<% function addS1S2() { %>
        xor    rdx, rdx
        mov    edx, eax
        add    edx, ecx
        jo     add_manageOverflow   ; rsi already is the 64bits result

        mov    [rdi], rdx       ; not necessary to adjust so just save and return
        ret

add_manageOverflow:                 ; Do the operation in 64 bits
        push   rsi
        movsx  rsi, eax
        movsx  rdx, ecx
        add    rsi, rdx
        call   rawCopyS2L
        pop    rsi
        ret
<% } %>


<% function addL1S2() { %>
        add rsi, 8
        movsx rdx, ecx
        add rdi, 8
        cmp rdx, 0
        <% const rawAddLabel = global.tmpLabel() %>
        jns <%= rawAddLabel %>
        neg rdx
        call rawSubLS
        sub rdi, 8
        sub rsi, 8
        ret
<%= rawAddLabel %>:
        call rawAddLS
        sub rdi, 8
        sub rsi, 8
        ret

<% } %>

<% function addS1L2() { %>
        lea rsi, [rdx + 8]
        movsx rdx, eax
        add rdi, 8
        cmp rdx, 0
        <% const rawAddLabel = global.tmpLabel() %>
        jns <%= rawAddLabel %>
        neg rdx
        call rawSubLS
        sub rdi, 8
        sub rsi, 8
        ret
<%= rawAddLabel %>:
        call rawAddLS
        sub rdi, 8
        sub rsi, 8
        ret
<% } %>

<% function addL1L2() { %>
        add rdi, 8
        add rsi, 8
        add rdx, 8
        call rawAddLL
        sub rdi, 8
        sub rsi, 8
        ret
<% } %>

;;;;;;;;;;;;;;;;;;;;;;
; add
;;;;;;;;;;;;;;;;;;;;;;
; Adds two elements of any kind
; Params:
;   rsi <= Pointer to element 1
;   rdx <= Pointer to element 2
;   rdi <= Pointer to result
; Modified Registers:
;    r8, r9, 10, r11, rax, rcx
;;;;;;;;;;;;;;;;;;;;;;
<%=name%>_add:
        mov    rax, [rsi]
        mov    rcx, [rdx]
        bt     rax, 63          ; Check if is short first operand
        jc     add_l1
        bt     rcx, 63          ; Check if is short second operand
        jc     add_s1l2

add_s1s2:                       ; Both operands are short
<%=     addS1S2()       %>
add_l1:
        bt     rcx, 63          ; Check if is short second operand
        jc     add_l1l2

;;;;;;;;
add_l1s2:
        bt     rax, 62          ; check if montgomery first
        jc     add_l1ms2
add_l1ns2:
<%=     global.setTypeDest("0x80"); %>
<%=     addL1S2(); %>

add_l1ms2:
        bt     rcx, 62          ; check if montgomery second
        jc     add_l1ms2m
add_l1ms2n:
<%=     global.setTypeDest("0xC0"); %>
<%=     global.toMont_b()       %>
<%=     addL1L2()         %>

add_l1ms2m:
<%=     global.setTypeDest("0xC0"); %>
<%=     addL1L2()         %>


;;;;;;;;
add_s1l2:
        bt     rcx, 62          ; check if montgomery first
        jc     add_s1l2m
add_s1l2n:
<%=     global.setTypeDest("0x80"); %>
<%=     addS1L2(); %>

add_s1l2m:
        bt     rax, 62          ; check if montgomery second
        jc     add_s1ml2m
add_s1nl2m:
<%=     global.setTypeDest("0xC0"); %>
<%=     global.toMont_a()       %>
<%=     addL1L2()         %>

add_s1ml2m:
<%=     global.setTypeDest("0xC0"); %>
<%=     addL1L2()         %>

;;;;
add_l1l2:
        bt     rax, 62          ; check if montgomery first
        jc     add_l1ml2
add_l1nl2:
        bt     rcx, 62          ; check if montgomery second
        jc     add_l1nl2m
add_l1nl2n:
<%=     global.setTypeDest("0x80"); %>
<%=     addL1L2()         %>

add_l1nl2m:
<%=     global.setTypeDest("0xC0"); %>
<%=     global.toMont_a(); %>
<%=     addL1L2()         %>

add_l1ml2:
        bt     rcx, 62          ; check if montgomery seconf
        jc     add_l1ml2m
add_l1ml2n:
<%=     global.setTypeDest("0xC0"); %>
<%=     global.toMont_b(); %>
<%=     addL1L2()         %>

add_l1ml2m:
<%=     global.setTypeDest("0xC0"); %>
<%=     addL1L2()         %>


;;;;;;;;;;;;;;;;;;;;;;
; rawAddLL
;;;;;;;;;;;;;;;;;;;;;;
; Adds two elements of type long
; Params:
;   rsi <= Pointer to the long data of element 1
;   rdx <= Pointer to the long data of element 2
;   rdi <= Pointer to the long data of result
; Modified Registers:
;    rax
;;;;;;;;;;;;;;;;;;;;;;
rawAddLL:
        ; Add component by component with carry
<% for (let i=0; i<n64; i++) { %>
        mov rax, [rsi + <%=i*8%>]
        <%= i==0 ? "add" : "adc" %> rax, [rdx + <%=i*8%>]
        mov [rdi + <%=i*8%>], rax
<% } %>
        jc rawAddLL_sq   ; if overflow, substract q

        ; Compare with q
<% for (let i=0; i<n64; i++) { %>
<%    if (i>0) { %>
        mov rax, [rdi + <%= (n64-i-1)*8 %>]
<%    } %>
        cmp rax, [q + <%= (n64-i-1)*8 %>]
        jc rawAddLL_done        ; q is bigget so done.
        jnz rawAddLL_sq         ; q is lower
<% } %>
        ; If equal substract q
rawAddLL_sq:
<% for (let i=0; i<n64; i++) { %>
        mov rax, [q + <%=i*8%>]
        <%= i==0 ? "sub" : "sbb" %> [rdi + <%=i*8%>], rax
<% } %>
rawAddLL_done:
        ret


;;;;;;;;;;;;;;;;;;;;;;
; rawAddLS
;;;;;;;;;;;;;;;;;;;;;;
; Adds two elements of type long
; Params:
;   rdi <= Pointer to the long data of result
;   rsi <= Pointer to the long data of element 1
;   rdx <= Value to be added
;;;;;;;;;;;;;;;;;;;;;;
rawAddLS:
        ; Add component by component with carry

        add rdx, [rsi]
        mov [rdi] ,rdx
<% for (let i=1; i<n64; i++) { %>
        mov rdx, 0
        adc rdx, [rsi + <%=i*8%>]
        mov [rdi + <%=i*8%>], rdx
<% } %>
        jc rawAddLS_sq   ; if overflow, substract q

        ; Compare with q
<% for (let i=0; i<n64; i++) { %>
        mov rax, [rdi + <%= (n64-i-1)*8 %>]
        cmp rax, [q + <%= (n64-i-1)*8 %>]
        jc rawAddLS_done        ; q is bigget so done.
        jnz rawAddLS_sq         ; q is lower
<% } %>
        ; If equal substract q
rawAddLS_sq:
<% for (let i=0; i<n64; i++) { %>
        mov rax, [q + <%=i*8%>]
        <%= i==0 ? "sub" : "sbb" %> [rdi + <%=i*8%>], rax
<% } %>
rawAddLS_done:
        ret