From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chen Gang Subject: Re: [PATCH] arch: m68k: math-emu: unsigned issue, 'unsigned long' will never be less than zero. Date: Thu, 30 May 2013 16:08:54 +0800 Message-ID: <51A70916.40101@asianux.com> References: <51A6F0C6.9050806@asianux.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-m68k-owner@vger.kernel.org List-Id: linux-m68k@vger.kernel.org To: Andreas Schwab Cc: Geert Uytterhoeven , linux-m68k@lists.linux-m68k.org, "linux-kernel@vger.kernel.org" , Linux-Arch On 05/30/2013 03:09 PM, Andreas Schwab wrote: > Chen Gang writes: > >> > 'oldmant.m32[1]' is 'unsigned long' which never '< 0', and the original >> > author wanted to check the highest bit whether set. >> > >> > So need type cast form 'unsigned long' to 'long'. > Better to make the bit test explicit, since it is not a sign bit. > That sounds good. >> > diff --git a/arch/m68k/math-emu/fp_arith.c b/arch/m68k/math-emu/fp_arith.c >> > index 08f286d..ef013c5 100644 >> > --- a/arch/m68k/math-emu/fp_arith.c >> > +++ b/arch/m68k/math-emu/fp_arith.c >> > @@ -519,7 +519,7 @@ static void fp_roundint(struct fp_ext *dest, int mode) >> > return; >> > break; >> > case 0x401e: >> > - if (!(oldmant.m32[1] >= 0)) >> > + if (!((long)oldmant.m32[1] >= 0)) > if (oldmant.m32[1] & 0x80000000) OK, I will send patch v2. Thanks. -- Chen Gang Asianux Corporation