public inbox for glibc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug libc/15512] New: __bswap_constant_16 not compiled when -Werror -Wsign-conversion is given
@ 2013-05-21 18:56 edjee at google dot com
  2014-06-13 17:41 ` [Bug libc/15512] " fweimer at redhat dot com
  0 siblings, 1 reply; 2+ messages in thread
From: edjee at google dot com @ 2013-05-21 18:56 UTC (permalink / raw)
  To: glibc-bugs

http://sourceware.org/bugzilla/show_bug.cgi?id=15512

             Bug #: 15512
           Summary: __bswap_constant_16 not compiled when -Werror
                    -Wsign-conversion is given
           Product: glibc
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: libc
        AssignedTo: unassigned@sourceware.org
        ReportedBy: edjee@google.com
                CC: drepper.fsp@gmail.com
    Classification: Unclassified


(Filling a bug, following Andreas' suggestion)

I'm using gcc and glibc on 32bit LE ARM (But all the other platforms that use
the architecture-independent bit/byteswap.h may have similar issue).

I found that the following code is not compiled, if I give -Werror
-Wsign-conversion arguments to gcc.


#include <stdio.h>
#include <byteswap.h>

int main(int argc, char *argv[]) {
  unsigned short int x = 0xbeef;
  printf("bswap_16(0x%x)=0x%x\n", x, bswap_16(x));
  return 0;
}


I saw compilation error "conversion to 'unsigned int' from 'int' may change the
sign of the result".

It looks like that a similar issue has been solved for x86 by
http://sourceware.org/git/?p=glibc.git;a=commit;h=69da074d7adfab7b57004a0dea9403a928e310a5
.
But the architecture-independent bit/byteswap.h has not been modified. In that
file, __bswap_constant_16 macro looks like this:

#define __bswap_constant_16(x) \
        ((unsigned short int)((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))


It seems that ((x) >> 8) promotes the data to signed value, even if x is
unsigned. So (((x) >> 8) && 0xffu) seems to fail when -Werror -Wsign-conversion
is given.

All the architecture-specific versions of byteswap.h uses 0xff but this file
uses 0xffu. Is this intended?

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.


^ permalink raw reply	[flat|nested] 2+ messages in thread

* [Bug libc/15512] __bswap_constant_16 not compiled when -Werror -Wsign-conversion is given
  2013-05-21 18:56 [Bug libc/15512] New: __bswap_constant_16 not compiled when -Werror -Wsign-conversion is given edjee at google dot com
@ 2014-06-13 17:41 ` fweimer at redhat dot com
  0 siblings, 0 replies; 2+ messages in thread
From: fweimer at redhat dot com @ 2014-06-13 17:41 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=15512

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
              Flags|                            |security-

-- 
You are receiving this mail because:
You are on the CC list for the bug.


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2014-06-13 17:41 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-05-21 18:56 [Bug libc/15512] New: __bswap_constant_16 not compiled when -Werror -Wsign-conversion is given edjee at google dot com
2014-06-13 17:41 ` [Bug libc/15512] " fweimer at redhat dot com

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).