djsundog/circuitpython
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

stmhal: Slightly improved memcpy; memset uses word store when aligned.

Browse Source
This commit is contained in:
Damien George 2014-09-13 00:12:41 +01:00
parent 5792500ccc
commit 32781cce6d
1 changed files with 36 additions and 30 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

66
stmhal/string0.c
View File

@ -30,40 +30,31 @@
#define likely(x) __builtin_expect((x), 1) #define likely(x) __builtin_expect((x), 1)
void *memcpy(void *dst, const void *src, size_t n) { void *memcpy(void *dst, const void *src, size_t n) {
if (likely(!((long)dst&3) && !((long)src&3))) { if (likely(!(((uint32_t)dst) & 3) && !(((uint32_t)src) & 3))) {
//copy words from aligned pointers first // pointers aligned
long *d = dst; uint32_t *d = dst;
const long *s = src; const uint32_t *s = src;
for (int i=(n>>2); i; i--) { // copy words first
for (size_t i = (n >> 2); i; i--) {
*d++ = *s++; *d++ = *s++;
} }
//copy remaining bytes if (n & 2) {
if (n&3) { // copy half-word
char *d8 = (char*)d; *(uint16_t*)d = *(const uint16_t*)s;
const char *s8 =(char*) s; d = (uint32_t*)((uint16_t*)d + 1);
s = (const uint32_t*)((const uint16_t*)s + 1);
switch (n&3) {
case 1:
*d8=*s8;
break;
case 2:
*d8++=*s8++;
*d8=*s8;
break;
case 3:
*d8++=*s8++;
*d8++=*s8++;
*d8=*s8;
break;
}
} }
if (n & 1) {
// copy byte
*((uint8_t*)d) = *((const uint8_t*)s);
}
} else { } else {
//unaligned access, copy bytes // unaligned access, copy bytes
char *d = dst; uint8_t *d = dst;
const char *s = src; const uint8_t *s = src;
for (; n; n--) { for (; n; n--) {
*d++ = *s++; *d++ = *s++;
@ -89,9 +80,24 @@ void *memmove(void *dest, const void *src, size_t n) {
} }
void *memset(void *s, int c, size_t n) { void *memset(void *s, int c, size_t n) {
uint8_t *s2 = s; if (c == 0 && ((uint32_t)s & 3) == 0) {
for (; n > 0; n--) { // aligned store of 0
*s2++ = c; uint32_t *s32 = s;
for (size_t i = n >> 2; i > 0; i--) {
*s32++ = 0;
}
if (n & 2) {
*((uint16_t*)s32) = 0;
s32 = (uint32_t*)((uint16_t*)s32 + 1);
}
if (n & 1) {
*((uint8_t*)s32) = 0;
}
} else {
uint8_t *s2 = s;
for (; n > 0; n--) {
*s2++ = c;
}
} }
return s; return s;
} }