#
✕ nick:jannau
#asahi-dev · 2023-04-13 07:30
jannaumaz: would it make sense to support something like https://lore.kernel.org/lkml/YE+avkajxbGJMOoM@krava/T/ ? that looks to me like more intuitive default behavior
#asahi-dev · 2023-04-13 07:21
jannauwith the pcie fix the bringup branch works on m1/m1 max/m2 with dcp firmware 12.3 and 13.2
#asahi-dev · 2023-04-13 07:20
jannaudid you see https://github.com/AsahiLinux/linux/pull/134 ? the ps_isp_sys change is maybe worth looking into further
#asahi-dev · 2023-04-13 07:17
jannaumarcan: np, changes are not complicated and I had them most of them since the m2 mac mini arrived
#asahi-dev · 2023-04-12 17:57
jannauworks for me on m2 mbp 13"
#asahi-dev · 2023-04-12 17:56
jannau`taskset -c $CORE perf stat sleep 1` with 0 and 4 for $CORE
#asahi-dev · 2023-04-12 17:54
jannaueric_engestrom: try -1
#asahi-dev · 2023-04-12 17:39
jannauon m2* it's apple_avalanche_pmu/cycles/ and apple_blizzard_pmu/cycles/
#asahi-dev · 2023-04-12 17:39
jannausee perf list about which cycle and instruction counters you have
#asahi-dev · 2023-04-12 17:38
jannauit should work on one or the other. I'm not sure how perf/ the kernel selects which counters are cycles/instructions but it mappes only to either one
#asahi-dev · 2023-04-12 17:36
jannaueric_engestrom: `perf stat sleep 1` is generally a nice way to test if perf works as expected. can you try that on both the performance and efficiency cores
#asahi-dev · 2023-04-12 17:31
jannauah, the compatible is missing a 'z'
#asahi-dev · 2023-04-12 17:27
jannausomehow the blizzard pmu went missing on m2 pro
#asahi-dev · 2023-04-12 17:22
jannau/proc/sys/kernel/perf_event_paranoid is set to allow user event sampling?
#asahi-dev · 2023-04-12 17:16
jannaudoes perf record -e cpu-clock ... work?
#asahi-dev · 2023-04-12 17:14
jannauthat said, a plain `perf record sleep 1` produces samples for me on m1 max
#asahi-dev · 2023-04-12 17:14
jannauif I remembered. with hw counters you need to taskset either efficiency or performance cores
#asahi-dev · 2023-04-12 17:06
jannaueric_engestrom: we have seen that before
#asahi-dev · 2023-04-12 17:06
jannaukettenis: t602x/bringup wifi/bt fixes are in https://github.com/AsahiLinux/linux/pull/134
#asahi-dev · 2023-04-12 14:28
jannaukettenis: it should overwrite the relevant bits but I forgot to look for the board type
#asahi-dev · 2023-04-11 23:12
jannaudo we have a problem with dart runtime PM? dart is probed first due to the iommu dependency, might immediately power-down since it is unused. if the device is not so far probed that it forces the pd on, the pd will power down
#asahi-dev · 2023-04-11 23:02
jannauadd pd_ignore_unused to the kernel command line or make the power-domain always-on. Is it shared between two or more devices?
#asahi-dev · 2023-04-11 22:39
jannauwith working Night Color (kwin_wayland 5.27.4)
#asahi-dev · 2023-04-11 22:37
jannaut602x/bringup works otherwise on mac studio with both 12.3 and 13.2 firmware
#asahi-dev · 2023-04-11 22:36
jannaumarcan: small PCIe copy'n'paste fix for pre t6020-pcie in https://github.com/AsahiLinux/linux/pull/134
#asahi-dev · 2023-04-11 12:24
jannaupossibly. there should be 3 insertions before D116
#asahi-dev · 2023-04-11 12:22
jannauI didn't look at driver or firmware
#asahi-dev · 2023-04-11 12:22
jannauipc.py can handle both 12.3 and 13.2 but needs the FW version set via env variable for 13.2
#asahi-dev · 2023-04-11 12:20
jannauI left the call/callback list in trace_dcp as is
#asahi-dev · 2023-04-11 12:19
jannaumy dcp 13.2 m1n1 work tree is in https://github.com/jannau/m1n1/tree/m2_dptx
#asahi-dev · 2023-04-11 12:18
jannauI think create_nvram_service was inserted as D112
#asahi-dev · 2023-04-11 12:15
jannauI guess either you or me have to look at macos traces
#asahi-dev · 2023-04-11 12:15
jannauso one of the new callbacks moving the others around. not used on m1*/m2
#asahi-dev · 2023-04-11 12:10
jannauplease add a noop callback for that, should be set_idle_caching_state_ap
#asahi-dev · 2023-04-11 12:02
jannaulol, I grepped for __builtin_clz and was then confused by 'return x ? sizeof(x) * 8 - __builtin_clz(x) : 0;' in builtin-fls.h. I didn't parse that fully and assumed it is a macro taking the width of the argument into account
#asahi-dev · 2023-04-11 11:52
jannauthe bus numbers have a gap at the disabled port
#asahi-dev · 2023-04-11 11:50
jannauor an OF iommu/pcie bug
#asahi-dev · 2023-04-11 11:50
jannaulooks like the dart <-> port mapping breaks if there are gaps in the ports
#asahi-dev · 2023-04-11 11:48
jannauyes
#asahi-dev · 2023-04-11 11:44
jannaumarcan: pcie is confused by disabling port01. ethernet and usb xhci work if port01 is left enabled. pcie0_dart_1 needs to be disabled though. SErrors on probe
#asahi-dev · 2023-04-11 11:24
jannauthat sounds plausible
#asahi-dev · 2023-04-11 11:14
jannauI didn't try hard though since the locked dart io-pgtbl interaction needs to be rewritten
#asahi-dev · 2023-04-11 11:13
jannauthe reserved mem code seems to be broken with 4 levels in a way I do not understand
#asahi-dev · 2023-04-11 11:12
jannauand https://github.com/jannau/linux/commit/d45faf548f63b4a334561376f838c56d1b63787e
#asahi-dev · 2023-04-11 11:12
jannaudo you have https://github.com/jannau/linux/commit/ca36a0932a5303abea73bc5ebecd40d2c6ef18e8
#asahi-dev · 2023-04-11 11:09
jannauwhen does the screen break? during linux dart probe or earlier?
#asahi-dev · 2023-04-11 11:08
jannaudoes it lock dart-disp0?
#asahi-dev · 2023-04-11 11:03
jannaudart_setup_pt_region() is only called from dcp_init() so it shouldn't be used
#asahi-dev · 2023-04-11 11:02
jannaubecause we don't touch the darts without the m1n1 change
#asahi-dev · 2023-04-11 11:01
jannauI thought that was just during fb mapping in display.c

plus ›