> the reason fileassoc lacked locking primitives is that I waited for
> ad@ to commit and stablize newlock2. I *discussed* this pr with ad@
> and yamt@, and, as a matter of fact, *prior* to newlock2, in a biglock
> kernel, it is impossible to trigger the NULL deref.

i've explained you that the PR is not (necessarily) about MP.
did you understand it?

(i'm not talking about this particular instance of NULL dereference.
i'm not interested in analyzing how the known to be broken code can behave.
i merely want to make sure that you know about the problem.  if you already
know, it's fine.)

YAMAMOTO Takashi