Theo de Raadt <dera...@openbsd.org> wrote:
> Stuart Henderson <s...@spacehopper.org> wrote:
>
> > On 2022/05/19 08:54, Theo de Raadt wrote:
> > > I have argued in the past that mlock() in our kernel should probably be
> > > a NOOP, return success all the time, and doing nothing.
> >
> > Would it make any sense to do that rather than abort if pledged
> > (say under stdio)?
> >
>
> Well I am talking about
>
> int
> sys_mlock(...)
> {
> return 0;
> }
>
> and deleting everything related to the concept of "locking physical memory
> resources", because everytime I see mlock() used, it is for the wrong
> reasons.
I mean, step back. Why is this library underneath firefox calling mlock?
The reason why it calls mlock() is not mentioned in the mlock manual page,
not even once.
