#
✕ nick:jannau
#asahi-dev · 2023-02-08 10:55
jannauthe trackpad generates ~10k irq/s
#asahi-dev · 2023-02-08 10:27
jannauto bad bluez seems to have no import device (with pairing key) functionality
#asahi-dev · 2023-02-08 10:26
jannaubt device/key parsing from nvram done. I guess I should write asahi-btsync in rust as well to avoid a stdout / shell interaction with device supplied name fields
#asahi-dev · 2023-02-08 10:04
jannausven: could it be a backchannel to pmgr / pmp?
#asahi-dev · 2023-02-07 00:12
jannauoh, t6021 support: https://github.com/AsahiLinux/m1n1/pull/285
#asahi-dev · 2023-02-06 08:59
jannauat least in nvram-proxy-data in the ADT but iirc the nvram content was identical
#asahi-dev · 2023-02-06 08:58
jannaumarcan: yes
#asahi-dev · 2023-02-06 08:49
jannaumarcan: re magic mouse: macos puts the bt keys in nvram. we need to come up with a way to let bluez import them
#asahi-dev · 2023-02-06 08:45
jannaumarcan: asmedia on both m2 and m2 pro
#asahi-dev · 2023-02-03 21:40
jannaut: for testing that the changes have the desired effect. the most interesting features are routing serial port to the SBU pins locally and remotely and resetting a connected mac
#asahi-dev · 2023-02-03 20:10
jannauI got just distracted by 13.2 / m2 mini support
#asahi-dev · 2023-02-03 20:09
jannauruntime pm shouldn't be hard to implement. I expect it's enough to shut the endpoint down and restart it at convenient places from dcp's suspend/resume functions
#asahi-dev · 2023-02-03 20:00
jannausven: I don't have much dcp changes outside of asahi-wip. rebased dp-altmode changes are in https://github.com/jannau/linux/tree/asahi-6.2-rc3-7_dp-altmode-wip (works only once due to missing runtime pm support for for dptx-port)
#asahi-dev · 2023-02-03 19:55
jannauack, just trying to find a way to limit the email recipients per patch automatically
#asahi-dev · 2023-02-03 19:51
jannauno, bt and wifi will at least need the new pci device ids added
#asahi-dev · 2023-02-03 19:50
jannauyes, maintainers and mailing lists for N subsystems
#asahi-dev · 2023-02-03 19:43
jannausigh, do I want to send the m2 devicetrees and 15 mostly trivial dt-binding changes to 36 recipients?
#asahi-dev · 2023-02-03 17:41
jannausame memory / mmio layout as m1 pro/max?
#asahi-dev · 2023-02-03 15:37
jannauat least it finished this time
#asahi-dev · 2023-02-03 14:27
jannauinstaller issue is reproducible. got slow later (around 89%), I can fetch files from cdn.asahilinux.org with several over 100 mbit/s
#asahi-dev · 2023-02-03 10:02
jannaum2 pro "aic-iack-offset" = 0xc000 so no 4 die m2 pro (and most likely m2 max) based SoCs
#asahi-dev · 2023-02-02 23:09
jannauif I'm not mistaken macos 13.2 updates / installs got quite a bit faster
#asahi-dev · 2023-02-02 23:08
jannauseems to be dead without finishing. looks broken to me
#asahi-dev · 2023-02-02 22:47
jannauat 79.99% now
#asahi-dev · 2023-02-02 22:37
jannautook maybe 30 minutes to progress from 73% to 78%
#asahi-dev · 2023-02-02 22:34
jannauextracting the asahi desktop root image is very slow on the m2 mac mini on macos 13.2 (local patched installer)
#asahi-dev · 2023-02-02 15:49
jannauI would be surprised if the HDMI port produces any output, m1n1 might even crash during display config
#asahi-dev · 2023-02-02 15:47
jannauyou have to start with getting m1n1 to run. see '#define BRINGUP' in config.h
#asahi-dev · 2023-02-02 13:22
jannauyes, I'll use the coremark results
#asahi-dev · 2023-02-02 13:21
jannausven: there's also the firmware but I suppose binary documentation is even worse than C/python code. it has one advantage though. it is by definition correct
#asahi-dev · 2023-02-02 13:19
jannauat least it doesn't produce a ton of warnings like dhrystone, even when compiled with --std=c89
#asahi-dev · 2023-02-02 13:16
jannauhandlerug: I fear there is only code. C and python code
#asahi-dev · 2023-02-02 13:15
jannaudhrystone value is a little suspicious since it's less than icestorm but the m2 e-cores should have become faster relative to the p-cores
#asahi-dev · 2023-02-02 13:12
jannauI get different results for coremark and dhrystone. e-core capacity of 710 for dhrystone and 756 for coremark
#asahi-dev · 2023-02-02 12:17
jannauwhat did we use to determine capacity-dmips-mhz? I remember coremark but Documentation/devicetree/bindings/arm/cpu-capacity.txt advises for Dhrystone
#asahi-dev · 2023-02-02 10:55
jannauhttps://www.ti.com/lit/ug/slvubh2b/slvubh2b.pdf is the TRM for another variant of the TI chip which has additional bits described and might be a little closer to the variant apple uses
#asahi-dev · 2023-02-02 07:42
jannauthanks j`ey, marcan. I was already off
#asahi-dev · 2023-02-01 21:57
jannauI will extract the upstreamable t8112 device tree bits tomorrow and prepare a patch set
#asahi-dev · 2023-02-01 21:55
jannauM2 mac mini device tree: https://github.com/AsahiLinux/linux/pull/110
#asahi-dev · 2023-02-01 17:40
jannauamazon.de sells it for €164 but with 14 days lead time
#asahi-dev · 2023-02-01 17:28
jannauseems to roughly end up at the same same price you would pay if the dock is bought directly in germany
#asahi-dev · 2023-01-31 23:04
jannaudcp logs that after a dptx hpd with the dcp iboot powered on and m1n1 waiting on hpd from dcp iboot
#asahi-dev · 2023-01-31 23:01
jannau"syslog: [nifiedPipeline.cpp:6906]VideoInterface::Plug callback: AP not ready yet, mark hot-plug pending" doesn't sound nice
#asahi-dev · 2023-01-31 14:13
jannaudptxport handling is probably incomplete but dcp stopped sending messages for that endport. in the macos trace everything is dcp initiated though
#asahi-dev · 2023-01-31 13:50
jannaumacos doesn't use the dp2hdmi-gpio.function-reset gpio
#asahi-dev · 2023-01-31 13:11
jannaustill a little chaotic on my part since I have the dptx-phy initialization and smc enable gpios in python and the rest in m1n1 triggered by a proxy display_configure after that
#asahi-dev · 2023-01-31 13:08
jannauas far as I can see dcp speaks with dp2hdmi, there's a reset gpio but I think macos doesn't use it
#asahi-dev · 2023-01-31 12:45
jannaudptx in m1n1 seems to finally work but still no hdmi output on the m2 mac mini
#asahi-dev · 2023-01-31 09:08
jannauwithout the read the tracer's output becomes confusing since rx and tx are out of order
#asahi-dev · 2023-01-31 09:06
jannaushouldn't it be safe to read from the mailbox in m1n1's afk implementation if the rx ringbuffer has data?

plus ›