From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <xry111@xry111.site>
Received: from xry111.site (xry111.site [89.208.246.23])
	by sourceware.org (Postfix) with ESMTPS id 9AF863858C74
	for <libc-alpha@sourceware.org>; Mon,  7 Nov 2022 14:38:34 +0000 (GMT)
DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 9AF863858C74
Authentication-Results: sourceware.org; dmarc=pass (p=reject dis=none) header.from=xry111.site
Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=xry111.site
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=xry111.site;
	s=default; t=1667831913;
	bh=s+PmJ9oqt9R63wxuI+Pw1CJsmNQqhsg8kCqQf4uCrqc=;
	h=Subject:From:To:Cc:Date:In-Reply-To:References:From;
	b=Hp4oh6PEAVUSfC2B9704C/U8tfQDs8OiyQIkkkGiHe8U9kVNyaPAvwGXpK+xnm7l7
	 SqSTKta+HcN+VQvjXC8JqVgjHhDwwBkKm5BUQZn+AJ2PcUdgariLi3ZET4wW+EhQSN
	 p8PVJRXLcYShJYp0u1fn4IJEpS44LFrtuj9Ot7ZA=
Received: from localhost.localdomain (xry111.site [IPv6:2001:470:683e::1])
	(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)
	 key-exchange ECDHE (P-256) server-signature ECDSA (P-384) server-digest SHA384)
	(Client did not present a certificate)
	(Authenticated sender: xry111@xry111.site)
	by xry111.site (Postfix) with ESMTPSA id 3CEF6668F0;
	Mon,  7 Nov 2022 09:38:31 -0500 (EST)
Message-ID: <a19e64a5599bc0cf3f01f453490baf2060a22d31.camel@xry111.site>
Subject: Re: [PATCH v2 01/10] LoongArch: Hard Float Support for functions
 {ll/l/ }rint{f/ }.
From: Xi Ruoyao <xry111@xry111.site>
To: Xiaolin Tang <tangxiaolin@loongson.cn>, adhemerval.zanella@linaro.org, 
	libc-alpha@sourceware.org, i.swmail@xen0n.name
Cc: caiyinyu@loongson.cn, xuchenghua@loongson.cn, chenglulu@loongson.cn
Date: Mon, 07 Nov 2022 22:38:29 +0800
In-Reply-To: <20221107140523.896735-2-tangxiaolin@loongson.cn>
References: <20221107140523.896735-1-tangxiaolin@loongson.cn>
	 <20221107140523.896735-2-tangxiaolin@loongson.cn>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.46.0 
MIME-Version: 1.0
X-Spam-Status: No, score=-0.5 required=5.0 tests=BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,FROM_SUSPICIOUS_NTLD,LIKELY_SPAM_FROM,PDS_OTHER_BAD_TLD,SPF_HELO_PASS,SPF_PASS,TXREP autolearn=no autolearn_force=no version=3.4.6
X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org
List-Id: <libc-alpha.sourceware.org>

On Mon, 2022-11-07 at 22:05 +0800, Xiaolin Tang wrote:

> +long long int
> +__llrint (double x)
> +{
> +=C2=A0 asm volatile ("frint.d \t%0, %1" : "=3Df" (x) : "f" (x));
> +=C2=A0 return x;
> +}

GCC trunk generates:

__llrint:
	frint.d 	$f0, $f0
	ftintrz.l.d $f0,$f0
	movfr2gr.d	$r4,$f0
	jr	$r1

It's not optimal.  The optimal way is:

__llrint:
	ftint.l.d 	$f0, $f0
	movfr2gr.d	$r4,$f0
	jr	$r1

I have a GCC patch to expand __builtin_llrint to ftint.l.d and
movgr2fr.d (with -fno-math-errno, which is enabled by Glibc building
system).

But I need some time to test my patch correctly: unfortunately I'm now
hitting some "random crashes" running Expect on LoongArch so I can't
complete GCC regression test.