Patchwork net/socket: Fix compiler warning (regression for MinGW)

login
register
mail settings
Submitter Stefan Weil
Date Sept. 20, 2012, 7:32 p.m.
Message ID <1348169539-24797-1-git-send-email-sw@weilnetz.de>
Download mbox | patch
Permalink /patch/185448/
State Superseded
Headers show

Comments

Stefan Weil - Sept. 20, 2012, 7:32 p.m.
Add a type cast which was removed by commit
213fd5087e2e4e2da10ad266df0ba950cf7618bf again.

Without it, MinGW compilers complain:

net/socket.c:136: warning:
 pointer targets in passing argument 2 of ‘sendto’ differ in signedness
/usr/lib/gcc/amd64-mingw32msvc/4.4.4/../../../../amd64-mingw32msvc/include/winsock2.h:1313: note:
 expected ‘const char *’ but argument is of type ‘const uint8_t *’

Signed-off-by: Stefan Weil <sw@weilnetz.de>
---
 net/socket.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Stefan Hajnoczi - Sept. 22, 2012, 4:32 p.m.
On Thu, Sep 20, 2012 at 09:32:19PM +0200, Stefan Weil wrote:
> Add a type cast which was removed by commit
> 213fd5087e2e4e2da10ad266df0ba950cf7618bf again.
> 
> Without it, MinGW compilers complain:
> 
> net/socket.c:136: warning:
>  pointer targets in passing argument 2 of ‘sendto’ differ in signedness
> /usr/lib/gcc/amd64-mingw32msvc/4.4.4/../../../../amd64-mingw32msvc/include/winsock2.h:1313: note:
>  expected ‘const char *’ but argument is of type ‘const uint8_t *’

Wow, that's messed up.  sendto() is POSIX and the prototype shouldn't be
const char *.

It's easy for someone to remove this cast in the future.  Please add a
comment explaining that it's needed because MinGW headers don't have the
POSIX version of sendto().

Thanks,
Stefan
Michael Tokarev - Sept. 22, 2012, 6:01 p.m.
On 22.09.2012 20:32, Stefan Hajnoczi wrote:
> On Thu, Sep 20, 2012 at 09:32:19PM +0200, Stefan Weil wrote:
>> Add a type cast which was removed by commit
>> 213fd5087e2e4e2da10ad266df0ba950cf7618bf again.
>>
>> Without it, MinGW compilers complain:
>>
>> net/socket.c:136: warning:
>>  pointer targets in passing argument 2 of ‘sendto’ differ in signedness
>> /usr/lib/gcc/amd64-mingw32msvc/4.4.4/../../../../amd64-mingw32msvc/include/winsock2.h:1313: note:
>>  expected ‘const char *’ but argument is of type ‘const uint8_t *’
> 
> Wow, that's messed up.  sendto() is POSIX and the prototype shouldn't be
> const char *.
> 
> It's easy for someone to remove this cast in the future.  Please add a
> comment explaining that it's needed because MinGW headers don't have the
> POSIX version of sendto().

There's qemu_recv() in qemu-common.h, for exactly the same purpose.
But qemu_recv() is much more evil, IMHO.

Thanks,

/mjt
Stefan Weil - Sept. 22, 2012, 6:17 p.m.
Am 22.09.2012 20:01, schrieb Michael Tokarev:
> On 22.09.2012 20:32, Stefan Hajnoczi wrote:
>> On Thu, Sep 20, 2012 at 09:32:19PM +0200, Stefan Weil wrote:
>>> Add a type cast which was removed by commit
>>> 213fd5087e2e4e2da10ad266df0ba950cf7618bf again.
>>>
>>> Without it, MinGW compilers complain:
>>>
>>> net/socket.c:136: warning:
>>>   pointer targets in passing argument 2 of ‘sendto’ differ in signedness
>>> /usr/lib/gcc/amd64-mingw32msvc/4.4.4/../../../../amd64-mingw32msvc/include/winsock2.h:1313: note:
>>>   expected ‘const char *’ but argument is of type ‘const uint8_t *’
>> Wow, that's messed up.  sendto() is POSIX and the prototype shouldn't be
>> const char *.
>>
>> It's easy for someone to remove this cast in the future.  Please add a
>> comment explaining that it's needed because MinGW headers don't have the
>> POSIX version of sendto().
> There's qemu_recv() in qemu-common.h, for exactly the same purpose.
> But qemu_recv() is much more evil, IMHO.
>
> Thanks,
>
> /mjt

Thanks for the hint.

I already sent v2 of my patch (with a comment as suggested by Stefan H.).

Of course we could also add a qemu_sendto to qemu-common.h.
Is there anybody who would prefer that solution?

Regards
Stefan W.
Blue Swirl - Sept. 22, 2012, 6:25 p.m.
On Sat, Sep 22, 2012 at 6:17 PM, Stefan Weil <sw@weilnetz.de> wrote:
> Am 22.09.2012 20:01, schrieb Michael Tokarev:
>
>> On 22.09.2012 20:32, Stefan Hajnoczi wrote:
>>>
>>> On Thu, Sep 20, 2012 at 09:32:19PM +0200, Stefan Weil wrote:
>>>>
>>>> Add a type cast which was removed by commit
>>>> 213fd5087e2e4e2da10ad266df0ba950cf7618bf again.
>>>>
>>>> Without it, MinGW compilers complain:
>>>>
>>>> net/socket.c:136: warning:
>>>>   pointer targets in passing argument 2 of ‘sendto’ differ in signedness
>>>>
>>>> /usr/lib/gcc/amd64-mingw32msvc/4.4.4/../../../../amd64-mingw32msvc/include/winsock2.h:1313:
>>>> note:
>>>>   expected ‘const char *’ but argument is of type ‘const uint8_t *’
>>>
>>> Wow, that's messed up.  sendto() is POSIX and the prototype shouldn't be
>>> const char *.
>>>
>>> It's easy for someone to remove this cast in the future.  Please add a
>>> comment explaining that it's needed because MinGW headers don't have the
>>> POSIX version of sendto().
>>
>> There's qemu_recv() in qemu-common.h, for exactly the same purpose.
>> But qemu_recv() is much more evil, IMHO.
>>
>> Thanks,
>>
>> /mjt
>
>
> Thanks for the hint.
>
> I already sent v2 of my patch (with a comment as suggested by Stefan H.).
>
> Of course we could also add a qemu_sendto to qemu-common.h.
> Is there anybody who would prefer that solution?

Yes.

>
> Regards
> Stefan W.
>

Patch

diff --git a/net/socket.c b/net/socket.c
index 69aad03..833f79f 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -131,7 +131,7 @@  static ssize_t net_socket_receive_dgram(NetClientState *nc, const uint8_t *buf,
     ssize_t ret;
 
     do {
-        ret = sendto(s->fd, buf, size, 0,
+        ret = sendto(s->fd, (const void *)buf, size, 0,
                      (struct sockaddr *)&s->dgram_dst,
                      sizeof(s->dgram_dst));
     } while (ret == -1 && errno == EINTR);