Do not use HP_TIMING_NOW for random bits

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.

resolv/res_mkquery.c

@@ -82,6 +82,7 @@
  * SOFTWARE.
  */
 
+#include <stdint.h>
 #include <sys/types.h>
 #include <sys/param.h>
 #include <netinet/in.h>
@@ -92,12 +93,7 @@
 #include <string.h>
 #include <sys/time.h>
 #include <shlib-compat.h>
-
-#include <hp-timing.h>
-#include <stdint.h>
-#if HP_TIMING_AVAIL
-# define RANDOM_BITS(Var) { uint64_t v64; HP_TIMING_NOW (v64); Var = v64; }
-#endif
+#include <random-bits.h>
 
 int
 __res_context_mkquery (struct resolv_context *ctx, int op, const char *dname,
@@ -120,16 +116,7 @@ __res_context_mkquery (struct resolv_context *ctx, int op, const char *dname,
   /* We randomize the IDs every time.  The old code just incremented
      by one after the initial randomization which still predictable if
      the application does multiple requests.  */
-  int randombits;
-#ifdef RANDOM_BITS
-  RANDOM_BITS (randombits);
-#else
-  struct timeval tv;
-  __gettimeofday (&tv, NULL);
-  randombits = (tv.tv_sec << 8) ^ tv.tv_usec;
-#endif
-
-  hp->id = randombits;
+  hp->id = random_bits ();
   hp->opcode = op;
   hp->rd = (ctx->resp->options & RES_RECURSE) != 0;
   hp->rcode = NOERROR;