From: Tom Lane <tgl@sss.pgh.pa.us>
To: Robert Henry <rrh.henry@gmail.com>
cc: pgsql-bugs@lists.postgresql.org
Subject: Re: postgres has no spinlock support on riscv rv64imafdc
In-reply-to: 
 <CAEYr_8nnwxQji0pQJYqFx7=gSM8x7jFJi16O-bLuVYvqvkjsLg@mail.gmail.com>
References: 
 <CAEYr_8nnwxQji0pQJYqFx7=gSM8x7jFJi16O-bLuVYvqvkjsLg@mail.gmail.com>
Comments: In-reply-to Robert Henry <rrh.henry@gmail.com>
	message dated "Thu, 17 Oct 2019 12:08:58 -0700"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <27500.1571382023.1@sss.pgh.pa.us>
Date: Fri, 18 Oct 2019 09:00:23 +0200
Message-ID: <27501.1571382023@sss.pgh.pa.us>
Precedence: bulk

Robert Henry <rrh.henry@gmail.com> writes:
> Building Postgres from scratch on the riscv board, no fancy cross
> compilation, I get:

> PostgreSQL does not have native spinlock support on this platform.  To
> continue the compilation, rerun configure using --disable-spinlocks.
> However, performance will be poor.  Please report this to
> pgsql-bugs@lists.postgresql.org.

The last we heard about RISC-V was this thread:

https://www.postgresql.org/message-id/flat/20161119230312.GA8656%40redhat.com

If things seem to work well now with --disable-spinlocks, maybe you
could test the patch suggested in that thread.  TBH, though, my preference
would be for some assembly code rather than relying on GCC builtins
as Richard's patch did.

			regards, tom lane