#ifndef __POLY1305ASM__
#define __POLY1305ASM__

/* field arithmetic used for computing poly1305 */

#define fe1305_mul_taun(m,t)				\
							\
	movq    m+0(%rsi),%rdx;    			\
							\
	mulx    t+0(%rdi),%r13,%r14;			\
	mulx    t+8(%rdi),%rbx,%r15;			\
	addq    %rbx,%r14;				\
							\
	mulx    t+16(%rdi),%rbx,%rax;			\
	adcq    %rbx,%r15;				\
	adcq	$0,%rax;				\
							\
	movq	%rax,120(%rsp);			\
	movq    m+8(%rsi),%rdx;    			\
							\
	mulx    t+0(%rdi),%rbx,%rbp;			\
	mulx    t+8(%rdi),%rax,%rcx;			\
	addq    %rax,%rbp;				\
							\
	mulx    t+16(%rdi),%rax,%rdx;			\
	adcq    %rcx,%rax;				\
	adcq    $0,%rdx;				\
							\
	addq    %rbx,%r14;				\
	adcq    %rbp,%r15;				\
	adcq    120(%rsp),%rax;			\
	adcq    $0,%rdx;				\
	    	    					\
	movq    %rdx,%rcx;				\
							\
	addq    t+0(%rdi),%r15;			\
	adcq    t+8(%rdi),%rax;			\
	adcq    t+16(%rdi),%rcx;			\
	
	
#define fe1305_mul_taunr(t)				\
							\
	movq    %r8,%rdx;    				\
							\
	mulx    t+0(%rdi),%r8,%r14;			\
	mulx    t+8(%rdi),%rbx,%r15;			\
	addq    %rbx,%r14;				\
							\
	mulx    t+16(%rdi),%rbx,%rax;			\
	adcq    %rbx,%r15;				\
	adcq	$0,%rax;				\
							\
	movq    %r9,%rdx;    				\
							\
	mulx    t+0(%rdi),%r9,%rbp;			\
	mulx    t+8(%rdi),%rbx,%r13;			\
	addq    %rbx,%rbp;				\
							\
	mulx    t+16(%rdi),%rbx,%rcx;			\
	adcq    %r13,%rbx;				\
	adcq    $0,%rcx;				\
							\
	addq    %r14,%r9;				\
	adcq    %rbp,%r15;				\
	adcq    %rbx,%rax;				\
	adcq    $0,%rcx;				\
							\
	movq    %r10,%rdx;    				\
							\
	mulx    t+0(%rdi),%r10,%r11;			\
	mulx    t+8(%rdi),%r14,%r13;			\
	addq    %r14,%r11;				\
							\
	mulx    t+16(%rdi),%r12,%rdx;			\
	adcq    %r13,%r12;				\
							\
	addq    %r15,%r10;				\
	adcq    %rax,%r11;				\
	adcq    %rcx,%r12;				\
	

#define fe1305_mul_tau(m,t)				\
							\
	movq    $0,%rcx;				\
	movq    m+0(%rsi),%rdx;    			\
							\
	mulx    t+0(%rdi),%r13,%r14;			\
	mulx    t+8(%rdi),%rbx,%r15;			\
	addq    %rbx,%r14;				\
	adcq    %rcx,%r15;				\
							\
	movq    m+8(%rsi),%rdx;    			\
							\
	mulx    t+0(%rdi),%rbx,%rbp;			\
	mulx    t+8(%rdi),%rdx,%rax;			\
	addq    %rdx,%rbp;				\
	adcq    %rcx,%rax;				\
							\
	addq    %rbx,%r14;				\
	adcq    %rbp,%r15;				\
	adcq    %rcx,%rax;				\
							\
	addq    t+0(%rdi),%r15;			\
	adcq    t+8(%rdi),%rax;			\
	adcq    %rcx,%rcx;				\
	
	
#define fe1305_mul_taur(t)				\
							\
	movq    $0,%r12;				\
	movq    %r8,%rdx;    				\
							\
	mulx    t+0(%rdi),%r8,%r14;			\
	mulx    t+8(%rdi),%rbx,%r15;			\
	addq    %rbx,%r14;				\
	adcq    %r12,%r15;				\
							\
	movq    %r9,%rdx;    				\
							\
	mulx    t+0(%rdi),%r9,%rbp;			\
	mulx    t+8(%rdi),%rdx,%r11;			\
	addq    %rdx,%rbp;				\
	adcq    %r12,%r11;				\
							\
	addq    %r14,%r9;				\
	adcq    %rbp,%r15;				\
	adcq    %r12,%r11;				\
							\
	movq    %r10,%rdx;    				\
							\
	mulx    t+0(%rdi),%r10,%rcx;			\
	mulx    t+8(%rdi),%rdx,%rbx;			\
	addq    %rdx,%rcx;				\
	adcq    %r12,%rbx;				\
							\
	addq    %r15,%r10;				\
	adcq    %rcx,%r11;				\
	adcq    %rbx,%r12;				\
	

#define fe1305_add_product()				\
							\
	addq	%r13,%r8;				\
	adcq	%r14,%r9;				\
	adcq	%r15,%r10;				\
	adcq	%rax,%r11;				\
	adcq	%rcx,%r12;				\


#define fe1305_add_msg_block(m)			\
							\
	addq    m+0(%rsi),%r8;				\
	adcq    m+8(%rsi),%r9;				\
	adcq    $1,%r10;				\


#define fe1305_reduce_5l()				\
							\
	movq    %r10,%r13;				\
	andq    mask2(%rip),%r10;			\
	andq    mask2c(%rip),%r13;			\
							\
	addq    %r13,%r8;				\
	adcq    %r11,%r9;				\
	adcq    %r12,%r10;				\
							\
	shrd    $2,%r11,%r13;				\
	shrd    $2,%r12,%r11;				\
	shrq    $2,%r12;				\
							\
	addq    %r13,%r8;				\
	adcq    %r11,%r9;				\
	adcq    %r12,%r10;				\
	
	
#define fe1305_reduce_4l()				\
							\
	movq    %r10,%r13;				\
							\
	andq    mask2(%rip),%r10;			\
	andq    mask2c(%rip),%r13;			\
							\
	addq    %r13,%r8;				\
	adcq    %r11,%r9;				\
	adcq    %r12,%r10;				\
							\
	shrd    $2,%r11,%r13;				\
	shrq    $2,%r11;				\
							\
	addq    %r13,%r8;				\
	adcq    %r11,%r9;				\
	adcq    %r12,%r10;				\


#define fe1305_reduce_3l()				\
							\
	movq    %r10,%r11;				\
	andq    mask2(%rip),%r10;			\
	shrq    $2,%r11;				\
							\
	imul    $5,%r11,%r11;				\
	addq    %r11,%r8;				\
	adcq    $0,%r9;				\
	adcq    $0,%r10;				\
	
#endif