From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <musl-return-21010-ml=inbox.vuxu.org@lists.openwall.com>
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on inbox.vuxu.org
X-Spam-Level: 
X-Spam-Status: No, score=-3.1 required=5.0 tests=HEADER_FROM_DIFFERENT_DOMAINS,
	MAILING_LIST_MULTI,RCVD_IN_DNSWL_MED,RCVD_IN_MSPIKE_H4,
	RCVD_IN_MSPIKE_WL,T_SCC_BODY_TEXT_LINE autolearn=ham
	autolearn_force=no version=3.4.4
Received: from second.openwall.net (second.openwall.net [193.110.157.125])
	by inbox.vuxu.org (Postfix) with SMTP id E14652B63E
	for <ml@inbox.vuxu.org>; Wed, 12 Jun 2024 18:50:20 +0200 (CEST)
Received: (qmail 19748 invoked by uid 550); 12 Jun 2024 16:50:14 -0000
Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm
Precedence: bulk
List-Post: <mailto:musl@lists.openwall.com>
List-Help: <mailto:musl-help@lists.openwall.com>
List-Unsubscribe: <mailto:musl-unsubscribe@lists.openwall.com>
List-Subscribe: <mailto:musl-subscribe@lists.openwall.com>
List-ID: <musl.lists.openwall.com>
Reply-To: musl@lists.openwall.com
Received: (qmail 19710 invoked from network); 12 Jun 2024 16:50:13 -0000
Date: Wed, 12 Jun 2024 18:50:04 +0200
From: Szabolcs Nagy <nsz@port70.net>
To: Meng Zhuo <mzh@mzh.io>
Cc: musl@lists.openwall.com
Message-ID: <20240612165004.GI3766212@port70.net>
Mail-Followup-To: Meng Zhuo <mzh@mzh.io>, musl@lists.openwall.com
References: <20240612160612.362246-1-mzh@mzh.io>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240612160612.362246-1-mzh@mzh.io>
Subject: Re: [musl] [PATCH v3] math: add riscv64 round/roundf

* Meng Zhuo <mzh@mzh.io> [2024-06-13 00:06:12 +0800]:

> ---
> First of all, sorry for any inconvenience.
> This is my first time try to contribute with mail list.
> ---
> v1 -> v3:
> * using isfinite to check NaN,+-inf and in range without raising INVALID
> * inlined copysign(f)
> ---
>  src/math/riscv64/round.c  | 20 ++++++++++++++++++++
>  src/math/riscv64/roundf.c | 20 ++++++++++++++++++++
>  2 files changed, 40 insertions(+)
>  create mode 100644 src/math/riscv64/round.c
>  create mode 100644 src/math/riscv64/roundf.c
> 
> diff --git a/src/math/riscv64/round.c b/src/math/riscv64/round.c
> new file mode 100644
> index 00000000..6d91c093
> --- /dev/null
> +++ b/src/math/riscv64/round.c
> @@ -0,0 +1,20 @@
> +#include <math.h>
> +
> +#if __riscv_flen >= 64
> +
> +double round(double x)
> +{
> +	if (!isfinite(x)) return x;

!isfinite(x) && fabs(x) >= 0x1p52

otherwise n can overflow below.

(note that the check can be a single cmp+branch on the bit representation of x)

> +	double tmp;
> +	long long n;
> +	__asm__ ("fcvt.l.d %0, %1, rmm" : "=r"(n) : "f"(x));
> +	__asm__ ("fcvt.d.l %0, %1" : "=f"(tmp) : "r"(n));
> +	__asm__ ("fsgnj.d %0, %1, %2" : "=f"(x) : "f"(tmp), "f"(x));

i'd add a comment that the sign bit is only copied to handle round(-0.0)

> +	return x;
> +}
> +
> +#else
> +
> +#include "../round.c"
> +
> +#endif
> diff --git a/src/math/riscv64/roundf.c b/src/math/riscv64/roundf.c
> new file mode 100644
> index 00000000..e67e791f
> --- /dev/null
> +++ b/src/math/riscv64/roundf.c
> @@ -0,0 +1,20 @@
> +#include <math.h>
> +
> +#if __riscv_flen >= 32
> +
> +float roundf(float x)
> +{
> +	if (!isfinite(x)) return x;

&& fabs(x) >= 0x1p23

> +	float tmp;
> +	long n;
> +	__asm__ ("fcvt.w.s %0, %1, rmm" : "=r"(n) : "f"(x));
> +	__asm__ ("fcvt.s.w %0, %1" : "=f"(tmp) : "r"(n));
> +	__asm__ ("fsgnj.s %0, %1, %2" : "=f"(x) : "f"(tmp), "f"(x));
> +	return x;
> +}
> +
> +#else
> +
> +#include "../roundf.c"
> +
> +#endif
> -- 
> 2.39.2