Skip to content

Commit

Permalink
gcm: Split FFI helpers into a separate submodule.
Browse files Browse the repository at this point in the history
Make it clearer which types are used in the FFI and start separating
the (unsafe) code that's used only for the FFI from the (safe) code
used in the Rust implementation.

`git difftool HEAD^1:src/aead/gcm.rs src/aead/gcm/ffi.rs`
  • Loading branch information
briansmith committed Jun 15, 2024
1 parent b454dfd commit 8d748be
Show file tree
Hide file tree
Showing 3 changed files with 224 additions and 163 deletions.
189 changes: 32 additions & 157 deletions src/aead/gcm.rs
Original file line number Diff line number Diff line change
Expand Up @@ -12,17 +12,25 @@
// OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
// CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.

use self::ffi::{Block, BLOCK_LEN, ZERO_BLOCK};
use super::{aes_gcm, Aad};
use crate::{
bits::{BitLength, FromByteLen as _},
constant_time, cpu, error,
cpu, error,
polyfill::{sliceutil::overwrite_at_start, ArraySplitMap as _},
};
use core::ops::BitXorAssign;
use cfg_if::cfg_if;

// GCM uses the same block type as AES.
use super::aes::{Block, BLOCK_LEN, ZERO_BLOCK};
cfg_if! {
if #[cfg(any(target_arch = "aarch64", target_arch = "x86_64"))] {
pub(super) use self::ffi::{HTable, Xi};
} else {
use self::ffi::{HTable, Xi};
}
}

#[macro_use]
mod ffi;
mod gcm_nohw;

#[derive(Clone)]
Expand All @@ -33,118 +41,21 @@ pub struct Key {
impl Key {
pub(super) fn new(h_be: Block, cpu_features: cpu::Features) -> Self {
let h: [u64; 2] = h_be.array_split_map(u64::from_be_bytes);

let mut key = Self {
h_table: HTable {
Htable: [U128 { hi: 0, lo: 0 }; HTABLE_LEN],
},
};
let h_table = &mut key.h_table;

match detect_implementation(cpu_features) {
let h_table = match detect_implementation(cpu_features) {
#[cfg(target_arch = "x86_64")]
Implementation::CLMUL if has_avx_movbe(cpu_features) => {
prefixed_extern! {
fn gcm_init_avx(HTable: &mut HTable, h: &[u64; 2]);
}
unsafe {
gcm_init_avx(h_table, &h);
}
}
Implementation::CLMUL if has_avx_movbe(cpu_features) => unsafe {
htable_new!(gcm_init_avx, &h, cou_features)
},

#[cfg(any(target_arch = "aarch64", target_arch = "x86_64", target_arch = "x86"))]
Implementation::CLMUL => {
prefixed_extern! {
fn gcm_init_clmul(Htable: &mut HTable, h: &[u64; 2]);
}
unsafe {
gcm_init_clmul(h_table, &h);
}
}
Implementation::CLMUL => unsafe { htable_new!(gcm_init_clmul, &h, cpu_features) },

#[cfg(any(target_arch = "aarch64", target_arch = "arm"))]
Implementation::NEON => {
prefixed_extern! {
fn gcm_init_neon(Htable: &mut HTable, h: &[u64; 2]);
}
unsafe {
gcm_init_neon(h_table, &h);
}
}

Implementation::Fallback => {
h_table.Htable[0] = gcm_nohw::init(h);
}
}

key
}
}

/// SAFETY:
/// * The function `$name` must meet the contract of the `f` paramweter of
/// `ghash()`.
#[cfg(any(
target_arch = "aarch64",
target_arch = "arm",
target_arch = "x86",
target_arch = "x86_64"
))]
macro_rules! ghash {
( $name:ident, $xi:expr, $h_table:expr, $input:expr, $cpu_features:expr ) => {{
prefixed_extern! {
fn $name(
xi: &mut Xi,
Htable: &HTable,
inp: *const u8,
len: crate::c::NonZero_size_t,
);
}
ghash($name, $xi, $h_table, $input, $cpu_features);
}};
}

/// SAFETY:
/// * `f` must read `len` bytes from `inp`; it may assume
/// that `len` is a (non-zero) multiple of `BLOCK_LEN`.
/// * `f` may inspect CPU features.
#[cfg(any(
target_arch = "aarch64",
target_arch = "arm",
target_arch = "x86",
target_arch = "x86_64"
))]
unsafe fn ghash(
f: unsafe extern "C" fn(
xi: &mut Xi,
Htable: &HTable,
inp: *const u8,
len: crate::c::NonZero_size_t,
),
xi: &mut Xi,
h_table: &HTable,
input: &[[u8; BLOCK_LEN]],
cpu_features: cpu::Features,
) {
use crate::polyfill::slice;
use core::num::NonZeroUsize;

let input = slice::flatten(input);
Implementation::NEON => unsafe { htable_new!(gcm_init_neon, &h, cpu_features) },

let input_len = match NonZeroUsize::new(input.len()) {
Some(len) => len,
None => {
return;
}
};

let _: cpu::Features = cpu_features;
// SAFETY:
// * There are `input_len: NonZeroUsize` bytes available at `input` for
// `f` to read.
// * CPU feature detection has been done.
unsafe {
f(xi, h_table, input.as_ptr(), input_len);
Implementation::Fallback => HTable::new_single_entry(gcm_nohw::init(h)),
};
Self { h_table }
}
}

Expand Down Expand Up @@ -209,7 +120,7 @@ impl<'key> Context<'key> {

pub fn update_blocks(&mut self, input: &[[u8; BLOCK_LEN]]) {
let xi = &mut self.Xi;
let h_table = &self.h_table;
let h_table = self.h_table;

match detect_implementation(self.cpu_features) {
#[cfg(target_arch = "x86_64")]
Expand Down Expand Up @@ -240,7 +151,7 @@ impl<'key> Context<'key> {
},

Implementation::Fallback => {
gcm_nohw::ghash(xi, h_table.Htable[0], input);
gcm_nohw::ghash(xi, h_table.first_entry(), input);
}
}
}
Expand All @@ -257,31 +168,21 @@ impl<'key> Context<'key> {
self.Xi.bitxor_assign(a);

let xi = &mut self.Xi;
let h_table = &self.h_table;
let h_table = self.h_table;

match detect_implementation(self.cpu_features) {
#[cfg(any(target_arch = "aarch64", target_arch = "x86_64", target_arch = "x86"))]
Implementation::CLMUL => {
prefixed_extern! {
fn gcm_gmult_clmul(xi: &mut Xi, Htable: &HTable);
}
unsafe {
gcm_gmult_clmul(xi, h_table);
}
}
Implementation::CLMUL => unsafe {
gmult!(gcm_gmult_clmul, xi, h_table, self.cpu_features)
},

#[cfg(any(target_arch = "aarch64", target_arch = "arm"))]
Implementation::NEON => {
prefixed_extern! {
fn gcm_gmult_neon(xi: &mut Xi, Htable: &HTable);
}
unsafe {
gcm_gmult_neon(xi, h_table);
}
}
Implementation::NEON => unsafe {
gmult!(gcm_gmult_neon, xi, h_table, self.cpu_features)
},

Implementation::Fallback => {
gcm_nohw::gmult(xi, h_table.Htable[0]);
gcm_nohw::gmult(xi, h_table.first_entry());
}
}
}
Expand All @@ -295,7 +196,7 @@ impl<'key> Context<'key> {
alen.copy_from_slice(&BitLength::<u64>::to_be_bytes(self.aad_len));
clen.copy_from_slice(&BitLength::<u64>::to_be_bytes(self.in_out_len));
self.update_block(block);
f(self.Xi.0, self.cpu_features)
f(self.Xi.into_block(), self.cpu_features)
}

#[cfg(target_arch = "x86_64")]
Expand All @@ -315,32 +216,6 @@ impl<'key> Context<'key> {
}
}

// The alignment is required by some assembly code.
#[derive(Clone)]
#[repr(C, align(16))]
pub(super) struct HTable {
Htable: [U128; HTABLE_LEN],
}

#[derive(Clone, Copy)]
#[repr(C)]
struct U128 {
hi: u64,
lo: u64,
}

const HTABLE_LEN: usize = 16;

#[repr(transparent)]
pub struct Xi(Block);

impl BitXorAssign<Block> for Xi {
#[inline]
fn bitxor_assign(&mut self, a: Block) {
self.0 = constant_time::xor_16(self.0, a)
}
}

#[allow(clippy::upper_case_acronyms)]
enum Implementation {
#[cfg(any(target_arch = "aarch64", target_arch = "x86_64", target_arch = "x86"))]
Expand Down
Loading

0 comments on commit 8d748be

Please sign in to comment.