crypto: blake2b - sync with blake2s implementation

Sync the BLAKE2b code with the BLAKE2s code as much as possible:

- Move a lot of code into new headers <crypto/blake2b.h> and
  <crypto/internal/blake2b.h>, and adjust it to be like the
  corresponding BLAKE2s code, i.e. like <crypto/blake2s.h> and
  <crypto/internal/blake2s.h>.

- Rename constants, e.g. BLAKE2B_*_DIGEST_SIZE => BLAKE2B_*_HASH_SIZE.

- Use a macro BLAKE2B_ALG() to define the shash_alg structs.

- Export blake2b_compress_generic() for use as a fallback.

This makes it much easier to add optimized implementations of BLAKE2b,
as optimized implementations can use the helper functions
crypto_blake2b_{setkey,init,update,final}() and
blake2b_compress_generic().  The ARM implementation will use these.

But this change is also helpful because it eliminates unnecessary
differences between the BLAKE2b and BLAKE2s code, so that the same
improvements can easily be made to both.  (The two algorithms are
basically identical, except for the word size and constants.)  It also
makes it straightforward to add a library API for BLAKE2b in the future
if/when it's needed.

This change does make the BLAKE2b code slightly more complicated than it
needs to be, as it doesn't actually provide a library API yet.  For
example, __blake2b_update() doesn't really need to exist yet; it could
just be inlined into crypto_blake2b_update().  But I believe this is
outweighed by the benefits of keeping the code in sync.

Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

2 files changed, 182 insertions, 0 deletions

diff --git a/include/crypto/blake2b.h b/include/crypto/blake2b.h
new file mode 100644
index 000000000000..18875f16f8ca
--- /dev/null
+++ b/include/crypto/blake2b.h
@@ -0,0 +1,67 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+
+#ifndef _CRYPTO_BLAKE2B_H
+#define _CRYPTO_BLAKE2B_H
+
+#include <linux/bug.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+
+enum blake2b_lengths {
+	BLAKE2B_BLOCK_SIZE = 128,
+	BLAKE2B_HASH_SIZE = 64,
+	BLAKE2B_KEY_SIZE = 64,
+
+	BLAKE2B_160_HASH_SIZE = 20,
+	BLAKE2B_256_HASH_SIZE = 32,
+	BLAKE2B_384_HASH_SIZE = 48,
+	BLAKE2B_512_HASH_SIZE = 64,
+};
+
+struct blake2b_state {
+	/* 'h', 't', and 'f' are used in assembly code, so keep them as-is. */
+	u64 h[8];
+	u64 t[2];
+	u64 f[2];
+	u8 buf[BLAKE2B_BLOCK_SIZE];
+	unsigned int buflen;
+	unsigned int outlen;
+};
+
+enum blake2b_iv {
+	BLAKE2B_IV0 = 0x6A09E667F3BCC908ULL,
+	BLAKE2B_IV1 = 0xBB67AE8584CAA73BULL,
+	BLAKE2B_IV2 = 0x3C6EF372FE94F82BULL,
+	BLAKE2B_IV3 = 0xA54FF53A5F1D36F1ULL,
+	BLAKE2B_IV4 = 0x510E527FADE682D1ULL,
+	BLAKE2B_IV5 = 0x9B05688C2B3E6C1FULL,
+	BLAKE2B_IV6 = 0x1F83D9ABFB41BD6BULL,
+	BLAKE2B_IV7 = 0x5BE0CD19137E2179ULL,
+};
+
+static inline void __blake2b_init(struct blake2b_state *state, size_t outlen,
+				  const void *key, size_t keylen)
+{
+	state->h[0] = BLAKE2B_IV0 ^ (0x01010000 | keylen << 8 | outlen);
+	state->h[1] = BLAKE2B_IV1;
+	state->h[2] = BLAKE2B_IV2;
+	state->h[3] = BLAKE2B_IV3;
+	state->h[4] = BLAKE2B_IV4;
+	state->h[5] = BLAKE2B_IV5;
+	state->h[6] = BLAKE2B_IV6;
+	state->h[7] = BLAKE2B_IV7;
+	state->t[0] = 0;
+	state->t[1] = 0;
+	state->f[0] = 0;
+	state->f[1] = 0;
+	state->buflen = 0;
+	state->outlen = outlen;
+	if (keylen) {
+		memcpy(state->buf, key, keylen);
+		memset(&state->buf[keylen], 0, BLAKE2B_BLOCK_SIZE - keylen);
+		state->buflen = BLAKE2B_BLOCK_SIZE;
+	}
+}
+
+#endif /* _CRYPTO_BLAKE2B_H */
diff --git a/include/crypto/internal/blake2b.h b/include/crypto/internal/blake2b.h
new file mode 100644
index 000000000000..982fe5e8471c
--- /dev/null
+++ b/include/crypto/internal/blake2b.h
@@ -0,0 +1,115 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+/*
+ * Helper functions for BLAKE2b implementations.
+ * Keep this in sync with the corresponding BLAKE2s header.
+ */
+
+#ifndef _CRYPTO_INTERNAL_BLAKE2B_H
+#define _CRYPTO_INTERNAL_BLAKE2B_H
+
+#include <crypto/blake2b.h>
+#include <crypto/internal/hash.h>
+#include <linux/string.h>
+
+void blake2b_compress_generic(struct blake2b_state *state,
+			      const u8 *block, size_t nblocks, u32 inc);
+
+static inline void blake2b_set_lastblock(struct blake2b_state *state)
+{
+	state->f[0] = -1;
+}
+
+typedef void (*blake2b_compress_t)(struct blake2b_state *state,
+				   const u8 *block, size_t nblocks, u32 inc);
+
+static inline void __blake2b_update(struct blake2b_state *state,
+				    const u8 *in, size_t inlen,
+				    blake2b_compress_t compress)
+{
+	const size_t fill = BLAKE2B_BLOCK_SIZE - state->buflen;
+
+	if (unlikely(!inlen))
+		return;
+	if (inlen > fill) {
+		memcpy(state->buf + state->buflen, in, fill);
+		(*compress)(state, state->buf, 1, BLAKE2B_BLOCK_SIZE);
+		state->buflen = 0;
+		in += fill;
+		inlen -= fill;
+	}
+	if (inlen > BLAKE2B_BLOCK_SIZE) {
+		const size_t nblocks = DIV_ROUND_UP(inlen, BLAKE2B_BLOCK_SIZE);
+		/* Hash one less (full) block than strictly possible */
+		(*compress)(state, in, nblocks - 1, BLAKE2B_BLOCK_SIZE);
+		in += BLAKE2B_BLOCK_SIZE * (nblocks - 1);
+		inlen -= BLAKE2B_BLOCK_SIZE * (nblocks - 1);
+	}
+	memcpy(state->buf + state->buflen, in, inlen);
+	state->buflen += inlen;
+}
+
+static inline void __blake2b_final(struct blake2b_state *state, u8 *out,
+				   blake2b_compress_t compress)
+{
+	int i;
+
+	blake2b_set_lastblock(state);
+	memset(state->buf + state->buflen, 0,
+	       BLAKE2B_BLOCK_SIZE - state->buflen); /* Padding */
+	(*compress)(state, state->buf, 1, state->buflen);
+	for (i = 0; i < ARRAY_SIZE(state->h); i++)
+		__cpu_to_le64s(&state->h[i]);
+	memcpy(out, state->h, state->outlen);
+}
+
+/* Helper functions for shash implementations of BLAKE2b */
+
+struct blake2b_tfm_ctx {
+	u8 key[BLAKE2B_KEY_SIZE];
+	unsigned int keylen;
+};
+
+static inline int crypto_blake2b_setkey(struct crypto_shash *tfm,
+					const u8 *key, unsigned int keylen)
+{
+	struct blake2b_tfm_ctx *tctx = crypto_shash_ctx(tfm);
+
+	if (keylen == 0 || keylen > BLAKE2B_KEY_SIZE)
+		return -EINVAL;
+
+	memcpy(tctx->key, key, keylen);
+	tctx->keylen = keylen;
+
+	return 0;
+}
+
+static inline int crypto_blake2b_init(struct shash_desc *desc)
+{
+	const struct blake2b_tfm_ctx *tctx = crypto_shash_ctx(desc->tfm);
+	struct blake2b_state *state = shash_desc_ctx(desc);
+	unsigned int outlen = crypto_shash_digestsize(desc->tfm);
+
+	__blake2b_init(state, outlen, tctx->key, tctx->keylen);
+	return 0;
+}
+
+static inline int crypto_blake2b_update(struct shash_desc *desc,
+					const u8 *in, unsigned int inlen,
+					blake2b_compress_t compress)
+{
+	struct blake2b_state *state = shash_desc_ctx(desc);
+
+	__blake2b_update(state, in, inlen, compress);
+	return 0;
+}
+
+static inline int crypto_blake2b_final(struct shash_desc *desc, u8 *out,
+				       blake2b_compress_t compress)
+{
+	struct blake2b_state *state = shash_desc_ctx(desc);
+
+	__blake2b_final(state, out, compress);
+	return 0;
+}
+
+#endif /* _CRYPTO_INTERNAL_BLAKE2B_H */