UPSTREAM: crypto: x86/poly1305 - unify Poly1305 state struct with generic code
In preparation of exposing a Poly1305 library interface directly from
the accelerated x86 driver, align the state descriptor of the x86 code
with the one used by the generic driver. This is needed to make the
library interface unified between all implementations.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
(cherry picked from commit ad8f5b88383ea685f2b8df2a12ee3e08089a1287)
Bug: 152722841
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: I4425d0b514eec8e1d3e88e9972f270cc17bf8064
diff --git a/arch/x86/crypto/poly1305_glue.c b/arch/x86/crypto/poly1305_glue.c
index d2a8116..1d13712 100644
--- a/arch/x86/crypto/poly1305_glue.c
+++ b/arch/x86/crypto/poly1305_glue.c
@@ -18,40 +18,14 @@
#include <asm/fpu/api.h>
#include <asm/simd.h>
-struct poly1305_simd_desc_ctx {
- struct poly1305_desc_ctx base;
- /* derived key u set? */
- bool uset;
-#ifdef CONFIG_AS_AVX2
- /* derived keys r^3, r^4 set? */
- bool wset;
-#endif
- /* derived Poly1305 key r^2 */
- u32 u[5];
- /* ... silently appended r^3 and r^4 when using AVX2 */
-};
-
asmlinkage void poly1305_block_sse2(u32 *h, const u8 *src,
const u32 *r, unsigned int blocks);
asmlinkage void poly1305_2block_sse2(u32 *h, const u8 *src, const u32 *r,
unsigned int blocks, const u32 *u);
-#ifdef CONFIG_AS_AVX2
asmlinkage void poly1305_4block_avx2(u32 *h, const u8 *src, const u32 *r,
unsigned int blocks, const u32 *u);
-static bool poly1305_use_avx2;
-#endif
-static int poly1305_simd_init(struct shash_desc *desc)
-{
- struct poly1305_simd_desc_ctx *sctx = shash_desc_ctx(desc);
-
- sctx->uset = false;
-#ifdef CONFIG_AS_AVX2
- sctx->wset = false;
-#endif
-
- return crypto_poly1305_init(desc);
-}
+static bool poly1305_use_avx2 __ro_after_init;
static void poly1305_simd_mult(u32 *a, const u32 *b)
{
@@ -67,53 +41,49 @@
static unsigned int poly1305_simd_blocks(struct poly1305_desc_ctx *dctx,
const u8 *src, unsigned int srclen)
{
- struct poly1305_simd_desc_ctx *sctx;
unsigned int blocks, datalen;
- BUILD_BUG_ON(offsetof(struct poly1305_simd_desc_ctx, base));
- sctx = container_of(dctx, struct poly1305_simd_desc_ctx, base);
-
if (unlikely(!dctx->sset)) {
datalen = crypto_poly1305_setdesckey(dctx, src, srclen);
src += srclen - datalen;
srclen = datalen;
}
-#ifdef CONFIG_AS_AVX2
- if (poly1305_use_avx2 && srclen >= POLY1305_BLOCK_SIZE * 4) {
- if (unlikely(!sctx->wset)) {
- if (!sctx->uset) {
- memcpy(sctx->u, dctx->r.r, sizeof(sctx->u));
- poly1305_simd_mult(sctx->u, dctx->r.r);
- sctx->uset = true;
+ if (IS_ENABLED(CONFIG_AS_AVX2) &&
+ poly1305_use_avx2 &&
+ srclen >= POLY1305_BLOCK_SIZE * 4) {
+ if (unlikely(dctx->rset < 4)) {
+ if (dctx->rset < 2) {
+ dctx->r[1] = dctx->r[0];
+ poly1305_simd_mult(dctx->r[1].r, dctx->r[0].r);
}
- memcpy(sctx->u + 5, sctx->u, sizeof(sctx->u));
- poly1305_simd_mult(sctx->u + 5, dctx->r.r);
- memcpy(sctx->u + 10, sctx->u + 5, sizeof(sctx->u));
- poly1305_simd_mult(sctx->u + 10, dctx->r.r);
- sctx->wset = true;
+ dctx->r[2] = dctx->r[1];
+ poly1305_simd_mult(dctx->r[2].r, dctx->r[0].r);
+ dctx->r[3] = dctx->r[2];
+ poly1305_simd_mult(dctx->r[3].r, dctx->r[0].r);
+ dctx->rset = 4;
}
blocks = srclen / (POLY1305_BLOCK_SIZE * 4);
- poly1305_4block_avx2(dctx->h.h, src, dctx->r.r, blocks,
- sctx->u);
+ poly1305_4block_avx2(dctx->h.h, src, dctx->r[0].r, blocks,
+ dctx->r[1].r);
src += POLY1305_BLOCK_SIZE * 4 * blocks;
srclen -= POLY1305_BLOCK_SIZE * 4 * blocks;
}
-#endif
+
if (likely(srclen >= POLY1305_BLOCK_SIZE * 2)) {
- if (unlikely(!sctx->uset)) {
- memcpy(sctx->u, dctx->r.r, sizeof(sctx->u));
- poly1305_simd_mult(sctx->u, dctx->r.r);
- sctx->uset = true;
+ if (unlikely(dctx->rset < 2)) {
+ dctx->r[1] = dctx->r[0];
+ poly1305_simd_mult(dctx->r[1].r, dctx->r[0].r);
+ dctx->rset = 2;
}
blocks = srclen / (POLY1305_BLOCK_SIZE * 2);
- poly1305_2block_sse2(dctx->h.h, src, dctx->r.r, blocks,
- sctx->u);
+ poly1305_2block_sse2(dctx->h.h, src, dctx->r[0].r,
+ blocks, dctx->r[1].r);
src += POLY1305_BLOCK_SIZE * 2 * blocks;
srclen -= POLY1305_BLOCK_SIZE * 2 * blocks;
}
if (srclen >= POLY1305_BLOCK_SIZE) {
- poly1305_block_sse2(dctx->h.h, src, dctx->r.r, 1);
+ poly1305_block_sse2(dctx->h.h, src, dctx->r[0].r, 1);
srclen -= POLY1305_BLOCK_SIZE;
}
return srclen;
@@ -163,10 +133,10 @@
static struct shash_alg alg = {
.digestsize = POLY1305_DIGEST_SIZE,
- .init = poly1305_simd_init,
+ .init = crypto_poly1305_init,
.update = poly1305_simd_update,
.final = crypto_poly1305_final,
- .descsize = sizeof(struct poly1305_simd_desc_ctx),
+ .descsize = sizeof(struct poly1305_desc_ctx),
.base = {
.cra_name = "poly1305",
.cra_driver_name = "poly1305-simd",
@@ -181,14 +151,14 @@
if (!boot_cpu_has(X86_FEATURE_XMM2))
return -ENODEV;
-#ifdef CONFIG_AS_AVX2
- poly1305_use_avx2 = boot_cpu_has(X86_FEATURE_AVX) &&
+ poly1305_use_avx2 = IS_ENABLED(CONFIG_AS_AVX2) &&
+ boot_cpu_has(X86_FEATURE_AVX) &&
boot_cpu_has(X86_FEATURE_AVX2) &&
cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM, NULL);
- alg.descsize = sizeof(struct poly1305_simd_desc_ctx);
+ alg.descsize = sizeof(struct poly1305_desc_ctx) + 5 * sizeof(u32);
if (poly1305_use_avx2)
alg.descsize += 10 * sizeof(u32);
-#endif
+
return crypto_register_shash(&alg);
}
diff --git a/crypto/poly1305_generic.c b/crypto/poly1305_generic.c
index 857a993..6241c79 100644
--- a/crypto/poly1305_generic.c
+++ b/crypto/poly1305_generic.c
@@ -25,7 +25,7 @@
poly1305_core_init(&dctx->h);
dctx->buflen = 0;
- dctx->rset = false;
+ dctx->rset = 0;
dctx->sset = false;
return 0;
@@ -43,7 +43,7 @@
srclen = datalen;
}
- poly1305_core_blocks(&dctx->h, &dctx->r, src,
+ poly1305_core_blocks(&dctx->h, dctx->r, src,
srclen / POLY1305_BLOCK_SIZE, 1);
}
@@ -95,7 +95,7 @@
dctx->buf[dctx->buflen++] = 1;
memset(dctx->buf + dctx->buflen, 0,
POLY1305_BLOCK_SIZE - dctx->buflen);
- poly1305_core_blocks(&dctx->h, &dctx->r, dctx->buf, 1, 0);
+ poly1305_core_blocks(&dctx->h, dctx->r, dctx->buf, 1, 0);
}
poly1305_core_emit(&dctx->h, digest);
diff --git a/include/crypto/internal/poly1305.h b/include/crypto/internal/poly1305.h
index cb58e61..04fa269 100644
--- a/include/crypto/internal/poly1305.h
+++ b/include/crypto/internal/poly1305.h
@@ -46,10 +46,10 @@
{
if (!dctx->sset) {
if (!dctx->rset && srclen >= POLY1305_BLOCK_SIZE) {
- poly1305_core_setkey(&dctx->r, src);
+ poly1305_core_setkey(dctx->r, src);
src += POLY1305_BLOCK_SIZE;
srclen -= POLY1305_BLOCK_SIZE;
- dctx->rset = true;
+ dctx->rset = 1;
}
if (srclen >= POLY1305_BLOCK_SIZE) {
dctx->s[0] = get_unaligned_le32(src + 0);
diff --git a/include/crypto/poly1305.h b/include/crypto/poly1305.h
index f5a4319..36b5886 100644
--- a/include/crypto/poly1305.h
+++ b/include/crypto/poly1305.h
@@ -22,20 +22,20 @@
};
struct poly1305_desc_ctx {
- /* key */
- struct poly1305_key r;
- /* finalize key */
- u32 s[4];
- /* accumulator */
- struct poly1305_state h;
/* partial buffer */
u8 buf[POLY1305_BLOCK_SIZE];
/* bytes used in partial buffer */
unsigned int buflen;
- /* r key has been set */
- bool rset;
- /* s key has been set */
+ /* how many keys have been set in r[] */
+ unsigned short rset;
+ /* whether s[] has been set */
bool sset;
+ /* finalize key */
+ u32 s[4];
+ /* accumulator */
+ struct poly1305_state h;
+ /* key */
+ struct poly1305_key r[1];
};
#endif