#
✕ nick:kettenis
#asahi-dev · 2021-03-09 16:00
kettenisunfortunately the kernel dies as the result of an SError exception as soon as it tries to exec init
#asahi-dev · 2021-03-09 15:59
kettenisloading u-boot works and it runs the OpenBSD bootloader which in turn loads the OpenBSD kernel
#asahi-dev · 2021-03-09 15:59
kettenisI extended the asahi linux device tree with the minimum nodes from the corellium device tree to bring up pcie
#asahi-dev · 2021-03-08 17:43
kettenisif even Intel can pull that off...
#asahi-dev · 2021-03-08 17:39
ketteniswhich is cool since it will help me transition away from preloader-m1
#asahi-dev · 2021-03-08 17:39
kettenisit happily loads my preloader-m1 + u-boot image
#asahi-dev · 2021-03-08 17:23
kettenischainloading macho files already works
#asahi-dev · 2021-03-07 22:56
kettenisworks here as well, and I can talk to it using the proxyclient code
#asahi-dev · 2021-03-07 16:53
kettenistime to get some food here; sleep well
#asahi-dev · 2021-03-07 16:44
ketteniscool
#asahi-dev · 2021-03-07 16:43
ketteniswill give it a spin later when my current native build finishes
#asahi-dev · 2021-03-07 16:41
kettenis(work as in it compiles)
#asahi-dev · 2021-03-07 16:40
kettenisseems to work; thanks!
#asahi-dev · 2021-03-07 16:13
kettenisI'll finish watching another will deacon talk then ;)
#asahi-dev · 2021-03-07 16:07
kettenisshould I let you fix a few things and see if there are any remaining issues a bit later?
#asahi-dev · 2021-03-07 16:04
kettenisso I'm using aarch64-none-elf- which in this case is the OpenBSD port of the "standard" linaro cross-compiler
#asahi-dev · 2021-03-07 16:03
kettenisI have some modifications now to malloc_conf.h that make dlmalloc happy
#asahi-dev · 2021-03-07 16:02
kettenisdlmalloc and libfdt are the main offenders
#asahi-dev · 2021-03-07 16:01
kettenisand <string.h> as well
#asahi-dev · 2021-03-07 16:00
kettenis<assert.h> and <stdlib.h>
#asahi-dev · 2021-03-07 15:58
ketteniswhere as I'm using a bare-metal cross-compiler
#asahi-dev · 2021-03-07 15:58
kettenisah, you are using a cross-linux compiler ;)
#asahi-dev · 2021-03-07 15:57
kettenisI don't think you are
#asahi-dev · 2021-03-07 15:57
ketteniscross-compiler I mean
#asahi-dev · 2021-03-07 15:57
kettenisyeah, I run into this because I am using a cross-compoler ;)
#asahi-dev · 2021-03-07 15:56
kettenisa "freestanding" C implementation is supposed to provide some of them
#asahi-dev · 2021-03-07 15:56
kettenisyeah host
#asahi-dev · 2021-03-07 15:30
kettenisouch, m1n1 depends quite heavily on the hos header files
#asahi-dev · 2021-03-06 16:43
kettenisother than that it is essentially just a "pci-host-ecam-generic" driver
#asahi-dev · 2021-03-06 16:42
kettenisI also have some code to establish MSI interrupts on behalf of devices
#asahi-dev · 2021-03-06 16:40
kettenisarnd: I need to do some MSI config because the u-boot driver skips that
#asahi-dev · 2021-03-06 10:36
kettenisso the OpenBSD driver for PCIe is pretty simple
#asahi-dev · 2021-03-06 10:36
kettenisright now u-boot applies the tunables and brings up PCIe
#asahi-dev · 2021-03-06 10:33
kettenisstill using the corellium preloader at this point because I need the tunables to bring up PCIe
#asahi-dev · 2021-03-06 10:32
kettenisj`ey: I could if I was using m1n1
#asahi-dev · 2021-03-06 10:32
kettenisand would mean some code uplication
#asahi-dev · 2021-03-06 10:31
kettenisso having u-boot parsing the apple device tree would mean m1n1 has to pass its address somehow
#asahi-dev · 2021-03-06 10:31
kettenisu-boot uses standard device trees
#asahi-dev · 2021-03-06 10:25
kettenisthe alternative is to export the tunables like corellium does
#asahi-dev · 2021-03-06 10:25
kettenisto apply the tunables you need to enable the clocks such that you can access the pcie host bridge registers
#asahi-dev · 2021-03-06 10:21
kettenisafter I'm done dealing with the 8-bit ASID fallout properly in the OpenBSD kernel, I'm going to look into using m1n1 such that I can get all the cores running
#asahi-dev · 2021-03-06 10:18
kettenissven: with that diff we could enable the pcie clock and apply the tunables?
#asahi-dev · 2021-02-21 13:01
kettenisbased on the corellium code I would expect the lowest two bits to specify the access permissions
#asahi-dev · 2021-02-20 22:16
kettenisdidn't manage to get the pcie DART working in bypass mode though
#asahi-dev · 2021-02-20 22:14
kettenishah, got OpenBSD to boot multi-user on the mini
#asahi-dev · 2021-02-17 20:17
kettenisanyway, I'll probably get to the point where I can play with bypass mode for the PCIe devices in a few days
#asahi-dev · 2021-02-17 20:15
kettenisif there is nothing to flush it will probably clear immediately
#asahi-dev · 2021-02-17 19:56
kettenisnot familliar with SMMU, but in the end it's just a page table of sorts and some registers to poke to flush cached translation entries
#asahi-dev · 2021-02-17 19:50
kettenis0x4 seems to be a "busy" bit
#asahi-dev · 2021-02-17 19:32
kettenisand a non-zero code in the other one

plus ›