#
✕ nick:sven
#asahi-dev · 2022-10-01 14:40
svencan you read 0x804 as well?
#asahi-dev · 2022-10-01 14:40
svenhuh
#asahi-dev · 2022-10-01 14:35
svenessentially unrouted dispext inputs are stuck at 1 there
#asahi-dev · 2022-10-01 14:34
svenok, would be interesting to see the initial value of that crossbar register i mentioned above
#asahi-dev · 2022-10-01 14:30
svenon a single die? huh
#asahi-dev · 2022-10-01 14:28
sveni just had an off-by-one error
#asahi-dev · 2022-10-01 14:28
svenah, yes, 4 makes sense.
#asahi-dev · 2022-10-01 14:26
svenand a ADT from a multi-die soc would also be very useful
#asahi-dev · 2022-10-01 14:26
svenjannau: can you run [hex(read32(u.adt["/arm-io/atc%d-dpxbar" % i].get_reg(0)[0] + 0x800)) for i in range(N)] on a multi die SoC and replace N with the number of atcphys there?
#asahi-dev · 2022-10-01 14:24
sven3 for each die?
#asahi-dev · 2022-10-01 14:23
svenhow many dispext/dcpext does the largest SoC have right now?
#asahi-dev · 2022-10-01 10:33
svenkonradybcio: t8103 boots with uncore_regs and fast_ipi set to false
#asahi-dev · 2022-10-01 10:29
svenlooks like drivers/mux is a very thin abstraction that would give us devicetree bindings for the display crossbar
#asahi-dev · 2022-10-01 10:14
sventhat probably confuses xorg
#asahi-dev · 2022-10-01 10:13
sveniirc the current driver always marks the screen as connected
#asahi-dev · 2022-10-01 09:12
svenI never tried that
#asahi-dev · 2022-10-01 09:12
svendid you trace macOS or m1n1 there?
#asahi-dev · 2022-10-01 09:07
sven*i didn‘t
#asahi-dev · 2022-10-01 09:07
svenit just happens so rarely that insider care to investigate
#asahi-dev · 2022-10-01 09:07
svenbut there is something that’s broken because I sometimes get that unmapped dart page thing
#asahi-dev · 2022-10-01 09:07
svendont think that ever happened here
#asahi-dev · 2022-10-01 09:06
svenhuh
#asahi-dev · 2022-10-01 08:54
svenhrm, I think I occasionally saw that invalid magic as well but just ignored it
#asahi-dev · 2022-10-01 08:45
svenSame here, so what exactly fails?
#asahi-dev · 2022-10-01 08:43
svenJust on t8103 though
#asahi-dev · 2022-10-01 08:43
svenuh, i think so
#asahi-dev · 2022-10-01 08:23
svenkonradybcio: I’ll test it later this weekend
#asahi-dev · 2022-10-01 08:22
svenmarcan: yup, I thought about that but don’t see how that’s possible with this architecture yet. Pretty sure the stream needs to be connected to DCP before we can tell anything
#asahi-dev · 2022-10-01 08:21
svenjannau: feel free to start the move, I probably won’t have much time before Monday
#asahi-dev · 2022-09-30 17:16
svenunfortunately this m1 only has two usb-c ports which makes it impossible to test this from python :(
#asahi-dev · 2022-09-30 17:16
svenoh well. guess i'll have to add support to the linux driver and see what happens if I do that with two sreens
#asahi-dev · 2022-09-30 17:06
svenit does start a DCPAVSplitDisplayVideoInterface though
#asahi-dev · 2022-09-30 17:04
svenso maybe it's literally just bring up both dptx and hope for the best
#asahi-dev · 2022-09-30 17:04
sveniomfb looks pretty normal
#asahi-dev · 2022-09-30 16:55
sveni guess CORE is actually ATCPHY_INPUT or something like that
#asahi-dev · 2022-09-30 16:54
svenso that's pretty much what I expected
#asahi-dev · 2022-09-30 16:54
sven[dptxport] [dcpdptx-port-epic1] > connectTo(target=0x8012 (CORE=0x2, ATC=0x1, DIE=0x0, CONNECTED=1), unk1=0x101)
#asahi-dev · 2022-09-30 16:54
sven[dptxport] [dcpdptx-port-epic0] > connectTo(target=0x8011 (CORE=0x1, ATC=0x1, DIE=0x0, CONNECTED=1), unk1=0x101)
#asahi-dev · 2022-09-30 16:53
svenweird. i should look into why that happens
#asahi-dev · 2022-09-30 16:47
svenyou can also try to just type "cont" and hope that nothing important is missing in the trace then
#asahi-dev · 2022-09-30 16:47
svenhrm, that's quite frequent. it very rarely happens here
#asahi-dev · 2022-09-30 16:47
svenoh
#asahi-dev · 2022-09-30 16:47
svenclosing the lid should be fine
#asahi-dev · 2022-09-30 16:47
svenfailed?
#asahi-dev · 2022-09-30 16:43
svensomething appears to be safe
#asahi-dev · 2022-09-30 16:29
svenoh, and you can stop the trace the moment something appears on your screen
#asahi-dev · 2022-09-30 16:27
svenjust reboot in those cases and try again
#asahi-dev · 2022-09-30 16:27
svennote that there sometimes seems to be a bug where it complains about not being able to read from some DART page and then falls back into a python shell
#asahi-dev · 2022-09-30 16:26
sventhis will also generate a *lot* of spam
#asahi-dev · 2022-09-30 16:26
svendottedmag: https://github.com/svenpeter42/m1n1/blob/dcp-tracer-2/proxyclient/hv/trace_dcp.py if you have some time. preferably with that display connected at boot time. no need to disconnect it or anything afterwards

plus ›