#asahi-dev · 2026-05-31 14:20
noopwafeljust ignore atcphy entirely (and make sure XHCI is enabled in your kernel, which was my mistake).#asahi-dev · 2026-05-31 14:19
noopwafelI think you don't even need it to do that.#asahi-dev · 2026-05-31 14:01
yukaonly reset dwc3?#asahi-dev · 2026-05-31 14:01
yukaHow much do I need the atcphy to do for usb2?#asahi-dev · 2026-05-31 13:30
yuka(only the spmi tipd stuff and atcphy from noopwafel rebased on current asahi-wip)#asahi-dev · 2026-05-31 13:30
yukaand no changes to the dwc3 or apple-dwc3 code#asahi-dev · 2026-05-31 13:29
yukamy dwc3 device tree node looks pretty much identical to the t600x / t602x ones#asahi-dev · 2026-05-31 13:29
yukadwc3-apple 702280000.usb: error -EBUSY: can't request region for resource [mem 0x70228c100-0x70228ccff]#asahi-dev · 2026-05-31 10:46
chadmedthe read command is similarly weird, the readback from the display does not look very much like a ddc packet other than the data inside of it#asahi-dev · 2026-05-31 10:45
chadmedand one of the formats does not have all of the data one would expect#asahi-dev · 2026-05-31 10:45
chadmedi need to look at this more. this is definitely the thing we want but the write command has two different formats and there's nothing obvious that distinguishes which one should be used. they are different enough that there should be#asahi-dev · 2026-05-31 09:32
yuka*thx#asahi-dev · 2026-05-31 09:32
jannaufixed#asahi-dev · 2026-05-31 09:28
yukaarch/arm64/boot/dts/apple/t6034-j516m.dts has j514m compatible#asahi-dev · 2026-05-31 09:28
jannautypo in t6034-j516m.dts#asahi-dev · 2026-05-31 09:28
yukaYup#asahi-dev · 2026-05-31 09:27
yukaon my uboot-booted device /proc/device-tree/soc/fifo@2acb14000/input@0/multi-touch/firmware-name contains apple/tpmtfw-j516m.bin even though it should be j514m#asahi-dev · 2026-05-31 09:26
yukajannau: the non-working touchpad when using uboot is maybe caused by somehow the wrong device tree being used#asahi-dev · 2026-05-31 09:08
JamesCalligeros[m]this is actually kind of dumb, there are two message formats that encode basically the same thing but with different layout/padding #asahi-dev · 2026-05-31 08:43
JamesCalligeros[m]Ill clean up what i have and PR#asahi-dev · 2026-05-31 08:42
JamesCalligeros[m]No stress, this looks like it could be it or at least related#asahi-dev · 2026-05-31 08:41
jannauJamesCalligeros[m]: I don't remember and can't find traces#asahi-dev · 2026-05-31 08:27
jannaunot to my knowledge, I'm not sure why we have broken t6031 cpufreq in tree#asahi-dev · 2026-05-31 07:54
chaos_princesssame question for t6030#asahi-dev · 2026-05-31 07:53
chaos_princessjust to check - does anyone have working cpufreq on t6031?#asahi-dev · 2026-05-31 07:52
yukasnooze6214: also see my rambling from wednesday evening about the usb-repeater-i2c#asahi-dev · 2026-05-31 07:46
yukathis dump is from linux, so it's uninitialized (see the first bytes of 2c: 09 21 = via labs usb vendor id)#asahi-dev · 2026-05-31 07:45
yukasnooze6214: I don't know how to put the Neo into gadget mode#asahi-dev · 2026-05-31 07:03
snooze6214[m]yuka: Ah neat :0 did you manage to dump it when the neo was in gadget mode somehow?#asahi-dev · 2026-05-31 06:54
yukasee https://yuyu.yuka.dev/public/j700-usb-i2c.txt#asahi-dev · 2026-05-31 06:53
yukasnooze2614 I can confirm you can reset the hub and it responds to i2c afterwards. I found it easier to do it under linux using the appropriate /dev/gpiochip*.#asahi-dev · 2026-05-31 06:51
JamesCalligeros[m]if i squint this looks like raw-ish ddc (0x37, 0x51, then a bunch of crap)#asahi-dev · 2026-05-31 06:51
JamesCalligeros[m]jannau: would you by any chance recall if that i2c-ish bridge was at group 1 cmds 9/10 on dpavserv?#asahi-dev · 2026-05-31 05:54
snooze6214[m]ah, thanks Nick Chan #asahi-dev · 2026-05-31 05:48
NickChan[m]<snooze6214[m]> "in the device tree, i see the..." <- The gpio is not at that address. You must take the parent node's ranges property into account to obtain the address #asahi-dev · 2026-05-31 05:32
snooze6214[m]* in the device tree, i see the following function-reset = 76:GPIO(0x8e, 0x101), and gpio0 is at address 0x000000010A000000. is gpio pins toggled via mmio? trying to do any read32/mask32 in that region triggers an exception for me :')#asahi-dev · 2026-05-31 05:23
snooze6214[m]context: There is a usb-hub on the i2c bus at address 0x2c on the mb neo, but i all register reads on it timeout. i want to reset that hub, maybe that'll work?#asahi-dev · 2026-05-31 05:21
snooze6214[m]in the device tree, i see the following function-reset = 76:GPIO(0x8e, 0x101), and gpio0 is at address 0x000000010A000000. is gpio pins toggled via mmio? trying to do any read32 in that region triggers an exception for me :')#asahi-dev · 2026-05-31 01:39
yukaI split up the rvbar change into two PRs for easier review: https://github.com/AsahiLinux/m1n1/pull/585 (chainload) https://github.com/AsahiLinux/m1n1/pull/577 (smp)#asahi-dev · 2026-05-31 01:19
yukait works ok until one command is split up into two packets#asahi-dev · 2026-05-31 01:19
yukaAha! max_packet_size on the usb endpoint seems to be 512 incl. the kis header#asahi-dev · 2026-05-30 23:59
jannaumeasured upload speed around 10 mbps#asahi-dev · 2026-05-30 23:57
yuka(just means you have to kill deviceinterfaced)#asahi-dev · 2026-05-30 23:56
yukaThis means it's likely also more reliable to use kisd on Mac OS#asahi-dev · 2026-05-30 23:53
yukabut tldr as jannau pointed out, the broken chainloading was caused by data loss, and I found that limiting the usb packets to 256 characters works around this#asahi-dev · 2026-05-30 23:52
yukaat least this is progress :D#asahi-dev · 2026-05-30 23:52
yukaI can chainload m1n1 and linux to M3 Max and A18 Pro with debugusb#asahi-dev · 2026-05-30 23:51
yukathe changes are currently at https://github.com/AsahiLinux/kisd/pull/8, pending further testing#asahi-dev · 2026-05-30 23:50
yukaI made some improvements to kisd, which should make debugusb somewhat usable as primary uart for interacting with m1n1