#asahi-dev · 2023-04-07 10:28
jannauI think it does nothing with the power-domain but even toggling the power-domain via reset in .remove doesn't brake anything. it SErrors if with USB_XHCI_PCI though#asahi-dev · 2023-04-07 10:03
jannauI've lost PCI_INIT_R, that explains why PCIE is only inited after `pci enum`#asahi-dev · 2023-04-07 09:57
jannauit does nothing weird directly but the port setup is called indirectly through the ecam init op#asahi-dev · 2023-04-07 09:36
jannauI fear we need a way to reset PCIe (ports). toggling power will probably work but I didn't manage to toggle power without running into SErrors (probably PCIe dart access)#asahi-dev · 2023-04-07 09:27
jannauwifi/bt and shdc work in both cases#asahi-dev · 2023-04-07 09:27
jannauat linux init time only the lan port is reported as up and works, xhci port is down and doesn't come up#asahi-dev · 2023-04-07 09:25
jannau`pci enum` at u-boot's prompt seems to bring the ports up and reports the lan and xhci ports as up#asahi-dev · 2023-04-07 09:24
jannaukettenis_: pcie is cursed. with PCIE_APPLE but !USB_XHCI_PCI in u-boot all parts are down when linux inits pcie. the port with the ethernet adapter doesn't come up#asahi-dev · 2023-04-06 21:37
jannauI guess the AttrPalmSizeThreshold should be removed if we disable libinput's palm rejection#asahi-dev · 2023-04-06 21:34
jannauthat matches the touchpads on apple m1 (pro/max) laptops#asahi-dev · 2023-04-06 21:33
jannaukit_ty_kate3: see "[Apple Laptop Touchpad (SPI)]" in https://gitlab.freedesktop.org/libinput/libinput/-/blob/main/quirks/50-system-apple.quirks #asahi-dev · 2023-04-06 21:31
jannauI think libinput has already quirks for apple devices though which iirc also match our devices#asahi-dev · 2023-04-06 21:31
jannauthe libinput list-devices output has nothing except the name for matching#asahi-dev · 2023-04-06 21:28
jannauin `udevadm info -t` search for ID_INPUT_TOUCHPAD=1#asahi-dev · 2023-04-06 21:28
jannauvendor id 0x05ac, product id is unclear since it comes from the device#asahi-dev · 2023-04-06 21:21
jannauname is "Apple Internal Keyboard / Trackpad"#asahi-dev · 2023-04-06 07:09
jannaumacos doesn't support routing dcp output to usb-c for clamshell mode on m2 laptops#asahi-dev · 2023-04-06 07:07
jannaui.e the panel on m2 laptops works even with macos 13.2 firmware and would expect the same for the m2 pro/max laptops#asahi-dev · 2023-04-06 07:05
jannauthe dp phy for dcp is only in the m2/m2 pro mac mini ADT#asahi-dev · 2023-04-06 07:04
jannauI wouldn't expect any issues on the laptops. iboot should still initialize the display and the dp phy seems to be handled by dcp firmware#asahi-dev · 2023-04-06 06:38
jannaudcp logs "AP not ready" on display hotplug events#asahi-dev · 2023-04-06 06:37
jannauI have the setup in m1n1 but apparently it looks like the dp port setup might be incompatible with the simplified dcb_iboot protocol used in m1n1#asahi-dev · 2023-04-06 06:35
jannauthe problem with display out on the m2 (pro) mac minis is that the dcp firmware doesn't handle the edp -> dp2hdmi setup anymore. dcp acts more like dcpext and its output can also be routed to the usb-c ports#asahi-dev · 2023-04-05 10:13
jannauno, the dependency is listed#asahi-dev · 2023-04-05 10:09
jannausigh, the alarm xf86-input-libinput pkg seems not to depend on libinput#asahi-dev · 2023-04-03 22:29
jannaukettenis_: I'll test if fw upload works in linux after disabling the u-boot usb hcd driver #asahi-dev · 2023-04-03 22:28
jannauadded to the Hypervisor page#asahi-dev · 2023-04-03 22:23
jannauthat becomes annoying when we can support the fw upload since it is powered on on boot so we don't need the pwr-enable gpio#asahi-dev · 2023-04-03 22:22
jannauand skip it since u-boot misses a driver for that#asahi-dev · 2023-04-03 22:21
jannauadhoc solution was to parse the smc gpios#asahi-dev · 2023-04-03 22:20
jannaufor the asmedia usb hcd I can only think of skipping it in u-boot #asahi-dev · 2023-04-03 22:19
jannauI'll send that tomorrow#asahi-dev · 2023-04-03 22:19
jannaukettenis_: only thing I've come up with that works is not init port with up link status in linux#asahi-dev · 2023-04-03 22:18
jannauChaosPrincess: "clpc=0" in bootargs#asahi-dev · 2023-04-02 14:51
jannaulibsensors is not needed#asahi-dev · 2023-04-02 14:49
jannaullvm might be a problem#asahi-dev · 2023-04-02 14:42
jannauthe flatpak is ideally build on the same system as the asahi linux PKGBUILDs without involvment of any github workflow#asahi-dev · 2023-04-02 14:41
jannausigh#asahi-dev · 2023-04-02 14:40
jannauideally this would not involve github in any way#asahi-dev · 2023-04-02 14:31
jannauthat can be an additional source for the mesa package#asahi-dev · 2023-04-02 14:19
jannaukettenis_: https://github.com/AsahiLinux/linux/commit/a636217494a787bbb92ae12976e086779b6694bf#asahi-dev · 2023-04-01 15:05
jannauI guess we want to use EPICProviderClass instead of interface-name#asahi-dev · 2023-04-01 14:17
jannauI wonder if they fixed this in 13.3 or 13.4 beta#asahi-dev · 2023-04-01 14:16
jannaubut there are of course variable length suffixes#asahi-dev · 2023-04-01 13:48
jannauI guess we should just ignore '^dispext[0-9]+:' for the service matching#asahi-dev · 2023-04-01 13:46
jannauwith that fixed dp audio still works on the 13.2 fw#asahi-dev · 2023-04-01 13:44
jannauat least it's consistent on all afk/epic endpoints now#asahi-dev · 2023-04-01 13:43
jannauI see those with the dcp firmware from macos 13.2#asahi-dev · 2023-04-01 13:37
jannausigh, apple: "dispext4294967295:dcpav-video-i" and "dispext4294967295:dcpav-audio-i"#asahi-dev · 2023-04-01 12:57
jannaupushed current state to https://github.com/jannau/linux/tree/dcpext-6.3-rc4-0_dcp-v13_2