Re: Reworks of CustomScan serialization/deserialization

From: Kouhei Kaigai <kaigai(at)ak(dot)jp(dot)nec(dot)com>
To: Robert Haas <robertmhaas(at)gmail(dot)com>
Cc: Petr Jelinek <petr(at)2ndquadrant(dot)com>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Reworks of CustomScan serialization/deserialization
Date: 2016-03-29 03:00:36
Message-ID: 9A28C8860F777E439AA12E8AEA7694F8011D7925@BPXM15GP.gisp.nec.co.jp
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

> -----Original Message-----
> From: pgsql-hackers-owner(at)postgresql(dot)org
> [mailto:pgsql-hackers-owner(at)postgresql(dot)org] On Behalf Of Robert Haas
> Sent: Tuesday, March 29, 2016 10:54 AM
> To: Kaigai Kouhei(海外 浩平)
> Cc: Petr Jelinek; pgsql-hackers(at)postgresql(dot)org
> Subject: Re: [HACKERS] Reworks of CustomScan serialization/deserialization
>
> On Mon, Mar 28, 2016 at 9:36 PM, Kouhei Kaigai <kaigai(at)ak(dot)jp(dot)nec(dot)com> wrote:
> > I don't have a strong reason to keep these stuff in separate files.
> > Both stuffs covers similar features and amount of code are enough small.
> > So, the attached v4 just merged custom-node.[ch] stuff into extensible.
> >
> > Once we put similar routines closely, it may be better to consolidate
> > these routines.
> > As long as EXTNODENAME_MAX_LEN == CUSTOM_NAME_MAX_LEN, both features
> > have identical structure layout, so it is easy to call an internal
> > common function to register or find out a table of callbacks according
> > to the function actually called by other modules.
> >
> > I'm inclined to think to replace EXTNODENAME_MAX_LEN and
> > CUSTOM_NAME_MAX_LEN by NAMEDATALEN again, to fit structure layout.
>
> I don't think that we need both EXTNODENAME_MAX_LEN and
> CUSTOM_NAME_MAX_LEN; we can use EXTNODENAME_MAX_LEN for both. I'm
> opposed to using NAMEDATALEN for anything unrelated to the size of a
> Name. If it's not being stored in a catalog, it doesn't need to care.
>
OK, I adjusted the v4 patch to use EXTNODENAME_MAX_LEN for both.

The structure of hash entry was revised as follows, then registered via
an internal common function: RegisterExtensibleNodeEntry, and found out
via also an internal common function: GetExtensibleNodeEntry.

typedef struct
{
char extnodename[EXTNODENAME_MAX_LEN];
const void *extnodemethods;
} ExtensibleNodeEntry;

ExtensibleNodeMethods and CustomScanMethods shall be stored in
'extensible_node_methods' and 'custom_scan_methods' separatedly.
The entrypoint functions calls above internal common functions with
appropriate HTAB variable.

It will be re-usable if we would have further extensible nodes in the
future versions.

Thanks,
--
NEC Business Creation Division / PG-Strom Project
KaiGai Kohei <kaigai(at)ak(dot)jp(dot)nec(dot)com>

Attachment Content-Type Size
pgsql-v9.6-custom-scan-serialization-reworks.5.patch application/octet-stream 15.7 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Andrew Dunstan 2016-03-29 03:12:05 Re: raw output from copy
Previous Message Robert Haas 2016-03-29 01:59:58 Re: Relation extension scalability