Re: [PATCH] Segfault in pool_do_auth() when failover races with a new connection

Hi Tatsuo,

Yes, this looks good. There's one small nit: there's some whitespace
noise in the patch (space before tab) on the line with
"my_main_node_id = *(volatile int *) &REAL_MAIN_NODE_ID;"

Best regards,
Emond

Op do 2 jul 2026 om 15:45 schreef Tatsuo Ishii <ishii(at)postgresql(dot)org>:
>
> Hi Emond,
>
> > The new patch looks fine to me. There's however one thing Claude
> > pointed out. According to it, my_main_node_id,
> > private_backend_status[], and my_backend_status[] are per-process
> > globals (copied on fork), not shared memory. So, making any of those
> > volatile should not make any difference.
>
> You are right. Attached is the v3 patch.
>
> Regards,
> --
> Tatsuo Ishii
> SRA OSS K.K.
> English: http://www.sraoss.co.jp/index_en/
> Japanese:http://www.sraoss.co.jp