-rw-r--r-- 4350 lib1305-20250407/crypto_onetimeauth/poly1305/amd64-maax-g4/poly1305_asm.h raw
#ifndef __POLY1305ASM__
#define __POLY1305ASM__

/* field arithmetic used for computing poly1305 */

#define fe1305_mul_taun(m,t)				\
							\
	xorq    %rcx,%rcx;				\
	movq    m+0(%rsi),%rdx;    			\
							\
	mulx    t+0(%rdi),%r13,%r14;			\
	mulx    t+8(%rdi),%rbx,%r15;			\
	adcx    %rbx,%r14;				\
							\
	mulx    t+16(%rdi),%rbx,%rax;			\
	adcx    %rbx,%r15;				\
	adcx	%rcx,%rax;				\
							\
	xorq    %rdx,%rdx;				\
	movq    m+8(%rsi),%rdx;    			\
							\
	mulx    t+0(%rdi),%rbx,%rbp;			\
	adcx    %rbx,%r14;				\
	adox    %rbp,%r15;				\
							\
	mulx    t+8(%rdi),%rbx,%rbp;			\
	adcx    %rbx,%r15;				\
	adox    %rbp,%rax;				\
							\
	mulx    t+16(%rdi),%rbx,%rbp;			\
	adcx    %rbx,%rax;				\
	adox    %rbp,%rcx;				\
	adcx    zero(%rip),%rcx;			\
	    	    					\
	xorq    %rdx,%rdx;				\
							\
	adcx    t+0(%rdi),%r15;			\
	adox	%rdx,%rax;				\
	adcx    t+8(%rdi),%rax;			\
	adox	%rdx,%rcx;				\
	adcx    t+16(%rdi),%rcx;			\


#define fe1305_mul_taunr(t)				\
							\
	xorq    %r15,%r15;				\
	movq    %r8,%rdx;    				\
							\
	mulx    t+0(%rdi),%r8,%r12;			\
	mulx    t+8(%rdi),%rbx,%r13;			\
	adcx    %rbx,%r12;				\
	adcx    %r15,%r13;				\
							\
	mulx    t+16(%rdi),%rbx,%r14;			\
	adcx    %rbx,%r13;				\
	adcx    %r15,%r14;				\
							\
	xorq    %rax,%rax;				\
	movq    %r9,%rdx;				\
	   						\
	mulx    t+0(%rdi),%r9,%rbp;			\
	adcx    %r12,%r9;				\
	adox    %rbp,%r13;				\
							\
	mulx    t+8(%rdi),%rbx,%rbp;			\
	adcx    %rbx,%r13;				\
	adox    %rbp,%r14;				\
	    						\
	mulx    t+16(%rdi),%rbx,%rbp;			\
	adcx    %rbx,%r14;				\
	adox    %rbp,%r15;				\
	adcx    %rax,%r15;				\
							\
	xorq    %rax,%rax;				\
	movq    %r10,%rdx;				\
	    						\
	mulx    t+0(%rdi),%r10,%rbp;			\
	adcx    %r13,%r10;				\
	adox    %rbp,%r14;				\
	    						\
	mulx    t+8(%rdi),%r11,%rbp;			\
	adcx    %r14,%r11;				\
	adox    %rbp,%r15;				\
							\
	mulx    t+16(%rdi),%r12,%rbp;			\
	adcx    %r15,%r12;				\
	

#define fe1305_mul_tau(m,t)				\
							\
	xorq    %rax,%rax;				\
	movq    m+0(%rsi),%rdx;    			\
							\
	mulx    t+0(%rdi),%r13,%r14;			\
	mulx    t+8(%rdi),%rbx,%r15;			\
	adcx    %rbx,%r14;				\
	adcx    %rax,%r15;				\
							\
	xorq    %rcx,%rcx;				\
	movq    m+8(%rsi),%rdx;			\
	   						\
	mulx    t+0(%rdi),%rbx,%rbp;			\
	adcx    %rbx,%r14;				\
	adox    %rbp,%r15;				\
	    						\
	mulx    t+8(%rdi),%rbx,%rbp;			\
	adcx    %rbx,%r15;				\
	adox    %rbp,%rax;				\
	adcx    %rcx,%rax;				\
							\
	xorq    %rdx,%rdx;				\
							\
	adcx    t+0(%rdi),%r15;			\
	adox	%rdx,%rax;				\
	adcx    t+8(%rdi),%rax;			\
	adox	%rdx,%rcx;				\
	adcx	%rdx,%rcx;				\
	

#define fe1305_mul_taur(t)				\
							\
	xorq    %r14,%r14;				\
	movq    %r8,%rdx; 				\
							\
	mulx    t+0(%rdi),%r8,%r12;			\
	mulx    t+8(%rdi),%rbx,%r13;			\
	adcx    %rbx,%r12;				\
	adcx    %r14,%r13;				\
							\
	xorq    %rax,%rax;				\
	movq    %r9,%rdx;				\
	   						\
	mulx    t+0(%rdi),%r9,%rbp;			\
	adcx    %r12,%r9;				\
	adox    %rbp,%r13;				\
	    						\
	mulx    t+8(%rdi),%rbx,%rbp;			\
	adcx    %rbx,%r13;				\
	adox    %rbp,%r14;				\
	adcx    %rax,%r14;				\
							\
	xorq    %r12,%r12;				\
	movq    %r10,%rdx;				\
	    						\
	mulx    t+0(%rdi),%r10,%rbp;			\
	adcx    %r13,%r10;				\
	adox    %rbp,%r14;				\
	    						\
	mulx    t+8(%rdi),%r11,%rbp;			\
	adcx    %r14,%r11;				\
	adox    %rbp,%r12;				\
	adcx    %rax,%r12;				\


#define fe1305_add_product()				\
							\
	xorq    %rdx,%rdx;				\
							\
	adcx	%r13,%r8;				\
	adcx	%r14,%r9;				\
	adcx	%r15,%r10;				\
	adcx	%rax,%r11;				\
	adcx	%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_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