Re: [Qemu-devel] [PATCH] bswap: fix compiler warning

[Alexey Kardashevskiy]

On 04/13/2013 03:36 AM, Stefan Weil wrote:
> Am 12.04.2013 18:47, schrieb David Gibson:
> > On Fri, Apr 12, 2013 at 12:51:51PM +0200, Andreas Färber wrote:
> > > Am 12.04.2013 03:41, schrieb Alexey Kardashevskiy:
> > > > The bswap functions use memcpy but the bswap.h header itself does not seem to
> > > > include it in some configuration such as cross compiling for powerpc64
> > > > on x86_64 machine (gcc 4.6.3 from ftp.kernel.org, headers/libs from FC18/ppc64),
> > > > the example warning is below.
> > > >
> > > > The patch explicitly includes string.h.
> > > >
> > > >    CC    ppc64-softmmu/hw/virtio/virtio.o
> > > > In file included from /home/alexey/qemu/include/libfdt_env.h:22:0,
> > > >                   from /home/alexey/qemu/../lib4qemu/usr/include/libfdt.h:54,
> > > >                   from /home/alexey/qemu/hw/nvram/spapr_nvram.c:25:
> > > > /home/alexey/qemu/include/qemu/bswap.h: In function 'lduw_p':
> > > > /home/alexey/qemu/include/qemu/bswap.h:244:5: warning: implicit declaration of 
> > > > function 'memcpy' [-Wimplicit-function-declaration]
> > > > /home/alexey/qemu/include/qemu/bswap.h:244:5: warning: incompatible implicit 
> > > > declaration of built-in function 'memcpy' [enabled by default]
> > > >
> > > > Signed-off-by: Alexey Kardashevskiy <address@hidden>
> > > > ---
> > > >   include/qemu/bswap.h |    2 ++
> > > >   1 file changed, 2 insertions(+)
> > > >
> > > > diff --git a/include/qemu/bswap.h b/include/qemu/bswap.h
> > > > index d3af35d..d50de0d 100644
> > > > --- a/include/qemu/bswap.h
> > > > +++ b/include/qemu/bswap.h
> > > > @@ -1,6 +1,8 @@
> > > >   #ifndef BSWAP_H
> > > >   #define BSWAP_H
> > > >
> > > > +#include <string.h>
> > > > +
> > > >   #include "config-host.h"
> > > >   #include <inttypes.h>
> > > >   #include <limits.h>
> > > Including string.h is certainly the right thing to do, but why do you
> > > single it out first?
> > Yeah, it probably shouldn't go above the config header.  Otherwise
> > it's the right thing.
>
> If string.h went directly after limits.h, we'd even preserve the
> alphabetic order of the system headers :-)

Ah. Misunderstood the original question :) Do I need to repost it?


--
Alexey