#
✕ nick:marcan
#asahi-dev · 2023-04-06 18:03
marcanoh I think that fifo thing is common these days
#asahi-dev · 2023-04-06 18:02
marcaninteresting
#asahi-dev · 2023-04-06 18:01
marcanChaosPrincess: oooh, nice! that was faster than I expected :)
#asahi-dev · 2023-04-06 17:52
marcanChaosPrincess: ack, remind me if I haven't done that by the weekend please (need to sleep now :p)
#asahi-dev · 2023-04-06 07:01
marcanI guess I might need to disassemble the firmware again...
#asahi-dev · 2023-04-06 07:01
marcanhow does it work on laptops then?
#asahi-dev · 2023-04-06 07:00
marcanugh...
#asahi-dev · 2023-04-05 17:30
marcanjannau: libinput snafu was my fault, we had an old package overlaid and I didn't think we'd end up with ABI breakage.
#asahi-dev · 2023-04-04 11:52
marcanhv.add_hvcall() etc to register handlers
#asahi-dev · 2023-04-04 11:52
marcanbrk 0x4242 is an EL0 hypercall in m1n1
#asahi-dev · 2023-04-04 11:52
marcanuse a `brk`, that gets caught by the hypervisor too
#asahi-dev · 2023-04-04 11:48
marcanOSes don't usually need to do VA to PA conversion via hardware, they have higher-level page management structures.
#asahi-dev · 2023-04-04 11:47
marcan*executing user code
#asahi-dev · 2023-04-04 11:46
marcanof course it will only work if EL0 is actually mapped, which might only be the case when actually executing kernel code. and then you'll have to look up how PAN interacts with that.
#asahi-dev · 2023-04-04 11:45
marcanother than that it should work
#asahi-dev · 2023-04-04 11:45
marcangordonfreeman: using the live SPSR_EL2 is probably wrong, it should use the saved SPSR from the HV context so it can survive exceptions at EL2. but at that point `el` might as well be an argument.
#asahi-dev · 2023-04-04 11:39
marcanah, for EL0... not sure about that
#asahi-dev · 2023-04-04 11:38
marcanhv_translate does work
#asahi-dev · 2023-04-02 17:10
marcanANE is very much a "side" thing right now so I'm comfortable leaving things up to you, I doubt it's going to become a critical piece of the ecosystem overnight (which is a good place to be when developing since you don't have to care about users as much :p)
#asahi-dev · 2023-04-02 17:08
marcaneiln: since I forgot to answer, before I go to sleep: feel free to move in-tree, that's where we want to be anyway. you can prepare a branch based on asahi-wip and I can rebase it into one more bits/ branch. we can enable it in actual asahi-dev then asahi builds whenver you think it's usable/worth having people test.
#asahi-dev · 2023-04-02 15:29
marcanand m1n1 only copies dynamic data
#asahi-dev · 2023-04-02 15:29
marcanbut that violates the principle that this is static data for any given device
#asahi-dev · 2023-04-02 15:29
marcanin principle yes, I guess
#asahi-dev · 2023-04-02 15:25
marcanAIUI that's on purpose, to make porting over new devices less of a pain
#asahi-dev · 2023-04-02 15:25
marcanthose numbers in the DTs are pretty much 1:1 from Apple's DT
#asahi-dev · 2023-04-02 15:25
marcanit'll just be a royal pain to make any changes
#asahi-dev · 2023-04-02 15:25
marcanI mean nothing stops you from replacing all the floats with the equivalent uint32 representations, it will work the same
#asahi-dev · 2023-04-02 15:23
marcanin fact the dynamic stuff m1n1 fills in isn't even floats (though it is computed using floats)
#asahi-dev · 2023-04-02 15:23
marcanno, only some of those are dynamic, m1n1 doesn't know the per-model values
#asahi-dev · 2023-04-02 15:22
marcanfeel free to come up with a better one ;)
#asahi-dev · 2023-04-02 15:20
marcanjust be glad m1n1 is taking care of the calculations involving exponentials (yes there's some of those those too)
#asahi-dev · 2023-04-02 15:19
marcanoh right, you want to do it in C :D
#asahi-dev · 2023-04-02 15:19
marcanhey lina pulled off softfloat in a couple hundred lines of rust and it's all operator overloaded and nice and ergonomic
#asahi-dev · 2023-04-02 15:18
marcangood luck not doing floats
#asahi-dev · 2023-04-02 15:18
marcanthe firmware wants floats
#asahi-dev · 2023-04-02 15:18
marcanthe kernel itself never uses floats
#asahi-dev · 2023-04-02 15:18
marcanonly dtc cares
#asahi-dev · 2023-04-02 15:18
marcanthat is already taken care of, it's softfloat
#asahi-dev · 2023-04-02 15:17
marcanyeah well it's either this or we start putting binary float32 hex constants in the DT which is worse
#asahi-dev · 2023-04-02 15:17
marcankettenis_: sounds like you're in a great position to start pushing through the float support then! :D
#asahi-dev · 2023-04-02 15:16
marcanideally also tell me what arch/ALARM packages I need as dependencies to do all this (flatpak-builder))
#asahi-dev · 2023-04-02 15:16
marcanthis isn't going to be very complicated anyway, if you can give me some flatpak yaml that can build the flatpak out of some relative directory of your choosing I can deal with the PKGBUILD side
#asahi-dev · 2023-04-02 15:14
marcanit's midnight here, I'm about to crash, and I'd prefer to keep discussion here since here is where the community is
#asahi-dev · 2023-04-02 15:06
marcanyeah so just do that :p
#asahi-dev · 2023-04-02 15:06
marcanor just make a symlink in the PKGBUILD...
#asahi-dev · 2023-04-02 15:05
marcanalso the $PWD is src/ when the prepare() / build() functions run
#asahi-dev · 2023-04-02 15:05
marcansrc/mesa-asahi-$_asahiver
#asahi-dev · 2023-04-02 15:04
marcansrc/
#asahi-dev · 2023-04-02 15:03
marcanbut the use case is arm64 :p
#asahi-dev · 2023-04-02 15:03
marcansure

plus ›