#asahi-dev · 2023-11-28 14:38
svenhuh, that does sound strange#asahi-dev · 2023-11-28 14:34
svenMTX should only matter if MTE is available but I don't think the M3 has that#asahi-dev · 2023-11-28 14:29
svenbefore mmu_configure? I don't think so#asahi-dev · 2023-11-28 14:29
svenhrm, how would TCR matter before the MMU is on?#asahi-dev · 2023-11-28 14:27
svenhuh#asahi-dev · 2023-11-28 14:27
svenor, wait, you first set it to the XNU value and then later m1n1 overwrites it and the MMU works then?#asahi-dev · 2023-11-28 14:26
svenhmm... how does the XNU TCR differ from ours? #asahi-dev · 2023-11-28 14:24
svenj`ey: i'm only asking dumb questions anyway because that usually helps when i'm stuck at weird bugs like that :D#asahi-dev · 2023-11-28 14:23
svenI don't remember if the initial pagetables already use SPRR but they might#asahi-dev · 2023-11-28 14:21
svenhave you checked if the initial xnu pagetables look weird in some way or how that TCR value XNU differs from the one inside m1n1?#asahi-dev · 2023-11-28 14:21
svennice :)#asahi-dev · 2023-11-28 14:17
svenI think it's just the call to mmu_init_sprr() which may do weird things if they changed those regs#asahi-dev · 2023-11-28 14:16
svenhrm, no, looking at it again we leave SPRR/GXF disabled in the MMU init and only prepare everything for the HV#asahi-dev · 2023-11-28 14:15
svenI'd try to remove the sprr/gxf init from m1n1's mmu and see if that changes anything. I think we only need that for the HV#asahi-dev · 2023-11-28 14:13
svenif the chickens are correct maybe they changed something with how SPRR/GXF works?#asahi-dev · 2023-11-28 14:12
svenhrm, i think when i wrote the original MMU code I didn't rely on any previous state of the architectural registers#asahi-dev · 2023-11-28 12:41
svenmainly because there were some reports of usb3 being a bit flaky but I think that's because I have the atcphy/dwc3 order wrong in the current kernel#asahi-dev · 2023-11-28 12:39
svencan anyone who has multiple M-series SoCs connect the same usb 3 device in the same orientation to all of them and trace atcphy? i want to confirm that there are no differences between the various generations#asahi-dev · 2023-11-27 10:20
svennext up is atcphy though I'll need a bit more clean up there and probably some testing downstream before I send that one out#asahi-dev · 2023-11-27 10:19
svennice :)#asahi-dev · 2023-11-26 15:36
svenit at least worked when I originally reverse engineered altmode#asahi-dev · 2023-11-26 15:36
svenuh, i think it should#asahi-dev · 2023-11-26 15:27
sven<sven> okay, pretty sure 0x24 is something like TPS_REG_USB4_STATUS and bit 23 inside TPS_REG_DATA_STATUS is "USB4 connection active" instead of TPS_DATA_STATUS_FORCE_LSX on the cd321x <-- double checked with the cd321x firmware and it looks like that's correct#asahi-dev · 2023-11-26 14:14
sveni don't even know if they exist for usb 4 yet#asahi-dev · 2023-11-26 14:13
svenyeah, i don't expect anyone to have it because those are very rare and really only needed if you need >5m afaict#asahi-dev · 2023-11-26 14:12
svenoh, the important part is "optical" and it can be either thunderbolt or usb 4#asahi-dev · 2023-11-26 14:11
svenit doesn't matter much because if no one has such a cable it just won't be support for now#asahi-dev · 2023-11-26 14:11
svenyou'll know if it's optical because it'll be very expensive and probably also >5m or so#asahi-dev · 2023-11-26 14:06
svendoes anyone here have an optical thunderbolt or usb 4 cable and the ability to run macos under the m1n1 hv?#asahi-dev · 2023-11-26 11:21
svenokay, pretty sure 0x24 is something like TPS_REG_USB4_STATUS and bit 23 inside TPS_REG_DATA_STATUS is "USB4 connection active" instead of TPS_DATA_STATUS_FORCE_LSX on the cd321x#asahi-dev · 2023-11-25 21:07
svenno idea what the last 4 bytes then are#asahi-dev · 2023-11-25 21:06
svenfirst 8 bits are some kind of status flags#asahi-dev · 2023-11-25 21:06
svenyup, looks like byte 2-5 in 0x24 are the contents of EUDO sent by the cd321x#asahi-dev · 2023-11-25 18:31
svenah, thanks! current candidate is 0x24#asahi-dev · 2023-11-25 16:15
svenanyone with a usb4 device (e.g. something goshen ridge based) here who's also running macos and can run hpmdiagnose before it's connected and then again afterwards for me?#asahi-dev · 2023-11-01 13:50
svenhmm.. so the cd321x firmware still supports overriding the automatically selected alt mode if some flag is set. just haven’t been able to figure out how to set that flag yet#asahi-dev · 2023-10-31 08:09
svensmells like some weird race :/#asahi-dev · 2023-10-31 08:09
svenit happens very reliably when i enable thunderbolt debug logs but it also happens occasionally if that's not enabled#asahi-dev · 2023-10-31 08:09
svenyup#asahi-dev · 2023-10-31 08:04
svenokay, looks like any pcie bridge reconfiguration actually has to be locked against any NHI tunnel reconfiguration (possible even any NHI register write to any peripheral) or the co-processor will die if you're unlucky :(#asahi-dev · 2023-10-29 12:02
svenit probably just doesn’t have a sign off because I forgot to add that flag when I ran git commit #asahi-dev · 2023-10-29 12:02
svensure #asahi-dev · 2023-10-29 11:51
svenbut there was also a function that told you if the display was tiled iirc#asahi-dev · 2023-10-29 11:51
svenhrm... it was related to DPTX_APCALL_SET_TILED_DISPLAY_HINTS somehow#asahi-dev · 2023-10-29 11:49
svenif two displays are connected to that port instead two separate dcpext dptx0 streams are connected to those ports#asahi-dev · 2023-10-29 11:49
sventhey're connected to the two dp-over-tbt ports of a single atcphy if a split panel display is connected there#asahi-dev · 2023-10-29 11:48
svenyeah, that's the split panel thing#asahi-dev · 2023-10-29 11:48
sventhat's used for split panel displays#asahi-dev · 2023-10-29 11:48
svenoh, inside dcp you mean