The getdents64 syscall adds on 32-but platforms padding which isn't needed
and not included in the userlevel data structure definition. We have to
avoid copying those padding bytes in the readdir64_r function.
Like the real header, the libc-internal wrapper for wchar.h needs to
undefine the macros so that if the header was already included before
the macros don't stay defined and cause problems later.
When doing i686-unknown-linux-gnu build configured with --enable-kernel=2.6.24,
there are several warnings like this:
../sysdeps/unix/sysv/linux/i386/fcntl.c:36:12: warning: ‘miss_F_GETOWN_EX’ defined but not used
It's already so marked in dl-sysdep.c. Failure to so mark
in the header file leads the compiler to believe that the
variable should be addressable via the .sdata section.
Signed-off-by: Richard Henderson <rth@twiddle.net>
GCC 4.5 warns about "extern void _end; &end;".
Use char[] instead, as that also doesn't fall foul
of a target's .sdata optimizations.
Signed-off-by: Richard Henderson <rth@twiddle.net>
When not using gethostbyname4 methods we immediately aborted the loop
over the nss modules on the first successful lookup. While this is
almost always what is wanted the nsswitch.conf file allows to select
something different.
The old implementation uses fd 0 to determine the login TTY. This
was needed because using /dev/tty it is not possible to deduce the
login TTY. For some time now there is the pseudo-file
/proc/self/loginuid which directly helps us to find the user. Prefer
using this file. It also works if stdin is closed, redirected, or
re-opened.
If we should use further fields in the e_ident array in the ELF header
of files we want zero to mean the default. Enforce zero for now so
that non-optional features indicated by nonzero values are not go
unnoticed when using an old dynamic linker.
