From: MASM fan <masmfan@gmail.com>
Date: Tue, 25 Nov 2014 20:38:28 +0000 (+0400)
Subject: GNU assembler compatibility
X-Git-Tag: nvc-v0.5.0~21
X-Git-Url: https://git.novaco.in/?a=commitdiff_plain;h=3857b51867976ea13bb578614518392a8b07f2a2;hp=734daea5c4dd864e2b86de555c4b628a8cfe856a;p=novacoin.git

GNU assembler compatibility
---

diff --git a/src/scrypt-arm.S b/src/scrypt-arm.S
index 12d94b0..65b9c7f 100644
--- a/src/scrypt-arm.S
+++ b/src/scrypt-arm.S
@@ -7,429 +7,284 @@
  * any later version.  See COPYING for more details.
  */
 
-
 #if defined(__arm__) && defined(__APCS_32__)
 
-#if defined(__linux__) && defined(__ELF__)
-.section .note.GNU-stack,"",%progbits
-#endif
-
-#if defined(__ARM_ARCH_5E__) || defined(__ARM_ARCH_5TE__) || \
-	defined(__ARM_ARCH_5TEJ__) || defined(__ARM_ARCH_6__) || \
-	defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || \
-	defined(__ARM_ARCH_6M__) || defined(__ARM_ARCH_6T2__) || \
-	defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__)
-#define __ARM_ARCH_5E_OR_6__
-#endif
-
-#if defined(__ARM_ARCH_5E_OR_6__) || defined(__ARM_ARCH_7__) || \
-	defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || \
-	defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7EM__)
-#define __ARM_ARCH_5E_OR_6_OR_7__
-#endif
-
-#ifdef __ARM_ARCH_5E_OR_6__
-
-#define scrypt_shuffle() \
-	add	lr, r0, #9*4; \
-	ldmia	r0, {r2-r7}; \
-	ldmia	lr, {r2, r8-r12, lr}; \
-	str	r3, [r0, #5*4]; \
-	str	r5, [r0, #15*4]; \
-	str	r6, [r0, #12*4]; \
-	str	r7, [r0, #1*4]; \
-	ldr r5, [r0, #7*4]; \
-	str	r2, [r0, #13*4]; \
-	str	r8, [r0, #2*4]; \
-	strd	r4, [r0, #10*4]; \
-	str	r9, [r0, #7*4]; \
-	str	r10, [r0, #4*4]; \
-	str	r11, [r0, #9*4]; \
-	str	lr, [r0, #3*4]; \
-	add	r2, r0, #64+0*4; \
-	add	lr, r0, #64+9*4; \
-	ldmia	r2, {r2-r7}; \
-	ldmia	lr, {r2, r8-r12, lr}; \
-	str	r3, [r0, #64+5*4]; \
-	str	r5, [r0, #64+15*4]; \
-	str	r6, [r0, #64+12*4]; \
-	str	r7, [r0, #64+1*4]; \
-	ldr r5, [r0, #64+7*4]; \
-	str	r2, [r0, #64+13*4]; \
-	str	r8, [r0, #64+2*4]; \
-	strd	r4, [r0, #64+10*4]; \
-	str	r9, [r0, #64+7*4]; \
-	str	r10, [r0, #64+4*4]; \
-	str	r11, [r0, #64+9*4]; \
-	str	lr, [r0, #64+3*4]; \
-
-
-#define salsa8_core_doubleround_body() \
-	add	r6, r2, r6; \
-	add	r7, r3, r7; \
-	eor	r10, r10, r6, ror #25; \
-	add	r6, r0, r4; \
-	eor	r11, r11, r7, ror #25; \
-	add	r7, r1, r5; \
-	strd	r10, [sp, #14*4]; \
-	eor	r12, r12, r6, ror #25; \
-	eor	lr, lr, r7, ror #25; \
-	ldrd	r6, [sp, #10*4]; \
-	add	r2, r10, r2; \
-	add	r3, r11, r3; \
-	eor	r6, r6, r2, ror #23; \
-	add	r2, r12, r0; \
-	eor	r7, r7, r3, ror #23; \
-	add	r3, lr, r1; \
-	strd	r6, [sp, #10*4]; \
-	eor	r8, r8, r2, ror #23; \
-	eor	r9, r9, r3, ror #23; \
-	ldrd	r2, [sp, #6*4]; \
-	add	r10, r6, r10; \
-	add	r11, r7, r11; \
-	eor	r2, r2, r10, ror #19; \
-	add	r10, r8, r12; \
-	eor	r3, r3, r11, ror #19; \
-	add	r11, r9, lr; \
-	eor	r4, r4, r10, ror #19; \
-	eor	r5, r5, r11, ror #19; \
-	ldrd	r10, [sp, #2*4]; \
-	add	r6, r2, r6; \
-	add	r7, r3, r7; \
-	eor	r10, r10, r6, ror #14; \
-	add	r6, r4, r8; \
-	eor	r11, r11, r7, ror #14; \
-	add	r7, r5, r9; \
-	eor	r0, r0, r6, ror #14; \
-	eor	r1, r1, r7, ror #14; \
-	ldrd	r6, [sp, #14*4]; \
-	strd	r2, [sp, #6*4]; \
-	strd	r10, [sp, #2*4]; \
-	add	r6, r11, r6; \
-	add	r7, r0, r7; \
-	eor	r4, r4, r6, ror #25; \
-	add	r6, r1, r12; \
-	eor	r5, r5, r7, ror #25; \
-	add	r7, r10, lr; \
-	eor	r2, r2, r6, ror #25; \
-	eor	r3, r3, r7, ror #25; \
-	strd	r2, [sp, #6*4]; \
-	add	r10, r3, r10; \
-	ldrd	r6, [sp, #10*4]; \
-	add	r11, r4, r11; \
-	eor	r8, r8, r10, ror #23; \
-	add	r10, r5, r0; \
-	eor	r9, r9, r11, ror #23; \
-	add	r11, r2, r1; \
-	eor	r6, r6, r10, ror #23; \
-	eor	r7, r7, r11, ror #23; \
-	strd	r6, [sp, #10*4]; \
-	add	r2, r7, r2; \
-	ldrd	r10, [sp, #14*4]; \
-	add	r3, r8, r3; \
-	eor	r12, r12, r2, ror #19; \
-	add	r2, r9, r4; \
-	eor	lr, lr, r3, ror #19; \
-	add	r3, r6, r5; \
-	eor	r10, r10, r2, ror #19; \
-	eor	r11, r11, r3, ror #19; \
-	ldrd	r2, [sp, #2*4]; \
-	add	r6, r11, r6; \
-	add	r7, r12, r7; \
-	eor	r0, r0, r6, ror #14; \
-	add	r6, lr, r8; \
-	eor	r1, r1, r7, ror #14; \
-	add	r7, r10, r9; \
-	eor	r2, r2, r6, ror #14; \
-	eor	r3, r3, r7, ror #14; \
-
-
-#define salsa8_core() \
-	ldmia	sp, {r0-r12, lr}; \
-	ldrd	r10, [sp, #14*4]; \
-	salsa8_core_doubleround_body(); \
-	ldrd	r6, [sp, #6*4]; \
-	strd	r2, [sp, #2*4]; \
-	strd	r10, [sp, #14*4]; \
-	salsa8_core_doubleround_body(); \
-	ldrd	r6, [sp, #6*4]; \
-	strd	r2, [sp, #2*4]; \
-	strd	r10, [sp, #14*4]; \
-	salsa8_core_doubleround_body(); \
-	ldrd	r6, [sp, #6*4]; \
-	strd	r2, [sp, #2*4]; \
-	strd	r10, [sp, #14*4]; \
-	salsa8_core_doubleround_body(); \
-	stmia	sp, {r0-r5}; \
-	strd	r8, [sp, #8*4]; \
-	str	r12, [sp, #12*4]; \
-	str	lr, [sp, #13*4]; \
-	strd	r10, [sp, #14*4]; \
-
-
-#else
-
-#define scrypt_shuffle() \
-
-
-#define salsa8_core_doubleround_body() \
-	ldr	r8, [sp, #8*4]; \
-	add	r11, r11, r10; \
-	ldr	lr, [sp, #13*4]; \
-	add	r12, r12, r3; \
-	eor	r2, r2, r11, ror #23; \
-	add	r11, r4, r0; \
-	eor	r7, r7, r12, ror #23; \
-	add	r12, r9, r5; \
-	str	r9, [sp, #9*4]; \
-	eor	r8, r8, r11, ror #23; \
-	str	r10, [sp, #14*4]; \
-	eor	lr, lr, r12, ror #23; \
-	ldr	r11, [sp, #11*4]; \
-	add	r9, lr, r9; \
-	ldr	r12, [sp, #12*4]; \
-	add	r10, r2, r10; \
-	eor	r1, r1, r9, ror #19; \
-	add	r9, r7, r3; \
-	eor	r6, r6, r10, ror #19; \
-	add	r10, r8, r4; \
-	str	r8, [sp, #8*4]; \
-	eor	r11, r11, r9, ror #19; \
-	str	lr, [sp, #13*4]; \
-	eor	r12, r12, r10, ror #19; \
-	ldr	r9, [sp, #10*4]; \
-	add	r8, r12, r8; \
-	ldr	r10, [sp, #15*4]; \
-	add	lr, r1, lr; \
-	eor	r0, r0, r8, ror #14; \
-	add	r8, r6, r2; \
-	eor	r5, r5, lr, ror #14; \
-	add	lr, r11, r7; \
-	eor	r9, r9, r8, ror #14; \
-	ldr	r8, [sp, #9*4]; \
-	eor	r10, r10, lr, ror #14; \
-	ldr	lr, [sp, #14*4]; \
-	add	r8, r9, r8; \
-	str	r9, [sp, #10*4]; \
-	add	lr, r10, lr; \
-	str	r10, [sp, #15*4]; \
-	eor	r11, r11, r8, ror #25; \
-	add	r8, r0, r3; \
-	eor	r12, r12, lr, ror #25; \
-	add	lr, r5, r4; \
-	eor	r1, r1, r8, ror #25; \
-	ldr	r8, [sp, #8*4]; \
-	eor	r6, r6, lr, ror #25; \
-	add	r9, r11, r9; \
-	ldr	lr, [sp, #13*4]; \
-	add	r10, r12, r10; \
-	eor	r8, r8, r9, ror #23; \
-	add	r9, r1, r0; \
-	eor	lr, lr, r10, ror #23; \
-	add	r10, r6, r5; \
-	str	r11, [sp, #11*4]; \
-	eor	r2, r2, r9, ror #23; \
-	str	r12, [sp, #12*4]; \
-	eor	r7, r7, r10, ror #23; \
-	ldr	r9, [sp, #9*4]; \
-	add	r11, r8, r11; \
-	ldr	r10, [sp, #14*4]; \
-	add	r12, lr, r12; \
-	eor	r9, r9, r11, ror #19; \
-	add	r11, r2, r1; \
-	eor	r10, r10, r12, ror #19; \
-	add	r12, r7, r6; \
-	str	r8, [sp, #8*4]; \
-	eor	r3, r3, r11, ror #19; \
-	str	lr, [sp, #13*4]; \
-	eor	r4, r4, r12, ror #19; \
-
-
-#define salsa8_core() \
-	ldmia	sp, {r0-r7}; \
-	ldr	r12, [sp, #15*4]; \
-	ldr	r8, [sp, #11*4]; \
-	ldr	lr, [sp, #12*4]; \
-	ldr	r9, [sp, #9*4]; \
-	add	r8, r8, r12; \
-	ldr	r11, [sp, #10*4]; \
-	add	lr, lr, r0; \
-	eor	r3, r3, r8, ror #25; \
-	add	r8, r5, r1; \
-	ldr	r10, [sp, #14*4]; \
-	eor	r4, r4, lr, ror #25; \
-	add	lr, r11, r6; \
-	eor	r9, r9, r8, ror #25; \
-	eor	r10, r10, lr, ror #25; \
-	salsa8_core_doubleround_body(); \
-	ldr	r11, [sp, #10*4]; \
-	add	r8, r9, r8; \
-	ldr	r12, [sp, #15*4]; \
-	add	lr, r10, lr; \
-	eor	r11, r11, r8, ror #14; \
-	add	r8, r3, r2; \
-	eor	r12, r12, lr, ror #14; \
-	add	lr, r4, r7; \
-	eor	r0, r0, r8, ror #14; \
-	ldr	r8, [sp, #11*4]; \
-	eor	r5, r5, lr, ror #14; \
-	ldr	lr, [sp, #12*4]; \
-	add	r8, r8, r12; \
-	str	r11, [sp, #10*4]; \
-	add	lr, lr, r0; \
-	str	r12, [sp, #15*4]; \
-	eor	r3, r3, r8, ror #25; \
-	add	r8, r5, r1; \
-	eor	r4, r4, lr, ror #25; \
-	add	lr, r11, r6; \
-	str	r9, [sp, #9*4]; \
-	eor	r9, r9, r8, ror #25; \
-	str	r10, [sp, #14*4]; \
-	eor	r10, r10, lr, ror #25; \
-	salsa8_core_doubleround_body(); \
-	ldr	r11, [sp, #10*4]; \
-	add	r8, r9, r8; \
-	ldr	r12, [sp, #15*4]; \
-	add	lr, r10, lr; \
-	eor	r11, r11, r8, ror #14; \
-	add	r8, r3, r2; \
-	eor	r12, r12, lr, ror #14; \
-	add	lr, r4, r7; \
-	eor	r0, r0, r8, ror #14; \
-	ldr	r8, [sp, #11*4]; \
-	eor	r5, r5, lr, ror #14; \
-	ldr	lr, [sp, #12*4]; \
-	add	r8, r8, r12; \
-	str	r11, [sp, #10*4]; \
-	add	lr, lr, r0; \
-	str	r12, [sp, #15*4]; \
-	eor	r3, r3, r8, ror #25; \
-	add	r8, r5, r1; \
-	eor	r4, r4, lr, ror #25; \
-	add	lr, r11, r6; \
-	str	r9, [sp, #9*4]; \
-	eor	r9, r9, r8, ror #25; \
-	str	r10, [sp, #14*4]; \
-	eor	r10, r10, lr, ror #25; \
-	salsa8_core_doubleround_body(); \
-	ldr	r11, [sp, #10*4]; \
-	add	r8, r9, r8; \
-	ldr	r12, [sp, #15*4]; \
-	add	lr, r10, lr; \
-	eor	r11, r11, r8, ror #14; \
-	add	r8, r3, r2; \
-	eor	r12, r12, lr, ror #14; \
-	add	lr, r4, r7; \
-	eor	r0, r0, r8, ror #14; \
-	ldr	r8, [sp, #11*4]; \
-	eor	r5, r5, lr, ror #14; \
-	ldr	lr, [sp, #12*4]; \
-	add	r8, r8, r12; \
-	str	r11, [sp, #10*4]; \
-	add	lr, lr, r0; \
-	str	r12, [sp, #15*4]; \
-	eor	r3, r3, r8, ror #25; \
-	add	r8, r5, r1; \
-	eor	r4, r4, lr, ror #25; \
-	add	lr, r11, r6; \
-	str	r9, [sp, #9*4]; \
-	eor	r9, r9, r8, ror #25; \
-	str	r10, [sp, #14*4]; \
-	eor	r10, r10, lr, ror #25; \
-	salsa8_core_doubleround_body(); \
-	ldr	r11, [sp, #10*4]; \
-	add	r8, r9, r8; \
-	ldr	r12, [sp, #15*4]; \
-	add	lr, r10, lr; \
-	str	r9, [sp, #9*4]; \
-	eor	r11, r11, r8, ror #14; \
-	eor	r12, r12, lr, ror #14; \
-	add	r8, r3, r2; \
-	str	r10, [sp, #14*4]; \
-	add	lr, r4, r7; \
-	str	r11, [sp, #10*4]; \
-	eor	r0, r0, r8, ror #14; \
-	str	r12, [sp, #15*4]; \
-	eor	r5, r5, lr, ror #14; \
-	stmia	sp, {r0-r7}; \
-
-
-#endif
-
-
-#define scrypt_core_macro1a_x4() \
-	ldmia	r0, {r4-r7}; \
-	ldmia	lr!, {r8-r11}; \
-	stmia	r1!, {r4-r7}; \
-	stmia	r3!, {r8-r11}; \
-	eor	r4, r4, r8; \
-	eor	r5, r5, r9; \
-	eor	r6, r6, r10; \
-	eor	r7, r7, r11; \
-	stmia	r0!, {r4-r7}; \
-	stmia	r12!, {r4-r7}; \
-
-
-#define scrypt_core_macro1b_x4() \
-	ldmia	r3!, {r8-r11}; \
-	ldmia	r2, {r4-r7}; \
-	eor	r8, r8, r4; \
-	eor	r9, r9, r5; \
-	eor	r10, r10, r6; \
-	eor	r11, r11, r7; \
-	ldmia	r0, {r4-r7}; \
-	stmia	r2!, {r8-r11}; \
-	eor	r4, r4, r8; \
-	eor	r5, r5, r9; \
-	eor	r6, r6, r10; \
-	eor	r7, r7, r11; \
-	ldmia	r1!, {r8-r11}; \
-	eor	r4, r4, r8; \
-	eor	r5, r5, r9; \
-	eor	r6, r6, r10; \
-	eor	r7, r7, r11; \
-	stmia	r0!, {r4-r7}; \
-	stmia	r12!, {r4-r7}; \
-
-
-#define scrypt_core_macro2_x4() \
-	ldmia	r12, {r4-r7}; \
-	ldmia	r0, {r8-r11}; \
-	add	r4, r4, r8; \
-	add	r5, r5, r9; \
-	add	r6, r6, r10; \
-	add	r7, r7, r11; \
-	stmia	r0!, {r4-r7}; \
-	ldmia	r2, {r8-r11}; \
-	eor	r4, r4, r8; \
-	eor	r5, r5, r9; \
-	eor	r6, r6, r10; \
-	eor	r7, r7, r11; \
-	stmia	r2!, {r4-r7}; \
-	stmia	r12!, {r4-r7}; \
-
-
-#define scrypt_core_macro3_x4() \
-	ldmia	r1!, {r4-r7}; \
-	ldmia	r0, {r8-r11}; \
-	add	r4, r4, r8; \
-	add	r5, r5, r9; \
-	add	r6, r6, r10; \
-	add	r7, r7, r11; \
-	stmia	r0!, {r4-r7}; \
-
-
-#define scrypt_core_macro3_x6() \
-	ldmia	r1!, {r2-r7}; \
-	ldmia	r0, {r8-r12, lr}; \
-	add	r2, r2, r8; \
-	add	r3, r3, r9; \
-	add	r4, r4, r10; \
-	add	r5, r5, r11; \
-	add	r6, r6, r12; \
-	add	r7, r7, lr; \
-	stmia	r0!, {r2-r7}; \
-
+.macro salsa8_core_doubleround_body
+	ldr	r8, [sp, #8*4]
+	add	r11, r11, r10
+	ldr	lr, [sp, #13*4]
+	add	r12, r12, r3
+	eor	r2, r2, r11, ror #23
+	add	r11, r4, r0
+	eor	r7, r7, r12, ror #23
+	add	r12, r9, r5
+	str	r9, [sp, #9*4]
+	eor	r8, r8, r11, ror #23
+	str	r10, [sp, #14*4]
+	eor	lr, lr, r12, ror #23
+	
+	ldr	r11, [sp, #11*4]
+	add	r9, lr, r9
+	ldr	r12, [sp, #12*4]
+	add	r10, r2, r10
+	eor	r1, r1, r9, ror #19
+	add	r9, r7, r3
+	eor	r6, r6, r10, ror #19
+	add	r10, r8, r4
+	str	r8, [sp, #8*4]
+	eor	r11, r11, r9, ror #19
+	str	lr, [sp, #13*4]
+	eor	r12, r12, r10, ror #19
+	
+	ldr	r9, [sp, #10*4]
+	add	r8, r12, r8
+	ldr	r10, [sp, #15*4]
+	add	lr, r1, lr
+	eor	r0, r0, r8, ror #14
+	add	r8, r6, r2
+	eor	r5, r5, lr, ror #14
+	add	lr, r11, r7
+	eor	r9, r9, r8, ror #14
+	ldr	r8, [sp, #9*4]
+	eor	r10, r10, lr, ror #14
+	ldr	lr, [sp, #14*4]
+	
+	
+	add	r8, r9, r8
+	str	r9, [sp, #10*4]
+	add	lr, r10, lr
+	str	r10, [sp, #15*4]
+	eor	r11, r11, r8, ror #25
+	add	r8, r0, r3
+	eor	r12, r12, lr, ror #25
+	add	lr, r5, r4
+	eor	r1, r1, r8, ror #25
+	ldr	r8, [sp, #8*4]
+	eor	r6, r6, lr, ror #25
+	
+	add	r9, r11, r9
+	ldr	lr, [sp, #13*4]
+	add	r10, r12, r10
+	eor	r8, r8, r9, ror #23
+	add	r9, r1, r0
+	eor	lr, lr, r10, ror #23
+	add	r10, r6, r5
+	str	r11, [sp, #11*4]
+	eor	r2, r2, r9, ror #23
+	str	r12, [sp, #12*4]
+	eor	r7, r7, r10, ror #23
+	
+	ldr	r9, [sp, #9*4]
+	add	r11, r8, r11
+	ldr	r10, [sp, #14*4]
+	add	r12, lr, r12
+	eor	r9, r9, r11, ror #19
+	add	r11, r2, r1
+	eor	r10, r10, r12, ror #19
+	add	r12, r7, r6
+	str	r8, [sp, #8*4]
+	eor	r3, r3, r11, ror #19
+	str	lr, [sp, #13*4]
+	eor	r4, r4, r12, ror #19
+.endm
+
+.macro salsa8_core
+	ldmia	sp, {r0-r7}
+	
+	ldr	r12, [sp, #15*4]
+	ldr	r8, [sp, #11*4]
+	ldr	lr, [sp, #12*4]
+	
+	ldr	r9, [sp, #9*4]
+	add	r8, r8, r12
+	ldr	r11, [sp, #10*4]
+	add	lr, lr, r0
+	eor	r3, r3, r8, ror #25
+	add	r8, r5, r1
+	ldr	r10, [sp, #14*4]
+	eor	r4, r4, lr, ror #25
+	add	lr, r11, r6
+	eor	r9, r9, r8, ror #25
+	eor	r10, r10, lr, ror #25
+	
+	salsa8_core_doubleround_body
+	
+	ldr	r11, [sp, #10*4]
+	add	r8, r9, r8
+	ldr	r12, [sp, #15*4]
+	add	lr, r10, lr
+	eor	r11, r11, r8, ror #14
+	str	r9, [sp, #9*4]
+	eor	r12, r12, lr, ror #14
+	add	r8, r3, r2
+	add	lr, r4, r7
+	str	r10, [sp, #14*4]
+	eor	r0, r0, r8, ror #14
+	ldr	r8, [sp, #11*4]
+	eor	r5, r5, lr, ror #14
+	ldr	lr, [sp, #12*4]
+	
+	add	r8, r8, r12
+	str	r11, [sp, #10*4]
+	add	lr, lr, r0
+	str	r12, [sp, #15*4]
+	eor	r3, r3, r8, ror #25
+	add	r8, r5, r1
+	eor	r4, r4, lr, ror #25
+	add	lr, r11, r6
+	eor	r9, r9, r8, ror #25
+	eor	r10, r10, lr, ror #25
+	
+	salsa8_core_doubleround_body
+	
+	ldr	r11, [sp, #10*4]
+	add	r8, r9, r8
+	ldr	r12, [sp, #15*4]
+	add	lr, r10, lr
+	eor	r11, r11, r8, ror #14
+	str	r9, [sp, #9*4]
+	eor	r12, r12, lr, ror #14
+	add	r8, r3, r2
+	add	lr, r4, r7
+	str	r10, [sp, #14*4]
+	eor	r0, r0, r8, ror #14
+	ldr	r8, [sp, #11*4]
+	eor	r5, r5, lr, ror #14
+	ldr	lr, [sp, #12*4]
+	
+	add	r8, r8, r12
+	str	r11, [sp, #10*4]
+	add	lr, lr, r0
+	str	r12, [sp, #15*4]
+	eor	r3, r3, r8, ror #25
+	add	r8, r5, r1
+	eor	r4, r4, lr, ror #25
+	add	lr, r11, r6
+	eor	r9, r9, r8, ror #25
+	eor	r10, r10, lr, ror #25
+	
+	salsa8_core_doubleround_body
+	
+	ldr	r11, [sp, #10*4]
+	add	r8, r9, r8
+	ldr	r12, [sp, #15*4]
+	add	lr, r10, lr
+	eor	r11, r11, r8, ror #14
+	str	r9, [sp, #9*4]
+	eor	r12, r12, lr, ror #14
+	add	r8, r3, r2
+	add	lr, r4, r7
+	str	r10, [sp, #14*4]
+	eor	r0, r0, r8, ror #14
+	ldr	r8, [sp, #11*4]
+	eor	r5, r5, lr, ror #14
+	ldr	lr, [sp, #12*4]
+	
+	add	r8, r8, r12
+	str	r11, [sp, #10*4]
+	add	lr, lr, r0
+	str	r12, [sp, #15*4]
+	eor	r3, r3, r8, ror #25
+	add	r8, r5, r1
+	eor	r4, r4, lr, ror #25
+	add	lr, r11, r6
+	eor	r9, r9, r8, ror #25
+	eor	r10, r10, lr, ror #25
+	
+	salsa8_core_doubleround_body
+	
+	ldr	r11, [sp, #10*4]
+	add	r8, r9, r8
+	ldr	r12, [sp, #15*4]
+	add	lr, r10, lr
+	str	r9, [sp, #9*4]
+	eor	r11, r11, r8, ror #14
+	eor	r12, r12, lr, ror #14
+	add	r8, r3, r2
+	str	r10, [sp, #14*4]
+	add	lr, r4, r7
+	str	r11, [sp, #10*4]
+	eor	r0, r0, r8, ror #14
+	str	r12, [sp, #15*4]
+	eor	r5, r5, lr, ror #14
+	
+	stmia	sp, {r0-r7}
+.endm
+
+
+.macro scrypt_core_macro1a_x4
+	ldmia	r0, {r4-r7}
+	ldmia	lr!, {r8-r11}
+	stmia	r1!, {r4-r7}
+	stmia	r3!, {r8-r11}
+	eor	r4, r4, r8
+	eor	r5, r5, r9
+	eor	r6, r6, r10
+	eor	r7, r7, r11
+	stmia	r0!, {r4-r7}
+	stmia	r12!, {r4-r7}
+.endm
+
+.macro scrypt_core_macro1b_x4
+	ldmia	r3!, {r8-r11}
+	ldmia	r2, {r4-r7}
+	eor	r8, r8, r4
+	eor	r9, r9, r5
+	eor	r10, r10, r6
+	eor	r11, r11, r7
+	ldmia	r0, {r4-r7}
+	stmia	r2!, {r8-r11}
+	eor	r4, r4, r8
+	eor	r5, r5, r9
+	eor	r6, r6, r10
+	eor	r7, r7, r11
+	ldmia	r1!, {r8-r11}
+	eor	r4, r4, r8
+	eor	r5, r5, r9
+	eor	r6, r6, r10
+	eor	r7, r7, r11
+	stmia	r0!, {r4-r7}
+	stmia	r12!, {r4-r7}
+.endm
+
+.macro scrypt_core_macro2_x4
+	ldmia	r12, {r4-r7}
+	ldmia	r0, {r8-r11}
+	add	r4, r4, r8
+	add	r5, r5, r9
+	add	r6, r6, r10
+	add	r7, r7, r11
+	stmia	r0!, {r4-r7}
+	ldmia	r2, {r8-r11}
+	eor	r4, r4, r8
+	eor	r5, r5, r9
+	eor	r6, r6, r10
+	eor	r7, r7, r11
+	stmia	r2!, {r4-r7}
+	stmia	r12!, {r4-r7}
+.endm
+
+.macro scrypt_core_macro3_x4
+	ldmia	r1!, {r4-r7}
+	ldmia	r0, {r8-r11}
+	add	r4, r4, r8
+	add	r5, r5, r9
+	add	r6, r6, r10
+	add	r7, r7, r11
+	stmia	r0!, {r4-r7}
+.endm
+
+.macro scrypt_core_macro3_x6
+	ldmia	r1!, {r2-r7}
+	ldmia	r0, {r8-r12, lr}
+	add	r2, r2, r8
+	add	r3, r3, r9
+	add	r4, r4, r10
+	add	r5, r5, r11
+	add	r6, r6, r12
+	add	r7, r7, lr
+	stmia	r0!, {r2-r7}
+.endm
 
 
 	.text
@@ -443,12 +298,7 @@
 scrypt_core:
 _scrypt_core:
 	stmfd	sp!, {r4-r11, lr}
-	mov	r12, sp
-	sub	sp, sp, #21*4
-	bic	sp, sp, #63
-	str	r12, [sp, #20*4]
-	
-	scrypt_shuffle()
+	sub	sp, sp, #20*4
 	
 	str	r0, [sp, #16*4]
 	add	r12, r1, #1024*32*4
@@ -457,94 +307,82 @@ scrypt_core_loop1:
 	add	lr, r0, #16*4
 	add	r3, r1, #16*4
 	mov	r12, sp
-	scrypt_core_macro1a_x4()
-	scrypt_core_macro1a_x4()
-	scrypt_core_macro1a_x4()
-	scrypt_core_macro1a_x4()
+	scrypt_core_macro1a_x4
+	scrypt_core_macro1a_x4
+	scrypt_core_macro1a_x4
+	scrypt_core_macro1a_x4
 	str	r1, [sp, #17*4]
 	
-	salsa8_core()
+	salsa8_core
 	
 	ldr	r0, [sp, #16*4]
 	mov	r12, sp
 	add	r2, r0, #16*4
-	scrypt_core_macro2_x4()
-	scrypt_core_macro2_x4()
-	scrypt_core_macro2_x4()
-	scrypt_core_macro2_x4()
+	scrypt_core_macro2_x4
+	scrypt_core_macro2_x4
+	scrypt_core_macro2_x4
+	scrypt_core_macro2_x4
 	
-	salsa8_core()
+	salsa8_core
 	
 	ldr	r0, [sp, #16*4]
 	mov	r1, sp
 	add	r0, r0, #16*4
-	scrypt_core_macro3_x6()
-	scrypt_core_macro3_x6()
+	scrypt_core_macro3_x6
+	scrypt_core_macro3_x6
 	ldr	r3, [sp, #17*4]
 	ldr	r12, [sp, #18*4]
-	scrypt_core_macro3_x4()
+	scrypt_core_macro3_x4
 	
 	add	r1, r3, #16*4
 	sub	r0, r0, #32*4
 	cmp	r1, r12
 	bne	scrypt_core_loop1
 	
-	ldr	r4, [r0, #16*4]
 	sub	r1, r1, #1024*32*4
 	str	r1, [sp, #17*4]
-	mov	r4, r4, lsl #32-10
 	mov	r12, #1024
-	add	r1, r1, r4, lsr #32-10-7
 scrypt_core_loop2:
+	str	r12, [sp, #18*4]
+	
+	ldr	r4, [r0, #16*4]
+	mov	r4, r4, lsl #32-10
+	add	r1, r1, r4, lsr #32-10-7
+	
 	add	r2, r0, #16*4
 	add	r3, r1, #16*4
-	str	r12, [sp, #18*4]
 	mov	r12, sp
-#ifdef __ARM_ARCH_5E_OR_6_OR_7__
-	pld [r1, #24*4]
-	pld [r1, #8*4]
-#endif
-	scrypt_core_macro1b_x4()
-	scrypt_core_macro1b_x4()
-	scrypt_core_macro1b_x4()
-	scrypt_core_macro1b_x4()
+	scrypt_core_macro1b_x4
+	scrypt_core_macro1b_x4
+	scrypt_core_macro1b_x4
+	scrypt_core_macro1b_x4
 	
-	salsa8_core()
+	salsa8_core
 	
 	ldr	r0, [sp, #16*4]
 	mov	r12, sp
 	add	r2, r0, #16*4
-	scrypt_core_macro2_x4()
-	scrypt_core_macro2_x4()
-	scrypt_core_macro2_x4()
-	scrypt_core_macro2_x4()
+	scrypt_core_macro2_x4
+	scrypt_core_macro2_x4
+	scrypt_core_macro2_x4
+	scrypt_core_macro2_x4
 	
-	salsa8_core()
+	salsa8_core
 	
 	ldr	r0, [sp, #16*4]
 	mov	r1, sp
-	ldr	r3, [sp, #17*4]
 	add	r0, r0, #16*4
-	scrypt_core_macro3_x4()
-	mov	r4, r4, lsl #32-10
-	add	r3, r3, r4, lsr #32-10-7
-	str	r3, [sp, #19*4]
-#ifdef __ARM_ARCH_5E_OR_6_OR_7__
-	pld	[r3, #16*4]
-	pld	[r3]
-#endif
-	scrypt_core_macro3_x6()
-	scrypt_core_macro3_x6()
+	scrypt_core_macro3_x6
+	scrypt_core_macro3_x6
+	scrypt_core_macro3_x4
 	
 	ldr	r12, [sp, #18*4]
 	sub	r0, r0, #32*4
-	ldr	r1, [sp, #19*4]
+	ldr	r1, [sp, #17*4]
 	subs	r12, r12, #1
 	bne	scrypt_core_loop2
 	
-	scrypt_shuffle()
-	
-	ldr	sp, [sp, #20*4]
+	add	sp, sp, #20*4
 #ifdef __thumb__
 	ldmfd	sp!, {r4-r11, lr}
 	bx	lr
diff --git a/src/scrypt-x86.S b/src/scrypt-x86.S
index 0e97e36..bfca2ed 100644
--- a/src/scrypt-x86.S
+++ b/src/scrypt-x86.S
@@ -1,363 +1,400 @@
-# Copyright 2011 pooler@litecoinpool.org
-# All rights reserved.
-#
-# Redistribution and use in source and binary forms, with or without
-# modification, are permitted provided that the following conditions
-# are met:
-# 1. Redistributions of source code must retain the above copyright
-#    notice, this list of conditions and the following disclaimer.
-# 2. Redistributions in binary form must reproduce the above copyright
-#    notice, this list of conditions and the following disclaimer in the
-#    documentation and/or other materials provided with the distribution.
-#
-# THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
-# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
-# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
-# ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
-# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
-# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
-# OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
-# HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
-# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
-# OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
-# SUCH DAMAGE.
-
-#if defined(__i386__)
+/*
+ * Copyright 2011-2012 pooler@litecoinpool.org
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
 
 #if defined(__linux__) && defined(__ELF__)
-.section .note.GNU-stack,"",%progbits
+	.section .note.GNU-stack,"",%progbits
 #endif
 
-#define gen_salsa8_core_quadround() \
-	movl	52(%esp), %ecx; \
-	movl	4(%esp), %edx; \
-	movl	20(%esp), %ebx; \
-	movl	8(%esp), %esi; \
-	leal	(%ecx, %edx), %edi; \
-	roll	$7, %edi; \
-	xorl	%edi, %ebx; \
-	movl	%ebx, 4(%esp); \
-	movl	36(%esp), %edi; \
-	leal	(%edx, %ebx), %ebp; \
-	roll	$9, %ebp; \
-	xorl	%ebp, %edi; \
-	movl	24(%esp), %ebp; \
-	movl	%edi, 8(%esp); \
-	addl	%edi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %ecx; \
-	movl	40(%esp), %ebx; \
-	movl	%ecx, 20(%esp); \
-	addl	%edi, %ecx; \
-	roll	$18, %ecx; \
-	leal	(%esi, %ebp), %edi; \
-	roll	$7, %edi; \
-	xorl	%edi, %ebx; \
-	movl	%ebx, 24(%esp); \
-	movl	56(%esp), %edi; \
-	xorl	%ecx, %edx; \
-	leal	(%ebp, %ebx), %ecx; \
-	roll	$9, %ecx; \
-	xorl	%ecx, %edi; \
-	movl	%edi, 36(%esp); \
-	movl	28(%esp), %ecx; \
-	movl	%edx, 28(%esp); \
-	movl	44(%esp), %edx; \
-	addl	%edi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %esi; \
-	movl	60(%esp), %ebx; \
-	movl	%esi, 40(%esp); \
-	addl	%edi, %esi; \
-	roll	$18, %esi; \
-	leal	(%ecx, %edx), %edi; \
-	roll	$7, %edi; \
-	xorl	%edi, %ebx; \
-	movl	%ebx, 44(%esp); \
-	movl	12(%esp), %edi; \
-	xorl	%esi, %ebp; \
-	leal	(%edx, %ebx), %esi; \
-	roll	$9, %esi; \
-	xorl	%esi, %edi; \
-	movl	%edi, 12(%esp); \
-	movl	48(%esp), %esi; \
-	movl	%ebp, 48(%esp); \
-	movl	64(%esp), %ebp; \
-	addl	%edi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %ecx; \
-	movl	16(%esp), %ebx; \
-	movl	%ecx, 16(%esp); \
-	addl	%edi, %ecx; \
-	roll	$18, %ecx; \
-	leal	(%esi, %ebp), %edi; \
-	roll	$7, %edi; \
-	xorl	%edi, %ebx; \
-	movl	32(%esp), %edi; \
-	xorl	%ecx, %edx; \
-	leal	(%ebp, %ebx), %ecx; \
-	roll	$9, %ecx; \
-	xorl	%ecx, %edi; \
-	movl	%edi, 32(%esp); \
-	movl	%ebx, %ecx; \
-	movl	%edx, 52(%esp); \
-	movl	28(%esp), %edx; \
-	addl	%edi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %esi; \
-	movl	40(%esp), %ebx; \
-	movl	%esi, 28(%esp); \
-	addl	%edi, %esi; \
-	roll	$18, %esi; \
-	leal	(%ecx, %edx), %edi; \
-	roll	$7, %edi; \
-	xorl	%edi, %ebx; \
-	movl	%ebx, 40(%esp); \
-	movl	12(%esp), %edi; \
-	xorl	%esi, %ebp; \
-	leal	(%edx, %ebx), %esi; \
-	roll	$9, %esi; \
-	xorl	%esi, %edi; \
-	movl	%edi, 12(%esp); \
-	movl	4(%esp), %esi; \
-	movl	%ebp, 4(%esp); \
-	movl	48(%esp), %ebp; \
-	addl	%edi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %ecx; \
-	movl	16(%esp), %ebx; \
-	movl	%ecx, 16(%esp); \
-	addl	%edi, %ecx; \
-	roll	$18, %ecx; \
-	leal	(%esi, %ebp), %edi; \
-	roll	$7, %edi; \
-	xorl	%edi, %ebx; \
-	movl	%ebx, 48(%esp); \
-	movl	32(%esp), %edi; \
-	xorl	%ecx, %edx; \
-	leal	(%ebp, %ebx), %ecx; \
-	roll	$9, %ecx; \
-	xorl	%ecx, %edi; \
-	movl	%edi, 32(%esp); \
-	movl	24(%esp), %ecx; \
-	movl	%edx, 24(%esp); \
-	movl	52(%esp), %edx; \
-	addl	%edi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %esi; \
-	movl	28(%esp), %ebx; \
-	movl	%esi, 28(%esp); \
-	addl	%edi, %esi; \
-	roll	$18, %esi; \
-	leal	(%ecx, %edx), %edi; \
-	roll	$7, %edi; \
-	xorl	%edi, %ebx; \
-	movl	%ebx, 52(%esp); \
-	movl	8(%esp), %edi; \
-	xorl	%esi, %ebp; \
-	leal	(%edx, %ebx), %esi; \
-	roll	$9, %esi; \
-	xorl	%esi, %edi; \
-	movl	%edi, 8(%esp); \
-	movl	44(%esp), %esi; \
-	movl	%ebp, 44(%esp); \
-	movl	4(%esp), %ebp; \
-	addl	%edi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %ecx; \
-	movl	20(%esp), %ebx; \
-	movl	%ecx, 4(%esp); \
-	addl	%edi, %ecx; \
-	roll	$18, %ecx; \
-	leal	(%esi, %ebp), %edi; \
-	roll	$7, %edi; \
-	xorl	%edi, %ebx; \
-	movl	36(%esp), %edi; \
-	xorl	%ecx, %edx; \
-	leal	(%ebp, %ebx), %ecx; \
-	roll	$9, %ecx; \
-	xorl	%ecx, %edi; \
-	movl	%edi, 20(%esp); \
-	movl	%ebx, %ecx; \
-	movl	%edx, 36(%esp); \
-	movl	24(%esp), %edx; \
-	addl	%edi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %esi; \
-	movl	28(%esp), %ebx; \
-	movl	%esi, 24(%esp); \
-	addl	%edi, %esi; \
-	roll	$18, %esi; \
-	leal	(%ecx, %edx), %edi; \
-	roll	$7, %edi; \
-	xorl	%edi, %ebx; \
-	movl	%ebx, 28(%esp); \
-	xorl	%esi, %ebp; \
-	movl	8(%esp), %esi; \
-	leal	(%edx, %ebx), %edi; \
-	roll	$9, %edi; \
-	xorl	%edi, %esi; \
-	movl	40(%esp), %edi; \
-	movl	%ebp, 8(%esp); \
-	movl	44(%esp), %ebp; \
-	movl	%esi, 40(%esp); \
-	addl	%esi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %ecx; \
-	movl	4(%esp), %ebx; \
-	movl	%ecx, 44(%esp); \
-	addl	%esi, %ecx; \
-	roll	$18, %ecx; \
-	leal	(%edi, %ebp), %esi; \
-	roll	$7, %esi; \
-	xorl	%esi, %ebx; \
-	movl	%ebx, 4(%esp); \
-	movl	20(%esp), %esi; \
-	xorl	%ecx, %edx; \
-	leal	(%ebp, %ebx), %ecx; \
-	roll	$9, %ecx; \
-	xorl	%ecx, %esi; \
-	movl	%esi, 56(%esp); \
-	movl	48(%esp), %ecx; \
-	movl	%edx, 20(%esp); \
-	movl	36(%esp), %edx; \
-	addl	%esi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %edi; \
-	movl	24(%esp), %ebx; \
-	movl	%edi, 24(%esp); \
-	addl	%esi, %edi; \
-	roll	$18, %edi; \
-	leal	(%ecx, %edx), %esi; \
-	roll	$7, %esi; \
-	xorl	%esi, %ebx; \
-	movl	%ebx, 60(%esp); \
-	movl	12(%esp), %esi; \
-	xorl	%edi, %ebp; \
-	leal	(%edx, %ebx), %edi; \
-	roll	$9, %edi; \
-	xorl	%edi, %esi; \
-	movl	%esi, 12(%esp); \
-	movl	52(%esp), %edi; \
-	movl	%ebp, 36(%esp); \
-	movl	8(%esp), %ebp; \
-	addl	%esi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %ecx; \
-	movl	16(%esp), %ebx; \
-	movl	%ecx, 16(%esp); \
-	addl	%esi, %ecx; \
-	roll	$18, %ecx; \
-	leal	(%edi, %ebp), %esi; \
-	roll	$7, %esi; \
-	xorl	%esi, %ebx; \
-	movl	32(%esp), %esi; \
-	xorl	%ecx, %edx; \
-	leal	(%ebp, %ebx), %ecx; \
-	roll	$9, %ecx; \
-	xorl	%ecx, %esi; \
-	movl	%esi, 32(%esp); \
-	movl	%ebx, %ecx; \
-	movl	%edx, 48(%esp); \
-	movl	20(%esp), %edx; \
-	addl	%esi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %edi; \
-	movl	24(%esp), %ebx; \
-	movl	%edi, 20(%esp); \
-	addl	%esi, %edi; \
-	roll	$18, %edi; \
-	leal	(%ecx, %edx), %esi; \
-	roll	$7, %esi; \
-	xorl	%esi, %ebx; \
-	movl	%ebx, 8(%esp); \
-	movl	12(%esp), %esi; \
-	xorl	%edi, %ebp; \
-	leal	(%edx, %ebx), %edi; \
-	roll	$9, %edi; \
-	xorl	%edi, %esi; \
-	movl	%esi, 12(%esp); \
-	movl	28(%esp), %edi; \
-	movl	%ebp, 52(%esp); \
-	movl	36(%esp), %ebp; \
-	addl	%esi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %ecx; \
-	movl	16(%esp), %ebx; \
-	movl	%ecx, 16(%esp); \
-	addl	%esi, %ecx; \
-	roll	$18, %ecx; \
-	leal	(%edi, %ebp), %esi; \
-	roll	$7, %esi; \
-	xorl	%esi, %ebx; \
-	movl	%ebx, 28(%esp); \
-	movl	32(%esp), %esi; \
-	xorl	%ecx, %edx; \
-	leal	(%ebp, %ebx), %ecx; \
-	roll	$9, %ecx; \
-	xorl	%ecx, %esi; \
-	movl	%esi, 32(%esp); \
-	movl	4(%esp), %ecx; \
-	movl	%edx, 4(%esp); \
-	movl	48(%esp), %edx; \
-	addl	%esi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %edi; \
-	movl	20(%esp), %ebx; \
-	movl	%edi, 20(%esp); \
-	addl	%esi, %edi; \
-	roll	$18, %edi; \
-	leal	(%ecx, %edx), %esi; \
-	roll	$7, %esi; \
-	xorl	%esi, %ebx; \
-	movl	%ebx, 48(%esp); \
-	movl	40(%esp), %esi; \
-	xorl	%edi, %ebp; \
-	leal	(%edx, %ebx), %edi; \
-	roll	$9, %edi; \
-	xorl	%edi, %esi; \
-	movl	%esi, 36(%esp); \
-	movl	60(%esp), %edi; \
-	movl	%ebp, 24(%esp); \
-	movl	52(%esp), %ebp; \
-	addl	%esi, %ebx; \
-	roll	$13, %ebx; \
-	xorl	%ebx, %ecx; \
-	movl	44(%esp), %ebx; \
-	movl	%ecx, 40(%esp); \
-	addl	%esi, %ecx; \
-	roll	$18, %ecx; \
-	leal	(%edi, %ebp), %esi; \
-	roll	$7, %esi; \
-	xorl	%esi, %ebx; \
-	movl	%ebx, 52(%esp); \
-	movl	56(%esp), %esi; \
-	xorl	%ecx, %edx; \
-	leal	(%ebp, %ebx), %ecx; \
-	roll	$9, %ecx; \
-	xorl	%ecx, %esi; \
-	movl	%esi, 56(%esp); \
-	addl	%esi, %ebx; \
-	movl	%edx, 44(%esp); \
-	roll	$13, %ebx; \
-	xorl	%ebx, %edi; \
-	movl	%edi, 60(%esp); \
-	addl	%esi, %edi; \
-	roll	$18, %edi; \
-	xorl	%edi, %ebp; \
-	movl	%ebp, 64(%esp); \
+#if defined(__i386__)
+	
+.macro scrypt_shuffle src, so, dest, do
+	movl	\so+60(\src), %eax
+	movl	\so+44(\src), %ebx
+	movl	\so+28(\src), %ecx
+	movl	\so+12(\src), %edx
+	movl	%eax, \do+12(\dest)
+	movl	%ebx, \do+28(\dest)
+	movl	%ecx, \do+44(\dest)
+	movl	%edx, \do+60(\dest)
+	movl	\so+40(\src), %eax
+	movl	\so+8(\src), %ebx
+	movl	\so+48(\src), %ecx
+	movl	\so+16(\src), %edx
+	movl	%eax, \do+8(\dest)
+	movl	%ebx, \do+40(\dest)
+	movl	%ecx, \do+16(\dest)
+	movl	%edx, \do+48(\dest)
+	movl	\so+20(\src), %eax
+	movl	\so+4(\src), %ebx
+	movl	\so+52(\src), %ecx
+	movl	\so+36(\src), %edx
+	movl	%eax, \do+4(\dest)
+	movl	%ebx, \do+20(\dest)
+	movl	%ecx, \do+36(\dest)
+	movl	%edx, \do+52(\dest)
+	movl	\so+0(\src), %eax
+	movl	\so+24(\src), %ebx
+	movl	\so+32(\src), %ecx
+	movl	\so+56(\src), %edx
+	movl	%eax, \do+0(\dest)
+	movl	%ebx, \do+24(\dest)
+	movl	%ecx, \do+32(\dest)
+	movl	%edx, \do+56(\dest)
+.endm
 
+.macro salsa8_core_gen_quadround
+	movl	52(%esp), %ecx
+	movl	4(%esp), %edx
+	movl	20(%esp), %ebx
+	movl	8(%esp), %esi
+	leal	(%ecx, %edx), %edi
+	roll	$7, %edi
+	xorl	%edi, %ebx
+	movl	%ebx, 4(%esp)
+	movl	36(%esp), %edi
+	leal	(%edx, %ebx), %ebp
+	roll	$9, %ebp
+	xorl	%ebp, %edi
+	movl	24(%esp), %ebp
+	movl	%edi, 8(%esp)
+	addl	%edi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %ecx
+	movl	40(%esp), %ebx
+	movl	%ecx, 20(%esp)
+	addl	%edi, %ecx
+	roll	$18, %ecx
+	leal	(%esi, %ebp), %edi
+	roll	$7, %edi
+	xorl	%edi, %ebx
+	movl	%ebx, 24(%esp)
+	movl	56(%esp), %edi
+	xorl	%ecx, %edx
+	leal	(%ebp, %ebx), %ecx
+	roll	$9, %ecx
+	xorl	%ecx, %edi
+	movl	%edi, 36(%esp)
+	movl	28(%esp), %ecx
+	movl	%edx, 28(%esp)
+	movl	44(%esp), %edx
+	addl	%edi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %esi
+	movl	60(%esp), %ebx
+	movl	%esi, 40(%esp)
+	addl	%edi, %esi
+	roll	$18, %esi
+	leal	(%ecx, %edx), %edi
+	roll	$7, %edi
+	xorl	%edi, %ebx
+	movl	%ebx, 44(%esp)
+	movl	12(%esp), %edi
+	xorl	%esi, %ebp
+	leal	(%edx, %ebx), %esi
+	roll	$9, %esi
+	xorl	%esi, %edi
+	movl	%edi, 12(%esp)
+	movl	48(%esp), %esi
+	movl	%ebp, 48(%esp)
+	movl	64(%esp), %ebp
+	addl	%edi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %ecx
+	movl	16(%esp), %ebx
+	movl	%ecx, 16(%esp)
+	addl	%edi, %ecx
+	roll	$18, %ecx
+	leal	(%esi, %ebp), %edi
+	roll	$7, %edi
+	xorl	%edi, %ebx
+	movl	32(%esp), %edi
+	xorl	%ecx, %edx
+	leal	(%ebp, %ebx), %ecx
+	roll	$9, %ecx
+	xorl	%ecx, %edi
+	movl	%edi, 32(%esp)
+	movl	%ebx, %ecx
+	movl	%edx, 52(%esp)
+	movl	28(%esp), %edx
+	addl	%edi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %esi
+	movl	40(%esp), %ebx
+	movl	%esi, 28(%esp)
+	addl	%edi, %esi
+	roll	$18, %esi
+	leal	(%ecx, %edx), %edi
+	roll	$7, %edi
+	xorl	%edi, %ebx
+	movl	%ebx, 40(%esp)
+	movl	12(%esp), %edi
+	xorl	%esi, %ebp
+	leal	(%edx, %ebx), %esi
+	roll	$9, %esi
+	xorl	%esi, %edi
+	movl	%edi, 12(%esp)
+	movl	4(%esp), %esi
+	movl	%ebp, 4(%esp)
+	movl	48(%esp), %ebp
+	addl	%edi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %ecx
+	movl	16(%esp), %ebx
+	movl	%ecx, 16(%esp)
+	addl	%edi, %ecx
+	roll	$18, %ecx
+	leal	(%esi, %ebp), %edi
+	roll	$7, %edi
+	xorl	%edi, %ebx
+	movl	%ebx, 48(%esp)
+	movl	32(%esp), %edi
+	xorl	%ecx, %edx
+	leal	(%ebp, %ebx), %ecx
+	roll	$9, %ecx
+	xorl	%ecx, %edi
+	movl	%edi, 32(%esp)
+	movl	24(%esp), %ecx
+	movl	%edx, 24(%esp)
+	movl	52(%esp), %edx
+	addl	%edi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %esi
+	movl	28(%esp), %ebx
+	movl	%esi, 28(%esp)
+	addl	%edi, %esi
+	roll	$18, %esi
+	leal	(%ecx, %edx), %edi
+	roll	$7, %edi
+	xorl	%edi, %ebx
+	movl	%ebx, 52(%esp)
+	movl	8(%esp), %edi
+	xorl	%esi, %ebp
+	leal	(%edx, %ebx), %esi
+	roll	$9, %esi
+	xorl	%esi, %edi
+	movl	%edi, 8(%esp)
+	movl	44(%esp), %esi
+	movl	%ebp, 44(%esp)
+	movl	4(%esp), %ebp
+	addl	%edi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %ecx
+	movl	20(%esp), %ebx
+	movl	%ecx, 4(%esp)
+	addl	%edi, %ecx
+	roll	$18, %ecx
+	leal	(%esi, %ebp), %edi
+	roll	$7, %edi
+	xorl	%edi, %ebx
+	movl	36(%esp), %edi
+	xorl	%ecx, %edx
+	leal	(%ebp, %ebx), %ecx
+	roll	$9, %ecx
+	xorl	%ecx, %edi
+	movl	%edi, 20(%esp)
+	movl	%ebx, %ecx
+	movl	%edx, 36(%esp)
+	movl	24(%esp), %edx
+	addl	%edi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %esi
+	movl	28(%esp), %ebx
+	movl	%esi, 24(%esp)
+	addl	%edi, %esi
+	roll	$18, %esi
+	leal	(%ecx, %edx), %edi
+	roll	$7, %edi
+	xorl	%edi, %ebx
+	movl	%ebx, 28(%esp)
+	xorl	%esi, %ebp
+	movl	8(%esp), %esi
+	leal	(%edx, %ebx), %edi
+	roll	$9, %edi
+	xorl	%edi, %esi
+	movl	40(%esp), %edi
+	movl	%ebp, 8(%esp)
+	movl	44(%esp), %ebp
+	movl	%esi, 40(%esp)
+	addl	%esi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %ecx
+	movl	4(%esp), %ebx
+	movl	%ecx, 44(%esp)
+	addl	%esi, %ecx
+	roll	$18, %ecx
+	leal	(%edi, %ebp), %esi
+	roll	$7, %esi
+	xorl	%esi, %ebx
+	movl	%ebx, 4(%esp)
+	movl	20(%esp), %esi
+	xorl	%ecx, %edx
+	leal	(%ebp, %ebx), %ecx
+	roll	$9, %ecx
+	xorl	%ecx, %esi
+	movl	%esi, 56(%esp)
+	movl	48(%esp), %ecx
+	movl	%edx, 20(%esp)
+	movl	36(%esp), %edx
+	addl	%esi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %edi
+	movl	24(%esp), %ebx
+	movl	%edi, 24(%esp)
+	addl	%esi, %edi
+	roll	$18, %edi
+	leal	(%ecx, %edx), %esi
+	roll	$7, %esi
+	xorl	%esi, %ebx
+	movl	%ebx, 60(%esp)
+	movl	12(%esp), %esi
+	xorl	%edi, %ebp
+	leal	(%edx, %ebx), %edi
+	roll	$9, %edi
+	xorl	%edi, %esi
+	movl	%esi, 12(%esp)
+	movl	52(%esp), %edi
+	movl	%ebp, 36(%esp)
+	movl	8(%esp), %ebp
+	addl	%esi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %ecx
+	movl	16(%esp), %ebx
+	movl	%ecx, 16(%esp)
+	addl	%esi, %ecx
+	roll	$18, %ecx
+	leal	(%edi, %ebp), %esi
+	roll	$7, %esi
+	xorl	%esi, %ebx
+	movl	32(%esp), %esi
+	xorl	%ecx, %edx
+	leal	(%ebp, %ebx), %ecx
+	roll	$9, %ecx
+	xorl	%ecx, %esi
+	movl	%esi, 32(%esp)
+	movl	%ebx, %ecx
+	movl	%edx, 48(%esp)
+	movl	20(%esp), %edx
+	addl	%esi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %edi
+	movl	24(%esp), %ebx
+	movl	%edi, 20(%esp)
+	addl	%esi, %edi
+	roll	$18, %edi
+	leal	(%ecx, %edx), %esi
+	roll	$7, %esi
+	xorl	%esi, %ebx
+	movl	%ebx, 8(%esp)
+	movl	12(%esp), %esi
+	xorl	%edi, %ebp
+	leal	(%edx, %ebx), %edi
+	roll	$9, %edi
+	xorl	%edi, %esi
+	movl	%esi, 12(%esp)
+	movl	28(%esp), %edi
+	movl	%ebp, 52(%esp)
+	movl	36(%esp), %ebp
+	addl	%esi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %ecx
+	movl	16(%esp), %ebx
+	movl	%ecx, 16(%esp)
+	addl	%esi, %ecx
+	roll	$18, %ecx
+	leal	(%edi, %ebp), %esi
+	roll	$7, %esi
+	xorl	%esi, %ebx
+	movl	%ebx, 28(%esp)
+	movl	32(%esp), %esi
+	xorl	%ecx, %edx
+	leal	(%ebp, %ebx), %ecx
+	roll	$9, %ecx
+	xorl	%ecx, %esi
+	movl	%esi, 32(%esp)
+	movl	4(%esp), %ecx
+	movl	%edx, 4(%esp)
+	movl	48(%esp), %edx
+	addl	%esi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %edi
+	movl	20(%esp), %ebx
+	movl	%edi, 20(%esp)
+	addl	%esi, %edi
+	roll	$18, %edi
+	leal	(%ecx, %edx), %esi
+	roll	$7, %esi
+	xorl	%esi, %ebx
+	movl	%ebx, 48(%esp)
+	movl	40(%esp), %esi
+	xorl	%edi, %ebp
+	leal	(%edx, %ebx), %edi
+	roll	$9, %edi
+	xorl	%edi, %esi
+	movl	%esi, 36(%esp)
+	movl	60(%esp), %edi
+	movl	%ebp, 24(%esp)
+	movl	52(%esp), %ebp
+	addl	%esi, %ebx
+	roll	$13, %ebx
+	xorl	%ebx, %ecx
+	movl	44(%esp), %ebx
+	movl	%ecx, 40(%esp)
+	addl	%esi, %ecx
+	roll	$18, %ecx
+	leal	(%edi, %ebp), %esi
+	roll	$7, %esi
+	xorl	%esi, %ebx
+	movl	%ebx, 52(%esp)
+	movl	56(%esp), %esi
+	xorl	%ecx, %edx
+	leal	(%ebp, %ebx), %ecx
+	roll	$9, %ecx
+	xorl	%ecx, %esi
+	movl	%esi, 56(%esp)
+	addl	%esi, %ebx
+	movl	%edx, 44(%esp)
+	roll	$13, %ebx
+	xorl	%ebx, %edi
+	movl	%edi, 60(%esp)
+	addl	%esi, %edi
+	roll	$18, %edi
+	xorl	%edi, %ebp
+	movl	%ebp, 64(%esp)
+.endm
 
 	.text
-	.align 32
-gen_salsa8_core:
-	gen_salsa8_core_quadround()
-	gen_salsa8_core_quadround()
+	.p2align 5
+salsa8_core_gen:
+	salsa8_core_gen_quadround
+	salsa8_core_gen_quadround
 	ret
 	
 	
 	.text
-	.align 32
+	.p2align 5
 	.globl scrypt_core
 	.globl _scrypt_core
 scrypt_core:
@@ -367,191 +404,191 @@ _scrypt_core:
 	pushl	%edi
 	pushl	%esi
 	
-	# Check for SSE2 availability
+	/* Check for SSE2 availability */
 	movl	$1, %eax
 	cpuid
 	andl	$0x04000000, %edx
-	jnz xmm_scrypt_core
+	jnz scrypt_core_sse2
 	
-gen_scrypt_core:
+scrypt_core_gen:
 	movl	20(%esp), %edi
 	movl	24(%esp), %esi
 	subl	$72, %esp
 	
-#define scrypt_core_macro1a(p, q) \
-	movl	p(%edi), %eax; \
-	movl	q(%edi), %edx; \
-	movl	%eax, p(%esi); \
-	movl	%edx, q(%esi); \
-	xorl	%edx, %eax; \
-	movl	%eax, p(%edi); \
-	movl	%eax, p(%esp); \
-
-	
-#define scrypt_core_macro1b(p, q) \
-	movl	p(%edi), %eax; \
-	xorl	p(%esi, %edx), %eax; \
-	movl	q(%edi), %ebx; \
-	xorl	q(%esi, %edx), %ebx; \
-	movl	%ebx, q(%edi); \
-	xorl	%ebx, %eax; \
-	movl	%eax, p(%edi); \
-	movl	%eax, p(%esp); \
-
-	
-#define scrypt_core_macro2(p, q) \
-	movl	p(%esp), %eax; \
-	addl	p(%edi), %eax; \
-	movl	%eax, p(%edi); \
-	xorl	q(%edi), %eax; \
-	movl	%eax, q(%edi); \
-	movl	%eax, p(%esp); \
-
-	
-#define scrypt_core_macro3(p, q) \
-	movl	p(%esp), %eax; \
-	addl	q(%edi), %eax; \
-	movl	%eax, q(%edi); \
-
+.macro scrypt_core_macro1a p, q
+	movl	\p(%edi), %eax
+	movl	\q(%edi), %edx
+	movl	%eax, \p(%esi)
+	movl	%edx, \q(%esi)
+	xorl	%edx, %eax
+	movl	%eax, \p(%edi)
+	movl	%eax, \p(%esp)
+.endm
+	
+.macro scrypt_core_macro1b p, q
+	movl	\p(%edi), %eax
+	xorl	\p(%esi, %edx), %eax
+	movl	\q(%edi), %ebx
+	xorl	\q(%esi, %edx), %ebx
+	movl	%ebx, \q(%edi)
+	xorl	%ebx, %eax
+	movl	%eax, \p(%edi)
+	movl	%eax, \p(%esp)
+.endm
+	
+.macro scrypt_core_macro2 p, q
+	movl	\p(%esp), %eax
+	addl	\p(%edi), %eax
+	movl	%eax, \p(%edi)
+	xorl	\q(%edi), %eax
+	movl	%eax, \q(%edi)
+	movl	%eax, \p(%esp)
+.endm
+	
+.macro scrypt_core_macro3 p, q
+	movl	\p(%esp), %eax
+	addl	\q(%edi), %eax
+	movl	%eax, \q(%edi)
+.endm
 	
 	leal	131072(%esi), %ecx
-gen_scrypt_core_loop1:
+scrypt_core_gen_loop1:
 	movl	%esi, 64(%esp)
 	movl	%ecx, 68(%esp)
 	
-	scrypt_core_macro1a(0, 64)
-	scrypt_core_macro1a(4, 68)
-	scrypt_core_macro1a(8, 72)
-	scrypt_core_macro1a(12, 76)
-	scrypt_core_macro1a(16, 80)
-	scrypt_core_macro1a(20, 84)
-	scrypt_core_macro1a(24, 88)
-	scrypt_core_macro1a(28, 92)
-	scrypt_core_macro1a(32, 96)
-	scrypt_core_macro1a(36, 100)
-	scrypt_core_macro1a(40, 104)
-	scrypt_core_macro1a(44, 108)
-	scrypt_core_macro1a(48, 112)
-	scrypt_core_macro1a(52, 116)
-	scrypt_core_macro1a(56, 120)
-	scrypt_core_macro1a(60, 124)
-	
-	call gen_salsa8_core
+	scrypt_core_macro1a	0, 64
+	scrypt_core_macro1a	4, 68
+	scrypt_core_macro1a	8, 72
+	scrypt_core_macro1a	12, 76
+	scrypt_core_macro1a	16, 80
+	scrypt_core_macro1a	20, 84
+	scrypt_core_macro1a	24, 88
+	scrypt_core_macro1a	28, 92
+	scrypt_core_macro1a	32, 96
+	scrypt_core_macro1a	36, 100
+	scrypt_core_macro1a	40, 104
+	scrypt_core_macro1a	44, 108
+	scrypt_core_macro1a	48, 112
+	scrypt_core_macro1a	52, 116
+	scrypt_core_macro1a	56, 120
+	scrypt_core_macro1a	60, 124
+	
+	call salsa8_core_gen
 	
 	movl	92(%esp), %edi
-	scrypt_core_macro2(0, 64)
-	scrypt_core_macro2(4, 68)
-	scrypt_core_macro2(8, 72)
-	scrypt_core_macro2(12, 76)
-	scrypt_core_macro2(16, 80)
-	scrypt_core_macro2(20, 84)
-	scrypt_core_macro2(24, 88)
-	scrypt_core_macro2(28, 92)
-	scrypt_core_macro2(32, 96)
-	scrypt_core_macro2(36, 100)
-	scrypt_core_macro2(40, 104)
-	scrypt_core_macro2(44, 108)
-	scrypt_core_macro2(48, 112)
-	scrypt_core_macro2(52, 116)
-	scrypt_core_macro2(56, 120)
-	scrypt_core_macro2(60, 124)
-	
-	call gen_salsa8_core
+	scrypt_core_macro2	0, 64
+	scrypt_core_macro2	4, 68
+	scrypt_core_macro2	8, 72
+	scrypt_core_macro2	12, 76
+	scrypt_core_macro2	16, 80
+	scrypt_core_macro2	20, 84
+	scrypt_core_macro2	24, 88
+	scrypt_core_macro2	28, 92
+	scrypt_core_macro2	32, 96
+	scrypt_core_macro2	36, 100
+	scrypt_core_macro2	40, 104
+	scrypt_core_macro2	44, 108
+	scrypt_core_macro2	48, 112
+	scrypt_core_macro2	52, 116
+	scrypt_core_macro2	56, 120
+	scrypt_core_macro2	60, 124
+	
+	call salsa8_core_gen
 	
 	movl	92(%esp), %edi
-	scrypt_core_macro3(0, 64)
-	scrypt_core_macro3(4, 68)
-	scrypt_core_macro3(8, 72)
-	scrypt_core_macro3(12, 76)
-	scrypt_core_macro3(16, 80)
-	scrypt_core_macro3(20, 84)
-	scrypt_core_macro3(24, 88)
-	scrypt_core_macro3(28, 92)
-	scrypt_core_macro3(32, 96)
-	scrypt_core_macro3(36, 100)
-	scrypt_core_macro3(40, 104)
-	scrypt_core_macro3(44, 108)
-	scrypt_core_macro3(48, 112)
-	scrypt_core_macro3(52, 116)
-	scrypt_core_macro3(56, 120)
-	scrypt_core_macro3(60, 124)
+	scrypt_core_macro3	0, 64
+	scrypt_core_macro3	4, 68
+	scrypt_core_macro3	8, 72
+	scrypt_core_macro3	12, 76
+	scrypt_core_macro3	16, 80
+	scrypt_core_macro3	20, 84
+	scrypt_core_macro3	24, 88
+	scrypt_core_macro3	28, 92
+	scrypt_core_macro3	32, 96
+	scrypt_core_macro3	36, 100
+	scrypt_core_macro3	40, 104
+	scrypt_core_macro3	44, 108
+	scrypt_core_macro3	48, 112
+	scrypt_core_macro3	52, 116
+	scrypt_core_macro3	56, 120
+	scrypt_core_macro3	60, 124
 	
 	movl	64(%esp), %esi
 	movl	68(%esp), %ecx
 	addl	$128, %esi
 	cmpl	%ecx, %esi
-	jne gen_scrypt_core_loop1
+	jne scrypt_core_gen_loop1
 
 	movl	96(%esp), %esi
 	movl	$1024, %ecx
-gen_scrypt_core_loop2:
+scrypt_core_gen_loop2:
 	movl	%ecx, 68(%esp)
 	
 	movl	64(%edi), %edx
 	andl	$1023, %edx
 	shll	$7, %edx
 	
-	scrypt_core_macro1b(0, 64)
-	scrypt_core_macro1b(4, 68)
-	scrypt_core_macro1b(8, 72)
-	scrypt_core_macro1b(12, 76)
-	scrypt_core_macro1b(16, 80)
-	scrypt_core_macro1b(20, 84)
-	scrypt_core_macro1b(24, 88)
-	scrypt_core_macro1b(28, 92)
-	scrypt_core_macro1b(32, 96)
-	scrypt_core_macro1b(36, 100)
-	scrypt_core_macro1b(40, 104)
-	scrypt_core_macro1b(44, 108)
-	scrypt_core_macro1b(48, 112)
-	scrypt_core_macro1b(52, 116)
-	scrypt_core_macro1b(56, 120)
-	scrypt_core_macro1b(60, 124)
-	
-	call gen_salsa8_core
+	scrypt_core_macro1b	0, 64
+	scrypt_core_macro1b	4, 68
+	scrypt_core_macro1b	8, 72
+	scrypt_core_macro1b	12, 76
+	scrypt_core_macro1b	16, 80
+	scrypt_core_macro1b	20, 84
+	scrypt_core_macro1b	24, 88
+	scrypt_core_macro1b	28, 92
+	scrypt_core_macro1b	32, 96
+	scrypt_core_macro1b	36, 100
+	scrypt_core_macro1b	40, 104
+	scrypt_core_macro1b	44, 108
+	scrypt_core_macro1b	48, 112
+	scrypt_core_macro1b	52, 116
+	scrypt_core_macro1b	56, 120
+	scrypt_core_macro1b	60, 124
+	
+	call salsa8_core_gen
 	
 	movl	92(%esp), %edi
-	scrypt_core_macro2(0, 64)
-	scrypt_core_macro2(4, 68)
-	scrypt_core_macro2(8, 72)
-	scrypt_core_macro2(12, 76)
-	scrypt_core_macro2(16, 80)
-	scrypt_core_macro2(20, 84)
-	scrypt_core_macro2(24, 88)
-	scrypt_core_macro2(28, 92)
-	scrypt_core_macro2(32, 96)
-	scrypt_core_macro2(36, 100)
-	scrypt_core_macro2(40, 104)
-	scrypt_core_macro2(44, 108)
-	scrypt_core_macro2(48, 112)
-	scrypt_core_macro2(52, 116)
-	scrypt_core_macro2(56, 120)
-	scrypt_core_macro2(60, 124)
-	
-	call gen_salsa8_core
+	scrypt_core_macro2	0, 64
+	scrypt_core_macro2	4, 68
+	scrypt_core_macro2	8, 72
+	scrypt_core_macro2	12, 76
+	scrypt_core_macro2	16, 80
+	scrypt_core_macro2	20, 84
+	scrypt_core_macro2	24, 88
+	scrypt_core_macro2	28, 92
+	scrypt_core_macro2	32, 96
+	scrypt_core_macro2	36, 100
+	scrypt_core_macro2	40, 104
+	scrypt_core_macro2	44, 108
+	scrypt_core_macro2	48, 112
+	scrypt_core_macro2	52, 116
+	scrypt_core_macro2	56, 120
+	scrypt_core_macro2	60, 124
+	
+	call salsa8_core_gen
 	
 	movl	92(%esp), %edi
 	movl	96(%esp), %esi
-	scrypt_core_macro3(0, 64)
-	scrypt_core_macro3(4, 68)
-	scrypt_core_macro3(8, 72)
-	scrypt_core_macro3(12, 76)
-	scrypt_core_macro3(16, 80)
-	scrypt_core_macro3(20, 84)
-	scrypt_core_macro3(24, 88)
-	scrypt_core_macro3(28, 92)
-	scrypt_core_macro3(32, 96)
-	scrypt_core_macro3(36, 100)
-	scrypt_core_macro3(40, 104)
-	scrypt_core_macro3(44, 108)
-	scrypt_core_macro3(48, 112)
-	scrypt_core_macro3(52, 116)
-	scrypt_core_macro3(56, 120)
-	scrypt_core_macro3(60, 124)
+	scrypt_core_macro3	0, 64
+	scrypt_core_macro3	4, 68
+	scrypt_core_macro3	8, 72
+	scrypt_core_macro3	12, 76
+	scrypt_core_macro3	16, 80
+	scrypt_core_macro3	20, 84
+	scrypt_core_macro3	24, 88
+	scrypt_core_macro3	28, 92
+	scrypt_core_macro3	32, 96
+	scrypt_core_macro3	36, 100
+	scrypt_core_macro3	40, 104
+	scrypt_core_macro3	44, 108
+	scrypt_core_macro3	48, 112
+	scrypt_core_macro3	52, 116
+	scrypt_core_macro3	56, 120
+	scrypt_core_macro3	60, 124
 	
 	movl	68(%esp), %ecx
 	subl	$1, %ecx
-	ja gen_scrypt_core_loop2
+	ja scrypt_core_gen_loop2
 	
 	addl	$72, %esp
 	popl	%esi
@@ -561,167 +598,114 @@ gen_scrypt_core_loop2:
 	ret
 
 
-#define xmm_salsa8_core_doubleround() \
-	movdqa	%xmm1, %xmm4; \
-	paddd	%xmm0, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$7, %xmm4; \
-	psrld	$25, %xmm5; \
-	pxor	%xmm4, %xmm3; \
-	pxor	%xmm5, %xmm3; \
-	movdqa	%xmm0, %xmm4; \
-	paddd	%xmm3, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$9, %xmm4; \
-	psrld	$23, %xmm5; \
-	pxor	%xmm4, %xmm2; \
-	movdqa	%xmm3, %xmm4; \
-	pshufd	$0x93, %xmm3, %xmm3; \
-	pxor	%xmm5, %xmm2; \
-	paddd	%xmm2, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$13, %xmm4; \
-	psrld	$19, %xmm5; \
-	pxor	%xmm4, %xmm1; \
-	movdqa	%xmm2, %xmm4; \
-	pshufd	$0x4e, %xmm2, %xmm2; \
-	pxor	%xmm5, %xmm1; \
-	paddd	%xmm1, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$18, %xmm4; \
-	psrld	$14, %xmm5; \
-	pxor	%xmm4, %xmm0; \
-	pshufd	$0x39, %xmm1, %xmm1; \
-	pxor	%xmm5, %xmm0; \
-	movdqa	%xmm3, %xmm4; \
-	paddd	%xmm0, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$7, %xmm4; \
-	psrld	$25, %xmm5; \
-	pxor	%xmm4, %xmm1; \
-	pxor	%xmm5, %xmm1; \
-	movdqa	%xmm0, %xmm4; \
-	paddd	%xmm1, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$9, %xmm4; \
-	psrld	$23, %xmm5; \
-	pxor	%xmm4, %xmm2; \
-	movdqa	%xmm1, %xmm4; \
-	pshufd	$0x93, %xmm1, %xmm1; \
-	pxor	%xmm5, %xmm2; \
-	paddd	%xmm2, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$13, %xmm4; \
-	psrld	$19, %xmm5; \
-	pxor	%xmm4, %xmm3; \
-	movdqa	%xmm2, %xmm4; \
-	pshufd	$0x4e, %xmm2, %xmm2; \
-	pxor	%xmm5, %xmm3; \
-	paddd	%xmm3, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$18, %xmm4; \
-	psrld	$14, %xmm5; \
-	pxor	%xmm4, %xmm0; \
-	pshufd	$0x39, %xmm3, %xmm3; \
-	pxor	%xmm5, %xmm0; \
-
-
-#define xmm_salsa8_core() \
-	xmm_salsa8_core_doubleround(); \
-	xmm_salsa8_core_doubleround(); \
-	xmm_salsa8_core_doubleround(); \
-	xmm_salsa8_core_doubleround(); \
-
+.macro salsa8_core_sse2_doubleround
+	movdqa	%xmm1, %xmm4
+	paddd	%xmm0, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$7, %xmm4
+	psrld	$25, %xmm5
+	pxor	%xmm4, %xmm3
+	movdqa	%xmm0, %xmm4
+	pxor	%xmm5, %xmm3
+	
+	paddd	%xmm3, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$9, %xmm4
+	psrld	$23, %xmm5
+	pxor	%xmm4, %xmm2
+	movdqa	%xmm3, %xmm4
+	pxor	%xmm5, %xmm2
+	pshufd	$0x93, %xmm3, %xmm3
+	
+	paddd	%xmm2, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$13, %xmm4
+	psrld	$19, %xmm5
+	pxor	%xmm4, %xmm1
+	movdqa	%xmm2, %xmm4
+	pxor	%xmm5, %xmm1
+	pshufd	$0x4e, %xmm2, %xmm2
 	
-	.align 32
-xmm_scrypt_core:
+	paddd	%xmm1, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$18, %xmm4
+	psrld	$14, %xmm5
+	pxor	%xmm4, %xmm0
+	movdqa	%xmm3, %xmm4
+	pxor	%xmm5, %xmm0
+	pshufd	$0x39, %xmm1, %xmm1
+	
+	paddd	%xmm0, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$7, %xmm4
+	psrld	$25, %xmm5
+	pxor	%xmm4, %xmm1
+	movdqa	%xmm0, %xmm4
+	pxor	%xmm5, %xmm1
+	
+	paddd	%xmm1, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$9, %xmm4
+	psrld	$23, %xmm5
+	pxor	%xmm4, %xmm2
+	movdqa	%xmm1, %xmm4
+	pxor	%xmm5, %xmm2
+	pshufd	$0x93, %xmm1, %xmm1
+	
+	paddd	%xmm2, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$13, %xmm4
+	psrld	$19, %xmm5
+	pxor	%xmm4, %xmm3
+	movdqa	%xmm2, %xmm4
+	pxor	%xmm5, %xmm3
+	pshufd	$0x4e, %xmm2, %xmm2
+	
+	paddd	%xmm3, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$18, %xmm4
+	psrld	$14, %xmm5
+	pxor	%xmm4, %xmm0
+	pshufd	$0x39, %xmm3, %xmm3
+	pxor	%xmm5, %xmm0
+.endm
+
+.macro salsa8_core_sse2
+	salsa8_core_sse2_doubleround
+	salsa8_core_sse2_doubleround
+	salsa8_core_sse2_doubleround
+	salsa8_core_sse2_doubleround
+.endm
+	
+	.p2align 5
+scrypt_core_sse2:
 	movl	20(%esp), %edi
 	movl	24(%esp), %esi
 	movl	%esp, %ebp
 	subl	$128, %esp
 	andl	$-16, %esp
 	
-	# shuffle 1st block to (%esp)
-	movl	60(%edi), %edx
-	movl	44(%edi), %ecx
-	movl	28(%edi), %ebx
-	movl	12(%edi), %eax
-	movl	%edx, 12(%esp)
-	movl	%ecx, 28(%esp)
-	movl	%ebx, 44(%esp)
-	movl	%eax, 60(%esp)
-	movl	40(%edi), %ecx
-	movl	24(%edi), %ebx
-	movl	8(%edi), %eax
-	movl	56(%edi), %edx
-	movl	%ecx, 8(%esp)
-	movl	%ebx, 24(%esp)
-	movl	%eax, 40(%esp)
-	movl	%edx, 56(%esp)
-	movl	20(%edi), %ebx
-	movl	4(%edi), %eax
-	movl	52(%edi), %edx
-	movl	36(%edi), %ecx
-	movl	%ebx, 4(%esp)
-	movl	%eax, 20(%esp)
-	movl	%edx, 36(%esp)
-	movl	%ecx, 52(%esp)
-	movl	0(%edi), %eax
-	movl	48(%edi), %edx
-	movl	32(%edi), %ecx
-	movl	16(%edi), %ebx
-	movl	%eax, 0(%esp)
-	movl	%edx, 16(%esp)
-	movl	%ecx, 32(%esp)
-	movl	%ebx, 48(%esp)
+	scrypt_shuffle %edi, 0, %esp, 0
+	scrypt_shuffle %edi, 64, %esp, 64
 	
-	# shuffle 2nd block to 64(%esp)
-	movl	124(%edi), %edx
-	movl	108(%edi), %ecx
-	movl	92(%edi), %ebx
-	movl	76(%edi), %eax
-	movl	%edx, 76(%esp)
-	movl	%ecx, 92(%esp)
-	movl	%ebx, 108(%esp)
-	movl	%eax, 124(%esp)
-	movl	104(%edi), %ecx
-	movl	88(%edi), %ebx
-	movl	72(%edi), %eax
-	movl	120(%edi), %edx
-	movl	%ecx, 72(%esp)
-	movl	%ebx, 88(%esp)
-	movl	%eax, 104(%esp)
-	movl	%edx, 120(%esp)
-	movl	84(%edi), %ebx
-	movl	68(%edi), %eax
-	movl	116(%edi), %edx
-	movl	100(%edi), %ecx
-	movl	%ebx, 68(%esp)
-	movl	%eax, 84(%esp)
-	movl	%edx, 100(%esp)
-	movl	%ecx, 116(%esp)
-	movl	64(%edi), %eax
-	movl	112(%edi), %edx
-	movl	96(%edi), %ecx
-	movl	80(%edi), %ebx
-	movl	%eax, 64(%esp)
-	movl	%edx, 80(%esp)
-	movl	%ecx, 96(%esp)
-	movl	%ebx, 112(%esp)
+	movdqa	96(%esp), %xmm6
+	movdqa	112(%esp), %xmm7
 	
 	movl	%esi, %edx
 	leal	131072(%esi), %ecx
-xmm_scrypt_core_loop1:
+scrypt_core_sse2_loop1:
 	movdqa	0(%esp), %xmm0
 	movdqa	16(%esp), %xmm1
 	movdqa	32(%esp), %xmm2
 	movdqa	48(%esp), %xmm3
 	movdqa	64(%esp), %xmm4
 	movdqa	80(%esp), %xmm5
-	movdqa	96(%esp), %xmm6
-	movdqa	112(%esp), %xmm7
+	pxor	%xmm4, %xmm0
+	pxor	%xmm5, %xmm1
 	movdqa	%xmm0, 0(%edx)
 	movdqa	%xmm1, 16(%edx)
+	pxor	%xmm6, %xmm2
+	pxor	%xmm7, %xmm3
 	movdqa	%xmm2, 32(%edx)
 	movdqa	%xmm3, 48(%edx)
 	movdqa	%xmm4, 64(%edx)
@@ -729,19 +713,11 @@ xmm_scrypt_core_loop1:
 	movdqa	%xmm6, 96(%edx)
 	movdqa	%xmm7, 112(%edx)
 	
-	pxor	%xmm4, %xmm0
-	pxor	%xmm5, %xmm1
-	pxor	%xmm6, %xmm2
-	pxor	%xmm7, %xmm3
-	movdqa	%xmm0, 0(%esp)
-	movdqa	%xmm1, 16(%esp)
-	movdqa	%xmm2, 32(%esp)
-	movdqa	%xmm3, 48(%esp)
-	xmm_salsa8_core()
-	paddd	0(%esp), %xmm0
-	paddd	16(%esp), %xmm1
-	paddd	32(%esp), %xmm2
-	paddd	48(%esp), %xmm3
+	salsa8_core_sse2
+	paddd	0(%edx), %xmm0
+	paddd	16(%edx), %xmm1
+	paddd	32(%edx), %xmm2
+	paddd	48(%edx), %xmm3
 	movdqa	%xmm0, 0(%esp)
 	movdqa	%xmm1, 16(%esp)
 	movdqa	%xmm2, 32(%esp)
@@ -749,61 +725,50 @@ xmm_scrypt_core_loop1:
 	
 	pxor	64(%esp), %xmm0
 	pxor	80(%esp), %xmm1
-	pxor	96(%esp), %xmm2
-	pxor	112(%esp), %xmm3
+	pxor	%xmm6, %xmm2
+	pxor	%xmm7, %xmm3
 	movdqa	%xmm0, 64(%esp)
 	movdqa	%xmm1, 80(%esp)
-	movdqa	%xmm2, 96(%esp)
-	movdqa	%xmm3, 112(%esp)
-	xmm_salsa8_core()
+	movdqa	%xmm2, %xmm6
+	movdqa	%xmm3, %xmm7
+	salsa8_core_sse2
 	paddd	64(%esp), %xmm0
 	paddd	80(%esp), %xmm1
-	paddd	96(%esp), %xmm2
-	paddd	112(%esp), %xmm3
+	paddd	%xmm2, %xmm6
+	paddd	%xmm3, %xmm7
 	movdqa	%xmm0, 64(%esp)
 	movdqa	%xmm1, 80(%esp)
-	movdqa	%xmm2, 96(%esp)
-	movdqa	%xmm3, 112(%esp)
 	
 	addl	$128, %edx
 	cmpl	%ecx, %edx
-	jne xmm_scrypt_core_loop1
+	jne scrypt_core_sse2_loop1
+	
+	movdqa	64(%esp), %xmm4
+	movdqa	80(%esp), %xmm5
 	
 	movl	$1024, %ecx
-xmm_scrypt_core_loop2:
+scrypt_core_sse2_loop2:
+	movd	%xmm4, %edx
 	movdqa	0(%esp), %xmm0
 	movdqa	16(%esp), %xmm1
 	movdqa	32(%esp), %xmm2
 	movdqa	48(%esp), %xmm3
-	movdqa	64(%esp), %xmm4
-	movdqa	80(%esp), %xmm5
-	movdqa	96(%esp), %xmm6
-	movdqa	112(%esp), %xmm7
-	movd	%xmm4, %edx
 	andl	$1023, %edx
 	shll	$7, %edx
 	pxor	0(%esi, %edx), %xmm0
 	pxor	16(%esi, %edx), %xmm1
 	pxor	32(%esi, %edx), %xmm2
 	pxor	48(%esi, %edx), %xmm3
-	pxor	64(%esi, %edx), %xmm4
-	pxor	80(%esi, %edx), %xmm5
-	pxor	96(%esi, %edx), %xmm6
-	pxor	112(%esi, %edx), %xmm7
-	movdqa	%xmm4, 64(%esp)
-	movdqa	%xmm5, 80(%esp)
-	movdqa	%xmm6, 96(%esp)
-	movdqa	%xmm7, 112(%esp)
 	
 	pxor	%xmm4, %xmm0
 	pxor	%xmm5, %xmm1
-	pxor	%xmm6, %xmm2
-	pxor	%xmm7, %xmm3
 	movdqa	%xmm0, 0(%esp)
 	movdqa	%xmm1, 16(%esp)
+	pxor	%xmm6, %xmm2
+	pxor	%xmm7, %xmm3
 	movdqa	%xmm2, 32(%esp)
 	movdqa	%xmm3, 48(%esp)
-	xmm_salsa8_core()
+	salsa8_core_sse2
 	paddd	0(%esp), %xmm0
 	paddd	16(%esp), %xmm1
 	paddd	32(%esp), %xmm2
@@ -813,94 +778,36 @@ xmm_scrypt_core_loop2:
 	movdqa	%xmm2, 32(%esp)
 	movdqa	%xmm3, 48(%esp)
 	
+	pxor	64(%esi, %edx), %xmm0
+	pxor	80(%esi, %edx), %xmm1
+	pxor	96(%esi, %edx), %xmm2
+	pxor	112(%esi, %edx), %xmm3
 	pxor	64(%esp), %xmm0
 	pxor	80(%esp), %xmm1
-	pxor	96(%esp), %xmm2
-	pxor	112(%esp), %xmm3
+	pxor	%xmm6, %xmm2
+	pxor	%xmm7, %xmm3
 	movdqa	%xmm0, 64(%esp)
 	movdqa	%xmm1, 80(%esp)
-	movdqa	%xmm2, 96(%esp)
-	movdqa	%xmm3, 112(%esp)
-	xmm_salsa8_core()
+	movdqa	%xmm2, %xmm6
+	movdqa	%xmm3, %xmm7
+	salsa8_core_sse2
 	paddd	64(%esp), %xmm0
 	paddd	80(%esp), %xmm1
-	paddd	96(%esp), %xmm2
-	paddd	112(%esp), %xmm3
+	paddd	%xmm2, %xmm6
+	paddd	%xmm3, %xmm7
+	movdqa	%xmm0, %xmm4
+	movdqa	%xmm1, %xmm5
 	movdqa	%xmm0, 64(%esp)
 	movdqa	%xmm1, 80(%esp)
-	movdqa	%xmm2, 96(%esp)
-	movdqa	%xmm3, 112(%esp)
 	
 	subl	$1, %ecx
-	ja xmm_scrypt_core_loop2
+	ja scrypt_core_sse2_loop2
 	
-	# re-shuffle 1st block back
-	movl	60(%esp), %edx
-	movl	44(%esp), %ecx
-	movl	28(%esp), %ebx
-	movl	12(%esp), %eax
-	movl	%edx, 12(%edi)
-	movl	%ecx, 28(%edi)
-	movl	%ebx, 44(%edi)
-	movl	%eax, 60(%edi)
-	movl	40(%esp), %ecx
-	movl	24(%esp), %ebx
-	movl	8(%esp), %eax
-	movl	56(%esp), %edx
-	movl	%ecx, 8(%edi)
-	movl	%ebx, 24(%edi)
-	movl	%eax, 40(%edi)
-	movl	%edx, 56(%edi)
-	movl	20(%esp), %ebx
-	movl	4(%esp), %eax
-	movl	52(%esp), %edx
-	movl	36(%esp), %ecx
-	movl	%ebx, 4(%edi)
-	movl	%eax, 20(%edi)
-	movl	%edx, 36(%edi)
-	movl	%ecx, 52(%edi)
-	movl	0(%esp), %eax
-	movl	48(%esp), %edx
-	movl	32(%esp), %ecx
-	movl	16(%esp), %ebx
-	movl	%eax, 0(%edi)
-	movl	%edx, 16(%edi)
-	movl	%ecx, 32(%edi)
-	movl	%ebx, 48(%edi)
-	
-	# re-shuffle 2nd block back
-	movl	124(%esp), %edx
-	movl	108(%esp), %ecx
-	movl	92(%esp), %ebx
-	movl	76(%esp), %eax
-	movl	%edx, 76(%edi)
-	movl	%ecx, 92(%edi)
-	movl	%ebx, 108(%edi)
-	movl	%eax, 124(%edi)
-	movl	104(%esp), %ecx
-	movl	88(%esp), %ebx
-	movl	72(%esp), %eax
-	movl	120(%esp), %edx
-	movl	%ecx, 72(%edi)
-	movl	%ebx, 88(%edi)
-	movl	%eax, 104(%edi)
-	movl	%edx, 120(%edi)
-	movl	84(%esp), %ebx
-	movl	68(%esp), %eax
-	movl	116(%esp), %edx
-	movl	100(%esp), %ecx
-	movl	%ebx, 68(%edi)
-	movl	%eax, 84(%edi)
-	movl	%edx, 100(%edi)
-	movl	%ecx, 116(%edi)
-	movl	64(%esp), %eax
-	movl	112(%esp), %edx
-	movl	96(%esp), %ecx
-	movl	80(%esp), %ebx
-	movl	%eax, 64(%edi)
-	movl	%edx, 80(%edi)
-	movl	%ecx, 96(%edi)
-	movl	%ebx, 112(%edi)
+	movdqa	%xmm6, 96(%esp)
+	movdqa	%xmm7, 112(%esp)
+	
+	scrypt_shuffle %esp, 0, %edi, 0
+	scrypt_shuffle %esp, 64, %edi, 64
 	
 	movl	%ebp, %esp
 	popl	%esi
diff --git a/src/scrypt-x86_64.S b/src/scrypt-x86_64.S
index 21ef9a3..36054f1 100644
--- a/src/scrypt-x86_64.S
+++ b/src/scrypt-x86_64.S
@@ -1,180 +1,195 @@
-# Copyright 2011-2012 pooler@litecoinpool.org
-# All rights reserved.
-#
-# Redistribution and use in source and binary forms, with or without
-# modification, are permitted provided that the following conditions
-# are met:
-# 1. Redistributions of source code must retain the above copyright
-#    notice, this list of conditions and the following disclaimer.
-# 2. Redistributions in binary form must reproduce the above copyright
-#    notice, this list of conditions and the following disclaimer in the
-#    documentation and/or other materials provided with the distribution.
-#
-# THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
-# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
-# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
-# ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
-# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
-# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
-# OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
-# HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
-# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
-# OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
-# SUCH DAMAGE.
-
-
-#if defined(__x86_64__)
+/*
+ * Copyright 2011-2012 pooler@litecoinpool.org
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
 
 #if defined(__linux__) && defined(__ELF__)
 	.section .note.GNU-stack,"",%progbits
 #endif
 
-#define scrypt_shuffle(src, so, dest, do) \
-	movl	so+60(src), %r8d; \
-	movl	so+44(src), %r9d; \
-	movl	so+28(src), %r10d; \
-	movl	so+12(src), %r11d; \
-	movl	%r8d, do+12(dest); \
-	movl	%r9d, do+28(dest); \
-	movl	%r10d, do+44(dest); \
-	movl	%r11d, do+60(dest); \
-	movl	so+40(src), %r8d; \
-	movl	so+8(src), %r9d; \
-	movl	so+48(src), %r10d; \
-	movl	so+16(src), %r11d; \
-	movl	%r8d, do+8(dest); \
-	movl	%r9d, do+40(dest); \
-	movl	%r10d, do+16(dest); \
-	movl	%r11d, do+48(dest); \
-	movl	so+20(src), %r8d; \
-	movl	so+4(src), %r9d; \
-	movl	so+52(src), %r10d; \
-	movl	so+36(src), %r11d; \
-	movl	%r8d, do+4(dest); \
-	movl	%r9d, do+20(dest); \
-	movl	%r10d, do+36(dest); \
-	movl	%r11d, do+52(dest); \
-	movl	so+0(src), %r8d; \
-	movl	so+24(src), %r9d; \
-	movl	so+32(src), %r10d; \
-	movl	so+56(src), %r11d; \
-	movl	%r8d, do+0(dest); \
-	movl	%r9d, do+24(dest); \
-	movl	%r10d, do+32(dest); \
-	movl	%r11d, do+56(dest); \
-
+#if defined(__x86_64__)
 
+.macro scrypt_shuffle src, so, dest, do
+	movl	\so+60(\src), %r8d
+	movl	\so+44(\src), %r9d
+	movl	\so+28(\src), %r10d
+	movl	\so+12(\src), %r11d
+	movl	%r8d, \do+12(\dest)
+	movl	%r9d, \do+28(\dest)
+	movl	%r10d, \do+44(\dest)
+	movl	%r11d, \do+60(\dest)
+	movl	\so+40(\src), %r8d
+	movl	\so+8(\src), %r9d
+	movl	\so+48(\src), %r10d
+	movl	\so+16(\src), %r11d
+	movl	%r8d, \do+8(\dest)
+	movl	%r9d, \do+40(\dest)
+	movl	%r10d, \do+16(\dest)
+	movl	%r11d, \do+48(\dest)
+	movl	\so+20(\src), %r8d
+	movl	\so+4(\src), %r9d
+	movl	\so+52(\src), %r10d
+	movl	\so+36(\src), %r11d
+	movl	%r8d, \do+4(\dest)
+	movl	%r9d, \do+20(\dest)
+	movl	%r10d, \do+36(\dest)
+	movl	%r11d, \do+52(\dest)
+	movl	\so+0(\src), %r8d
+	movl	\so+24(\src), %r9d
+	movl	\so+32(\src), %r10d
+	movl	\so+56(\src), %r11d
+	movl	%r8d, \do+0(\dest)
+	movl	%r9d, \do+24(\dest)
+	movl	%r10d, \do+32(\dest)
+	movl	%r11d, \do+56(\dest)
+.endm
 
-#define salsa8_core_gen_doubleround() \
-	movq	72(%rsp), %r15; \
-	leaq	(%r14, %rdx), %rbp; \
-	roll	$7, %ebp; \
-	xorl	%ebp, %r9d; \
-	leaq	(%rdi, %r15), %rbp; \
-	roll	$7, %ebp; \
-	xorl	%ebp, %r10d; \
-	leaq	(%rdx, %r9), %rbp; \
-	roll	$9, %ebp; \
-	xorl	%ebp, %r11d; \
-	leaq	(%r15, %r10), %rbp; \
-	roll	$9, %ebp; \
-	xorl	%ebp, %r13d; \
-	leaq	(%r9, %r11), %rbp; \
-	roll	$13, %ebp; \
-	xorl	%ebp, %r14d; \
-	leaq	(%r10, %r13), %rbp; \
-	roll	$13, %ebp; \
-	xorl	%ebp, %edi; \
-	leaq	(%r11, %r14), %rbp; \
-	roll	$18, %ebp; \
-	xorl	%ebp, %edx; \
-	leaq	(%r13, %rdi), %rbp; \
-	roll	$18, %ebp; \
-	xorl	%ebp, %r15d; \
-	movq	48(%rsp), %rbp; \
-	movq	%r15, 72(%rsp); \
-	leaq	(%rax, %rbp), %r15; \
-	roll	$7, %r15d; \
-	xorl	%r15d, %ebx; \
-	leaq	(%rbp, %rbx), %r15; \
-	roll	$9, %r15d; \
-	xorl	%r15d, %ecx; \
-	leaq	(%rbx, %rcx), %r15; \
-	roll	$13, %r15d; \
-	xorl	%r15d, %eax; \
-	leaq	(%rcx, %rax), %r15; \
-	roll	$18, %r15d; \
-	xorl	%r15d, %ebp; \
-	movq	88(%rsp), %r15; \
-	movq	%rbp, 48(%rsp); \
-	leaq	(%r12, %r15), %rbp; \
-	roll	$7, %ebp; \
-	xorl	%ebp, %esi; \
-	leaq	(%r15, %rsi), %rbp; \
-	roll	$9, %ebp; \
-	xorl	%ebp, %r8d; \
-	leaq	(%rsi, %r8), %rbp; \
-	roll	$13, %ebp; \
-	xorl	%ebp, %r12d; \
-	leaq	(%r8, %r12), %rbp; \
-	roll	$18, %ebp; \
-	xorl	%ebp, %r15d; \
-	movq	%r15, 88(%rsp); \
-	movq	72(%rsp), %r15; \
-	leaq	(%rsi, %rdx), %rbp; \
-	roll	$7, %ebp; \
-	xorl	%ebp, %edi; \
-	leaq	(%r9, %r15), %rbp; \
-	roll	$7, %ebp; \
-	xorl	%ebp, %eax; \
-	leaq	(%rdx, %rdi), %rbp; \
-	roll	$9, %ebp; \
-	xorl	%ebp, %ecx; \
-	leaq	(%r15, %rax), %rbp; \
-	roll	$9, %ebp; \
-	xorl	%ebp, %r8d; \
-	leaq	(%rdi, %rcx), %rbp; \
-	roll	$13, %ebp; \
-	xorl	%ebp, %esi; \
-	leaq	(%rax, %r8), %rbp; \
-	roll	$13, %ebp; \
-	xorl	%ebp, %r9d; \
-	leaq	(%rcx, %rsi), %rbp; \
-	roll	$18, %ebp; \
-	xorl	%ebp, %edx; \
-	leaq	(%r8, %r9), %rbp; \
-	roll	$18, %ebp; \
-	xorl	%ebp, %r15d; \
-	movq	48(%rsp), %rbp; \
-	movq	%r15, 72(%rsp); \
-	leaq	(%r10, %rbp), %r15; \
-	roll	$7, %r15d; \
-	xorl	%r15d, %r12d; \
-	leaq	(%rbp, %r12), %r15; \
-	roll	$9, %r15d; \
-	xorl	%r15d, %r11d; \
-	leaq	(%r12, %r11), %r15; \
-	roll	$13, %r15d; \
-	xorl	%r15d, %r10d; \
-	leaq	(%r11, %r10), %r15; \
-	roll	$18, %r15d; \
-	xorl	%r15d, %ebp; \
-	movq	88(%rsp), %r15; \
-	movq	%rbp, 48(%rsp); \
-	leaq	(%rbx, %r15), %rbp; \
-	roll	$7, %ebp; \
-	xorl	%ebp, %r14d; \
-	leaq	(%r15, %r14), %rbp; \
-	roll	$9, %ebp; \
-	xorl	%ebp, %r13d; \
-	leaq	(%r14, %r13), %rbp; \
-	roll	$13, %ebp; \
-	xorl	%ebp, %ebx; \
-	leaq	(%r13, %rbx), %rbp; \
-	roll	$18, %ebp; \
-	xorl	%ebp, %r15d; \
-	movq	%r15, 88(%rsp); \
 
+.macro salsa8_core_gen_doubleround
+	movq	72(%rsp), %r15
+	
+	leaq	(%r14, %rdx), %rbp
+	roll	$7, %ebp
+	xorl	%ebp, %r9d
+	leaq	(%rdi, %r15), %rbp
+	roll	$7, %ebp
+	xorl	%ebp, %r10d
+	leaq	(%rdx, %r9), %rbp
+	roll	$9, %ebp
+	xorl	%ebp, %r11d
+	leaq	(%r15, %r10), %rbp
+	roll	$9, %ebp
+	xorl	%ebp, %r13d
+	
+	leaq	(%r9, %r11), %rbp
+	roll	$13, %ebp
+	xorl	%ebp, %r14d
+	leaq	(%r10, %r13), %rbp
+	roll	$13, %ebp
+	xorl	%ebp, %edi
+	leaq	(%r11, %r14), %rbp
+	roll	$18, %ebp
+	xorl	%ebp, %edx
+	leaq	(%r13, %rdi), %rbp
+	roll	$18, %ebp
+	xorl	%ebp, %r15d
+	
+	movq	48(%rsp), %rbp
+	movq	%r15, 72(%rsp)
+	
+	leaq	(%rax, %rbp), %r15
+	roll	$7, %r15d
+	xorl	%r15d, %ebx
+	leaq	(%rbp, %rbx), %r15
+	roll	$9, %r15d
+	xorl	%r15d, %ecx
+	leaq	(%rbx, %rcx), %r15
+	roll	$13, %r15d
+	xorl	%r15d, %eax
+	leaq	(%rcx, %rax), %r15
+	roll	$18, %r15d
+	xorl	%r15d, %ebp
+	
+	movq	88(%rsp), %r15
+	movq	%rbp, 48(%rsp)
+	
+	leaq	(%r12, %r15), %rbp
+	roll	$7, %ebp
+	xorl	%ebp, %esi
+	leaq	(%r15, %rsi), %rbp
+	roll	$9, %ebp
+	xorl	%ebp, %r8d
+	leaq	(%rsi, %r8), %rbp
+	roll	$13, %ebp
+	xorl	%ebp, %r12d
+	leaq	(%r8, %r12), %rbp
+	roll	$18, %ebp
+	xorl	%ebp, %r15d
+	
+	movq	%r15, 88(%rsp)
+	movq	72(%rsp), %r15
+	
+	leaq	(%rsi, %rdx), %rbp
+	roll	$7, %ebp
+	xorl	%ebp, %edi
+	leaq	(%r9, %r15), %rbp
+	roll	$7, %ebp
+	xorl	%ebp, %eax
+	leaq	(%rdx, %rdi), %rbp
+	roll	$9, %ebp
+	xorl	%ebp, %ecx
+	leaq	(%r15, %rax), %rbp
+	roll	$9, %ebp
+	xorl	%ebp, %r8d
+	
+	leaq	(%rdi, %rcx), %rbp
+	roll	$13, %ebp
+	xorl	%ebp, %esi
+	leaq	(%rax, %r8), %rbp
+	roll	$13, %ebp
+	xorl	%ebp, %r9d
+	leaq	(%rcx, %rsi), %rbp
+	roll	$18, %ebp
+	xorl	%ebp, %edx
+	leaq	(%r8, %r9), %rbp
+	roll	$18, %ebp
+	xorl	%ebp, %r15d
+	
+	movq	48(%rsp), %rbp
+	movq	%r15, 72(%rsp)
+	
+	leaq	(%r10, %rbp), %r15
+	roll	$7, %r15d
+	xorl	%r15d, %r12d
+	leaq	(%rbp, %r12), %r15
+	roll	$9, %r15d
+	xorl	%r15d, %r11d
+	leaq	(%r12, %r11), %r15
+	roll	$13, %r15d
+	xorl	%r15d, %r10d
+	leaq	(%r11, %r10), %r15
+	roll	$18, %r15d
+	xorl	%r15d, %ebp
+	
+	movq	88(%rsp), %r15
+	movq	%rbp, 48(%rsp)
+	
+	leaq	(%rbx, %r15), %rbp
+	roll	$7, %ebp
+	xorl	%ebp, %r14d
+	leaq	(%r15, %r14), %rbp
+	roll	$9, %ebp
+	xorl	%ebp, %r13d
+	leaq	(%r14, %r13), %rbp
+	roll	$13, %ebp
+	xorl	%ebp, %ebx
+	leaq	(%r13, %rbx), %rbp
+	roll	$18, %ebp
+	xorl	%ebp, %r15d
+	
+	movq	%r15, 88(%rsp)
+.endm
 
 	.text
 	.p2align 6
@@ -211,10 +226,10 @@ salsa8_core_gen:
 	shrq	$32, %r15
 	movq	%r15, 88(%rsp)
 	
-	salsa8_core_gen_doubleround()
-	salsa8_core_gen_doubleround()
-	salsa8_core_gen_doubleround()
-	salsa8_core_gen_doubleround()
+	salsa8_core_gen_doubleround
+	salsa8_core_gen_doubleround
+	salsa8_core_gen_doubleround
+	salsa8_core_gen_doubleround
 	
 	shlq	$32, %rdi
 	xorq	%rdi, %rdx
@@ -289,39 +304,30 @@ _scrypt_core:
 	movq	%rdx, %rsi
 #endif
 
+.macro scrypt_core_cleanup
 #if defined(WIN64)
-#define scrypt_core_cleanup() \
-	popq	%rsi; \
-	popq	%rdi; \
-	movdqa	8(%rsp), %xmm6; \
-	movdqa	24(%rsp), %xmm7; \
-	movdqa	40(%rsp), %xmm8; \
-	movdqa	56(%rsp), %xmm9; \
-	movdqa	72(%rsp), %xmm10; \
-	movdqa	88(%rsp), %xmm11; \
-	movdqa	104(%rsp), %xmm12; \
-	movdqa	120(%rsp), %xmm13; \
-	movdqa	136(%rsp), %xmm14; \
-	movdqa	152(%rsp), %xmm15; \
-	addq	$176, %rsp; \
-	popq	%r15; \
-	popq	%r14; \
-	popq	%r13; \
-	popq	%r12; \
-	popq	%rbp; \
-	popq	%rbx; \
-	
-#else
-#define scrypt_core_cleanup() \
-	popq	%r15; \
-	popq	%r14; \
-	popq	%r13; \
-	popq	%r12; \
-	popq	%rbp; \
-	popq	%rbx; \
-	
+	popq	%rsi
+	popq	%rdi
+	movdqa	8(%rsp), %xmm6
+	movdqa	24(%rsp), %xmm7
+	movdqa	40(%rsp), %xmm8
+	movdqa	56(%rsp), %xmm9
+	movdqa	72(%rsp), %xmm10
+	movdqa	88(%rsp), %xmm11
+	movdqa	104(%rsp), %xmm12
+	movdqa	120(%rsp), %xmm13
+	movdqa	136(%rsp), %xmm14
+	movdqa	152(%rsp), %xmm15
+	addq	$176, %rsp
 #endif
-
+	popq	%r15
+	popq	%r14
+	popq	%r13
+	popq	%r12
+	popq	%rbp
+	popq	%rbx
+.endm
+	
 	/* GenuineIntel processors have fast SIMD */
 	xorl	%eax, %eax
 	cpuid
@@ -462,81 +468,88 @@ scrypt_core_gen_loop2:
 	movdqa	%xmm15, 112(%rdi)
 	
 	addq	$136, %rsp
-	scrypt_core_cleanup()
+	scrypt_core_cleanup
 	ret
 
 
-#define salsa8_core_xmm_doubleround() \
-	movdqa	%xmm1, %xmm4; \
-	paddd	%xmm0, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$7, %xmm4; \
-	psrld	$25, %xmm5; \
-	pxor	%xmm4, %xmm3; \
-	movdqa	%xmm0, %xmm4; \
-	pxor	%xmm5, %xmm3; \
-	paddd	%xmm3, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$9, %xmm4; \
-	psrld	$23, %xmm5; \
-	pxor	%xmm4, %xmm2; \
-	movdqa	%xmm3, %xmm4; \
-	pxor	%xmm5, %xmm2; \
-	pshufd	$0x93, %xmm3, %xmm3; \
-	paddd	%xmm2, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$13, %xmm4; \
-	psrld	$19, %xmm5; \
-	pxor	%xmm4, %xmm1; \
-	movdqa	%xmm2, %xmm4; \
-	pxor	%xmm5, %xmm1; \
-	pshufd	$0x4e, %xmm2, %xmm2; \
-	paddd	%xmm1, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$18, %xmm4; \
-	psrld	$14, %xmm5; \
-	pxor	%xmm4, %xmm0; \
-	movdqa	%xmm3, %xmm4; \
-	pxor	%xmm5, %xmm0; \
-	pshufd	$0x39, %xmm1, %xmm1; \
-	paddd	%xmm0, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$7, %xmm4; \
-	psrld	$25, %xmm5; \
-	pxor	%xmm4, %xmm1; \
-	movdqa	%xmm0, %xmm4; \
-	pxor	%xmm5, %xmm1; \
-	paddd	%xmm1, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$9, %xmm4; \
-	psrld	$23, %xmm5; \
-	pxor	%xmm4, %xmm2; \
-	movdqa	%xmm1, %xmm4; \
-	pxor	%xmm5, %xmm2; \
-	pshufd	$0x93, %xmm1, %xmm1; \
-	paddd	%xmm2, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$13, %xmm4; \
-	psrld	$19, %xmm5; \
-	pxor	%xmm4, %xmm3; \
-	movdqa	%xmm2, %xmm4; \
-	pxor	%xmm5, %xmm3; \
-	pshufd	$0x4e, %xmm2, %xmm2; \
-	paddd	%xmm3, %xmm4; \
-	movdqa	%xmm4, %xmm5; \
-	pslld	$18, %xmm4; \
-	psrld	$14, %xmm5; \
-	pxor	%xmm4, %xmm0; \
-	pshufd	$0x39, %xmm3, %xmm3; \
-	pxor	%xmm5, %xmm0; \
-
-
-#define salsa8_core_xmm() \
-	salsa8_core_xmm_doubleround(); \
-	salsa8_core_xmm_doubleround(); \
-	salsa8_core_xmm_doubleround(); \
-	salsa8_core_xmm_doubleround(); \
+.macro salsa8_core_xmm_doubleround
+	movdqa	%xmm1, %xmm4
+	paddd	%xmm0, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$7, %xmm4
+	psrld	$25, %xmm5
+	pxor	%xmm4, %xmm3
+	movdqa	%xmm0, %xmm4
+	pxor	%xmm5, %xmm3
+	
+	paddd	%xmm3, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$9, %xmm4
+	psrld	$23, %xmm5
+	pxor	%xmm4, %xmm2
+	movdqa	%xmm3, %xmm4
+	pxor	%xmm5, %xmm2
+	pshufd	$0x93, %xmm3, %xmm3
+	
+	paddd	%xmm2, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$13, %xmm4
+	psrld	$19, %xmm5
+	pxor	%xmm4, %xmm1
+	movdqa	%xmm2, %xmm4
+	pxor	%xmm5, %xmm1
+	pshufd	$0x4e, %xmm2, %xmm2
+	
+	paddd	%xmm1, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$18, %xmm4
+	psrld	$14, %xmm5
+	pxor	%xmm4, %xmm0
+	movdqa	%xmm3, %xmm4
+	pxor	%xmm5, %xmm0
+	pshufd	$0x39, %xmm1, %xmm1
+	
+	paddd	%xmm0, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$7, %xmm4
+	psrld	$25, %xmm5
+	pxor	%xmm4, %xmm1
+	movdqa	%xmm0, %xmm4
+	pxor	%xmm5, %xmm1
+	
+	paddd	%xmm1, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$9, %xmm4
+	psrld	$23, %xmm5
+	pxor	%xmm4, %xmm2
+	movdqa	%xmm1, %xmm4
+	pxor	%xmm5, %xmm2
+	pshufd	$0x93, %xmm1, %xmm1
+	
+	paddd	%xmm2, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$13, %xmm4
+	psrld	$19, %xmm5
+	pxor	%xmm4, %xmm3
+	movdqa	%xmm2, %xmm4
+	pxor	%xmm5, %xmm3
+	pshufd	$0x4e, %xmm2, %xmm2
+	
+	paddd	%xmm3, %xmm4
+	movdqa	%xmm4, %xmm5
+	pslld	$18, %xmm4
+	psrld	$14, %xmm5
+	pxor	%xmm4, %xmm0
+	pshufd	$0x39, %xmm3, %xmm3
+	pxor	%xmm5, %xmm0
+.endm
 
+.macro salsa8_core_xmm
+	salsa8_core_xmm_doubleround
+	salsa8_core_xmm_doubleround
+	salsa8_core_xmm_doubleround
+	salsa8_core_xmm_doubleround
+.endm
 	
 	.p2align 6
 scrypt_core_xmm:
@@ -615,7 +628,7 @@ scrypt_core_xmm_loop1:
 	movdqa	%xmm9, %xmm1
 	movdqa	%xmm10, %xmm2
 	movdqa	%xmm11, %xmm3
-	salsa8_core_xmm()
+	salsa8_core_xmm
 	paddd	%xmm0, %xmm8
 	paddd	%xmm1, %xmm9
 	paddd	%xmm2, %xmm10
@@ -629,7 +642,7 @@ scrypt_core_xmm_loop1:
 	movdqa	%xmm13, %xmm1
 	movdqa	%xmm14, %xmm2
 	movdqa	%xmm15, %xmm3
-	salsa8_core_xmm()
+	salsa8_core_xmm
 	paddd	%xmm0, %xmm12
 	paddd	%xmm1, %xmm13
 	paddd	%xmm2, %xmm14
@@ -657,7 +670,7 @@ scrypt_core_xmm_loop2:
 	movdqa	%xmm9, %xmm1
 	movdqa	%xmm10, %xmm2
 	movdqa	%xmm11, %xmm3
-	salsa8_core_xmm()
+	salsa8_core_xmm
 	paddd	%xmm0, %xmm8
 	paddd	%xmm1, %xmm9
 	paddd	%xmm2, %xmm10
@@ -675,7 +688,7 @@ scrypt_core_xmm_loop2:
 	movdqa	%xmm13, %xmm1
 	movdqa	%xmm14, %xmm2
 	movdqa	%xmm15, %xmm3
-	salsa8_core_xmm()
+	salsa8_core_xmm
 	paddd	%xmm0, %xmm12
 	paddd	%xmm1, %xmm13
 	paddd	%xmm2, %xmm14
@@ -739,7 +752,7 @@ scrypt_core_xmm_loop2:
 	movdqa	%xmm14, 96(%rdi)
 	movdqa	%xmm13, 112(%rdi)
 	
-	scrypt_core_cleanup()
+	scrypt_core_cleanup
 	ret
-	
+
 #endif