macro: add DIV_ROUND_UP()

This macro calculates A / B but rounds up instead of down. We explicitly
do *NOT* use:
        (A + B - 1) / A
as it suffers from an integer overflow, even though the passed values are
properly tested against overflow. Our test-cases show this behavior.

Instead, we use:
        A / B + !!(A % B)

Note that on "Real CPUs" this does *NOT* result in two divisions. Instead,
instructions like idivl@x86 provide both, the quotient and the remainder.
Therefore, both algorithms should perform equally well (I didn't verify
this, though).
This commit is contained in:
David Herrmann 2014-12-29 17:51:36 +01:00
parent 5ef378c1c5
commit 180a60bc87
2 changed files with 45 additions and 0 deletions

View File

@ -197,6 +197,17 @@ static inline unsigned long ALIGN_POWER2(unsigned long u) {
UNIQ_T(X,xq); \
})
/* [(x + y - 1) / y] suffers from an integer overflow, even though the
* computation should be possible in the given type. Therefore, we use
* [x / y + !!(x % y)]. Note that on "Real CPUs" a division returns both the
* quotient and the remainder, so both should be equally fast. */
#define DIV_ROUND_UP(_x, _y) \
__extension__ ({ \
const typeof(_x) __x = (_x); \
const typeof(_y) __y = (_y); \
(__x / __y + !!(__x % __y)); \
})
#define assert_se(expr) \
do { \
if (_unlikely_(!(expr))) \

View File

@ -145,6 +145,39 @@ static void test_alloca(void) {
assert_se(!memcmp(t, zero, 997));
}
static void test_div_round_up(void) {
int div;
/* basic tests */
assert_se(DIV_ROUND_UP(0, 8) == 0);
assert_se(DIV_ROUND_UP(1, 8) == 1);
assert_se(DIV_ROUND_UP(8, 8) == 1);
assert_se(DIV_ROUND_UP(12, 8) == 2);
assert_se(DIV_ROUND_UP(16, 8) == 2);
/* test multiple evaluation */
div = 0;
assert_se(DIV_ROUND_UP(div++, 8) == 0 && div == 1);
assert_se(DIV_ROUND_UP(++div, 8) == 1 && div == 2);
assert_se(DIV_ROUND_UP(8, div++) == 4 && div == 3);
assert_se(DIV_ROUND_UP(8, ++div) == 2 && div == 4);
/* overflow test with exact division */
assert_se(sizeof(0U) == 4);
assert_se(0xfffffffaU % 10U == 0U);
assert_se(0xfffffffaU / 10U == 429496729U);
assert_se(DIV_ROUND_UP(0xfffffffaU, 10U) == 429496729U);
assert_se((0xfffffffaU + 10U - 1U) / 10U == 0U);
assert_se(0xfffffffaU / 10U + !!(0xfffffffaU % 10U) == 429496729U);
/* overflow test with rounded division */
assert_se(0xfffffffdU % 10U == 3U);
assert_se(0xfffffffdU / 10U == 429496729U);
assert_se(DIV_ROUND_UP(0xfffffffdU, 10U) == 429496730U);
assert_se((0xfffffffdU + 10U - 1U) / 10U == 0U);
assert_se(0xfffffffdU / 10U + !!(0xfffffffdU % 10U) == 429496730U);
}
static void test_first_word(void) {
assert_se(first_word("Hello", ""));
assert_se(first_word("Hello", "Hello"));
@ -1357,6 +1390,7 @@ int main(int argc, char *argv[]) {
test_max();
test_container_of();
test_alloca();
test_div_round_up();
test_first_word();
test_close_many();
test_parse_boolean();