about summary refs log tree commit diff
path: root/sysdeps/i386/i586
diff options
context:
space:
mode:
Diffstat (limited to 'sysdeps/i386/i586')
-rw-r--r--sysdeps/i386/i586/add_n.S133
-rw-r--r--sysdeps/i386/i586/addmul_1.S34
-rw-r--r--sysdeps/i386/i586/mul_1.S20
-rw-r--r--sysdeps/i386/i586/sub_n.S133
-rw-r--r--sysdeps/i386/i586/submul_1.S16
5 files changed, 160 insertions, 176 deletions
diff --git a/sysdeps/i386/i586/add_n.S b/sysdeps/i386/i586/add_n.S
index f52f9c60bc..f214c8cb36 100644
--- a/sysdeps/i386/i586/add_n.S
+++ b/sysdeps/i386/i586/add_n.S
@@ -1,7 +1,7 @@
 /* Pentium __mpn_add_n -- Add two limb vectors of the same length > 0 and store
    sum in a third limb vector.
 
-Copyright (C) 1992, 1994, 1995 Free Software Foundation, Inc.
+Copyright (C) 1992, 1994, 1995, 1996 Free Software Foundation, Inc.
 
 This file is part of the GNU MP Library.
 
@@ -30,13 +30,6 @@ the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
 #include "sysdep.h"
 #include "asm-syntax.h"
 
-#define t1	%eax
-#define t2	%edx
-#define src1	%esi
-#define src2	%ebp
-#define dst	%edi
-#define x	%ebx
-
 .text
 	ALIGN (3)
 	.globl C_SYMBOL_NAME(__mpn_add_n)
@@ -46,85 +39,85 @@ C_SYMBOL_NAME(__mpn_add_n:)
 	pushl	%ebx
 	pushl	%ebp
 
-	movl	20(%esp),dst		/* res_ptr */
-	movl	24(%esp),src1		/* s1_ptr */
-	movl	28(%esp),src2		/* s2_ptr */
+	movl	20(%esp),%edi		/* res_ptr */
+	movl	24(%esp),%esi		/* s1_ptr */
+	movl	28(%esp),%ebp		/* s2_ptr */
 	movl	32(%esp),%ecx		/* size */
 
-	movl	(src2),x
+	movl	(%ebp),%ebx
 
 	decl	%ecx
-	movl	%ecx,t2
+	movl	%ecx,%edx
 	shrl	$3,%ecx
-	andl	$7,t2
+	andl	$7,%edx
 	testl	%ecx,%ecx		/* zero carry flag */
 	jz	Lend
-	pushl	t2
+	pushl	%edx
 
 	ALIGN (3)
-Loop:	movl	28(dst),%eax		/* fetch destination cache line */
-	leal	32(dst),dst
-
-L1:	movl	(src1),t1
-	movl	4(src1),t2
-	adcl	x,t1
-	movl	4(src2),x
-	adcl	x,t2
-	movl	8(src2),x
-	movl	t1,-32(dst)
-	movl	t2,-28(dst)
-
-L2:	movl	8(src1),t1
-	movl	12(src1),t2
-	adcl	x,t1
-	movl	12(src2),x
-	adcl	x,t2
-	movl	16(src2),x
-	movl	t1,-24(dst)
-	movl	t2,-20(dst)
-
-L3:	movl	16(src1),t1
-	movl	20(src1),t2
-	adcl	x,t1
-	movl	20(src2),x
-	adcl	x,t2
-	movl	24(src2),x
-	movl	t1,-16(dst)
-	movl	t2,-12(dst)
-
-L4:	movl	24(src1),t1
-	movl	28(src1),t2
-	adcl	x,t1
-	movl	28(src2),x
-	adcl	x,t2
-	movl	32(src2),x
-	movl	t1,-8(dst)
-	movl	t2,-4(dst)
-
-	leal	32(src1),src1
-	leal	32(src2),src2
+Loop:	movl	28(%edi),%eax		/* fetch destination cache line */
+	leal	32(%edi),%edi
+
+L1:	movl	(%esi),%eax
+	movl	4(%esi),%edx
+	adcl	%ebx,%eax
+	movl	4(%ebp),%ebx
+	adcl	%ebx,%edx
+	movl	8(%ebp),%ebx
+	movl	%eax,-32(%edi)
+	movl	%edx,-28(%edi)
+
+L2:	movl	8(%esi),%eax
+	movl	12(%esi),%edx
+	adcl	%ebx,%eax
+	movl	12(%ebp),%ebx
+	adcl	%ebx,%edx
+	movl	16(%ebp),%ebx
+	movl	%eax,-24(%edi)
+	movl	%edx,-20(%edi)
+
+L3:	movl	16(%esi),%eax
+	movl	20(%esi),%edx
+	adcl	%ebx,%eax
+	movl	20(%ebp),%ebx
+	adcl	%ebx,%edx
+	movl	24(%ebp),%ebx
+	movl	%eax,-16(%edi)
+	movl	%edx,-12(%edi)
+
+L4:	movl	24(%esi),%eax
+	movl	28(%esi),%edx
+	adcl	%ebx,%eax
+	movl	28(%ebp),%ebx
+	adcl	%ebx,%edx
+	movl	32(%ebp),%ebx
+	movl	%eax,-8(%edi)
+	movl	%edx,-4(%edi)
+
+	leal	32(%esi),%esi
+	leal	32(%ebp),%ebp
 	decl	%ecx
 	jnz	Loop
 
-	popl	t2
+	popl	%edx
 Lend:
-	decl	t2			/* test t2 w/o clobbering carry */
+	decl	%edx			/* test %edx w/o clobbering carry */
 	js	Lend2
-	incl	t2
+	incl	%edx
 Loop2:
-	leal	4(dst),dst
-	movl	(src1),t1
-	adcl	x,t1
-	movl	4(src2),x
-	movl	t1,-4(dst)
-	leal	4(src1),src1
-	leal	4(src2),src2
-	decl	t2
+	leal	4(%edi),%edi
+	movl	(%esi),%eax
+	adcl	%ebx,%eax
+	movl	4(%ebp),%ebx
+	movl	%eax,-4(%edi)
+	leal	4(%esi),%esi
+	leal	4(%ebp),%ebp
+	decl	%edx
 	jnz	Loop2
 Lend2:
-	movl	(src1),t1
-	adcl	x,t1
-	movl	t1,(dst)
+	movl	(%esi),%eax
+	adcl	%ebx,%eax
+	movl	%eax,(%edi)
 
 	sbbl	%eax,%eax
 	negl	%eax
diff --git a/sysdeps/i386/i586/addmul_1.S b/sysdeps/i386/i586/addmul_1.S
index b222840591..5bf2603cab 100644
--- a/sysdeps/i386/i586/addmul_1.S
+++ b/sysdeps/i386/i586/addmul_1.S
@@ -1,7 +1,7 @@
 /* Pentium __mpn_addmul_1 -- Multiply a limb vector with a limb and add
    the result to a second limb vector.
 
-Copyright (C) 1992, 1994 Free Software Foundation, Inc.
+Copyright (C) 1992, 1994, 1996 Free Software Foundation, Inc.
 
 This file is part of the GNU MP Library.
 
@@ -32,12 +32,12 @@ the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
 
 #define res_ptr edi
 #define s1_ptr esi
+#define size ecx
 #define s2_limb ebp
 
 	TEXT
 	ALIGN (3)
 	GLOBL	C_SYMBOL_NAME(__mpn_addmul_1)
-	.type	C_SYMBOL_NAME(__mpn_addmul_1),@function
 C_SYMBOL_NAME(__mpn_addmul_1:)
 
 	INSN1(push,l	,R(edi))
@@ -47,38 +47,36 @@ C_SYMBOL_NAME(__mpn_addmul_1:)
 
 	INSN2(mov,l	,R(res_ptr),MEM_DISP(esp,20))
 	INSN2(mov,l	,R(s1_ptr),MEM_DISP(esp,24))
-	INSN2(mov,l	,R(ecx),MEM_DISP(esp,28))
+	INSN2(mov,l	,R(size),MEM_DISP(esp,28))
 	INSN2(mov,l	,R(s2_limb),MEM_DISP(esp,32))
 
-	INSN2(lea,l	,R(res_ptr),MEM_INDEX(res_ptr,ecx,4))
-	INSN2(lea,l	,R(s1_ptr),MEM_INDEX(s1_ptr,ecx,4))
-	INSN1(neg,l	,R(ecx))
-	INSN2(xor,l	,R(edx),R(edx))
+	INSN2(lea,l	,R(res_ptr),MEM_INDEX(res_ptr,size,4))
+	INSN2(lea,l	,R(s1_ptr),MEM_INDEX(s1_ptr,size,4))
+	INSN1(neg,l	,R(size))
+	INSN2(xor,l	,R(ebx),R(ebx))
 	ALIGN (3)
-Loop:
-	INSN2(mov,l	,R(ebx),R(edx))
-	INSN2(mov,l	,R(eax),MEM_INDEX(s1_ptr,ecx,4))
+
+Loop:	INSN2(adc,l	,R(ebx),$0)
+	INSN2(mov,l	,R(eax),MEM_INDEX(s1_ptr,size,4))
 
 	INSN1(mul,l	,R(s2_limb))
 
 	INSN2(add,l	,R(eax),R(ebx))
-	INSN2(mov,l	,R(ebx),MEM_INDEX(res_ptr,ecx,4))
+	INSN2(mov,l	,R(ebx),MEM_INDEX(res_ptr,size,4))
 
 	INSN2(adc,l	,R(edx),$0)
 	INSN2(add,l	,R(ebx),R(eax))
 
-	INSN2(adc,l	,R(edx),$0)
-	INSN2(mov,l	,MEM_INDEX(res_ptr,ecx,4),R(ebx))
+	INSN2(mov,l	,MEM_INDEX(res_ptr,size,4),R(ebx))
+	INSN1(inc,l	,R(size))
 
-	INSN1(inc,l	,R(ecx))
+	INSN2(mov,l	,R(ebx),R(edx))
 	INSN1(jnz,	,Loop)
 
-
-	INSN2(mov,l	,R(eax),R(edx))
+	INSN2(adc,l	,R(ebx),$0)
+	INSN2(mov,l	,R(eax),R(ebx))
 	INSN1(pop,l	,R(ebp))
 	INSN1(pop,l	,R(ebx))
 	INSN1(pop,l	,R(esi))
 	INSN1(pop,l	,R(edi))
 	ret
-Lfe1:
-	.size	C_SYMBOL_NAME(__mpn_addmul_1),Lfe1-C_SYMBOL_NAME(__mpn_addmul_1)
diff --git a/sysdeps/i386/i586/mul_1.S b/sysdeps/i386/i586/mul_1.S
index 2b7258e130..048c0601f2 100644
--- a/sysdeps/i386/i586/mul_1.S
+++ b/sysdeps/i386/i586/mul_1.S
@@ -1,7 +1,7 @@
 /* Pentium __mpn_mul_1 -- Multiply a limb vector with a limb and store
    the result in a second limb vector.
 
-Copyright (C) 1992, 1994 Free Software Foundation, Inc.
+Copyright (C) 1992, 1994, 1996 Free Software Foundation, Inc.
 
 This file is part of the GNU MP Library.
 
@@ -53,24 +53,24 @@ C_SYMBOL_NAME(__mpn_mul_1:)
 	INSN2(lea,l	,R(res_ptr),MEM_INDEX(res_ptr,size,4))
 	INSN2(lea,l	,R(s1_ptr),MEM_INDEX(s1_ptr,size,4))
 	INSN1(neg,l	,R(size))
-	INSN2(xor,l	,R(edx),R(edx))
+	INSN2(xor,l	,R(ebx),R(ebx))
 	ALIGN (3)
-Loop:
-	INSN2(mov,l	,R(ebx),R(edx))
+
+Loop:	INSN2(adc,l	,R(ebx),$0)
 	INSN2(mov,l	,R(eax),MEM_INDEX(s1_ptr,size,4))
 
 	INSN1(mul,l	,R(s2_limb))
 
-	INSN2(add,l	,R(eax),R(ebx))
-
-	INSN2(adc,l	,R(edx),$0)
-	INSN2(mov,l	,MEM_INDEX(res_ptr,size,4),R(eax))
+	INSN2(add,l	,R(ebx),R(eax))
 
+	INSN2(mov,l	,MEM_INDEX(res_ptr,size,4),R(ebx))
 	INSN1(inc,l	,R(size))
-	INSN1(jnz,	,Loop)
 
+	INSN2(mov,l	,R(ebx),R(edx))
+	INSN1(jnz,	,Loop)
 
-	INSN2(mov,l	,R(eax),R(edx))
+	INSN2(adc,l	,R(ebx),$0)
+	INSN2(mov,l	,R(eax),R(ebx))
 	INSN1(pop,l	,R(ebp))
 	INSN1(pop,l	,R(ebx))
 	INSN1(pop,l	,R(esi))
diff --git a/sysdeps/i386/i586/sub_n.S b/sysdeps/i386/i586/sub_n.S
index 9c964a82f3..cd158a5469 100644
--- a/sysdeps/i386/i586/sub_n.S
+++ b/sysdeps/i386/i586/sub_n.S
@@ -1,7 +1,7 @@
 /* Pentium __mpn_sub_n -- Subtract two limb vectors of the same length > 0
    and store difference in a third limb vector.
 
-Copyright (C) 1992, 1994, 1995 Free Software Foundation, Inc.
+Copyright (C) 1992, 1994, 1995, 1996 Free Software Foundation, Inc.
 
 This file is part of the GNU MP Library.
 
@@ -30,13 +30,6 @@ the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
 #include "sysdep.h"
 #include "asm-syntax.h"
 
-#define t1	%eax
-#define t2	%edx
-#define src1	%esi
-#define src2	%ebp
-#define dst	%edi
-#define x	%ebx
-
 .text
 	ALIGN (3)
 	.globl C_SYMBOL_NAME(__mpn_sub_n)
@@ -46,85 +39,85 @@ C_SYMBOL_NAME(__mpn_sub_n:)
 	pushl	%ebx
 	pushl	%ebp
 
-	movl	20(%esp),dst		/* res_ptr */
-	movl	24(%esp),src1		/* s1_ptr */
-	movl	28(%esp),src2		/* s2_ptr */
+	movl	20(%esp),%edi		/* res_ptr */
+	movl	24(%esp),%esi		/* s1_ptr */
+	movl	28(%esp),%ebp		/* s2_ptr */
 	movl	32(%esp),%ecx		/* size */
 
-	movl	(src2),x
+	movl	(%ebp),%ebx
 
 	decl	%ecx
-	movl	%ecx,t2
+	movl	%ecx,%edx
 	shrl	$3,%ecx
-	andl	$7,t2
+	andl	$7,%edx
 	testl	%ecx,%ecx		/* zero carry flag */
 	jz	Lend
-	pushl	t2
+	pushl	%edx
 
 	ALIGN (3)
-Loop:	movl	28(dst),%eax		/* fetch destination cache line */
-	leal	32(dst),dst
-
-L1:	movl	(src1),t1
-	movl	4(src1),t2
-	sbbl	x,t1
-	movl	4(src2),x
-	sbbl	x,t2
-	movl	8(src2),x
-	movl	t1,-32(dst)
-	movl	t2,-28(dst)
-
-L2:	movl	8(src1),t1
-	movl	12(src1),t2
-	sbbl	x,t1
-	movl	12(src2),x
-	sbbl	x,t2
-	movl	16(src2),x
-	movl	t1,-24(dst)
-	movl	t2,-20(dst)
-
-L3:	movl	16(src1),t1
-	movl	20(src1),t2
-	sbbl	x,t1
-	movl	20(src2),x
-	sbbl	x,t2
-	movl	24(src2),x
-	movl	t1,-16(dst)
-	movl	t2,-12(dst)
-
-L4:	movl	24(src1),t1
-	movl	28(src1),t2
-	sbbl	x,t1
-	movl	28(src2),x
-	sbbl	x,t2
-	movl	32(src2),x
-	movl	t1,-8(dst)
-	movl	t2,-4(dst)
-
-	leal	32(src1),src1
-	leal	32(src2),src2
+Loop:	movl	28(%edi),%eax		/* fetch destination cache line */
+	leal	32(%edi),%edi
+
+L1:	movl	(%esi),%eax
+	movl	4(%esi),%edx
+	sbbl	%ebx,%eax
+	movl	4(%ebp),%ebx
+	sbbl	%ebx,%edx
+	movl	8(%ebp),%ebx
+	movl	%eax,-32(%edi)
+	movl	%edx,-28(%edi)
+
+L2:	movl	8(%esi),%eax
+	movl	12(%esi),%edx
+	sbbl	%ebx,%eax
+	movl	12(%ebp),%ebx
+	sbbl	%ebx,%edx
+	movl	16(%ebp),%ebx
+	movl	%eax,-24(%edi)
+	movl	%edx,-20(%edi)
+
+L3:	movl	16(%esi),%eax
+	movl	20(%esi),%edx
+	sbbl	%ebx,%eax
+	movl	20(%ebp),%ebx
+	sbbl	%ebx,%edx
+	movl	24(%ebp),%ebx
+	movl	%eax,-16(%edi)
+	movl	%edx,-12(%edi)
+
+L4:	movl	24(%esi),%eax
+	movl	28(%esi),%edx
+	sbbl	%ebx,%eax
+	movl	28(%ebp),%ebx
+	sbbl	%ebx,%edx
+	movl	32(%ebp),%ebx
+	movl	%eax,-8(%edi)
+	movl	%edx,-4(%edi)
+
+	leal	32(%esi),%esi
+	leal	32(%ebp),%ebp
 	decl	%ecx
 	jnz	Loop
 
-	popl	t2
+	popl	%edx
 Lend:
-	decl	t2			/* test t2 w/o clobbering carry */
+	decl	%edx			/* test %edx w/o clobbering carry */
 	js	Lend2
-	incl	t2
+	incl	%edx
 Loop2:
-	leal	4(dst),dst
-	movl	(src1),t1
-	sbbl	x,t1
-	movl	4(src2),x
-	movl	t1,-4(dst)
-	leal	4(src1),src1
-	leal	4(src2),src2
-	decl	t2
+	leal	4(%edi),%edi
+	movl	(%esi),%eax
+	sbbl	%ebx,%eax
+	movl	4(%ebp),%ebx
+	movl	%eax,-4(%edi)
+	leal	4(%esi),%esi
+	leal	4(%ebp),%ebp
+	decl	%edx
 	jnz	Loop2
 Lend2:
-	movl	(src1),t1
-	sbbl	x,t1
-	movl	t1,(dst)
+	movl	(%esi),%eax
+	sbbl	%ebx,%eax
+	movl	%eax,(%edi)
 
 	sbbl	%eax,%eax
 	negl	%eax
diff --git a/sysdeps/i386/i586/submul_1.S b/sysdeps/i386/i586/submul_1.S
index 14bfe54e24..440f64f358 100644
--- a/sysdeps/i386/i586/submul_1.S
+++ b/sysdeps/i386/i586/submul_1.S
@@ -1,7 +1,7 @@
 /* Pentium __mpn_submul_1 -- Multiply a limb vector with a limb and subtract
    the result from a second limb vector.
 
-Copyright (C) 1992, 1994 Free Software Foundation, Inc.
+Copyright (C) 1992, 1994, 1996 Free Software Foundation, Inc.
 
 This file is part of the GNU MP Library.
 
@@ -53,10 +53,10 @@ C_SYMBOL_NAME(__mpn_submul_1:)
 	INSN2(lea,l	,R(res_ptr),MEM_INDEX(res_ptr,size,4))
 	INSN2(lea,l	,R(s1_ptr),MEM_INDEX(s1_ptr,size,4))
 	INSN1(neg,l	,R(size))
-	INSN2(xor,l	,R(edx),R(edx))
+	INSN2(xor,l	,R(ebx),R(ebx))
 	ALIGN (3)
-Loop:
-	INSN2(mov,l	,R(ebx),R(edx))
+
+Loop:	INSN2(adc,l	,R(ebx),$0)
 	INSN2(mov,l	,R(eax),MEM_INDEX(s1_ptr,size,4))
 
 	INSN1(mul,l	,R(s2_limb))
@@ -67,14 +67,14 @@ Loop:
 	INSN2(adc,l	,R(edx),$0)
 	INSN2(sub,l	,R(ebx),R(eax))
 
-	INSN2(adc,l	,R(edx),$0)
 	INSN2(mov,l	,MEM_INDEX(res_ptr,size,4),R(ebx))
-
 	INSN1(inc,l	,R(size))
-	INSN1(jnz,	,Loop)
 
+	INSN2(mov,l	,R(ebx),R(edx))
+	INSN1(jnz,	,Loop)
 
-	INSN2(mov,l	,R(eax),R(edx))
+	INSN2(adc,l	,R(ebx),$0)
+	INSN2(mov,l	,R(eax),R(ebx))
 	INSN1(pop,l	,R(ebp))
 	INSN1(pop,l	,R(ebx))
 	INSN1(pop,l	,R(esi))