Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

chacha20: adding 64-bit counter support #359

Open
wants to merge 11 commits into
base: master
Choose a base branch
from
Open

chacha20: adding 64-bit counter support #359

Show file tree
Hide file tree
Changes from 7 commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
fcc8872
added 64-bit counter support for ChaCha20Legacy, and a test
nstilt1 May 30, 2024
aa22cb6
failing neon tests currently
nstilt1 May 30, 2024
14312d8
fixed neon code to pass 64-bit counter tests. Tested with multiple TE…
nstilt1 Jun 7, 2024
f984d8e
removed unused macro and unnecessary test
nstilt1 Jun 7, 2024
dedf194
adjusted legacy counter test to try different amounts of blocks
nstilt1 Jun 8, 2024
61acc54
added test to see if it would fail, and it did
nstilt1 Jun 8, 2024
66ae108
turned add64! into add_counter! that accepts a bool; now it passes tests
nstilt1 Jun 8, 2024
bef4ccd
removed Variant::USES_32_BIT_COUNTER; about to check neon code
nstilt1 Aug 2, 2024
1064554
fixed neon code
nstilt1 Aug 2, 2024
b71cf9f
Merge branch 'master' into counter_support_2.0
nstilt1 Aug 2, 2024
388017d
removed unnecessary qualifications
nstilt1 Aug 2, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
47 changes: 42 additions & 5 deletions Cargo.lock
View file
Open in desktop

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

51 changes: 37 additions & 14 deletions chacha20/src/backends/avx2.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,8 +1,8 @@
use crate::Rounds;
use crate::{Rounds, Variant};
use core::marker::PhantomData;

#[cfg(feature = "rng")]
use crate::{ChaChaCore, Variant};
use crate::ChaChaCore;

#[cfg(feature = "cipher")]
use crate::{
Expand Down Expand Up @@ -33,10 +33,11 @@ const N: usize = PAR_BLOCKS / 2;
#[inline]
#[target_feature(enable = "avx2")]
#[cfg(feature = "cipher")]
pub(crate) unsafe fn inner<R, F>(state: &mut [u32; STATE_WORDS], f: F)
pub(crate) unsafe fn inner<R, F, V>(state: &mut [u32; STATE_WORDS], f: F)
where
R: Rounds,
F: StreamClosure<BlockSize = U64>,
V: Variant
{
let state_ptr = state.as_ptr() as *const __m128i;
let v = [
Expand All @@ -45,21 +46,33 @@ where
_mm256_broadcastsi128_si256(_mm_loadu_si128(state_ptr.add(2))),
];
let mut c = _mm256_broadcastsi128_si256(_mm_loadu_si128(state_ptr.add(3)));
c = _mm256_add_epi32(c, _mm256_set_epi32(0, 0, 0, 1, 0, 0, 0, 0));
if V::USES_U32_COUNTER {
c = _mm256_add_epi32(c, _mm256_set_epi32(0, 0, 0, 1, 0, 0, 0, 0));
} else {
c = _mm256_add_epi64(c, _mm256_set_epi64x(0, 1, 0, 0));
}
let mut ctr = [c; N];
for i in 0..N {
ctr[i] = c;
c = _mm256_add_epi32(c, _mm256_set_epi32(0, 0, 0, 2, 0, 0, 0, 2));
if V::USES_U32_COUNTER {
c = _mm256_add_epi32(c, _mm256_set_epi32(0, 0, 0, 2, 0, 0, 0, 2));
} else {
c = _mm256_add_epi64(c, _mm256_set_epi64x(0, 2, 0, 2));
}
}
let mut backend = Backend::<R> {
let mut backend = Backend::<R, V> {
v,
ctr,
_pd: PhantomData,
_variant: PhantomData
};

f.call(&mut backend);

state[12] = _mm256_extract_epi32(backend.ctr[0], 0) as u32;
if !V::USES_U32_COUNTER {
state[13] = _mm256_extract_epi32(backend.ctr[0], 1) as u32;
}
}

#[inline]
Expand All @@ -83,41 +96,47 @@ where
ctr[i] = c;
c = _mm256_add_epi32(c, _mm256_set_epi32(0, 0, 0, 2, 0, 0, 0, 2));
}
let mut backend = Backend::<R> {
let mut backend = Backend::<R, V> {
v,
ctr,
_pd: PhantomData,
_variant: PhantomData
};

backend.rng_gen_par_ks_blocks(buffer);

core.state[12] = _mm256_extract_epi32(backend.ctr[0], 0) as u32;
}

struct Backend<R: Rounds> {
struct Backend<R: Rounds, V: Variant> {
v: [__m256i; 3],
ctr: [__m256i; N],
_pd: PhantomData<R>,
_variant: PhantomData<V>
}

#[cfg(feature = "cipher")]
impl<R: Rounds> BlockSizeUser for Backend<R> {
impl<R: Rounds, V: Variant> BlockSizeUser for Backend<R, V> {
type BlockSize = U64;
}

#[cfg(feature = "cipher")]
impl<R: Rounds> ParBlocksSizeUser for Backend<R> {
impl<R: Rounds, V: Variant> ParBlocksSizeUser for Backend<R, V> {
type ParBlocksSize = U4;
}

#[cfg(feature = "cipher")]
impl<R: Rounds> StreamBackend for Backend<R> {
impl<R: Rounds, V: Variant> StreamBackend for Backend<R, V> {
#[inline(always)]
fn gen_ks_block(&mut self, block: &mut Block) {
unsafe {
let res = rounds::<R>(&self.v, &self.ctr);
for c in self.ctr.iter_mut() {
*c = _mm256_add_epi32(*c, _mm256_set_epi32(0, 0, 0, 1, 0, 0, 0, 1));
if V::USES_U32_COUNTER {
*c = _mm256_add_epi32(*c, _mm256_set_epi32(0, 0, 0, 1, 0, 0, 0, 1));
} else {
*c = _mm256_add_epi64(*c, _mm256_set_epi64x(0, 1, 0, 1));
}
}

let res0: [__m128i; 8] = core::mem::transmute(res[0]);
Expand All @@ -136,7 +155,11 @@ impl<R: Rounds> StreamBackend for Backend<R> {

let pb = PAR_BLOCKS as i32;
for c in self.ctr.iter_mut() {
*c = _mm256_add_epi32(*c, _mm256_set_epi32(0, 0, 0, pb, 0, 0, 0, pb));
if V::USES_U32_COUNTER {
*c = _mm256_add_epi32(*c, _mm256_set_epi32(0, 0, 0, pb, 0, 0, 0, pb));
} else {
*c = _mm256_add_epi64(*c, _mm256_set_epi64x(0, pb as i64, 0, pb as i64));
}
}

let mut block_ptr = blocks.as_mut_ptr() as *mut __m128i;
Expand All @@ -153,7 +176,7 @@ impl<R: Rounds> StreamBackend for Backend<R> {
}

#[cfg(feature = "rng")]
impl<R: Rounds> Backend<R> {
impl<R: Rounds, V: Variant> Backend<R, V> {
#[inline(always)]
fn rng_gen_par_ks_blocks(&mut self, blocks: &mut [u32; 64]) {
unsafe {
Expand Down
Loading
Loading