X-Git-Url: https://git.ucc.asn.au/?a=blobdiff_plain;f=Kernel%2Farch%2Fx86%2Flib.c;h=c499c7a8ed06be6f664f8b39663bc355c1b7fbaa;hb=3887bf8941fe33071d061961f213a7da360af02a;hp=2334417c787b43ac5090d438e4a93608164ef327;hpb=8bc40333b1401d7616b225945fee53d972c2f418;p=tpg%2Facess2.git

diff --git a/Kernel/arch/x86/lib.c b/Kernel/arch/x86/lib.c
index 2334417c..c499c7a8 100644
--- a/Kernel/arch/x86/lib.c
+++ b/Kernel/arch/x86/lib.c
@@ -2,7 +2,7 @@
  * AcessOS Microkernel Version
  * lib.c
  */
-#include <common.h>
+#include <acess.h>
 
 // === CODE ===
 void Spinlock(int *lock)
@@ -49,36 +49,67 @@ Uint32 ind(Uint16 Port)
 }
 
 /**
- * \fn void *memset(void *Dest, int Val, Uint Num)
- * \brief Do a byte set of Dest
+ * \fn void *memset(void *Dest, int Val, size_t Num)
+ * \brief Do a byte granuality set of Dest
  */
-void *memset(void *Dest, int Val, Uint Num)
+void *memset(void *Dest, int Val, size_t Num)
 {
-	__asm__ __volatile__ ("rep stosb" :: "D" (Dest), "a" (Val), "c" (Num));
+	Uint32	val = Val&0xFF;
+	val |= val << 8;
+	val |= val << 16;
+	__asm__ __volatile__ (
+		"rep stosl;\n\t"
+		"mov %3, %%ecx;\n\t"
+		"rep stosb"
+		:: "D" (Dest), "a" (val), "c" (Num/4), "r" (Num&3));
 	return Dest;
 }
 /**
- * \fn void *memsetd(void *Dest, Uint Val, Uint Num)
+ * \brief Set double words
  */
-void *memsetd(void *Dest, Uint Val, Uint Num)
+void *memsetd(void *Dest, Uint32 Val, size_t Num)
 {
 	__asm__ __volatile__ ("rep stosl" :: "D" (Dest), "a" (Val), "c" (Num));
 	return Dest;
 }
 
+/**
+ * \fn int memcmp(const void *m1, const void *m2, size_t Num)
+ * \brief Compare two pieces of memory
+ */
+int memcmp(const void *m1, const void *m2, size_t Num)
+{
+	while(Num--)
+	{
+		if(*(Uint8*)m1 != *(Uint8*)m2)	break;
+		m1 ++;
+		m2 ++;
+	}
+	return *(Uint8*)m1 - *(Uint8*)m2;
+}
 
 /**
- * \fn void *memcpy(void *Dest, void *Src, Uint Num)
+ * \fn void *memcpy(void *Dest, const void *Src, size_t Num)
+ * \brief Copy \a Num bytes from \a Src to \a Dest
  */
-void *memcpy(void *Dest, void *Src, Uint Num)
+void *memcpy(void *Dest, const void *Src, size_t Num)
 {
-	__asm__ __volatile__ ("rep movsb" :: "D" (Dest), "S" (Src), "c" (Num));
+	if( ((Uint)Dest & 3) || ((Uint)Src & 3) )
+		__asm__ __volatile__ ("rep movsb" :: "D" (Dest), "S" (Src), "c" (Num));
+	else {
+		__asm__ __volatile__ (
+			"rep movsl;\n\t"
+			"mov %3, %%ecx;\n\t"
+			"rep movsb"
+			:: "D" (Dest), "S" (Src), "c" (Num/4), "r" (Num&3));
+	}
 	return Dest;
 }
 /**
- * \fn void *memcpyd(void *Dest, void *Src, Uint Num)
+ * \fn void *memcpyd(void *Dest, const void *Src, size_t Num)
+ * \brief Copy \a Num DWORDs from \a Src to \a Dest
  */
-void *memcpyd(void *Dest, void *Src, Uint Num)
+void *memcpyd(void *Dest, const void *Src, size_t Num)
 {
 	__asm__ __volatile__ ("rep movsl" :: "D" (Dest), "S" (Src), "c" (Num));
 	return Dest;
@@ -90,10 +121,15 @@ void *memcpyd(void *Dest, void *Src, Uint Num)
  */
 Uint64 __udivdi3(Uint64 Num, Uint64 Den)
 {
-	Uint64	ret = 0;
+	Uint64	P[2];
+	Uint64	q = 0;
+	 int	i;
 	
-	if(Den == 0)	__asm__ __volatile__ ("int $0x0");	// Call Div by Zero Error
-	if(Den == 1)	return Num;	// Speed Hacks
+	if(Den == 0)	__asm__ __volatile__ ("int $0x0");
+	// Common speedups
+	if(Num <= 0xFFFFFFFF && Den <= 0xFFFFFFFF)
+		return (Uint32)Num / (Uint32)Den;
+	if(Den == 1)	return Num;
 	if(Den == 2)	return Num >> 1;	// Speed Hacks
 	if(Den == 4)	return Num >> 2;	// Speed Hacks
 	if(Den == 8)	return Num >> 3;	// Speed Hacks
@@ -101,19 +137,34 @@ Uint64 __udivdi3(Uint64 Num, Uint64 Den)
 	if(Den == 32)	return Num >> 5;	// Speed Hacks
 	if(Den == 1024)	return Num >> 10;	// Speed Hacks
 	if(Den == 2048)	return Num >> 11;	// Speed Hacks
+	if(Den == 4096)	return Num >> 12;
+	if(Num < Den)	return 0;
+	if(Num < Den*2)	return 1;
+	if(Num == Den*2)	return 2;
 	
-	if(Num >> 32 == 0 && Den >> 32 == 0)
-		return (Uint32)Num / (Uint32)Den;
-	
-	//Log("__udivdi3: (Num={0x%x:%x}, Den={0x%x:%x})",
-	//	Num>>32, Num&0xFFFFFFFF,
-	//	Den>>32, Den&0xFFFFFFFF);
-	
-	while(Num > Den) {
-		ret ++;
-		Num -= Den;
+	// Restoring division, from wikipedia
+	// http://en.wikipedia.org/wiki/Division_(digital)
+	P[0] = Num;	P[1] = 0;
+	for( i = 64; i--; )
+	{
+		// P <<= 1;
+		P[1] = (P[1] << 1) | (P[0] >> 63);
+		P[0] = P[0] << 1;
+		
+		// P -= Den << 64
+		P[1] -= Den;
+		
+		// P >= 0
+		if( !(P[1] & (1ULL<<63)) ) {
+			q |= (Uint64)1 << (63-i);
+		}
+		else {
+			//q |= 0 << (63-i);
+			P[1] += Den;
+		}
 	}
-	return ret;
+	
+	return q;
 }
 
 /**
@@ -131,19 +182,39 @@ Uint64 __umoddi3(Uint64 Num, Uint64 Den)
 	if(Den == 32)	return Num & 31;	// Speed Hacks
 	if(Den == 1024)	return Num & 1023;	// Speed Hacks
 	if(Den == 2048)	return Num & 2047;	// Speed Hacks
+	if(Den == 4096)	return Num & 4095;	// Speed Hacks
 	
 	if(Num >> 32 == 0 && Den >> 32 == 0)
 		return (Uint32)Num % (Uint32)Den;
 	
-	while(Num > Den)
-		Num -= Den;
-	return Num;
+	return Num - __udivdi3(Num, Den) * Den;
+}
+
+Uint16 LittleEndian16(Uint16 Val)
+{
+	return Val;
+}
+Uint16 BigEndian16(Uint16 Val)
+{
+	return ((Val&0xFF)<<8) | ((Val>>8)&0xFF);
+}
+Uint32 LittleEndian32(Uint32 Val)
+{
+	return Val;
+}
+Uint32 BigEndian32(Uint32 Val)
+{
+	return ((Val&0xFF)<<24) | ((Val&0xFF00)<<8) | ((Val>>8)&0xFF00) | ((Val>>24)&0xFF);
 }
 
 // --- EXPORTS ---
 EXPORT(memcpy);	EXPORT(memset);
+EXPORT(memcmp);
 //EXPORT(memcpyw);	EXPORT(memsetw);
 EXPORT(memcpyd);	EXPORT(memsetd);
 EXPORT(inb);	EXPORT(inw);	EXPORT(ind);
 EXPORT(outb);	EXPORT(outw);	EXPORT(outd);
 EXPORT(__udivdi3);	EXPORT(__umoddi3);
+
+EXPORT(LittleEndian16);	EXPORT(BigEndian16);
+EXPORT(LittleEndian32);	EXPORT(BigEndian32);