crypto: x86/camellia-aesni-avx - add more optimized XTS code

Add more optimized XTS code for camellia-aesni-avx, for smaller stack usage
and small boost for speed.

tcrypt results, with Intel i5-2450M:
        enc     dec
16B     1.10x   1.01x
64B     0.82x   0.77x
256B    1.14x   1.10x
1024B   1.17x   1.16x
8192B   1.10x   1.11x

Since XTS is practically always used with data blocks of size 512 bytes or
more, I chose to not make use of camellia-2way for block sized smaller than
256 bytes. This causes slower result in tcrypt for 64 bytes.

Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

1 files changed, 50 insertions, 41 deletions

diff --git a/arch/x86/crypto/camellia_aesni_avx_glue.c b/arch/x86/crypto/camellia_aesni_avx_glue.c
index 96cbb6068fce..4ff7ed47b3db 100644
--- a/arch/x86/crypto/camellia_aesni_avx_glue.c
+++ b/arch/x86/crypto/camellia_aesni_avx_glue.c
@@ -1,7 +1,7 @@
 /*
  * Glue Code for x86_64/AVX/AES-NI assembler optimized version of Camellia
  *
- * Copyright © 2012 Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
+ * Copyright © 2012-2013 Jussi Kivilinna <jussi.kivilinna@iki.fi>
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -37,6 +37,23 @@ asmlinkage void camellia_cbc_dec_16way(struct camellia_ctx *ctx, u8 *dst,
 asmlinkage void camellia_ctr_16way(struct camellia_ctx *ctx, u8 *dst,
 				   const u8 *src, le128 *iv);
 
+asmlinkage void camellia_xts_enc_16way(struct camellia_ctx *ctx, u8 *dst,
+				       const u8 *src, le128 *iv);
+asmlinkage void camellia_xts_dec_16way(struct camellia_ctx *ctx, u8 *dst,
+				       const u8 *src, le128 *iv);
+
+static void camellia_xts_enc(void *ctx, u128 *dst, const u128 *src, le128 *iv)
+{
+	glue_xts_crypt_128bit_one(ctx, dst, src, iv,
+				  GLUE_FUNC_CAST(camellia_enc_blk));
+}
+
+static void camellia_xts_dec(void *ctx, u128 *dst, const u128 *src, le128 *iv)
+{
+	glue_xts_crypt_128bit_one(ctx, dst, src, iv,
+				  GLUE_FUNC_CAST(camellia_dec_blk));
+}
+
 static const struct common_glue_ctx camellia_enc = {
 	.num_funcs = 3,
 	.fpu_blocks_limit = CAMELLIA_AESNI_PARALLEL_BLOCKS,
@@ -69,6 +86,19 @@ static const struct common_glue_ctx camellia_ctr = {
 	} }
 };
 
+static const struct common_glue_ctx camellia_enc_xts = {
+	.num_funcs = 2,
+	.fpu_blocks_limit = CAMELLIA_AESNI_PARALLEL_BLOCKS,
+
+	.funcs = { {
+		.num_blocks = CAMELLIA_AESNI_PARALLEL_BLOCKS,
+		.fn_u = { .xts = GLUE_XTS_FUNC_CAST(camellia_xts_enc_16way) }
+	}, {
+		.num_blocks = 1,
+		.fn_u = { .xts = GLUE_XTS_FUNC_CAST(camellia_xts_enc) }
+	} }
+};
+
 static const struct common_glue_ctx camellia_dec = {
 	.num_funcs = 3,
 	.fpu_blocks_limit = CAMELLIA_AESNI_PARALLEL_BLOCKS,
@@ -101,6 +131,19 @@ static const struct common_glue_ctx camellia_dec_cbc = {
 	} }
 };
 
+static const struct common_glue_ctx camellia_dec_xts = {
+	.num_funcs = 2,
+	.fpu_blocks_limit = CAMELLIA_AESNI_PARALLEL_BLOCKS,
+
+	.funcs = { {
+		.num_blocks = CAMELLIA_AESNI_PARALLEL_BLOCKS,
+		.fn_u = { .xts = GLUE_XTS_FUNC_CAST(camellia_xts_dec_16way) }
+	}, {
+		.num_blocks = 1,
+		.fn_u = { .xts = GLUE_XTS_FUNC_CAST(camellia_xts_dec) }
+	} }
+};
+
 static int ecb_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
 		       struct scatterlist *src, unsigned int nbytes)
 {
@@ -261,54 +304,20 @@ static int xts_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
 		       struct scatterlist *src, unsigned int nbytes)
 {
 	struct camellia_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
-	be128 buf[CAMELLIA_AESNI_PARALLEL_BLOCKS];
-	struct crypt_priv crypt_ctx = {
-		.ctx = &ctx->crypt_ctx,
-		.fpu_enabled = false,
-	};
-	struct xts_crypt_req req = {
-		.tbuf = buf,
-		.tbuflen = sizeof(buf),
-
-		.tweak_ctx = &ctx->tweak_ctx,
-		.tweak_fn = XTS_TWEAK_CAST(camellia_enc_blk),
-		.crypt_ctx = &crypt_ctx,
-		.crypt_fn = encrypt_callback,
-	};
-	int ret;
-
-	desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
-	ret = xts_crypt(desc, dst, src, nbytes, &req);
-	camellia_fpu_end(crypt_ctx.fpu_enabled);
 
-	return ret;
+	return glue_xts_crypt_128bit(&camellia_enc_xts, desc, dst, src, nbytes,
+				     XTS_TWEAK_CAST(camellia_enc_blk),
+				     &ctx->tweak_ctx, &ctx->crypt_ctx);
 }
 
 static int xts_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
 		       struct scatterlist *src, unsigned int nbytes)
 {
 	struct camellia_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
-	be128 buf[CAMELLIA_AESNI_PARALLEL_BLOCKS];
-	struct crypt_priv crypt_ctx = {
-		.ctx = &ctx->crypt_ctx,
-		.fpu_enabled = false,
-	};
-	struct xts_crypt_req req = {
-		.tbuf = buf,
-		.tbuflen = sizeof(buf),
 
-		.tweak_ctx = &ctx->tweak_ctx,
-		.tweak_fn = XTS_TWEAK_CAST(camellia_enc_blk),
-		.crypt_ctx = &crypt_ctx,
-		.crypt_fn = decrypt_callback,
-	};
-	int ret;
-
-	desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
-	ret = xts_crypt(desc, dst, src, nbytes, &req);
-	camellia_fpu_end(crypt_ctx.fpu_enabled);
-
-	return ret;
+	return glue_xts_crypt_128bit(&camellia_dec_xts, desc, dst, src, nbytes,
+				     XTS_TWEAK_CAST(camellia_enc_blk),
+				     &ctx->tweak_ctx, &ctx->crypt_ctx);
 }
 
 static struct crypto_alg cmll_algs[10] = { {