|
From: Jeroen N. W. <jn...@xs...> - 2004-10-11 17:42:04
|
There seems to be a problem in corecheck/tests/as_mmap.stderr.exp (CVS). It expects valgrind to report problems with syscall mmap2, but program as_mmap.c uses syscall mmap, which valgrind reports correctly. If this indeed is a problem, the patch below can be used to fix it. Cheers, Jeroen. Index: corecheck/tests/as_mmap.stderr.exp =================================================================== RCS file: /home/kde/valgrind/corecheck/tests/as_mmap.stderr.exp,v retrieving revision 1.1 diff -u -r1.1 as_mmap.stderr.exp --- corecheck/tests/as_mmap.stderr.exp 16 Dec 2003 02:14:00 -0000 1.1 +++ corecheck/tests/as_mmap.stderr.exp 11 Oct 2004 17:36:00 -0000 @@ -1,9 +1,9 @@ -Warning: client syscall mmap2 tried to modify addresses 0x0-0x10000 +Warning: client syscall mmap tried to modify addresses 0x0-0x10000 mmap @ 0x00000000: Cannot allocate memory -Warning: client syscall mmap2 tried to modify addresses 0xB0000000-0xB0010000 +Warning: client syscall mmap tried to modify addresses 0xB0000000-0xB0010000 mmap @ top: Cannot allocate memory -Warning: client syscall mmap2 tried to modify addresses 0xB8000000-0xB8010000 +Warning: client syscall mmap tried to modify addresses 0xB8000000-0xB8010000 mmap @ top+.5G: Cannot allocate memory ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0) |
|
From: Tom H. <th...@cy...> - 2004-10-11 18:23:55
|
In message <228...@we...>
"Jeroen N. Witmond" <jn...@xs...> wrote:
> There seems to be a problem in corecheck/tests/as_mmap.stderr.exp (CVS).
> It expects valgrind to report problems with syscall mmap2, but program
> as_mmap.c uses syscall mmap, which valgrind reports correctly.
The problem is that as_mmap.c uses the glibc mmap() function so
there is no way of knowing whether mmap or mmap2 will be used.
Depending on how it is built, and what arguments are used, glibc
sometimes uses the old mmap system call and sometimes used the new
mmap2 system call.
> If this indeed is a problem, the patch below can be used to fix it.
Unfortunately that would break it on all my test systems ;-)
The best fix is probably to add an alternate set of results. Either
that or to add a custom filter to fixup the output.
Tom
--
Tom Hughes (th...@cy...)
Software Engineer, Cyberscience Corporation
http://www.cyberscience.com/
|