kiper220/mimalloc
1
0
Fork 0
mirror of https://github.com/microsoft/mimalloc.git synced 2025-07-06 19:38:41 +03:00
Code Issues Projects Releases Packages Wiki Activity Actions

merge from dev

Browse source
This commit is contained in:
Daan Leijen 2021-01-30 16:37:38 -08:00
commit 36b7a3cb03
6 changed files with 79 additions and 26 deletions
Download patch file Download diff file Expand all files Collapse all files

66
include/mimalloc-internal.h
View file

@ -180,21 +180,6 @@ bool _mi_page_is_valid(mi_page_t* page);
#endif
// -----------------------------------------------------------------------------------
// On windows x86/x64 with msvc/clang-cl, use `rep movsb` for `memcpy` (issue #201)
// -----------------------------------------------------------------------------------
#if defined(_WIN32) && (defined(_M_IX86) || defined(_M_X64))
#include <intrin.h>
static inline void _mi_memcpy_rep_movsb(void* d, const void* s, size_t n) {
__movsb((unsigned char*)d, (const unsigned char*)s, n);
}
#define _mi_memcpy(d,s,n) _mi_memcpy_rep_movsb(d,s,n)
#else
#define _mi_memcpy(d,s,n) memcpy(d,s,n)
#endif
/* -----------------------------------------------------------
Inlined definitions
----------------------------------------------------------- */
@ -997,4 +982,55 @@ static inline size_t mi_bsr(uintptr_t x) {
}
// ---------------------------------------------------------------------------------
// Provide our own `_mi_memcpy` for potential performance optimizations.
//
// For now, only on Windows with msvc/clang-cl we optimize to `rep movsb` if
// we happen to run on x86/x64 cpu's that have "fast short rep movsb" (FSRM) support
// (AMD Zen3+ (~2020) or Intel Ice Lake+ (~2017). See also issue #201 and pr #253.
// ---------------------------------------------------------------------------------
#if defined(_WIN32) && (defined(_M_IX86) || defined(_M_X64))
#include <intrin.h>
#include <string.h>
extern bool _mi_cpu_has_fsrm;
static inline void _mi_memcpy(void* dst, const void* src, size_t n) {
if (_mi_cpu_has_fsrm) {
__movsb((unsigned char*)dst, (const unsigned char*)src, n);
}
else {
memcpy(dst, src, n); // todo: use noinline?
}
}
#else
#include <string.h>
static inline void _mi_memcpy(void* dst, const void* src, size_t n) {
memcpy(dst, src, n);
}
#endif
// -------------------------------------------------------------------------------
// The `_mi_memcpy_aligned` can be used if the pointers are machine-word aligned
// This is used for example in `mi_realloc`.
// -------------------------------------------------------------------------------
#if (__GNUC__ >= 4) || defined(__clang__)
// On GCC/CLang we provide a hint that the pointers are word aligned.
#include <string.h>
static inline void _mi_memcpy_aligned(void* dst, const void* src, size_t n) {
mi_assert_internal(((uintptr_t)dst % MI_INTPTR_SIZE == 0) && ((uintptr_t)src % MI_INTPTR_SIZE == 0));
void* adst = __builtin_assume_aligned(dst, MI_INTPTR_SIZE);
const void* asrc = __builtin_assume_aligned(src, MI_INTPTR_SIZE);
memcpy(adst, asrc, n);
}
#else
// Default fallback on `_mi_memcpy`
static inline void _mi_memcpy_aligned(void* dst, const void* src, size_t n) {
mi_assert_internal(((uintptr_t)dst % MI_INTPTR_SIZE == 0) && ((uintptr_t)src % MI_INTPTR_SIZE == 0));
_mi_memcpy(dst, src, n);
}
#endif
#endif