Windows x64 と Linux x64 における引数のレジスタ・スタック配置の比較

目的

WindowsとLinuxで、引数渡しにどのような違いがあるのかを調べます。

Cソース

gcc -O0 -S -masm=intel test.c -o test.s

test.c

void func(
    char a1, char a2, char a3, char a4,
    char a5, char a6, char a7, char a8,
    char a9, char a10, char a11, char a12
) {
    // 何もしない。
}

int main() {
    // 適当な値を渡して呼び出す
    func(1,2,3,4,5,6,7,8,9,10,11,12);
    return 0;
}

生成されたアセンブリ

Windows10

test.s

test.s

	.file	"test.c"
	.intel_syntax noprefix
	.text
	.globl	func
	.def	func;	.scl	2;	.type	32;	.endef
	.seh_proc	func
func:
	push	rbp
	.seh_pushreg	rbp
	push	r13
	.seh_pushreg	r13
	push	r12
	.seh_pushreg	r12
	push	rdi
	.seh_pushreg	rdi
	push	rsi
	.seh_pushreg	rsi
	push	rbx
	.seh_pushreg	rbx
	sub	rsp, 40
	.seh_stackalloc	40
	lea	rbp, 32[rsp]
	.seh_setframe	rbp, 32
	.seh_endprologue
	mov	eax, edx
	mov	r13d, r8d
	mov	r12d, r9d
	mov	edi, DWORD PTR 96[rbp]
	mov	esi, DWORD PTR 104[rbp]
	mov	ebx, DWORD PTR 112[rbp]
	mov	r11d, DWORD PTR 120[rbp]
	mov	r10d, DWORD PTR 128[rbp]
	mov	r9d, DWORD PTR 136[rbp]
	mov	r8d, DWORD PTR 144[rbp]
	mov	edx, DWORD PTR 152[rbp]
	mov	BYTE PTR 64[rbp], cl
	mov	BYTE PTR 72[rbp], al
	mov	eax, r13d
	mov	BYTE PTR 80[rbp], al
	mov	eax, r12d
	mov	BYTE PTR 88[rbp], al
	mov	eax, edi
	mov	BYTE PTR -4[rbp], al
	mov	eax, esi
	mov	BYTE PTR -8[rbp], al
	mov	eax, ebx
	mov	BYTE PTR -12[rbp], al
	mov	eax, r11d
	mov	BYTE PTR -16[rbp], al
	mov	eax, r10d
	mov	BYTE PTR -20[rbp], al
	mov	eax, r9d
	mov	BYTE PTR -24[rbp], al
	mov	eax, r8d
	mov	BYTE PTR -28[rbp], al
	mov	eax, edx
	mov	BYTE PTR -32[rbp], al
	nop
	add	rsp, 40
	pop	rbx
	pop	rsi
	pop	rdi
	pop	r12
	pop	r13
	pop	rbp
	ret
	.seh_endproc
	.globl	main
	.def	main;	.scl	2;	.type	32;	.endef
	.seh_proc	main
main:
	push	rbp
	.seh_pushreg	rbp
	mov	rbp, rsp
	.seh_setframe	rbp, 0
	sub	rsp, 96
	.seh_stackalloc	96
	.seh_endprologue
	call	__main
	mov	DWORD PTR 88[rsp], 12
	mov	DWORD PTR 80[rsp], 11
	mov	DWORD PTR 72[rsp], 10
	mov	DWORD PTR 64[rsp], 9
	mov	DWORD PTR 56[rsp], 8
	mov	DWORD PTR 48[rsp], 7
	mov	DWORD PTR 40[rsp], 6
	mov	DWORD PTR 32[rsp], 5
	mov	r9d, 4
	mov	r8d, 3
	mov	edx, 2
	mov	ecx, 1
	call	func
	mov	eax, 0
	add	rsp, 96
	pop	rbp
	ret
	.seh_endproc
	.def	__main;	.scl	2;	.type	32;	.endef
	.ident	"GCC: (MinGW-W64 x86_64-ucrt-posix-seh, built by Brecht Sanders, r4) 15.1.0"

Ubuntu

test.s

test.s

	.file	"test.c"
	.intel_syntax noprefix
	.text
	.globl	func
	.type	func, @function
func:
.LFB0:
	.cfi_startproc
	endbr64
	push	rbp
	.cfi_def_cfa_offset 16
	.cfi_offset 6, -16
	mov	rbp, rsp
	.cfi_def_cfa_register 6
	push	r13
	push	r12
	push	rbx
	.cfi_offset 13, -24
	.cfi_offset 12, -32
	.cfi_offset 3, -40
	mov	eax, ecx
	mov	r13d, r8d
	mov	r12d, r9d
	mov	ebx, DWORD PTR 16[rbp]
	mov	r11d, DWORD PTR 24[rbp]
	mov	r10d, DWORD PTR 32[rbp]
	mov	r9d, DWORD PTR 40[rbp]
	mov	r8d, DWORD PTR 48[rbp]
	mov	ecx, DWORD PTR 56[rbp]
	mov	BYTE PTR -28[rbp], dil
	mov	BYTE PTR -32[rbp], sil
	mov	BYTE PTR -36[rbp], dl
	mov	BYTE PTR -40[rbp], al
	mov	eax, r13d
	mov	BYTE PTR -44[rbp], al
	mov	eax, r12d
	mov	BYTE PTR -48[rbp], al
	mov	eax, ebx
	mov	BYTE PTR -52[rbp], al
	mov	eax, r11d
	mov	BYTE PTR -56[rbp], al
	mov	eax, r10d
	mov	BYTE PTR -60[rbp], al
	mov	eax, r9d
	mov	BYTE PTR -64[rbp], al
	mov	eax, r8d
	mov	BYTE PTR -68[rbp], al
	mov	eax, ecx
	mov	BYTE PTR -72[rbp], al
	nop
	pop	rbx
	pop	r12
	pop	r13
	pop	rbp
	.cfi_def_cfa 7, 8
	ret
	.cfi_endproc
.LFE0:
	.size	func, .-func
	.globl	main
	.type	main, @function
main:
.LFB1:
	.cfi_startproc
	endbr64
	push	rbp
	.cfi_def_cfa_offset 16
	.cfi_offset 6, -16
	mov	rbp, rsp
	.cfi_def_cfa_register 6
	push	12
	push	11
	push	10
	push	9
	push	8
	push	7
	mov	r9d, 6
	mov	r8d, 5
	mov	ecx, 4
	mov	edx, 3
	mov	esi, 2
	mov	edi, 1
	call	func
	add	rsp, 48
	mov	eax, 0
	leave
	.cfi_def_cfa 7, 8
	ret
	.cfi_endproc
.LFE1:
	.size	main, .-main
	.ident	"GCC: (Ubuntu 11.4.0-1ubuntu1~22.04.2) 11.4.0"
	.section	.note.GNU-stack,"",@progbits
	.section	.note.gnu.property,"a"
	.align 8
	.long	1f - 0f
	.long	4f - 1f
	.long	5
0:
	.string	"GNU"
1:
	.align 8
	.long	0xc0000002
	.long	3f - 2f
2:
	.long	0x3
3:
	.align 8
4:

WindowsとLinuxの違いを比較

項目	Windows x64	Linux x64
レジスタに入る引数	4個	6個
使用レジスタ	RCX, RDX, R8, R9	RDI, RSI, RDX, RCX, R8, R9
残りの引数	スタック	スタック

Windows

	mov	DWORD PTR 88[rsp], 12
	mov	DWORD PTR 80[rsp], 11
	mov	DWORD PTR 72[rsp], 10
	mov	DWORD PTR 64[rsp], 9
	mov	DWORD PTR 56[rsp], 8
	mov	DWORD PTR 48[rsp], 7
	mov	DWORD PTR 40[rsp], 6
	mov	DWORD PTR 32[rsp], 5
	mov	r9d, 4
	mov	r8d, 3
	mov	edx, 2
	mov	ecx, 1

Linux

	push	12
	push	11
	push	10
	push	9
	push	8
	push	7
	mov	r9d, 6
	mov	r8d, 5
	mov	ecx, 4
	mov	edx, 3
	mov	esi, 2
	mov	edi, 1

動作環境

Windows

OS: Windows10
Version: 10.0.19045.6332
CPU: Intel Core i7-10750H, 6 cores, 12 threads
GCC: 15.1.0 (MinGW-W64 x86_64-ucrt-posix-seh, built by Brecht Sanders, r4)
Architecture: x86_64

Ubuntu

OS: Ubuntu 22.04.5 LTS (Jammy Jellyfish)
Kernel: Linux 6.8.0-84-generic
CPU: Intel Core i5-8250U, 4 cores, 8 threads, Little Endian
GCC: 12.3.0
Architecture: x86_64