Feedback-ID: id4a34324:Fastmail
Date: Tue, 18 Mar 2025 11:14:12 -0400
From: Andres Freund <andres@anarazel.de>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-committers@lists.postgresql.org
Subject: Re: pgsql: aio: Add core asynchronous I/O infrastructure
Message-ID: <ecqk4e772iqtwuwfu56yr3cv5wnzug6fpqvmwjcwlduqlom5xr@2fzz3cdycdyb>
References: <E1tuJo6-0030Bd-0C@gemulon.postgresql.org>
 <93029.1742310216@sss.pgh.pa.us>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <93029.1742310216@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/ecqk4e772iqtwuwfu56yr3cv5wnzug6fpqvmwjcwlduqlom5xr%402fzz3cdycdyb>
Precedence: bulk

Hi,

On 2025-03-18 11:03:36 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > aio: Add core asynchronous I/O infrastructure
> 
> Some of the buildfarm is mildly unhappy with this.
> So far I see
> 
>  ayu           | 2025-03-18 13:08:04 | aio_callback.c:83:12: warning: comparison of constant 1 with expression of type 'PgAioHandleCallbackID' (aka 'enum PgAioHandleCallbackID') is always false [-Wtautological-constant-out-of-range-compare]
>  ayu           | 2025-03-18 13:08:04 | aio_callback.c:190:56: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]
>  ayu           | 2025-03-18 13:08:04 | aio_callback.c:220:56: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]
>  ayu           | 2025-03-18 13:08:04 | aio_callback.c:274:56: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]
>  ayu           | 2025-03-18 13:08:04 | aio_target.c:50:41: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]
>  ayu           | 2025-03-18 13:08:04 | aio_target.c:110:41: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]
>  demoiselle    | 2025-03-18 03:43:46 | aio_callback.c:83:12: warning: comparison of constant 1 with expression of type 'PgAioHandleCallbackID' (aka 'enum PgAioHandleCallbackID') is always false [-Wtautological-constant-out-of-range-compare]
>  demoiselle    | 2025-03-18 03:43:46 | aio_callback.c:190:56: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]
>  demoiselle    | 2025-03-18 03:43:46 | aio_callback.c:220:56: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]
>  demoiselle    | 2025-03-18 03:43:46 | aio_callback.c:274:56: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]
>  demoiselle    | 2025-03-18 03:43:46 | aio_target.c:50:41: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]
>  demoiselle    | 2025-03-18 03:43:46 | aio_target.c:110:41: warning: comparison of constant 1 with expression of type 'PgAioTargetID' (aka 'enum PgAioTargetID') is always true [-Wtautological-constant-out-of-range-compare]

Thanks for noticing that.


> I don't have that compiler handy to check, but maybe changes like
> 
> -	if (cb_id >= lengthof(aio_handle_cbs))
> +	if ((size_t) cb_id >= lengthof(aio_handle_cbs))
> 
> would silence these?

I wonder if we should instead either ask those buildfarm animals to be
disabled or have the warning manually disabled. I don't think it's a good
investment on our part to work towards warning cleanliness on clang 4 and 5,
on distros from 2017 and and 2018 respectively.


> Or you could just drop all those tests and have faith in the type-safety of
> the enums.

The issue isn't really the type-safety of the enums (not that I would have a
lot of trust in that), it's whether the relevant callback/target actually have
been added to the the callback/target arrays.

Greetings,

Andres Freund