#asahi-dev · 2026-04-14 14:53
tryhard[m]latest version via apt#asahi-dev · 2026-04-14 14:53
tryhard[m]that's the latest ubuntu version#asahi-dev · 2026-04-14 14:53
jannaubut 1.25 should be recent enough: PAHOLE_VERSION >= 124#asahi-dev · 2026-04-14 14:52
jannauv1.30 here#asahi-dev · 2026-04-14 14:51
jannauthe full condition for CONFIG RUST is `(!DEBUG_INFO_BTF [=n] || PAHOLE_HAS_LANG_EXCLUDE [=y] && !LTO [=n])`#asahi-dev · 2026-04-14 14:51
tryhard[m]1.25#asahi-dev · 2026-04-14 14:50
jannauwhich pahole version do you have?#asahi-dev · 2026-04-14 14:48
tryhard[m]`&& (!DEBUG_INFO_BTF [=y]`#asahi-dev · 2026-04-14 14:48
chaos_princessi personally build with LLVM=1, maybe that is relevant?#asahi-dev · 2026-04-14 14:47
tryhard[m]yeah#asahi-dev · 2026-04-14 14:45
chaos_princessdoes rust depend on it or sth?#asahi-dev · 2026-04-14 14:44
tryhard[m]i did that but then it disables my pmp#asahi-dev · 2026-04-14 14:43
jannaudisable DEBUG_INFO_BTF#asahi-dev · 2026-04-14 14:43
chaos_princessi just no longer remember what exactly#asahi-dev · 2026-04-14 14:43
chaos_princessif you turn some of those off, it should build#asahi-dev · 2026-04-14 14:42
chaos_princessgo into menuconfig and search for BTF, or UNWIND#asahi-dev · 2026-04-14 14:42
tryhard[m]i don't understand#asahi-dev · 2026-04-14 14:41
chaos_princessi think its something related to unwinding#asahi-dev · 2026-04-14 14:40
tryhard[m]sent a code block: https://matrix.org/oftc/media/v1/media/download/AYdF-MbHQG-bQYWyp3l8CG3M3SI97D-jHt0ef5qGx4rr-LXM59-TvS9QEONCMtGAI1gcRS6BagyRHwCyVNU7eFNCed1IjOaQAG1hdHJpeC5vcmcvWGtPVG1qSmZhRGpQdWNwb0xFV1VuVUZl#asahi-dev · 2026-04-14 14:40
svenboot cpu vs. secondaries also shouldn't make a difference, so diffing sysregs like chaos_princess suggested is also a good idea#asahi-dev · 2026-04-14 14:39
svenoh yes, absolutely#asahi-dev · 2026-04-14 14:39
svenotherwise i guess we'll have to add a quirk and patch out all wfis very early in the boot sequence :/#asahi-dev · 2026-04-14 14:39
yukaWe should first confirm it exactly by doing wfi in m1n1 and testing the register state afterwards#asahi-dev · 2026-04-14 14:38
svenso if it does that on m4 i'll file a cheeky radar and see what happens :D#asahi-dev · 2026-04-14 14:38
svenyeah, that's my reading as well#asahi-dev · 2026-04-14 14:38
jannau"If the system is configured such that the WFI or WFIT instruction can be completed, then the architecture requires that the architectural state is not lost." means it can't lose some state but keep PC#asahi-dev · 2026-04-14 14:38
sven*lose#asahi-dev · 2026-04-14 14:38
sven"If the system is configured such that the WFI or WFIT instruction can be completed, then the architecture requires that the architectural state is not lost." so that very much sounds to me like wfi isn't allowed to use state if the expectation is that execution will resume afterwards at next next instruction#asahi-dev · 2026-04-14 14:36
j`eyRHJPJY above it also mentions state#asahi-dev · 2026-04-14 14:36
sven\o/#asahi-dev · 2026-04-14 14:36
yukaI have all cores up :)#asahi-dev · 2026-04-14 14:35
sventhat doesn't quite sound like normal wfi to me#asahi-dev · 2026-04-14 14:35
svenhave been saved by software as part of the powerdown sequence before the WFI." that's the full quote#asahi-dev · 2026-04-14 14:35
sven"In some implementations, based on the configuration of system specific registers, WFI can be used as part of a powerdown sequence where no interrupts will cause WFI wakeup events, and restoration of power involves resetting of the PE. In those cases, the WFI is permitted to cause a loss of architectural state, as it is assumed that this state will#asahi-dev · 2026-04-14 14:35
svenhmmm true#asahi-dev · 2026-04-14 14:35
j`eybut Im not sure if 'powerdown sequence' is some other documented thing#asahi-dev · 2026-04-14 14:35
svenalso surprised how fast you found that! i was looking as well#asahi-dev · 2026-04-14 14:35
yukaSo I guess just nobody else implemented this?#asahi-dev · 2026-04-14 14:34
svenoh well#asahi-dev · 2026-04-14 14:34
sven" In those cases, the WFI is permitted to cause a loss of architectural state" :(#asahi-dev · 2026-04-14 14:34
j`eysven: https://developer.arm.com/documentation/ddi0487/maa/-Part-D-The-AArch64-System-Level-Architecture/-Chapter-D1-The-AArch64-System-Level-Programmers--Model/-D1-7-Mechanisms-for-entering-a-low-power-state/-D1-7-2-Wait-for-Interrupt-mechanism wwyxs#asahi-dev · 2026-04-14 14:30
sveni mean something along the lines of "so you took away our toys (SPRR/GXF) to make this thing look like an architectural ARM machine, how about you also make wfi work correctly?" ;)#asahi-dev · 2026-04-14 14:29
yukaDo they claim that this is ARM ISA compliant anywhere?#asahi-dev · 2026-04-14 14:28
sven*allowed#asahi-dev · 2026-04-14 14:28
svenis wfi allows to lose state according to the ARM ARM? if that's not allowed and we can confirm it does on m4 i'm gonna file a cheeky feedback assistant report / radar :D#asahi-dev · 2026-04-14 14:24
yukabut yeah maybe there is some new msr controlling this#asahi-dev · 2026-04-14 14:24
yukaI checked the CYC_OVRD and it's 0 in all cpus and it can't be set#asahi-dev · 2026-04-14 14:24
chaos_princessdon't we have a msr scanner? maybe diff all the things?#asahi-dev · 2026-04-14 14:24
chaos_princessboot cpu is not supposed to be special, so that is probably a difference in some msr#asahi-dev · 2026-04-14 14:22
yukaor, like, always _may_ lose register state, on the boot cpu it doesn't