Please explain some of your code

[playday3008]

What does it do (except calling W::call)?
Why do you order registers in union like that?
What is that magic constants, like: 0x38 and 0x88

WidescreenFixesPack/includes/assembly64.hpp

Lines 249 to 297 in dab621f

	inline uintptr_t make_reg_pack_and_call()
	{
	static reg_pack pack = { 0 };
	auto fn = injector::asm_invoke([](x86::Assembler& a)
	{
	pushad();
	a.mov(r15, &pack);
	for (size_t i = 0; i < reg_pack::reg_rf; i++)
	{
	a.mov(r14, qword_ptr(rsp, i * sizeof(uint64_t)));
	a.mov(qword_ptr(r15, i * sizeof(uint64_t)), r14);
	}
	a.mov(r14, rsp);
	a.add(r14, 0x88);
	a.mov(r14, qword_ptr(r14));
	a.mov(qword_ptr(r15, reg_pack::reg_ret_ptr * sizeof(uint64_t)), r14);
	a.mov(rcx, &pack);
	a.call(W::call);
	popad();
	a.mov(rax, &pack);
	a.mov(r15, qword_ptr(rax, 0 * sizeof(uint64_t)));
	a.mov(r14, qword_ptr(rax, 1 * sizeof(uint64_t)));
	a.mov(r13, qword_ptr(rax, 2 * sizeof(uint64_t)));
	a.mov(r12, qword_ptr(rax, 3 * sizeof(uint64_t)));
	a.mov(r11, qword_ptr(rax, 4 * sizeof(uint64_t)));
	a.mov(r10, qword_ptr(rax, 5 * sizeof(uint64_t)));
	a.mov(r9, qword_ptr(rax, 6 * sizeof(uint64_t)));
	a.mov(r8, qword_ptr(rax, 7 * sizeof(uint64_t)));
	a.mov(rbp, qword_ptr(rax, 8 * sizeof(uint64_t)));
	a.mov(rsp, qword_ptr(rax, 9 * sizeof(uint64_t)));
	a.pushfq();
	a.add(rsp, 0x38);
	a.popfq();
	a.mov(rdi, qword_ptr(rax, 10 * sizeof(uint64_t)));
	a.mov(rsi, qword_ptr(rax, 11 * sizeof(uint64_t)));
	a.mov(rbx, qword_ptr(rax, 12 * sizeof(uint64_t)));
	a.mov(rdx, qword_ptr(rax, 13 * sizeof(uint64_t)));
	a.mov(rcx, qword_ptr(rax, 14 * sizeof(uint64_t)));
	a.mov(rax, qword_ptr(rax, 15 * sizeof(uint64_t)));
	a.ret();
	});
	return (uintptr_t)fn;
	}

[ThirteenAG]

Maybe check original source for x86? https://github.com/thelink2012/injector/blob/master/include/injector/assembly.hpp#L96
It has comments and overall more readable.

[playday3008]

I've checked, but assembly code not similar at all, and I would like to know what this code do

[playday3008]

And in x86 version registers have predefined order (described in Intel® 64 and IA-32 Architectures Software Developer’s Manual (Volume 2: Instruction Set Reference, A-Z))
But for x64, there's no official documentation about register push/pop order

[playday3008]

Also I'm a newbie in assembler, and I would like to know, what is 0x88 and 0x38, for me, it's just a magic constants

[ThirteenAG]

0x88 is the size of all registers, 0x38 I got by trial and error I think, don't remember now.

[ThirteenAG]

What you mean it's not similar, it's the same. The point is to save registers before calling and restore them after.

But for x64, there's no official documentation about register push/pop order

Don't know why does that matter.

[playday3008]

Similar != doing the same thing,
Similar == implementation

[ThirteenAG]

Implementation obviously is gonna be different between x64 and x86.

[playday3008]

Yes, I know, and I would like to understand these differences