#asahi-dev · 2026-05-28 08:28
jannauI'd continue for now, a working driver is better than none. From the working driver either support can be added (there are other platforms with v4l2 drivers) or it can be used as base to rewrite the driver#asahi-dev · 2026-05-28 07:52
jannausnooze6214[m]: usb device mode doesn't work on the neo since there's a usb2 hub (there ought a way to bypass the hub but we don't know how)#asahi-dev · 2026-05-23 22:48
jannaucpufreq and M3 Pro/Max pcie are missing for bare minimal M3* support#asahi-dev · 2026-05-23 22:02
jannauu-boot stores the initialization packets and writes them into the remote (rmt) fifo so that linux will be able to read them#asahi-dev · 2026-05-23 22:01
jannausame on m3*. it's broken after initializing it in u-boot and then booting linux#asahi-dev · 2026-05-23 21:53
jannauno, this is about mtp dockchannel and explains the broken hand-over from u-boot to linux#asahi-dev · 2026-05-23 21:30
jannaus/around //#asahi-dev · 2026-05-23 21:30
jannaudid apple switch the local / remote dockchanel mmio regions on m3around ?#asahi-dev · 2026-05-23 20:33
jannauerr, missing 0x1_0000_0000#asahi-dev · 2026-05-23 20:32
jannau+ 0x2_1000_0000#asahi-dev · 2026-05-23 14:24
jannauyes, no pcie for t603x#asahi-dev · 2026-05-23 14:14
jannauno, but HPM were i2c devices and the interface smells like macos could have an i2c in userspace API#asahi-dev · 2026-05-23 13:21
jannaumight not even need a kext, https://github.com/AsahiLinux/macvdmtool/blob/main/AppleHPMLib.h#asahi-dev · 2026-05-23 13:08
jannauhttps://www.via-labs.com/product_show.php?id=126#asahi-dev · 2026-05-23 13:07
jannauany information on the hub's maker and model?#asahi-dev · 2026-05-23 13:02
jannaufl0_id: I did already ^^^#asahi-dev · 2026-05-23 13:01
jannauI've tagged and pushed asahi-7.0.10-1 to github. rust builds will now fail without toolchain and additional m3 pro/max hardware support: nvme, mtp and smc#asahi-dev · 2026-05-21 22:01
jannauand freedesktop-sdk-25.08.12 is released with non-broken mesa. 24.08 still has mesa-26.0.5. hopefully not used by most frequently used flatpaks#asahi-dev · 2026-05-21 21:48
jannauI added spmi-pmic, smc, mtp and nvme device nodes for M3 Pro/Max/Ultra to asahi-wip#asahi-dev · 2026-05-21 17:42
jannaufl0_id: do you have a m3 max device tree with more hw than in asahi-wip (not pcie but smc, mtp and or nvme)#asahi-dev · 2026-05-21 15:47
jannauM3* installer support https://github.com/AsahiLinux/asahi-installer/pull/424 (not usable on its own yet)#asahi-dev · 2026-05-21 10:18
jannaubacklight brightness control was always a DCP feature (and not from the beginning), GPIO was only on/off#asahi-dev · 2026-05-20 18:53
jannaukettenis: I'm adding M3 Pro, Max and Ultra mem maps to u-boot and mach-apple/board.c gets to large for my taste. Do you have a preference on to organize them. I'm thinking of moving each of them to their own file like "mem_maps/t8103.h"#asahi-dev · 2026-05-20 17:36
jannauI'd guess median patch count for non LTS stable releases is 200 - 400#asahi-dev · 2026-05-20 17:34
jannauI suppose this is in part the higher fixes workload due to "AI" tools#asahi-dev · 2026-05-20 17:28
jannau1145! I'm only cc-ed on this single one#asahi-dev · 2026-05-20 17:23
jannaufun: "[PATCH 7.0 0574/1146] phy: apple: apple: Use local variable for ioremap return value" from the 7.0.10 stable reviews#asahi-dev · 2026-05-18 21:23
jannauI got packages but there wasn't interesting in the payload. get descriptor packages for the "debug usb" string where captured#asahi-dev · 2026-05-18 21:19
jannauI hadn't luck with spotting anything in wireshark usb captures on macos#asahi-dev · 2026-05-18 21:18
jannauyuka: works#asahi-dev · 2026-05-18 20:09
jannaum4 pro I get the m1n1 log but iface.nop() still fails#asahi-dev · 2026-05-18 20:01
jannausame issue with a clean m1n1 build#asahi-dev · 2026-05-18 19:55
jannauit works partly on my m3, I get the iboot boot log but iface.nop() fails. m1n1 is new enough but might contain random hacks#asahi-dev · 2026-05-18 19:47
jannauah, it helps if m1n1 on the target is new enough to have dockchannel uart#asahi-dev · 2026-05-18 19:41
jannauis the product id the same for all macs in debug usb (0x1881)? we probably want https://paste.centos.org/view/2f5e2933#asahi-dev · 2026-05-18 19:12
jannauI've tagged asahi-7.0.9-1 (just a rebase)#asahi-dev · 2026-05-17 20:33
jannauusing 32 bit reads does not help#asahi-dev · 2026-05-17 19:53
jannauno, using a large timeout does not work, it doesn't even read 25% of the sent data#asahi-dev · 2026-05-17 19:48
jannauthe ~170ms wait happens every 0x800 bytes#asahi-dev · 2026-05-17 19:18
jannaualso the other direction is fast. my initial benchmark was memread() which was ~6mbps iirc#asahi-dev · 2026-05-17 19:17
jannaularge pauses seem evenly spaced. I can't imagine that it so slow that it can't even reach 115200bps#asahi-dev · 2026-05-17 19:14
jannauI'd expect chainload to work with a very large timeout#asahi-dev · 2026-05-17 19:08
jannaunot sure why it would do that. maybe there are some fifo maintenance ops required#asahi-dev · 2026-05-17 19:06
jannaunot sure about that. the chainload issue I've seen was dockchannel_uart_getbyte() spending hundreds of milliseconds in the DATA_RX_COUNT loop every 2048 byte#asahi-dev · 2026-05-17 13:48
jannauthere are subtle differences in atcphy. tryting to hack around them hasn't change behavior#asahi-dev · 2026-05-17 13:45
jannautrace diff against macos looks ok-ish. hard to say since macos does on connect a USB3 -> USB3_DP transition#asahi-dev · 2026-05-17 13:40
jannauhttps://paste.debian.net/hidden/1cae8b10#asahi-dev · 2026-05-17 13:35
jannausven: "System FW version: unknown (mBoot-18000.120.36)" I assume that is 26.5#asahi-dev · 2026-05-17 13:30
jannauhttps://paste.debian.net/hidden/a67d4eef#asahi-dev · 2026-05-17 13:29
jannaumaybe just newer ACIO FW with more logging? I have "AppleCIOFirmware-475~290__2026-04-18-18:18" despite the latest installed macOS being 13.5