git: 71ec05a21257 - main - sort: Cache value of MB_CUR_MAX
Mark Johnston
markj at FreeBSD.org
Thu May 13 13:35:19 UTC 2021
The branch main has been updated by markj:
URL: https://cgit.FreeBSD.org/src/commit/?id=71ec05a21257e159f40d54e26ad0011bb19b5134
commit 71ec05a21257e159f40d54e26ad0011bb19b5134
Author: Cyril Zhang <cyril at freebsdfoundation.org>
AuthorDate: 2021-05-13 12:55:06 +0000
Commit: Mark Johnston <markj at FreeBSD.org>
CommitDate: 2021-05-13 13:33:19 +0000
sort: Cache value of MB_CUR_MAX
Every usage of MB_CUR_MAX results in a call to __mb_cur_max. This is
inefficient and redundant. Caching the value of MB_CUR_MAX in a global
variable removes these calls and speeds up the runtime of sort. For
numeric sorting, runtime is almost halved in some tests.
PR: 255551
PR: 255840
Reviewed by: markj
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D30170
---
usr.bin/sort/bwstring.c | 54 ++++++++++++++++++++++++------------------------
usr.bin/sort/bwstring.h | 9 ++++----
usr.bin/sort/radixsort.c | 4 ++--
usr.bin/sort/sort.c | 6 +++++-
usr.bin/sort/sort.h | 6 ++++++
5 files changed, 45 insertions(+), 34 deletions(-)
diff --git a/usr.bin/sort/bwstring.c b/usr.bin/sort/bwstring.c
index cab21324c4b3..f6200c53c83e 100644
--- a/usr.bin/sort/bwstring.c
+++ b/usr.bin/sort/bwstring.c
@@ -59,7 +59,7 @@ initialise_months(void)
char *tmp;
size_t len;
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
if (cmonths == NULL) {
char *m;
@@ -143,7 +143,7 @@ void
bwsprintf(FILE *f, struct bwstring *bws, const char *prefix, const char *suffix)
{
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
fprintf(f, "%s%s%s", prefix, bws->data.cstr, suffix);
else
fprintf(f, "%s%S%s", prefix, bws->data.wstr, suffix);
@@ -158,14 +158,14 @@ const void* bwsrawdata(const struct bwstring *bws)
size_t bwsrawlen(const struct bwstring *bws)
{
- return ((MB_CUR_MAX == 1) ? bws->len : SIZEOF_WCHAR_STRING(bws->len));
+ return ((mb_cur_max == 1) ? bws->len : SIZEOF_WCHAR_STRING(bws->len));
}
size_t
bws_memsize(const struct bwstring *bws)
{
- return ((MB_CUR_MAX == 1) ? (bws->len + 2 + sizeof(struct bwstring)) :
+ return ((mb_cur_max == 1) ? (bws->len + 2 + sizeof(struct bwstring)) :
(SIZEOF_WCHAR_STRING(bws->len + 1) + sizeof(struct bwstring)));
}
@@ -175,7 +175,7 @@ bws_setlen(struct bwstring *bws, size_t newlen)
if (bws && newlen != bws->len && newlen <= bws->len) {
bws->len = newlen;
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
bws->data.cstr[newlen] = '\0';
else
bws->data.wstr[newlen] = L'\0';
@@ -190,14 +190,14 @@ bwsalloc(size_t sz)
{
struct bwstring *ret;
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
ret = sort_malloc(sizeof(struct bwstring) + 1 + sz);
else
ret = sort_malloc(sizeof(struct bwstring) +
SIZEOF_WCHAR_STRING(sz + 1));
ret->len = sz;
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
ret->data.cstr[ret->len] = '\0';
else
ret->data.wstr[ret->len] = L'\0';
@@ -218,7 +218,7 @@ bwsdup(const struct bwstring *s)
else {
struct bwstring *ret = bwsalloc(s->len);
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
memcpy(ret->data.cstr, s->data.cstr, (s->len));
else
memcpy(ret->data.wstr, s->data.wstr,
@@ -242,7 +242,7 @@ bwssbdup(const wchar_t *str, size_t len)
ret = bwsalloc(len);
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
for (size_t i = 0; i < len; ++i)
ret->data.cstr[i] = (unsigned char) str[i];
else
@@ -263,7 +263,7 @@ bwscsbdup(const unsigned char *str, size_t len)
ret = bwsalloc(len);
if (str) {
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
memcpy(ret->data.cstr, str, len);
else {
mbstate_t mbs;
@@ -277,7 +277,7 @@ bwscsbdup(const unsigned char *str, size_t len)
memset(&mbs, 0, sizeof(mbs));
while (cptr < len) {
- size_t n = MB_CUR_MAX;
+ size_t n = mb_cur_max;
if (n > len - cptr)
n = len - cptr;
@@ -334,7 +334,7 @@ bwscpy(struct bwstring *dst, const struct bwstring *src)
nums = dst->len;
dst->len = nums;
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
memcpy(dst->data.cstr, src->data.cstr, nums);
dst->data.cstr[dst->len] = '\0';
} else {
@@ -363,7 +363,7 @@ bwsncpy(struct bwstring *dst, const struct bwstring *src, size_t size)
nums = size;
dst->len = nums;
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
memcpy(dst->data.cstr, src->data.cstr, nums);
dst->data.cstr[dst->len] = '\0';
} else {
@@ -398,7 +398,7 @@ bwsnocpy(struct bwstring *dst, const struct bwstring *src, size_t offset,
if (nums > size)
nums = size;
dst->len = nums;
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
memcpy(dst->data.cstr, src->data.cstr + offset,
(nums));
dst->data.cstr[dst->len] = '\0';
@@ -420,7 +420,7 @@ size_t
bwsfwrite(struct bwstring *bws, FILE *f, bool zero_ended)
{
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
size_t len = bws->len;
if (!zero_ended) {
@@ -478,7 +478,7 @@ bwsfgetln(FILE *f, size_t *len, bool zero_ended, struct reader_buffer *rb)
eols = zero_ended ? btowc('\0') : btowc('\n');
- if (!zero_ended && (MB_CUR_MAX > 1)) {
+ if (!zero_ended && (mb_cur_max > 1)) {
wchar_t *ret;
ret = fgetwln(f, len);
@@ -494,7 +494,7 @@ bwsfgetln(FILE *f, size_t *len, bool zero_ended, struct reader_buffer *rb)
}
return (bwssbdup(ret, *len));
- } else if (!zero_ended && (MB_CUR_MAX == 1)) {
+ } else if (!zero_ended && (mb_cur_max == 1)) {
char *ret;
ret = fgetln(f, len);
@@ -523,7 +523,7 @@ bwsfgetln(FILE *f, size_t *len, bool zero_ended, struct reader_buffer *rb)
}
rb->fgetwln_z_buffer[*len] = 0;
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
while (!feof(f)) {
int c;
@@ -603,7 +603,7 @@ bwsncmp(const struct bwstring *bws1, const struct bwstring *bws2,
if (len < cmp_len)
cmp_len = len;
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
const unsigned char *s1, *s2;
s1 = bws1->data.cstr + offset;
@@ -696,7 +696,7 @@ bwscoll(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset)
len1 -= offset;
len2 -= offset;
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
const unsigned char *s1, *s2;
s1 = bws1->data.cstr + offset;
@@ -852,7 +852,7 @@ bwstod(struct bwstring *s0, bool *empty)
{
double ret = 0;
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
unsigned char *end, *s;
char *ep;
@@ -909,7 +909,7 @@ int
bws_month_score(const struct bwstring *s0)
{
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
const unsigned char *end, *s;
s = s0->data.cstr;
@@ -949,7 +949,7 @@ struct bwstring *
ignore_leading_blanks(struct bwstring *str)
{
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
unsigned char *dst, *end, *src;
src = str->data.cstr;
@@ -1005,7 +1005,7 @@ ignore_nonprinting(struct bwstring *str)
{
size_t newlen = str->len;
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
unsigned char *dst, *end, *src;
unsigned char c;
@@ -1058,7 +1058,7 @@ dictionary_order(struct bwstring *str)
{
size_t newlen = str->len;
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
unsigned char *dst, *end, *src;
unsigned char c;
@@ -1109,7 +1109,7 @@ struct bwstring *
ignore_case(struct bwstring *str)
{
- if (MB_CUR_MAX == 1) {
+ if (mb_cur_max == 1) {
unsigned char *end, *s;
s = str->data.cstr;
@@ -1137,7 +1137,7 @@ void
bws_disorder_warnx(struct bwstring *s, const char *fn, size_t pos)
{
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
warnx("%s:%zu: disorder: %s", fn, pos + 1, s->data.cstr);
else
warnx("%s:%zu: disorder: %ls", fn, pos + 1, s->data.wstr);
diff --git a/usr.bin/sort/bwstring.h b/usr.bin/sort/bwstring.h
index 9d62eaf37f62..b63bb97ab93f 100644
--- a/usr.bin/sort/bwstring.h
+++ b/usr.bin/sort/bwstring.h
@@ -38,6 +38,7 @@
#include <sysexits.h>
#include <wchar.h>
+#include "sort.h"
#include "mem.h"
extern bool byte_sort;
@@ -109,7 +110,7 @@ static inline bwstring_iterator
bws_end(struct bwstring *bws)
{
- return ((MB_CUR_MAX == 1) ?
+ return ((mb_cur_max == 1) ?
(bwstring_iterator) (bws->data.cstr + bws->len) :
(bwstring_iterator) (bws->data.wstr + bws->len));
}
@@ -118,7 +119,7 @@ static inline bwstring_iterator
bws_iterator_inc(bwstring_iterator iter, size_t pos)
{
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
return ((unsigned char *) iter) + pos;
else
return ((wchar_t*) iter) + pos;
@@ -128,7 +129,7 @@ static inline wchar_t
bws_get_iter_value(bwstring_iterator iter)
{
- if (MB_CUR_MAX == 1)
+ if (mb_cur_max == 1)
return *((unsigned char *) iter);
else
return *((wchar_t*) iter);
@@ -137,7 +138,7 @@ bws_get_iter_value(bwstring_iterator iter)
int
bws_iterator_cmp(bwstring_iterator iter1, bwstring_iterator iter2, size_t len);
-#define BWS_GET(bws, pos) ((MB_CUR_MAX == 1) ? ((bws)->data.cstr[(pos)]) : (bws)->data.wstr[(pos)])
+#define BWS_GET(bws, pos) ((mb_cur_max == 1) ? ((bws)->data.cstr[(pos)]) : (bws)->data.wstr[(pos)])
void initialise_months(void);
diff --git a/usr.bin/sort/radixsort.c b/usr.bin/sort/radixsort.c
index 4993566aeb77..4c448fad69e9 100644
--- a/usr.bin/sort/radixsort.c
+++ b/usr.bin/sort/radixsort.c
@@ -258,7 +258,7 @@ add_leaf(struct sort_level *sl, struct sort_list_item *item)
static inline int
get_wc_index(struct sort_list_item *sli, size_t level)
{
- const size_t wcfact = (MB_CUR_MAX == 1) ? 1 : sizeof(wchar_t);
+ const size_t wcfact = (mb_cur_max == 1) ? 1 : sizeof(wchar_t);
const struct key_value *kv;
const struct bwstring *bws;
@@ -331,7 +331,7 @@ free_sort_level(struct sort_level *sl)
static void
run_sort_level_next(struct sort_level *sl)
{
- const size_t wcfact = (MB_CUR_MAX == 1) ? 1 : sizeof(wchar_t);
+ const size_t wcfact = (mb_cur_max == 1) ? 1 : sizeof(wchar_t);
struct sort_level *slc;
size_t i, sln, tosort_num;
diff --git a/usr.bin/sort/sort.c b/usr.bin/sort/sort.c
index d5c18552d914..0fbbd9284aad 100644
--- a/usr.bin/sort/sort.c
+++ b/usr.bin/sort/sort.c
@@ -99,6 +99,8 @@ struct sort_opts sort_opts_vals;
bool debug_sort;
bool need_hint;
+size_t mb_cur_max;
+
#if defined(SORT_THREADS)
unsigned int ncpu = 1;
size_t nthreads = 1;
@@ -305,7 +307,7 @@ conv_mbtowc(wchar_t *wc, const char *c, const wchar_t def)
if (wc && c) {
int res;
- res = mbtowc(wc, c, MB_CUR_MAX);
+ res = mbtowc(wc, c, mb_cur_max);
if (res < 1)
*wc = def;
}
@@ -322,6 +324,8 @@ set_locale(void)
setlocale(LC_ALL, "");
+ mb_cur_max = MB_CUR_MAX;
+
lc = localeconv();
if (lc) {
diff --git a/usr.bin/sort/sort.h b/usr.bin/sort/sort.h
index b472fd16c098..96b36a65f4e4 100644
--- a/usr.bin/sort/sort.h
+++ b/usr.bin/sort/sort.h
@@ -127,4 +127,10 @@ extern struct sort_opts sort_opts_vals;
extern struct sort_mods * const default_sort_mods;
+/*
+ * Cached value of MB_CUR_MAX. Because MB_CUR_MAX is used often throughout the program,
+ * this avoids frequent calls to __mb_cur_max.
+ */
+extern size_t mb_cur_max;
+
#endif /* __BSD_SORT_H__ */
More information about the dev-commits-src-all
mailing list