ipv6: rewrite core ipv6 methods to operate on u128s

[Alextopher]

I was doing some assembly analysis and saw that the methods on Ipv6Network were not optimizing nearly as well as the equivalent Ipv4Network methods. This PR rewrites the core methods on Ipv6Network to use u128 similarly to how Ipv4Network uses u32.

While preforming this change I addressed some clippy lints and made some more common-sense additions.

Results on .contains()

Before:

.section .text.ipnetwork::ipv6::Ipv6Network::contains,"ax",@progbits
	.globl	ipnetwork::ipv6::Ipv6Network::contains
	.p2align	4, 0x90
	.type	ipnetwork::ipv6::Ipv6Network::contains,@function
ipnetwork::ipv6::Ipv6Network::contains:
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 226
		pub fn contains(&self, ip: Ipv6Addr) -> bool {
	.cfi_startproc
	push rbp
	.cfi_def_cfa_offset 16
	push r15
	.cfi_def_cfa_offset 24
	push r14
	.cfi_def_cfa_offset 32
	push rbx
	.cfi_def_cfa_offset 40
	.cfi_offset rbx, -40
	.cfi_offset r14, -32
	.cfi_offset r15, -24
	.cfi_offset rbp, -16
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/net/ip_addr.rs : 1373
		let [a, b, c, d, e, f, g, h] = unsafe { transmute::<_, [u16; 8]>(self.octets) };
	movzx edx, word ptr [rsi]
	movzx r11d, byte ptr [rdi + 16]
	xor eax, eax
	mov r8d, r11d
	sub r8b, 16
	mov r9d, 16
	mov r10d, r11d
	cmovae r10d, r9d
	mov ecx, -65536
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 206
		let mask = !(0xffff >> set_bits) as u16;
	sarx ebx, ecx, r10d
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 208
		chunk[1] = mask as u8;
	movzx ebp, bh
	vmovd xmm0, ebp
	vpinsrb xmm0, xmm0, ebx, 1
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/num/mod.rs : 493
		uint_impl! {
	movzx r8d, r8b
	cmovb r8d, eax
	cmp r8b, 16
	cmovae r8d, r9d
	mov r10d, r11d
	sub r10b, 32
	movzx r10d, r10b
	cmovb r10d, eax
	cmp r10b, 16
	cmovae r10d, r9d
	mov ebx, r11d
	sub bl, 48
	movzx ebx, bl
	cmovb ebx, eax
	cmp bl, 16
	cmovae ebx, r9d
	mov ebp, r11d
	sub bpl, 64
	movzx ebp, bpl
	cmovb ebp, eax
	cmp bpl, 16
	cmovae ebp, r9d
	mov r14d, r11d
	sub r14b, 80
	movzx r14d, r14b
	cmovb r14d, eax
	cmp r14b, 16
	cmovae r14d, r9d
	mov r15d, r11d
	sub r15b, 96
	movzx r15d, r15b
	cmovb r15d, eax
	cmp r15b, 16
	cmovae r15d, r9d
	sub r11b, 112
	movzx r11d, r11b
	cmovb r11d, eax
	cmp r11b, 16
	cmovae r11d, r9d
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/net/ip_addr.rs : 1373
		let [a, b, c, d, e, f, g, h] = unsafe { transmute::<_, [u16; 8]>(self.octets) };
	vmovd r9d, xmm0
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 234
		.all(|(mask, (a, b))| a & mask == b & mask)
	xor dx, word ptr [rdi]
	test dx, r9w
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/iter/traits/iterator.rs : 2462
		accum = f(accum, x)?;
	jne .LBB39_14
	sarx eax, ecx, r8d
	mov edx, eax
	shr edx, 8
	vpinsrb xmm0, xmm0, edx, 2
	vpinsrb xmm0, xmm0, eax, 3
	sarx eax, ecx, r10d
	mov edx, eax
	shr edx, 8
	vpinsrb xmm0, xmm0, edx, 4
	vpinsrb xmm5, xmm0, eax, 5
	sarx eax, ecx, ebx
	mov edx, eax
	shr edx, 8
	vpinsrb xmm0, xmm5, edx, 6
	vpinsrb xmm4, xmm0, eax, 7
	sarx eax, ecx, ebp
	mov edx, eax
	shr edx, 8
	vpinsrb xmm0, xmm4, edx, 8
	vpinsrb xmm3, xmm0, eax, 9
	sarx eax, ecx, r14d
	mov edx, eax
	shr edx, 8
	vpinsrb xmm0, xmm3, edx, 10
	vpinsrb xmm2, xmm0, eax, 11
	sarx eax, ecx, r15d
	mov edx, eax
	shr edx, 8
	vpinsrb xmm0, xmm2, edx, 12
	vpinsrb xmm1, xmm0, eax, 13
	sarx eax, ecx, r11d
	mov ecx, eax
	shr ecx, 8
	vpinsrb xmm0, xmm1, ecx, 14
	vpinsrb xmm0, xmm0, eax, 15
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/net/ip_addr.rs : 1373
		let [a, b, c, d, e, f, g, h] = unsafe { transmute::<_, [u16; 8]>(self.octets) };
	vpextrw eax, xmm0, 1
	movzx ecx, word ptr [rsi + 2]
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 234
		.all(|(mask, (a, b))| a & mask == b & mask)
	xor cx, word ptr [rdi + 2]
	test ecx, eax
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/iter/traits/iterator.rs : 2462
		accum = f(accum, x)?;
	jne .LBB39_2
	movzx eax, word ptr [rdi + 4]
	movzx ecx, word ptr [rsi + 4]
	vpextrw edx, xmm5, 2
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 234
		.all(|(mask, (a, b))| a & mask == b & mask)
	xor ecx, eax
	test cx, dx
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/iter/traits/iterator.rs : 2462
		accum = f(accum, x)?;
	jne .LBB39_2
	movzx eax, word ptr [rdi + 6]
	movzx ecx, word ptr [rsi + 6]
	vpextrw edx, xmm4, 3
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 234
		.all(|(mask, (a, b))| a & mask == b & mask)
	xor ecx, eax
	test cx, dx
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/iter/traits/iterator.rs : 2462
		accum = f(accum, x)?;
	jne .LBB39_2
	movzx eax, word ptr [rdi + 8]
	movzx ecx, word ptr [rsi + 8]
	vpextrw edx, xmm3, 4
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 234
		.all(|(mask, (a, b))| a & mask == b & mask)
	xor ecx, eax
	test cx, dx
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/iter/traits/iterator.rs : 2462
		accum = f(accum, x)?;
	jne .LBB39_2
	movzx eax, word ptr [rdi + 10]
	movzx ecx, word ptr [rsi + 10]
	vpextrw edx, xmm2, 5
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 234
		.all(|(mask, (a, b))| a & mask == b & mask)
	xor ecx, eax
	test cx, dx
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/iter/traits/iterator.rs : 2462
		accum = f(accum, x)?;
	jne .LBB39_2
	movzx eax, word ptr [rdi + 12]
	movzx ecx, word ptr [rsi + 12]
	vpextrw edx, xmm1, 6
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 234
		.all(|(mask, (a, b))| a & mask == b & mask)
	xor ecx, eax
	test cx, dx
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/iter/traits/iterator.rs : 2462
		accum = f(accum, x)?;
	je .LBB39_13
.LBB39_2:
	xor eax, eax
.LBB39_14:
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 235
		}
	pop rbx
	.cfi_def_cfa_offset 32
	pop r14
	.cfi_def_cfa_offset 24
	pop r15
	.cfi_def_cfa_offset 16
	pop rbp
	.cfi_def_cfa_offset 8
	ret
.LBB39_13:
	.cfi_def_cfa_offset 40
	movzx eax, word ptr [rdi + 14]
	movzx ecx, word ptr [rsi + 14]
	vpextrw edx, xmm0, 7
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 234
		.all(|(mask, (a, b))| a & mask == b & mask)
	xor ecx, eax
	test cx, dx
	sete al
	jmp .LBB39_14

After:

.section .text.ipnetwork::ipv6::Ipv6Network::contains,"ax",@progbits
	.globl	ipnetwork::ipv6::Ipv6Network::contains
	.p2align	4, 0x90
	.type	ipnetwork::ipv6::Ipv6Network::contains,@function
ipnetwork::ipv6::Ipv6Network::contains:
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 242
		pub fn contains(&self, ip: Ipv6Addr) -> bool {
	.cfi_startproc
	movzx eax, byte ptr [rdi + 16]
	xor ecx, ecx
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 190
		let mask = u128::MAX << (IPV6_BITS - self.prefix);
	xor edx, edx
	sub dl, al
	mov rax, -1
	shlx r8, rax, rdx
	test dl, 64
	cmove rcx, r8
	cmovne rax, r8
		// /rustc/25ef9e3d85d934b27d9dada2f9dd52b1dc63bb04/library/core/src/num/mod.rs : 1205
		uint_impl! {
	bswap rax
	bswap rcx
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 208
		let network = u128::from(self.addr) & mask;
	mov rdx, qword ptr [rdi]
	mov rdi, qword ptr [rdi + 8]
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 248
		(ip & mask) == net
	xor rdx, qword ptr [rsi]
	and rdx, rax
	xor rdi, qword ptr [rsi + 8]
	and rdi, rcx
	or rdi, rdx
	sete al
		// /home/mahonec/p/github.com/Alextopher/ipnetwork/src/ipv6.rs : 249
		}
	ret

[achanda]

Thanks! This is great.