This is the mail archive of the
mailing list for the Cygwin project.
Re: Broken autoconf mmap test
On 25 March 2011 10:17, Corinna Vinschen wrote:
>> If you guys think that cheating for autoconf on 64 bit systems is a good
>> idea, I'm willing to implement it.
> Here's a patch which works fine for me:
> Index: mmap.cc
> RCS file: /cvs/src/src/winsup/cygwin/mmap.cc,v
> retrieving revision 1.166
> diff -u -p -r1.166 mmap.cc
> --- mmap.cc Â Â 18 Mar 2011 13:56:56 -0000 Â Â Â1.166
> +++ mmap.cc Â Â 25 Mar 2011 10:17:09 -0000
> @@ -801,6 +801,38 @@ mmap64 (void *addr, size_t len, int prot
> Â Â Â /* mmap /dev/zero is like MAP_ANONYMOUS. */
> Â Â Â if (fh->get_device () == FH_ZERO)
> Â Â Â Âflags |= MAP_ANONYMOUS;
> + Â Â Â/* The autoconf mmap test maps a file of size 1 byte. ÂIt then tests
> + Â Â Â Âevery byte of the entire mapped page of 64K for 0-bytes since that's
> + Â Â Â Âwhat POSIX requires. ÂThe problem is, we can't create that mapping on
> + Â Â Â Â64 bit systems. ÂThe file mapping will be only a single page, 4K, and
> + Â Â Â Âsince 64 bit systems don't support the AT_ROUND_TO_PAGE flag, the
> + Â Â Â Âremainder of the 64K slot will result in a SEGV when accessed.
> + Â Â Â ÂSo, what we do here is cheating for autoconf. ÂWe know exactly what
> + Â Â Â Âautoconf is doing. ÂThe file is called conftest.txt, it has a size of
> + Â Â Â Â1 byte, the requested mapping size is 64K, the requested protection
> + Â Â Â Âis PROT_READ | PROT_WRITE, the mapping is MAP_SHARED, the offset is 0.
> + Â Â Â ÂIf all these requirements are given, we just return an anonymous map
> + Â Â Â Âof 64K. ÂThis will help to get over the autoconf test even on 64 bit
> + Â Â Â Âsystems. */
This should mention why cheating seems justified, i.e. the assumption
that "real" programs don't depend on being able to access past EOF.