From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on inbox.vuxu.org
X-Spam-Level: 
X-Spam-Status: No, score=-3.3 required=5.0 tests=MAILING_LIST_MULTI,
	RCVD_IN_DNSWL_MED,RCVD_IN_MSPIKE_H3,RCVD_IN_MSPIKE_WL autolearn=ham
	autolearn_force=no version=3.4.4
Received: (qmail 17008 invoked from network); 20 Nov 2020 18:02:02 -0000
Received: from mother.openwall.net (195.42.179.200)
  by inbox.vuxu.org with ESMTPUTF8; 20 Nov 2020 18:02:02 -0000
Received: (qmail 13980 invoked by uid 550); 20 Nov 2020 18:01:56 -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 13962 invoked from network); 20 Nov 2020 18:01:56 -0000
Date: Fri, 20 Nov 2020 19:01:44 +0100
From: Szabolcs Nagy <nsz@port70.net>
To: Jinliang Li <jinliang.li@linux.alibaba.com>
Cc: musl@lists.openwall.com
Message-ID: <20201120180144.GN1370092@port70.net>
Mail-Followup-To: Jinliang Li <jinliang.li@linux.alibaba.com>,
	musl@lists.openwall.com
References: <20201120104503.65390-1-jinliang.li@linux.alibaba.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20201120104503.65390-1-jinliang.li@linux.alibaba.com>
Subject: Re: [musl] [PATCH] Fix build error for vabs.f64 and vsqrt.f64

* Jinliang Li <jinliang.li@linux.alibaba.com> [2020-11-20 18:45:03 +0800]:
> To fix the build error "Error: selected fpu does not support instruction vabs.f64 or vsqrt.f64",
> when single precision fpu, e.g. -mfpu=fpv5-sp-d16 is selected.
> 
> Check if double precision fpu is supported for vabs.f64 and vsqrt.f64.

yes i think this patch should work:

both vfp and softfp call abi should work with single prec fpu.

the fp register state and load/store instructions should be
compatible between double prec and single prec only fpu so
setjmp/longjmp/fenv asm should work too.

so it is enough to fix up math asm to support single prec only fpu.

> ---
>  src/math/arm/fabs.c | 2 +-
>  src/math/arm/sqrt.c | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/src/math/arm/fabs.c b/src/math/arm/fabs.c
> index f890520a..6e1d367d 100644
> --- a/src/math/arm/fabs.c
> +++ b/src/math/arm/fabs.c
> @@ -1,6 +1,6 @@
>  #include <math.h>
>  
> -#if __ARM_PCS_VFP
> +#if __ARM_PCS_VFP && __ARM_FP&8
>  
>  double fabs(double x)
>  {
> diff --git a/src/math/arm/sqrt.c b/src/math/arm/sqrt.c
> index 874af960..567e2e91 100644
> --- a/src/math/arm/sqrt.c
> +++ b/src/math/arm/sqrt.c
> @@ -1,6 +1,6 @@
>  #include <math.h>
>  
> -#if __ARM_PCS_VFP || (__VFP_FP__ && !__SOFTFP__)
> +#if (__ARM_PCS_VFP || (__VFP_FP__ && !__SOFTFP__)) && (__ARM_FP&8)
>  
>  double sqrt(double x)
>  {
> -- 
> 2.28.0