diff options
Diffstat (limited to 'vendor/sha2')
-rw-r--r-- | vendor/sha2/.cargo-checksum.json | 2 | ||||
-rw-r--r-- | vendor/sha2/CHANGELOG.md | 6 | ||||
-rw-r--r-- | vendor/sha2/Cargo.toml | 3 | ||||
-rw-r--r-- | vendor/sha2/src/sha256.rs | 3 | ||||
-rw-r--r-- | vendor/sha2/src/sha256/loongarch64_asm.rs | 227 | ||||
-rw-r--r-- | vendor/sha2/src/sha512.rs | 3 | ||||
-rw-r--r-- | vendor/sha2/src/sha512/loongarch64_asm.rs | 242 |
7 files changed, 484 insertions, 2 deletions
diff --git a/vendor/sha2/.cargo-checksum.json b/vendor/sha2/.cargo-checksum.json index 2716fafc7..7c069038f 100644 --- a/vendor/sha2/.cargo-checksum.json +++ b/vendor/sha2/.cargo-checksum.json @@ -1 +1 @@ -{"files":{"CHANGELOG.md":"61b6ae2d035d98d2b5612a5c1c40238f4fa7e7bb3062daacf677abb4f7c60eab","Cargo.toml":"7d0b0694d8fd9d50c060a871ae98f80f392473086ec0723aa2017081c8654dce","LICENSE-APACHE":"a9040321c3712d8fd0b09cf52b17445de04a23a10165049ae187cd39e5c86be5","LICENSE-MIT":"b4eb00df6e2a4d22518fcaa6a2b4646f249b3a3c9814509b22bd2091f1392ff1","README.md":"b7af562922e4a631657acf264772d2af2b72a08d9bbc5fbcf56d9324f9027708","benches/mod.rs":"c32d9f91a541821ea988c14eee710963e623ef1edf69b02b41a29bc44e04ba95","src/consts.rs":"2f820349fa7cbf9fecc1d4aabbd1a721bb1badc3f32ef9e903826960b6f42523","src/core_api.rs":"73b160d98bfa6737688875ad73da5e3c2c93582604dc313d208200e12fdab676","src/lib.rs":"9d0ec0ba86a801bd9b2024f0b84ee322a26c7376a623dd61210e0eb9d6355aa1","src/sha256.rs":"cfc2b62a412112e471781a770793f0ba0466594b2e37001334562f3d95f340ce","src/sha256/aarch64.rs":"18121a25867a575fec8ef64da763693ece4e3e3e84da095254b8471234c6f1f8","src/sha256/soft.rs":"98e765a8e8dfa0af31f2b76570f212e6b3099522bf300e1554cbbd9fd5d02960","src/sha256/x86.rs":"70f1597f2029522b35bfd026df0a8908f086523ab2a80ba3ef35e6231b56353c","src/sha512.rs":"b0c94cf6e1a4a8efb8ccc494da96fdf805b1745fef7614875df5b4db9ee186da","src/sha512/aarch64.rs":"2ed929329a0fa66180e4726d028713a49f99cc223e635078fc1f3252a44981e0","src/sha512/soft.rs":"0183ad89418b886859d2afa9bf061bc92759ae337c1d26147b4300042e63ef42","src/sha512/x86.rs":"c7dd8bdf3212e1e8c4cc9cc6b380dc0468f79dcfd0f61a445d0d38cead45a03a","tests/data/sha224.blb":"59b185972521af418fd49a079de3d5f5bed74cd76d80473da51cab3faee6c7d0","tests/data/sha256.blb":"bb096934bb7e43e41ce143d211397afca6fcdfe243a39811688ea31aae6f800a","tests/data/sha384.blb":"e8fe66c07ba336fae2c0aa4c87cb768f41bd4ed318ee1a36fbde0a68581946ec","tests/data/sha512.blb":"1cc0e86571f2f4e3bc81438ce7b6c25c118d2d7437355240113f59cbb782c8d6","tests/data/sha512_224.blb":"b02dd46741db1034112e0888d0cdb233a21b9a82c319456f806bbaae49acf440","tests/data/sha512_256.blb":"95195b758e362d92ff0cebebac4cca696512ea5811b635243bc70e29164e5786","tests/mod.rs":"61be596fd9b45a8db345950ff2ed6f87eaf4d239ac156885f36e819da0597644"},"package":"479fb9d862239e610720565ca91403019f2f00410f1864c5aa7479b950a76ed8"}
\ No newline at end of file +{"files":{"CHANGELOG.md":"b7b0a14409ac2880f86fe50d9584acc81f2346ebcb4e46a9e2235b54ac5b02ef","Cargo.toml":"5fdf94b86fc47d105d2f2cc55c6346d15e7f3d2d7ea92031b1ce2d24276e7778","LICENSE-APACHE":"a9040321c3712d8fd0b09cf52b17445de04a23a10165049ae187cd39e5c86be5","LICENSE-MIT":"b4eb00df6e2a4d22518fcaa6a2b4646f249b3a3c9814509b22bd2091f1392ff1","README.md":"b7af562922e4a631657acf264772d2af2b72a08d9bbc5fbcf56d9324f9027708","benches/mod.rs":"c32d9f91a541821ea988c14eee710963e623ef1edf69b02b41a29bc44e04ba95","src/consts.rs":"2f820349fa7cbf9fecc1d4aabbd1a721bb1badc3f32ef9e903826960b6f42523","src/core_api.rs":"73b160d98bfa6737688875ad73da5e3c2c93582604dc313d208200e12fdab676","src/lib.rs":"9d0ec0ba86a801bd9b2024f0b84ee322a26c7376a623dd61210e0eb9d6355aa1","src/sha256.rs":"78e84eea5d517554aa5a10860bf2ce5013ca26d529e78643cd59062546e0746f","src/sha256/aarch64.rs":"18121a25867a575fec8ef64da763693ece4e3e3e84da095254b8471234c6f1f8","src/sha256/loongarch64_asm.rs":"79e2d5e3c039581e2319f8789de9ed13a8dd819ebffd13532dbd83448c7ad662","src/sha256/soft.rs":"98e765a8e8dfa0af31f2b76570f212e6b3099522bf300e1554cbbd9fd5d02960","src/sha256/x86.rs":"70f1597f2029522b35bfd026df0a8908f086523ab2a80ba3ef35e6231b56353c","src/sha512.rs":"1b19c23c63e9cfca8b42fd9e108a8570dd03e22a37d4d6f499f2fa5e566cb2de","src/sha512/aarch64.rs":"2ed929329a0fa66180e4726d028713a49f99cc223e635078fc1f3252a44981e0","src/sha512/loongarch64_asm.rs":"58a7b54d95a0e037ba80570d96ffe0dd7c0014c7fcb45b90725e522cc4992d8a","src/sha512/soft.rs":"0183ad89418b886859d2afa9bf061bc92759ae337c1d26147b4300042e63ef42","src/sha512/x86.rs":"c7dd8bdf3212e1e8c4cc9cc6b380dc0468f79dcfd0f61a445d0d38cead45a03a","tests/data/sha224.blb":"59b185972521af418fd49a079de3d5f5bed74cd76d80473da51cab3faee6c7d0","tests/data/sha256.blb":"bb096934bb7e43e41ce143d211397afca6fcdfe243a39811688ea31aae6f800a","tests/data/sha384.blb":"e8fe66c07ba336fae2c0aa4c87cb768f41bd4ed318ee1a36fbde0a68581946ec","tests/data/sha512.blb":"1cc0e86571f2f4e3bc81438ce7b6c25c118d2d7437355240113f59cbb782c8d6","tests/data/sha512_224.blb":"b02dd46741db1034112e0888d0cdb233a21b9a82c319456f806bbaae49acf440","tests/data/sha512_256.blb":"95195b758e362d92ff0cebebac4cca696512ea5811b635243bc70e29164e5786","tests/mod.rs":"61be596fd9b45a8db345950ff2ed6f87eaf4d239ac156885f36e819da0597644"},"package":"793db75ad2bcafc3ffa7c68b215fee268f537982cd901d132f89c6343f3a3dc8"}
\ No newline at end of file diff --git a/vendor/sha2/CHANGELOG.md b/vendor/sha2/CHANGELOG.md index a552266e5..a5182bcad 100644 --- a/vendor/sha2/CHANGELOG.md +++ b/vendor/sha2/CHANGELOG.md @@ -5,6 +5,12 @@ All notable changes to this project will be documented in this file. The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). +## 0.10.8 (2023-09-26) +### Added +- `asm!`-based backend for LoongArch64 targets gated behind `loongarch64_asm` feature [#507] + +[#507]: https://github.com/RustCrypto/hashes/pull/507 + ## 0.10.7 (2023-06-15) ### Added - AArch64 Neon-based backend ([#490]) diff --git a/vendor/sha2/Cargo.toml b/vendor/sha2/Cargo.toml index c19f3241d..0a9655ce0 100644 --- a/vendor/sha2/Cargo.toml +++ b/vendor/sha2/Cargo.toml @@ -12,7 +12,7 @@ [package] edition = "2018" name = "sha2" -version = "0.10.7" +version = "0.10.8" authors = ["RustCrypto Developers"] description = """ Pure Rust implementation of the SHA-2 hash function family @@ -59,6 +59,7 @@ asm-aarch64 = ["asm"] compress = [] default = ["std"] force-soft = [] +loongarch64_asm = [] oid = ["digest/oid"] std = ["digest/std"] diff --git a/vendor/sha2/src/sha256.rs b/vendor/sha2/src/sha256.rs index a45331e17..8f8287836 100644 --- a/vendor/sha2/src/sha256.rs +++ b/vendor/sha2/src/sha256.rs @@ -17,6 +17,9 @@ cfg_if::cfg_if! { mod soft; mod aarch64; use aarch64::compress; + } else if #[cfg(all(feature = "loongarch64_asm", target_arch = "loongarch64"))] { + mod loongarch64_asm; + use loongarch64_asm::compress; } else { mod soft; use soft::compress; diff --git a/vendor/sha2/src/sha256/loongarch64_asm.rs b/vendor/sha2/src/sha256/loongarch64_asm.rs new file mode 100644 index 000000000..c80fce8bb --- /dev/null +++ b/vendor/sha2/src/sha256/loongarch64_asm.rs @@ -0,0 +1,227 @@ +//! LoongArch64 assembly backend + +macro_rules! c { + ($($l:expr)*) => { + concat!($($l ,)*) + }; +} + +macro_rules! rounda { + ($i:literal, $a:literal, $b:literal, $c:literal, $d:literal, $e:literal, $f:literal, $g:literal, $h:literal) => { + c!( + "ld.w $a5, $a1, (" $i " * 4);" + "revb.2h $a5, $a5;" + "rotri.w $a5, $a5, 16;" + roundtail!($i, $a, $b, $c, $d, $e, $f, $g, $h) + ) + }; +} + +macro_rules! roundb { + ($i:literal, $a:literal, $b:literal, $c:literal, $d:literal, $e:literal, $f:literal, $g:literal, $h:literal) => { + c!( + "ld.w $a4, $sp, (((" $i " - 15) & 0xF) * 4);" + "ld.w $a5, $sp, (((" $i " - 16) & 0xF) * 4);" + "ld.w $a6, $sp, (((" $i " - 7) & 0xF) * 4);" + "add.w $a5, $a5, $a6;" + "rotri.w $a6, $a4, 18;" + "srli.w $a7, $a4, 3;" + "rotri.w $a4, $a4, 7;" + "xor $a6, $a6, $a7;" + "xor $a4, $a4, $a6;" + "add.w $a5, $a5, $a4;" + "ld.w $a4, $sp, (((" $i " - 2) & 0xF) * 4);" + "rotri.w $a6, $a4, 19;" + "srli.w $a7, $a4, 10;" + "rotri.w $a4, $a4, 17;" + "xor $a6, $a6, $a7;" + "xor $a4, $a4, $a6;" + "add.w $a5, $a5, $a4;" + roundtail!($i, $a, $b, $c, $d, $e, $f, $g, $h) + ) + }; +} + +macro_rules! roundtail { + ($i:literal, $a:literal, $b:literal, $c:literal, $d:literal, $e:literal, $f:literal, $g:literal, $h:literal) => { + c!( + // Part 0 + "rotri.w $a6, " $e ", 11;" + "rotri.w $a7, " $e ", 25;" + "rotri.w $a4, " $e ", 6;" + "xor $a6, $a6, $a7;" + "xor $a4, $a4, $a6;" + "xor $a6, " $g ", " $f ";" + "ld.w $a7, $a3, " $i " * 4;" + "and $a6, $a6, " $e ";" + "xor $a6, $a6, " $g ";" + "add.w $a4, $a4, $a6;" + "add.w $a4, $a4, $a7;" + "add.w " $h ", " $h ", $a5;" + "add.w " $h ", " $h ", $a4;" + // Part 1 + "add.w " $d ", " $d ", " $h ";" + // Part 2 + "rotri.w $a6, " $a ", 13;" + "rotri.w $a7, " $a ", 22;" + "rotri.w $a4, " $a ", 2;" + "xor $a6, $a6, $a7;" + "xor $a4, $a4, $a6;" + "add.w " $h ", " $h ", $a4;" + "or $a4, " $c ", " $b ";" + "and $a6, " $c ", " $b ";" + "and $a4, $a4, " $a ";" + "or $a4, $a4, $a6;" + "add.w " $h ", " $h ", $a4;" + "st.w $a5, $sp, ((" $i " & 0xF) * 4);" + ) + }; +} + +pub fn compress(state: &mut [u32; 8], blocks: &[[u8; 64]]) { + if blocks.is_empty() { + return; + } + + unsafe { + core::arch::asm!( + // Allocate scratch stack space + "addi.d $sp, $sp, -64;", + + // Load state + "ld.w $t0, $a0, 0", + "ld.w $t1, $a0, 4", + "ld.w $t2, $a0, 8", + "ld.w $t3, $a0, 12", + "ld.w $t4, $a0, 16", + "ld.w $t5, $a0, 20", + "ld.w $t6, $a0, 24", + "ld.w $t7, $a0, 28", + + "42:", + + // Do 64 rounds of hashing + rounda!( 0, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + rounda!( 1, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + rounda!( 2, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + rounda!( 3, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + rounda!( 4, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + rounda!( 5, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + rounda!( 6, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + rounda!( 7, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + rounda!( 8, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + rounda!( 9, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + rounda!(10, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + rounda!(11, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + rounda!(12, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + rounda!(13, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + rounda!(14, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + rounda!(15, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(16, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(17, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(18, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(19, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(20, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(21, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(22, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(23, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(24, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(25, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(26, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(27, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(28, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(29, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(30, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(31, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(32, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(33, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(34, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(35, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(36, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(37, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(38, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(39, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(40, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(41, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(42, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(43, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(44, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(45, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(46, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(47, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(48, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(49, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(50, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(51, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(52, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(53, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(54, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(55, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(56, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(57, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(58, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(59, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(60, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(61, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(62, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(63, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + + // Update state registers + "ld.w $a4, $a0, 0", // a + "ld.w $a5, $a0, 4", // b + "ld.w $a6, $a0, 8", // c + "ld.w $a7, $a0, 12", // d + "add.w $t0, $t0, $a4", + "add.w $t1, $t1, $a5", + "add.w $t2, $t2, $a6", + "add.w $t3, $t3, $a7", + "ld.w $a4, $a0, 16", // e + "ld.w $a5, $a0, 20", // f + "ld.w $a6, $a0, 24", // g + "ld.w $a7, $a0, 28", // h + "add.w $t4, $t4, $a4", + "add.w $t5, $t5, $a5", + "add.w $t6, $t6, $a6", + "add.w $t7, $t7, $a7", + + // Save updated state + "st.w $t0, $a0, 0", + "st.w $t1, $a0, 4", + "st.w $t2, $a0, 8", + "st.w $t3, $a0, 12", + "st.w $t4, $a0, 16", + "st.w $t5, $a0, 20", + "st.w $t6, $a0, 24", + "st.w $t7, $a0, 28", + + // Looping over blocks + "addi.d $a1, $a1, 64", + "addi.d $a2, $a2, -1", + "bnez $a2, 42b", + + // Restore stack register + "addi.d $sp, $sp, 64", + + in("$a0") state, + inout("$a1") blocks.as_ptr() => _, + inout("$a2") blocks.len() => _, + in("$a3") crate::consts::K32.as_ptr(), + + // Clobbers + out("$a4") _, + out("$a5") _, + out("$a6") _, + out("$a7") _, + out("$t0") _, + out("$t1") _, + out("$t2") _, + out("$t3") _, + out("$t4") _, + out("$t5") _, + out("$t6") _, + out("$t7") _, + + options(preserves_flags), + ); + } +} diff --git a/vendor/sha2/src/sha512.rs b/vendor/sha2/src/sha512.rs index af4178c0b..dfe0b454f 100644 --- a/vendor/sha2/src/sha512.rs +++ b/vendor/sha2/src/sha512.rs @@ -19,6 +19,9 @@ cfg_if::cfg_if! { mod soft; mod aarch64; use aarch64::compress; + } else if #[cfg(all(feature = "loongarch64_asm", target_arch = "loongarch64"))] { + mod loongarch64_asm; + use loongarch64_asm::compress; } else { mod soft; use soft::compress; diff --git a/vendor/sha2/src/sha512/loongarch64_asm.rs b/vendor/sha2/src/sha512/loongarch64_asm.rs new file mode 100644 index 000000000..557089def --- /dev/null +++ b/vendor/sha2/src/sha512/loongarch64_asm.rs @@ -0,0 +1,242 @@ +//! LoongArch64 assembly backend + +macro_rules! c { + ($($l:expr)*) => { + concat!($($l ,)*) + }; +} + +macro_rules! rounda { + ($i:literal, $a:literal, $b:literal, $c:literal, $d:literal, $e:literal, $f:literal, $g:literal, $h:literal) => { + c!( + "ld.d $a5, $a1, (" $i " * 8);" + "revb.d $a5, $a5;" + roundtail!($i, $a, $b, $c, $d, $e, $f, $g, $h) + ) + }; +} + +macro_rules! roundb { + ($i:literal, $a:literal, $b:literal, $c:literal, $d:literal, $e:literal, $f:literal, $g:literal, $h:literal) => { + c!( + "ld.d $a4, $sp, (((" $i " - 15) & 0xF) * 8);" + "ld.d $a5, $sp, (((" $i " - 16) & 0xF) * 8);" + "ld.d $a6, $sp, (((" $i " - 7) & 0xF) * 8);" + "add.d $a5, $a5, $a6;" + "rotri.d $a6, $a4, 8;" + "srli.d $a7, $a4, 7;" + "rotri.d $a4, $a4, 1;" + "xor $a6, $a6, $a7;" + "xor $a4, $a4, $a6;" + "add.d $a5, $a5, $a4;" + "ld.d $a4, $sp, (((" $i " - 2) & 0xF) * 8);" + "rotri.d $a6, $a4, 61;" + "srli.d $a7, $a4, 6;" + "rotri.d $a4, $a4, 19;" + "xor $a6, $a6, $a7;" + "xor $a4, $a4, $a6;" + "add.d $a5, $a5, $a4;" + roundtail!($i, $a, $b, $c, $d, $e, $f, $g, $h) + ) + }; +} + +macro_rules! roundtail { + ($i:literal, $a:literal, $b:literal, $c:literal, $d:literal, $e:literal, $f:literal, $g:literal, $h:literal) => { + c!( + // Part 0 + "rotri.d $a6, " $e ", 18;" + "rotri.d $a7, " $e ", 41;" + "rotri.d $a4, " $e ", 14;" + "xor $a6, $a6, $a7;" + "xor $a4, $a4, $a6;" + "xor $a6, " $g ", " $f ";" + "ld.d $a7, $a3, " $i " * 8;" + "and $a6, $a6, " $e ";" + "xor $a6, $a6, " $g ";" + "add.d $a4, $a4, $a6;" + "add.d $a4, $a4, $a7;" + "add.d " $h ", " $h ", $a5;" + "add.d " $h ", " $h ", $a4;" + // Part 1 + "add.d " $d ", " $d ", " $h ";" + // Part 2 + "rotri.d $a6, " $a ", 39;" + "rotri.d $a7, " $a ", 34;" + "rotri.d $a4, " $a ", 28;" + "xor $a6, $a6, $a7;" + "xor $a4, $a4, $a6;" + "add.d " $h ", " $h ", $a4;" + "or $a4, " $c ", " $b ";" + "and $a6, " $c ", " $b ";" + "and $a4, $a4, " $a ";" + "or $a4, $a4, $a6;" + "add.d " $h ", " $h ", $a4;" + "st.d $a5, $sp, ((" $i " & 0xF) * 8);" + ) + }; +} + +pub fn compress(state: &mut [u64; 8], blocks: &[[u8; 128]]) { + if blocks.is_empty() { + return; + } + + unsafe { + core::arch::asm!( + // Allocate scratch stack space + "addi.d $sp, $sp, -128;", + + // Load state + "ld.d $t0, $a0, 0", + "ld.d $t1, $a0, 8", + "ld.d $t2, $a0, 16", + "ld.d $t3, $a0, 24", + "ld.d $t4, $a0, 32", + "ld.d $t5, $a0, 40", + "ld.d $t6, $a0, 48", + "ld.d $t7, $a0, 56", + + "42:", + + // Do 64 rounds of hashing + rounda!( 0, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + rounda!( 1, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + rounda!( 2, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + rounda!( 3, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + rounda!( 4, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + rounda!( 5, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + rounda!( 6, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + rounda!( 7, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + rounda!( 8, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + rounda!( 9, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + rounda!(10, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + rounda!(11, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + rounda!(12, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + rounda!(13, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + rounda!(14, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + rounda!(15, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(16, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(17, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(18, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(19, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(20, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(21, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(22, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(23, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(24, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(25, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(26, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(27, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(28, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(29, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(30, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(31, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(32, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(33, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(34, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(35, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(36, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(37, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(38, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(39, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(40, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(41, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(42, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(43, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(44, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(45, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(46, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(47, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(48, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(49, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(50, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(51, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(52, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(53, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(54, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(55, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(56, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(57, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(58, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(59, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(60, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(61, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(62, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(63, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(64, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(65, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(66, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(67, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(68, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(69, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(70, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(71, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + roundb!(72, "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7"), + roundb!(73, "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6"), + roundb!(74, "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4", "$t5"), + roundb!(75, "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3", "$t4"), + roundb!(76, "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2", "$t3"), + roundb!(77, "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1" , "$t2"), + roundb!(78, "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0" , "$t1"), + roundb!(79, "$t1" , "$t2", "$t3", "$t4", "$t5", "$t6", "$t7", "$t0"), + + // Update state registers + "ld.d $a4, $a0, 0", // a + "ld.d $a5, $a0, 8", // b + "ld.d $a6, $a0, 16", // c + "ld.d $a7, $a0, 24", // d + "add.d $t0, $t0, $a4", + "add.d $t1, $t1, $a5", + "add.d $t2, $t2, $a6", + "add.d $t3, $t3, $a7", + "ld.d $a4, $a0, 32", // e + "ld.d $a5, $a0, 40", // f + "ld.d $a6, $a0, 48", // g + "ld.d $a7, $a0, 56", // h + "add.d $t4, $t4, $a4", + "add.d $t5, $t5, $a5", + "add.d $t6, $t6, $a6", + "add.d $t7, $t7, $a7", + + // Save updated state + "st.d $t0, $a0, 0", + "st.d $t1, $a0, 8", + "st.d $t2, $a0, 16", + "st.d $t3, $a0, 24", + "st.d $t4, $a0, 32", + "st.d $t5, $a0, 40", + "st.d $t6, $a0, 48", + "st.d $t7, $a0, 56", + + // Looping over blocks + "addi.d $a1, $a1, 128", + "addi.d $a2, $a2, -1", + "bnez $a2, 42b", + + // Restore stack register + "addi.d $sp, $sp, 128", + + in("$a0") state, + inout("$a1") blocks.as_ptr() => _, + inout("$a2") blocks.len() => _, + in("$a3") crate::consts::K64.as_ptr(), + + // Clobbers + out("$a4") _, + out("$a5") _, + out("$a6") _, + out("$a7") _, + out("$t0") _, + out("$t1") _, + out("$t2") _, + out("$t3") _, + out("$t4") _, + out("$t5") _, + out("$t6") _, + out("$t7") _, + + options(preserves_flags), + ); + } +} |