From: Trond Myklebust Subject: Re: [PATCH -next] nfs: fix ISO C90 warning Date: Wed, 16 Dec 2009 18:01:37 -0500 Message-ID: <1261004497.13878.9.camel@localhost> References: <20091215165309.58e98eea.sfr@canb.auug.org.au> <20091216142334.6a0ba115.randy.dunlap@oracle.com> <1261003219.13878.6.camel@localhost> <20091216144700.34b80b4b.randy.dunlap@oracle.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Cc: Randy Dunlap , Stephen Rothwell , linux-next@vger.kernel.org, LKML , linux-nfs@vger.kernel.org To: Thomas Gleixner Return-path: In-Reply-To: Sender: linux-next-owner@vger.kernel.org List-ID: On Wed, 2009-12-16 at 23:58 +0100, Thomas Gleixner wrote: > On Wed, 16 Dec 2009, Randy Dunlap wrote: > > On Wed, 16 Dec 2009 17:40:19 -0500 Trond Myklebust wrote: > > > > > On Wed, 2009-12-16 at 14:23 -0800, Randy Dunlap wrote: > > > > From: Randy Dunlap > > > > > > > > Fix gcc ISO C90 warning: > > > > > > > > fs/nfs/callback.c:356: warning: ISO C90 forbids mixed declarations and code > > > > > > > > Signed-off-by: Randy Dunlap > > > > --- > > > > fs/nfs/callback.c | 2 +- > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > --- linux-next-20091215.orig/fs/nfs/callback.c > > > > +++ linux-next-20091215/fs/nfs/callback.c > > > > @@ -352,8 +352,8 @@ static int check_gss_callback_principal( > > > > static int nfs_callback_authenticate(struct svc_rqst *rqstp) > > > > { > > > > struct nfs_client *clp; > > > > - RPC_IFDEBUG(char buf[RPC_MAX_ADDRBUFLEN]); > > > > int ret = SVC_OK; > > > > + RPC_IFDEBUG(char buf[RPC_MAX_ADDRBUFLEN]); > > > > > > > > > > What version of gcc is giving rise to this warning? > > > > > gcc --version > > gcc (GCC) 4.2.1 (SUSE Linux) > > > > > RPC_IFDEBUG is a macro that either evaluates to its argument, or to > > > nothing, depending on whether or not RPC_DEBUG is defined or not. In > > > neither case should it evaluate to anything illegal under C90 rules > > > afaics. > > > > Yep. Odd warning. > > Not really. If the debug macro evaluates to nothing then you have: > > struct nfs_client *clp; > ; > int ret = SVC_OK; > > So you have a stray semicolon, which is interpreted as an empty code > line. That qualifies for the mixed declaration and code case :) > > I know it's nitpicking, but ... Ah... I see what you mean. So really what we should do is just move that semicolon inside the macro. That would change the !RPC_DEBUG case to struct nfs_client *clp; int ret = SVC_OK; which is 100% legal... Cheers Trond