From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:44255) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VqpIV-0000IT-P5 for qemu-devel@nongnu.org; Wed, 11 Dec 2013 14:18:24 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VqpIN-0000Ey-0D for qemu-devel@nongnu.org; Wed, 11 Dec 2013 14:18:15 -0500 From: Tom Musta Date: Wed, 11 Dec 2013 13:16:31 -0600 Message-Id: <1386789398-5239-12-git-send-email-tommusta@gmail.com> In-Reply-To: <1386789398-5239-1-git-send-email-tommusta@gmail.com> References: <1386789398-5239-1-git-send-email-tommusta@gmail.com> Subject: [Qemu-devel] [V2 PATCH 11/18] softfloat: Fix float64_to_uint32 List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: Tom Musta , qemu-ppc@nongnu.org The float64_to_uint32 has several flaws: - for numbers between 2**32 and 2**64, the inexact exception flag may get incorrectly set. In this case, only the invalid flag should be set. test pattern: 425F81378DC0CD1F / 0x1.f81378dc0cd1fp+38 - for numbers between 2**63 and 2**64, incorrect results may be produced: test pattern: 43EAAF73F1F0B8BD / 0x1.aaf73f1f0b8bdp+63 This patch re-implements float64_to_uint32 to re-use the float64_to_uint64 routine (instead of float64_to_int64). For the saturation case, the inexact bit is explicitly cleared before raising the invalid flag. Signed-off-by: Tom Musta --- fpu/softfloat.c | 10 ++++------ 1 files changed, 4 insertions(+), 6 deletions(-) diff --git a/fpu/softfloat.c b/fpu/softfloat.c index 1003e59..6a8b6f5 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -6578,15 +6578,13 @@ uint_fast16_t float32_to_uint16_round_to_zero(float32 a STATUS_PARAM) uint32 float64_to_uint32( float64 a STATUS_PARAM ) { - int64_t v; + uint64_t v; uint32 res; - v = float64_to_int64(a STATUS_VAR); - if (v < 0) { - res = 0; - float_raise( float_flag_invalid STATUS_VAR); - } else if (v > 0xffffffff) { + v = float64_to_uint64(a STATUS_VAR); + if (v > 0xffffffff) { res = 0xffffffff; + STATUS(float_exception_flags) &= ~float_flag_inexact; float_raise( float_flag_invalid STATUS_VAR); } else { res = v; -- 1.7.1