 480 seconds]
13:39jannauchaos_princess: can you try adding "apple,min-state = <4>;" to ps_disp_fe? I don't understand what they did with disp the power-states
13:42sofus has joined #asahi-dev
13:53jannaugood, not sure what to do about it
14:00chaos_princessunfortunately, dcs still can't go as low as macos does, we can only do 8, while macos does 6
14:07oliver has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
14:08oliver has joined #asahi-dev
14:09rodmac has joined #asahi-dev
14:15sven[ 0.025462] smp: Bringing up secondary CPUs ...
14:15sven[ 0.026035] psci: calling into PSCI-via-EFI now
14:15svenhello, world from m1n1!
14:15sven[ 0.026035] psci: calling into PSCI-via-EFI done
14:15sven[ 5.090406] CPU1: failed to come online
14:16svenprogress \o/
14:16svenit's just a nop for now but the kernel -> m1n1 -> kernel calls work now and the stub can still access all mmio
14:30jannaum1n1 writes directly to the uart inside the psci call? \o/
14:30svenrequires a rather hacky kernel patch because there's no way to express "we really need Device-nGnRnE" in the memory table
14:47chaos_princessare you using kernel memory mappings?
14:48svenno, you can define special EFI runtime mappings that the kernel enables before jumping to m1n1 again
14:48svenso it's just the usual 1:1 mapping
14:49mazguess we'd have to each the EFI code in the kernel to generate "special" mappings on Apple HW...
14:50svenyeah, right now i have this hack here but that's something we can bikeshed on the mailing list
14:50sven+ if (of_machine_is_compatible("apple,arm-platform"))
14:50sven+ prot = __pgprot(PROT_DEVICE_nGnRnE);
14:51mazyup, I expected something of the sort. seems relatively minor, in the grand scheme of things.
14:52chaos_princesswhy not jump into m1n1 with mmu off?
14:53svenneeds more invasive patches because that's not how efi runtime services are supposed to work
14:54mazyou still want the rest of the kernel to be there. and you need the MMU on for coherency (at least architecturally, AS may be more lenient here).
14:57nickchani believe the nGnRnE requirement is new in A14/M1
14:57sveni only care about M1+ here
14:57nickchanbut I previous gen still *work* with nGnRnE so i think it's fine
14:57svenanything before that can always be a follow-up patch
14:57nickchanwell you can't say that if you use 'of_machine_is_compatible("apple,arm-platform")'; not like that's an issue here
14:57svenwhy not?
14:58sventhis will only exist for M1+ for now and you can fix it later for <M1 then
14:58nickchanthere is no need to fix anything
14:58nickchanso this is fine
14:58sveneven better then
14:58nickchanthey still work with nGnRnE im just saying nGnRE works on previous gen
14:59nickchanextra context for your commit messages
15:11kaleryeblueberryoboe has joined #asahi-dev
15:19kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
15:24kaleryeblueberryoboe has joined #asahi-dev
15:32jannauchaos_princess: how much pmgr log do you need? Is just the initialisation enough? macos panics easily on t8122
15:32chaos_princessis there a pmgr misc on t8122 at all?
15:33jannauprobably not, I thought you were looking into PMP
15:34jannaut6030 might not have pmgr misc either
15:34chaos_princessswitched since yu-ka was looking into it
15:34chaos_princesswell, t6030 might, idk. the thing i need should be rather early
15:36sofus has quit [Ping timeout: 480 seconds]
15:41jannauwhat would I be looking for? PMGR_SOC_MISC?
15:42chaos_princessidk actually, something that has the "right vibes", on t6031 it was sth like MPMGRClkgen
15:42sofus has joined #asahi-dev
15:44jannaulogs sent via priv msg
15:46chaos_princessit probably does not have it
15:50jannauyes, it feels more like an extended M3 than a chopped down M3 Max
15:53chaos_princessand i was wrong, dcp still dies
15:53chaos_princessyou just need to wait longer for it to happen
16:10sofus has quit [Ping timeout: 480 seconds]
16:19sofus has joined #asahi-dev
16:28kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
16:35kaleryeblueberryoboe has joined #asahi-dev
16:41slowlydev has joined #asahi-dev
16:41mripard_ has joined #asahi-dev
16:43alexanderian has joined #asahi-dev
16:43mripard has quit [Ping timeout: 480 seconds]
16:43alexanderian has left #asahi-dev [#asahi-dev]
16:48XeR has quit [Remote host closed the connection]
16:48XeR has joined #asahi-dev
16:48kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
16:53mripard has joined #asahi-dev
16:54mripard_ has quit [Ping timeout: 480 seconds]
16:56kaleryeblueberryoboe has joined #asahi-dev
17:04kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
17:06slowlydev has quit [Remote host closed the connection]
17:12kaleryeblueberryoboe has joined #asahi-dev
17:20kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
17:27kaleryeblueberryoboe has joined #asahi-dev
17:29thunfisch has quit [Quit: frrrp!]
17:38kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
17:45kaleryeblueberryoboe has joined #asahi-dev
17:47Pilot[PENMaxi][m] has joined #asahi-dev
17:54kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
18:01kaleryeblueberryoboe has joined #asahi-dev
18:09kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
18:14rodmac has quit [Quit: leaving]
18:15gargantua has joined #asahi-dev
18:17kaleryeblueberryoboe has joined #asahi-dev
18:24steinsgate has joined #asahi-dev
18:28kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
18:36kaleryeblueberryoboe has joined #asahi-dev
18:46kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
18:48yukaI have m1n1 in m1n1 hv on M4 working :-)
18:48yukaisn't actually too difficult, not sure what was missing when I tried it the other day
18:48yukanow adding the obtaining cpu features in proxyclient thing
18:49yukawhat does tgtypes mean?
18:49yukaas in proxyclient/m1n1/tgtypes.py
18:49yukaI'm unsure if the cpu features struct belongs there
18:53kaleryeblueberryoboe has joined #asahi-dev
18:58svenno idea :D
18:59svenproxyclient contains quite a bunch of hacks anyway since we don't ship that to end-users
19:02kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
19:09kaleryeblueberryoboe has joined #asahi-dev
19:10yukaany reason we always proxy msrs into gl2?
19:11yukaI would assume for anything mac os would do, EL2 would be enough
19:12yukagenerally no idea what happens if an m1n1 guest tries to do genter pre-M4
19:14sventhere are some sysregs that can only be written from gl2 once its enabled
19:14svenat least HCR_EL2 but i think there were a few more
19:14svenif gxf isn't enabled just writing to them from EL2 is fine
19:15yukaAh, makes sense
19:15yukathen I will make it if apple_sysregs_unlocked
19:17Pilot[PENMaxi][m]Hide your wives and your valuables there are mauraders out there who have government contracts to pillage.
19:18kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
19:21kaleryeblueberryoboe has joined #asahi-dev
19:22sofus has quit [Quit: Lost terminal]
19:26svenyeah, that should be fine
19:27svenoh, and if a guest on M1-M3 calls genter from EL1 it'll end up in GL1 fwiw
19:27svenunless you set some bit in some impdef sysreg to trap genter, then you'll end up in GL2 or EL2
19:29kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
19:36yukaI think for smp hv we need some new code
19:36yukasince we can not set rvbar
19:37kaleryeblueberryoboe has joined #asahi-dev
19:38john-cabaj has joined #asahi-dev
19:38john-cabaj has quit []
19:45kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
19:52kaleryeblueberryoboe has joined #asahi-dev
19:56hdbngr has joined #asahi-dev
20:01kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
20:07kaleryeblueberryoboe has joined #asahi-dev
20:18kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
20:20hdbngr has quit [Ping timeout: 480 seconds]
20:22kaleryeblueberryoboe has joined #asahi-dev
20:22truita has quit [Remote host closed the connection]
20:25truita has joined #asahi-dev
20:32kaleryeblueberryoboe has quit [Ping timeout: 480 seconds]
20:40kaleryeblueberryoboe has joined #asahi-dev
20:40yukagot it working
20:41yukawith smp
20:42hardy94 has joined #asahi-dev
20:42hardy94 has quit []
21:02hardy94 has joined #asahi-dev
21:02hdbngr has joined #asahi-dev
21:04yukaMulti-dockchannel-uart would be nice now
21:04yukafor hv on the macbook neo
21:06yukaseems like the other dockchannel uarts are at /arm-io/dockchannel-uart[0] + 0x[1-7]4004
21:06yukaso let's use them
21:07kaleryeblueberryoboe has quit [Remote host closed the connection]
21:07kaleryeblueberryoboe has joined #asahi-dev
21:14yuka(the 4004 being the DATA_TX8, so really 0x[1-7]0000)
21:23kaleryeblueberryoboe has quit [Remote host closed the connection]
21:24kaleryeblueberryoboe has joined #asahi-dev
21:44steinsgate has quit [Quit: Connection closed for inactivity]
21:46kaleryeblueberryoboe has quit [Remote host closed the connection]
21:58hdbngr has quit [Ping timeout: 480 seconds]
22:09gargantua has quit [Ping timeout: 480 seconds]
22:39cow123 has quit [Read error: No route to host]
22:41thrypuro[m] has joined #asahi-dev
22:41cow123 has joined #asahi-dev
22:46thrypuro[m]1 has joined #asahi-dev
23:01fl0id_oftc is now known as fl0_id_afk
23:16fl0_id_afk is now known as fl0id_oftc
23:16fl0id_oftc is now known as fl0_id_afk
23:16StanfordAppliedCyber[m]do we care what firmware we use for these devices as long as it supports the registers we need? i'm pretty sure the answer is no, just confirming that
23:32JamesCalligeros[m]what do you mean? we need to target specific versions because the ABI is unstable
23:33JamesCalligeros[m]thats why m1 and m2 series machines all install the macos 13.5 fw bundle
23:33StanfordAppliedCyber[m]i mean for m1 there's no reason to update to 26.5 fw because the hardware is the same regardless, it's just a new firmware abi that we have to RE for no real benefit
23:38JamesCalligeros[m]yeah pmuch, unless the 13.5 ipsw goes away at some point
23:41StanfordAppliedCyber[m]well, time to port my hv to 26.6 beta 2 and hope that fixes smp magically
23:41StanfordAppliedCyber[m]given we can write to that sysreg now this seems like a more appealing target anyways
23:44StanfordAppliedCyber[m]whoever made smp so cursed on my machine needs a stern talking to
23:48hdbngr has joined #asahi-dev
23:59hdbngr has quit [Ping timeout: 480 seconds]
…