RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-31 15:48:12 +00:00
Code Issues Activity Actions

Add a fast memcpy() using MMX when we're moving >= 1KB.

Browse source 
This is a nice speedup for WindowServer. I'll eventually have to do this
with SSE but the kernel doesn't support SSE yet so this is it for now.
This commit is contained in:
Andreas Kling 2019-02-07 08:46:52 +01:00
parent e29060620f
commit 1f159eaab0
6 changed files with 70 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

50
AK/StdLibExtras.cpp Normal file
View file

@ -0,0 +1,50 @@
#include <AK/StdLibExtras.h>
#include <AK/Assertions.h>
#include <AK/Types.h>
#include <AK/kstdio.h>
void* mmx_memcpy(void* dest, const void* src, size_t len)
{
ASSERT(len >= 1024);
auto* dest_ptr = (byte*)dest;
auto* src_ptr = (const byte*)src;
if ((dword)dest_ptr & 7) {
dword prologue = 8 - ((dword)dest_ptr & 7);
asm volatile(
"rep movsb\n"
:: "S"(src_ptr), "D"(dest_ptr), "c"(prologue)
: "memory"
);
len -= prologue;
}
for (dword i = len / 64; i; --i) {
asm volatile(
"movq (%0), %%mm0\n"
"movq 8(%0), %%mm1\n"
"movq 16(%0), %%mm2\n"
"movq 24(%0), %%mm3\n"
"movq 32(%0), %%mm4\n"
"movq 40(%0), %%mm5\n"
"movq 48(%0), %%mm6\n"
"movq 56(%0), %%mm7\n"
"movq %%mm0, (%1)\n"
"movq %%mm1, 8(%1)\n"
"movq %%mm2, 16(%1)\n"
"movq %%mm3, 24(%1)\n"
"movq %%mm4, 32(%1)\n"
"movq %%mm5, 40(%1)\n"
"movq %%mm6, 48(%1)\n"
"movq %%mm7, 56(%1)\n"
:: "r" (src_ptr), "r" (dest_ptr) : "memory");
src_ptr += 64;
dest_ptr += 64;
}
asm volatile("emms":::"memory");
// Whatever remains we'll have to memcpy.
len %= 64;
if (len)
memcpy(dest_ptr, src_ptr, len);
return dest;
}