#
✕ nick:Jamie[m]
#asahi-re · 2022-10-04 13:04
Jamie[m]yeah exactly I was just finding details of that
#asahi-re · 2022-10-04 13:02
Jamie[m]the macos driver *does* do complex parsing of video formats to produce this spaghetti
#asahi-re · 2022-10-04 12:59
Jamie[m]chadmed, rcombs: r.e. ping a few days ago in #-dev, I didn't get a frame just by poking from M1 exactly: I was replaying all the massive complex buffers from a trace of the mac driver
#asahi-re · 2022-10-04 12:54
Jamie[m]compared to the spaghetti of hardcoded offsets
#asahi-re · 2022-10-04 12:53
Jamie[m]diy firmware and/or direct register access both jumped out as potentially sane options to me too
#asahi-re · 2022-10-04 12:52
Jamie[m]I believe in the H264 case it contains a couple of arrays with one element for each slice used in the frame
#asahi-re · 2022-10-04 12:52
Jamie[m]I remember the 0xB7504 thing being a rounder number though, unless you're talking about size rather than alignment
#asahi-re · 2022-10-04 12:51
Jamie[m]and yeah input frame data is in stream 0, as is decoded framebuffer oufput
#asahi-re · 2022-10-04 12:50
Jamie[m]I agree with all that assessment of avd rqou_
#asahi-re · 2022-09-08 01:00
Jamie[m]ooo that samsung link sounds interesting
#asahi-re · 2022-09-07 17:10
Jamie[m]rqou looked at it more recently that me, not sure if they haven any notes
#asahi-re · 2022-09-07 17:10
Jamie[m]shared my notes with your github
#asahi-re · 2022-09-07 17:07
Jamie[m]chadmed: yep, where are you up to?
#asahi-re · 2022-07-21 13:19
Jamie[m]pretty odd
#asahi-re · 2022-07-21 13:19
Jamie[m]I did notice the extra mailboxes in the patterns of registers that exist
#asahi-re · 2022-07-21 13:18
Jamie[m]makes sense
#asahi-re · 2022-07-20 15:58
Jamie[m](it was a fork of your m1n1 which added t8110 dart support for prores)
#asahi-re · 2022-07-20 15:58
Jamie[m]oh my code is on the prores branch in my m1n1 repo btw
#asahi-re · 2022-07-20 15:55
Jamie[m]once I understood those, I was gonna go back a level and mess around with offsets from the 0x60-byte mailbox messages into the cool zone descriptors, to figure out the relationships between those pointers as well
#asahi-re · 2022-07-20 15:52
Jamie[m](which is where the actual video data is read and written)
#asahi-re · 2022-07-20 15:51
Jamie[m]to figure out the pointers to the different blocks of data in dart stream 0
#asahi-re · 2022-07-20 15:51
Jamie[m]dart stream 1, which has big 0xbc000-size descriptors for each frame
#asahi-re · 2022-07-20 15:50
Jamie[m]my plan was then to start messing around with offsets in what i think my notes refer to as "the cool zone"
#asahi-re · 2022-07-20 15:48
Jamie[m]and iirc confirmed that it was successfully producing a picture in memory
#asahi-re · 2022-07-20 15:47
Jamie[m]then in proxyclient/experiments/avd.py I'm replaying register interactions and overwriting the memory contents with those dumps
#asahi-re · 2022-07-20 15:46
Jamie[m]in trace_avd_new.py i'm dumping out the entire DART-mapped ranges before and after each frame
#asahi-re · 2022-07-20 15:45
Jamie[m]and now i know my backups work :)
#asahi-re · 2022-07-20 15:44
Jamie[m]alright m1n1 shared as well
#asahi-re · 2022-07-20 15:37
Jamie[m]rqou_: if you were looking at the kext to grab the firmware, I assume you also noticed the tunables structures which include register names?
#asahi-re · 2022-07-20 15:31
Jamie[m]i gotta untrain myself from using hard-reset when i wanna change what commit a branch is pointing to lol
#asahi-re · 2022-07-20 15:30
Jamie[m]s/in gc/in fsck/
#asahi-re · 2022-07-20 15:30
Jamie[m]but should be easier to recover the whole directory from my backups
#asahi-re · 2022-07-20 15:30
Jamie[m]so they exist as dangling files i can access in gc
#asahi-re · 2022-07-20 15:30
Jamie[m]nah staged
#asahi-re · 2022-07-20 15:27
Jamie[m]sigh
#asahi-re · 2022-07-20 15:27
Jamie[m]also i just git hard reset over my m1n1 experiments
#asahi-re · 2022-07-20 15:26
Jamie[m]they start out comprehensible and devolve into random ravings and big lists of numbers
#asahi-re · 2022-07-20 15:26
Jamie[m]github's renderer doesn't include single newlines so you might wanna view it as plain text
#asahi-re · 2022-07-20 15:25
Jamie[m]rqou_: shared my notes with you on github
#asahi-re · 2022-07-20 15:10
Jamie[m]it's not even aligned to anything more than word size, just a random blob in the middle of its memory
#asahi-re · 2022-07-20 15:10
Jamie[m]it's a pretty messy interface, macos hardcodes the offset of the memory region in the m3 which it knows is free for writing those mailbox messages to
#asahi-re · 2022-07-20 15:08
Jamie[m]and yeah i agree with that assessment of the mailbox
#asahi-re · 2022-07-20 15:08
Jamie[m]ah crap giving you the notes totally slipped my mind rqou_
#asahi-re · 2022-07-17 10:40
Jamie[m]or maybe you will have reproduced all my work by that point :)
#asahi-re · 2022-07-17 10:38
Jamie[m]sorry didn’t get chance to assemble my notes, gotta sleep now, will find them tomorrow
#asahi-re · 2022-07-17 10:37
Jamie[m]oh yeah idk what’s up with the first half of it
#asahi-re · 2022-07-17 00:00
Jamie[m](although AVE might be completely un-looked-at too?)
#asahi-re · 2022-07-16 23:58
Jamie[m]and i now seem to have contracted long covid, so i’m not gonna have the energy for it any time soon either
#asahi-re · 2022-07-16 23:57
Jamie[m]I haven’t had time outside work for it for 6 months
#asahi-re · 2022-07-16 23:57
Jamie[m]rqou_: nobody’s actively looking at AVD

plus ›