2026-07-02
00:00StanfordAppliedCyber[m]also, i did look at the m3 sptm, and the answewr is that this endpoint just looks very different, the sysreg solution i thought worked earlier does in fact not work (i read the asm wrong)
00:02StanfordAppliedCyber[m]i havent fully looked into how m3 sptm works, but i'm not optimistic we can just copy the m3 sptm method
00:18StanfordAppliedCyber[m]specifically the m3 just lacks the entire codepath that the a18 pro sptm binary has
00:38cow123 has quit [Ping timeout: 480 seconds]
00:46cow123 has joined #asahi-dev
00:56StanfordAppliedCyber[m]i did have one idea: figure out how xnu allocates pages and specifically where it gets the entropy, patch the entropy to some fixed value so that page allocation becomes deterministic, then in the sptm emulator make sure that page is always mapped NC forever
00:59StanfordAppliedCyber[m]this has obvious risk though, and i'd like to see if anyone else has any better ideas before doing this
01:08StanfordAppliedCyber[m]or use breakpoints on the allocator and for specifically this call return some address, i suppose that also works
01:15pb17 has quit [Ping timeout: 480 seconds]
01:20StanfordAppliedCyber[m]also, i did try the dumb thing of just adding a sleep to the retype endpoint and that didn't solve the problem, unsurprisingly
01:20StanfordAppliedCyber[m]i didn't try just accessing a bunch of other pages to try to flush the fabric that way, i suppose i should try that
01:32hdbngr has quit [Ping timeout: 480 seconds]
01:40pb17 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]
02:08nela has quit [Ping timeout: 480 seconds]
02:12nela has joined #asahi-dev
02:14ciara has joined #asahi-dev
02:41donaldh has quit [Ping timeout: 480 seconds]
02:56jn has joined #asahi-dev
03:00starcatmeow has joined #asahi-dev
03:00jnn has quit [Ping timeout: 480 seconds]
03:18r0ni has quit [Quit: Textual IRC Client: www.textualapp.com]
03:18turboretor6 has joined #asahi-dev
03:23turboretor has quit [Ping timeout: 480 seconds]
03:23turboretor6 is now known as turboretor
03:28integralpilot has joined #asahi-dev
03:28integralpiloti was messing around with random NVRAM vars so now i have to DFU revive LOL
03:29JamesCalligeros[m]yeah be careful with nvram. a lot of the documented nvram vars are intel only and will send you to dfu if you apply them to apple silicon
03:32integralpilotin my case i saw that there was one var auto-boot=TRUE and was like "well it annoys me when it turns on automatically when i open the lid, i'll change it to FALSE" and now i get DFU. its reviving and should be fine, but lesson learned i guess
03:39marvin24 has joined #asahi-dev
03:40integralpilot"Downloading Mac software update - 60MB of 19.77 GB - 2 hours remaining" :(
03:45integralpilotah thank you, good to know for when it finishes reviving!
03:46NickChan[m]auto-boot false should send to iboot recovery shouldn't it
03:47NickChan[m]in which case... (full message at <https://matrix.org/oftc/media/v1/media/download/AZGUH4kLBBXK0oIZaUyA8vBs7vJqQ-Vl6bLN9tEN12KJGIjtMfPLaf8ghlfhSD1skp65PlIw_pqU3TIKvtNhHExCefaQnBdwAG1hdHJpeC5vcmcvWGNwbVhESG9FcUxCanJrSFhYTHZJZWlT>)
04:00dustcurve has quit [Remote host closed the connection]
04:00dustcurve has joined #asahi-dev
04:01nora_ has joined #asahi-dev
04:03nora has quit [Ping timeout: 480 seconds]
04:04integralpilotyes, it does go to iboot recovery. thank you so much for that, should save me a lot of time vs reviving, will try now
04:15integralpilotit did work, thanks!
04:18dustcurve has quit [Remote host closed the connection]
04:18dustcurve has joined #asahi-dev
04:34donaldh has joined #asahi-dev
04:42integralpilot has quit [Quit: integralpilot]
04:43integralpilot has joined #asahi-dev
04:49mattia013 has joined #asahi-dev
06:36sofus has joined #asahi-dev
06:39sofus has quit []
06:39Sobek[m]Checking a thing, boot loop on Asahi after updating macOS 27 beta is a known issue due to the battery firmware change ? Or that would result in different symptoms ?
06:39Sobek[m](I really want to turn anyone who installs macOS beta and surprises pikatchu when it doesn’t work into a newt(on). Obviously they didn’t give exact versions)
06:39integralpilotyes, the bootloop is a result of the battery firmware change
06:40sofus has joined #asahi-dev
06:40integralpilotyou can either: 1. stop macsmc-power from loading, 2. apply the patch & rebuild kernel, 3. DFU revive to get rid of the newer SMC firmware
06:41Sobek[m]Yaaay, so that issue is basically a duplicate bug. Do we have a Asahi-radar number to close the issue as duplicate of asahi-radar:xyz ? 😈
06:46hdbngr has joined #asahi-dev
06:50chaos_princessSobek[m]: 517
06:53pb17 has quit [Ping timeout: 480 seconds]
07:03Sobek[m]integralpilot: FYI see https://social.treehouse.systems/@AsahiLinux/116813000429256653 there are now suitable kernels.
07:05pb17 has joined #asahi-dev
07:18integralpilotthat's good, thanks for LMK! im not the most up to date with when fedora does it packaging and stuff
07:19integralpilotok about to start upstream wifi test v2, it _should_ work this time
07:20Sobek[m](I had to be corrected by chaos princess, I'm definitely sharing the updated info around. Kind of shocked to get two different people reportingt he battery issues in the span of an hour at most).
07:21JamesCalligeros[m]Im not surprised people want to try macos 27 but like we have said multiple times to not do so for now
07:24StanfordAppliedCyber[m]it appears that on a18 pro there's some sections of the preloaded firmware we have to patch otherwise agx crashes
07:26Sobek[m](Uh, what ? Is this an apple bug ? Or more likely a bug on our side ? The patching firmware doesn't sound like a solution that will fly with the kernel in the long run, right ?)
07:27StanfordAppliedCyber[m]does BtpG or StpG or BlpP or S2xG ring a bell to anyone?
07:27hdbngr_ has joined #asahi-dev
07:27StanfordAppliedCyber[m]i'm trying to figure it out right now
07:28enverbalalic has joined #asahi-dev
07:29starcatmeow has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
07:30StanfordAppliedCyber[m]yeah i'll need it
07:33svenis this something xnu does now? I’d expect all coproc firmware to be read only by the time we get control
07:34chaos_princessbootargs?
07:34StanfordAppliedCyber[m]no, this isn't something xnu does, i think it's something iboot is NOT doing
07:34hdbngr has quit [Ping timeout: 480 seconds]
07:34StanfordAppliedCyber[m]i don't have great answers here
07:34StanfordAppliedCyber[m]here's what i know:
07:34integralpilotas in, iboot only doesnt do when booting us and does it for macos?
07:35StanfordAppliedCyber[m]this is a hypothesis to be clear
07:35StanfordAppliedCyber[m]the gpu firmware is crashing
07:36StanfordAppliedCyber[m]let me grab the exact crash right now so i don't say something wrong
07:36StanfordAppliedCyber[m]SAPTAccessViolation addr=0x600
07:36enverbalalic has quit [Ping timeout: 480 seconds]
07:37StanfordAppliedCyber[m]err_status 0x80000008
07:37StanfordAppliedCyber[m]axi_req 0xe1243
07:37StanfordAppliedCyber[m]sid 1/0
07:38StanfordAppliedCyber[m]well for obvious reasons i thought maybe my dart or uat handling wasn't good, ie, one of my endpoints was just messed up, some memory wasn't mapped properly, crash
07:39chaos_princessmaybe you are missing axi2af bridge configs?
07:39chaos_princessthey are needed on m1/2
07:39StanfordAppliedCyber[m]what are these configs?
07:40integralpilot has quit [Quit: integralpilot]
07:40chaos_princesspokes to pmgr mmio that do... stuff. Without those stuff breaks in subtle ways, like caches going non-coherent
07:41StanfordAppliedCyber[m]can you point me to some files in m1n1 or linux?
07:44StanfordAppliedCyber[m]hmmmm, well obviously i havent added t8140 support to that file, but i'd imagine that xnu takes care of thta for me
07:44chaos_princessit is not impossible that this setup was moved into sptm
07:45StanfordAppliedCyber[m]sptm does zero mmio
07:45StanfordAppliedCyber[m]i'm extremely confident of that
07:45chaos_princessdarts?
07:45StanfordAppliedCyber[m]i don't think sptm is at all relevant to understanding hardware (other than the iommus obviously)
07:48integralpilot has joined #asahi-dev
07:48integralpilotok so wifi is still not working but there's progress!
07:48integralpilotwe successfully load the firmware and it boots and it starts talking to us and giving us log messages from the wifi
07:48StanfordAppliedCyber[m]this is m3?
07:48integralpilotyes. wifi is already working downstream, but i am working on upstreaming it right now
07:49integralpilotso the current results are with upstream 7.2-rc1 with my current WIP patches for wifi
07:49StanfordAppliedCyber[m]oh, nice, what wifi chip?
07:49chaos_princessnot yours :P
07:50chaos_princessbrcm something-or-another
07:50integralpilotbrcm4388 i think
07:50StanfordAppliedCyber[m]yeah my chip already has the driver upstream :)
07:51chaos_princessit is an off the shelf mediatek one
07:52chaos_princessand yes, it is probably apple-flavoured and will need a few changes here and there
07:52StanfordAppliedCyber[m]it may be slightly custom
07:52Sobek[m](Also, have we got any macs with Apple homegrown radio chips ? Or not yet but probably coming soon)
07:52chaos_princessSobek[m]: m5 pro/max
07:52StanfordAppliedCyber[m]but it's probably add some stuff to the existing driver with mostly stuff we understand
07:52StanfordAppliedCyber[m]not have fun with a fully custom driver
07:53StanfordAppliedCyber[m]whats the radio chip used for?
07:53chaos_princesswifi/bt/thread
07:53StanfordAppliedCyber[m]thread?
07:53chaos_princessiot radio
07:53Sobek[m]chaos_princess: (More *fun* challenges in addition to SPTM and raw mode disabling everything under the sun).
07:54StanfordAppliedCyber[m]what other fun stuff is in m5 pro/max?
07:54StanfordAppliedCyber[m]tbh the a18 isn't that crazy
07:54chaos_princessnew gpu. but vanilla m5 also has it
07:54StanfordAppliedCyber[m]like there's a lot of craziness, but i've been able to solve most of my issues save for smp and debugusb
07:54chaos_princessand i think that's it?
07:55StanfordAppliedCyber[m]great, i just bought a m5 so i'ma look at that first :)
07:55StanfordAppliedCyber[m]comes in the next 2 weeks
07:56Sobek[m]sven: Yeah, especially since Apple probably put a lot in firmware. It would be hilarious if M5 Pro / Max get reliable radio before their brcm using siblings.
07:58Sobek[m]sven: I think it's reliable now, but I was thinking more of M5 / M4 macs which are still work in progress (siblings more than ancestors).
07:58chaos_princessm4 shouldn't need changes
07:58chaos_princessm2 needed them since that brought 6ghz
07:58StanfordAppliedCyber[m]used on ebay
07:58StanfordAppliedCyber[m]got my club to pay for one
07:58integralpilotclub?
07:58StanfordAppliedCyber[m]stanford applied cyber is a stanford club focused on cybersecruity
07:59svenand there’s no way we get a full wifi/bt driver ready before adding some small things here and there for brcm again
07:59integralpilotoh, cool!
07:59StanfordAppliedCyber[m]we're privileged enough to have budget for interesting projects like this, even though it's not our main focus
08:00integralpilotthats awesome!
08:03yukaSobek: I got an M5 Pro at pre-increase market price and submitted it as reimbursement to the asahi opencollective, pending processing
08:04yukanot sure anyone got the M5 Max but it shouldn't be too much different, and a friend has one which I can probably borrow for a week when it comes to that
08:04StanfordAppliedCyber[m]knock wood no m3 pro vs max situation :)
08:04StanfordAppliedCyber[m]the cpu tile on m5 pro and max is the same right?
08:04StanfordAppliedCyber[m]just the gpu tile is bigger?
08:06integralpilotoooh i found the wifi issue!!
08:07integralpilotrebuilding now, it was a tiny change (-7 / +8) but i think it should be working now
08:09StanfordAppliedCyber[m]how does the linux dev loop work? like just stream the kernel to m1n1 every test?
08:09integralpilotyuka: no, i got that working properly. it was with how we talk to the firmware
08:09chaos_princessthat is one option.
08:09integralpilotthat's what i do
08:09chaos_princessOr you can `make install` and reboot
08:09StanfordAppliedCyber[m]is there an easy way to pop a serial shell on linux over the uart?
08:09integralpilotif you have m1n1 in EL2
08:09StanfordAppliedCyber[m]i presume so
08:10chaos_princessyes, m1n1 virtual serial works
08:10integralpilotthen linux in EL1, m1n1 sets up a virutal seiral
08:10integralpilotthats what i do and it works rly well
08:10StanfordAppliedCyber[m]is that marcan's "world's stupidest vuart" file?
08:10StanfordAppliedCyber[m]yeah, that's what i plan to do
08:10StanfordAppliedCyber[m]i'ma need to figure out how it works shortly
08:10StanfordAppliedCyber[m]for weird logistical reasons it's rare i get hands on time with the mac so the dev workflow si remote
08:11StanfordAppliedCyber[m]i'm currently talking to you over an rdp session
08:11integralpilotbasically, its just a normal tty device. on the linux running in EL1 it appears as /dev/ttySAC0. in the connected linux device, it appears as /dev/ttyACM1
08:11chaos_princessit just emulates a hardware uart
08:11integralpilotyou can just echo & cat to it as needed
08:11integralpilot(i.e. i always do dmesg > /dev/ttySAC0 as part of my init script)
08:12integralpilotfor quick dev work i dont recommend having a full fedora or anything
08:12integralpilotwhat i use is alpine mini rootfs
08:13StanfordAppliedCyber[m]yeah i'm not, i don't care about actually using linux on these
08:13StanfordAppliedCyber[m]i was just gonna make a nice initrd with all the utilities i'd need
08:13StanfordAppliedCyber[m]appreciate the advice though
08:14StanfordAppliedCyber[m]for my actual work unfortunately x86 is nonnegotiable
08:14integralpilotperfect then!
08:15integralpilotthe wifi fix did not work and same error :(
08:19StanfordAppliedCyber[m]best of luck soldier, don't let the mac win
08:20integralpilotlol thx!!
08:39integralpilotmy cat is very interested in the wifi debugging and is now sitting on my mac
08:51donaldh has quit [Ping timeout: 480 seconds]
08:59chaos_princessm3 pro is materially different from m3 max instead of being a chopped max like in m1/2
09:00chaos_princesslike, it is more similar to m4 in certain ways
09:05r0ni has joined #asahi-dev
09:10bennofs has quit [Quit: The Lounge - https://thelounge.chat]
09:13integralpilotjannau: thanks! i got that part already working from the kernel, its just talking with the firmware once its booted that im solving (i am about to test smth now that should hopefully work)
09:13integralpilotAmi[m]: ill give her a little bit of ham rn now from you!
09:13bennofs3 has joined #asahi-dev
09:14integralpilotupdate: she liked it!
09:38bennofs3 has quit [Quit: The Lounge - https://thelounge.chat]
09:39bennofs3 has joined #asahi-dev
09:48integralpilotok wifi update we get further, now the issue is when we ask firmware for BRCM_C_GET_BANDLIST the firmware then rejects it with -52.
09:48integralpiloti think i see how to fix it, apple did some unqiue stuff here compared to the plain version and we need a bigger buffer and prehaps some parsing changes
09:53integralpiloti meant i see how to fix it by which stuff i need to add from downstream to adapt to the apple version
09:54MuratIlyas[m] has joined #asahi-dev
09:55integralpilotjust need to reshape a little bit to be better for upstream imo
10:03integralpilotIT WORKS NOW!!!
10:03integralpilotfinally
10:03integralpilotwlan0 appears and it probes right
10:03integralpilotlets see if it can connect
10:03beastd|2 has joined #asahi-dev
10:06integralpilotthere's a warning about the 6GHz band the firmware reports back being unrecognised
10:06chaos_princessa lot of downstream changes are specifically about 6ghz
10:06integralpilotshould i add 6Ghz support now (for this patchset) or should it wait untill a later patchset after base wifi works
10:07integralpiloti am thinking just skip it for now?
10:07beastd has quit [Ping timeout: 480 seconds]
10:12JamesCalligeros[m]No harm in getting everything working imo
10:12JamesCalligeros[m]But it might be a big patchset
10:14integralpilotyeah, thats my concern
10:15integralpilothow busy / picky / would not like big patchsets are the network maintainers
10:15jamie__ has quit [Remote host closed the connection]
10:15jamie__ has joined #asahi-dev
10:16integralpilotyep, sounds good then
10:16jannauthose are two different things and iirc there were complaints about too large brcmfmac series specifically
10:18enverbalalic has joined #asahi-dev
10:18integralpilotoh thanks, definently should split then
10:26enverbalalic has quit [Ping timeout: 480 seconds]
10:29integralpilothow big is too big?
10:29integralpilotabout +600 is needed to make scanning for networks work properly (without this the firmware still initalises fine)
10:30integralpilotso im debating for the initial series: just have it initalise it right (can't connect to anything yet) vs add that in too
10:42Guest12229 has quit [Ping timeout: 480 seconds]
10:45enverbalalic has joined #asahi-dev
10:45enick_590 has joined #asahi-dev
10:53donaldh has joined #asahi-dev
10:56integralpiloti think it is?
10:56integralpilotim not sure how to even test it though, the wifi in my house is just 2.4Ghz or 5Ghz
11:01dustcurve has quit [Remote host closed the connection]
11:01dustcurve has joined #asahi-dev
11:09dustcurve has quit [Remote host closed the connection]
11:09dustcurve has joined #asahi-dev
11:09chadmedi used to run iwd exclusively but it started behaving poorly and having weird failure modes and dropouts so i dumped it and went back :(
11:11yukatook me ~5 tries (restarting iwd etc, switching networks) to connect to the 6GHz successfully
11:16integralpilotjust fixed an error with sending the PSK to the firmware, about to try if the upstream patches work now (which i really think they should)
11:18integralpilothad to re-write the code for it though, because i dont like how it was done downstream and hopefully its a bit nicer
11:19chadmedi want to clean up the asoc machine driver before submitting it too but i know a lot of the mess is load bearing
11:19integralpilotnot sure how to test 802.1X? what is that again sorry
11:20integralpilotwell if i stand outside my school maybe?
11:20integralpilot(joking)
11:20integralpilotseriously, ill see if its in the wifi settings for my router
11:21chaos_princessdoes your school have eduroam
11:22integralpilotchaos_princess i think thats only for unis
11:23chadmednah most large institutions have it, all the qld health facilities and most of the private hospitals up here do
11:23integralpilotwe dont
11:23integralpilotour normal wifi is 802.1x though
11:24integralpilotoh yeah i guess that makes sense cause like medical students?
11:24integralpilotinteresting
11:24chadmeddont go to school on your break lol, i just set up 802.1x for wireless and ethernet at $work recently so worst case scenario i can take a mac to work one day next week
11:27integralpilotthank you!
11:29integralpilotonce im done (which is hopefully tonight) ill push my patches ontop of v7.2-rc1 to GH if that works?
11:31integralpilotahhh but those use spi hid right
11:31integralpilotif dc might have worked :(
11:32integralpilotah yeah can just use serial to get fedora shell (that's what i'm doing now)
11:32chaos_princessyou should probably take the m2
11:32chaos_princessas it uses a newer chip
11:37aites[m]Heya all! I was on this matrix room not so long ago and I was trying to figure out how to boot macos under the m1n1 hypervisor - Just came to say thank you for the help chaos_princess and co, I finally managed to sort it out, for some reason my macbook only lets me install older versions of macos like ventura 13.5 after I've wiped the drive, I think for some reason having an asahi partition made it not want to install. Either way, after
11:38* aites[m]uploaded a video: (8548KiB) < https://matrix.org/oftc/media/v1/media/download/ATUcBcreMqe75JkfJmJRXQBt4p64j15s6oGXMbdPowEeLDbhQuFRkD690APdy4fWenLZ08lxcpktx-LwPoawdi5CefargyzQAG1vemlsbGEub3JnLzEzNzQ0Yzc2MDg4ODI1Njc5NmNmY2NmOTYzOTYzNzliZjFlNDYyMGQyMDcyNjQ1OTc2NTg5MzM2NTc2 >
12:18integralpilot has quit [Remote host closed the connection]
12:51mripard has quit [Quit: WeeChat 4.9.2]
12:52mripard has joined #asahi-dev
12:55floklimischa85: nice on t6041! can you try https://cyberchaos.dev/yuka/linux/-/commit/3bf49bf384d90f91503369de26a09380bd88bf77 is also sufficient?
13:01amarioguy has joined #asahi-dev
13:14integralpilot has joined #asahi-dev
13:15integralpilotim giving up for the night on wifi - still having some errors when trying to join network
13:15integralpilot has quit []
13:48jsk has joined #asahi-dev
13:51alessaba has joined #asahi-dev
13:52jsk has quit [Remote host closed the connection]
14:05alessaba has quit [Ping timeout: 480 seconds]
14:06alessaba has joined #asahi-dev
14:13yurix has quit [Remote host closed the connection]
14:14enverbalalic has quit [Ping timeout: 480 seconds]
14:15yurix has joined #asahi-dev
14:21alessaba has quit [Ping timeout: 480 seconds]
14:22alessaba has joined #asahi-dev
14:30enverbalalic has joined #asahi-dev
14:52enverbalalic has quit [Ping timeout: 480 seconds]
14:59pb17 has quit [Ping timeout: 480 seconds]
15:01pb17 has joined #asahi-dev
15:12rodmac has joined #asahi-dev
15:19enverbalalic has joined #asahi-dev
15:41rodmac has quit [Quit: leaving]
15:42enverbalalic has quit [Ping timeout: 480 seconds]
16:35svenhah, i just realized that we can do the psci switch more or less without breaking users: add a psci=on/off m1n1 bootarg that adjusts the device tree to psci or spin-table and once a os with psci-via-efi support has booted just switch the bootarg inside m1n1
17:58mischa85flokli: i'm not sure. i have WFI nopped in the HV. cannot succesfully boot Linux on baremetal yet
18:13mischa85yuka: full smp with the hv wfi-trap brings up all 16 cores. with the trap off + idle=nop it borks at guest-secondary-entry. single-core idle=nop works fine.
18:14jannaufor fedora-asahi-remix update-m1n1 could only add psci=on when all installed kernels support it based on the kernel config of the installed kernels
18:20enverbalalic has joined #asahi-dev
18:23svendepending on how fast i finish this and get it upstream we could also just say that for m4+ it always defaults to psci=on
18:28enverbalalic has quit [Ping timeout: 480 seconds]
18:38enverbalalic has joined #asahi-dev
18:46enverbalalic has quit [Ping timeout: 480 seconds]
18:59boubou has joined #asahi-dev
18:59boubou is now known as Guest12504
19:02rosalyntg_ has joined #asahi-dev
19:02rosalyntg has quit [Remote host closed the connection]
19:22enverbalalic has joined #asahi-dev
19:37Calandracas_ has joined #asahi-dev
19:41Calandracas has quit [Ping timeout: 480 seconds]
19:42flokli has quit [Quit: WeeChat 4.9.1]
19:43flokli has joined #asahi-dev
19:46Calandracas_ has quit [Ping timeout: 480 seconds]
20:39amarioguy has quit [Ping timeout: 480 seconds]
20:47starcatmeow has joined #asahi-dev
21:48hardy94 has joined #asahi-dev
22:26pb17 has quit [Ping timeout: 480 seconds]
22:32SternXD has joined #asahi-dev
22:37pb17 has joined #asahi-dev
22:41starcatmeow has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
23:10leah has quit [Ping timeout: 480 seconds]
23:14ninelore has quit [Ping timeout: 480 seconds]
23:29ninelore has joined #asahi-dev