#
✕ nick:nickchan
#asahi-dev · 2024-09-12 09:35
nickchanedit: swap a7 and desktop chips above
#asahi-dev · 2024-09-12 09:34
nickchanalphabetical A9 -> (desktop chips) -> A7 -> A8 -> A10 isn't very readable for the asahi-soc reviwer, I think
#asahi-dev · 2024-09-12 09:33
nickchanim more inclined to have a logical order
#asahi-dev · 2024-09-12 09:32
nickchanthese changes are to be merged through asahi-soc i think
#asahi-dev · 2024-09-12 09:31
nickchanv2 will be this weekend
#asahi-dev · 2024-09-12 09:28
nickchanI think you should make a reply on the mailing list
#asahi-dev · 2024-09-12 09:28
nickchansven: Well, the thing is the existing entries are *not* consistent
#asahi-dev · 2024-09-12 02:25
nickchannicolas17: chassis-type is a relatively recent thing
#asahi-dev · 2024-09-11 18:46
nickchanhttps://lore.kernel.org/asahi/a3849bc0-000a-4290-b2c3-71a50a750495@gmail.com
#asahi-dev · 2024-09-11 18:46
nickchanCan we get a clarification on how Asahi-related devicetree bindings are supposed to be sorted? It seems to be a mix of alphabetical and logical (SoC type and generation) sorting right now.
#asahi-dev · 2024-09-11 15:32
nickchanwell DT spec accepts github PRs so that's easy!
#asahi-dev · 2024-09-11 14:28
nickchanI think for Apple TV a new chassis-type is actually needed
#asahi-dev · 2024-09-11 14:11
nickchanwell, it was added to the makefile then removed by the commit to add t8011
#asahi-dev · 2024-09-11 13:59
nickchant8010-n112.dts not being in the makefile and missing an include
#asahi-dev · 2024-09-11 13:58
nickchan... right stuff went wrong it really isn't possible to get things right the first type
#asahi-dev · 2024-09-11 13:32
nickchanlike apple tv
#asahi-dev · 2024-09-11 13:32
nickchanI guess I can use embedded when nothing else fits
#asahi-dev · 2024-09-11 13:32
nickchanchassis-type: desktop, laptop, convertible, server, tablet, handset, watch, embedded
#asahi-dev · 2024-09-11 13:31
nickchani mean homepage worked fine for the key at the bottom of the samsung galaxy a7
#asahi-dev · 2024-09-11 13:30
nickchanahaha then those wouldn't fit and KEY_HOMEPAGE should be used
#asahi-dev · 2024-09-11 13:29
nickchanbut Apple TV ?
#asahi-dev · 2024-09-11 13:29
nickchannow for iPhones, iPads and iPod touches it is pretty obvious
#asahi-dev · 2024-09-11 13:28
nickchanthere's also the issue of chassis-type
#asahi-dev · 2024-09-11 13:22
nickchanand that seems to be best fit so far
#asahi-dev · 2024-09-11 13:22
nickchanarch/arm64/boot/dts/qcom/msm8939-samsung-a7.dts uses KEY_HOMEPAGE
#asahi-dev · 2024-09-11 13:21
nickchanchadmed: well there's KEY_HOME, KEY_HOMEPAGE and KEY_MENU
#asahi-dev · 2024-09-11 11:43
nickchanit's probably better to assign KEY_MENU to home button, not KEY_LEFTMETA
#asahi-dev · 2024-09-11 11:15
nickchanI suppose the summit driver is for backlight control brightness/on/off
#asahi-dev · 2024-09-11 11:13
nickchananything lower is different though
#asahi-dev · 2024-09-11 11:13
nickchana11 has a "lcd,summit", i should see if it works with the existing driver
#asahi-dev · 2024-09-11 09:51
nickchanthese DT patches are for 6.13, sending them early because with so many things to add here something is bound to go wrong
#asahi-dev · 2024-09-11 09:50
nickchanoops
#asahi-dev · 2024-09-06 10:44
nickchansven: can you check if the last two commits in the above branch affects serial on M1 (or above)?
#asahi-dev · 2024-09-06 10:43
nickchanhttps://github.com/asdfugil/linux-apple/commits/apple/bits/010-serial
#asahi-dev · 2024-09-06 10:42
nickchanhowever, I do not know if M1 would be happy about that
#asahi-dev · 2024-09-06 10:42
nickchannow setting these bits doesn't seem to break A10
#asahi-dev · 2024-09-06 10:42
nickchanthey seem to use bit 3 in UTRSTAT as RXTO, which isenabled by bit 11 in UCON
#asahi-dev · 2024-09-06 10:41
nickchanfixed serial rx for a7-a9
#asahi-dev · 2024-09-05 15:49
nickchangoing to add an override in system_supports_32bit_el0
#asahi-dev · 2024-09-05 15:18
nickchanbecause of the msr emulation that does need to be changed
#asahi-dev · 2024-09-05 15:17
nickchanwhat's needed is changing the sanitized value from the pfr0 register
#asahi-dev · 2024-09-05 15:12
nickchanwait no that functions just says a given pfr0 value has 32 bit el0
#asahi-dev · 2024-09-05 15:10
nickchanwhich reads the register value with the function id_aa64pfr0_32bit_el0() also in cpufeature.h
#asahi-dev · 2024-09-05 15:10
nickchansystem_supports_32bit_el0() in arch/arm64/include/asm/cpufeature.h does
#asahi-dev · 2024-09-05 15:09
nickchanmaz: has_32bit_el0() does not control availability of 32-bit execution
#asahi-dev · 2024-09-05 14:46
nickchannone of Apple's arm64 chips supported aarch32 EL1 so yeah it is not needed
#asahi-dev · 2024-09-05 14:39
nickchanmaz: it has EL0 and EL1
#asahi-dev · 2024-09-05 14:13
nickchanmaz: was trying to edit arch/arm64/kernel/cpu_errata.c but I guess not
#asahi-dev · 2024-09-05 13:43
nickchanhigh powered p-states is ok, low ones aren't
#asahi-dev · 2024-09-05 13:42
nickchanright the issue is basically the core can only exec 32 bit in certain p-states

plus ›