; void leaf_call(int b, int c, int d, int e, int f, int g, int h)
; {
; }
;
; void nonleaf_call(int a, int b, int c, int d, int e, int f, int g, int h)
; {
;     /* use some local data */
;     char x[10] = { 'L' };
;     leaf_call(b, c, d, e, f, g, h);
; }
;
; int main()
; {
;     nonleaf_call(0, 1, 2, 3, 4, 5, 6, 7);
;     return 0;
; }



; output from godbolt compiler explorer w/ msvc 19.0

leaf_call PROC
        mov     DWORD PTR [rsp+32], r9d
        mov     DWORD PTR [rsp+24], r8d
        mov     DWORD PTR [rsp+16], edx
        mov     DWORD PTR [rsp+8], ecx
        ret     0
leaf_call ENDP

x$ = 64
b$ = 104
c$ = 112
d$ = 120
e$ = 128
f$ = 136
g$ = 144
h$ = 152
nonleaf_call PROC
        mov     DWORD PTR [rsp+32], r9d   ; | |
        mov     DWORD PTR [rsp+24], r8d   ; | |
        mov     DWORD PTR [rsp+16], edx   ; | | spill
        mov     DWORD PTR [rsp+8], ecx    ; | |
        push    rdi                       ; |
        sub     rsp, 80                   ; | prolog
        mov     BYTE PTR x$[rsp], 76      ; 'L' -> local area (of local array's space)
        lea     rax, QWORD PTR x$[rsp+1]  ; |
        mov     rdi, rax                  ; |
        xor     eax, eax                  ; | zero-init rest of local array's space
        mov     ecx, 9                    ; |
        rep stosb                         ; |
        mov     eax, DWORD PTR h$[rsp]    ; arg 6 (fetched from caller's frame param area), and ...
        mov     DWORD PTR [rsp+48], eax   ; ... "pushed" onto stack
        mov     eax, DWORD PTR g$[rsp]    ; arg 5 (fetched from caller's frame param area), and ...
        mov     DWORD PTR [rsp+40], eax   ; ... "pushed" onto stack
        mov     eax, DWORD PTR f$[rsp]    ; arg 4 (fetched from caller's frame param area), and ...
        mov     DWORD PTR [rsp+32], eax   ; ... "pushed" onto stack
        mov     r9d, DWORD PTR e$[rsp]    ; arg 3
        mov     r8d, DWORD PTR d$[rsp]    ; arg 2
        mov     edx, DWORD PTR c$[rsp]    ; arg 1
        mov     ecx, DWORD PTR b$[rsp]    ; arg 0
        call    leaf_call                 ; push return addr and call
        add     rsp, 80                   ; |
        pop     rdi                       ; | epilog
        ret     0                         ; |
nonleaf_call ENDP

main    PROC
        sub     rsp, 72                   ; prolog
        mov     DWORD PTR [rsp+56], 7     ; "push" arg 7 onto stack
        mov     DWORD PTR [rsp+48], 6     ; "push" arg 6 onto stack
        mov     DWORD PTR [rsp+40], 5     ; "push" arg 5 onto stack
        mov     DWORD PTR [rsp+32], 4     ; "push" arg 4 onto stack
        mov     r9d, 3                    ; arg 3
        mov     r8d, 2                    ; arg 2
        mov     edx, 1                    ; arg 1
        xor     ecx, ecx                  ; arg 0
        call    nonleaf_call              ; push return addr and call
        xor     eax, eax                  ; return value
        add     rsp, 72                   ; |
        ret     0                         ; | epilog
main    ENDP



; ---------- structs by value, struct in first call on reg arg boundary ---------->
;
; struct A { int i, j; long long l; };
;
; void leaf_call(int b, int c, int d, int e, struct A f, int g, int h)
; {
; }
;
; void nonleaf_call(int a, int b, int c, int d, int e, struct A f, int g, int h)
; {
;     /* use some local data */
;     char x[220] = { 'L' };
;     leaf_call(b, c, d, e, f, g, h);
; }
;
; int main()
; {
;     struct A a = {5, 6, 7ll};
;     nonleaf_call(0, 1, 2, 3, 4, a, 8, 9);
;     return 0;
; }



; output from godbolt compiler explorer w/ msvc 19.0

leaf_call PROC
        mov     DWORD PTR [rsp+32], r9d
        mov     DWORD PTR [rsp+24], r8d
        mov     DWORD PTR [rsp+16], edx
        mov     DWORD PTR [rsp+8], ecx
        ret     0
leaf_call ENDP

$T1 = 64
x$ = 80
b$ = 344
c$ = 352
d$ = 360
e$ = 368
f$ = 376
g$ = 384
h$ = 392
nonleaf_call PROC
        mov     DWORD PTR [rsp+32], r9d
        mov     DWORD PTR [rsp+24], r8d
        mov     DWORD PTR [rsp+16], edx
        mov     DWORD PTR [rsp+8], ecx
        push    rsi
        push    rdi
        sub     rsp, 312                      ; 00000138H
        mov     BYTE PTR x$[rsp], 76                  ; 0000004cH
        lea     rax, QWORD PTR x$[rsp+1]
        mov     rdi, rax
        xor     eax, eax
        mov     ecx, 219                      ; 000000dbH
        rep stosb
        lea     rax, QWORD PTR $T1[rsp]
        mov     rdi, rax
        mov     rsi, QWORD PTR f$[rsp]
        mov     ecx, 16
        rep movsb
        mov     eax, DWORD PTR h$[rsp]
        mov     DWORD PTR [rsp+48], eax
        mov     eax, DWORD PTR g$[rsp]
        mov     DWORD PTR [rsp+40], eax
        lea     rax, QWORD PTR $T1[rsp]
        mov     QWORD PTR [rsp+32], rax
        mov     r9d, DWORD PTR e$[rsp]
        mov     r8d, DWORD PTR d$[rsp]
        mov     edx, DWORD PTR c$[rsp]
        mov     ecx, DWORD PTR b$[rsp]
        call    leaf_call
        add     rsp, 312                      ; 00000138H
        pop     rdi
        pop     rsi
        ret     0
nonleaf_call ENDP

a$ = 64
$T1 = 80
main    PROC
        push    rsi
        push    rdi
        sub     rsp, 104                      ; 00000068H
        mov     DWORD PTR a$[rsp], 5
        mov     DWORD PTR a$[rsp+4], 6
        mov     QWORD PTR a$[rsp+8], 7
        lea     rax, QWORD PTR $T1[rsp]
        lea     rcx, QWORD PTR a$[rsp]
        mov     rdi, rax
        mov     rsi, rcx
        mov     ecx, 16
        rep movsb
        mov     DWORD PTR [rsp+56], 9
        mov     DWORD PTR [rsp+48], 8
        lea     rax, QWORD PTR $T1[rsp]
        mov     QWORD PTR [rsp+40], rax
        mov     DWORD PTR [rsp+32], 4
        mov     r9d, 3
        mov     r8d, 2
        mov     edx, 1
        xor     ecx, ecx
        call    nonleaf_call
        xor     eax, eax
        add     rsp, 104                      ; 00000068H
        pop     rdi
        pop     rsi
        ret     0
main    ENDP



; ---------- structs by value, complex example (multiple structs, partly passed via regs) ---------->
;
; struct A { int i, j; float f; };
; struct B { double d; long long l; };
;
; void leaf_call(int b, struct A c, struct B d, int e, int f, struct A g, struct B h, int i, int j)
; {
; }
;
; void nonleaf_call(int a, int b, struct A c, struct B d, int e, int f, struct A g, struct B h, int i, int j)
; {
;     /* use some local data */
;     char x[220] = { 'L' };
;     leaf_call(b, c, d, e, f, g, h, i, j);
; }
;
; int main()
; {
;     struct A a = {2, 3, 4.f};
;     struct B b = {5., 6ll};
;     struct A c = {9, 10, 11.f};
;     struct B d = {12., 13ll};
;     nonleaf_call(0, 1, a, b, 7, 8, c, d, 14, 15);
;     return 0;
; }



; output from godbolt compiler explorer w/ msvc 19.0

__real@41300000 DD 041300000r                   ; |
__real@40800000 DD 040800000r                   ; |
__real@4028000000000000 DQ 04028000000000000r   ; | fp data
__real@4014000000000000 DQ 04014000000000000r   ; |

leaf_call PROC
        mov     DWORD PTR [rsp+32], r9d
        mov     QWORD PTR [rsp+24], r8
        mov     QWORD PTR [rsp+16], rdx
        mov     DWORD PTR [rsp+8], ecx
        ret     0
leaf_call ENDP

$T1 = 80
$T2 = 96
$T3 = 112
$T4 = 128
x$ = 144
b$ = 408
c$ = 416
d$ = 424
e$ = 432
f$ = 440
g$ = 448
h$ = 456
i$ = 464
j$ = 472
nonleaf_call PROC
        mov     QWORD PTR [rsp+32], r9
        mov     QWORD PTR [rsp+24], r8
        mov     DWORD PTR [rsp+16], edx
        mov     DWORD PTR [rsp+8], ecx
        push    rsi
        push    rdi
        sub     rsp, 376
        mov     BYTE PTR x$[rsp], 76
        lea     rax, QWORD PTR x$[rsp+1]
        mov     rdi, rax
        xor     eax, eax
        mov     ecx, 219
        rep stosb
        lea     rax, QWORD PTR $T3[rsp]
        mov     rdi, rax
        mov     rsi, QWORD PTR h$[rsp]
        mov     ecx, 16
        rep movsb
        lea     rax, QWORD PTR $T1[rsp]
        mov     rdi, rax
        mov     rsi, QWORD PTR g$[rsp]
        mov     ecx, 12
        rep movsb
        lea     rax, QWORD PTR $T4[rsp]
        mov     rdi, rax
        mov     rsi, QWORD PTR d$[rsp]
        mov     ecx, 16
        rep movsb
        lea     rax, QWORD PTR $T2[rsp]
        mov     rdi, rax
        mov     rsi, QWORD PTR c$[rsp]
        mov     ecx, 12
        rep movsb
        mov     eax, DWORD PTR j$[rsp]
        mov     DWORD PTR [rsp+64], eax
        mov     eax, DWORD PTR i$[rsp]
        mov     DWORD PTR [rsp+56], eax
        lea     rax, QWORD PTR $T3[rsp]
        mov     QWORD PTR [rsp+48], rax
        lea     rax, QWORD PTR $T1[rsp]
        mov     QWORD PTR [rsp+40], rax
        mov     eax, DWORD PTR f$[rsp]
        mov     DWORD PTR [rsp+32], eax
        mov     r9d, DWORD PTR e$[rsp]
        lea     r8, QWORD PTR $T4[rsp]
        lea     rdx, QWORD PTR $T2[rsp]
        mov     ecx, DWORD PTR b$[rsp]
        call    leaf_call
        add     rsp, 376
        pop     rdi
        pop     rsi
        ret     0
nonleaf_call ENDP

c$ = 80
a$ = 96
$T1 = 112
$T2 = 128
d$ = 144
b$ = 160
$T3 = 176
$T4 = 192
main    PROC
        push    rsi
        push    rdi
        sub     rsp, 216
        mov     DWORD PTR a$[rsp], 2
        mov     DWORD PTR a$[rsp+4], 3
        movss   xmm0, DWORD PTR __real@40800000
        movss   DWORD PTR a$[rsp+8], xmm0
        movsd   xmm0, QWORD PTR __real@4014000000000000
        movsd   QWORD PTR b$[rsp], xmm0
        mov     QWORD PTR b$[rsp+8], 6
        mov     DWORD PTR c$[rsp], 9
        mov     DWORD PTR c$[rsp+4], 10
        movss   xmm0, DWORD PTR __real@41300000
        movss   DWORD PTR c$[rsp+8], xmm0
        movsd   xmm0, QWORD PTR __real@4028000000000000
        movsd   QWORD PTR d$[rsp], xmm0
        mov     QWORD PTR d$[rsp+8], 13
        lea     rax, QWORD PTR $T3[rsp]
        lea     rcx, QWORD PTR d$[rsp]
        mov     rdi, rax
        mov     rsi, rcx
        mov     ecx, 16
        rep movsb
        lea     rax, QWORD PTR $T1[rsp]
        lea     rcx, QWORD PTR c$[rsp]
        mov     rdi, rax
        mov     rsi, rcx
        mov     ecx, 12
        rep movsb
        lea     rax, QWORD PTR $T4[rsp]
        lea     rcx, QWORD PTR b$[rsp]
        mov     rdi, rax
        mov     rsi, rcx
        mov     ecx, 16
        rep movsb
        lea     rax, QWORD PTR $T2[rsp]
        lea     rcx, QWORD PTR a$[rsp]
        mov     rdi, rax
        mov     rsi, rcx
        mov     ecx, 12
        rep movsb
        mov     DWORD PTR [rsp+72], 15
        mov     DWORD PTR [rsp+64], 14
        lea     rax, QWORD PTR $T3[rsp]
        mov     QWORD PTR [rsp+56], rax
        lea     rax, QWORD PTR $T1[rsp]
        mov     QWORD PTR [rsp+48], rax
        mov     DWORD PTR [rsp+40], 8
        mov     DWORD PTR [rsp+32], 7
        lea     r9, QWORD PTR $T4[rsp]
        lea     r8, QWORD PTR $T2[rsp]
        mov     edx, 1
        xor     ecx, ecx
        call    nonleaf_call
        xor     eax, eax
        add     rsp, 216
        pop     rdi
        pop     rsi
        ret     0
main    ENDP



; ---------- returning tiny struct by value (passes via regs) ---------->
;
; struct A { unsigned char a; };
;
; struct A call(unsigned char c)
; {
;     struct A r = {c};
;     return r;
; }
;
; int main()
; {
;     struct A a = call(123);
;     return 0;
; }



; output from godbolt compiler explorer w/ msvc 19.0

r$ = 0
c$ = 32
call    PROC
        mov     BYTE PTR [rsp+8], cl   ; |         spill
        sub     rsp, 24                ; | prolog
        movzx   eax, BYTE PTR c$[rsp]  ; in arg 0 (fetched from spill area) -> eax, then ...
        mov     BYTE PTR r$[rsp], al   ; ... -> struct in local area (top of stack, as leaf call and thus no reserved spill area)
        movzx   eax, BYTE PTR r$[rsp]  ; reget same value into eax to return (small) struct via reg (a bit pointless to refetch)
        add     rsp, 24                ; |
        ret     0                      ; | epilog
call    ENDP

a$ = 32
main    PROC
        sub     rsp, 56                ; prolog
        mov     cl, 123                ; arg 0
        call    call                   ; push return addr and call
        mov     BYTE PTR a$[rsp], al   ; write struct data to local area (123)
        xor     eax, eax               ; return value
        add     rsp, 56                ; |
        ret     0                      ; | epilog
main    ENDP



; ---------- C++ trivial and non-trivial aggrs passed to C funcs ---------->
;
; struct Trivial { int a; };
; struct NonTrivial { int a; NonTrivial() : a(0) {} NonTrivial(const NonTrivial& rhs) : a(rhs.a) { } };
;
; extern "C" {
; 
;     void f1(struct Trivial s)    { }
;     void f2(struct NonTrivial s) { }
;
;     void f()
;     {
;         struct Trivial t;
;         struct NonTrivial n;
;         int a=1;
;         a += 123;
;         f1(t);
;         a -= 123;
;         f2(n);
;         a -= 12;
;     }
; }



; output from godbolt compiler explorer w/ msvc 19.0

        ; ... snip, removed code of ctor and copy ctor ...

f1      PROC
        mov     DWORD PTR [rsp+8], ecx
        ret     0
f1      ENDP

f2      PROC
        mov     QWORD PTR [rsp+8], rcx
        ret     0
f2      ENDP

a$ = 32
n$ = 36
t$ = 40
$T1 = 44
$T2 = 48
f       PROC
$LN3:
        sub     rsp, 72                                    ; prolog
        lea     rcx, QWORD PTR n$[rsp]                     ; \ this ptr (NULL)
        call    NonTrivial::NonTrivial(void)               ; | NonTrivial::NonTrivial() / ctor
        mov     DWORD PTR a$[rsp], 1                       ; a = 1
        mov     eax, DWORD PTR a$[rsp]                     ; |
        add     eax, 123                                   ; | a += 123
        mov     DWORD PTR a$[rsp], eax                     ; /
        mov     ecx, DWORD PTR t$[rsp]                     ; f1 arg 0 (struct Trivial), via reg as small struct
        call    f1                                         ; call f1(struct Trivial)
        mov     eax, DWORD PTR a$[rsp]                     ; |
        sub     eax, 123                                   ; | a -= 123
        mov     DWORD PTR a$[rsp], eax                     ; /
        lea     rax, QWORD PTR $T1[rsp]                    ; @@@ unsure
        mov     QWORD PTR $T2[rsp], rax                    ; ... @@@
        lea     rdx, QWORD PTR n$[rsp]                     ; \               ptr to dest of copy of n
        mov     rcx, QWORD PTR $T2[rsp]                    ; | copy n        ptr to n
        call    NonTrivial::NonTrivial(NonTrivial const &) ; /               NonTrivial::NonTrivial(const NonTrivial&) / copy ctor
        mov     rcx, rax                                   ; f2 arg 0 (ptr to copy of struct NonTrivial), via ptr as non-trivial
        call    f2                                         ; call f2(struct NonTrivial)
        mov     eax, DWORD PTR a$[rsp]                     ; |
        sub     eax, 12                                    ; | a -= 12
        mov     DWORD PTR a$[rsp], eax                     ; /
        add     rsp, 72                                    ; \
        ret     0                                          ; | epilog
f       ENDP



; ---------- C++ trivial and non-trivial aggrs as return values ---------->
;
; struct Trivial { int a; };
; struct NonTrivial { int a; NonTrivial() : a(0) {} NonTrivial(const NonTrivial& rhs) : a(rhs.a) { } };
; 
; extern "C" {
;     struct Trivial    f1() { return Trivial(); }
; }
; 
;     struct NonTrivial f2() { return NonTrivial(); }
; 
; extern "C" {
;     void f()
;     {
;         int a=1;
;         a += 123;
;         struct Trivial t = f1();
;         a -= 123;
;         struct NonTrivial n = f2();
;         a -= 12;
;     }
; }



; output from godbolt compiler explorer w/ msvc 19.0

        ; ... snip, removed code of ctor and copy ctor ...

$T1 = 0
f1      PROC
$LN3:
        push    rdi
        sub     rsp, 16
        lea     rax, QWORD PTR $T1[rsp]
        mov     rdi, rax
        xor     eax, eax
        mov     ecx, 4
        rep stosb
        mov     eax, DWORD PTR $T1[rsp]
        add     rsp, 16
        pop     rdi
        ret     0
f1      ENDP

__$ReturnUdt$ = 48
NonTrivial f2(void) PROC
$LN3:
        mov     QWORD PTR [rsp+8], rcx
        sub     rsp, 40
        mov     rcx, QWORD PTR __$ReturnUdt$[rsp]
        call    NonTrivial::NonTrivial(void)
        mov     rax, QWORD PTR __$ReturnUdt$[rsp]
        add     rsp, 40
        ret     0
NonTrivial f2(void) ENDP

a$ = 32
$T1 = 36
t$ = 40
n$ = 44
f       PROC
$LN3:
        sub     rsp, 56                 ; prolog
        mov     DWORD PTR a$[rsp], 1    ; a = 1
        mov     eax, DWORD PTR a$[rsp]  ; |
        add     eax, 123                ; | a += 123
        mov     DWORD PTR a$[rsp], eax  ; |
        call    f1                      ; call f1()
        mov     DWORD PTR $T1[rsp], eax ;
        mov     eax, DWORD PTR $T1[rsp] ;
        mov     DWORD PTR t$[rsp], eax  ;
        mov     eax, DWORD PTR a$[rsp]  ; |
        sub     eax, 123                ; | a -= 123
        mov     DWORD PTR a$[rsp], eax  ; |
        lea     rcx, QWORD PTR n$[rsp]  ; ptr to space to hold non-triv retval
        call    NonTrivial f2(void)     ; call f2()
        mov     eax, DWORD PTR a$[rsp]  ; |
        sub     eax, 12                 ; | a -= 12
        mov     DWORD PTR a$[rsp], eax  ; /
        add     rsp, 56                 ; \ epilog
        ret     0                       ; |
f       ENDP                             



; vim: ft=asm