#asahi · 2021-09-18 12:33
kettenisI butchered the u-boot device trees a bit as well when adding the j293 device tree#asahi · 2021-09-18 12:31
kettenisnot sure if the dt schema stuff checks for the names#asahi · 2021-09-18 12:31
kettenis(masquerading as gpio)#asahi · 2021-09-18 12:30
kettenisI believe pinctrl_xxx labels are preferred since these are pinctrl nodes#asahi · 2021-09-18 12:29
kettenisfrankly, that makes sense#asahi · 2021-09-18 12:26
kettenisalyssa: yes that is why I picked those names#asahi · 2021-09-18 12:24
kettenisI think we need to trick marcan into coordinating some of the DT additions#asahi · 2021-09-18 12:18
ketteniswishes I would find the OpenBSD bug that causes SMP crashes#asahi · 2021-09-18 12:08
kettenisj_ey: where are you keeping the pinctrl bits?#asahi · 2021-09-18 12:07
kettenisthe last commit of that series is sweet ;)#asahi · 2021-09-17 12:44
kettenis0x800#asahi · 2021-09-17 12:43
kettenisthe "enable" bit#asahi · 2021-09-17 12:43
kettenisor does it set and clear it around transactions?#asahi · 2021-09-17 12:43
kettenisdoes macos just set that bit and leave it on?#asahi · 2021-09-17 12:19
kettenisah, apparently systemd is in the EFI bootloader game now as well#asahi · 2021-09-17 12:07
kettenismaybe they still have code to support the original pasemi chips#asahi · 2021-09-17 12:07
kettenissmells like it#asahi · 2021-09-17 11:14
kettenisah, there is required-opps to encode dependencies between OPP table nodes#asahi · 2021-09-17 11:10
kettenisfun#asahi · 2021-09-17 11:09
ketteniss/svan/sven/#asahi · 2021-09-17 11:09
kettenissvan: you mentioned resetting the PHY is really broken in device mode, but even in host mode that doesn't work reliably?#asahi · 2021-09-17 11:07
kettenismarcan: existing DT bindings in this area live in Documentation/devicetree/bindings/{opp,cpufreq}#asahi · 2021-09-17 11:03
kettenissven: wonder how I'm going to implement *that* in OpenBSD#asahi · 2021-09-17 10:59
kettenisif m1n1 puts things in a reasonable state, Linux would still run if this stuff doesn't work on a new machine#asahi · 2021-09-17 10:58
kettenismaybe it isn't too bad to rely on the driver here#asahi · 2021-09-17 10:57
kettenisbut I guess that's what you want to avoid#asahi · 2021-09-17 10:57
kettenisI fear the default way existing platforms handle that is by putting the magic values in the driver#asahi · 2021-09-17 10:55
kettenisdon't think they allow encoding magic register values#asahi · 2021-09-17 10:55
kettenisnot sure the standard OPP tables would work#asahi · 2021-09-17 10:54
kettenisbut I'm not sure there is a mechanism to describe dependencies#asahi · 2021-09-17 10:53
kettenisthey also get used for GPU and memory controllers#asahi · 2021-09-17 10:52
kettenisThe way "pstates" are typically represented in the DT is through OPP tables that describe the frequency and the associated voltage#asahi · 2021-09-17 10:51
kettenisI think the DT should describe all the supported states#asahi · 2021-09-16 18:18
kettenisouch#asahi · 2021-09-16 18:16
kettenisdon't really think that would be particularly helpful#asahi · 2021-09-16 18:16
kettenisso i just misunderstood what you meant#asahi · 2021-09-16 18:15
kettenisah#asahi · 2021-09-16 18:15
kettenissven: The cover letter of the v2 mailbox series said you added some more examples, but I didn't spot those...#asahi · 2021-09-16 18:12
kettenisthat fix would probably only fix color issues though#asahi · 2021-09-16 18:09
kettenisI may have fixed a framebuffer problem today (if you booted using EFI)#asahi · 2021-09-16 18:07
kettenisstarted upstreaming my u-boot code#asahi · 2021-09-16 15:06
kettenisthe mini should be fine; it has a fan#asahi · 2021-09-16 08:32
kettenisdon't think the current brcmfmac driver has a way to use that stuff#asahi · 2021-09-16 08:31
kettenismarcan: not sure what the wifi calibration stuff is for#asahi · 2021-09-14 13:09
ketteniss/bot/not/#asahi · 2021-09-14 13:09
kettenisbot in a way that is directly exposed to the OS#asahi · 2021-09-13 13:35
kettenisbut the U-Boot implementation will do what we need#asahi · 2021-09-13 13:34
kettenismarcan: it seems some x86 machine will use ELF/BOOT/BOOTX64.EFI on media that is actually removable#asahi · 2021-09-13 10:25
kettenisSo as long as you don't try to install another OS that overwrites that file, things will just work#asahi · 2021-09-13 10:22
kettenisI'm fairly confident that the same happens on arm64 systems (with EFI/BOOT/BOOTAA64.EFI)