amd64: mostly depessimize copystr

[mjg]

Summary

The asm version should probably get removed and be re-coded in C. It can however work as a template for copyinstr changes later.

commit message;

amd64: mostly depessimize copystr

• remove a forward branch in the common case
• replace xchg + lodsb/stosb loop with simple movs

A simple test on Intel(R) Core(TM) i7-4600U CPU @ 2.10GH copying /foo/bar/baz in a loop
goes from 295715863 ops/s to 465807408.

Test Plan

verified the same result with:

#include <string.h>
#include <stdio.h>

int     copystr(const void * kfaddr,
            void * kdaddr, size_t len,
            size_t * lencopied);


int
main(void)
{
	char src[10], dst[10];
	int i, r;
	size_t c;

	strcpy(src, "abcde");
	for (i = 0; i < 10; i++) {
		memset(dst, 'A', sizeof(dst));
		r = copystr(src, dst, i, &c);
		printf("i=%d r=%d c=%ld dst=[%10s]\n", i, r, c, dst);
	}
	return (0);
}

[kib]

sys/amd64/amd64/support.S

1030 ↗

(On Diff #48338)

movb

1034 ↗

(On Diff #48338)

testb