#asahi-dev · 2022-09-21 18:55
svenyes#asahi-dev · 2022-09-21 15:40
svenit should work then #asahi-dev · 2022-09-21 15:40
svenyup#asahi-dev · 2022-09-21 15:39
svenassuming it still applies and works on the current upstream kernel #asahi-dev · 2022-09-21 15:38
svenoh, yes. With that patch it should be fine#asahi-dev · 2022-09-21 15:38
sventhey can’t do dma without a working iommu #asahi-dev · 2022-09-21 15:37
svenI’m afraid so :(#asahi-dev · 2022-09-21 15:37
sven16k page size?#asahi-dev · 2022-09-21 15:37
svenhuh#asahi-dev · 2022-09-21 15:37
svendo they show up on lspci?#asahi-dev · 2022-09-21 15:36
svenon the Mac mini they work at usb3 speed with an upstream kernel#asahi-dev · 2022-09-21 15:35
svenI think so#asahi-dev · 2022-09-21 15:35
sventhey’re just some common controller attached via pcie #asahi-dev · 2022-09-21 15:35
svenusb type a should just work #asahi-dev · 2022-09-21 15:34
svenI could upstream that reset quirk but then we’d maybe have to drop it later if it turns out I have to do crazy stuff like make dwc3 a child of the phy or something like that#asahi-dev · 2022-09-21 15:34
svenyup#asahi-dev · 2022-09-21 15:33
svenit’s all entangled with dpalt mode and thunderbolt #asahi-dev · 2022-09-21 15:33
svenI’m not convinced they are correct yet #asahi-dev · 2022-09-21 14:41
sven(And disappears by the time it sends the reply for the hotplug function if you don’t open it in time)#asahi-dev · 2022-09-21 14:41
sveninstead I had to call open and getUnit on a new service that appears #asahi-dev · 2022-09-21 14:40
sveninstead I had to call#asahi-dev · 2022-09-21 14:40
sventhen I also had to send “hot plug change detected” which is another epic function but not wait for the reply #asahi-dev · 2022-09-21 14:40
sveneither marcan got some of those connevredTo,.. epic functions slightly wrong or apple changed them during an update #asahi-dev · 2022-09-21 14:31
sventhis is just gonna be something else that equally dumb like the fix that got HPD working but will still takes days to figure out I bet :(#asahi-dev · 2022-09-21 14:30
svenunsurprisingly no difference between my Mac mini and my mba either #asahi-dev · 2022-09-21 08:48
svenThat pmgr function also doesn’t seem to be called#asahi-dev · 2022-09-21 08:47
svendoesn’t look like there’s any pmgr magic #asahi-dev · 2022-09-21 07:04
svenI had to double check because my repository has become a bit messy at this point because I started trying random things because nothing was working :D#asahi-dev · 2022-09-21 05:25
svenpretty sure that’s what I currently use #asahi-dev · 2022-09-21 04:45
svengood idea, let me trace that later #asahi-dev · 2022-09-21 04:41
svenyes#asahi-dev · 2022-09-20 20:20
svenbut I’ve wanted to run m1n1 in el1 to check the atcphy bringup anyway #asahi-dev · 2022-09-20 20:19
svenI think there the more interesting stuff happens to the other endpoints #asahi-dev · 2022-09-20 20:12
svenill compare DCP and dcpext tomorrow then and see if there’s any difference #asahi-dev · 2022-09-20 20:11
svenhah, didn’t know about that#asahi-dev · 2022-09-20 19:59
svenah, that explains it. it just requested one of those 3 regs#asahi-dev · 2022-09-20 19:50
svendisp0 is also hardcoded in sr_mapDeviceMemoryWithIndex but changing that to dispext0 also doesn't make a difference#asahi-dev · 2022-09-20 19:47
svenfrom that I got bit=3, 0x23b738018 and 0x23bc3c000 for the regs#asahi-dev · 2022-09-20 19:47
svenso that's from my macos trace#asahi-dev · 2022-09-20 19:47
sven00000020 6e 02 00 00 00 00 00 00 00 73 11 00 04 00 00 00 |n........s......|#asahi-dev · 2022-09-20 19:47
sven00000030 42 4c 61 79 65 72 5f 4d 00 00 00 00 |BLayer_M.... |#asahi-dev · 2022-09-20 19:47
sven00000010 00 c0 c3 3b 02 00 00 00 72 61 63 74 03 00 00 00 |...;....ract....|#asahi-dev · 2022-09-20 19:47
sven00000000 00 ff ff ff 01 00 52 47 18 80 73 3b 02 00 00 00 |......RG..s;....|#asahi-dev · 2022-09-20 19:46
svenI changed them to what I observed in my macOS trace and nothing changed#asahi-dev · 2022-09-20 19:43
svenalso bit=3 instead of 2 there but that also doesn't help#asahi-dev · 2022-09-20 19:40
svenyeah, it's on#asahi-dev · 2022-09-20 19:39
sven0x23b738018 is passed instead of 0x23b738014 in rt_bandwidth_setup_ap but that doesn't seem to make a difference#asahi-dev · 2022-09-20 19:37
sventrue#asahi-dev · 2022-09-20 19:36
svenah.. could be just mislabled ADT or something then#asahi-dev · 2022-09-20 19:33
sventhat... sounds wrong.. shouldn't it change the dispext clk?