On Sun, 3 Apr 2016, Philip Guenther wrote: > Thinking about this more, I believe these savectx() calls date from when > cpu_fork() worked completely differently, with the child actually > returning back through cpu_fork(). Now, the child's stack frame is > updated to make it start in proc_trampoline() when it's first switched > to, so the registers that savectx() saves (stack/frame pointers, maybe > callee-saved registers) are useless to the child. > > On that theory, I'm going to test the mips64 part of this diff. I don't > have arm or m88k to test; anyone?
loongson appears to be happy with this, completing a build and release, and Patrick says the arm version seems happy. Any oks? Philip
