On Tue, Oct 03, 2017 at 11:09:08PM +0530, Gargi Sharma wrote:
> On Tue, Oct 3, 2017 at 9:45 PM, Oleg Nesterov <o...@redhat.com> wrote:
> > On 10/02, Andrew Morton wrote:
> >>
> >> From: Alexey Dobriyan <adobri...@gmail.com>
> >> Subject: pid: delete struct pidmap::nr_free
> >>
> >> There is a check in pid allocation code to skip a full page:
> >>
> >> if (likely(atomic_read(&map->nr_free))) {
> >> ...
> >>
> >> In practice it doesn't do anything. To skip a pidmap page one has to have
> >> 32K consecutive pids allocated at the same time which doesn't happen.
> >>
> >> Currently the price is _every_ fork/exit on every system being slower than
> >> necessary.
> >
> > Agreed, I too never understood how can this counter help.
> >
> > Add Gargi and Rik, the next version of "Replace PID bitmap allocation with
> > IDR
> > API" can conflict with this and the previous pid-delete-reserved_pids.patch.
>
> I think this patch will become obsolete as pidmap will be removed. As
> for the 1/2
> patch of Alexey's series, I'll incorporate it so that rolled over PIDs
^^^^^^^^^^^
> start from 1 instead of RESERVED_PIDS.
I saw IDR pids, lots of code removed indeed.
Incorporation is not really necessary as RESERVED_PIDS is an independent change.
Just send it as 2/2 or whatever the number or rebase on top of -mm.
