#asahi-dev · 2026-07-01 00:06
StanfordAppliedCyber[m]not to mention, if my understanding is correct, it's not needed for linux#asahi-dev · 2026-07-01 00:00
StanfordAppliedCyber[m]i thought apple's position, though not explicitly states, was anything arm defined would work, anything impdef/apple custom is locked#asahi-dev · 2026-06-30 23:59
StanfordAppliedCyber[m]also why do we expect apple to unlock SYS_IMP_APL_VM_TMR_FIQ_ENA_EL2 at all? after all it's an impdef register#asahi-dev · 2026-06-30 23:52
StanfordAppliedCyber[m]so all m5 class socs are still borked?#asahi-dev · 2026-06-30 09:41
StanfordAppliedCyber[m]that repo is a year old? i dsaw it but didn't see gpu stuff#asahi-dev · 2026-06-30 09:41
StanfordAppliedCyber[m]like do we know isa details, presumably firmware abi is still a giant question mark (and i don't even have an m3 to look)#asahi-dev · 2026-06-30 09:41
StanfordAppliedCyber[m]chaos_princess: can you link to the m3 work? i'm very interested in the gpu#asahi-dev · 2026-06-30 09:03
StanfordAppliedCyber[m]gracias chaos princess#asahi-dev · 2026-06-30 09:03
StanfordAppliedCyber[m]oh that does look useful#asahi-dev · 2026-06-30 08:58
StanfordAppliedCyber[m]and the clipboard is messed up#asahi-dev · 2026-06-30 08:58
StanfordAppliedCyber[m]i'd copy paste to you the errors but i'm actually working over an RDP session#asahi-dev · 2026-06-30 08:57
StanfordAppliedCyber[m]one is some UNEXP_RT_ something#asahi-dev · 2026-06-30 08:57
StanfordAppliedCyber[m]like, one is some dart disp0 thin#asahi-dev · 2026-06-30 08:56
StanfordAppliedCyber[m]like i have three issues all around memory mappings#asahi-dev · 2026-06-30 08:56
StanfordAppliedCyber[m]the agx issue is a memory cashing thing#asahi-dev · 2026-06-30 08:56
StanfordAppliedCyber[m]i can literally get a trace#asahi-dev · 2026-06-30 08:56
StanfordAppliedCyber[m]dcp works#asahi-dev · 2026-06-30 08:56
StanfordAppliedCyber[m]but i'm very, very late into boot#asahi-dev · 2026-06-30 08:56
StanfordAppliedCyber[m]i'm admittedly very confused what's happenning because iboot thinks that this is a 26.5 install#asahi-dev · 2026-06-30 08:55
StanfordAppliedCyber[m]wdym?#asahi-dev · 2026-06-30 08:54
StanfordAppliedCyber[m]i'm not looking forward to upstreaming this hypervisor#asahi-dev · 2026-06-30 08:54
StanfordAppliedCyber[m]the windowserver progress bar is 50% full crashing on some agx thingy#asahi-dev · 2026-06-30 08:53
StanfordAppliedCyber[m]i still need to create a new volume, but that can wait#asahi-dev · 2026-06-30 08:53
StanfordAppliedCyber[m]i managed to get past the sep regression usig some mega cursed adt hacks#asahi-dev · 2026-06-30 08:14
StanfordAppliedCyber[m]it's in a locked room that i cant get into for a few days#asahi-dev · 2026-06-30 08:14
StanfordAppliedCyber[m]unfortunately i'm actually not in person and not able to touch the neo for a few days lol#asahi-dev · 2026-06-30 08:13
StanfordAppliedCyber[m]because the local policy is causing an sep panic#asahi-dev · 2026-06-30 08:13
StanfordAppliedCyber[m]i'll probably create a new volume but yeah that's likely the sanest idea#asahi-dev · 2026-06-30 08:13
StanfordAppliedCyber[m]yes i had a 26.5 m1n1 install before#asahi-dev · 2026-06-30 08:12
StanfordAppliedCyber[m]i believe it's one container, two volumes, and the boot picker screen shows two options#asahi-dev · 2026-06-30 08:12
StanfordAppliedCyber[m]admittedly i'm not sure what's up with the whole container vs volume thing#asahi-dev · 2026-06-30 08:11
StanfordAppliedCyber[m]yes#asahi-dev · 2026-06-30 08:11
StanfordAppliedCyber[m]and there's my custom-macos volume, which is where i installed 26.6 beta 2, downgraded the security, and installed m1n1#asahi-dev · 2026-06-30 08:11
StanfordAppliedCyber[m]yes, there's the macintosh hd volume which is a fully secure macos container#asahi-dev · 2026-06-30 08:10
StanfordAppliedCyber[m]and it's handing me an 26.5 adt which i have to take values from and patch into the 26.6 beta 2 adt from the ipsw#asahi-dev · 2026-06-30 08:10
StanfordAppliedCyber[m]but for some reason my local policy says 26.5 when i installed m1n1?#asahi-dev · 2026-06-30 08:10
StanfordAppliedCyber[m]weird issue-- i installed 26.6 beta 2 into some volume, booted into that volume, and it boots successfully sand says 26.6 beta 2#asahi-dev · 2026-06-25 06:44
StanfordAppliedCyber[m]what i don't get is why reverting to a known good snapshot and running an exact test that worked 5 days ago oesn't work#asahi-dev · 2026-06-25 06:44
StanfordAppliedCyber[m]endpoint 0, tag 8, type 0x04, command 0x08, param 0x4000 (presumably the size of the ool buffers)#asahi-dev · 2026-06-25 06:39
StanfordAppliedCyber[m]the SEP is trying to configure the OOL buffers#asahi-dev · 2026-06-25 06:36
StanfordAppliedCyber[m]right now a command on endpoint 8 is what's actually hanging#asahi-dev · 2026-06-25 06:34
StanfordAppliedCyber[m]does anyone know what preconnditions are required for sep endpoing 18 to be published?#asahi-dev · 2026-06-25 06:33
StanfordAppliedCyber[m]i have been unable to reprooduce the run since then#asahi-dev · 2026-06-25 06:33
StanfordAppliedCyber[m]with one notable exception: earlier today i had a single run that made it past the sep issue to a known later boot wall (after dcp takes over the display)#asahi-dev · 2026-06-25 06:33
StanfordAppliedCyber[m]but in the past 2 days i have: rolled back the 1tr object to the known good one, rolled back my branch to a known good commit, and tested aknown good hypervisor with a known good 1tr boot object and i can't get past the sep failure#asahi-dev · 2026-06-25 06:32
StanfordAppliedCyber[m]so the obvious culprit is the new 1tr object#asahi-dev · 2026-06-25 06:31
StanfordAppliedCyber[m]the regression started after i installed a new 1tr object that changed the memory map#asahi-dev · 2026-06-25 06:31
StanfordAppliedCyber[m]tl;dr my boot, which was previously all the way to windowserver taking over the display, now hangs on the sks endpoint never being enabled#asahi-dev · 2026-06-25 06:30
StanfordAppliedCyber[m]i am incredibly confused about some regression i'm facing right now and i'm wondering if anyone can comment#asahi-dev · 2026-06-24 02:34
StanfordAppliedCyber[m](this is in contrast to sptm btw, which has lots of both)