#
…

2026-07-03

00:50asahime[m] has joined #asahi-dev
01:03runxiyu_ has quit [Remote host closed the connection]
01:18runxiyu_ has joined #asahi-dev
01:24Calandracas has joined #asahi-dev
01:25runxiyu_ has quit [Remote host closed the connection]
01:26mischa85 has quit [Quit: Lost terminal]
01:31sofus has quit [Ping timeout: 480 seconds]
01:33enverbalalic has quit [Quit: enverbalalic]
01:39sofus has joined #asahi-dev
01:41runxiyu_ has joined #asahi-dev
02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-dev
02:33madmatt0675 has joined #asahi-dev
02:36picnoir has quit [Ping timeout: 480 seconds]
02:37madmatt067 has quit [Ping timeout: 480 seconds]
02:37madmatt0675 is now known as madmatt067
02:44leah has joined #asahi-dev
02:47starcatmeow has joined #asahi-dev
02:51picnoir has joined #asahi-dev
03:46marvin24 has quit [Ping timeout: 480 seconds]
04:00nora has joined #asahi-dev
04:02nora_ has quit [Ping timeout: 480 seconds]
04:43nela has quit [Ping timeout: 480 seconds]
04:46nela has joined #asahi-dev
05:09madmatt0672 has joined #asahi-dev
05:13madmatt067 has quit [Ping timeout: 480 seconds]
05:13madmatt0672 is now known as madmatt067
05:49SternXD has quit [Remote host closed the connection]
05:50sofustime to re a 12252 byte struct :p (vp9 frame symbol counts)
05:55mattia013 has quit [Ping timeout: 480 seconds]
06:16StanfordAppliedCyber[m]honestly for times like these i just wish we had a dedicated tainted dev to jump on the grenade, just look at the firmware/kext, and return us a struct
06:16StanfordAppliedCyber[m]feels like that would save a lot of effort
06:20sofusyeah, although in this case it sounds worse than it is
06:25StanfordAppliedCyber[m]this would actually be a good use of all the llm people who keep coming around lol
06:26StanfordAppliedCyber[m]since it's hard to trust their code and it's usually crap anyways, writing docs would actually be useful as long as they're 90% accurate and not tainted
06:27myriaI mean, I'd be tempted to do something like that. Not really much else help I can be given I don't have much hardware otherwise
06:31StanfordAppliedCyber[m]if you're fine with jumping on a grenade for us there actually is something that i think could be helpful for the gpu side, to be clear though it would immediately taint you and prevent you from ever writing gpu code (at least kernel level) even if you want to
06:32JamesCalligeros[m]Not necessarily. It’s a case by case thing
07:02StanfordAppliedCyber[m]i'm literally writing a bump allocator of noncachable pages right now
07:02StanfordAppliedCyber[m]this is some of the dumbest code of my life
07:04StanfordAppliedCyber[m]literally just set a hardware breakpoint on the kernelcache location that retypes pages to page table, and then we grab a new page that's marked NC from the start and just give xnu that
07:06teaman[m] has joined #asahi-dev
07:16freetolga has joined #asahi-dev
07:26LarstiQStanfordAppliedCyber[m]: ah, so you haven't written an `is_even` function that has cases for the first 100 numbers from 0? :)
08:22bennofs3 has quit [Quit: The Lounge - https://thelounge.chat]
08:23bennofs3 has joined #asahi-dev
09:10StanfordAppliedCyber[m]well i managed to sidestep the custom gxf instruction
09:10StanfordAppliedCyber[m]my horrible bump allocator works
09:10StanfordAppliedCyber[m]got 10 boots in a row without the type 24 fabric access violation returning
09:15StanfordAppliedCyber[m]turns out you cant have cache issues if you simply never cache things :)
09:17StanfordAppliedCyber[m]this wouldn't be possible if we were to virtualize sptm btw
09:17StanfordAppliedCyber[m]virtualized sptm i think this is a fatal error that we can't get around without some intensive surgery on the sptm binary, which also defeats the point of virtualizing it
09:18StanfordAppliedCyber[m]whereas because i'm emulating sptm i can just always return a NC page for all mappings
09:31StanfordAppliedCyber[m]^ it turns out i don't need hardware breakpoints after all
09:32madmatt0676 has joined #asahi-dev
09:33svenuh, no, virtual SPTM without SPRR requires shadow pagetables anyway and we can map things however we want with no patches to the SPTM binary there
09:35svensecond stage attrs also supersede first stage ones iirc so even without shadow pagetables this would be easy
09:35pb17 has quit [Ping timeout: 480 seconds]
09:36madmatt067 has quit [Ping timeout: 480 seconds]
09:36madmatt0676 is now known as madmatt067
09:38pb17 has joined #asahi-dev
09:39svenwhat is that „custom cache instruction“ btw?
09:41StanfordAppliedCyber[m]re: memory mappings, let me think about that, i'm not convinced, though it is currently 2am and my brain isn't functional
09:41StanfordAppliedCyber[m]re: custom cache instructions, sptm has some new instructions present only on m4+, i was discussing them a few days ago
09:42svenyes, which instruction specifically?
09:42StanfordAppliedCyber[m]tl;dr 00201401 appears to be a flush of all pages, 00201320 appears to be a flush by page number, 00201300 is the same thing by address
09:43Sobek[m]Flushing from TLBs, or normal caches ?
09:43StanfordAppliedCyber[m]in sptm it basically keeps track of dirty pages, if below some threshold it calls 00201320 on all of them, otherwise it calls 00201401 to flush all of them
09:44StanfordAppliedCyber[m]no clue, but i don't think it's just tlbs
09:44sven00201401 is just genter or gexit with argument 1 instead of 0
09:44sventhink hvc #1 instead of hvc #0
09:45StanfordAppliedCyber[m]sorry, 0x00201328 was flush by page number
09:45StanfordAppliedCyber[m]yeah i copy pasted the wrong instructions
09:45StanfordAppliedCyber[m]this is why you don't do work at 3am
09:45StanfordAppliedCyber[m]let me find the right ones
09:48StanfordAppliedCyber[m]wait no 00201401 is correct, are you sure that's genter with argument 1? i don't think genter takes arguments at all, regardless, that instruction appears directly in the sptm binary, idk why genter would appear in sptm itself
09:48StanfordAppliedCyber[m]wait ofc it takes args, i'm sleepy
09:49StanfordAppliedCyber[m]but i can verify it appears in sptm itself, in a codepath adjacent to 0x00201328, i'm literally looking at it in ghidra right now
09:50svenmight be gexit though I don’t think that one takes arguments
09:50svenbut it doesn’t make much sense to have 00201400 for gexit and then 00201401 for something entirely else
09:51StanfordAppliedCyber[m]yeah 0x00201420 is genter
09:51StanfordAppliedCyber[m]so this is a different instruction
09:51StanfordAppliedCyber[m]i'm hallucinating but not that bad
09:51StanfordAppliedCyber[m]also it's definitely not just another gexit
09:51StanfordAppliedCyber[m]looking at the code next to it there's no way that's an exit instruction
09:51svenweird
09:57starcatmeow has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
10:17beastd|2 has quit [Quit: KVIrc 5.2.10 Quasar http://www.kvirc.net/]
10:31jamie_ has joined #asahi-dev
10:37madmatt0676 has joined #asahi-dev
10:38jamie__ has quit [Ping timeout: 480 seconds]
10:40madmatt0670 has joined #asahi-dev
10:41madmatt067 has quit [Ping timeout: 480 seconds]
10:41madmatt0670 is now known as madmatt067
10:45madmatt0676 has quit [Ping timeout: 480 seconds]
11:00freetolgaThis is still discussion of "how do we RE m4/a18 pro/m5" right
11:00freetolgaor general RE progress
11:00freetolgaFinding how to write drivers and dts
11:00chaos_princesssptm is mostly relevant on m4, yes
11:01freetolgaIm worried m4 is like maxwell nvidia where its too new and locked down while being too old and locked down
11:02chaos_princessit will remain a problem for everything from m4 onwards
11:02chadmedit is no more "lokced down" than any other soc, the problem is specifically getting macos to run properly outside of the special "macos boot mode"
11:02freetolgaUnless apple introduces some HAL in sptm layer that every os can simply consume and all motivation to RE is lost
11:03chaos_princessthis will not happen
11:03chaos_princesslinux can't run with sptm
11:03chadmedsptm is tightly tightly couple to xnu's idea of memory management, linux will never ever run under it
11:03freetolgaif apple brings official windows would they just start a vm from 1TR with DMA to make it feel real?
11:04freetolgathat was what i fear basically
11:04freetolgaThey make some sandbox that is lightweight
11:04chadmedwho cares what they do for windows
11:04integralpilot has joined #asahi-dev
11:04freetolgaok ok i wont clutter
11:04freetolgagood luck doing sptm
11:04freetolgaIm bad with advanced arm devices
11:05chadmedsptm isnt really anything special. its just a little "extra kernel" that the real kernel has to get authorisation from to do certain things
11:05chadmedit has no bearing on whether or not we can run linux as it is completely disabled in the "not macos" boot mode
11:06chadmedbut this is the problem - m1n1 runs in the "not macos" boot mode, and when we start xnu in the hypervisor it expects sptm to already be running. the problem is not "defeating
11:06freetolgaIs it possible to RE something like pci or gpu without going further in sptim emulation?
11:06chadmed sptm, its getting it running in that boot mode
11:06freetolgaIt sounds like it gets pretty far
11:06chadmedpcie is a bad example because its a complicated block with like 12 grillion tunables that need to be found and set
11:07chadmedbut yes some basic stuff like the interrupt controller, smc, whatever probably already works with just some devicetree changes
11:07freetolgaI think someone got those working yeah
11:07chaos_princessanother point for why sptm is not special - vanilla arm64 defines EL3 and S-EL{0,1,2}, and the memory accesses from those are marked with a secure bit that the hw can use to make decisions. apple arm64 is nothing like that, gl2 (the mode sptm runs in) only changes how page tables are interpreted, and the hardware can't discriminate
11:08freetolgaI kinda just wanna be a user but i also recognize that this is a free of charge projecr
11:09freetolgaIts my fault to get a m4 instead of m2
11:25teaman[m]i'm also an M4 user here, if i can help in some way (whether becoming a tainted or non tainted dev) im all ears
11:33freetolgateaman[m]: the easiest thing rn is to port hte installer to macos 26 as in macos 15 m1n1 doesnt work at all
11:33freetolgayou should isntall macos 26 and make macos 26 vm in utm to test installer
11:33freetolgaotherwise you keep losing space in your macos internal ssd
11:33freetolgafrom the partitions
11:35chaos_princessinstaller is like the last step
11:35freetolgaits about porting to macos 26 vm
11:35freetolganot the m4 machine
11:35freetolgathe recovery preparation failed in macos 26 iirc
11:35svenmaybe don’t suggest what people should do if you’re not a developer in the project yourself
11:38chaos_princessi am completely confused as to why you are trying to "port to macos 26 vm"
11:39freetolgai repeated what i heard before but idc
11:40chaos_princessplease don't play broken telephone, thanks
11:41freetolga has left #asahi-dev [#asahi-dev]
12:18madmatt0674 has joined #asahi-dev
12:23madmatt067 has quit [Ping timeout: 480 seconds]
12:23madmatt0674 is now known as madmatt067
12:34eConnah has joined #asahi-dev
12:43pb17 has quit [Ping timeout: 480 seconds]
12:50eConnah has quit [Quit: eConnah]
12:51pb17 has joined #asahi-dev
12:52eConnah has joined #asahi-dev
13:10eConnah has quit [Quit: eConnah]
13:10eConnah has joined #asahi-dev
13:11runxiyu_ has quit [Remote host closed the connection]
13:14integralpilot has quit [Ping timeout: 480 seconds]
13:17integralpilot has joined #asahi-dev
13:18runxiyu_ has joined #asahi-dev
13:29nela has quit [Quit: bye!]
13:53yukathe stuff people keep generating using llms... It would be so funny if it wasn't so sad
13:56svenwhat happened this time? :/
13:56yuka"DockChannel UART that carries the firmware/XNU debug console out the USB-C SBU pins (debug USB)."
13:57yukafrom some commit message of a commit not immediately marked as AI-assisted/generated
13:57ravikant_ has joined #asahi-dev
13:57Sobek[m]freetolga: Is there any sort of rumors to THAT effect ??
13:58svenno, and let's please keep speculation out of here
13:58chaos_princessyuka: what repo is it under, and does anyone need to be ejected
14:00yukachaos_princess: it's in someone's fork of AsahiLinux/linux. they didn't advertise the repo or commits anywhere in official spaces, I checkout out the repo after they talked about interesting things here in irc.
14:00yuka*checked out
14:01sveni honestly wouldn't mind people using LLMs that much if it weren't almost always those kinds of people who don't want to put it any effort and just play "slow meat-based proxy" between their favorite LLM and us :(
14:01yukaI'm not sure where the line is in terms of bringing the llm generated content into asahi spaces
14:01chaos_princessoh, if it is private repo - then whatever
14:01yukajust it's not marked as such and other people have already started discovering and using it
14:02chaos_princessat least it is not that rando that misconfigured their llm harness and kept opening/closing prs to tiny-dfr
14:02sveni don't really care what people do outside of our repos
14:02yukaso anyother PSA, https://github.com/mischa85/linux / https://github.com/mischa85/m1n1 are slow, please don't copy stuff from them I guess
14:03yuka*are slop
14:03yuka> Sequence derived from analysing the macOS AppleT6040PCIe
14:04chaos_princessoh, thats a full "please go away"
14:04lynndotpy72 has quit [Quit: bye bye]
14:05nela has joined #asahi-dev
14:05lynndotpy72 has joined #asahi-dev
14:07integralpilot> Reverse-engineered from the macOS AppleANS3CGv2Controller driver
14:07integralpilotthis gets worse and worse the more i look i think im just closing that repo now to be safe
14:08svenyeah, it's obvious that we can't take any contributions from there at this point due to multiple reasons
14:08yukaback to a more fun topic
14:10yukaI was annoyed by constantly juggling around my minimal linux device trees for the M4/M5 devices
14:11yukaso I decided to commit them and add some extra info for expectation management
14:11yuka(as in, for each device: this boots / does not boot in single core / smp, ...)
14:12svenyup. i think i may have finally found a fix for #463 and #480 including an explanation that maybe makes sense. i've been hitting that crap again recently but it's been very reliable now with an additional dsb sy and a cache invalidation
14:12chaos_princesswell, we have a lot of iphone device trees which are also in "theoretically boots" stage, so don't worry about expectation management
14:12NickChan[m]that isn't Asahi Linux is it
14:13yukawell it's in upstream
14:13sveni'll happily merge any initial M4/M5/whatever device tree that's on LKML that at least boots a single core and has serial and whatever
14:13yukaThat's the thing
14:13NickChan[m]out of 69 of those dts files i say maybe 30 of them is tested
14:13yukaThe samsung serial is very weird on A18 Pro / M5*
14:14yukait needs at least a new compatible
14:14yukaAnd I don't think we should use it anymore
14:14chaos_princesswell, we sent dockchannel, so that's a serial too
14:14yukaExactly
14:15nickchanyuka: right those files are trickled from upstream
14:16yukathere's no need to hurry any of this, but the path towards it would be dockchannel merged upstream -> dockchannel-uart proposed -> initial A18 Pro / M5* dts with dockchannel-uart
14:17nickchananyways they boot in practice just with very little IO
14:17nickchanactually, I retract that statement because I do not test the upstream a lot
14:18nickchanit's basically tested once when I want to send something and then not tested (testing stays downstream)
14:20yukawhat do you run on those devices anyways? pmOS?
14:20nickchanyuka: 1. pmOS 2. a specific crafted ramdisk to test a bunch of stuff
14:21sveni'll also take a device tree without serial if that makes life easier tbh
14:21enverbalalic has joined #asahi-dev
14:22nickchana device tree where the only form of I/O is watchdog
14:22yukawell you can always add an earlycon=
14:22yukaeven it's not in the device tree
14:22nickchan...this is actually the state of some of the T2 device trees if one doesn't want to solder a test point
14:23nickchanwhich existed solely to test the pesky spi with the spi flash
14:23integralpilotaren't T2 macs x86/acpi? or are t2 dts for running linux on the t2 itself (is that even possible?)
14:23enverbalalicon t6041 at least i have not been able to get any trace of serial through vdmtool no matter what i try, it really looks like dockchannel is the only way to get logs out of it. i was meaning to write a dc-uart earlycon driver last night but got sad when i saw the repos that yuka posted above
14:24rodmac has joined #asahi-dev
14:24nickchandoes every m1+ dockchannel has an interrupt property?
14:27chaos_princessintegralpilot: you can run linux on t2 itself yes
14:27yukaI might send some very minimal M4* and A18 Pro device trees soon then. I would wait longer for M5 since there are still devices being released with that and I'm getting headaches from the M5 Pro / M5 Max. Iirc the M5 Max I had on hands a few months ago advertised as t6050, same as my M5 Pro.
14:28chaos_princessand they are acpi-but-weird
14:28integralpilotoh, interesting
14:30yukaenverbalalic yes, you were using an M2 Macbook on the other end I believe?
14:30yukanot all debuguarts work with all debuguarts
14:31yukaM5 Pro <-> M1 Pro also wasn't able to get any trace of serial
14:31yukaflokli had the same
14:31yukawith M2 <-> M4 Pro
14:31yukaand these quirks are the reason I believe we should no longer consider debuguart the main serial console for those devices
14:32yukathey are semi broken, don't work with high baudrates, or even at all depending on which other macbook or serial chip you have on the other devices
14:32yuka* on the other end
14:33enverbalalicyeah, an m2 on the other end
14:33enverbalalictime to write that dc earlycon then
14:38integralpilot has quit [Quit: integralpilot]
14:41svenmight be a good idea to look at those machines with an oscilloscope
14:46yukaWas there any machine with different variants (cpu core count) but using the same SoC designation?
14:46yukaAs counter-example, M3 Max (14-core) is t6034 while M3 Max (16-core) is t6031
14:47chaos_princessm1 and m2 pros
14:47chaos_princessboth 10 and 8 are t6000
14:47yukaand were they both j314s too?
14:47chaos_princessin fact, t6034 is the weird one
14:48enverbalalici think m4 max 14 and 16 are both t6041 as well
14:48yukaso they use the same device tree, and then m1n1 detects the disabled cores?
14:48yukagood to know
14:49yukabut on M1/M2/M4, Pro and Max were still different SoC ids
14:49yukat6000 vs t6001, t6020 vs t6021, t6040 vs t6041
14:49mischa85 has joined #asahi-dev
14:49yukaM5 Pro and M5 Max are both t6050
14:50yukabut have different midr part numbers
14:56runxiyu_ is now known as runxiyu
14:56runxiyu is now known as runxiyu_
15:08mischa85nice, some slander. let me be clear, i did and will not try to bring any of this upstream. this is my own personal repo, and i even tagged one commit as "RE-derived, DO NOT UPSTREAM". this is strictly RE territory. i'll tag the other commits as well with capitals, but at this point i'm also okay with putting the repo on private to prevent anyone from seeing whatever happens on there, and eventually
15:08mischa85produce some clean-room documentation. did i already say i went in with a hazmat suit and am tainted? i have strictly talked about it within #asahi-re context. tldr; it's llm slop, not ready for upstream in any way, shape or form.
15:13chaos_princessas you have said, it is llm-generated, which we do not accept
15:15jannauI suspect slight clock de-sync issues between device generations. m3 ultra's uart still syncs roughtly till 375000 bps but breaks above. I plan to look at it with glasgow's auto bps uart widget at it
15:16mischa85chaos_princess: i wasn't trying to get any code accepted, that's the whole point
15:16chaos_princesswell, it won't be, so whats the problem
15:16svenjannau: yeah, something like that is my suspicion as well. i have an oscilloscope and maybe it's about time i get some m4 or m5 machine as well :-)
15:17jannauyuka: 14 core m3 max is special beacuse it has only 3/4 of the memeory controllers compared to the full m3 max
15:17mischa85chaos_princess: i have that exact same question
15:18jannauif I only had a device with working thunderbolt. *scnr*
15:19sveni'm really at a loss what's going on there
15:19sventhe mmio traces look similar enough and yet it just refuses to work in linux
15:19svenprobably some timing bullshit
15:20enverbalalic has quit [Ping timeout: 480 seconds]
15:20chaos_princessthere was a working version, right?
15:20chaos_princesshow annoying would a bisect be?
15:21jannauack. I have patches which avoid some of the differences, without changes
15:21svenchaos_princess: very because of the whole dwc3 glue driver and atcphy changes
15:21sventhe bisect result will probably be "old atcphy+dwc3 works, new one doesn"t"
15:22enverbalalic has joined #asahi-dev
15:22jannaubisect between two independent branches doesn't work very well
15:23jannauwhat apperently works is replacing atc with the 6.17 downstream atc driver
15:23jannauworks in the sense that usb3 still works
15:24svenhuh, interesting :D
15:37mattia013 has joined #asahi-dev
16:06ravikant_ has quit [Ping timeout: 480 seconds]
16:10svenhttps://github.com/AsahiLinux/m1n1/pull/613 not sure if that makes sense but this reliably fixed those weird secondary core bringup crashes for me
16:11svennow let's see how i broke this psci-via-efi thing while cleaning it up :/
16:15svenchaos_princess: now that i looked at it again, is there a reason that pmgr-misc thing can only be built-in and not a module?
16:16amarioguy has joined #asahi-dev
16:16chaos_princessi don't see one actually, let me try building it as module and see what breaks
16:19svenack, also don't see a reason why this has to be built-in
16:44enick_590 has quit [Quit: Bridge terminating on SIGTERM]
16:44rhysmdnz has quit [Quit: Bridge terminating on SIGTERM]
16:49rhysmdnz has joined #asahi-dev
16:49Guest12554 has joined #asahi-dev
16:57hamr84 has quit [Quit: The Lounge - https://thelounge.chat]
17:24croissant` has joined #asahi-dev
17:31croissant has quit [Ping timeout: 480 seconds]
17:53runxiyu_ has quit [Remote host closed the connection]
17:53runxiyu_ has joined #asahi-dev
17:54enverbalalic has quit [Ping timeout: 480 seconds]
18:06rodmac has quit [Quit: leaving]
18:18ddxtanx has quit [Remote host closed the connection]
18:28mischa85this expression looks like it's being evaluated in 32-bits. so trouble when a soc tries to boost higher than ~4.2 GHz
18:28yukaI'm quite sure I had pmgr on t8140 working before
18:28yukabut now I crash at arch_counter_get_cntpct+0xc/0x28
18:33yukaI think this might be caused by iBoot updates/changes, but not 100% sure
18:34yukaarch_counter_get_cntpct seems to be quite simple
18:38svenmischa85: nothing before m4 can even boost that high iirc and that highest p-state on m4 might even need cluster sleep support
18:39mischa85sven: yes i realise. but it's something that's going to bite someone sooner or later
18:40svenright, but whoever implements that is obviously going to notice at that point *shrug*
18:40chaos_princessshould be fixed tbh, just not something that matters for current socs
18:41svensure, just don't see it as a priority and it might be a bit more than a simple fix though since that opp api seems to take a pointer to a long
18:41chaos_princessno? the fix is easy, just slap UL on all constants
18:42chaos_princessthe problem is that frequency is an int
18:42svenoh, du
18:47yukathe pc is at "mrs x0, cntpct_el0" but m1n1 can read CNTPCT_EL0 just fine
18:49yukaI commented out arch_counter_enforce_ordering and disabled the "Enhanced Counter Virtualization" feature, to simplify the code even further
19:18ConnorAlecks[m] has joined #asahi-dev
19:27yukaI will have to confirm this on 26.5.1 on monday when I'm back home and have physical access to the machine
19:27yukaso for now no pmgr
19:28croissant` has quit []
19:28croissant has joined #asahi-dev
19:51yukaOh wow
19:51yukanow I got "Unable to handle kernel write to read-only memory at virtual address" for the first time on the Neo :(
20:17qyliss has quit [Quit: bye]
20:19qyliss has joined #asahi-dev
20:21enverbalalic has joined #asahi-dev
20:36enverbalalic has quit [Ping timeout: 480 seconds]
20:39StanfordAppliedCyber[m]<teaman[m]> "i'm also an M4 user here, if i..." <- send me a dm, i have a large static re task if you're fine with tainting yourself
20:43starcatmeow has joined #asahi-dev
20:43enverbalalic has joined #asahi-dev
20:43StanfordAppliedCyber[m]sven: agreed, people using llms poorly is very annoying
20:51enverbalalic has quit [Ping timeout: 480 seconds]
21:14enverbalalic has joined #asahi-dev
21:18cy8aer has quit [Remote host closed the connection]
21:19leah is now known as Guest12563
21:19leah has joined #asahi-dev
21:20Guest12563 has quit [Ping timeout: 480 seconds]
21:22enverbalalic has quit [Ping timeout: 480 seconds]
21:23enverbalalic has joined #asahi-dev
21:25cy8aer has joined #asahi-dev
21:27yukaWhatever jannau did for the i2c version of "tps6598x: Add compile time config to enable debugusb", it also fixes it for spmi
21:29mattia013 has quit [Quit: mattia013]
21:40yukaStanfordAppliedCyber: this might be interesting for you too, the Neo can now self-enter debugusb and as such you don't need a second mac
21:41StanfordAppliedCyber[m]my debugusb doesn't appear to work at all unfortunately, but once i get to the desktop priority #1 will be getting that to work
21:41StanfordAppliedCyber[m]so i can upstream the HV properly
21:45mischa85 has quit [Quit: Lost terminal]
21:48mischa85 has joined #asahi-dev
22:00enverbalalicthere is no way to make a regular PC be able to use debugusb, right ? Requires special stuff from mac usb controllers ?
22:00svendebugusb works with a normal pc as a host
22:01svenbut you also really only need it for early bringup
22:01svenonce the hv works you just get a regular usb serial (cdc acm) device for the guest
22:03jannauand on the neo until we find out how to bypass the usb hub for usb device
22:05jannauyuka: I'm not sure I follow. how did my i2c cd321x change fixed anything for spmi usb-c controllers?
22:15r0ni has quit [Ping timeout: 480 seconds]
22:22yukajannau: when I last tried this, the entire code for enter_kis didn't exist, and I came up with my own
22:23yukajannau: my own code had weird failure modes, LOCK didn't error, but also didn't work properly, possibly what you are now solving with the "... and soft reset if it fails"
22:26yukajannau: and now, after the rebase, I have your enter_kis code and can confirm it works over the existing spmi transport code of the PR#594
22:30StanfordAppliedCyber[m]sven: proudly building my hypervisor ovre a 115200 uart :)
22:33enverbalalicwait if i'm understanding this correctly - with what jannau did for m1n1 to self-enter debugusb and a linux driver to dump earlycon to it, i can get debugusb output on a normal PC ? is there any existing docs for that ? no tuxvdmtool
22:39yukaenverbalalic: the part which isn't fully documented yet: you need to build m1n1 with USE_DEBUG_USB (comment in the line in config.h) and install this as custom boot object (like described in the m1n1 user guide)
22:40weirdtreething has quit [Read error: Connection reset by peer]
22:45hardy94 has quit [Remote host closed the connection]
22:46hardy94 has joined #asahi-dev
22:47pb17 has quit [Ping timeout: 480 seconds]
22:52hardy94 has quit [Remote host closed the connection]
22:56pb17 has joined #asahi-dev
23:16SternXD has joined #asahi-dev
23:39enverbalalic has quit [Quit: enverbalalic]
23:45chadmed has quit [Quit: Konversation terminated!]
…