#asahi · 2026-11-15 19:15
jannauI thought I did the same `#ifdef USE_PMP` for the additional dcpext#asahi · 2026-11-15 13:02
jannaufedora doesn't package asahi-va-driver yet and I think avd-fw isn't yet available to users#asahi · 2026-11-13 18:32
jannauvrfm: M1 and M2 support only 36-bit PARange, the VM's IPA size is derived from this, see https://gitlab.com/qemu-project/qemu/-/commit/1ab445af8cd99343f29032b5944023ad7d8edebf for reference#asahi · 2026-11-13 10:22
jannauthat means M6 is < Armv9.6. "From Armv9.6, the value 0b0001 is not permitted."#asahi · 2026-11-13 10:22
jannauchaos_princess: ID_AA64ISAR2_EL1: … CSSC=0x1 …, so feat_cssc is supported but feat_cmpbr is not. https://gist.github.com/jannau/d96eb2fbd778f911ca039d50f81cf5a6#asahi · 2026-11-13 10:09
jannauI'll check#asahi · 2026-11-13 10:08
jannauwikipedia claims M5 is ARMv9.2-A (no information for M6)#asahi · 2026-11-13 10:08
jannauI don't I've looked at the feature registers yet. determining the exact instruction set version from that is also annoying#asahi · 2026-11-09 15:19
jannauon fedora yes#asahi · 2026-11-09 15:18
jannauthe symlink might be required for update-m1n1 though. on fedora it picks the dtbs from that symlink#asahi · 2026-11-09 09:37
jannauah, that just old information. avd works know on all SoCs to the same extent#asahi · 2026-11-09 09:32
jannausinan[m]: m1 / m2 pro have the same level of support. M3 devices have an av1 decoder, which might be what you're refering to#asahi · 2026-11-09 09:31
jannauPate[m]: I would say no but without knowing which dock this is exactly it's hard to say anything. I assume it is a Revodok #asahi · 2026-11-09 09:25
jannauPate[m]: Startech wasn't a recommendation for you#asahi · 2026-11-08 20:39
jannaukitlith: if available you can also try to charge with different usb-c charger and cable. the usb-c ports can be used for charging as well. phone usb-c charger will work but you need to wait a few minutes before turning the macbook on#asahi · 2026-11-08 16:13
jannauusb and lan on these devices will need PCIe tunnelling since they are thunderbolt3 with an ASMedia ASM1042A xHCI#asahi · 2026-11-08 16:12
jannauthere seems to be a HDMI variant as well https://www.startech.com/en-de/docking-stations-multiports/tb3dkm2hdl#asahi · 2026-11-08 16:11
jannaupossibly https://www.startech.com/en-de/docking-stations-multiports/tb3dkm2dpl with extra LAN and USB3 5Gbps port. The first is tested, the second is not but from description it should use 2 DP tunnels as well#asahi · 2026-11-08 16:06
jannauworkaround would a dock with thunderbolt/usb4 downstream port(s) and https://www.startech.com/en-de/display-video-adapters/tb32dp14#asahi · 2026-11-08 10:47
jannaut8122 will be annoying as we still haven't figured out why the boot core is much slower#asahi · 2026-11-08 09:43
jannaucan you give more details about your hubs? Ideally they should be supported if they work under macos or other linux PCs unless they require PCIe tunnelling over thunderbolt/usb4#asahi · 2026-11-08 09:40
jannauhard to do recommendations because there are just too many and I wouldn't even bet that nameguarantees it has identical components inside#asahi · 2026-11-08 09:38
jannauthunderbolt3 docks and other usb4/tb4 docks probably will start soon-ish (likely within weeks to months)#asahi · 2026-11-08 09:37
jannauusb3 hubs (usually less expensive) with network should work. thunderbolt4 / usb4 hubs/docks might have working usb3 and ethernet ports (if the ethernet nic is USB3) with 7.2 kernels#asahi · 2026-11-07 11:57
jannauavd-fw 0.2 will work for now#asahi · 2026-11-07 11:56
jannauAnEagle[m]: it was intended as "no backwards compatibility to avd-fw before 0.2"#asahi · 2026-11-04 11:14
jannauthat's unrelated to hardware support, just a dependency of installed software #asahi · 2026-11-03 23:09
jannauthere are other interfaces to energy/power counters than smc which we don't support though#asahi · 2026-11-03 23:07
jannaus/right/expected/#asahi · 2026-10-29 13:28
jannaulook at arch/arm64/configs/asahi.config you probably need CONFIG_RESET_APPLE_CIO=m and CONFIG_USB4_APPLE_SOC=m#asahi · 2026-10-29 13:06
jannaueasier solution might be to update the kernel. 7.2 is available since yesterday#asahi · 2026-10-29 13:05
jannaunot directly related to initramfs but to the kernel you want to boot. running update-m1n1 is most likely not enough. the uses the DTBs in /boot/dtb (a symlink). If that still points to the new kernel it won't help#asahi · 2026-10-29 12:59
jannauthe fairydust kernel used is based on 7.2? not aware that any DTS change should affect internel keyboard though#asahi · 2026-10-29 12:51
jannaudoes USB work on a non-fairydust kernel?#asahi · 2026-10-29 12:51
jannauit is USB 3.2 Gen2 Type-C#asahi · 2026-10-28 08:35
jannauthat was years ago so might be worth retesting. not sure I expect changes given that kde hasn't yet switched to upower for managing the thresholds (that would get rid of the authentication for changing them)#asahi · 2026-10-28 08:31
jannaucharge_control_start_threshold is broken. it is only there since some software (iirc kde) expected charge_control_end_threshold and charge_control_start_threshold#asahi · 2026-10-28 08:29
jannauI think charge_control_end_threshold is ABI wise ok. The driver is alloved to not support arbitrary values and is expected to round to next supported one#asahi · 2026-10-28 08:24
jannauwe do not want to get rid of it (the functionality)#asahi · 2026-10-26 19:39
jannauthe dark display during boot is fixed in asahi-7.2.12-3#asahi · 2026-10-24 21:55
jannauI've tested on the 13-inch mbp m2. check whether CONFIG_RESET_APPLE_CIO and CONFIG_USB4_APPLE_SOC are enabled (as module)#asahi · 2026-10-24 21:38
jannauNiklasReimer[m]: which device?#asahi · 2026-10-14 22:03
jannauyour data is unrecoverable though. it is not erased from the flash but it's impossible to restore. the SSD controller writes data encrypted with a key which is supposedly not extractable from the hardware#asahi · 2026-10-14 22:01
jannaubtrfs defaults to "discard=async" which issues trim. it doesn't actually erases data in the flash cells but informs the SSD controller that the bloacks are unused. reads will always return 0 without accessing any flash#asahi · 2026-10-14 10:21
jannauno m5 imacs and m6 imacs are rumored to be announced in 2 weeks#asahi · 2026-10-12 12:56
jannauand restore your backups#asahi · 2026-10-08 07:19
jannauMichaelReeves[m]: karma in https://bodhi.fedoraproject.org/ is for individual package updates#asahi · 2026-10-08 07:18
jannauAnEagle[m]: what do you mean by really low resolution? A possible change is returning to 2x scaling for retina displays#asahi · 2026-10-08 07:17
jannaueruditehermit: should be fine if you want to use pre-release software (more release candidate quality at this point)#asahi · 2026-10-07 14:58
jannauwe were already discussing this in the the fedora asahi matrix channel. let's continue there, especially when posting screenshots