Skip to content

Commit

Permalink
md5: Add inline-asm backend for LoongArch64 targets (#505)
Browse files Browse the repository at this point in the history
  • Loading branch information
@heiher
heiher committed Sep 22, 2023
1 parent 7aba4b5 commit a24e8b7
Show file tree
Hide file tree
Showing 7 changed files with 369 additions and 169 deletions.
1 change: 1 addition & 0 deletions Cargo.lock
View file

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

5 changes: 5 additions & 0 deletions md5/Cargo.toml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,7 @@ name = "md5"

[dependencies]
digest = "0.10.7"
cfg-if = "1.0"

[target.'cfg(any(target_arch = "x86", target_arch = "x86_64"))'.dependencies]
md5-asm = { version = "0.5", optional = true }
Expand All @@ -28,4 +29,8 @@ hex-literal = "0.2.2"
default = ["std"]
std = ["digest/std"]
asm = ["md5-asm"] # WARNING: this feature SHOULD NOT be enabled by library crates
# Use assembly backend for LoongArch64 targets
# WARNING: Bumps MSRV to 1.72. This feature SHOULD NOT be enabled by library crates
loongarch64_asm = []
oid = ["digest/oid"] # Enable OID support. WARNING: Bumps MSRV to 1.57
force-soft = [] # Force software implementation
175 changes: 12 additions & 163 deletions md5/src/compress.rs
View file
Original file line number Diff line number Diff line change
@@ -1,165 +1,14 @@
#![allow(clippy::many_single_char_names, clippy::unreadable_literal)]
use core::convert::TryInto;

const RC: [u32; 64] = [
// round 1
0xd76aa478, 0xe8c7b756, 0x242070db, 0xc1bdceee, 0xf57c0faf, 0x4787c62a, 0xa8304613, 0xfd469501,
0x698098d8, 0x8b44f7af, 0xffff5bb1, 0x895cd7be, 0x6b901122, 0xfd987193, 0xa679438e, 0x49b40821,
// round 2
0xf61e2562, 0xc040b340, 0x265e5a51, 0xe9b6c7aa, 0xd62f105d, 0x02441453, 0xd8a1e681, 0xe7d3fbc8,
0x21e1cde6, 0xc33707d6, 0xf4d50d87, 0x455a14ed, 0xa9e3e905, 0xfcefa3f8, 0x676f02d9, 0x8d2a4c8a,
// round 3
0xfffa3942, 0x8771f681, 0x6d9d6122, 0xfde5380c, 0xa4beea44, 0x4bdecfa9, 0xf6bb4b60, 0xbebfbc70,
0x289b7ec6, 0xeaa127fa, 0xd4ef3085, 0x04881d05, 0xd9d4d039, 0xe6db99e5, 0x1fa27cf8, 0xc4ac5665,
// round 4
0xf4292244, 0x432aff97, 0xab9423a7, 0xfc93a039, 0x655b59c3, 0x8f0ccc92, 0xffeff47d, 0x85845dd1,
0x6fa87e4f, 0xfe2ce6e0, 0xa3014314, 0x4e0811a1, 0xf7537e82, 0xbd3af235, 0x2ad7d2bb, 0xeb86d391,
];

#[inline(always)]
fn op_f(w: u32, x: u32, y: u32, z: u32, m: u32, c: u32, s: u32) -> u32 {
((x & y) | (!x & z))
.wrapping_add(w)
.wrapping_add(m)
.wrapping_add(c)
.rotate_left(s)
.wrapping_add(x)
}
#[inline(always)]
fn op_g(w: u32, x: u32, y: u32, z: u32, m: u32, c: u32, s: u32) -> u32 {
((x & z) | (y & !z))
.wrapping_add(w)
.wrapping_add(m)
.wrapping_add(c)
.rotate_left(s)
.wrapping_add(x)
}

#[inline(always)]
fn op_h(w: u32, x: u32, y: u32, z: u32, m: u32, c: u32, s: u32) -> u32 {
(x ^ y ^ z)
.wrapping_add(w)
.wrapping_add(m)
.wrapping_add(c)
.rotate_left(s)
.wrapping_add(x)
}

#[inline(always)]
fn op_i(w: u32, x: u32, y: u32, z: u32, m: u32, c: u32, s: u32) -> u32 {
(y ^ (x | !z))
.wrapping_add(w)
.wrapping_add(m)
.wrapping_add(c)
.rotate_left(s)
.wrapping_add(x)
}

#[inline]
pub fn compress_block(state: &mut [u32; 4], input: &[u8; 64]) {
let mut a = state[0];
let mut b = state[1];
let mut c = state[2];
let mut d = state[3];

let mut data = [0u32; 16];
for (o, chunk) in data.iter_mut().zip(input.chunks_exact(4)) {
*o = u32::from_le_bytes(chunk.try_into().unwrap());
}

// round 1
a = op_f(a, b, c, d, data[0], RC[0], 7);
d = op_f(d, a, b, c, data[1], RC[1], 12);
c = op_f(c, d, a, b, data[2], RC[2], 17);
b = op_f(b, c, d, a, data[3], RC[3], 22);

a = op_f(a, b, c, d, data[4], RC[4], 7);
d = op_f(d, a, b, c, data[5], RC[5], 12);
c = op_f(c, d, a, b, data[6], RC[6], 17);
b = op_f(b, c, d, a, data[7], RC[7], 22);

a = op_f(a, b, c, d, data[8], RC[8], 7);
d = op_f(d, a, b, c, data[9], RC[9], 12);
c = op_f(c, d, a, b, data[10], RC[10], 17);
b = op_f(b, c, d, a, data[11], RC[11], 22);

a = op_f(a, b, c, d, data[12], RC[12], 7);
d = op_f(d, a, b, c, data[13], RC[13], 12);
c = op_f(c, d, a, b, data[14], RC[14], 17);
b = op_f(b, c, d, a, data[15], RC[15], 22);

// round 2
a = op_g(a, b, c, d, data[1], RC[16], 5);
d = op_g(d, a, b, c, data[6], RC[17], 9);
c = op_g(c, d, a, b, data[11], RC[18], 14);
b = op_g(b, c, d, a, data[0], RC[19], 20);

a = op_g(a, b, c, d, data[5], RC[20], 5);
d = op_g(d, a, b, c, data[10], RC[21], 9);
c = op_g(c, d, a, b, data[15], RC[22], 14);
b = op_g(b, c, d, a, data[4], RC[23], 20);

a = op_g(a, b, c, d, data[9], RC[24], 5);
d = op_g(d, a, b, c, data[14], RC[25], 9);
c = op_g(c, d, a, b, data[3], RC[26], 14);
b = op_g(b, c, d, a, data[8], RC[27], 20);

a = op_g(a, b, c, d, data[13], RC[28], 5);
d = op_g(d, a, b, c, data[2], RC[29], 9);
c = op_g(c, d, a, b, data[7], RC[30], 14);
b = op_g(b, c, d, a, data[12], RC[31], 20);

// round 3
a = op_h(a, b, c, d, data[5], RC[32], 4);
d = op_h(d, a, b, c, data[8], RC[33], 11);
c = op_h(c, d, a, b, data[11], RC[34], 16);
b = op_h(b, c, d, a, data[14], RC[35], 23);

a = op_h(a, b, c, d, data[1], RC[36], 4);
d = op_h(d, a, b, c, data[4], RC[37], 11);
c = op_h(c, d, a, b, data[7], RC[38], 16);
b = op_h(b, c, d, a, data[10], RC[39], 23);

a = op_h(a, b, c, d, data[13], RC[40], 4);
d = op_h(d, a, b, c, data[0], RC[41], 11);
c = op_h(c, d, a, b, data[3], RC[42], 16);
b = op_h(b, c, d, a, data[6], RC[43], 23);

a = op_h(a, b, c, d, data[9], RC[44], 4);
d = op_h(d, a, b, c, data[12], RC[45], 11);
c = op_h(c, d, a, b, data[15], RC[46], 16);
b = op_h(b, c, d, a, data[2], RC[47], 23);

// round 4
a = op_i(a, b, c, d, data[0], RC[48], 6);
d = op_i(d, a, b, c, data[7], RC[49], 10);
c = op_i(c, d, a, b, data[14], RC[50], 15);
b = op_i(b, c, d, a, data[5], RC[51], 21);

a = op_i(a, b, c, d, data[12], RC[52], 6);
d = op_i(d, a, b, c, data[3], RC[53], 10);
c = op_i(c, d, a, b, data[10], RC[54], 15);
b = op_i(b, c, d, a, data[1], RC[55], 21);

a = op_i(a, b, c, d, data[8], RC[56], 6);
d = op_i(d, a, b, c, data[15], RC[57], 10);
c = op_i(c, d, a, b, data[6], RC[58], 15);
b = op_i(b, c, d, a, data[13], RC[59], 21);

a = op_i(a, b, c, d, data[4], RC[60], 6);
d = op_i(d, a, b, c, data[11], RC[61], 10);
c = op_i(c, d, a, b, data[2], RC[62], 15);
b = op_i(b, c, d, a, data[9], RC[63], 21);

state[0] = state[0].wrapping_add(a);
state[1] = state[1].wrapping_add(b);
state[2] = state[2].wrapping_add(c);
state[3] = state[3].wrapping_add(d);
}

#[inline]
pub fn compress(state: &mut [u32; 4], blocks: &[[u8; 64]]) {
for block in blocks {
compress_block(state, block)
cfg_if::cfg_if! {
if #[cfg(feature = "force-soft")] {
mod soft;
pub use soft::compress;
} else if #[cfg(all(feature = "asm", any(target_arch = "x86", target_arch = "x86_64")))] {
pub use md5_asm::compress;
} else if #[cfg(all(feature = "loongarch64_asm", target_arch = "loongarch64"))] {
mod loongarch64_asm;
pub use loongarch64_asm::compress;
} else {
mod soft;
pub use soft::compress;
}
}
14 changes: 14 additions & 0 deletions md5/src/compress/consts.rs
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,14 @@
pub const RC: [u32; 64] = [
// round 1
0xd76aa478, 0xe8c7b756, 0x242070db, 0xc1bdceee, 0xf57c0faf, 0x4787c62a, 0xa8304613, 0xfd469501,
0x698098d8, 0x8b44f7af, 0xffff5bb1, 0x895cd7be, 0x6b901122, 0xfd987193, 0xa679438e, 0x49b40821,
// round 2
0xf61e2562, 0xc040b340, 0x265e5a51, 0xe9b6c7aa, 0xd62f105d, 0x02441453, 0xd8a1e681, 0xe7d3fbc8,
0x21e1cde6, 0xc33707d6, 0xf4d50d87, 0x455a14ed, 0xa9e3e905, 0xfcefa3f8, 0x676f02d9, 0x8d2a4c8a,
// round 3
0xfffa3942, 0x8771f681, 0x6d9d6122, 0xfde5380c, 0xa4beea44, 0x4bdecfa9, 0xf6bb4b60, 0xbebfbc70,
0x289b7ec6, 0xeaa127fa, 0xd4ef3085, 0x04881d05, 0xd9d4d039, 0xe6db99e5, 0x1fa27cf8, 0xc4ac5665,
// round 4
0xf4292244, 0x432aff97, 0xab9423a7, 0xfc93a039, 0x655b59c3, 0x8f0ccc92, 0xffeff47d, 0x85845dd1,
0x6fa87e4f, 0xfe2ce6e0, 0xa3014314, 0x4e0811a1, 0xf7537e82, 0xbd3af235, 0x2ad7d2bb, 0xeb86d391,
];
Loading

0 comments on commit a24e8b7

Please sign in to comment.