#asahi-dev · 2023-09-07 13:28
marcanthen we can also put the silly hardcoded table in m1n1#asahi-dev · 2023-09-07 13:28
marcanworking on it#asahi-dev · 2023-09-07 13:27
marcanI'm saying we just alloc the heap in m1n1 and treat it like another region, so it just works#asahi-dev · 2023-09-07 13:27
marcanlinux does the mapping, m1n1 just tells it via reserved-regions#asahi-dev · 2023-09-07 13:27
marcanoh on power off?#asahi-dev · 2023-09-07 13:17
marcanI'm trying out this idea and cleaning up the m1n1 side a bit#asahi-dev · 2023-09-07 13:17
marcant8103 m1 air, should be the same as you#asahi-dev · 2023-09-07 13:11
marcaneiln: I have this idea that we just carve out the heap in m1n1 and drop it in as a reserved region and let the linux side be oblivious to it#asahi-dev · 2023-09-07 12:24
marcanargs.unk_size = 0x1800000; that sure looks like the heap size thing#asahi-dev · 2023-09-07 12:18
marcannow let me see if I can clean up the DART stuff a bit#asahi-dev · 2023-09-07 12:18
marcanwith that removed it works \o/#asahi-dev · 2023-09-07 12:16
marcanor was this broken in the asahi branch still? (I rebased on asahi-wip, please always use that as a base; asahi is more so random packagers don't package random dev kernels, but I keep forgetting to update asahi...)#asahi-dev · 2023-09-07 12:14
marcaneiln: what's up with apple_isp_resv_region? that stuff is already handled in common code, no? it actually fails for me because of_iommu already took care of it and the mappings end up duplicated...#asahi-dev · 2023-09-07 11:49
marcangoing to poke around on an M1 air and see if I can clean up a bit#asahi-dev · 2023-09-07 11:07
marcaneiln: left some comments on the m1n1 PR, but I'm tempted to hack on this a bit myself if you don't mind#asahi-dev · 2023-09-06 06:58
marcanyeah, it looks like all the overrides in the Apple Silicon range are completely boring#asahi-dev · 2023-09-06 06:56
marcanso I guess we don't need to worry about magic color profiles then#asahi-dev · 2023-09-06 06:55
marcanI think it's just supposed to be Display P3 and DCP/DISP do the actual calibration behind the scenes#asahi-dev · 2023-09-06 06:55
marcanand then I realized switching between the custom color profile and Display P3 in system settings does nothing...#asahi-dev · 2023-09-06 06:55
marcanhuh, I spent way too much time trying to figure out how to map from machine to the display profile override plists... only to find at least some/most of the AS ones are basically empty#asahi-dev · 2023-09-05 15:27
marcanyeah, I should fix that. not sure how it got left out.#asahi-dev · 2023-09-05 09:16
marcanI forgot we still hardcode mem maps there... :p#asahi-dev · 2023-09-05 09:15
marcanyeah#asahi-dev · 2023-09-03 18:54
marcant8103 is tested and improved, so just t8112 left#asahi-dev · 2023-09-03 18:54
marcanI should get some sleep, but if you get a chance t8112 tests would be good too :)#asahi-dev · 2023-09-03 18:54
marcanexcellent#asahi-dev · 2023-09-03 18:46
marcanI'll push 1.3.6 to prod, seems to be ok#asahi-dev · 2023-09-03 18:32
marcanyes#asahi-dev · 2023-09-03 18:30
marcantagged v1.3.6 with the t8112 fix then#asahi-dev · 2023-09-03 18:28
marcanI'm very confused as to how that m1n1 change could break selinux#asahi-dev · 2023-09-03 18:26
marcanI pushed a fix for t8112 that may or may not matter#asahi-dev · 2023-09-03 18:25
marcandid it work the second time?#asahi-dev · 2023-09-03 18:25
marcanoh yeah that would happen#asahi-dev · 2023-09-03 18:18
marcanthat sounds... weird#asahi-dev · 2023-09-03 18:07
marcanjannau: try https://github.com/clamchowder/Microbenchmarks/tree/master/MemoryLatency , that shows a definite improvement on t6000 around the point where it spills from L2 to SLC#asahi-dev · 2023-09-03 17:59
marcanjannau: try 1.3.5 on t602x#asahi-dev · 2023-09-03 16:55
marcanfound another bug on t600x that would make it not really work, heh#asahi-dev · 2023-09-03 16:49
marcansigh#asahi-dev · 2023-09-03 16:49
marcanno surprise there, I just winged that#asahi-dev · 2023-09-03 16:49
marcanwhoops#asahi-dev · 2023-09-03 16:40
marcanmight break t8112/t602x though#asahi-dev · 2023-09-03 16:40
marcanah so this at least came after ultra, less likely to break the world then#asahi-dev · 2023-09-03 16:19
marcanpushed m1n1 v1.3.4 to asahi-dev with a really, really stupid bugfix that should improve performance but needs testing on all major chip variants to make sure I didn't break the world#asahi-dev · 2023-09-02 14:28
marcanpushed kernel/mesa updates (to dev and prod, passed a basic smoke test)#asahi-dev · 2023-09-01 18:48
marcanno need for a KDK, you can just run the retail kernel#asahi-dev · 2023-09-01 15:58
marcanand only bothered to upgrade all the new chassis variants#asahi-dev · 2023-09-01 15:57
marcanthe m1 air/pro and m2 pro are ancient case designs, they probably kept whatever old cameras those always had#asahi-dev · 2023-09-01 15:57
marcannot terribly surprising#asahi-dev · 2023-09-01 15:57
marcaneiln: I guess for the new chassis design they upgraded it#asahi-dev · 2023-09-01 07:36
marcanxkb has a bunch of complicated features to support this stuff