#asahi · 2021-12-25 09:20
marcanthe display panel controller#asahi · 2021-12-25 09:19
marcanthe SPI controller is done, there isn't really anything left to do there#asahi · 2021-12-25 09:19
marcanand Touch ID, and the display TCON#asahi · 2021-12-24 18:54
marcanor rather missing feature#asahi · 2021-12-24 18:54
marcanpmgr issue is just due to a hack in the MCA driver#asahi · 2021-12-24 18:30
marcanokay, okay, 33 years ago, you win :p#asahi · 2021-12-24 18:27
marcanyes#asahi · 2021-12-24 18:27
marcanI don't think IOMMUs existed back then :p#asahi · 2021-12-24 18:25
marcansven came up with some heroic patches to work around this and allow Linux to use 16K pages for IOMMUs while using 4K pages for the CPU, but they aren't ready yet#asahi · 2021-12-24 18:24
marcanthese machines are not designed for running with 4K pages in kernelmode#asahi · 2021-12-24 18:24
marcanyes, the IOMMUs are 16K only#asahi · 2021-12-24 18:23
marcandevices will show up but the drivers will all error out#asahi · 2021-12-24 18:23
marcanthat is expected#asahi · 2021-12-24 18:23
marcanyes, DMA doesn't work#asahi · 2021-12-24 18:19
marcanoh, just saw, heh#asahi · 2021-12-24 17:57
marcanI don't see how that error can happen other than if DMA allocs are horribly broken, which they would be on a 4K kernel#asahi · 2021-12-24 17:56
marcanthat would explain that kind of error#asahi · 2021-12-24 17:56
marcan4K kernels are not supported and will not work for PCIe devices yet, until the patchset for that gets in shape#asahi · 2021-12-24 17:56
marcanmps: wait, is this a 4K kernel or a 16K kernel?#asahi · 2021-12-24 10:27
marcanmps: yes#asahi · 2021-12-24 10:27
marcanunless you want to see me eating cheese fondue with a friend and opening presents :p#asahi · 2021-12-24 10:25
marcanthough I probably won't work on this until after xmas :p#asahi · 2021-12-24 10:25
marcanmps: can you enable all the debug options and paste a full dmesg of at least the brcmfmac range? CONFIG_BRCM_TRACING=y CONFIG_BRCMDBG=y#asahi · 2021-12-24 06:14
marcanpushed wifi/take4 which unbreaks t6000 machines (which broke due to the T2 changes)#asahi · 2021-12-24 03:11
marcantpw_rules: should be 64K tbh#asahi · 2021-12-23 18:28
marcanyou could try reverting d795c6f4901#asahi · 2021-12-23 18:12
marcanI'll test this later on more machines though#asahi · 2021-12-23 18:12
marcanstill the same chip and firmware#asahi · 2021-12-23 18:10
marcanmps: I did increase the ringbuffer size, but I find it hard to see how it could behave differently on the Air and on the Mini, unless it's a firmware difference#asahi · 2021-12-23 18:09
marcanlandscape15[m]: it's not just going to be a one-click installer; there will definitely be at least basic flexibility, possibly behind a toggle, if not an outright expert menu at some point#asahi · 2021-12-23 16:25
marcan(dunno if I will have a mode to do it without the EFI partition, seems kinda pointless since most people will want that anyway, but I will have a mode to re-create an existing stub with another version)#asahi · 2021-12-23 16:25
marcanof course it will allow you to just create the stub+EFI partition if that's all you want, it's not going to force you to install a rootfs#asahi · 2021-12-23 16:24
marcanlandscape15[m]: dev version of what? the final installer will do all the repartitioning for you including resizing macos and creating the EFI partition and such#asahi · 2021-12-23 14:52
marcanright now it just has that module that isn't otherwise used that you need to invoke by hand#asahi · 2021-12-23 14:52
marcanthat will happen once the installer is a real installer#asahi · 2021-12-23 14:52
marcanthe installer does not, and besides that needs distro support#asahi · 2021-12-23 14:52
marcanthe module in there yes#asahi · 2021-12-23 14:50
marcan(and I don't want to go having the firmware extractor decide whether to add the var or not based on platform)#asahi · 2021-12-23 14:50
marcanbut I just now realize that doing that on dongles which *do* have a MAC in SROM like on T2s overrides it#asahi · 2021-12-23 14:50
marcanpreviously I was just having a fake one in the nvram file and then setting the real one using firmware APIs on startup#asahi · 2021-12-23 14:49
marcanlandscape15[m]: you need to have *some* MAC address in NVRAM, otherwise the M1 machine firmwares won't start without a MAC at all#asahi · 2021-12-23 14:32
marcanthis should make the MAC work properly on both T2 and M1#asahi · 2021-12-23 14:31
marcanyou also need to update the asahi-installer repo and re-extract the firmware#asahi · 2021-12-23 14:31
marcanok, pushed wifi/take3 with a change to how MAC address handling is done#asahi · 2021-12-23 12:53
marcanin a way that affects M1#asahi · 2021-12-23 12:53
marcanwait for take3, I'm going to change the MAC handling#asahi · 2021-12-23 12:31
marcanfixed, fetch and try again#asahi · 2021-12-23 12:22
marcanwait I think I just messed up#asahi · 2021-12-23 12:22
marcanor not?#asahi · 2021-12-23 12:22
marcan.... that's a bug in the broadcom headers