libbb: add bit counting function, use where appropriate

Although "naive" counting function is not too slow and is smaller,
using it on e.g. each of 1024 words of CPU mask feels wrong.

function                                             old     new   delta
bb_popcnt_32                                           -      52     +52
get_prefix                                           323     321      -2
nproc_main                                           206     199      -7
d4_run_script                                        739     731      -8
ipcalc_main                                          533     507     -26
------------------------------------------------------------------------------
(add/remove: 2/0 grow/shrink: 0/4 up/down: 52/-43)              Total: 9 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

1 files changed, 3 insertions, 6 deletions

diff --git a/coreutils/nproc.c b/coreutils/nproc.c
index a0d818c59..df63bf57a 100644
--- a/coreutils/nproc.c
+++ b/coreutils/nproc.c
@@ -56,13 +56,10 @@ int nproc_main(int argc UNUSED_PARAM, char **argv UNUSED_PARAM)
                 unsigned long *mask = get_malloc_cpu_affinity(0, &sz);
                 sz /= sizeof(long);
                 for (i = 0; i < sz; i++) {
-                        unsigned long m = mask[i];
-                        while (m) {
-                                if (m & 1)
-                                        count++;
-                                m >>= 1;
-                        }
+                        if (mask[i] != 0) /* most mask[i] are usually 0 */
+                                count += bb_popcnt_long(mask[i]);
                 }
+                IF_FEATURE_CLEAN_UP(free(mask);)
         }
 
         IF_LONG_OPTS(count -= ignore;)