#asahi-dev · 2025-03-18 00:24
nickchanfl0_id: the cpu start interface has remained compatible for like some 10 (?) years#asahi-dev · 2025-03-17 04:22
nickchanbut then again, that machine barely has any reason to exist#asahi-dev · 2025-03-17 04:22
nickchanis the j180d pcie slots like not functional without smc#asahi-dev · 2025-03-15 18:38
nickchani find the cpu start offset by bruteforcing#asahi-dev · 2025-03-15 04:58
nickchan>s3_1_c15_c1_5 is _EL1, s3_4_c15_c14_5 is probably _EL2 and s3_4_c15_c14_6 is probably _EL12 #asahi-dev · 2025-03-15 04:52
nickchanand as I said before some new registers like the one that controlled timer frequency has VHE schematics which is another source of confusion#asahi-dev · 2025-03-15 04:51
nickchanand the next one s3_0_c15_c2_4 should be HID31?#asahi-dev · 2025-03-15 04:50
nickchanfl0_id: s3_0_c15_c2_3 should be HID30#asahi-dev · 2025-03-07 05:14
nickchannicolas17: it's all using existing data#asahi-dev · 2025-03-07 04:13
nickchanabout the wiki#asahi-dev · 2025-03-07 04:13
nickchanchannel#asahi-dev · 2025-03-07 04:13
nickchansomeone shoule go update the chanel topic#asahi-dev · 2025-03-07 04:13
nickchananyways#asahi-dev · 2025-03-07 01:58
nickchansven: will you apply the a7-a11 cache nodes for this cycle? I already mentioned it should be fine on the list that it is only used for reporting to userspace in sysfs#asahi-dev · 2025-03-06 18:15
nickchanand delay may be above 32-bit and i am not sure the kernel it would like that#asahi-dev · 2025-03-06 18:11
nickchananyways this means that I would have to schedule_delayed_work() up to 194 days and i have no idea how i am going to test that in a short time#asahi-dev · 2025-03-06 18:09
nickchanpresumably since a12 otherwise dtk would have massive issues with time#asahi-dev · 2025-03-06 18:08
nickchan>sets the offset itself on m1#asahi-dev · 2025-03-06 18:06
nickchanwould not be suprising if smc sets the offset itself on m1 so that doesn't happen#asahi-dev · 2025-03-06 18:05
nickchansince whatever time setting set by other software on a7-a11 is overwritten on iOS boot, which is needed so the rtc does not go wrong even if someone reboots the device right when the counter is about to overflow#asahi-dev · 2025-03-06 18:03
nickchansmc may be doing some heavy lifting on m1#asahi-dev · 2025-03-06 18:03
nickchanand OSes will overwrite each other's time configuration, evident if setting time from iOS restore ramdisk#asahi-dev · 2025-03-06 18:02
nickchanon a7-a11 the counter can only count like 194 days and cannot be reset so it requires OS interevention to stay correct#asahi-dev · 2025-03-06 17:58
nickchanalso#asahi-dev · 2025-03-06 17:50
nickchanbut on m1 the rtc definitely has more bits for the offset#asahi-dev · 2025-03-06 17:49
nickchanfunnily enough a11 is still subject to year 2038 problem#asahi-dev · 2025-03-06 17:44
nickchanspmi pmu rtc should exist on m1 too but asahi device tree makes no reference to it#asahi-dev · 2025-03-06 17:01
nickchanmore information tends comes up as research is conducted#asahi-dev · 2025-03-06 17:00
nickchanokay pmu reset panic counter does apply on a7-a11#asahi-dev · 2025-03-05 15:10
nickchanwell, it could be komodo but i don't think this is the case either#asahi-dev · 2025-03-05 15:07
nickchanand hidra smells more like m5 to me#asahi-dev · 2025-03-05 15:06
nickchanno hints in codenames #asahi-dev · 2025-03-05 15:06
nickchannothing lines up other than the cpid 6032 and 6033#asahi-dev · 2025-03-05 14:56
nickchanstill better than nvidia#asahi-dev · 2025-03-05 14:52
nickchanthis must be a different die#asahi-dev · 2025-03-05 14:52
nickchanm3 ultra with maximum **512** GB of unified memory#asahi-dev · 2025-03-05 03:37
nickchanactually it's not in torvald's tree so it might be changable?#asahi-dev · 2025-03-05 03:33
nickchanand i am pretty sure that compatible comes from a certain string in apple's drivers#asahi-dev · 2025-03-05 03:33
nickchanchadmed: upstreamed so it's there forever#asahi-dev · 2025-03-05 03:32
nickchanwell I could, what I mean is I would have to declare a non-functional adp in the fdt#asahi-dev · 2025-03-05 03:32
nickchanhowever, I can't use the MIPI part by itself#asahi-dev · 2025-03-05 03:18
nickchanthis seals it being incompatible #asahi-dev · 2025-03-05 03:17
nickchansorry, I mean 0x218001030#asahi-dev · 2025-03-05 03:15
nickchanscreen size is found in a few places like 0x218001020#asahi-dev · 2025-03-05 03:15
nickchan0x218200000 appears to be a clone of 0x218000000 #asahi-dev · 2025-03-05 03:14
nickchansearched ADP_SCREEN_SIZE in a11's display pipe's mem range#asahi-dev · 2025-03-05 01:31
nickchanchaos_princess: weird#asahi-dev · 2025-03-04 14:18
nickchanbetas are unsigned regularly though signed for a lot longer compared to *most* iOS versions#asahi-dev · 2025-03-04 03:46
nickchanno notable new logic I think#asahi-dev · 2025-03-04 03:45
nickchanmostly interpolating existing M1 and M2 support