We have this weird bug in m1n1 that I first hit about a month ago. It’s likely a race and/or cache issue somewhere, only happens when running something under the hypervisor (normal Linux boots aren’t affected) and there a various ways to work around it (e.g. compile with gcc instead of clang 
)
I *thought* we figured out what’s going on yesterday night: core starts with the wrong stack and overwrites the stack of another core which then crashes when trying to jump to a pointer from the stack that was overwritten. The symptoms (and even the pointer it tries to jump to and the oberwritten stack) perfectly fit except that I looked at the SPs of all cores and they all look reasonable 
GitHubinstruction abort with FAR: 0x2000000 usually inside mmu_setup_secondary · Issue #463 · AsahiLinux/m1n1
200-300fps 
Oh noes! 
Meanwhile, the rest of us are still trying to remember what Asahi Linux even is. 
Asahi Linux bravely attempts to tackle the insurmountable task of writing a Linux progress report that's even more thrilling than watching paint dry 
. With the riveting excitement of "administrative work" and "upstream patch submission," this is sure to be a blockbuster hit in the land of sleep 
.
