svn commit: r304436 - in head: . sys/netinet

[Slawa Olhovchenkov]

On Fri, Aug 26, 2016 at 02:42:27PM -0700, Adrian Chadd wrote:

> On 26 August 2016 at 14:36, Slawa Olhovchenkov <slw at zxy.spb.ru> wrote:
> > On Fri, Aug 26, 2016 at 02:32:00PM -0700, Adrian Chadd wrote:
> >
> >> Hi,
> >>
> >> It's pcb lock contention.
> >
> > Not sure: only 5% of all time.
> > And same 5% for tcbhashsize = 65K and 256K.
> > Or you talk about some more thin effect?
> 
> You're in the inpcb lock from multiple places.

For tcp case I am found only 3 places: in_pcbdrop in_pcbremlists in_pcblookup_hash

> the tcbhashsize doesnt influence the pcb lock contention - it just
> affects how long you take doing lookups. iF your hash table is too
> small then you end up doing lots of O(n) walks of a hash bucket to
> find a pcb entry. :)

Hmm. I am not clearly understund you.
How originate pcb lock contention?
How I can see this (in pmc profile, in dtarce probese)?