Re: Separating Buffer LWlocks

On 2015-09-08 14:31:53 -0400, Robert Haas wrote:
> On Tue, Sep 8, 2015 at 2:19 PM, Andres Freund <andres@anarazel.de> wrote:
> > On 2015-09-08 14:15:32 -0400, Robert Haas wrote:
> >> We could do that, but I'm not sure just calling LWLockNewTrancheId()
> >> for all of the tranches would be so bad either.
> >
> > To me that seems either fragile or annoying to use. If all backends call
> > LWLockNewTrancheId() we need to a be sure the callbacks are always going
> > to be called in the same order.
> 
> How is that going to work?  The counter is in shared memory.

Oh, right. It's just the tranche definitions themselves that live in
backend private memory.

> > Otherwise everyone needs to store the
> > tranche in shared memory (like xlog.c now does) which I find to be a
> > rather annoying requirement.
> 
> Yes, everyone would need to do that.  If that's too annoying to live
> with, then we can adopt your suggestion.

I think having to add a separate ShmemInitStruct() just to be able to
store the tranche id falls under too annoying.

Greetings,

Andres Freund