#asahi-dev · 2025-12-18 22:00
jannauoliver_: did you look at the values of `ts64_unk`? at least at some point it looked like it might be a timestamp as well. I don't remember the details though#asahi-dev · 2025-12-18 21:54
jannauthanks, I'll probably will only merge it for 6.18#asahi-dev · 2025-12-18 21:14
jannauthe application sets min and max refresh rate properties#asahi-dev · 2025-12-18 21:12
jannaunot sure how to determine with the internal displays that they are actually running with VRR#asahi-dev · 2025-12-18 21:11
jannauchaos_princess: the driver needs to set a vrr property on the connector and I think we need a EDID#asahi-dev · 2025-12-18 21:09
jannauideally only for the 120HZ macbook modes#asahi-dev · 2025-12-18 21:08
jannausetting them to the meaningless value 120#asahi-dev · 2025-12-18 21:06
jannaupower saving. probably doesn't matter overall at the current state of Pro/Max macbooks#asahi-dev · 2025-12-18 21:03
jannaumeh. not really closer to VRR then :(. let's set them to 120 for now and enable the 120Hz mode#asahi-dev · 2025-12-18 12:03
jannaubut yes. for a start tou could look at do a trace_device and/or trace_range in the hypervisor#asahi-dev · 2025-12-18 12:01
jannaunot quite. the reg values have to be translated by based on /arm-io/ranges m1n1's adt code already does this#asahi-dev · 2025-12-17 17:09
jannauoliver_: https://gist.github.com/jannau/522cccad57274dbefe9ae174c6df1143#asahi-dev · 2025-12-17 16:55
jannauoliver_: HV has `show_timestamps` which reads CNTPCT_EL0#asahi-dev · 2025-12-17 16:52
jannauthe other columns are ts1, ts2 - t1, ts2?#asahi-dev · 2025-12-17 16:49
jannauno need for traces I have my own. the timestamp in the beginning are CNTVCT_EL0?#asahi-dev · 2025-12-17 16:47
jannauI ordered some cheapish photo diodes with over the top specs. the rest I have on hand#asahi-dev · 2025-12-17 16:44
jannauare these are offsets from the first value?#asahi-dev · 2025-12-16 17:53
jannauheading out now#asahi-dev · 2025-12-16 17:52
jannauhave you tried different usb ports? is it possible to try a different cable or host?#asahi-dev · 2025-12-16 17:45
jannauI assume there is nothing in lsusb or the kernel log?#asahi-dev · 2025-12-16 17:43
jannauviolette: what kind cable? was the cable connected before the mac booted?#asahi-dev · 2025-12-16 14:16
jannaumakes note to update the GPU driver#asahi-dev · 2025-12-16 14:16
jannaunevermind, I wasn't aware that the offset doesn't apply to EL2#asahi-dev · 2025-12-16 14:02
jannauI guess it'll break clearly if dcp has its own offset#asahi-dev · 2025-12-16 13:47
jannauyou might want to experiment with arch_timer_read_cntpct_el0() for the current time. based on the GPU the coprocessors probably share the physical counter. not sure how well received using it in driver code is upstream (my guess is not all all)#asahi-dev · 2025-12-16 12:41
jannauwould be nice if the input latency could be avoided#asahi-dev · 2025-12-16 12:41
jannauideally I want to measure the time from DRM KMS page flip to screen update but I guess the wait time for the flip can be averaged out#asahi-dev · 2025-12-16 12:34
jannauVRR might work already with this once the driver exports the right properties#asahi-dev · 2025-12-16 12:32
jannauhas anyone a good idea to measure latency? but I guess if there's no corruption it means the surfaces aren't held multiple frame updates#asahi-dev · 2025-12-16 12:29
jannauoliver_: great work. when I traced macos the timestamp fields weren't set consistently. iirc they were often unset#asahi-dev · 2025-12-14 16:44
jannauyes, downstream only. I don't think I'll submit rust rtkit this cycle#asahi-dev · 2025-12-14 16:17
jannauAPPLE_TUNABLE creates merge conflicts in drivers/soc/apple/{Kconfig,Makefile} currently avoided by sheer luck. The additions in various did not share context so far#asahi-dev · 2025-12-14 15:21
jannausven: git am complains about the trailing empty line in drivers/phy/apple/Kconfig#asahi-dev · 2025-12-14 11:45
jannaurc releases on sunday morning are nice. I have also a bunch of simple outstanding patches I need to send#asahi-dev · 2025-12-14 10:57
jannauchadmed_: might be worth keeping an eye on https://gitlab.freedesktop.org/pipewire/pipewire/-/issues/5027 to ensure changes don't break our audio setup#asahi-dev · 2025-12-13 08:39
jannauoh, I didn't know you could run a smaller version installassistant as long as the major version matches#asahi-dev · 2025-12-13 05:51
jannauoliver_: for tracing the latest macos 13.x version should be fine. For m2 pro/max you would need dfu restore to install 13.5. For the other devices you need to install macos 12 first and then upgrade to 13.5. otherwise you end up with the latest 13.x version anyway#asahi-dev · 2025-12-12 21:30
jannauthe epic method numbers are wrong for v13.5 firmware. I have fixes locally but still need to verify them#asahi-dev · 2025-12-12 21:28
jannauoliver: https://github.com/AsahiLinux/m1n1/pull/519 for now. You might still have to comment tracing of some or all EPIC endpoints#asahi-dev · 2025-12-12 20:35
jannauhttps://lpc.events/event/19/contributions/2023/ covers the same topic#asahi-dev · 2025-12-12 20:35
jannauhttps://lpc.events/event/19/contributions/2125/ is interesting for the PCIe device power control missing upstream#asahi-dev · 2025-12-12 20:28
jannaudo we have an upstreamable solution for the shared speaker reset gpio? if not https://lpc.events/event/19/contributions/2013/ would be interesting#asahi-dev · 2025-12-12 16:52
jannaupossible an issue from the aop tracer. I might have fixed that only locally. I'll check for unmerged DCP trace change later today#asahi-dev · 2025-12-12 13:27
jannauwhich device are you tracing?#asahi-dev · 2025-12-12 13:26
jannauotherwise I'll take a look what dcp tracing change I have locally#asahi-dev · 2025-12-12 13:26
jannauthat might just required for parse_log though. we should change the default#asahi-dev · 2025-12-12 13:25
jannauoliver: try exporting AGX_FWVER=V13_5#asahi-dev · 2025-12-12 11:07
jannauit has at the bottom a a commented line to write the log for parse_log.py#asahi-dev · 2025-12-12 11:06
jannauoliver: see proxyclient/hv/trace_dcp.py you use that with `./tools/run_guest.py -m hv/trace_dcp.py ...`#asahi-dev · 2025-12-07 13:49
jannauhttps://github.com/AsahiLinux/m1n1/pull/518