#
…

2026-08-09

00:03hdbngr has quit [Ping timeout: 480 seconds]
00:07chadmed has quit [Quit: Konversation terminated!]
00:09chadmed has joined #asahi-dev
00:10hdbngr has joined #asahi-dev
00:19hdbngr has quit [Ping timeout: 480 seconds]
00:26hdbngr has joined #asahi-dev
00:34hdbngr has quit [Ping timeout: 480 seconds]
00:42hdbngr has joined #asahi-dev
00:44chadmedsofus: do we know how to make AVD export interchange compressed buffers?
00:50hdbngr has quit [Ping timeout: 480 seconds]
00:58hdbngr has joined #asahi-dev
01:01___nick___ has quit []
01:03___nick___ has joined #asahi-dev
01:04___nick___ has quit []
01:06hdbngr has quit [Ping timeout: 480 seconds]
01:06___nick___ has joined #asahi-dev
01:09hdbngr has joined #asahi-dev
01:18hdbngr has quit [Ping timeout: 480 seconds]
01:25hdbngr has joined #asahi-dev
01:34hdbngr has quit [Ping timeout: 480 seconds]
01:41hdbngr has joined #asahi-dev
01:49hdbngr has quit [Ping timeout: 480 seconds]
01:57hdbngr has joined #asahi-dev
02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-dev
02:00ciara has quit [Remote host closed the connection]
02:05hdbngr has quit [Ping timeout: 480 seconds]
02:12ciara has joined #asahi-dev
02:12hdbngr has joined #asahi-dev
02:21hdbngr has quit [Ping timeout: 480 seconds]
02:31hdbngr has joined #asahi-dev
02:39hdbngr has quit [Ping timeout: 480 seconds]
02:47hdbngr has joined #asahi-dev
02:49oolong has joined #asahi-dev
03:05hdbngr has quit [Ping timeout: 480 seconds]
03:13hdbngr has joined #asahi-dev
03:22hdbngr has quit [Ping timeout: 480 seconds]
03:29bgb has quit [Remote host closed the connection]
03:31bgb has joined #asahi-dev
03:31hdbngr has joined #asahi-dev
03:38incog has joined #asahi-dev
03:44hdbngr has quit [Ping timeout: 480 seconds]
03:51pb17 has quit [Ping timeout: 480 seconds]
03:53hdbngr has joined #asahi-dev
03:59nora has joined #asahi-dev
04:01pb17 has joined #asahi-dev
04:01nora_ has quit [Ping timeout: 480 seconds]
04:01hdbngr has quit [Ping timeout: 480 seconds]
04:02incog has quit []
04:10hdbngr has joined #asahi-dev
04:19hdbngr has quit [Ping timeout: 480 seconds]
04:27hdbngr has joined #asahi-dev
04:35hdbngr has quit [Ping timeout: 480 seconds]
04:44hdbngr has joined #asahi-dev
05:01hdbngr has quit [Ping timeout: 480 seconds]
05:08hdbngr has joined #asahi-dev
05:17hdbngr has quit [Ping timeout: 480 seconds]
05:26hdbngr has joined #asahi-dev
05:35hdbngr has quit [Ping timeout: 480 seconds]
05:38LarstiQ(the peanut gallery appreciates the context and progress)
05:43hdbngr has joined #asahi-dev
05:49sofuschadmed: no, do you know if i need to to anything special to make macos use them? (for tracing)
05:50chadmedit should be using them already as the surface they end up scanned out on is an interchange surface
05:51hdbngr has quit [Ping timeout: 480 seconds]
05:51sofusi dont think i ever traced an actual video playing... will do later
05:52chadmedno rush, still need to get a bunch of other dcp-related stuff merged first
05:53chadmedim just thinking _potentially_ some of the artefacts/slowness could be related to copying big linear buffers to the gpu. both the gpu and dcp can read interchange
05:53hdbngr has joined #asahi-dev
05:56chadmedthere may be cases where the gpu cant read interchange (i dont think anyone tested G13 but potentially pre-M-series AGX too) but that doesnt matter for direct scanout since all the dcp variants we support can read them
06:00sofushow would it work from a driver perspective? would it just be adding a new pixel format?
06:02hdbngr has quit [Ping timeout: 480 seconds]
06:03chadmedim not entirely sure. does v4l2 support format modifiers in the same way drm does?
06:03sofusim not sure, but i dont think so
06:07chadmedoh god what a mess, looks like each vendor just creates their own pixfmts lol
06:09hdbngr has joined #asahi-dev
06:16sofusi suppose we could just add our own and add special handling in the vaapi-wrapper
06:18hdbngr has quit [Ping timeout: 480 seconds]
06:23ArcaneNibble has quit [Quit: ZNC 1.10.2 - https://znc.in]
06:24ArcaneNibble has joined #asahi-dev
06:25chadmedyeah but not important until the gpu and dcp are sorted out
06:25chadmedwas just curious if we knew how to make avd do it
06:27hdbngr has joined #asahi-dev
06:35hdbngr has quit [Ping timeout: 480 seconds]
06:44hdbngr has joined #asahi-dev
06:52hdbngr has quit [Ping timeout: 480 seconds]
06:54dcavalca8569797667858944 has quit [Read error: Connection reset by peer]
06:54dcavalca8569797667858944 has joined #asahi-dev
06:54hdbngr has joined #asahi-dev
07:03hdbngr has quit [Ping timeout: 480 seconds]
07:12hdbngr has joined #asahi-dev
07:20hdbngr has quit [Ping timeout: 480 seconds]
07:28hdbngr has joined #asahi-dev
07:37hdbngr has quit [Ping timeout: 480 seconds]
07:44hdbngr has joined #asahi-dev
07:45svenokay, I guess the „wait for the reset thingy to finish“ poll is just broken then
07:45svenbut good to know the rest works then
07:53hdbngr has quit [Ping timeout: 480 seconds]
07:55hdbngr has joined #asahi-dev
08:07k70 has quit [Quit: gone]
08:17svenchaos_princess: maybe fixed, i was confused by readl_poll_timeout again. if this doesn't work i'll just add a udelay to the reset controller
08:19svennow do i want to look into how m2 does this or do i want to write dt-bindings? :(
08:23sofuschadmed: either im dumb or there was no change to the submissions/instructions, maybe there is an enable_interchange register or something
08:27svenanyone got all m2 adt templates for me?
08:27svenah wait.. maybe they're
08:27svenyeah, nvm, i already have them
08:28chaos_princessthat package has m3 too if you are interested
08:28sveni'll do M1/M2 for now since atcphy is the same there
08:29svenand then dt-bindings :(
08:29k70 has joined #asahi-dev
08:47hdbngr has quit [Ping timeout: 480 seconds]
08:48svenhm... M2 has iop,mxwrap-acio now instead of iop,m3wrap-v2-acio
08:48svencould be the same thing, could be something new. guess i'll just stick to M1 for now
08:59hdbngr has joined #asahi-dev
09:07hdbngr has quit [Ping timeout: 480 seconds]
09:38svenhttps://github.com/AsahiLinux/linux/commits/b4/cio-reset/ assuming the poll works now that's the reset controller series
09:38nickchandon't worry about pre M-series AGX too much, it would be... some time at least
09:39nickchanA11 is the first AGX, previous is powerVR I think
09:45nickchanspeaking of nvme probably could simplify m1n1's code by using a queue size of 1
09:45nickchanwould make adding a11 nvme less ugly code wise
09:52picnoir has quit [Ping timeout: 480 seconds]
09:53mmediouni[m]nickchan: wasn't there AGX0? realworldtech claims that the Apple-designed shader core shipped in the A8: https://www.realworldtech.com/apple-custom-gpu/
09:54nickchanthat would match with metal support
09:55nickchanthe thing is I have heard of *both* A8 and A11 and I do not know which is which
10:02picnoir has joined #asahi-dev
10:18svenm2 I can’t guess, I need a trace there
10:19chaos_princessdon't have any of those
10:19picnoir has quit [Ping timeout: 480 seconds]
10:20picnoir has joined #asahi-dev
10:20svenfrom a quick glance I think at least the mailbox has changed
10:21svenno idea if it’s a new variant or just the regular „other“ one used everywhere else
10:37hardy94 has joined #asahi-dev
11:07agmitron[m] has joined #asahi-dev
11:10agmitron[m]Hi. Are PMP usage reports welcome?
11:10agmitron[m]Is there a place I can report my observations about using PMP?
11:10agmitron[m]Device: Macbook Pro 16” M2 Max
12:17Extern has joined #asahi-dev
12:18svenalright, sent out the reset controller series
12:20svenfor m2 I could use a trace of all pages related to the acio nodes
12:20svendon’t Need the atcphy side for now
12:21svenjust raw register access is fine
12:25LarstiQsven: plain m2 or m2 pro also?
12:25svenany m2 is fine
12:25svenonce I understand a single one I can probably guess the rest again
12:27LarstiQprepping for cooking atm but I can maybe give it a try later today
13:19JPeisachsven: i see the SoC CIO patch, is there a specific way I can test it besides running it on a system?
13:19svenyou can only test it together with the rest of the tbt stuff
13:20sventbt-reset-wip branch in our linux fork on github
13:20chaos_princess+ replace the reset_controller_reset with reset_controller_deassert in tb patch
13:21svenoh, have I not pushed that yet? gimme a moment
13:25svenpushed
13:33chaos_princesshave you tried it only with host -> device?
13:34svenI tried host host but tipd decided it would rather negotiate usb3 for some reason
13:34chaos_princessthere is something going wrong with host to host
13:34chaos_princesshere it negotiated usb4
13:36svenyeah, 3. this wasn’t two Mac hosts though but an old x86 one I have
13:36svenwhat exactly goes wrong?
13:37chaos_princessboth devices link up, see each other, negotiate usb4net, and then when packets are sent they disappear into shadow realm
13:38svenugh. tracing that isn’t going to be fun
14:14BlueMatt has quit []
14:15kraem has quit [Remote host closed the connection]
14:16kraem has joined #asahi-dev
14:16JPeisachso before I read your message I applied the patch to fairydust, which I didn’t notice any changes or breakage. then I built and tested on the tbt branch and there is no keyboard/mouse input at all. not sure what could have lead to that
14:17JPeisachoh i2c bus stuck errors
14:17chaos_princessisn't the tbt branch on 7.2 with no asahi patches beyond what is in upstream
14:18chaos_princessso, if you want things to work, rebase it on asahi-wip
14:18svenand you can only test that patch if you actually connect a tbt device
14:19JPeisachwell that sucks because at my high school I just graduated from there was a ton of those I could use. I don’t have any handy and im on vacation anyway so im basically just regression testing ig
14:19sventhere’s essentially no way this can cause a regression
14:19chaos_princessnah, just booting will not run any new code
14:19sventhe reset controller is inert until you have the rest of the tbt patches and connect a tbt device
14:20chaos_princessneed either actual tb devices, or two tb hosts connected with a tb-capable cable
14:20BlueMatt has joined #asahi-dev
14:20JPeisachclosest thing I would have is an official Apple thunderbolt thingy that has a USB-A, HDMI, and something else adapter but that’s at home
14:20JPeisachalright so I’ll just review the patches on the mailing list and call it a day
14:21JPeisachsorry for the noise
14:32nafod has quit [Remote host closed the connection]
14:40BlueMatt_ has joined #asahi-dev
14:45BlueMatt has quit [Ping timeout: 480 seconds]
14:59BlueMatt_ is now known as BlueMatt
15:13nnym has quit [Ping timeout: 480 seconds]
15:22nnym has joined #asahi-dev
15:28hardy94 has quit [Quit: hardy94]
15:31nnym has quit [Ping timeout: 480 seconds]
15:36mx08 has quit [Quit: WeeChat 3.8]
15:37mx08 has joined #asahi-dev
15:39nnym has joined #asahi-dev
15:47nnym has quit [Ping timeout: 480 seconds]
15:56nnym has joined #asahi-dev
16:03liv[m]sven: how would i give you this information? just compile a kernel from a specific branch and plug in tb devices and send dmesg log?
16:03svenwhich information? the m2 trace?
16:04nnym has quit [Ping timeout: 480 seconds]
16:05liv[m]i have a m2 max mbp 14"
16:05svenhave you used the hypervisor before with a secondary machine?
16:05liv[m]so anything you need that i can provide without too much hassle
16:05liv[m]sven: no
16:05svenso that may or may not be 'too much hassle' then
16:07nnym has joined #asahi-dev
16:07svenit's essentially this here, https://asahilinux.org/docs/sw/m1n1-hypervisor/, but there are enough people on here who both have an m2 and have all that setup and there's no need to rush
16:13liv[m]sven: it does seem a little dangerous for me to do on my main machine
16:13svenyeah, no worries
16:13liv[m]however i feel it should be possible so ill try eventually
16:15SandwichBtw has joined #asahi-dev
16:20svenit's mostly useful for developers fwiw
16:33chaos_princessjust don't do it on your main install, and it is pretty safe
16:35liv[m]if i brick a 2nd macos install, can i just remove it from the first one without having to do a dfu??
16:42chaos_princessyes, just delete via disk utility
16:43liv[m]ah alright
16:43liv[m]thanks :)
16:44chaos_princesslike, the "dangerous" parts are the nvme controller (for obvious reasons), and the SEP (it is involved with disk encryption keys)
16:44chaos_princessrest you can poke at and get a machine reset at worst
16:44hdbngr has joined #asahi-dev
16:45sventhe most annoying part probably is that a secondary full macos installation takes quite a bit of space
16:46LarstiQnot really a way to slim that down I suppose?
16:47sveni think you get it down to like 60 (or 80?) gb
16:47svenpossibly even lower but then you actually have to put in effort :D
16:49Rafay has joined #asahi-dev
16:49LarstiQin principle I have that free on the linux side, don't remember the osx side. This is going to take a bit more effor than today but I do want to get it done
16:49liv[m]sven: damn, i thought it was like half of that...
16:49__nick__ has joined #asahi-dev
16:50liv[m]LarstiQ: same boat here :p
16:52RafayI have it set it up as well on mac mini m2, but I can just trace tbt host-to-host. If you would like to trace that, let me know the range you need.
16:52___nick___ has quit [Ping timeout: 480 seconds]
16:53hdbngr has quit [Ping timeout: 480 seconds]
16:55Carlos has quit [Ping timeout: 480 seconds]
16:55Carlos has joined #asahi-dev
16:58chaos_princessinstall it into the same container as your primary one if you are feeling spicy
17:02LarstiQI'd much rather backup everything and reimage and take a bit longer :P
17:12svenRafay: host to host is fine, i just need the initial bringup which is the same. let me find the nodes i'm interested in
17:15svenso i need /arm-io/{acioN,dart-acioN,acio-cpuN} where N is the port you're going to use
17:15svenor just trace all of them if you don't know which port it is
17:16svenit'll likely be a fairly long trace
17:27hdbngr has joined #asahi-dev
17:43hdbngr has quit [Ping timeout: 480 seconds]
17:44hdbngr has joined #asahi-dev
17:52Z750 has quit [Quit: bye]
17:57RafayI have a trace from couple days ago of the required stuff here: https://pastes.io/38kHwhA7 . But it has other stuff like PMGR, atcphy1, and cd3217 tracing mixed in. If there's too much garbage, let me know, I can get a fresh trace later.
18:00svensounds perfect! i'll just grep for the relevant parts :)
18:01svenlol, wtf, pastes.io wants money to allow downloading? :D
18:01svendon't they know copy/paste works just fine
18:02Z750 has joined #asahi-dev
18:04sventhat's from a mac mini, right?
18:05Z750 has quit []
18:07RafayYeah for some reason. I tried pastebin but it has a small size limit. This is was the only thing that popped in the results.
18:07RafayYeah it's from a mac mini.
18:07svenok, let's see
18:09svennah, no worries, i just copy/pasted it :)
18:10svenit was just funny to me that they want to charge money for downloading the thing they already sent to my browser
18:10Z750 has joined #asahi-dev
18:13svennice, mailbox really looks like the normal ASC mailbox instead of that special m3 one
18:13svenwait..
18:14svenor is that a new variant
18:15svennope, i'm just confused. it's the "normal" one
18:17sri has joined #asahi-dev
18:19RafayIn the adt it's written as iop,mxwrap-acio .
18:19svenyeah, that's why i got suspicious
18:19svenbut from looking at the trace it looks like the normal asc mailbox
18:32svenchaos_princess: can you change https://github.com/AsahiLinux/linux/blob/tbt-reset-wip/drivers/thunderbolt/apple.c#L89 to 0x4000 and see if that keeps packets out of the shadow realm? :D
18:33chaos_princessyea, give me a bit
18:44rany has quit [Remote host closed the connection]
18:44hdbngr has quit [Ping timeout: 480 seconds]
18:45svenon top of tbt-reset-wip
18:46rany has joined #asahi-dev
18:47chaos_princessare the "allocating {T,R}X ring -1 of size 256" messages helpful?
18:48chaos_princess(ignore the -1, i am still on hacky fix of hop id allocation)
18:52RafayLemme try. Will take anywhere from a few minutes to a few hours though. First time compiling asahi.
18:53svensure, no worries. there's no rush here :)
18:54hardy94 has joined #asahi-dev
18:54hdbngr has joined #asahi-dev
18:54svenchaos_princess: so i suspect that something is wrong with anything that's not the control ring (ring 0). i just don't know what is wrong
18:54svenfrom what i can tell all the setup happens on ring 0 and that works
18:54svenand the moment it tries to use any other ring something breaks
18:55svenand from Rafay's trace i noticed that the offset is wrong which would've nicely explained all that
19:02svena trace from macos establishing USB4NET and sending some packets would be helpful
19:03hdbngr has quit [Ping timeout: 480 seconds]
19:11lion328 has joined #asahi-dev
19:11svenchaos_princess: can you try ring_iowrite32options(ring, ring->hop ? 0x28 : 2, 4); instead of https://github.com/AsahiLinux/linux/blob/98588460927272c731c6018af25463986fdf9e0e/drivers/thunderbolt/nhi.c#L677
19:12chaos_princessack, give me a bit
19:12lion328 has quit [Remote host closed the connection]
19:12svenif it's not that i may have to write a real tracer for this :/
19:16donaldh has quit [Ping timeout: 480 seconds]
19:18chaos_princessdoesn't look like that's what it is
19:18svendon't see any difference between the rings other than that
19:23hdbngr has joined #asahi-dev
19:26sri has quit [Quit: Leaving]
19:29Extern has quit [Remote host closed the connection]
19:29svenjust to confirm, this was together with that 0x4000 fix, right?
19:29Extern has joined #asahi-dev
19:32hdbngr has quit [Ping timeout: 480 seconds]
19:41hdbngr has joined #asahi-dev
19:42sventhere's CONFIG_USB4_DMA_TEST, still trying to figure out what it does but that could be useful to track down the problem
19:54hardy94 has quit [Quit: hardy94]
19:58lion328 has joined #asahi-dev
19:59hdbngr has quit [Ping timeout: 480 seconds]
20:05__nick__ has quit [Remote host closed the connection]
20:08bgb has quit [Remote host closed the connection]
20:08bgb has joined #asahi-dev
20:10hdbngr has joined #asahi-dev
20:19hdbngr has quit [Ping timeout: 480 seconds]
20:26hdbngr has joined #asahi-dev
20:36hdbngr has quit [Ping timeout: 480 seconds]
20:37hdbngr has joined #asahi-dev
20:47Yushi has quit [Remote host closed the connection]
20:48Yushi has joined #asahi-dev
21:00hdbngr has quit [Ping timeout: 480 seconds]
21:18hdbngr has joined #asahi-dev
21:20bgb has quit [Remote host closed the connection]
21:21bgb has joined #asahi-dev
21:21donaldh has joined #asahi-dev
21:26hdbngr has quit [Ping timeout: 480 seconds]
21:27chadmed has quit [Quit: Konversation terminated!]
21:27chadmed has joined #asahi-dev
21:38hdbngr has joined #asahi-dev
21:47hdbngr has quit [Ping timeout: 480 seconds]
21:59crabbedhaloablut has quit []
22:00crabbedhaloablut has joined #asahi-dev
22:01hdbngr has joined #asahi-dev
22:11hdbngr has quit [Ping timeout: 480 seconds]
22:19hdbngr has joined #asahi-dev
22:22sri has joined #asahi-dev
22:28hdbngr has quit [Ping timeout: 480 seconds]
22:30WindowPain has quit [Remote host closed the connection]
22:32WindowPain has joined #asahi-dev
22:33apc has quit [Quit: The Lounge - https://thelounge.chat]
22:34apc has joined #asahi-dev
23:07hdbngr has joined #asahi-dev
23:13SandwichBtw has quit [Ping timeout: 480 seconds]
23:16hdbngr has quit [Ping timeout: 480 seconds]
23:16catred has joined #asahi-dev
23:48sri has quit [Remote host closed the connection]
23:57hdbngr has joined #asahi-dev
…