#asahi-dev · 2026-08-18 20:18
jannaumaybe if the driver creates a domain and doesn't use the default domain?#asahi-dev · 2026-08-18 20:16
jannauI think it is not fully correct because 0x00x_xxxx_xxxx would be valid IOVA as well with that mask#asahi-dev · 2026-08-18 20:11
jannauso we would need to specify the dma_mask has hardware data in the device table#asahi-dev · 2026-08-18 20:09
jannauit works but requires `dma_set_mask_and_coherent(&pdev->dev, (1ULL << 40) | DMA_BIT_MASK(36));`#asahi-dev · 2026-08-18 20:05
jannauit does not work#asahi-dev · 2026-08-18 19:59
jannaulet me check. it would be a little annoying since it's a solution that only works for tbt#asahi-dev · 2026-08-18 19:47
jannaunot sure how everyone feels about inscreasing #iommu-cells#asahi-dev · 2026-08-18 19:47
jannauagreed, it feels like it should belong to the node which uses the dart#asahi-dev · 2026-08-18 19:43
jannauso "apple,dma-range = <0x100 0x0 0x1 0x0>;" seems to work#asahi-dev · 2026-08-18 19:42
jannauiova are as expected but it looks like acio has only a 36-bit address space (which still starts at 0x100_0000_0000#asahi-dev · 2026-08-18 19:25
jannauno, I suspect linux uses 0xeffff8000 and acio somehow offsets it by vm-base (0x100_0000_0000) anyway#asahi-dev · 2026-08-18 19:22
jannauwith "apple,dma-range = <0x100 0x0 0x2ff 0x0>;"#asahi-dev · 2026-08-18 19:22
jannauiirc it allocates top down so I would expect iovas of the form 0x3fe_xxxx_xxxx#asahi-dev · 2026-08-18 19:19
jannaunot sure linux' dma iommu code does the right thing#asahi-dev · 2026-08-18 19:18
jannau0x10effff8000 and 0x10effffc000#asahi-dev · 2026-08-18 19:16
jannaustill fails with that (and set_dma_mask_and_coherent(DMA_BIT_MASK(42)))#asahi-dev · 2026-08-18 18:57
jannauno, it probably needs "iommu: apple-dart: Support specifying the DMA aperture in the DT"#asahi-dev · 2026-08-18 18:52
jannaut6001 still works#asahi-dev · 2026-08-18 18:43
jannaudevicetree looks good, I assumed it was a t602x thing but should test if t600x regressed#asahi-dev · 2026-08-18 18:39
jannaudoesn't work though, I get a dart translatrion error for stream 1 and it fails to allocate host router due to a timeout#asahi-dev · 2026-08-18 18:33
jannauthe mailbox is correct and acio boots. was probably caused by the tunables mixup#asahi-dev · 2026-08-18 18:12
jannauit's 0x701108000#asahi-dev · 2026-08-18 18:12
jannauyes, looks like it#asahi-dev · 2026-08-18 18:10
jannauMMIO: W.4 0x28e08a02c (error-handler[2], offset 0xa02c) = 0x2, then poll until it reads 0, and then M3 ctrl = 0x2 (START=1)#asahi-dev · 2026-08-18 18:10
jannauI think your magic pmgr offset is correct#asahi-dev · 2026-08-18 18:01
jannauwhich was required for the old version#asahi-dev · 2026-08-18 18:01
jannau"kboot: Change name of acio alias" as local change#asahi-dev · 2026-08-18 18:01
jannaukinda my fault. My m1n1 still has your "#asahi-dev · 2026-08-18 17:54
jannaum1n1 logged "Unable to find alias usb4_1_acio" and the linux driver failed to load with missing tunables. I haven't looked further#asahi-dev · 2026-08-18 17:50
jannausure, do you have a good guess where it is or do I better trace pmgr completely?#asahi-dev · 2026-08-18 17:40
jannauoh, so we need to fix m1n1#asahi-dev · 2026-08-18 17:39
jannauuse https://gist.github.com/jannau/38035e43d862bf8a41b75573a5e227dd if github works#asahi-dev · 2026-08-18 17:38
jannauthey overlap. usb4_?_acio aliases are missing but there there is complete set of usb4-?-* aliases with "-"#asahi-dev · 2026-08-18 17:35
jannauresults in an hung irq handler https://paste.debian.net/hidden/59485152#asahi-dev · 2026-08-18 17:34
jannausven: m2 pro: "thunderbolt-apple-acio f01ac0000.cio: M3 RTKit failed to boot: -62" after the obvious incomplete fix for the aliases: https://paste.centos.org/view/68e73df7#asahi-dev · 2026-08-17 12:30
jannaumany parts of the driver hold references to the drm device when they ideally shouldn't#asahi-dev · 2026-08-17 12:28
jannauupstream gpuvm bindings and explicit device lifetime handling vs. years of technical debt#asahi-dev · 2026-08-17 12:22
jannaurebasing the gpu driver onto v7.2 is pain. And I pulled in the merge request for 7.3 as well. I suspect it will overall be less pain than doing v7.2 and v7.3 separately#asahi-dev · 2026-08-17 07:32
jannaui.e. the os specific part, i.e. iboot and dcp fw are version matched#asahi-dev · 2026-08-17 07:31
jannauisn't dcp firmware loaded by iboot2#asahi-dev · 2026-08-17 07:31
jannauyuka: iboot never handled dptx. on M1* with macOS 11 (which initialised the desktop HDMI port) it was handled in the dcp firmware. It wasn't that different from how the laptops worked#asahi-dev · 2026-08-16 12:59
jannauhave you already done the torture tests alyssa proposed in mesa against vk/opengl CTS?#asahi-dev · 2026-08-16 10:56
jannauI don't think this is too ugly and should make the shared-gpio-proxy work#asahi-dev · 2026-08-16 10:55
jannaujust change the flags in devm_gpiod_get_optional to GPIOD_OUT_LOW and potentionally set it high earlier if something expects to interact with it before tas27*_codec_probe.#asahi-dev · 2026-08-16 10:45
jannau"treat the output-high value as a reference counted, logical "GPIO enabled" setting, meaning: the GPIO value is set to high when the first user requests it to be high and back to low once the last user stops "voting" for high."#asahi-dev · 2026-08-16 10:44
jannaufrom the description of e992d54c6f970b382ffeacd7c88f68b94a3c6caf the shared gpio proxy should work#asahi-dev · 2026-08-16 10:40
jannauif avaoidable we should not use a reset controller for shutdown#asahi-dev · 2026-08-16 10:37
jannauyes, we want shutdown-gpios and not reset-gpios. I don't think the reset controler will work for this#asahi-dev · 2026-08-16 10:20
jannauthe tas27xx dt-bindings example even uses the gpio pin for reset and shutdown#asahi-dev · 2026-08-16 10:14
jannauhuh, both drivers also use shutdown-gpios and map this to sdz