#asahi · 2021-06-19 06:14
marcanamw: re run_guest.py, open up a picocom on the secondary USB CDC-ACM device#asahi · 2021-06-19 06:12
marcanVinDuv: the problem with running m1n1/linux under the HV is there is no vUART input support yet; been meaning to add that...#asahi · 2021-06-19 06:10
marcanSystem Preferences > Security & Privacy > FileVault I believe#asahi · 2021-06-19 06:09
marcanunless they changed it in 11.4, which they might have#asahi · 2021-06-19 06:08
marcanI believe the security model is that if you have FileVault enabled, you need to authenticate before being allowed to run your own code#asahi · 2021-06-19 06:08
marcanencryption#asahi · 2021-06-19 06:06
marcanamw: do you have FileVault enabled?#asahi · 2021-06-19 00:36
marcanbloom: feel free to set +M if we get another attack#asahi · 2021-06-18 11:45
marcankettenis: I think DCP is already running#asahi · 2021-06-18 11:14
marcanyes#asahi · 2021-06-18 09:39
marcanj_ey: yeah, you're correct :)#asahi · 2021-06-18 08:56
marcanat which point it is more privileged than ELx, but it has to be bootstrapped like that first#asahi · 2021-06-18 08:56
marcanand can be enabled by ELx, and locked#asahi · 2021-06-18 08:56
marcaneach GLx is a side level for ELx#asahi · 2021-06-18 08:56
marcanbytefire_: the GL levels are not part of the normal ELx order#asahi · 2021-06-18 05:27
marcangoing to stream for a bit now#asahi · 2021-06-18 05:26
marcanyeah, I think I know how I want to do it#asahi · 2021-06-18 05:24
marcanthat might be cleaner to implement (though less flexible)#asahi · 2021-06-18 05:24
marcanor maybe I can have a "shape" for register arrays#asahi · 2021-06-18 05:23
marcanfine for now though#asahi · 2021-06-18 05:23
marcanI don't like the irq group duplication but I ran into something similar with DART; I'll see about adding explicit register lists, so that combined with the array support we can have 2D arrays#asahi · 2021-06-18 05:23
marcanfair#asahi · 2021-06-18 05:22
marcanis that because 0 is "no group" or so?#asahi · 2021-06-18 05:22
marcanare there really 7 IRQ groups?#asahi · 2021-06-18 05:22
marcanyeah, I think so#asahi · 2021-06-18 05:20
marcanlgtm#asahi · 2021-06-18 05:18
marcanoh, because padding?#asahi · 2021-06-18 05:17
marcanwhy have all the IRQ groups as separate arrays and not one combined one?#asahi · 2021-06-18 05:14
marcanregisters may have overlapping sparse ranges#asahi · 2021-06-18 05:14
marcanjannau: the if is necessary#asahi · 2021-06-18 04:36
marcanthere's the page size issue which may or may not matter depending on how it's implemented#asahi · 2021-06-18 04:36
marcanuse likely yes, pass through probably; not sure how passthrough interacts with IOMMU drivers#asahi · 2021-06-18 01:47
marcan(and it'll probably run faster on Linux once the graphic bits are done, unless Apple and Parallels implement vGIC support before then :p)#asahi · 2021-06-18 01:47
marcanbrentr123[m]: considering how much faster the M1 is than most arm64 windows machines, it ought to be running *faster* than "full" speed :P#asahi · 2021-06-17 17:36
marcananyway, off to sleep :)#asahi · 2021-06-17 17:35
marcanthat used to be a bool trace (on/off)#asahi · 2021-06-17 17:35
marcanre that "is True" thing, backwards compat#asahi · 2021-06-17 17:35
marcanI'm likely to deprecate vm_hooks altogether yes#asahi · 2021-06-17 17:34
marcanI still use that hook for non-Tracer clients, currently that PMGR hack, but that will probably go away#asahi · 2021-06-17 17:33
marcanso there's no point in using a token to tell things apart at the SPTE level#asahi · 2021-06-17 17:33
marcanand it turns out I need to do a RangeMap lookup for every MMIO event anyway to find the observe-only ones anyway#asahi · 2021-06-17 17:32
marcanoriginally I planned to have a table of hooks (the C side passes back an index into that) but trying to build it into this framework will run into issues with garbage in that list, since it is only ever appended to#asahi · 2021-06-17 17:32
marcanthe one that actually handles all this#asahi · 2021-06-17 17:32
marcan0 is the "internal" hook#asahi · 2021-06-17 17:30
marcanbut damn does it work well (once I ironed out the kinks)#asahi · 2021-06-17 17:30
marcanyeah it feels like a terrible idea and a great idea at the same time#asahi · 2021-06-17 17:29
marcancorrect :)#asahi · 2021-06-17 17:29
marcanyes#asahi · 2021-06-17 17:26
marcanI think it turned out quite pretty#asahi · 2021-06-17 17:26
marcanj_ey: so this is what I ended up with :-) https://github.com/AsahiLinux/m1n1/blob/main/proxyclient/m1n1/trace/asc.py