#
…

2021-04-03

00:34DanB91 has joined #asahi-gpu
00:46DanB91 has quit [Remote host closed the connection]
01:20Emantor has quit [Quit: ZNC - http://znc.in]
01:23zkrx has quit [Ping timeout: 265 seconds]
01:24Emantor has joined #asahi-gpu
01:48zkrx has joined #asahi-gpu
01:58DanB91 has joined #asahi-gpu
02:08phiologe has quit [Ping timeout: 250 seconds]
02:09phiologe has joined #asahi-gpu
02:43* bloomgave in and started doing development on the mac
03:02bloomAlso, I'm not sure what /System/PrivateFrameworks/IOGPU.framework/Versions/A/Resources/IOGPUCommonSignposts.plist is but it might be interesting.
03:03bloomA quick Google search indicates "signposts" are a macOS tracing thing?
03:03DanB91 has quit [Remote host closed the connection]
03:03DanB91 has joined #asahi-gpu
03:04DanB91 has quit [Remote host closed the connection]
03:05DanB91 has joined #asahi-gpu
03:25DanB91 has quit [Remote host closed the connection]
03:26DanB91 has joined #asahi-gpu
03:27bloomI like Instruments :'o
03:30aquijoule_ has joined #asahi-gpu
03:30bloomInteresting columns '[resource] ID' and '[parent] IDs', explains so much of the weirdness.
03:31DanB91 has quit [Ping timeout: 240 seconds]
03:32richbridger has quit [Ping timeout: 252 seconds]
03:33odmir has quit [Remote host closed the connection]
03:35odmir has joined #asahi-gpu
03:39odmir has quit [Ping timeout: 250 seconds]
04:00bloommatched up an ID in the command stream to what instrmuents calls "Encoder ID"
04:00bloomthis is getting.. more complicated
04:01bloomLooks like a _lot_ of these magic numbers solely exist to support profiling...
04:03bloommetal-io-surface-access shows WindowServer making use of a "83b&" pixel format for the display (4k if it matters) .. guessing that's compressed tiled something or other
04:05odmir has joined #asahi-gpu
04:05bloomCuriously Instruments at the low level involes a ton of S-expressions
04:07bloomAnyway, the big take aways are:
04:08bloom - the word at offset 0x28 of the response to ALLOCATE_MEM is the resource ID
04:09bloom - unk6 grabs a LOT of IDs (as returned), within the massive interval I think userspace allocates the IDs it actually needs for things like command buffers?
04:10bloom- the thing at offset 0xe4 of my compute trace is the Encoder ID... probably just used to match up with Instruments, buy it's much too late here to test that now.
04:18marcanbloom: that firmware should be loaded by iBoot2, before m1n1 runs
04:19bloommarcan: noted.
04:19marcanit should get copied into ./Firmware/Preboot/C1EB0F38-C67D-43AB-BA6F-3CFA88FC9E6F/restore/Firmware/agx/armfw_g13g.im4p for installation/recovery purposes
04:20DanB91 has joined #asahi-gpu
04:20bloomsee 2 lines later :p
04:20marcanand then on boot provisioning that should end up in AGX.img4 next to iBoot2
04:20marcanor something like that
04:21marcanand yes, RTKit is their RTOS
04:23marcanoh actually GFX.img4 I believe
04:23marcanso the trick is to look for pre-loaded: 1 in the ADT
04:23marcan firmware-name: GFX
04:24marcanthat's under iop-gfx-nub
04:24marcanwhich is under sgx (yes, sgx, hi powervr)
04:24marcanall those pre-loaded: 1 nodes are things iBoot2 loads firmware for us for
04:26marcaneither way, we can assume 1TR will be there, deleting 1TR is not a supported use case for us
04:26marcanand 1TR *is* a full macOS install with all necessary firmware (currently, wifi as far as we know)
04:27marcanso I do not expect us to ever have any redistribution issues, unless Apple starts having some hardware that 1TR lacks drivers for but needs firmware
04:27bloommarcan: Hmm, so, the GNU/Linux side doesn't need to know about firmware blobs, so it's like they're not really there.
04:27marcanyup, exactly
04:27bloomWe can apply for RYF for the Mac Mini, then?
04:27marcanI guess!
04:27marcanexcept for the wifi
04:27bloomWell, as long as we pull out the wifi chi
04:27bloom--ninja'd
04:28marcan*registers a company to re-sell Apple M1 laptops with the wifi chip nerfed
04:29marcanwonder if there's a way to destroy it with e.g. firmware misconfiguring power, then we wouldn't even need to take the things apart
04:29bloomoh, that's a good question
04:29bloomdefinitely worth an engineering investigation
04:29bloomfor your freedom
04:29marcanwill prioritize it
04:38odmir has quit [Ping timeout: 258 seconds]
04:39DanB91 has quit [Remote host closed the connection]
04:40DanB91 has joined #asahi-gpu
04:45DanB91 has quit [Ping timeout: 268 seconds]
05:11DanB91 has joined #asahi-gpu
05:16DanB91 has quit [Ping timeout: 265 seconds]
05:29odmir has joined #asahi-gpu
06:02odmir has quit [Ping timeout: 250 seconds]
06:53odmir has joined #asahi-gpu
07:26odmir has quit [Ping timeout: 250 seconds]
08:17odmir has joined #asahi-gpu
08:47DanB91 has joined #asahi-gpu
08:50odmir has quit [Ping timeout: 258 seconds]
08:52DanB91 has quit [Ping timeout: 240 seconds]
09:14odmir has joined #asahi-gpu
09:47odmir has quit [Ping timeout: 245 seconds]
10:15odmir has joined #asahi-gpu
10:47odmir has quit [Ping timeout: 250 seconds]
10:51sven has quit [Quit: ZNC 1.7.4 - https://znc.in]
10:52sven has joined #asahi-gpu
11:57marcan has quit [Quit: Now where's my screwdriver...]
11:58marcan has joined #asahi-gpu
12:47DanB91 has joined #asahi-gpu
13:28DanB91 has quit [Remote host closed the connection]
13:28DanB91 has joined #asahi-gpu
13:39DanB91 has quit [Remote host closed the connection]
13:39DanB91 has joined #asahi-gpu
13:45DanB91 has quit [Ping timeout: 240 seconds]
14:07DanB91 has joined #asahi-gpu
14:11DanB91 has quit [Remote host closed the connection]
14:32DanB91 has joined #asahi-gpu
15:03DanB91 has quit [Remote host closed the connection]
15:03DanB91 has joined #asahi-gpu
15:08DanB91 has quit [Ping timeout: 250 seconds]
16:17balrogbloom: won't the fact that the firmware is upgradeable from flash make it impossible to get it RYF-certified?
16:17balrogor rather, loaded from flash
16:18davidrysk[m](also that's me, just haven't been using matrix lately)
16:26Baughn_ has joined #asahi-gpu
16:27DanB91 has joined #asahi-gpu
16:31Baughn_ has quit [Quit: ZNC 1.6.2+deb1 - http://znc.in]
16:32Baughn has quit []
16:32Baughn has joined #asahi-gpu
16:33DanB91 has quit [Ping timeout: 268 seconds]
16:43linuxgemini has joined #asahi-gpu
16:54linuxgemini has quit [Quit: Ping timeout (120 seconds)]
17:10DanB91 has joined #asahi-gpu
17:39linuxgemini has joined #asahi-gpu
17:41linuxgemini has quit [Client Quit]
17:56linuxgemini has joined #asahi-gpu
18:47DanB91 has quit [Remote host closed the connection]
18:48mxw39 has quit [Quit: Konversation terminated!]
18:48DanB91 has joined #asahi-gpu
18:49mxw39 has joined #asahi-gpu
18:53DanB91 has quit [Ping timeout: 268 seconds]
19:02DanB91 has joined #asahi-gpu
19:32DanB91 has quit [Remote host closed the connection]
19:33DanB91 has joined #asahi-gpu
19:40DanB91 has quit [Ping timeout: 240 seconds]
20:15DanB91 has joined #asahi-gpu
20:20DanB91 has quit [Ping timeout: 268 seconds]
20:23DanB91 has joined #asahi-gpu
20:25DanB91 has quit [Client Quit]
21:04odmir has joined #asahi-gpu
21:21bpye has quit [Ping timeout: 240 seconds]
21:40bpye has joined #asahi-gpu
21:58odmir has quit []
22:01odmir has joined #asahi-gpu
22:15bloomWhy does xcodebuild trigger every time I run git after cold boot...?
23:03bloomok, pushed some stuff sorting out cmdbuf/encoder IDs
23:03bloomthere is so much infrastructure here seemingly only to appease the Apple kernel
23:06bloomuh oh, command bfufers are erroring with code 10.. whatever that means
23:06bloomdespite appearing to work. this is.. bad
23:08bloommeh, been doing that since the first commit.. not a regression anyway
23:22* bloomfalling victim to the streetlight effect..
23:33bloomProbably got everything I'm going to get from Instruments
…