#asahi-re · 2021-01-16 18:19
j`eyassuming you dont have a uart connected#asahi-re · 2021-01-16 18:19
j`eyBluerise: you could try commenting out the printfs/uarts#asahi-re · 2021-01-16 18:15
j`eyright, thats what I meant by "it's disabled"#asahi-re · 2021-01-16 18:15
Bluerisethe lowest-spec mini#asahi-re · 2021-01-16 18:15
BlueriseAh, yeah. But by default it's on, he's got to disable it explicitly#asahi-re · 2021-01-16 18:15
j`eyBluerise: what m1 are you on?#asahi-re · 2021-01-16 18:14
BlueriseSo I'll reinstall big sur and update it to beta 2 and re-test#asahi-re · 2021-01-16 18:14
j`eyBluerise: https://github.com/AsahiLinux/m1n1/blob/main/src/main.c#L38#asahi-re · 2021-01-16 18:14
Bluerisej`ey: in the stream he says the WDT is on#asahi-re · 2021-01-16 18:14
Blueriseyrlf: I think he couldn't boot his beta1 anymore, so maybe I should just update/install beta2 on both#asahi-re · 2021-01-16 17:59
Blueriseyup#asahi-re · 2021-01-16 17:59
j`eyBluerise: using the latest m1n1?#asahi-re · 2021-01-16 17:56
yrlfBluerise: didn't marcan say something about that at the beginning of last (not today's) stream?#asahi-re · 2021-01-16 17:55
Bluerisenope, nothing on the screen, simply reboots#asahi-re · 2021-01-16 17:53
Shizi'd say just the main (if you want it to continue booting)#asahi-re · 2021-01-16 17:51
Blueriseor only the main install?#asahi-re · 2021-01-16 17:51
Bluerisedoes my second macos install,the one where I replace the kernel, need the newest beta as well?#asahi-re · 2021-01-16 17:50
Blueriseheh, I did not enable the logo, let me retry with it on.#asahi-re · 2021-01-16 17:50
j`eybut the watchdog is disabled, so it shouldnt just reboot..#asahi-re · 2021-01-16 17:49
j`eydid you enable the logo?#asahi-re · 2021-01-16 17:49
Blueriseright now it just reboots#asahi-re · 2021-01-16 17:49
Bluerisemarcan: should m1n1.macho show something when bootet in a mini?#asahi-re · 2021-01-16 16:51
uberusha1sorry about that#asahi-re · 2021-01-16 16:51
uberusha1 /j #radare#asahi-re · 2021-01-15 04:43
marcan*refers#asahi-re · 2021-01-15 04:43
marcanmy gut feeling is that this is okay; conversely, I don't think bloom needs this, so why go there. it would be nice to have a lawyer's opinion#asahi-re · 2021-01-15 04:41
marcandisassembling and decompiling referse specifically to the actual object code#asahi-re · 2021-01-15 04:41
marcanon paper it does not, no#asahi-re · 2021-01-15 04:39
davidrysk[m]one could argue that listing exported symbols and linking against the userspace graphics driver binaries is not disassembling or decompiling them#asahi-re · 2021-01-15 04:39
davidrysk[m]and, for marcan, the policy says "contributors must not disassemble or decompile userspace graphics driver binaries", but does this fall under that prohibition#asahi-re · 2021-01-15 04:37
davidrysk[m]is linking and calling exported functions that follow the same effective API as LLVM in an Apple private framework in order to assemble LLVM IR into GPU-architecture code in violation of EULA#asahi-re · 2021-01-15 04:36
davidrysk[m]I expect this to be more a EULA/RE problem than a copyright problem#asahi-re · 2021-01-15 04:29
marcanI don't have a good answer personally#asahi-re · 2021-01-15 04:29
marcanI'm actually going to try to see if I can get a lawyer involved with us to answer questions like this#asahi-re · 2021-01-15 04:17
mjg59But you're probably into speak to a lawyer territory#asahi-re · 2021-01-15 04:17
mjg59You'd have some argument that under US law it's just a list of facts and so not copyrightable#asahi-re · 2021-01-15 04:16
mjg59Probably jurisdiction dependent#asahi-re · 2021-01-15 03:47
TellowKrinkle[m]I'm going to guess probably not, but not sure#asahi-re · 2021-01-15 03:47
TellowKrinkle[m]For the instruction metadata, I think an equivalent question would be "if a GPU manufacturer included a JSON file in their gpu driver with a list of registers and instruction scheduling data, would it be okay to use that when developing your driver?"#asahi-re · 2021-01-15 03:25
davidrysk[m]thanks!#asahi-re · 2021-01-15 03:25
davidrysk[m]cool :)#asahi-re · 2021-01-15 03:13
TellowKrinkle[m]Alright I posted it#asahi-re · 2021-01-15 03:03
TellowKrinkle[m]I'm making a post on that issue right now#asahi-re · 2021-01-15 03:03
davidrysk[m]it might be totally ok#asahi-re · 2021-01-15 03:03
davidrysk[m]yeah just run it by marcan if he doesn't see this#asahi-re · 2021-01-15 03:02
TellowKrinkle[m]You do need to call the AGX2 target initialization functions, which are not part of the normal LLVM headers (since the public version has no AGX2 target)#asahi-re · 2021-01-15 03:00
davidrysk[m]the only reason you'd need the symbol list is to figure out which symbols to call, but if they're plain llvm symbols and you can use llvm headers...#asahi-re · 2021-01-15 02:58
TellowKrinkle[m]Never mind looks like you don't need the symbol list, clang magically finds the dylib in the dyld cache, so we don't need the symbol list in the repo#asahi-re · 2021-01-15 02:51
TellowKrinkle[m](By it I mean the asmcheck program, not llc)#asahi-re · 2021-01-15 02:49
TellowKrinkle[m](It prints the associated instruction metadata from llvm tablegen)