From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:42944)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1Uu1yz-0002rA-3u
	for qemu-devel@nongnu.org; Tue, 02 Jul 2013 10:55:07 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1Uu1yt-0005Zr-20
	for qemu-devel@nongnu.org; Tue, 02 Jul 2013 10:55:02 -0400
Received: from mx1.redhat.com ([209.132.183.28]:59927)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1Uu1ys-0005ZQ-Mm
	for qemu-devel@nongnu.org; Tue, 02 Jul 2013 10:54:58 -0400
Message-ID: <51D2E9BC.1060401@redhat.com>
Date: Tue, 02 Jul 2013 16:54:52 +0200
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <CANd1uZm7GFAUcNT=qUQjtAHOp2N8Mrv7=K2gidQrn-_szYK1Tg@mail.gmail.com>
In-Reply-To: <CANd1uZm7GFAUcNT=qUQjtAHOp2N8Mrv7=K2gidQrn-_szYK1Tg@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH] int128: optimize
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Jay Foad <jay.foad@gmail.com>
Cc: qemu-devel@nongnu.org

Il 02/07/2013 14:46, Jay Foad ha scritto:
>>  static inline Int128 int128_neg(Int128 a)
>>  {
>> -    a.lo = ~a.lo;
>> -    a.hi = ~a.hi;
>> -    return int128_add(a, int128_one());
>> +    uint64_t lo = -a.lo;
>> +    return (Int128) { lo, ~a.hi + !lo };
>>  }
> 
> This leaves int128_one unused. (Also the temporary lo seems a bit
> pointless, since you could just as well write -a.lo and !a.lo.)

The unused int128_one is not a problem, someone might find a use later.

>>  static inline bool int128_ge(Int128 a, Int128 b)
>>  {
>> -    return int128_nonneg(int128_sub(a, b));
>> +    return a.hi > b.hi || (a.hi == b.hi && a.lo >= b.lo);
>>  }
> 
> This is a bug fix. The old version gives the wrong answer when a and b
> are both large and have opposite signs.

We don't really use Int128's that are bigger than 2^64, but you are
right.  I'll add this to the commit message.

Paolo