RE: AIX support

From: Srirama Kucherlapati <sriram(dot)rk(at)in(dot)ibm(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Peter Eisentraut <peter(at)eisentraut(dot)org>, "pgsql-hackers(at)lists(dot)postgresql(dot)org" <pgsql-hackers(at)lists(dot)postgresql(dot)org>, Heikki Linnakangas <hlinnaka(at)iki(dot)fi>, Tristan Partin <tristan(at)partin(dot)io>, AIX PG user <postgres-ibm-aix(at)wwpdl(dot)vnet(dot)ibm(dot)com>, Aditya Kamath <Aditya(dot)Kamath1(at)ibm(dot)com>
Subject: RE: AIX support
Date: 2026-01-29 16:01:38
Message-ID: SJ4PPFB817783268EB83C0384D7652748C1DB9EA@SJ4PPFB81778326.namprd15.prod.outlook.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi Tom,

Thank you for taking the time to review the patch and share your feedback.

We are currently working on the comments related to performing a regular build

without relying on environment variables, as well as updating the
documentation. At the moment, we are encountering an issue where some of the
flags defined in src/template/aix are not being propagated into the Makefiles
in the subdirectories. Additionally, we are seeing linker problems due to
OBJECT_MODE and LDFLAGS not being exported. We are working
through these issues and will provide an update once they are resolved.

>So one TODO item for your patch is to restore (and update)
>the AIX-specific portions of the documentation, particularly
>the build instructions.

We are updating the AIX‑specific documentation accordingly. Would you prefer these details to be added in configure or included in installation.sgml?

>After re-reading what used to be in installation.sgml, I'm inclined
>to suggest that we just say that only 64-bit builds using gcc are
>supported on AIX. Surely the number of users still wanting 32-bit
>Postgres on AIX is indistinguishable from zero.

We are now building only 64‑bit on AIX by default, aligning both OBJECT_MODE and the -maix64 flag in configure as well as in src/template/aix.

> Were you perhaps building with --disable-rpath?
I haven’t used this flag.
> (It also said to set LDFLAGS="-Wl,-bbigtoc", but that doesn’t
> seem to be necessary anymore?)
I think we had to use this when additional packages like icu, readline, zlib were included.

> However, my build doesn't pass the core regression tests. There
> is one failure in float8.out, which on investigation is because
>lgamma(NaN) is failing (returning ERANGE) instead of silently
>returning NaN.

This is a known issue and our AIX team had fixed it, I have requested for a iFix. I’ll provide. It as soon as its available.

>so I don't know why it doesn't like __int128. This is very new code,
>so there may be a non-AIX-specific issue for us to fix.
This seems to be supported on 64Bit only, when I tried to build with -maxi64 it got build.

>Another thing I noticed: ps status isn't working. (All the postmaster
>child processes look the same in "ps auxww".) I suppose you need to
>restore the AIX support in ps_status.c.
I have updated this file.

> FWIW, attached is a delta patch showing where I am now (on top of
> your v11 patch and the pgstat_slru.c hack I showed earlier).

Thank you for the patch.

I will keep you updated once all the changes have been finalized.

Thank you once again for your suggestions and feedback.

Warm regards,

Sriram.

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Gilles Darold 2026-01-29 16:07:04 Re: Pasword expiration warning
Previous Message Ashutosh Bapat 2026-01-29 15:41:20 Re: Improvements and refactoring in shmem.c