#asahi-re · 2021-01-11 16:58
davidrysk[m]but isn't clear whether that's just for kexts#asahi-re · 2021-01-11 16:57
davidrysk[m]the manpage for kmutil uses the term "boot kext collection"#asahi-re · 2021-01-11 16:57
davidrysk[m](about you being tired lol)#asahi-re · 2021-01-11 16:57
davidrysk[m]:(#asahi-re · 2021-01-11 16:56
davidrysk[m]marcan: are you looking at the kernel or the kernelcollection?#asahi-re · 2021-01-11 16:56
davidrysk[m]too basic :)#asahi-re · 2021-01-11 16:55
davidrysk[m]https://github.com/corkami/pics/tree/master/binary/macho101#asahi-re · 2021-01-11 16:55
davidrysk[m]uh I know corkami made a poster for mach-o#asahi-re · 2021-01-11 16:55
davidrysk[m]no, it uses LC_MAIN#asahi-re · 2021-01-11 16:54
davidrysk[m]that's not how a typical (userspace) mach-o encodes its entry point#asahi-re · 2021-01-11 16:54
davidrysk[m]https://twitter.com/marcan42/status/1348668351019565056#asahi-re · 2021-01-11 16:46
davidrysk[m]yeah#asahi-re · 2021-01-11 16:45
davidrysk[m]https://crna.cc/b/4? -- kinda interesting#asahi-re · 2021-01-11 16:39
davidrysk[m]oop, PongoOS directly includes APSL-licensed Apple headers :(#asahi-re · 2021-01-11 16:37
davidrysk[m]wonders if PongoOS has done some of this work already#asahi-re · 2021-01-11 16:30
davidrysk[m]marcan: https://twitter.com/marcan42/status/1348668349404782592 -- are you looking at the built t8101 kernel?#asahi-re · 2021-01-10 03:57
davidrysk[m]the IOKit documentation from the PPC Mac OS X era talks about it: https://developer.apple.com/library/archive/documentation/DeviceDrivers/Conceptual/IOKitFundamentals/DataMgmt/DataMgmt.html#asahi-re · 2021-01-10 03:45
davidrysk[m]I wonder just how many engineers involved with PowerPC bringup came back to Apple to get M1 going :)#asahi-re · 2021-01-10 00:09
davidrysk[m]https://github.com/bazad/devicetree-parse doesn't crash for me on the MBP one#asahi-re · 2021-01-09 19:39
davidrysk[m]but in some cases you still need maximum performance#asahi-re · 2021-01-09 19:39
davidrysk[m]they're heavily leveraging IOMMUs to reduce the exposure that Thunderbolt and PCIe would otherwise cause#asahi-re · 2021-01-09 19:37
davidrysk[m]that said, there still are uses for kernel/kext programming#asahi-re · 2021-01-09 19:37
davidrysk[m]I kinda want to see them move their own drivers into userspace as much as possible too, in order to reduce kernel attack surface#asahi-re · 2021-01-09 19:36
davidrysk[m]and since Apple controls the hardware down to the architecture level now, they can modify the hardware design as needed to avoid the typical problems with microkernels#asahi-re · 2021-01-09 19:35
davidrysk[m]it makes lots of sense to move as much of the driver stuff as possible to userspace#asahi-re · 2021-01-09 19:31
davidrysk[m]though I do hope they implement a proper FUSE layer in macOS 12#asahi-re · 2021-01-09 19:31
davidrysk[m]I expect to see it for a few specific use cases: primarily filesystems and stuff that requires maximum performance#asahi-re · 2021-01-09 19:30
davidrysk[m]modwizcode: they wrote a fair amount of documentation for it#asahi-re · 2021-01-09 19:30
davidrysk[m](which isn't what we're doing here)#asahi-re · 2021-01-09 19:29
davidrysk[m]they'll be incredibly useful for kext dev#asahi-re · 2021-01-09 19:29
davidrysk[m]I believe the t8101 debug kernels are waiting for configure-boot to be finished#asahi-re · 2021-01-09 19:29
davidrysk[m](including for the t8020 kernels for the DTK)#asahi-re · 2021-01-09 19:29
davidrysk[m]the dwarf files for some kernels can be found in the KDKs#asahi-re · 2021-01-09 19:29
davidrysk[m]dSYMs contain a dwarf file#asahi-re · 2021-01-09 19:28
davidrysk[m]yep#asahi-re · 2021-01-09 19:28
davidrysk[m]with mach-o, full debug information is stored separately from the binaries#asahi-re · 2021-01-09 19:28
davidrysk[m]hm, probably just un-stripped#asahi-re · 2021-01-09 19:26
davidrysk[m]that's my point#asahi-re · 2021-01-09 19:26
davidrysk[m]well, yep#asahi-re · 2021-01-09 19:26
davidrysk[m](-l)#asahi-re · 2021-01-09 19:26
davidrysk[m]Shiz: did you see how much crap bputil prints out if you enable debugging?#asahi-re · 2021-01-09 19:25
davidrysk[m]due to dynamic runtime crap, objc/swift compiled binaries necessarily have lots of strings in them#asahi-re · 2021-01-09 19:24
davidrysk[m]are they using objc or swift?#asahi-re · 2021-01-09 19:24
davidrysk[m]Shiz: > re:string, libbootpolicy.dylib is *extremely* verbose#asahi-re · 2021-01-09 19:19
davidrysk[m](well, linux with btrfs or xfs)#asahi-re · 2021-01-09 19:19
davidrysk[m](copy on write, like cp --reflink=always on linux)#asahi-re · 2021-01-09 19:18
davidrysk[m]oh yeah tip if you're limited for disk space: cp -c will make a reflink copy#asahi-re · 2021-01-09 19:11
davidrysk[m]there used to be darwinbuild but it's dead#asahi-re · 2021-01-09 19:11
davidrysk[m]a lot of this stuff depends on other crap that's also on opensource.apple.com but building it correctly is a huge mess#asahi-re · 2021-01-09 19:10
davidrysk[m]but maybe you had a better fix