#
✕ nick:kettenis
#asahi-dev · 2022-01-04 13:28
kettenisyeah, probing is fine
#asahi-dev · 2022-01-04 13:24
kettenisfunction 2 doesn't exist, but reading its config space while the device is "alive" results in no more wifi interrupts
#asahi-dev · 2022-01-04 13:22
kettenisbluetooth is function 1
#asahi-dev · 2022-01-04 12:34
kettenishmm, attempting to read from config space of the (non-existing) function 2 of the BCM4378 PCI device makes the device crap out
#asahi-dev · 2022-01-03 15:26
kettenisso other drivers can do the same thing if needed
#asahi-dev · 2022-01-03 15:26
kettenisit probably doesn't at the moment, but the nvme driver now runs some code before handover to shut itself down as far as we can
#asahi-dev · 2022-01-03 15:03
kettenisand apparently an API for programs to influence those decisions
#asahi-dev · 2022-01-03 11:13
kettenisthen forget about what I just said for now ;)
#asahi-dev · 2022-01-03 11:12
kettenisyou removed those to avoid too much debug spam?
#asahi-dev · 2022-01-03 11:11
kettenisand it shouldn't be needed on the t6000 branch anyway since it doesn't have the pmgr driver yet
#asahi-dev · 2022-01-03 11:11
kettenisno
#asahi-dev · 2022-01-03 11:09
kettenisto fix the power domain issues you'd need to add some u-boot,dm-pre-reloc properties
#asahi-dev · 2022-01-03 11:06
kettenisshould be harmless
#asahi-dev · 2022-01-03 11:06
kettenisoh, you enabled some debugging output
#asahi-dev · 2022-01-03 11:03
kettenisdefenitely not expected
#asahi-dev · 2022-01-02 15:07
kettenisyeah, mailbox and nvme compatibles need to be reconciled still
#asahi-dev · 2022-01-02 14:35
kettenisso not entirely sure if what I just did really helps
#asahi-dev · 2022-01-02 14:34
kettenisoh wait, github, only lets you do that on commits
#asahi-dev · 2022-01-02 13:31
kettenisis it useful if I annotate those in the github repo?
#asahi-dev · 2022-01-02 13:31
kettenismarcan: I noticed some small merge botches in the t8103 dts on the asahi branch
#asahi-dev · 2022-01-01 19:13
kettenisfixes committed, but it will probably be a couple of days before they start showing up in snapshots
#asahi-dev · 2022-01-01 12:17
kettenissure
#asahi-dev · 2022-01-01 12:00
kettenisturned out the be a bug in the "lazy" FPU handling
#asahi-dev · 2022-01-01 11:59
kettenisbeen barking up the wrong tree all along
#asahi-dev · 2022-01-01 11:58
kettenisyay, finally figured out the OpenBSD SMP issue
#asahi-dev · 2021-12-31 14:00
ketteniswhether it actually works is a different question of course
#asahi-dev · 2021-12-31 14:00
kettenisthe relevant u-boot code seems to support that
#asahi-dev · 2021-12-31 11:34
ketteniscould be that get_page_table_size() returns a value that is too small
#asahi-dev · 2021-12-31 11:33
kettenisdon't see something obviously wrong with the t6000_mem_map
#asahi-dev · 2021-12-31 11:28
kettenisso it makes it into build_mem_map() probably
#asahi-dev · 2021-12-31 11:23
kettenisthis should work with the hypervisor uart emulation
#asahi-dev · 2021-12-31 11:23
ketteniswhen you re-enable the debug uart, you'll also need to change the base address to te riht address
#asahi-dev · 2021-12-31 11:15
kettenisyou'll nedd a device tree from asahi branch of the asahi github linux repo
#asahi-dev · 2021-12-31 11:15
kettenisthat'd be great; I don't have the hardware myself
#asahi-dev · 2021-12-31 11:07
ketteniswhere working means that you'll get some framebuffer output
#asahi-dev · 2021-12-31 11:06
kettenisthat may have a chance of working on the M1 Pro/Max
#asahi-dev · 2021-12-31 11:06
kettenishttps://github.com/kettenis/u-boot/tree/t6000
#asahi-dev · 2021-12-31 09:55
kettenisthe bits for enabling bypass mode in the t6000 dart didn't change isn't it?
#asahi-dev · 2021-12-31 09:34
kettenismeanwhile, I've figured out how to set up the memory map in u-boot "late"
#asahi-dev · 2021-12-31 09:25
kettenisah, but it disables single-stepping conditionally
#asahi-dev · 2021-12-31 09:24
kettenishmm, looks like linux messes with with MDSCR_EL1 as well when entering/leaving the kernel
#asahi-dev · 2021-12-30 22:59
kettenisI should probably change OpenBSD such that it doesn't touch that register unconditionally in each trap
#asahi-dev · 2021-12-30 22:58
kettenissven: that doesn't seem to handle any of the standard MSRs
#asahi-dev · 2021-12-30 22:14
kettenisnot immediately obvious to me how to stop the hypervisor do these prints
#asahi-dev · 2021-12-30 22:13
kettenislots of those running OpenBSD under the hypervisor
#asahi-dev · 2021-12-30 22:13
kettenis[cpu0] Pass: msr MDSCR_EL1, x2 = 1000 (OK) (MDSCR_EL1)
#asahi-dev · 2021-12-30 22:13
kettenis[cpu0] Pass: mrs x2, MDSCR_EL1 = 1000 (MDSCR_EL1)
#asahi-dev · 2021-12-29 23:23
kettenisbut I've been busy hacking my own stuff
#asahi-dev · 2021-12-29 23:23
kettenismy plan was to actually put that diff on my branch
#asahi-dev · 2021-12-29 23:20
kettenisnot yet; I assume you've tested it and it works ;)

plus ›