From patchwork Mon Jan 28 18:52:33 2013 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Richard Henderson X-Patchwork-Id: 216302 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id 04F8E2C0097 for ; Tue, 29 Jan 2013 05:53:10 +1100 (EST) Received: from localhost ([::1]:56888 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TztpM-0000gb-3e for incoming@patchwork.ozlabs.org; Mon, 28 Jan 2013 13:53:08 -0500 Received: from eggs.gnu.org ([208.118.235.92]:41981) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Tztoz-0000Lc-PD for qemu-devel@nongnu.org; Mon, 28 Jan 2013 13:52:52 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Tztow-0005Jj-Sz for qemu-devel@nongnu.org; Mon, 28 Jan 2013 13:52:45 -0500 Received: from mail-qa0-f52.google.com ([209.85.216.52]:35882) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Tztow-0005Jd-Of for qemu-devel@nongnu.org; Mon, 28 Jan 2013 13:52:42 -0500 Received: by mail-qa0-f52.google.com with SMTP id bs12so1138377qab.18 for ; Mon, 28 Jan 2013 10:52:42 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=x-received:sender:from:to:subject:date:message-id:x-mailer :in-reply-to:references; bh=aOq75veIqp9uuIMFdGuJbfOMVW14lhhj7rwcJGq1Reo=; b=Wm8N6AlwtaJL0mtRZ7ufJaYJgshDbvRGHYtHAAU2TII6AWUKCPn426L460Reo9Cq01 Hnhd+6RYTdwy8legw+sjBbvwIUDil5TorDWtASs2Z5balaVYUbbdtAcpZ0wCak1y7t61 7LOXVhhgeRul0ES7DlEmI/qek59iIFaIGNHlJpETjmeoefwtcuDIpZrqbikE7j/+0Isy iPisFU95aHnaUAfHSBWAKthfNjzd/v7pN0KSnunrjlJjNLLbRyfuV+jWt6CfqT+SXGok r/UucOccokKKqC6/lHeBwBnzVRckDscZQ9Fg29cKesLeQcjyVmIGdZq5nOZG5D2zhx28 oE7Q== X-Received: by 10.49.84.104 with SMTP id x8mr20053227qey.5.1359399162187; Mon, 28 Jan 2013 10:52:42 -0800 (PST) Received: from anchor.twiddle.home.com (50-194-63-110-static.hfc.comcastbusiness.net. [50.194.63.110]) by mx.google.com with ESMTPS id bb8sm6203883qeb.5.2013.01.28.10.52.40 (version=TLSv1 cipher=RC4-SHA bits=128/128); Mon, 28 Jan 2013 10:52:41 -0800 (PST) From: Richard Henderson To: qemu-devel@nongnu.org Date: Mon, 28 Jan 2013 10:52:33 -0800 Message-Id: <1359399154-13050-2-git-send-email-rth@twiddle.net> X-Mailer: git-send-email 1.7.11.7 In-Reply-To: <1359399154-13050-1-git-send-email-rth@twiddle.net> References: <1359399154-13050-1-git-send-email-rth@twiddle.net> X-detected-operating-system: by eggs.gnu.org: GNU/Linux 3.x [fuzzy] X-Received-From: 209.85.216.52 Subject: [Qemu-devel] [PATCH 1/2] host-utils: Use __int128 for mul[us]64 X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Sender: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Replace some x86_64 specific inline assembly with something that all 64-bit hosts ought to optimize well. At worst this becomes a call to the gcc __multi3 routine, which is no worse than our implementation in util/host-utils.c. With gcc 4.7, we get identical code generation for x86_64. We now get native multiplication on ia64 and s390x hosts. With minor improvements to gcc we can get it for ppc64 as well. Signed-off-by: Richard Henderson --- configure | 20 ++++++++++++++++++++ include/qemu/host-utils.h | 17 ++++++++--------- util/host-utils.c | 4 ++-- 3 files changed, 30 insertions(+), 11 deletions(-) diff --git a/configure b/configure index b7635e4..ecf1cbc 100755 --- a/configure +++ b/configure @@ -3150,6 +3150,22 @@ if compile_prog "" "" ; then cpuid_h=yes fi +######################################## +# check if __int128 is usable. + +int128=no +cat > $TMPC << EOF +int main (void) { + __int128 a = 0; + unsigned __int128 b = 1; + a = a + b; + a = a * b; + return 0; +} +EOF +if compile_prog "" "" ; then + int128=yes +fi ########################################## # End of CC checks @@ -3692,6 +3708,10 @@ if test "$cpuid_h" = "yes" ; then echo "CONFIG_CPUID_H=y" >> $config_host_mak fi +if test "$int128" = "yes" ; then + echo "CONFIG_INT128=y" >> $config_host_mak +fi + if test "$glusterfs" = "yes" ; then echo "CONFIG_GLUSTERFS=y" >> $config_host_mak fi diff --git a/include/qemu/host-utils.h b/include/qemu/host-utils.h index 81c9a75..01f6610 100644 --- a/include/qemu/host-utils.h +++ b/include/qemu/host-utils.h @@ -27,22 +27,21 @@ #include "qemu/compiler.h" /* QEMU_GNUC_PREREQ */ -#if defined(__x86_64__) -#define __HAVE_FAST_MULU64__ +#ifdef CONFIG_INT128 static inline void mulu64(uint64_t *plow, uint64_t *phigh, uint64_t a, uint64_t b) { - __asm__ ("mul %0\n\t" - : "=d" (*phigh), "=a" (*plow) - : "a" (a), "0" (b)); + unsigned __int128 r = (unsigned __int128)a * b; + *plow = r; + *phigh = r >> 64; } -#define __HAVE_FAST_MULS64__ + static inline void muls64(uint64_t *plow, uint64_t *phigh, int64_t a, int64_t b) { - __asm__ ("imul %0\n\t" - : "=d" (*phigh), "=a" (*plow) - : "a" (a), "0" (b)); + __int128 r = (__int128)a * b; + *plow = r; + *phigh = r >> 64; } #else void muls64(uint64_t *phigh, uint64_t *plow, int64_t a, int64_t b); diff --git a/util/host-utils.c b/util/host-utils.c index 5e3915a..2d06a2c 100644 --- a/util/host-utils.c +++ b/util/host-utils.c @@ -30,7 +30,7 @@ //#define DEBUG_MULDIV /* Long integer helpers */ -#if !defined(__x86_64__) +#ifndef CONFIG_INT128 static void add128 (uint64_t *plow, uint64_t *phigh, uint64_t a, uint64_t b) { *plow += a; @@ -102,4 +102,4 @@ void muls64 (uint64_t *plow, uint64_t *phigh, int64_t a, int64_t b) a, b, *phigh, *plow); #endif } -#endif /* !defined(__x86_64__) */ +#endif /* !CONFIG_INT128 */