The twalk function is very difficult to use in a multi-threaded
program because there is no way to pass external state to the
iterator function.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>

Keep these functions compatible with Gnulib while adding
__time64_t support.  The basic idea is to move private API
declarations from include/time.h to time/mktime-internal.h, since
the former file cannot easily be shared with Gnulib whereas the
latter can.
Also, do some other minor cleanup while in the neighborhood.
* include/time.h: Include stdbool.h, time/mktime-internal.h.
(__mktime_internal): Move this prototype to time/mktime-internal.h,
since Gnulib needs it.
(__localtime64_r, __gmtime64_r) [__TIMESIZE == 64]:
Move these macros to time/mktime-internal.h, since Gnulib needs them.
(__mktime64, __timegm64) [__TIMESIZE != 64]: New prototypes.
(in_time_t_range): New static function.
* posix/bits/types.h (__time64_t) [__TIMESIZE == 64 && !defined __LIBC]:
Do not define as a macro in this case, so that portable code is
less tempted to use __time64_t.
* time/mktime-internal.h: Rewrite so that it does both glibc
and Gnulib work.  Include time.h if not _LIBC.
(mktime_offset_t) [!_LIBC]: Define for gnulib.
(__time64_t, __gmtime64_r, __localtime64_r, __mktime64, __timegm64)
[!_LIBC || __TIMESIZE == 64]: New macros, mostly moved here
from include/time.h.
(__gmtime_r, __localtime_r, __mktime_internal) [!_LIBC]:
New macros, taken from GNulib.
(__mktime_internal): New prototype, moved here from include/time.h.
* time/mktime.c (mktime_min, mktime_max, convert_time)
(ranged_convert, __mktime_internal, __mktime64):
* time/timegm.c (__timegm64):
Use __time64_t, not time_t.
* time/mktime.c: Stop worrying about whether time_t is floating-point.
(__mktime64) [! (_LIBC && __TIMESIZE != 64)]:
Rename from mktime.
(mktime) [_LIBC && __TIMESIZE != 64]: New function.
* time/timegm.c [!_LIBC]: Include libc-config.h, not config.h,
for libc_hidden_def.
Include errno.h.
(__timegm64) [! (_LIBC && __TIMESIZE != 64)]:
Rename from timegm.
(timegm) [_LIBC && __TIMESIZE != 64]: New function.

First cut at publicizing __time64_t

alloc_buffer_next is useful for peeking to the remaining part of the
buffer and update it, with subsequent allocation (once the length
is known) using alloc_buffer_alloc_bytes.  This is not as robust
as the other interfaces, but it allows using alloc_buffer with
string-writing interfaces such as snprintf and ns_name_ntop.

This allows an architecture to use the old generic implementation
and also set explicit loop unrolling.

Checked on aarch64-linux-gnu.

	* include/loop_unroll.h: New file.
	* wcsmbs/wcscpy (__wcscpy): Add option to use loop unrolling
	besides generic implementation.

This patch removes the HP_TIMING_BITS usage for fast random bits and replace
with clock_gettime (CLOCK_MONOTONIC).  It has unspecified starting time and
nano-second accuracy, so its randomness is significantly better than
gettimeofday.

Althoug it should incur in more overhead (specially for architecture that
support hp-timing), the symbol is also common implemented as a vDSO.

Checked on aarch64-linux-gnu, x86_64-linux-gnu, and i686-linux-gnu. I also
checked on a i686-gnu build.

	* include/random-bits.h: New file.
	* resolv/res_mkquery.c [HP_TIMING_AVAIL] (RANDOM_BITS,
	(__res_context_mkquery): Remove usage hp-timing usage and replace with
	random_bits.
	* resolv/res_send.c [HP_TIMING_AVAIL] (nameserver_offset): Likewise.
	* sysdeps/posix/tempname.c [HP_TIMING_AVAIL] (__gen_tempname):
	Likewise.

With clock_getres, clock_gettime, and clock_settime refactor to remove the
generic CLOCK_PROCESS_CPUTIME_ID and CLOCK_THREAD_CPUTIME_ID support through
hp-timing, there is no usage of internal __get_clockfreq.  This patch removes
both generic and Linux implementation..

Checked with a build against aarch64-linux-gnu, i686-linux-gnu, ia64-linux-gnu,
sparc64-linux-gnu, powerpc-linux-gnu-power4.

	* include/libc-internal.h (__get_clockfreq): Remove prototype.
	* rt/Makefile (clock-routines): Remove get_clockfreq.
	* rt/get_clockfreq.c: Remove file.
	* sysdeps/unix/sysv/linux/i386/get_clockfreq.c: Likewise.
	* sysdeps/unix/sysv/linux/ia64/get_clockfreq.c: Likewise.
	* sysdeps/unix/sysv/linux/sparc/sparc64/get_clockfreq.c: Likewise.
	* sysdeps/unix/sysv/linux/powerpc/get_clockfreq.c: Move code to ...
	* sysdeps/unix/sysv/linux/powerpc/get_timebase_freq.c: ... here.

The stub implementations are turned into compat symbols.

Linux actually has two reserved system call numbers (for getpmsg
and putpmsg), but these system calls have never been implemented,
and there are no plans to implement them, so this patch replaces
the wrappers with the generic stubs.

According to <https://bugzilla.redhat.com/show_bug.cgi?id=436349>,
the presence of the XSI STREAMS declarations is a minor portability
hazard because they are not actually implemented.

This commit does not change the TIRPC support code in
sunrpc/rpc_svcout.c.  It uses additional XTI functionality and
therefore never worked with glibc.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>

inttypes.h and stdint.h are in sysdeps/generic, but there are no other
versions of these headers anywhere in the source tree, so they aren’t
actually system-dependent.  Move them to the subdirectory that
installs them (stdlib).

Reviewed-by: Joseph Myers <joseph@codesourcery.com>

	* sysdeps/generic/inttypes.h, sysdeps/generic/stdint.h:
	Move to stdlib.
	* include/inttypes.h: Adjust to match.
	* include/stdint.h: New wrapper.

On platforms where long double may have the same format as double
(-mlong-double-64), error and error_at_line do not take that into
account and might produce wrong output if a long double conversion is
requested by the format string ('%Lf').  This patch adds compatibility
functions for this situation and redirects calls via header magic.

Tested for powerpc, powerpc64 and powerpc64le.

When support for long double format with 128-bits (-mlong-double-128)
was added for platforms where long double had the same format as double,
such as powerpc, compatibility versions for the functions listed in the
commit title were missed.  Since the older format of long double can
still be used (with -mlong-double-64), using these functions with a
format string that requests the printing of long double variables will
produce wrong outputs.

This patch adds the missing compatibility functions and header magic to
redirect calls to them when -mlong-double-64 is in use.

Tested for powerpc, powerpc64 and powerpc64le.