#asahi-dev · 2026-03-02 07:44
jannaueasiest way to fix this is probably to set the current CTM (or identity if unset) when enabling the crtc / on the first swap#asahi-dev · 2026-03-02 07:04
jannauI'd expect clients are not allowed to change properties while the crtc is disabled. I don't think this is what happens though. I suspect kwin sets the CTM to identity when it disables the crtc and that gets ignored#asahi-dev · 2026-03-01 23:04
jannauuefi-only build and deply to bunny CI for asahi-installer: https://github.com/AsahiLinux/asahi-installer/pull/414#asahi-dev · 2026-03-01 21:12
jannauno, Apple is lying in the ADT#asahi-dev · 2026-03-01 16:44
jannauI've pushed a v6.19.5 based tree to asahi-wip, not tagged. please test#asahi-dev · 2026-03-01 14:14
jannauadding explicit tests for this would be a good idea. I have macos code which reports the metadata value for uncompressed tiles so creating framebuffers with arbitrary content should be possible#asahi-dev · 2026-03-01 14:08
jannauchadmed: is interchange used for the cursor overlay? I don't think anything I spotted in the dcp PR should cause this#asahi-dev · 2026-03-01 14:06
jannauoliver: stay safe#asahi-dev · 2026-03-01 10:34
jannauchadmed: I think oliver tests on t6000 or t6001 so not that surprising#asahi-dev · 2026-03-01 10:30
jannauI've pushed `asahi-6.18.15-1` to github (plain rebase onto v6.18.15)#asahi-dev · 2026-02-28 20:24
jannauand a unaligned u32 type because the struct has a single u16 in the middle#asahi-dev · 2026-02-28 20:22
jannauseems to work with 1 additional byte to make the size a multiple of 4#asahi-dev · 2026-02-28 20:02
jannauI assumed we can't change PDMConfig as that would change the lenght of the epic message#asahi-dev · 2026-02-28 20:00
jannauI can't count. I counted a 5th u8 in the middle of DecimatorConfig#asahi-dev · 2026-02-28 19:53
jannauchaos_princess: how sure are you about the layout of aop's PDMConfig and DecimatorConfig? (pin_)init with unaligned fields is unsound. Using repr(C, packed) end up with an alignment of 1 as the structs have an odd size#asahi-dev · 2026-02-28 17:22
jannauilyaas_: are you using gnome?#asahi-dev · 2026-02-28 14:45
jannauah, I deleted a random line accidentally#asahi-dev · 2026-02-28 13:34
jannau"?#asahi-dev · 2026-02-28 13:34
jannauhow did I manage to break aop in v6.19: "apple_aop 24ac00000.aop: Attempting to send message with no txbuf at endpoint 38#asahi-dev · 2026-02-22 09:27
jannausiirko[m]: use raw binaries for run_guest.py as well. i.e. m1n1.bin and `run_guest.py -r`#asahi-dev · 2026-02-21 21:53
jannausiirko[m]: then use m1n1's hypervisor. boot bundles of m1n1.bin + dtbs + gzipped u-boot-nodtb.bin with tools/run_guest.py#asahi-dev · 2026-02-21 21:42
jannauI use `tio -m INLCRNL,OCRNL -b 500000 /dev/m1n1-sec` instead of picocom-sec.sh. can deal on its own with absent device nodes#asahi-dev · 2026-02-21 21:41
jannauin that case ACM1 as provided by m1n1 over USB will be used#asahi-dev · 2026-02-21 21:40
jannauthere won't be anything on the m1n1 provided cdc acm ports unless you boot u-boot under HV#asahi-dev · 2026-02-21 21:39
jannausiirko[m]: it will log on hw uart, so on another mac on ttySAC (connect to dfu port on both system) or the central scrutinizer one#asahi-dev · 2026-02-21 14:58
jannauplease do not edit messages. it's very annoying on irc side#asahi-dev · 2026-02-20 22:16
jannaumemory map comes mostly from the ranges node in the ADT and limiting sizes for pcie/usb#asahi-dev · 2026-02-20 22:14
jannauI haven't yet added the new "apple,t8103-*" base compatibles to u-boot so those have to be added as well#asahi-dev · 2026-02-20 22:13
jannausiirko[m]: not that demanding. I think it's mostly adding new compatibles and a new mem-layout in arch/arm/mach-apple/board.c since m3's physical mem also starts at 1 << 40#asahi-dev · 2026-02-20 21:16
jannauthe author date of that commit is "Date: Mon Apr 17 20:41:13 2023 +0900"#asahi-dev · 2026-02-20 19:40
jannausiirko[m]: yes, u-boot needs explicit m3 support#asahi-dev · 2026-02-20 09:50
jannauinitramfs shouldn't be necessary when everything is buit-in (apple nvme and sart for nvme#asahi-dev · 2026-02-15 16:23
jannauand bits/040-dwc3 is empty on top of v6.19#asahi-dev · 2026-02-15 11:42
jannaumerged,, let's see if that works#asahi-dev · 2026-02-15 11:42
jannauI made an uefi-only payload with with the 6.18 devicetrees. That one has to be manually uploaded. Do we want to change that as well? Build is in https://github.com/AsahiLinux/m1n1/actions/runs/22034235279/job/63664361797 (not yet tested)#asahi-dev · 2026-02-15 11:38
jannaumade the secrets even more opaque so it won't be even visible that it is missing#asahi-dev · 2026-02-15 11:36
jannauhttps://github.com/AsahiLinux/asahi-installer/pull/410#asahi-dev · 2026-02-15 11:03
jannauI'm in the progress of fixing it#asahi-dev · 2026-02-15 11:03
jannauyes#asahi-dev · 2026-02-15 10:55
jannauhow does github render secrects in the log. I would expect something different that an empty string if the secret exists: `-H "AccessKey: " \` #asahi-dev · 2026-02-15 10:50
jannauI've merged the installer push workflow but upload failed: https://github.com/AsahiLinux/asahi-installer/actions/runs/22034335855/job/63664639745#asahi-dev · 2026-02-14 20:12
jannauI've tagged and pushed asahi-6.18.10-1#asahi-dev · 2026-02-14 09:16
jannauat least if the thing driving it is in the SoC#asahi-dev · 2026-02-14 09:14
jannaudidn't marcan suspect the eusb repeater is there as level shifter? so it would be required in the debugusb path as well#asahi-dev · 2026-02-14 09:09
jannaubtw I started working on usb plug detection / re-init in m1n1 so that's possible to switch from debugusb to to usb device mode/cdc_acm#asahi-dev · 2026-02-14 09:07
jannauhave you tried switching modes? but isn't the eUSB repeater the problem#asahi-dev · 2026-02-13 21:30
jannauinstall from OTA is committed: 2c5ab624030b ("stub, util: Support installing from OTA images")#asahi-dev · 2026-02-13 21:24
jannaudid marcan finish that? at some point we discovered that the dcp bug wasn't fixed in the OTA#asahi-dev · 2026-02-13 21:22
jannaufirmware is likely identical though but the installer needs the ipsw. the 14.6 install-assistent isn't available anymore from apple :(#asahi-dev · 2026-02-12 18:51
jannaujust have $SOMEONE ask for it in the next 10 days, standard terms apply ;)