#
✕ nick:sven
#asahi-dev · 2026-07-08 11:50
sveni expected that to crash
#asahi-dev · 2026-07-08 11:50
svenyup, same :D
#asahi-dev · 2026-07-08 11:49
sveni expect that to be only useful for directly chainloading to the kernel fwiw, u-boot (or whatever) probably needs a different protocol to essentially pass along that new table i made up + the runtime mappings we need
#asahi-dev · 2026-07-08 11:45
svenpretty sure it violates the spec though
#asahi-dev · 2026-07-08 11:45
svenyes, it's the bare minimum to make it happy
#asahi-dev · 2026-07-08 09:01
svenso 10 iphone 8 users then? :D
#asahi-dev · 2026-07-08 08:57
sven:D
#asahi-dev · 2026-07-08 08:49
svencitation required
#asahi-dev · 2026-07-08 08:49
svenif we could at least file bugs easily
#asahi-dev · 2026-07-08 08:49
svenyeah :(
#asahi-dev · 2026-07-08 08:47
sveni get that they had to disable GXF/SPRR and whatever but did they really have to also break all kinds of other random crap and just not test this mode? :(
#asahi-dev · 2026-07-08 08:46
sven:(
#asahi-dev · 2026-07-08 07:48
sven:)
#asahi-dev · 2026-07-08 07:31
svenit "just" needs to pass along the runtime memory descriptors and that additional table
#asahi-dev · 2026-07-08 06:02
svennice!
#asahi-dev · 2026-07-07 21:03
svenwe still don’t know how to wake up from when all cores are off
#asahi-dev · 2026-07-07 21:03
svenright now that psci code just does the same as our downstream cpuidle driver
#asahi-dev · 2026-07-07 20:52
svenbut it shouldn’t be *much* worse
#asahi-dev · 2026-07-07 20:52
svenhm, I’d expect psci-efi to be a bit worse because it has to switch pagetables unless I messed up somewhere
#asahi-dev · 2026-07-07 20:45
svenand I guess m4+ would only have deep wfi
#asahi-dev · 2026-07-07 20:45
sveni think I took some from the cpuidle driver but I’d just treat them as math.randint for now
#asahi-dev · 2026-07-07 20:44
svenoh, I made those latencies up
#asahi-dev · 2026-07-07 18:37
svenand roughly half of that is probably documentaiton
#asahi-dev · 2026-07-07 18:37
sven8 files changed, 261 insertions(+), 10 deletions(-) <-- i'm still surprised how small that diffstat ended up
#asahi-dev · 2026-07-07 18:35
svenhttps://github.com/AsahiLinux/linux/commits/efi-psci/ and that's what i'll send as a RFC in a few days
#asahi-dev · 2026-07-07 18:21
sveni adjusted the ld script and as long as m1n1 is loaded with 64k alignment everything should work I think
#asahi-dev · 2026-07-07 17:35
sven:D
#asahi-dev · 2026-07-07 17:32
svenif it breaks you just get to fix it yourself ;)
#asahi-dev · 2026-07-07 17:31
svenas long as it doesn’t include SPTM (26+ iirc) it might work
#asahi-dev · 2026-07-07 17:31
svenjust try it and see what happens
#asahi-dev · 2026-07-07 16:45
svenhttps://github.com/AsahiLinux/m1n1/pull/618 and the m1n1 side, not interested in detailed reviews yet
#asahi-dev · 2026-07-07 16:31
svenlet me try to complain again
#asahi-dev · 2026-07-07 16:31
sven:(
#asahi-dev · 2026-07-07 08:08
svenyuka: yeah, that could've been it then. it only reproduces when all the stars align which kinda fits to my memory ordering / stale cache line from speculation theory
#asahi-dev · 2026-07-07 07:35
sven this issue was a weird crash when enabling the MMU on secondary cores
#asahi-dev · 2026-07-07 07:34
sveni think that was unrelated
#asahi-dev · 2026-07-07 07:13
svenactually, I’ve spent enough hours debugging this crap. let me just merge it and be done with it :D
#asahi-dev · 2026-07-07 07:10
svenack
#asahi-dev · 2026-07-07 06:38
svenI think I’ll just reduce it to dsb sy and leave a comment about the cache thing if this happens again
#asahi-dev · 2026-07-07 06:33
sventhe kernel’s head.S also does dc ivac with mmu off
#asahi-dev · 2026-07-07 06:05
svenI think it’s really just the dsb sy that‘s enough to fix this
#asahi-dev · 2026-07-07 06:04
sventested stack invalidation before and after MMU setup and unsurprisingly both work
#asahi-dev · 2026-07-06 20:19
svennice!
#asahi-dev · 2026-07-06 19:09
svenanyway, i'll leave it as-is for now
#asahi-dev · 2026-07-06 19:08
svenyeah, probably just s/16k/64k/ in the ld script
#asahi-dev · 2026-07-06 19:05
svenjust need some proof of concept for the RFC series anyway
#asahi-dev · 2026-07-06 19:05
sveni'll leave it as-is for now
#asahi-dev · 2026-07-06 19:05
sventhe end of .text / beginning of .data is only 16 KiB aligned and what i care about here
#asahi-dev · 2026-07-06 19:04
svenah, good
#asahi-dev · 2026-07-06 19:04
svengetting all this into u-boot or other 2nd stage loaders is going to be more fun

plus ›