#asahi-dev · 2022-11-17 22:36
robherjannau: no serial port on MBP 14 though?#asahi-dev · 2022-11-17 22:33
robherhow does one get the kernel log early? 'earlycon=efi' doesn't seem to work or I'm dying before earlycon is up.#asahi-dev · 2022-11-16 18:08
robhersven: phys tend to have shorter lifetimes given they are somewhat tied to process nodes, so probably less value in using "apple,atcphy" than it would be for a purely digital IP block.#asahi-dev · 2022-10-19 22:09
robherI'm guessing fw_devlink is not handling iommu-map dependency nor is deferred probe working for it.#asahi-dev · 2022-10-19 22:08
robherNo splats and wifi works now with dart built-in!#asahi-dev · 2022-10-19 21:58
robherOF: /soc/pcie@590000000: no iommu-map translation for id 0x0 on (null)#asahi-dev · 2022-10-19 21:57
robherOr just another problem...#asahi-dev · 2022-10-19 21:57
robherjannau: I think the problem may be PCI probes before the DART.#asahi-dev · 2022-10-19 21:34
robherActually, the NULL deref happened the current (sucessful) boot too...#asahi-dev · 2022-10-19 21:33
robherjannau: It's the config from Asahi for the most part.#asahi-dev · 2022-10-19 21:28
robherHere's a boot with the NULL deref: https://www.irccloud.com/pastebin/6EhNxoL3/#asahi-dev · 2022-10-19 21:13
robherI thought maybe there was both dart@1234 and iommu@1234 which dtc only complains about with 'W=1', but no complaint there. Though 'dart' should be fixed...#asahi-dev · 2022-10-19 21:04
robherThe dts looks sane to me...#asahi-dev · 2022-10-19 20:24
robherIn this case, things seem to be working. Some boots have no audio nor USB and there's a NULL ptr backtrace from the DART driver. I don't seem to have captured a log from that.#asahi-dev · 2022-10-19 20:15
robherhttps://www.irccloud.com/pastebin/bmB2pjMO/#asahi-dev · 2022-10-19 20:15
robheriommu probing seems flaky on current asahi-wip:#asahi-dev · 2022-09-27 13:27
robhermarcan: Yeah, I think so.#asahi-dev · 2022-09-27 13:25
robherright#asahi-dev · 2022-09-27 13:23
robhermarcan: For a virtual device, no need to put in DT. Just need the initcall to create platform device based on of_machine_is_compatible() and then it's probe can iterate over nodes by compatible.#asahi-dev · 2022-09-27 13:20
robhermarcan: The performance-domain binding can do the mapping.#asahi-dev · 2022-09-27 12:19
robhermarcan: Fix the OS? cpufreq has always been problematic though often that's because it's more a virtual device on top of other drivers.#asahi-dev · 2022-09-18 01:20
robhermarcan: I'm offline the next week so will have to look at it later.#asahi-dev · 2022-09-16 14:35
robherjannau: NP, I may be biased now with my new m1pro. ;)#asahi-dev · 2022-09-15 12:27
robhermarcan: platform maintainer is who should shout if the dts files start having warnings. Either you or I can take the dart binding if it comes to that...#asahi-dev · 2022-09-01 21:52
robherjannau: A fix is pushed to dtschema.#asahi-dev · 2022-09-01 20:17
robherGreat!#asahi-dev · 2022-09-01 20:17
robherHopefully someone will fix to keep apple platforms warning free.#asahi-dev · 2022-09-01 20:16
robherThere is also this one: interrupt-controller@23b100000: affinities:p-core-pmu-affinity: 'fiq-index' is a required property#asahi-dev · 2022-09-01 20:12
robherjannau: humm, definitely something weird happening there. Will investigate.#asahi-dev · 2022-09-01 19:48
robherBut I'd expect it to be fine until maxItems is > 1.#asahi-dev · 2022-09-01 19:47
robherIt is probably going to need custom decoding (in https://github.com/devicetree-org/dt-schema/blob/main/dtschema/dtb.py) given it's kind of special.#asahi-dev · 2022-09-01 19:45
robherjannau: Do you have more detail?#asahi-dev · 2022-06-30 13:17
robhermarcan: nice! That all sounds fine to me.#asahi-dev · 2022-06-02 15:19
robheralyssa: Helps people from giving acks that will never be applied or applying the patch to another tree.#asahi-dev · 2022-03-10 14:56
robhermacstudio?#asahi-dev · 2022-03-10 14:51
robhermarcan: Though if you just used macros for labels that would be a lot better than the Armada crap.#asahi-dev · 2022-03-10 14:50
robhermarcan: Right. Raise the question on the devicetree-compiler list.#asahi-dev · 2022-03-10 14:36
robhermarcan: if we can't express what's needed in dts syntax, we can always add to the syntax. #asahi-dev · 2022-03-10 14:26
robhermarcan: for things on buses, 'ranges' should help you. 'aliases' shouldn't be shared anyways. They usually should be board specific.#asahi-dev · 2022-03-10 13:53
robhermarcan: sub-directories are up to you, but I think would be an overkill. arch/arm/boot/dts is what is unwieldy. :(#asahi-dev · 2022-03-10 13:51
robhermarcan: we generally discourage macros for constructing DT nodes, but for includes whatever works for you.#asahi-dev · 2022-02-28 14:00
robherphys are pretty tied to process nodes and therefore change in almost every soc.#asahi-dev · 2022-02-09 13:58
robhermarcan: BTW, are you aware of power/reset/syscon-poweroff.yaml? Might be useful for shutdown here.#asahi-dev · 2022-02-09 13:48
robherI have to wonder if Apple wasn't disciplined enough for consistent register offsets, is the rest of the h/w functionality consistent? The answer for most vendors is very much a NO.#asahi-dev · 2022-02-09 13:40
robherWith an MFD node, it would define the full range of the block and then the child nodes define reg within the MFD range, but not overlapping with child node 'reg'.#asahi-dev · 2022-02-09 13:38
robherYou might want to have a mfd type node with rtc being a child node. I don't think having something like 'reg = <4 4>, <20 4>, <0 0x1000>;' would be a good idea. If you have that, I'd just have a custom offsets property with the offsets you need.#asahi-dev · 2022-02-09 13:31
robherare the rtc regs interleaved with other things?#asahi-dev · 2022-02-09 13:28
robherhumm, it would be a bit odd if the reg entries overlap themselves.#asahi-dev · 2022-02-09 13:25
robherI'm okay with the offsets in 'reg' if you want to go that route.#asahi-dev · 2022-02-09 13:22
robherAre all the masks consistent or you need properties for those too?