#asahi · 2021-09-11 13:01
marcanI don't have a strong opinion there, I guess it comes down to what that buys us#asahi · 2021-09-11 13:01
marcanseparate rtkit node is basically what macos does, fwiw#asahi · 2021-09-11 12:59
marcan< sven> the other two irqs are also symmetric fwiw. i think they literally just hooked up the same HW block to both processors <- just like the wii :p#asahi · 2021-09-10 11:55
marcanI do not miss scheduling meetings those 2 weeks while I was working at G#asahi · 2021-09-06 16:45
marcanI get the feeling they hardcode half these things and that devices table is an incomplete mess#asahi · 2021-09-06 16:45
marcanalso -V usually means power/voltage I think?#asahi · 2021-09-06 16:44
marcanall zeros? there are like literally no ids there#asahi · 2021-09-06 16:44
marcan #246 CIO0_RECONFIG-V id: 337 gate: 0 psreg: 0 flags: 00000010 unk: 0x000 unk3: 0 0 0 0 @ #asahi · 2021-09-06 16:42
marcanheh#asahi · 2021-09-06 16:42
marcanso I wouldn't expect syslogs#asahi · 2021-09-06 16:42
marcanthere are few strings in there#asahi · 2021-09-06 16:42
marcanalso the cio firmware is like 100kB and I'm not even sure if it has any printfs#asahi · 2021-09-06 16:40
marcansven: \o/#asahi · 2021-09-06 16:40
marcanyes#asahi · 2021-09-06 16:40
marcangeneric power domain#asahi · 2021-09-06 16:39
marcanfwiw, the reason I haven't sent the clocks patch yet is I need to give genpd a good look, seeing as apparently those pmgr bits are more than just clock gates#asahi · 2021-09-06 03:26
marcanthat would also allow us to stop caring about decompression of initramfs payloads, and leave it to linux, which saves boot memory and probably time if Linux's decompressor is better#asahi · 2021-09-06 03:25
marcaneven moreso really#asahi · 2021-09-06 03:25
marcanso is cpio#asahi · 2021-09-06 03:24
marcanit might be time to start considering using an actual archive format (cpio?)#asahi · 2021-09-06 03:23
marcanit's a bit awkward#asahi · 2021-09-06 03:23
marcansince it can't locate any payloads after that#asahi · 2021-09-06 03:23
marcanj_ey: yes, uncompressed kernels need to come *last* for this reason#asahi · 2021-09-04 12:27
marcanhttps://www.st.com/resource/en/technical_article/ta0356-usb-typec-and-power-delivery-displayport-alternate-mode-stmicroelectronics.pdf if you want to see actual protocol details while still not being the spec#asahi · 2021-09-04 12:26
marcandottedmag: https://www.ti.com/lit/wp/slyy109a/slyy109a.pdf is probably as close to something digestible as you're going to get#asahi · 2021-09-04 12:23
marcansven: we already had to deal with that spec for the serial port stuff :p#asahi · 2021-09-04 12:22
marcaninstead of apple quirks we may have to deal with broken broadcom silicon :p#asahi · 2021-09-04 12:22
marcanalyssa: don't speak about apple quirks that fast... tg3 on my intel iMac is unstable, especially with the IOMMU turned on :p#asahi · 2021-09-02 17:53
marcanthis is pretty small though, it looks like it could be mainly sampling, not control, but I'm not sure#asahi · 2021-09-02 17:52
marcanthe PMP firmware definitely talks SPMI and ooooh here are the strings of the perf counters I was not finding in the ADT#asahi · 2021-09-02 17:51
marcanwell, that might be a good reason to implement SMP in the HV#asahi · 2021-09-02 17:50
marcanlol yeah#asahi · 2021-09-02 17:50
marcankettenis_: yeah, possibly#asahi · 2021-09-02 17:50
marcanso that could be done in software#asahi · 2021-09-02 17:50
marcanI know multiple devices in the ADT reference the same power rails to set a voltage#asahi · 2021-09-02 17:50
marcanI'm not sure if there's "voting" here per se, at the hw level#asahi · 2021-09-02 17:49
marcanmaybe this is also all automagic stuff controlled by the pmu or something...#asahi · 2021-09-02 17:49
marcanI only found like 4 proper cross-refs from the devices table to clocks, everything else is null, which is bizarre#asahi · 2021-09-02 17:49
marcanI still don't know how clocks are controlled, what regs they map to#asahi · 2021-09-02 17:48
marcan(that's the "everything but gpu/cpu" domain basically)#asahi · 2021-09-02 17:48
marcane.g. any SoC block requiring high performance can put the whole SoC power domain into a higher voltage state#asahi · 2021-09-02 17:47
marcanthey also use this device concept to trigger voltage/performance states for some devices#asahi · 2021-09-02 17:47
marcanand now I wonder if clocks is the right abstraction in linux for this, though it might be the best we can do for now#asahi · 2021-09-02 17:46
marcanit looks like this is a relatively high level abstraction for what power/idle state a given device is in (most of which only have 2 states)#asahi · 2021-09-02 17:46
marcani.e. we know we don't need to touch this to spin up cores, yet the ps regs for each CPU and the clusters overall accurately represent the current state of what CPUs have been spun up#asahi · 2021-09-02 17:45
marcanso there is clearly abstraction going on#asahi · 2021-09-02 17:45
marcanand there's things like the power states for the CPUs represent the currently running CPUs, but they obviously aren't turned on *that way*#asahi · 2021-09-02 17:45
marcanyes (which crashes the HV, but definitely pokes a pile of state regs)#asahi · 2021-09-02 17:44
marcanit's starting to sound like that#asahi · 2021-09-02 17:44
marcanI could only see actual power state changes with something obvious like pulling up a video that enables the video decoder