Is LKL usage of setjmp correct?

[liuyuan10]

As far as I know, setjmp must be used inline so the the stack is still alive when long jmp happens. "And some esoteric tasks (generally requiring the use of assembly code, such as setjmp and certain kinds of locking) must be performed inline or they won't work right."

But the way we use setjmp in LKL is not inlined so I'm worried this may be cause of segment fault I observed (at very low frequency). I can't explain why it works though.

Also I don't find a way to inline the setjmp into LKL kernel code.

[tavip]

Yes, your are right. In the initial implementation I used a different host ops API for setjmp to avoid this issue, but during cleanups I've forgot why I did it that way and scraped it :( I will post a PR to fix this issue soon.