#asahi-dev · 2026-11-03 19:42
jmax[m]So cool! And a changelog entry!#asahi-dev · 2026-10-29 05:39
jmax[m]I thought there's been a kernel patch for this in asahi-wip (and upstream?) for a long time now#asahi-dev · 2026-10-13 10:19
jmax[m]Looks so professional and well put together!#asahi-dev · 2026-10-09 07:40
jmax[m]Lol#asahi-dev · 2026-10-09 07:40
jmax[m]Ah cool cool#asahi-dev · 2026-10-09 07:39
jmax[m]James Calligeros: curious if dxvk works under muvm for the same setup or not. Would tell us if the crashes when dxvk is used are p3s related or just wow64 as you mentioned #asahi-dev · 2026-09-26 18:20
jmax[m]woohoo! very cool! the cursed m1n1 on m4 screenshot seals the deal!#asahi-dev · 2026-08-31 09:37
jmax[m]Probably not worth a v2 :D#asahi-dev · 2026-08-31 09:35
jmax[m]nit: 's/Unlinke/Unlike/g' in the patch#asahi-dev · 2026-07-27 09:09
jmax[m]🎉📸🎥#asahi-dev · 2026-07-26 03:06
jmax[m]I'm#asahi-dev · 2026-07-26 03:05
jmax[m]🎉 to dcp on 14.8.3 fw! This is on M2 currently?#asahi-dev · 2026-07-25 19:54
jmax[m](checks time in Europe. almost 10pm.. oof!)#asahi-dev · 2026-07-23 20:30
jmax[m]Makes sense. Didn't know we could run SPTM in raw object mode. Thanks for the explanations!#asahi-dev · 2026-07-23 20:25
jmax[m]Ah! And darts are managed directly by XNU (via mmio based driving and configuration of the darts) and not by SPTM?#asahi-dev · 2026-07-23 20:23
jmax[m]CPU level page table mappings (since we can intercept CPU EL2 faults) how do we use this approach for device mappings (like GPU) when we can't trap device faults gracefully? I assume XNU uses SPTM to map both CPU pages and Device pages. Hopefully my question is not complete nonsense lol#asahi-dev · 2026-07-23 20:23
jmax[m]This may be a dumb question (I'm only understanding like 10% of what is being discussed here about sprr shadow pages). Is this to allow for running XNU under m1n1 in raw object mode? And this effort is to make it so we can teach m1n1 which page tables XNU needs so that m1n1 can map them for XNU (instead of SPTM doing it)? And we do that through brute force by mapping the tables only when XNU traps when trying to access them? If this handles#asahi-dev · 2026-06-17 17:46
jmax[m]Yeah I think the best first step would be to create the absolutely smallest repro (a super localized diff, for example a single byte change in a single instruction, which may affect it), since it seems to be super sensitive to any kind of change to code structure or data access patterns. But yeah I'd imagine things going into weird states as cores race to enable the MMU and their caches.#asahi-dev · 2026-06-17 17:46
jmax[m]Maybe it's one of these messy things related to multiple layers of caches? Like a core speculating and bringing the data into its own _cluster's_ L2 then another core in the same cluster reading the stale value?#asahi-dev · 2026-06-17 17:46
jmax[m]I remember you running into this a long while ago. Was it happening on the same SoC you're hitting it on now or is it still the same SoC as before?#asahi-dev · 2026-06-17 16:30
jmax[m]Is the sensitivity (crashing vs not) mostly related to code layout/memory addresses? Or is it related to latencies? (I.e. if you only change a loop counter in some delay function, does the issue go away)?#asahi-dev · 2026-06-17 16:30
jmax[m]Thinking if it's only related to code layout/vars/stack use and not timing, you could just change the MMU mappings for all regions to be Device to disable speculation and see if the issue goes away (as an experiment)#asahi-dev · 2026-06-17 16:30
jmax[m]A shot in the dark, Can we use percpu stacks and globals? Or is that already the case but we have instances where cores share pointers to date in their own stacks? Or do we not even do that and we suspect cores are snooping each other's caches through speculation?