#
✕ nick:kettenis_
#asahi-dev · 2023-04-02 15:17
kettenis_floating-point in kernels is problematic
#asahi-dev · 2023-04-02 15:17
kettenis_I also have the feeling that this idea is going to meet some significant resistence
#asahi-dev · 2023-04-02 15:17
kettenis_marcan: the device tree floating-point is annoying since it means I can't build the asahi-wip device trees for testing with u-boot and OpenBSD anymore
#asahi-dev · 2023-04-02 15:00
kettenis_32-bit x86, seriously?
#asahi-dev · 2023-04-02 14:01
kettenis_and I can't find anything in the device tree schema that suggests that floating-point numbers are supported
#asahi-dev · 2023-04-02 14:01
kettenis_my version of dtc doesn't like that
#asahi-dev · 2023-04-02 14:00
kettenis_apple,avg-power-ki-only = <2.4>;
#asahi-dev · 2023-02-06 08:48
kettenis_I looked at it for a bit, but not very hard as i have been ill for the last 5 days
#asahi-dev · 2023-02-06 08:46
kettenis_looks like the offset of the register to start the secondary cores moved again
#asahi-dev · 2023-02-06 08:45
kettenis_yes, I got the m2pro mini with 10GbE to maximise hardware diversity
#asahi-dev · 2023-02-06 08:43
kettenis_yup
#asahi-dev · 2023-02-06 08:43
kettenis_then I can make a head start with u-boot support (including the asmedia usb)
#asahi-dev · 2023-02-06 08:43
kettenis_marcan: if you have some time to look at the low-level cpu core initialization (chickens, spin up) that would be great
#asahi-dev · 2023-02-03 17:44
kettenis_but the offsets to spin up the secondary CPUs appear to be different
#asahi-dev · 2023-02-03 17:43
kettenis_I think so
#asahi-dev · 2023-02-03 17:23
kettenis_serial port on the m2 pro mac mini works just like on the m1 mac mini
#asahi-dev · 2023-01-24 14:36
kettenis_I wonder if (and hope that) the serial port will work
#asahi-dev · 2023-01-24 13:52
kettenis_jannau: LOL
#asahi-dev · 2023-01-24 12:22
kettenis_I should get my M2 Pro mini in about two weeks if Apple's prediction is right
#asahi-dev · 2023-01-24 12:21
kettenis_u-boot will probably need changes for the M2 Pro/Max
#asahi-dev · 2023-01-24 10:42
kettenis_total? or just Apple Silicon machines? ;)
#asahi-dev · 2023-01-22 22:15
kettenis_one of the reasons I suggested that approach ;)
#asahi-dev · 2023-01-22 22:11
kettenis_linux may already have a quirk table
#asahi-dev · 2023-01-22 22:10
kettenis_might not include the function number in the RID
#asahi-dev · 2023-01-22 14:59
kettenis_thanks!
#asahi-dev · 2023-01-22 14:53
kettenis_prepares us for an upcoming u-boot change
#asahi-dev · 2023-01-22 14:52
kettenis_https://github.com/AsahiLinux/m1n1/pull/276
#asahi-dev · 2023-01-22 14:52
kettenis_marcan: since we're talking pull requests:
#asahi-dev · 2023-01-22 11:57
kettenis_I suppose I can always ignore the stuff in the DT if it conflicts with my OCD ;)
#asahi-dev · 2023-01-22 11:47
kettenis_having a consistent set of descriptions for those keys may be a challenge ("name of the power rail" vs. "what they're for")
#asahi-dev · 2023-01-22 11:38
kettenis_of course if Apple starts using the same key for wildly different purposes this won't work
#asahi-dev · 2023-01-22 11:38
kettenis_seems to work fine so far, although I only support a limited set of keys for now
#asahi-dev · 2023-01-22 11:37
kettenis_in the openbsd aplsmc(4) driver I simply have a list of keys we might be interested in and check which of them exist when the driver attaches
#asahi-dev · 2023-01-21 14:42
kettenis_PCIe has a switched topology with each downstream port of the switch getting its own bus number
#asahi-dev · 2023-01-21 14:40
kettenis_that would limit you to 64 bus numbers, but that is probably plenty
#asahi-dev · 2023-01-21 14:36
kettenis_given the way PCIe topologies work, you probably just want to map PCIe bus numbers onto SIDs
#asahi-dev · 2023-01-21 14:14
kettenis_ah, right
#asahi-dev · 2023-01-21 14:13
kettenis_the obvious thing to do actually is just pass the full RID to the iommu driver and let it figure out what SID to use
#asahi-dev · 2023-01-21 14:12
kettenis_how many SIDs does the iommu have?
#asahi-dev · 2023-01-21 09:17
kettenis_sven: are the pci bridges set up correctly?
#asahi-dev · 2023-01-20 14:52
kettenis_just looks like normal PCIe hotplug to the os
#asahi-dev · 2023-01-20 14:52
kettenis_inlcuding hotplug
#asahi-dev · 2023-01-20 14:52
kettenis_turns out that apart from a bug in the pci bridge resource management it just worked
#asahi-dev · 2023-01-20 14:51
kettenis_I'm kinda disappointed though; one of the reasons I got the x1 was because I wanted to look at what was needed in OpenBSD to have thunderbolt support
#asahi-dev · 2023-01-20 14:50
kettenis_it is a gen9 so fairly recent
#asahi-dev · 2023-01-20 14:48
kettenis_s/x9/x1/
#asahi-dev · 2023-01-20 14:48
kettenis_guess the firmware handles everything there
#asahi-dev · 2023-01-20 14:48
kettenis_nope; but it nevertheless works on my x9
#asahi-dev · 2023-01-20 14:46
kettenis_but no thunderbolt support in openbsd yet of course
#asahi-dev · 2023-01-20 14:46
kettenis_heh, so I have the same stack of dongles here now

plus ›