#
✕ nick:marcan
#asahi-dev · 2022-10-31 14:30
marcanthis whole thing is stupid
#asahi-dev · 2022-10-31 14:30
marcanso if he doesn't play ball, we try to sneak it in because using MFD is clearly the correct solution, regardless of what the maintainer says; and if he blocks that, then we have no choice but to reinvent the MFD wheel because a Linux subsystem maintainer has decided some people shouldn't use his subsystem
#asahi-dev · 2022-10-31 14:29
marcanhe probably thinks this should be using regmaps or shared memory or something like most other MFD devices
#asahi-dev · 2022-10-31 14:29
marcanhe doesn't know what he wants, and he hasn't bothered to actually read and understand how the driver works
#asahi-dev · 2022-10-31 14:27
marcanrmk: if Lee doesn't reply I'd rather just resubmit it as platform and not Cc him, because he clearly doesn't want to work with us and I'm not interested in playing a guessing game with no feedback
#asahi-dev · 2022-10-31 13:36
marcansomething something linux power domains?
#asahi-dev · 2022-10-31 13:36
marcanI'm going with dumb then
#asahi-dev · 2022-10-31 13:36
marcansven: and do you get the same values back in linux?
#asahi-dev · 2022-10-31 13:34
marcan(or just add random delays)
#asahi-dev · 2022-10-31 13:34
marcanmaybe do any reads and print the values and cross check
#asahi-dev · 2022-10-31 13:34
marcandoes macos read anything? maybe it's polling for a bit to flip but when you trace it's slow enough it always is already done?
#asahi-dev · 2022-10-31 13:33
marcansven: for PLLs and things like that delays would often be necessary, yeah
#asahi-dev · 2022-10-31 06:57
marcanso it doesn't matter if we don't expect AS users to use xHCI in the initramfs for installers because it's broken in u-boot right now; it needs the firmware because the module is going to be in there anyway, and if it loads without the firmware then xHCI won't work *at all*
#asahi-dev · 2022-10-31 06:56
marcanand that will imply that on *this* platform you need the firmware
#asahi-dev · 2022-10-31 06:56
marcanbut even without that, there is a general issue because distros will want to put xhci_pci in the initramfs, so that on *other* platforms you can use those controllers at least
#asahi-dev · 2022-10-31 06:55
marcantpw_rules: it's conceivable we will add something to u-boot to load xHCI firmware in the future (once PCI xHCI works at all)
#asahi-dev · 2022-10-29 11:57
marcanrandom message improvements and probably other things I forgot
#asahi-dev · 2022-10-29 11:57
marcanalso, I promoted M2 to non-expert, and dropped 12.1 from the ipsw list
#asahi-dev · 2022-10-29 11:57
marcanthe repair option also works inline from paired mode (if paired recoveryOS actually works)
#asahi-dev · 2022-10-29 11:56
marcanotherwise it'll trigger a reboot/step2.sh path
#asahi-dev · 2022-10-29 11:56
marcanif you run m1n1 upgrade from recoveryOS paired with asahi, it'll do it all inline
#asahi-dev · 2022-10-29 11:55
marcanm1n1 upgrade does what it says on the tin
#asahi-dev · 2022-10-29 11:55
marcanrepair fixes botched step2.sh and cases where apple nuked our boot policy (it just redoes the last step of the installation)
#asahi-dev · 2022-10-29 11:55
marcanimproved APFS resize limits checking, the new firmware extractors, firmware.cpio, and the new repair and m1n1 upgrade options
#asahi-dev · 2022-10-29 11:55
marcanok, pushed new (hopefully working) asahi-installer to alx.sh/dev, with all the random things we've picked up over time
#asahi-dev · 2022-10-29 11:03
marcanprobably some syntax errors there :p
#asahi-dev · 2022-10-29 11:02
marcanchadmed: untested patch: https://mrcn.st/p/baH3NSlh
#asahi-dev · 2022-10-29 10:51
marcanlooks like the way it works is it looks for regulators to find the count, if it doesn't find them, it assumes 1 for the purposes of parsing the table, which is what you'd want
#asahi-dev · 2022-10-29 10:49
marcanwhich might need some rework, but that's kernel dev for you :)
#asahi-dev · 2022-10-29 10:48
marcanopp_parse_supplies just needs to be fixed so it doesn't assume everything needs a voltage
#asahi-dev · 2022-10-29 10:48
marcanusing the OPP is correct
#asahi-dev · 2022-10-29 10:48
marcan_get_dt_power still does the same thing with the OPP
#asahi-dev · 2022-10-29 10:47
marcanthat won't work
#asahi-dev · 2022-10-29 10:46
marcanthat needs fixing
#asahi-dev · 2022-10-29 10:46
marcanyeah, opp_parse_supplies
#asahi-dev · 2022-10-29 10:45
marcanif that doesn't work, that's a bug
#asahi-dev · 2022-10-29 10:45
marcanchadmed: Look at dev_pm_opp_of_register_em. It looks for microwatt, if it finds it, it elides the calculations.
#asahi-dev · 2022-10-29 10:40
marcanthat makes sense
#asahi-dev · 2022-10-29 10:39
marcanhow does EAS work exactly anyway?
#asahi-dev · 2022-10-29 10:38
marcanso maybe that binding should be changed, and just have that field dropped to a single u32 representing total power for this entry
#asahi-dev · 2022-10-29 10:38
marcanwhich is different from a power level you'd use for EAS
#asahi-dev · 2022-10-29 10:38
marcanas in some kind of cap
#asahi-dev · 2022-10-29 10:38
marcanthe binding implies this is to configure regulators
#asahi-dev · 2022-10-29 10:38
marcanOTOH, that entire binding is arguably silly
#asahi-dev · 2022-10-29 10:37
marcan(in case of multiple regulators)
#asahi-dev · 2022-10-29 10:37
marcanu32 array, more specifically
#asahi-dev · 2022-10-29 10:37
marcansee drivers/opp/of.c for how to parse it
#asahi-dev · 2022-10-29 10:36
marcanchadmed: the bindings don't say anything about types for these fields, which is arguably a bug, but I think the intent is clearly that opp-microwatt is u32
#asahi-dev · 2022-10-29 10:33
marcanwhat's the problem exactly?
#asahi-dev · 2022-10-29 10:33
marcanthe DT properties should have constant sizes regardless of arch

plus ›