#
#asahi-dev · 2026-06-18 14:57
sveni only care about M1+ here
#asahi-dev · 2026-06-18 14:57
nickchani believe the nGnRnE requirement is new in A14/M1
#asahi-dev · 2026-06-18 14:54
mazyou 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).
#asahi-dev · 2026-06-18 14:53
svenneeds more invasive patches because that's not how efi runtime services are supposed to work
#asahi-dev · 2026-06-18 14:52
chaos_princesswhy not jump into m1n1 with mmu off?
#asahi-dev · 2026-06-18 14:51
mazyup, I expected something of the sort. seems relatively minor, in the grand scheme of things.
#asahi-dev · 2026-06-18 14:50
sven+ prot = __pgprot(PROT_DEVICE_nGnRnE);
#asahi-dev · 2026-06-18 14:50
sven+ if (of_machine_is_compatible("apple,arm-platform"))
#asahi-dev · 2026-06-18 14:50
svenyeah, right now i have this hack here but that's something we can bikeshed on the mailing list
#asahi-dev · 2026-06-18 14:49
mazguess we'd have to each the EFI code in the kernel to generate "special" mappings on Apple HW...
#asahi-dev · 2026-06-18 14:48
svenso it's just the usual 1:1 mapping
#asahi-dev · 2026-06-18 14:48
svenno, you can define special EFI runtime mappings that the kernel enables before jumping to m1n1 again
#asahi-dev · 2026-06-18 14:47
chaos_princessare you using kernel memory mappings?
#asahi-dev · 2026-06-18 14:30
svenrequires a rather hacky kernel patch because there's no way to express "we really need Device-nGnRnE" in the memory table
#asahi-dev · 2026-06-18 14:30
svenyup
#asahi-dev · 2026-06-18 14:30
jannaum1n1 writes directly to the uart inside the psci call? \o/
#asahi-dev · 2026-06-18 14:16
svenit's just a nop for now but the kernel -> m1n1 -> kernel calls work now and the stub can still access all mmio
#asahi-dev · 2026-06-18 14:16
svenprogress \o/
#asahi-dev · 2026-06-18 14:15
sven[ 5.090406] CPU1: failed to come online
#asahi-dev · 2026-06-18 14:15
sven[ 0.026035] psci: calling into PSCI-via-EFI done
#asahi-dev · 2026-06-18 14:15
svenhello, world from m1n1!
#asahi-dev · 2026-06-18 14:15
sven[ 0.026035] psci: calling into PSCI-via-EFI now
#asahi-dev · 2026-06-18 14:15
sven[ 0.025462] smp: Bringing up secondary CPUs ...
#asahi-dev · 2026-06-18 14:00
chaos_princessunfortunately, dcs still can't go as low as macos does, we can only do 8, while macos does 6
#asahi-dev · 2026-06-18 13:53
jannaugood, not sure what to do about it
#asahi-dev · 2026-06-18 13:46
chaos_princessthat was it
#asahi-dev · 2026-06-18 13:39
jannauchaos_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
#asahi-dev · 2026-06-18 11:36
chadmedyuka: https://github.com/AsahiLinux/docs/pull/269
#asahi-dev · 2026-06-18 09:01
chaos_princessdisplay still can't survive pmgr_misc even with min-state
#asahi-dev · 2026-06-18 08:57
sofusthat would kinda make sense
#asahi-dev · 2026-06-18 08:56
JamesCalligeros[m]Would expect buf and not bfr but who knows
#asahi-dev · 2026-06-18 08:56
JamesCalligeros[m]Buffer?
#asahi-dev · 2026-06-18 08:56
sofuswhat about BFR? There is always one DMA_BFR_CONFIG with the same name after an DMA_CONFIG
#asahi-dev · 2026-06-18 08:51
sofusits one of the things they are inconsistent with
#asahi-dev · 2026-06-18 08:50
JamesCalligeros[m]Read and write
#asahi-dev · 2026-06-18 08:50
chaos_princessread/write?
#asahi-dev · 2026-06-18 08:50
sofusits actually all dma config, any clue what RD and WR could mean?
#asahi-dev · 2026-06-18 08:47
JamesCalligeros[m]Yeah
#asahi-dev · 2026-06-18 08:47
chaos_princessmy take is "this is probably fine in this case"
#asahi-dev · 2026-06-18 08:47
chaos_princesswell, AVC_MACROBLOCK_SIZE is also probably fine
#asahi-dev · 2026-06-18 08:46
JamesCalligeros[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
#asahi-dev · 2026-06-18 08:45
chaos_princessyou can make an argument that the name is someone's interpretation of how to write the hw description in a concise way
#asahi-dev · 2026-06-18 08:44
sofusthanks, the names are super inconsistent, some tunatables equal, but one or two names are different
#asahi-dev · 2026-06-18 08:44
chaos_princess... debatable
#asahi-dev · 2026-06-18 08:36
JamesCalligeros[m]Register names that tell you what the register does are facts about the hardware and not novel/original code
#asahi-dev · 2026-06-18 08:36
JamesCalligeros[m]NAL but yes “facts about the hardware” are not copyrightable
#asahi-dev · 2026-06-18 08:36
nickchansee marcan's explaination: https://lore.kernel.org/asahi/3cd6bf37-f335-470b-9b3f-25628067e0f1@marcan.st/
#asahi-dev · 2026-06-18 08:35
nickchanit's not any different from the register names copied from proprietary datasheet in all places in linux
#asahi-dev · 2026-06-18 08:35
sofusnice
#asahi-dev · 2026-06-18 08:34
nickchanthis should be fine

plus ›