From 2f8cca2d6daa1422cd4a1286bc0d42e29145bdf1 Mon Sep 17 00:00:00 2001 From: Balazs Gerofi Date: Wed, 23 Nov 2016 08:51:22 +0900 Subject: [PATCH] memcpy(): faster version using ASM rep; movsl --- arch/x86/kernel/include/arch-string.h | 23 +++++++++++++++++++++++ kernel/syscall.c | 9 +++++---- lib/include/string.h | 7 +++++++ 3 files changed, 35 insertions(+), 4 deletions(-) create mode 100644 arch/x86/kernel/include/arch-string.h diff --git a/arch/x86/kernel/include/arch-string.h b/arch/x86/kernel/include/arch-string.h new file mode 100644 index 00000000..c80ab1ea --- /dev/null +++ b/arch/x86/kernel/include/arch-string.h @@ -0,0 +1,23 @@ +#ifndef _ASM_X86_STRING_H +#define _ASM_X86_STRING_H + +#define ARCH_FAST_MEMCPY + +static inline void *__inline_memcpy(void *to, const void *from, size_t n) +{ + unsigned long d0, d1, d2; + asm volatile("rep ; movsl\n\t" + "testb $2,%b4\n\t" + "je 1f\n\t" + "movsw\n" + "1:\ttestb $1,%b4\n\t" + "je 2f\n\t" + "movsb\n" + "2:" + : "=&c" (d0), "=&D" (d1), "=&S" (d2) + : "0" (n / 4), "q" (n), "1" ((long)to), "2" ((long)from) + : "memory"); + return to; +} + +#endif diff --git a/kernel/syscall.c b/kernel/syscall.c index 91648158..fd026107 100644 --- a/kernel/syscall.c +++ b/kernel/syscall.c @@ -7909,10 +7909,11 @@ retry_lookup: } rva = phys_to_virt(rphys); - - memcpy((op == PROCESS_VM_READ) ? local_iov[li].iov_base + loff : rva, - (op == PROCESS_VM_READ) ? rva : local_iov[li].iov_base + loff, - to_copy); + + fast_memcpy( + (op == PROCESS_VM_READ) ? local_iov[li].iov_base + loff : rva, + (op == PROCESS_VM_READ) ? rva : local_iov[li].iov_base + loff, + to_copy); copied += to_copy; dkprintf("local_iov[%d]: 0x%lx %s remote_iov[%d]: 0x%lx, %lu copied, psize: %lu, rpage_left: %lu\n", diff --git a/lib/include/string.h b/lib/include/string.h index 74044102..948eecb5 100644 --- a/lib/include/string.h +++ b/lib/include/string.h @@ -14,6 +14,7 @@ #define __STRING_H #include +#include size_t strlen(const char *p); size_t strnlen(const char *p, size_t maxlen); @@ -29,6 +30,12 @@ void *memcpy_long(void *dest, const void *src, size_t n); int memcmp(const void *s1, const void *s2, size_t n); void *memset(void *s, int n, size_t l); +#ifdef ARCH_FAST_MEMCPY +#define fast_memcpy __inline_memcpy +#else +#define fast_memcpy memcpy +#endif + extern int snprintf(char * buf, size_t size, const char *fmt, ...); extern int sprintf(char * buf, const char *fmt, ...); extern int sscanf(const char * buf, const char * fmt, ...);