Re: Add pg_stat_vfdcache view for VFD cache statistics

On Tue, Sep 8, 2026 at 6:33 PM David Geier <geidav(dot)pg(at)gmail(dot)com> wrote:

> >>>> However, I'm wondering if the better approach wouldn't be to change
> fd.c
> >>>> to use a long-lived memory context. Then all bookkeeping would happen
> >>>> automatically and the memory size could simply be reported via
> existing
> >>>> memory context stats infrastructure.
> >>>>
> >>>> Not entirely sure though if there's some roadblock when switching to a
> >>>> memory context.
> >>> I don't see any issue with this either. However, the only benefit we
> >> would
> >>> gain is using existing infrastructure but only for backend vfd cache
> >> memory
> >>> (i.e cache_bytes).
> >>> Everything else stays the same (counters, cluster-wide memory);
> >> therefore,
> >>> if there's no other benefit to replacing with memory contexts, maybe
> it's
> >>> not worth it.
> >>
> >> The biggest benefit in my view is consistency with the rest of
> PostgreSQL.
> >> That is from a usage point of view as well as from a coding point of
> view.
> >> If you want, I can give that a try and share a patch with you if
> >> successful.
> >>
> > Yes of course, I’d be happy to take a look.
> Attached is the patch. It's pretty small and passes regress tests.
>
Nice one.
Few comments on it:
#1:
+ oldcontext = MemoryContextSwitchTo(VfdCxt);
+ newVfdCache = repalloc_array(VfdCache, Vfd, newCacheSize);
+ MemoryContextSwitchTo(oldcontext);

Can't we just do this?
+ newVfdCache = repalloc_array(VfdCache, Vfd, newCacheSize);

because repalloc doesn't need CurrentMemoryContext.

#2:
+ newDescs = MemoryContextAllocExtended(VfdCxt,
+ newMax * sizeof(AllocateDesc), MCXT_ALLOC_NO_OOM);
if (newDescs == NULL)
return false;
+ memcpy(newDescs, allocatedDescs, maxAllocatedDescs *
sizeof(AllocateDesc));
+ pfree(allocatedDescs);

We can also just replace it with:
+ newDescs = repalloc_array_extended(allocatedDescs, AllocateDesc,
+ newMax, MCXT_ALLOC_NO_OOM);
Correct?

Attached are v7 patches containing your patch for memory contexts with
these small suggestions and pg_stat_vfdcache patches like before.

Regards,
Ayoub