On Tue, Oct 14, 2014 at 11:31 AM, Andres Freund <andres@2ndquadrant.com> wrote:
>> It doesn't look particularly dangerous to me. Famous last words.
>
>> Basically, I think what we're doing right now is holding the buffer
>> mapping lock so that the buffer can't be renamed under us while we're
>> pinning it.
>
> What I'm afraid of is that there's hidden assumptions about the
> consistency provided by the mapping locks.
That's certainly worth checking for, but I think the only code that
needs to be checked is the code that would formerly have run while
holding said lock. And there isn't that much of that.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company