Обсуждение: Re: [COMMITTERS] pgsql: Invent a memory context reset/delete callback mechanism.

On Fri, 2015-02-27 at 22:17 +0000, Tom Lane wrote:
> In passing, per discussion, rearrange some boolean fields in struct
> MemoryContextData so as to avoid wasted padding space.  For safety,
> this requires making allowInCritSection's existence unconditional;
> but I think that's a better approach than what was there anyway.

I notice that this uses the bytes in MemoryContextData that I was hoping
to use for the memory accounting patch (for memory-bounded HashAgg).

Leaving no space for even a 32-bit value (without AllocSetContext
growing beyond the magical 192-byte number Andres mentioned) removes
most of the options for memory accounting. The only things I can think
of are:

  1. Move a few less-frequently-accessed fields out to an auxiliary
structure (Tomas proposed something similar); or
  2. Walk the memory contexts, but use some kind of
estimation/extrapolation so that it doesn't need to be done for every
input tuple of HashAgg.

Thoughts?

Regards,
    Jeff Davis

Jeff Davis <pgsql@j-davis.com> writes:
> On Fri, 2015-02-27 at 22:17 +0000, Tom Lane wrote:
>> In passing, per discussion, rearrange some boolean fields in struct
>> MemoryContextData so as to avoid wasted padding space.  For safety,
>> this requires making allowInCritSection's existence unconditional;
>> but I think that's a better approach than what was there anyway.

> I notice that this uses the bytes in MemoryContextData that I was hoping
> to use for the memory accounting patch (for memory-bounded HashAgg).

Meh.  I thought Andres' complaint about that was unfounded anyway ;-).
But I don't really see why a memory accounting patch should be expected to
have zero footprint.
        regards, tom lane

* Tom Lane (tgl@sss.pgh.pa.us) wrote:
> Jeff Davis <pgsql@j-davis.com> writes:
> > On Fri, 2015-02-27 at 22:17 +0000, Tom Lane wrote:
> >> In passing, per discussion, rearrange some boolean fields in struct
> >> MemoryContextData so as to avoid wasted padding space.  For safety,
> >> this requires making allowInCritSection's existence unconditional;
> >> but I think that's a better approach than what was there anyway.
>
> > I notice that this uses the bytes in MemoryContextData that I was hoping
> > to use for the memory accounting patch (for memory-bounded HashAgg).
>
> Meh.  I thought Andres' complaint about that was unfounded anyway ;-).
> But I don't really see why a memory accounting patch should be expected to
> have zero footprint.

For my 2c, I agree that it's a bit ugly to waste space due to padding,
but I'm all about improving the memory accounting and would feel that's
well worth having a slightly larger MemoryContextData.

In other words, I agree with Tom that it doesn't need to have a zero
footprint, but disagree about wasting space due to padding. :D
Thanks!
    Stephen