#
✕ nick:nickchan
#asahi-dev · 2026-01-20 08:56
nickchanive seen the same register first referenced as acntrdir_el1 but later agtcntrdir_el1 and the encoding is the same
#asahi-dev · 2026-01-20 08:55
nickchanthough it seems that apple changed from A... to AGT...
#asahi-dev · 2026-01-20 08:51
nickchanBut if you run strings on a kernel then you can see it
#asahi-dev · 2026-01-20 08:51
nickchanfor example, agtcntrdir_el1 isn't anywhere in the public sources
#asahi-dev · 2026-01-20 08:45
nickchanthis is not all need to play with hw to know what the rest of the arch register clones are
#asahi-dev · 2026-01-20 08:45
nickchanhttps://github.com/apple-oss-distributions/xnu/blob/43a90889846e00bfb5cf1d255cdc0a701a1e05a4/osfmk/arm64/machine_routines.c#L2324
#asahi-dev · 2026-01-20 08:44
nickchaninstead of the arch ones
#asahi-dev · 2026-01-20 08:44
nickchanso for the other bits maybe go check their effect on *those* registers instead
#asahi-dev · 2026-01-20 08:43
nickchanShiz: 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
#asahi-dev · 2026-01-20 08:42
nickchanbut ofc that's redacted
#asahi-dev · 2026-01-20 08:42
nickchanit must setting some of the bits in the hypervisor
#asahi-dev · 2026-01-20 08:41
nickchanso if i had to guess bit[0] would have something to do with EL0
#asahi-dev · 2026-01-20 08:41
nickchanShiz: setting bit[1] sets the corresponding EL's timer to 24 MHz subject to VHE caveats
#asahi-dev · 2026-01-20 07:55
nickchani see you set them in the PR
#asahi-dev · 2026-01-20 07:54
nickchanseems like jannau figured them out
#asahi-dev · 2026-01-20 07:53
nickchanso what does bit[0] of AGTCNTRDIR_EL{1,2,12} do after all
#asahi-dev · 2026-01-18 16:06
nickchando some math and the UART_CLOCK can be recovered if it changed
#asahi-dev · 2026-01-18 16:05
nickchanwhat would be useful is reading the default UBRDIV when it's 115200
#asahi-dev · 2026-01-18 15:58
nickchanhttps://github.com/AsahiLinux/m1n1/blob/717054aaf02c5419cd59fe9c78bae65afda70713/proxyclient/m1n1/proxyutils.py#L543
#asahi-dev · 2026-01-18 15:58
nickchanyuka: it assumes the other side is already 1500000
#asahi-dev · 2026-01-18 15:56
nickchanyuka: the proxyclient does
#asahi-dev · 2026-01-18 15:55
nickchantry a lower baud then
#asahi-dev · 2026-01-18 12:21
nickchanACPW doesn't exist on t8015 so keeping the current check would be good
#asahi-dev · 2026-01-18 12:16
nickchani don't get the logic there
#asahi-dev · 2026-01-18 12:16
nickchanjannay: why must AC-n and AC-i be present if only ACPW is present
#asahi-dev · 2026-01-18 12:11
nickchanforgejo is decent
#asahi-dev · 2026-01-18 12:08
nickchanmaybe ML really isn't the best way to do code development, if only there's web UI integrated with git... oh wait
#asahi-dev · 2026-01-18 12:08
nickchanmaybe ML really isn't the best way to do code development, if only there's web UI integrated with git... oh wait
#asahi-dev · 2026-01-18 11:58
nickchanand i already found a better way of checking some keys instead of hardcoding compatible, in any case that's quite ready
#asahi-dev · 2026-01-18 11:57
nickchanif you want to see why i needed dynamic allocation in practice
#asahi-dev · 2026-01-18 11:57
nickchanbtw i have a working prototype now (absolutely not readying though with things like 5V1A hardcode): https://github.com/HoolockLinux/linux/blob/hoolock/drivers/power/supply/macsmc-power.c
#asahi-dev · 2026-01-18 11:54
nickchancause with static allocation can't really get away if there's ever any more vairance
#asahi-dev · 2026-01-18 11:53
nickchanmaybe I can send a reply
#asahi-dev · 2026-01-18 11:53
nickchancan technically get away with static allocation if just for M1/M2/M3 and decrementing num_properties but that way no new property can be added
#asahi-dev · 2026-01-18 10:35
nickchanI mean I use that a lot in m1n1
#asahi-dev · 2026-01-18 10:35
nickchanor are you talking about linux
#asahi-dev · 2026-01-18 10:35
nickchanjannau: huh? I use 1.5M uart a lot
#asahi-dev · 2026-01-17 15:43
nickchanokay i will use 64k paging in uboot so i can ship only 1 bin
#asahi-dev · 2026-01-17 14:47
nickchanworks on 11.2
#asahi-dev · 2026-01-17 14:47
nickchanthere is
#asahi-dev · 2026-01-17 14:45
nickchanand yes this means DTK's rosetta is cursed
#asahi-dev · 2026-01-17 14:45
nickchanthen in a14 4k paging is added again
#asahi-dev · 2026-01-17 14:45
nickchan4k paging isn't support from a9-a13
#asahi-dev · 2026-01-17 14:41
nickchanturns out it isn't too hard to add 16K and 64K paging to uboot
#asahi-dev · 2026-01-17 14:34
nickchanShiz: was thinking about bigger chips that has 3 clusters or more
#asahi-dev · 2026-01-17 12:36
nickchanon m1
#asahi-dev · 2026-01-17 12:36
nickchanand uboot has dart translation support so that's working with 4K page on the CPU MMU and 16K IOMMU
#asahi-dev · 2026-01-17 12:32
nickchanbut it will not work on hardware that actually does not have 4K paging
#asahi-dev · 2026-01-17 12:32
nickchana mircale that works on m1
#asahi-dev · 2026-01-17 12:32
nickchanof course uboot assumes 4K everywhere

plus ›