/* The following code tries to optimize the copy by using unsigned *alignment.Thiswillworkfineifbothsourceanddestinationare *alignedonthesameboundary.However,iftheyarealignedon *differentboundariesshiftswillbenecessary.Thismightresultin *badperformanceonMicroBlazesystemswithoutabarrelshifter.
*/ const uint32_t *i_src;
uint32_t *i_dst;
if (likely(c >= 4)) { unsigned value, buf_hold;
/* Align the destination to a word boundary. */ /* This is done in an endian independent manner. */ switch ((unsignedlong)dst & 3) { case1:
*dst++ = *src++;
--c;
fallthrough; case2:
*dst++ = *src++;
--c;
fallthrough; case3:
*dst++ = *src++;
--c;
}
i_dst = (void *)dst;
/* Choose a copy scheme based on the source */ /* alignment relative to destination. */ switch ((unsignedlong)src & 3) { case0x0: /* Both byte offsets are aligned */
i_src = (constvoid *)src;
for (; c >= 4; c -= 4)
*i_dst++ = *i_src++;
src = (constvoid *)i_src; break; case0x1: /* Unaligned - Off by 1 */ /* Word align the source */
i_src = (constvoid *) ((unsigned)src & ~3); #ifndef __MICROBLAZEEL__ /* Load the holding buffer */
buf_hold = *i_src++ << 8;
for (; c >= 4; c -= 4) {
value = *i_src++;
*i_dst++ = buf_hold | value >> 24;
buf_hold = value << 8;
} #else /* Load the holding buffer */
buf_hold = (*i_src++ & 0xFFFFFF00) >> 8;
for (; c >= 4; c -= 4) {
value = *i_src++;
*i_dst++ = buf_hold | ((value & 0xFF) << 24);
buf_hold = (value & 0xFFFFFF00) >> 8;
} #endif /* Realign the source */
src = (constvoid *)i_src;
src -= 3; break; case0x2: /* Unaligned - Off by 2 */ /* Word align the source */
i_src = (constvoid *) ((unsigned)src & ~3); #ifndef __MICROBLAZEEL__ /* Load the holding buffer */
buf_hold = *i_src++ << 16;
for (; c >= 4; c -= 4) {
value = *i_src++;
*i_dst++ = buf_hold | value >> 16;
buf_hold = value << 16;
} #else /* Load the holding buffer */
buf_hold = (*i_src++ & 0xFFFF0000) >> 16;
for (; c >= 4; c -= 4) {
value = *i_src++;
*i_dst++ = buf_hold | ((value & 0xFFFF) << 16);
buf_hold = (value & 0xFFFF0000) >> 16;
} #endif /* Realign the source */
src = (constvoid *)i_src;
src -= 2; break; case0x3: /* Unaligned - Off by 3 */ /* Word align the source */
i_src = (constvoid *) ((unsigned)src & ~3); #ifndef __MICROBLAZEEL__ /* Load the holding buffer */
buf_hold = *i_src++ << 24;
for (; c >= 4; c -= 4) {
value = *i_src++;
*i_dst++ = buf_hold | value >> 8;
buf_hold = value << 24;
} #else /* Load the holding buffer */
buf_hold = (*i_src++ & 0xFF000000) >> 24;
for (; c >= 4; c -= 4) {
value = *i_src++;
*i_dst++ = buf_hold | ((value & 0xFFFFFF) << 8);
buf_hold = (value & 0xFF000000) >> 24;
} #endif /* Realign the source */
src = (constvoid *)i_src;
src -= 1; break;
}
dst = (void *)i_dst;
}
/* Finish off any remaining bytes */ /* simple fast copy, ... unless a cache boundary is crossed */ switch (c) { case3:
*dst++ = *src++;
fallthrough; case2:
*dst++ = *src++;
fallthrough; case1:
*dst++ = *src++;
}
Die Informationen auf dieser Webseite wurden
nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit,
noch Qualität der bereit gestellten Informationen zugesichert.
Bemerkung:
Die farbliche Syntaxdarstellung und die Messung sind noch experimentell.