#asahi-dev · 2022-06-04 16:03
jannauI was previously on "display: Report time spent modesetting". I'll check if it's a "regression"#asahi-dev · 2022-06-04 15:59
jannauI think it was before I rebased onto the gdbserver merge#asahi-dev · 2022-06-04 14:12
jannaualmost slow enough to see single character updates in the earlycon up to the linux mem init at least#asahi-dev · 2022-06-04 14:11
jannauwas there a change in the hypervisor that could make linux boot or the vuart very slow?#asahi-dev · 2022-06-04 11:12
jannauthe same might happen for the second smp_call on the same core. it will wait until the first call has finished#asahi-dev · 2022-06-04 11:10
jannausmp_call_sync waits for the result and will timeout if the bench code takes more than 3 seconds#asahi-dev · 2022-06-04 10:45
jannauif not why wouldn't it racy#asahi-dev · 2022-06-04 10:45
jannaudo you have a smp_wait or any other kind of synchronization with the code you're running#asahi-dev · 2022-06-04 08:18
jannausmp_call waits just before the the secondary core enters the called function#asahi-dev · 2022-06-04 08:16
jannaubut the second smp_call can cause a proxy timeout if the first is still running (on the same core)#asahi-dev · 2022-06-04 08:14
jannaushould just clobber the previous return value#asahi-dev · 2022-06-04 08:12
jannaunot sure what smp_call does if the previous call is still running on the same core#asahi-dev · 2022-06-04 08:10
jannauyou could run it on a secondary core with smp_call()#asahi-dev · 2022-06-04 08:08
jannauchadmed: proxy timeout? how long is the estimated run time?#asahi-dev · 2022-05-31 18:31
jannauI still have the failure case that the dcp modeset suceeds because it comes after the unplug. The monitor doesn't come up though#asahi-dev · 2022-05-31 18:28
jannaugood night#asahi-dev · 2022-05-31 18:23
jannauyes, stage 1 doesn't do anything. Done explictly to make experimenting with it easier#asahi-dev · 2022-05-31 18:22
jannauthe monitor cycles with 500 - 700 ms for HPD de-assert/assert after modeset i it was in power save mode#asahi-dev · 2022-05-31 18:15
jannaumarcan: https://paste.debian.net/1242599/#asahi-dev · 2022-05-31 18:14
jannaumarcan: sometimes silently, sometimes with "EPIC: IOP returned 0xe00002bc" in the mode set#asahi-dev · 2022-05-31 17:52
jannaumacos still panics with an reallocated frame buffer on the studio#asahi-dev · 2022-05-31 17:49
jannaudelay between first stage modeset and 2nd stage modeset for custom resolution seems to be long enough for my setup though#asahi-dev · 2022-05-31 17:45
jannauwe might have another timing problem. the second modeset fails if it is too close to first one leaving a broken display#asahi-dev · 2022-05-31 17:36
jannauworks with my picky display#asahi-dev · 2022-05-31 17:23
jannauhigher resolution displays than that are hopefully connected via displayport anyway#asahi-dev · 2022-05-31 17:20
jannau1 1/2 tables should be enough for 5120x2160#asahi-dev · 2022-05-31 17:16
jannauthanks, I guess I never tested the l2-tt stuff without fb reallocation and macos. that looked so obvious that macos should be able to handle that that I never considered that would be the problem#asahi-dev · 2022-05-31 06:38
jannaubut if the display does disconnect without user interaction after modeset it's probably within a couple of 100 ms#asahi-dev · 2022-05-31 06:36
jannauyes, but I'm not sure how much that helps if the monitor indeed disconnects within 500 ms after mode set. we have to test what happens#asahi-dev · 2022-05-30 21:21
jannauVinDuv: you mean like https://github.com/jannau/m1n1/commit/dafd2da58908f38c72cfb5f40e0b4e18c8034d32#asahi-dev · 2022-05-30 20:02
jannauI assume dart-dcp remains locked and keeps its config#asahi-dev · 2022-05-30 20:01
jannauso on disp0 side the same behavior as when dcp crashes with a 1920x1200 fb on the mac studio#asahi-dev · 2022-05-30 20:00
jannaugood night#asahi-dev · 2022-05-30 16:45
jannauand of course I can't reproduce it anymore. there are 9 syslog messages at shutdown. one of them looks like it reported after the shutdown. #asahi-dev · 2022-05-30 16:20
jannauyes, I think there are more than 8 messages. dcp produces several syslog messages over an HDMI poweroff/poweron cycle#asahi-dev · 2022-05-29 22:13
jannaujeffmiw: CONFIG_GPIO_MACSMC simpledrm will not load if the backlight driver is not present and for some reason it's not deferred#asahi-dev · 2022-05-29 22:05
jannaujeffmiw: is smc-gpio built-in in your config? it has if simpledrm is built-in#asahi-dev · 2022-05-29 19:57
jannauah, the man page sounds so universal#asahi-dev · 2022-05-29 16:28
jannauadding devicetree support or PCI IDs?#asahi-dev · 2022-05-29 16:27
jannauis there anything to do to get https://man.openbsd.org/aq.4 working?#asahi-dev · 2022-05-29 16:26
jannausure, will do. hopefully soon#asahi-dev · 2022-05-29 10:14
jannaukettenis: openbsd71 boots on the m1 ultra (with the fixed devicetree) nvme and front usb ports (2nd die) work#asahi-dev · 2022-05-28 19:39
jannautriggers unfortunately SErrors with with dart/dwc3 as modules #asahi-dev · 2022-05-28 19:38
jannaukettenis: the m1 ultra device tree you merged for u-boot needs this: https://github.com/jannau/linux/commit/8c2b2b7d7281c09f7fb721fa0eeb6dab5116bd07#asahi-dev · 2022-05-28 17:28
jannauoh joy, resume for the locked dart#asahi-dev · 2022-05-26 20:15
jannauso it's indeed the same issue I see with the linux dcp driver. the display disconnects after the modeset#asahi-dev · 2022-05-26 19:49
jannauthe trigger is that there is an un-acked syslog message when the display did not came up. it's not there if the display is initialized#asahi-dev · 2022-05-26 19:36
jannauI wasn't able to test that since the missing syslog ack blocked dcp#asahi-dev · 2022-05-26 19:34
jannauyes, it's annoying. it might already work with chainloading although it's probably too fast. so #183 would allow me to add a delay before the second modeset#asahi-dev · 2022-05-26 18:50
jannaunot sure if it blocks only (that long) because the display is uninitialized or if there are otherwise no syslog messages