#
…

2026-06-04

00:03liangqi has quit [Ping timeout: 480 seconds]
00:04voidah has quit [Ping timeout: 480 seconds]
00:04voidah has joined #asahi-dev
00:13voidah has quit [Ping timeout: 480 seconds]
00:29liangqi has joined #asahi-dev
00:34nicenickname6161 has joined #asahi-dev
00:37liangqi has quit [Ping timeout: 480 seconds]
00:42hardy94 has joined #asahi-dev
00:54Calandracas_ has joined #asahi-dev
00:56liangqi has joined #asahi-dev
01:00Calandracas__ has quit [Ping timeout: 480 seconds]
01:04liangqi has quit [Ping timeout: 480 seconds]
01:05voidah has joined #asahi-dev
01:08hardy94 has quit [Remote host closed the connection]
01:13liangqi has joined #asahi-dev
01:15voidah has quit [Ping timeout: 480 seconds]
01:21Lyude has quit [Remote host closed the connection]
01:21liangqi has quit [Ping timeout: 480 seconds]
01:22Lyude has joined #asahi-dev
01:30liangqi has joined #asahi-dev
01:38liangqi has quit [Ping timeout: 480 seconds]
01:56liangqi has joined #asahi-dev
02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-dev
02:05liangqi has quit [Ping timeout: 480 seconds]
02:07voidah has joined #asahi-dev
02:13liangqi has joined #asahi-dev
02:17voidah has quit [Ping timeout: 480 seconds]
02:19michaelros44445 has joined #asahi-dev
02:29liangqi has quit [Ping timeout: 480 seconds]
02:38liangqi has joined #asahi-dev
02:38michaelros44445 has quit [Read error: Connection reset by peer]
02:39michaelros44445 has joined #asahi-dev
02:40voidah has joined #asahi-dev
02:48tobhe_ has joined #asahi-dev
02:49liangqi has quit [Ping timeout: 480 seconds]
02:54tobhe has quit [Ping timeout: 480 seconds]
02:58liangqi has joined #asahi-dev
03:06liangqi has quit [Ping timeout: 480 seconds]
03:11nicenickname6161 has quit [Ping timeout: 480 seconds]
03:17MikeM64 has quit [Quit: ZNC 1.9.1+deb2+b3 - https://znc.in]
03:25voidah has quit [Ping timeout: 480 seconds]
03:32liangqi has joined #asahi-dev
03:34voidah has joined #asahi-dev
03:37StanfordAppliedCyber[m]i can confirm btw that there's some apple custom funkiness in userspace and is NOT in the kernel
03:37StanfordAppliedCyber[m]the kernel hasn't required any patching other than genter --> hvc
03:37StanfordAppliedCyber[m]userspace is currently dying on SIGILL
03:37StanfordAppliedCyber[m]i'll figure out how the regular arm fallback paths work, hoping it's just some sysctls
03:40liangqi has quit [Ping timeout: 480 seconds]
03:42voidah has quit [Ping timeout: 480 seconds]
03:49nora has joined #asahi-dev
03:51nora_ has quit [Ping timeout: 480 seconds]
03:59nicolas17StanfordAppliedCyber[m]: any idea what process? or are you not even getting past launchd yet?
04:00StanfordAppliedCyber[m]fwiw i'm not super worried about most of userspace
04:00StanfordAppliedCyber[m]the kernel works fine and my emulation is good
04:10StanfordAppliedCyber[m]ive heard from multiple people that the userspace in vm images and bare metal boot images is the same, the only difference is the kernel
04:11StanfordAppliedCyber[m]if this is true that means if the kernel works fine there's a way of turning off all the apple bs in userspace
04:11StanfordAppliedCyber[m]oh ffs i'm hitting SIGILL on a fucking sysreg
04:15StanfordAppliedCyber[m]AND IT'S TIMER BULLSHIT
04:21nela has quit [Ping timeout: 480 seconds]
04:22michaelros44445 has quit [Ping timeout: 480 seconds]
04:26NickChan[m]we know from m3 that there's a whole set of parallel registers to arm's arch timer registers
04:26NickChan[m]it's first added in M3/A15
04:26StanfordAppliedCyber[m]yeah i found all of them in this hypervisor work
04:26StanfordAppliedCyber[m]fixed this particular error though
04:26StanfordAppliedCyber[m]launchd is now exiting for a different reason which i need to figure out
04:27r0ni has quit [Ping timeout: 480 seconds]
04:27nela has joined #asahi-dev
04:29eluks has quit [Remote host closed the connection]
04:29eluks has joined #asahi-dev
04:31michaelros44445 has joined #asahi-dev
04:38StanfordAppliedCyber[m]are there any resources from the m3 learnings?
04:39voidah has joined #asahi-dev
04:45StanfordAppliedCyber[m]i'm not sure why but writing to some of the new timer registers in el2 fault, despite me doing exactly what xnu is doing
04:46NickChan[m]StanfordAppliedCyber[m]: they are probably all locked
04:46StanfordAppliedCyber[m]...............but xnu writes them just fine?
04:46StanfordAppliedCyber[m]i find it hard to believe it's a raw vs macho boot difference
04:46StanfordAppliedCyber[m]is there precedent for that?
04:48voidah has quit [Ping timeout: 480 seconds]
04:49michaelros44445 has quit [Read error: Connection reset by peer]
04:49michaelros44445 has joined #asahi-dev
04:51michaelros44445 has quit [Read error: Connection reset by peer]
04:51michaelros44445 has joined #asahi-dev
04:58michaelros44445 has quit [Read error: Connection reset by peer]
04:58michaelros44445 has joined #asahi-dev
05:29pb17 has quit [Ping timeout: 480 seconds]
05:39pb17 has joined #asahi-dev
05:45liangqi has joined #asahi-dev
05:50r0ni has joined #asahi-dev
05:53liangqi has quit [Ping timeout: 480 seconds]
06:00michaelros44445 has quit [Ping timeout: 480 seconds]
06:02liangqi has joined #asahi-dev
06:06ddxtanx_ has joined #asahi-dev
06:06ddxtanx has quit [Read error: Connection reset by peer]
06:14StanfordAppliedCyber[m]actually yeah, does anyone know if there's a precedent for apple impdef registers behaving differently in macho boot vs raw boot? launchd is faulting trying to read ACNTVCT_EL0, the mrs x8 ACNTVCT_EL0 is being decoded as an illegal instruction and launchd is getting killed
06:14JamesCalligeros[m]Raw boot disables/locks a bunch of stuff including some instructions and some sysregs
06:14JamesCalligeros[m]Some of these will be bugs that apple is aware of and is fixing
06:15StanfordAppliedCyber[m]oh so this is expected?
06:15liangqi has quit [Ping timeout: 480 seconds]
06:15svenyeah, it’s the same mechanism that locks out gxf
06:15StanfordAppliedCyber[m]are there any known solutions for this other than waiting for apple to fix them?
06:15svenand they won’t fix all registers
06:16StanfordAppliedCyber[m]amazing, time to think of something interesting
06:16svenso far I’ve only filed a radar about that aic one
06:16sventhe timer regs will likely remain disabled because they’re not required for eg linux
06:16StanfordAppliedCyber[m]i think i'm going to do something horrible
06:17JamesCalligeros[m]I mean we’re already stubbing and emulating sptm, in for a penny…
06:17Sobek[m]Hmm, how can this be fixed ? Chicken bits / iBoot code change are enough ?
06:17StanfordAppliedCyber[m]you may be able to patch all the register accesses to svc, add a new syscall to xnu that just calls hvc (or find some way of making a hypercall directly from el0) and then software emulate the access
06:18JamesCalligeros[m]iboot does it before jumping to its boot payload so yes apple will “fix” what they want to fix with iboot updates
06:18StanfordAppliedCyber[m]what's "that aic one" btw? not familiar with that
06:18NickChan[m]the fiq virt timer
06:18JamesCalligeros[m]But as sven said, they wont “fix” what they dont consider broken (which is anything that is not required for strict arm spec compliance)
06:18habiby has quit [Remote host closed the connection]
06:18StanfordAppliedCyber[m]oh, i hit that, we have some horrible fiq handling code in my branch currently
06:18NickChan[m]doesn't have much to do with the actual aic it's however handled in linux's aic driver
06:19jannauwe probably need to fill a radar about CYC_OVRD as well
06:19habiby has joined #asahi-dev
06:19NickChan[m]as is the current state windows guests will also be broken even after all of that is fixed
06:20StanfordAppliedCyber[m]one of my friends is a very senior engineer who lead a lot of the sptm work and controls a lot of boot, may talk to him about it (obviously he hasn't told me anything that isn't public)
06:20jannauyou could try if we can trap those. there are unknown bits in HACR_EL2 and AHCR_EL2 might be another hyp config reg
06:20StanfordAppliedCyber[m]see if we can escalate some of the radars
06:20JamesCalligeros[m]Dear god i hope you did not talk to them to inform your development work…
06:20jannauthose == the apple counter sysregs
06:21NickChan[m]hacr_el2 is probably raz/wi; apple at least managed to have that behavior on actlr_el1
06:22NickChan[m]evident from the "advertise Apple TSO but it's broken" in some of the logs
06:24liangqi has joined #asahi-dev
06:27voidah has joined #asahi-dev
06:32ddxtanx has joined #asahi-dev
06:32ddxtanx_ has quit [Read error: Connection reset by peer]
06:35voidah has quit [Ping timeout: 480 seconds]
06:36liangqi has quit [Ping timeout: 480 seconds]
06:36StanfordAppliedCyber[m]may also be able to mess with commpage
06:36StanfordAppliedCyber[m]this is definitely a problem we can engineer a solution for
06:38Sobek[m](Wait, Apple TSO is not available in Raw boot now ?)
06:39StanfordAppliedCyber[m]TSO ISN'T AVAILABLE IN RAW BOOT?
06:39NickChan[m]definitely not available so no more gaming
06:39NickChan[m]no more windows guests cause no hacr
06:40StanfordAppliedCyber[m]..............................................................................................................................ironically microsoft could save us here, if they wanted windows to work on arm macs then apple would def work with them to make it happen
06:40StanfordAppliedCyber[m]actually no because that'd just be a vm in macos
06:40StanfordAppliedCyber[m]bare metal windows sounds horrifying for both apple and microsoft
06:41JamesCalligeros[m]lrcpc should be available right
06:41JamesCalligeros[m]Bare metal arm windows relies on acpi and a bunch of other awful shit, its not happening
06:41Sobek[m]JamesCalligeros[m]: Is that a fence instruction that's used when TSO mode isn't a thing ?
06:41StanfordAppliedCyber[m]yeah i realized that a second after i sent the message
06:43NickChan[m](all of that not withstanding putting linux in EL1 in macho boot mode which is perfectly possible)
06:43Sobek[m]NickChan[m]: But the whole SPTM makes it impractical right ?
06:44StanfordAppliedCyber[m]to put it mildly, yes, upstreaming that would be an utter nightmare
06:44NickChan[m]Sobek[m]: EL1 avoids the SPTM issues; imagine a Linux VM on macOS but you pass through more and more hardware
06:44NickChan[m]Linux in real EL2 under SPTM is actually impossible because userspace can't have RWX
06:45NickChan[m]but of course one would still need some thin hypervisor in EL2 to virtualize all the sptm stuffs so it's a very cursed VM
06:47Sobek[m]The question is, if we showed a PoC of that, what would ARM and Apple do ?
06:47Sobek[m]Or would it get them to stop being so vindictive about raw boot and Apple extensions ?
06:47Sobek[m]Would ARM strong-arm Apple more into breaking that approach / disabling all extensions in that case too ?
06:48NickChan[m]this gets you similar architectural capability as a pcc research guest, where booting a custom sptm is documented
06:51StanfordAppliedCyber[m]okay well patching commpage fixed the first issue but new sysreg
06:52StanfordAppliedCyber[m]i think it's clear i'm going to have to settle for modifying the sigill handler
06:52StanfordAppliedCyber[m]probably need some el1 helper to make things fast
06:59pb17 has quit [Ping timeout: 480 seconds]
07:00pb17 has joined #asahi-dev
07:11nela has quit [Ping timeout: 480 seconds]
07:12r0ni has quit [Ping timeout: 480 seconds]
07:12nela has joined #asahi-dev
07:18ddxtanx_ has joined #asahi-dev
07:18ddxtanx has quit [Read error: Connection reset by peer]
07:21ddxtanx has joined #asahi-dev
07:21ddxtanx_ has quit [Read error: Connection reset by peer]
07:22voidah has joined #asahi-dev
07:22chadmedim not entirely sure all of this was done at arm's behest
07:28Sobek[m]Fair, though TSO being out is very consistent with ARM past actions on the LKML. (I don't think crippling TSO mode on Asahi was in Apple's interest. (Unless this is part of a plan to phase out Rosetta for Linux too ? Apple did say Rosetta for Linux wasn't going away like Rosetta on macOS).
07:30ddxtanx has quit [Remote host closed the connection]
07:30voidah has quit [Ping timeout: 480 seconds]
07:32ddxtanx has joined #asahi-dev
07:38StanfordAppliedCyber[m]i was told that userspace is the same between the macos vm images and the bare metal installs
07:38StanfordAppliedCyber[m]how does this work with the custom sysregs
07:38StanfordAppliedCyber[m]is that information just outdated at this point and apple has different userspace for bare metal boot?
07:39foltz1 has joined #asahi-dev
07:45foltz has quit [Ping timeout: 480 seconds]
07:45foltz1 is now known as foltz
07:45r0ni has joined #asahi-dev
07:46yukajannau: I think even mac os does no longer use CYC_OVRD on these? Do you mean ACC_OVRD?
07:46NickChan[m]different userspace isn't compatible with booting macOS from an external drive
07:46NickChan[m]it must all be the same
07:48StanfordAppliedCyber[m]well then how do the custom sysregs/memory encryption work?
07:49jannauyuka: yes, just half read what you said then
07:49NickChan[m]check commpage and sysctl possibly
07:50StanfordAppliedCyber[m]yeah i'm doing that rn, commpage already had a fix for the timer issue
07:50StanfordAppliedCyber[m]just force it to use the syscall
07:50StanfordAppliedCyber[m]and xnu uses our emulated values
07:51liangqi has joined #asahi-dev
07:54liangqi has quit [Remote host closed the connection]
07:54liangqi has joined #asahi-dev
07:54liangqi has quit [Remote host closed the connection]
07:54liangqi has joined #asahi-dev
07:58asahi has joined #asahi-dev
08:08gargantua has quit [Ping timeout: 480 seconds]
08:20picnoir has quit [Ping timeout: 480 seconds]
08:23bear[m]1 has joined #asahi-dev
08:28picnoir has joined #asahi-dev
08:45voidah has joined #asahi-dev
08:45StanfordAppliedCyber[m]i think that there actually are ways of turning all this off
08:45StanfordAppliedCyber[m]there's just a lot of switches
08:45StanfordAppliedCyber[m]some of them are in the data section of the kernelcache macho that i'm just going to patch to 0
08:47kloenk has quit [Remote host closed the connection]
08:48kloenk has joined #asahi-dev
08:53voidah has quit [Ping timeout: 480 seconds]
08:59svenis that aic fiq virt thingy fixed now btw? i'll poke them again otherwise
09:01svenand did someone check that TSO really isn't available in raw boot mode anymore? or are we just guessing so far?
09:01svenand gaming still works without TSO, it's just a bit slower *shrug*
09:04StanfordAppliedCyber[m]can you link me to the radar so i can make sure i know what you're talking about exactly? all ik is that i had to add a bunch of fiq stuff to my hypervisor
09:04sveni can't link you because all radars are internal
09:04svenit's just that fiq virt mask thingy used in the aic driver
09:04kloenk has quit [Remote host closed the connection]
09:05svenSYS_IMP_APL_VM_TMR_FIQ_ENA_EL2
09:05sventhat one: #define SYS_IMP_APL_VM_TMR_FIQ_ENA_EL2 sys_reg(3, 5, 15, 1, 3)
09:05kloenk has joined #asahi-dev
09:05svenSobek[m]: don't assume any intention for raw boot mode, it's just very untested and they probably started with "everything disabled" and just put in a few exceptions
09:06svenand likely missed quite a bunch of stuff
09:06jannauyuka, chaos_princess: did you already benchmark cpu cores for capacity-dmips-mhz on m3 max?
09:06chaos_princessoh, forgot to ask - so how do you even do that?
09:07jannauI'm currently doing it on m3 and and the perf cores are slower than the efficiency cores (both at max frequency)
09:08jannauDocumentation/devicetree/bindings/cpu/cpu-capacity.txt
09:09asahi has quit [Ping timeout: 480 seconds]
09:09jannauI'm using coremark instead whrystome though since 40 year old C code is annoying
09:09jannaudhrystone
09:09chaos_princessok, reading failure on my part, i read it and failed to notice the part where you are supposed to run dhrystone
09:09chadmedistr we didnt use dhrystone for m1 or m2
09:10StanfordAppliedCyber[m]sven: this is exactly what my friend at apple said. he told me they didn't test it at all internally and apparently there were some conversations like "should we test this, we are shipping it" "sure, but after you do all your other work"
09:10chadmedthe actual benchmark is kind of immaterial, youre supposed to fix the pcores to 1024 and normalise everything relative to that
09:10svenStanfordAppliedCyber[m]: yeah, i know. i got it from people at apple as well
09:11svenunfortunately we are their QA and have to wait months until the bugs are fixed then :(
09:11chadmedwe might have used coremark instead of dhrystone
09:11StanfordAppliedCyber[m]so just like their regular QA process :)
09:11svenand their internal process for un-restricting sysregs in iboot is apparently especially annoying
09:13StanfordAppliedCyber[m]yeah i heard that a lot of their engineers actually like asahi linux, and there's been some meetings that are basically "should we put resources into supporting them"
09:13StanfordAppliedCyber[m]apparently they really like that we will never require customer support
09:13StanfordAppliedCyber[m]not to mention are paying for their products
09:16pjv has joined #asahi-dev
09:21roxfan has quit [Ping timeout: 480 seconds]
09:21NickChan[m]sven: I went off "CPU features: CPU advertises Apple TSO but it is broken, ignoring" in a message on matrix in the re channel
09:22NickChan[m]in the actual code it means "TSO bit exists in AIDR_EL1 but TSO bit is RAZ/WI in ACTLR_EL1"
09:23svencould try to file a radar about that as well if it's confirmed and see what happens :D
09:23NickChan[m]i guess it helps that the bit is documented on apple's site
09:23sveni don't have high hopes but it's worth a try
09:31jannaum3 pro looks better but I doubt 8.784 vs 10.570 coremark iterations/s per MHz are expected
09:31pjv has quit [Quit: pjv]
09:38roxfan has joined #asahi-dev
09:39chadmedecore still faster than pcore?
09:42jannauno, p-cores are faster both absolute and relative to their max freq but I'm not trusting the results. e-core at 83% of p-core at iso frequency seems to high
09:44jannauand 56% of the absolute values with both at their max freqency
09:49chadmedit wouldnt surprise me tbh. firestorm and icestorm were pretty close, so were avalanche and blizzard
09:50jannauoh, I should have looked at the m1/m2 capacity-dmips-mhz values. m2 with 756 / 1024
09:50chadmedblizz is like 74% of avalanche
09:51jannauwhy had I 300-400 in mind for the e-cores?
09:51chadmedbecause thats what crappy ecores usually come out at :P
09:54jannau"The CPU performance cores and efficiency cores are 30 percent and 50 percent faster than those in M1"
09:57jannauthat would put the e-cores at 80% of the p-cores based on our m1 capacity-dmips-mhz
09:58jannauso the 83% I get from coremark could be correct
09:58voidah has joined #asahi-dev
10:01chadmedyeah it sounds reasonable to me, especially if its reproducible
10:07voidah has quit [Ping timeout: 480 seconds]
10:11michaelros44445 has joined #asahi-dev
10:48voidah has joined #asahi-dev
10:56voidah has quit [Ping timeout: 480 seconds]
11:03jannauit's only core 4 (active at m1n1 startup) which is inexplicable slow on M3. 17688 vs. 42851 Iterations/Sec for core 4 vs. core 5, 6, 7
11:09jannauchaos_princess, yuka: see commit messages how I determined capacity-dmips-mhz
11:10jannausurprisingly quite a difference between t8122 and t6030
11:10voidah has joined #asahi-dev
11:12flokli has quit [Quit: WeeChat 4.9.0]
11:12flokli has joined #asahi-dev
11:19voidah has quit [Ping timeout: 480 seconds]
11:34jannaubad performance on the boot core is not related to core switching
11:43yukayou took the coremark number and then relativized them to 1024?
11:43chadmedthat is the correct way. the fastest cores in the system are by convention always 1024, and the others are normalised relative to that
11:44yukaSo then my E-cores are 591
11:45jannaudid you normalize them to the (max) frequency
11:45yukaAh, no
11:46chadmedthe measure is per mhz, you need to peg one core to its top pstate and then clamp the coremark task to it
11:47voidah has joined #asahi-dev
11:49yuka933 then
11:51jannaufaster e-cores at a lower freq and slower p-cores compared to m3: https://paste.debian.net/hidden/1791542a
11:57liangqi has quit []
11:58chadmed has quit [Quit: Konversation terminated!]
11:59chadmed has joined #asahi-dev
12:04tryhard[m]chaos_princess: where do you ignore devpwr requests before full boot in pmp v2?
12:06chaos_princessas you can see, i do set it in the shared sram area, i just don't wait for pmp to ack that
12:06chaos_princessthis specific bit is probably different with rtkit pmp
12:07berzrk has joined #asahi-dev
12:07tryhard[m]this status bit doesnt really exist
12:25chaos_princessso synthesize it based on rtkit messages
12:27gargantua has joined #asahi-dev
12:30tryhard[m]i think ill just use atomicbool
12:32tryhard[m]anybody understands what may be causing this? https://pastebin.com/AHc22Hb0
12:33tryhard[m]i think it's something related to the power domains?
12:57nicolas has joined #asahi-dev
12:57Jijdejiiedjjiedijede has joined #asahi-dev
12:58nicolas has quit [Remote host closed the connection]
12:58alysssa has joined #asahi-dev
13:13asahi has joined #asahi-dev
13:13tryhard[m]uh, i'm feel kinda dumb i didn't figure it out before. its just that pmp isnt fully boot so gfx isnt on and thus init the gpu causes a crasah
13:14alysssa has quit [Remote host closed the connection]
13:14Jijdejiiedjjiedijede has quit [Remote host closed the connection]
13:14tryhard[m]chaos_princess, how did you overcome this? i need to delay every soc init to after pmp
13:15chaos_princessi told you - i track whether pmp has finished booting, and if it haven't - i just set the bits and not wait for acks
13:16chaos_princessin your case - it will be harder, you will have to track your own set of things that should be reported to pmp, and when it is at the state when it is ready to accept them - send them all via a rtkit message
13:16tryhard[m]but i cant send the required rtkit msgs before its fully boot
13:17chaos_princessright, that is what i am saying - mark it as pending in your driver, and once it is ready - send
13:17tryhard[m]yeah but the socs dont care if i track it right?
13:17chaos_princessyep, you track it driver-side, not pmp-side
13:17tryhard[m]i mean, in this instance, gpu doesnt wait for gfx
13:18chaos_princessit should wait
13:18tryhard[m]the issue is it doesnt
13:18chaos_princesshow many pm domains do you have on gpu
13:18chaos_princessso, pmp_report_gfx?
13:19chaos_princessshow your enable function
13:19tryhard[m]turning it back to ps_gfx resolves the crash obviosuly
13:19tryhard[m]on which driver
13:19tryhard[m]report or pmp
13:19chaos_princessyou have a dependency on ps_gfx on pmp_report_gfx, right?
13:20chaos_princessshow send_devpwr
13:22tryhard[m]i kinda ignore devpwrs until pmp is fully booted but it shouldn't matter right now because gpu should wait
13:23chaos_princess... but your send_devpwr does not send anything
13:23tryhard[m]sure but if you see in the logs, gpu is crashing before full init
13:25tryhard[m]if its correct
13:25chaos_princessadd logging to pmgr driver to see if you are actually causing pmgr changes
13:25chaos_princessand to report drivers too, why not
13:26tryhard[m]chaos_princess: because of devpwr?
13:27tryhard[m]there wont be any changes before of devpwr msgs because obviosuly they're not being sent right now
13:29chaos_princessi am asking you to add logging to pmgr, not pmp
13:31tryhard[m]but like why would i care about pmgr changes if i dont cause them?
13:35voidah has quit [Ping timeout: 480 seconds]
13:38chaos_princessbecause you should be
13:38chaos_princessnot turning on the pmgr is what kills the gpu
14:11marvin24 has quit [Ping timeout: 480 seconds]
14:34tryhard[m]mhm. where should i add logging in pmgr?
14:38michaelros444455 has joined #asahi-dev
14:38chaos_princessin enable/disable callbacks
14:39chaos_princessi think you can just stick #define DEBUG at the top of the file and it will do it itself
14:44michaelros44445 has quit [Read error: Connection reset by peer]
14:47tryhard[m]defining DEBUG doesnt print anything
14:47tryhard[m]and adding to enable/disable also dont print
14:48sven#define dev_dbg dev_err at the very top should definitely work
14:48sveni think there's also some kernel commandline arg to enable debug for some files but i just can't be bothered to remember that one
14:50michaelros444455 has quit [Ping timeout: 480 seconds]
14:50tryhard[m]enable/disable doesnt really execute i guess
14:50svendo you have your full code somewhere?
14:50tryhard[m]it makes sense because i changed the power domains no?
14:51tryhard[m]lemme push
14:52marvin24 has joined #asahi-dev
14:54asahi has quit [Ping timeout: 480 seconds]
15:01jannau"debug" on the kernel command line is needed for dev_dbg, which is annoying because systemd components interpret that as well
15:01tryhard[m]already send that :D
15:12sventhere’s some dynamic debug thing that only enabled dev_dbg for a single file iirc
15:15MikeM64 has joined #asahi-dev
15:18tryhard[m]nevertheless inserting printfs on enable/disable function should show
15:27tryhard[m]maybe i should move the full initialization to the probe function? that way all socs will wait
15:27ralim8 has joined #asahi-dev
15:27voidah has joined #asahi-dev
15:32tpw_rules has quit [Ping timeout: 480 seconds]
15:33ralim has quit [Ping timeout: 480 seconds]
15:33ralim8 is now known as ralim
15:34chaos_princesspmp takes forever to boot, any approach that waits for it to finish is note feasible
15:40tryhard[m]alright i managed to do something
15:40tryhard[m]now devpwr is being sent before
15:40tryhard[m]although the crash still happens
15:41tryhard[m]its obviously happening when i'm changing from <&ps_gfx> to <&pmp_report_gfx> for gpu
15:41sofus has joined #asahi-dev
15:41tryhard[m]otherwise it doesnt happen... why is that?
15:44tryhard[m]ok actually devpwr isnt even being sent for gfx...
15:46michaelros444455 has joined #asahi-dev
15:49michaelros4444558 has joined #asahi-dev
15:53michaelros4444559 has joined #asahi-dev
15:53michaelros444455 has quit [Read error: Connection reset by peer]
15:53michaelros4444559 is now known as michaelros444455
15:57tpw_rules has joined #asahi-dev
16:00michaelros4444558 has quit [Ping timeout: 480 seconds]
16:05gargantua has quit [Ping timeout: 480 seconds]
16:06alysssa has joined #asahi-dev
16:07Jijdejiiedjjiedijede has joined #asahi-dev
16:07voidah has quit [Ping timeout: 480 seconds]
16:07voidah has joined #asahi-dev
16:08michaelros444455 has quit [Ping timeout: 480 seconds]
16:08pb17 has quit [Ping timeout: 480 seconds]
16:10pb17 has joined #asahi-dev
16:15alysssa has quit [Remote host closed the connection]
16:16alysssa has joined #asahi-dev
16:16tryhard[m]devpwr is being sent for dispdfr. how is that related to gfx?
16:20alysssa has quit [Remote host closed the connection]
16:20Jijdejiiedjjiedijede has quit [Remote host closed the connection]
16:32chaos_princessthis is the touchbar display controller
16:33gargantua has joined #asahi-dev
16:53MikeM64 has quit [Quit: ZNC 1.9.1+deb2+b3 - https://znc.in]
16:54tryhard[m]maybe the gpu needs to be probing in a certain order?
16:54tryhard[m]like after some soc?
16:54tryhard[m]i see it usually probes after usb and before smc
16:54tryhard[m]chaos_princess: also, why would it ask for it?
16:55tryhard[m]m1 doesnt have a touchbar
16:55chaos_princessthe short answer is that you did something wrong
16:55chaos_princessidk what exactly, but it should not do that
16:55chaos_princessand no, order does not matter
16:56Ardu0us_ has joined #asahi-dev
16:59Ardu0us has quit [Ping timeout: 480 seconds]
17:00tryhard[m]ill just put status = disabled
17:00tryhard[m]or even just remove that node
17:15sofus has quit [Remote host closed the connection]
17:28sofus has joined #asahi-dev
17:43lukeleppan91 has quit [Read error: Connection reset by peer]
17:43lukeleppan91 has joined #asahi-dev
17:59regis has joined #asahi-dev
18:00regis has quit []
18:12yukaSpeakers are working nicely on t514m, thanks chaos_princess
18:35ddxtanx has quit [Remote host closed the connection]
18:44ArcaneNibble has quit [Quit: ZNC 1.10.2 - https://znc.in]
18:45ArcaneNibble has joined #asahi-dev
19:02Jijdejiiedjjiedijede has joined #asahi-dev
19:02alysssa has joined #asahi-dev
19:06nicolas17sven: maybe you can put your aic radar on https://openradar.appspot.com/
19:11sveneh.. it wants my google account to login. i'd rather not
19:11Jijdejiiedjjiedijede has quit [Remote host closed the connection]
19:12svenFB21831949 / Write access to sys_reg(3, 5, 15, 1, 3) / APL_INTENABLE_EL2 faults in non-Apple-entitled boot mode
19:12sventhat's it essentially :D
19:12svenjust had to create it to send it to someone i know
19:18sofus has quit [Read error: Connection reset by peer]
19:20sofus has joined #asahi-dev
19:22alysssa has quit [Remote host closed the connection]
19:25amarioguyreading asahi backlog about windows guests....man all paths really are leading back to "kernel patching is the sustainable path" despite everything i've ever read saying no...
19:25amarioguy(on m4/m5, because of SPTM troubles)
19:29amarioguyoh wtf just read again, hacr's gone?? ughhh
19:32amarioguy> Bare metal arm windows relies on acpi and a bunch of other awful shit, its not happening
19:33amarioguyfwiw, i'm not entirely sure you *need* the cursed parts of ACPI, all you actually need is a pseudo-DT in ACPI format, and have drivers carry the rest
19:35amarioguythe actually awful part is the lack of ARM PMU support unironically, because it means that any instance of those reads if Apple doesn't add those MSRs in future cores will fault, and must be changed
19:49lacamar has joined #asahi-dev
19:53lacamar_ has quit [Ping timeout: 480 seconds]
20:01pjv has joined #asahi-dev
20:18yurix has quit [Read error: Connection reset by peer]
20:18yurix has joined #asahi-dev
20:22fumnanya has joined #asahi-dev
20:27fumnanya has quit []
20:53sofus has quit [Quit: leaving]
21:19voidah has quit [Ping timeout: 480 seconds]
21:22voidah has joined #asahi-dev
21:50pb17 has quit [Ping timeout: 480 seconds]
21:52pb17 has joined #asahi-dev
22:05pjv has quit [Quit: pjv]
22:09gargantua has quit [Ping timeout: 480 seconds]
22:18gargantua has joined #asahi-dev
22:29hub2 has quit [Quit: Ping timeout (120 seconds)]
22:29hub2 has joined #asahi-dev
23:22voidah has quit [Ping timeout: 480 seconds]
23:29voidah has joined #asahi-dev
…