ferritls-core 0.8.1

//! SHA-2 家族：SHA-256 / SHA-384 / SHA-512（FIPS 180-4）。
//!
//! FIPS 批准；上电自检覆盖（M5）。`Sha256` 的块压缩经 [`crate::ops`]
//! **函数分发**（未安装 → 软件直连，`new()` 不查任何全局；安装 SHA-NI
//! 后端后只替换压缩函数指针，上下文保持哑结构——OpenSSL 模式，见
//! [`crate::ops`] 模块文档）。SHA-384/512 无 x86 硬件指令对应，保持
//! 软件直连。
//!
//! **显式省略**：SHA-224/SHA-512/224/SHA-512/256（FIPS 180-4 其余三个
//! 变体）——TLS 1.2/1.3 与本模块全部消费者只用 256/384/512，边界内
//! 不留无消费者的变体；且 256 侧为手写结构体（512 侧经 `sha512_family!`
//! 宏实例化），补 SHA-224 需先宏化。若需要，按 512 侧宏化 → 加变体
//! → SHAVS 向量的顺序扩展。
//!
//! 向量：NIST CAVP SHAVS（`tests/sha2.rs`，含 "abc"/空串/两块消息 KAT 与
//! 流式一致性检查）。

const K256: [u32; 64] = [
    0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5, 0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5,
    0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3, 0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174,
    0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc, 0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da,
    0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7, 0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967,
    0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13, 0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85,
    0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3, 0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070,
    0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5, 0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3,
    0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208, 0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2,
];

const IV256: [u32; 8] = [
    0x6a09e667, 0xbb67ae85, 0x3c6ef372, 0xa54ff53a, 0x510e527f, 0x9b05688c, 0x1f83d9ab, 0x5be0cd19,
];

/// SHA-256 摘要器（流式）。
#[derive(Clone)]
pub struct Sha256 {
    h: [u32; 8],
    buf: [u8; 64],
    buf_len: usize,
    total: u64,
}

impl Sha256 {
    /// 摘要输出字节数。
    pub const OUTPUT_LEN: usize = 32;
    /// 压缩函数块字节数。
    pub const BLOCK_LEN: usize = 64;

    /// 创建新的摘要器。
    pub fn new() -> Self {
        Self {
            h: IV256,
            buf: [0; 64],
            buf_len: 0,
            total: 0,
        }
    }

    /// 吸入数据。可多次调用。
    pub fn update(&mut self, mut data: &[u8]) {
        let compress = crate::ops::current_sha256_compress();
        self.total = self.total.wrapping_add(data.len() as u64);
        if self.buf_len > 0 {
            let take = (64 - self.buf_len).min(data.len());
            self.buf[self.buf_len..self.buf_len + take].copy_from_slice(&data[..take]);
            self.buf_len += take;
            data = &data[take..];
            if self.buf_len == 64 {
                let block = self.buf;
                compress(&mut self.h, &block);
                self.buf_len = 0;
            }
        }
        while data.len() >= 64 {
            let mut block = [0u8; 64];
            block.copy_from_slice(&data[..64]);
            compress(&mut self.h, &block);
            data = &data[64..];
        }
        if !data.is_empty() {
            self.buf[..data.len()].copy_from_slice(data);
            self.buf_len = data.len();
        }
    }

    /// 结束并输出摘要。消耗 `self` 以便内部状态被清零。
    ///
    /// 填充直接写入缓冲（0x80、零、64 位大端长度），数据等价于
    /// 旧实现「update(0x80) 后逐字节补零」的路径，但只取一次压缩
    /// 函数、最多压缩两次。
    pub fn finalize(mut self) -> [u8; 32] {
        let compress = crate::ops::current_sha256_compress();
        let bit_len = self.total.wrapping_mul(8);
        self.buf[self.buf_len] = 0x80;
        self.buf_len += 1;
        if self.buf_len > 56 {
            // 0x80 放不进本块：本块补零压掉，长度单独成块。
            self.buf[self.buf_len..].fill(0);
            let block = self.buf;
            compress(&mut self.h, &block);
            self.buf = [0u8; 64];
        } else {
            self.buf[self.buf_len..56].fill(0);
        }
        self.buf[56..64].copy_from_slice(&bit_len.to_be_bytes());
        let block = self.buf;
        compress(&mut self.h, &block);

        let mut out = [0u8; 32];
        for (i, w) in self.h.iter().enumerate() {
            out[i * 4..i * 4 + 4].copy_from_slice(&w.to_be_bytes());
        }
        out
    }

    /// 一次性摘要便捷函数。
    pub fn one_shot(data: &[u8]) -> [u8; 32] {
        let mut h = Self::new();
        h.update(data);
        h.finalize()
    }
}

impl Default for Sha256 {
    fn default() -> Self {
        Self::new()
    }
}

pub(crate) fn compress256(h: &mut [u32; 8], block: &[u8; 64]) {
    let mut w = [0u32; 64];
    for i in 0..16 {
        w[i] = u32::from_be_bytes([
            block[i * 4],
            block[i * 4 + 1],
            block[i * 4 + 2],
            block[i * 4 + 3],
        ]);
    }

    let [mut a, mut b, mut c, mut d, mut e, mut f, mut g, mut hh] = *h;
    for i in 0..64 {
        // 软流水：把 w[i+16] 的计算提前 16 轮到其被使用之前——独立
        // 加法/旋转链与压缩轮关键路径并行，省去独立展开循环的串行
        // 前导延迟（2026-09-17 P3：默认与宽 ISA 全面快于两段式 ~14%
        // 与滚动窗口融合 ~9–10%，见 BENCHMARKS §5.5/§5.6）。64 轮
        // 定长循环经完全展开后该条件为编译期常量，无运行期分支。
        if i < 48 {
            let s0 = w[i + 1].rotate_right(7) ^ w[i + 1].rotate_right(18) ^ (w[i + 1] >> 3);
            let s1 = w[i + 14].rotate_right(17) ^ w[i + 14].rotate_right(19) ^ (w[i + 14] >> 10);
            w[i + 16] = w[i]
                .wrapping_add(s0)
                .wrapping_add(w[i + 9])
                .wrapping_add(s1);
        }
        let s1 = e.rotate_right(6) ^ e.rotate_right(11) ^ e.rotate_right(25);
        let ch = (e & f) ^ (!e & g);
        let t1 = hh
            .wrapping_add(s1)
            .wrapping_add(ch)
            .wrapping_add(K256[i])
            .wrapping_add(w[i]);
        let s0 = a.rotate_right(2) ^ a.rotate_right(13) ^ a.rotate_right(22);
        let maj = (a & b) ^ (a & c) ^ (b & c);
        let t2 = s0.wrapping_add(maj);
        hh = g;
        g = f;
        f = e;
        e = d.wrapping_add(t1);
        d = c;
        c = b;
        b = a;
        a = t1.wrapping_add(t2);
    }
    h[0] = h[0].wrapping_add(a);
    h[1] = h[1].wrapping_add(b);
    h[2] = h[2].wrapping_add(c);
    h[3] = h[3].wrapping_add(d);
    h[4] = h[4].wrapping_add(e);
    h[5] = h[5].wrapping_add(f);
    h[6] = h[6].wrapping_add(g);
    h[7] = h[7].wrapping_add(hh);
}

const K512: [u64; 80] = [
    0x428a2f98d728ae22,
    0x7137449123ef65cd,
    0xb5c0fbcfec4d3b2f,
    0xe9b5dba58189dbbc,
    0x3956c25bf348b538,
    0x59f111f1b605d019,
    0x923f82a4af194f9b,
    0xab1c5ed5da6d8118,
    0xd807aa98a3030242,
    0x12835b0145706fbe,
    0x243185be4ee4b28c,
    0x550c7dc3d5ffb4e2,
    0x72be5d74f27b896f,
    0x80deb1fe3b1696b1,
    0x9bdc06a725c71235,
    0xc19bf174cf692694,
    0xe49b69c19ef14ad2,
    0xefbe4786384f25e3,
    0x0fc19dc68b8cd5b5,
    0x240ca1cc77ac9c65,
    0x2de92c6f592b0275,
    0x4a7484aa6ea6e483,
    0x5cb0a9dcbd41fbd4,
    0x76f988da831153b5,
    0x983e5152ee66dfab,
    0xa831c66d2db43210,
    0xb00327c898fb213f,
    0xbf597fc7beef0ee4,
    0xc6e00bf33da88fc2,
    0xd5a79147930aa725,
    0x06ca6351e003826f,
    0x142929670a0e6e70,
    0x27b70a8546d22ffc,
    0x2e1b21385c26c926,
    0x4d2c6dfc5ac42aed,
    0x53380d139d95b3df,
    0x650a73548baf63de,
    0x766a0abb3c77b2a8,
    0x81c2c92e47edaee6,
    0x92722c851482353b,
    0xa2bfe8a14cf10364,
    0xa81a664bbc423001,
    0xc24b8b70d0f89791,
    0xc76c51a30654be30,
    0xd192e819d6ef5218,
    0xd69906245565a910,
    0xf40e35855771202a,
    0x106aa07032bbd1b8,
    0x19a4c116b8d2d0c8,
    0x1e376c085141ab53,
    0x2748774cdf8eeb99,
    0x34b0bcb5e19b48a8,
    0x391c0cb3c5c95a63,
    0x4ed8aa4ae3418acb,
    0x5b9cca4f7763e373,
    0x682e6ff3d6b2b8a3,
    0x748f82ee5defb2fc,
    0x78a5636f43172f60,
    0x84c87814a1f0ab72,
    0x8cc702081a6439ec,
    0x90befffa23631e28,
    0xa4506cebde82bde9,
    0xbef9a3f7b2c67915,
    0xc67178f2e372532b,
    0xca273eceea26619c,
    0xd186b8c721c0c207,
    0xeada7dd6cde0eb1e,
    0xf57d4f7fee6ed178,
    0x06f067aa72176fba,
    0x0a637dc5a2c898a6,
    0x113f9804bef90dae,
    0x1b710b35131c471b,
    0x28db77f523047d84,
    0x32caab7b40c72493,
    0x3c9ebe0a15c9bebc,
    0x431d67c49c100d4c,
    0x4cc5d4becb3e42b6,
    0x597f299cfc657e2a,
    0x5fcb6fab3ad6faec,
    0x6c44198c4a475817,
];

const IV512: [u64; 8] = [
    0x6a09e667f3bcc908,
    0xbb67ae8584caa73b,
    0x3c6ef372fe94f82b,
    0xa54ff53a5f1d36f1,
    0x510e527fade682d1,
    0x9b05688c2b3e6c1f,
    0x1f83d9abfb41bd6b,
    0x5be0cd19137e2179,
];

const IV384: [u64; 8] = [
    0xcbbb9d5dc1059ed8,
    0x629a292a367cd507,
    0x9159015a3070dd17,
    0x152fecd8f70e5939,
    0x67332667ffc00b31,
    0x8eb44a8768581511,
    0xdb0c2e0d64f98fa7,
    0x47b5481dbefa4fa4,
];

macro_rules! sha512_family {
    ($name:ident, $out:expr, $iv:expr, $doc:expr) => {
        #[doc = $doc]
        #[derive(Clone)]
        pub struct $name {
            h: [u64; 8],
            buf: [u8; 128],
            buf_len: usize,
            total: u64,
        }

        impl $name {
            /// 摘要输出字节数。
            pub const OUTPUT_LEN: usize = $out;
            /// 压缩函数块字节数。
            pub const BLOCK_LEN: usize = 128;

            /// 创建新的摘要器。
            pub fn new() -> Self {
                Self {
                    h: $iv,
                    buf: [0; 128],
                    buf_len: 0,
                    total: 0,
                }
            }

            /// 吸入数据。可多次调用。
            pub fn update(&mut self, mut data: &[u8]) {
                self.total = self.total.wrapping_add(data.len() as u64);
                if self.buf_len > 0 {
                    let take = (128 - self.buf_len).min(data.len());
                    self.buf[self.buf_len..self.buf_len + take].copy_from_slice(&data[..take]);
                    self.buf_len += take;
                    data = &data[take..];
                    if self.buf_len == 128 {
                        let block = self.buf;
                        compress512(&mut self.h, &block);
                        self.buf_len = 0;
                    }
                }
                while data.len() >= 128 {
                    let mut block = [0u8; 128];
                    block.copy_from_slice(&data[..128]);
                    compress512(&mut self.h, &block);
                    data = &data[128..];
                }
                if !data.is_empty() {
                    self.buf[..data.len()].copy_from_slice(data);
                    self.buf_len = data.len();
                }
            }

            /// 结束并输出摘要。
            pub fn finalize(mut self) -> [u8; $out] {
                let bit_len = (self.total as u128).wrapping_mul(8);
                self.update(&[0x80]);
                while self.buf_len != 112 {
                    self.update(&[0]);
                }
                self.buf[112..128].copy_from_slice(&bit_len.to_be_bytes());
                let block = self.buf;
                compress512(&mut self.h, &block);

                let mut out = [0u8; $out];
                for (i, w) in self.h.iter().take($out / 8).enumerate() {
                    out[i * 8..i * 8 + 8].copy_from_slice(&w.to_be_bytes());
                }
                out
            }

            /// 一次性摘要便捷函数。
            pub fn one_shot(data: &[u8]) -> [u8; $out] {
                let mut h = Self::new();
                h.update(data);
                h.finalize()
            }
        }

        impl Default for $name {
            fn default() -> Self {
                Self::new()
            }
        }
    };
}

sha512_family!(
    Sha384,
    48,
    IV384,
    "SHA-384 摘要器（流式；SHA-512 算法 + 截断 48 字节）。"
);
sha512_family!(Sha512, 64, IV512, "SHA-512 摘要器（流式）。");

fn compress512(h: &mut [u64; 8], block: &[u8; 128]) {
    let mut w = [0u64; 80];
    for i in 0..16 {
        let mut b = [0u8; 8];
        b.copy_from_slice(&block[i * 8..i * 8 + 8]);
        w[i] = u64::from_be_bytes(b);
    }
    for i in 16..80 {
        let s0 = w[i - 15].rotate_right(1) ^ w[i - 15].rotate_right(8) ^ (w[i - 15] >> 7);
        let s1 = w[i - 2].rotate_right(19) ^ w[i - 2].rotate_right(61) ^ (w[i - 2] >> 6);
        w[i] = w[i - 16]
            .wrapping_add(s0)
            .wrapping_add(w[i - 7])
            .wrapping_add(s1);
    }

    let [mut a, mut b, mut c, mut d, mut e, mut f, mut g, mut hh] = *h;
    for i in 0..80 {
        let s1 = e.rotate_right(14) ^ e.rotate_right(18) ^ e.rotate_right(41);
        let ch = (e & f) ^ (!e & g);
        let t1 = hh
            .wrapping_add(s1)
            .wrapping_add(ch)
            .wrapping_add(K512[i])
            .wrapping_add(w[i]);
        let s0 = a.rotate_right(28) ^ a.rotate_right(34) ^ a.rotate_right(39);
        let maj = (a & b) ^ (a & c) ^ (b & c);
        let t2 = s0.wrapping_add(maj);
        hh = g;
        g = f;
        f = e;
        e = d.wrapping_add(t1);
        d = c;
        c = b;
        b = a;
        a = t1.wrapping_add(t2);
    }
    h[0] = h[0].wrapping_add(a);
    h[1] = h[1].wrapping_add(b);
    h[2] = h[2].wrapping_add(c);
    h[3] = h[3].wrapping_add(d);
    h[4] = h[4].wrapping_add(e);
    h[5] = h[5].wrapping_add(f);
    h[6] = h[6].wrapping_add(g);
    h[7] = h[7].wrapping_add(hh);
}