On Wed, Nov 23, 2016 at 7:07 AM, Thomas Munro
<thomas.munro@enterprisedb.com> wrote:
> Those let you create an area in existing memory (in a DSM segment,
> traditional inherited shmem). The in-place versions will stlll create
> DSM segments on demand as required, though I suppose if you wanted to
> prevent that you could with dsa_set_size_limit(area, size). One
> complication is that of course the automatic detach feature doesn't
> work if you're in some random piece of memory. I have exposed
> dsa_on_dsm_detach, so that there is a way to hook it up to the detach
> hook for a pre-existing DSM segment, but that's the caller's
> responibility.
shm_mq_attach() made the opposite decision about how to solve this
problem, and frankly I think that API is a lot more convenient: if the
first argument to shm_mq_attach() happens to be located inside of a
DSM, you can pass the DSM as the second argument and it registers the
on_dsm_detach() hook for you. If not, you can pass NULL and deal with
it in some other way. But this makes the common case very simple.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company