>>I'd appreciate it if someone more familiar with the MIPS cache code could
>>look at how it interacts with the fstatfs() system call.
>
>Andrew was muttering something about pmap_prefer problems on the mips
>mailing lists. Maybe this has something to do with it?

It'd be a serious resentment if pmap was really the cause.   It's VITA flaw.
I have to switch my brain cell back to NetBSD kernel mode to look the code.

Toru Nishimura/ALKYL Technology