[PATCH 0/7] crypto: OCTEON MD5 bugfix + SHA modules

[PATCH 0/7] crypto: OCTEON MD5 bugfix + SHA modules

[Aaro Koskinen]

Hi,

The first patch is a bug fix for OCTEON MD5 aimed for 4.0-rc cycle.

The remaining patches add SHA1/SHA256/SHA512 modules. I have tested
these on the following OCTEON boards and CPUs with 4.0-rc2:

	D-Link DSR-1000N:	CN5010p1.1-500-SCP
	EdgeRouter Lite:	CN5020p1.1-500-SCP
	EdgeRouter Pro:		CN6120p1.1-1000-NSP

All selftests are passing. With tcrypt, I get the following numbers
on speed compared to the generic modules:

	SHA1:

test  0 (   16 byte blocks,   16 bytes per update,   1 updates): 1.25x faster
test  1 (   64 byte blocks,   16 bytes per update,   4 updates): 1.20x faster
test  2 (   64 byte blocks,   64 bytes per update,   1 updates): 1.47x faster
test  3 (  256 byte blocks,   16 bytes per update,  16 updates): 1.15x faster
test  4 (  256 byte blocks,   64 bytes per update,   4 updates): 1.56x faster
test  5 (  256 byte blocks,  256 bytes per update,   1 updates): 2.27x faster
test  6 ( 1024 byte blocks,   16 bytes per update,  64 updates): 1.13x faster
test  7 ( 1024 byte blocks,  256 bytes per update,   4 updates): 2.74x faster
test  8 ( 1024 byte blocks, 1024 bytes per update,   1 updates): 3.60x faster
test  9 ( 2048 byte blocks,   16 bytes per update, 128 updates): 1.13x faster
test 10 ( 2048 byte blocks,  256 bytes per update,   8 updates): 2.87x faster
test 11 ( 2048 byte blocks, 1024 bytes per update,   2 updates): 3.90x faster
test 12 ( 2048 byte blocks, 2048 bytes per update,   1 updates): 4.18x faster
test 13 ( 4096 byte blocks,   16 bytes per update, 256 updates): 1.13x faster
test 14 ( 4096 byte blocks,  256 bytes per update,  16 updates): 2.95x faster
test 15 ( 4096 byte blocks, 1024 bytes per update,   4 updates): 4.09x faster
test 16 ( 4096 byte blocks, 4096 bytes per update,   1 updates): 4.57x faster
test 17 ( 8192 byte blocks,   16 bytes per update, 512 updates): 1.13x faster
test 18 ( 8192 byte blocks,  256 bytes per update,  32 updates): 2.99x faster
test 19 ( 8192 byte blocks, 1024 bytes per update,   8 updates): 4.20x faster
test 20 ( 8192 byte blocks, 4096 bytes per update,   2 updates): 4.72x faster
test 21 ( 8192 byte blocks, 8192 bytes per update,   1 updates): 4.73x faster

	SHA256:

test  0 (   16 byte blocks,   16 bytes per update,   1 updates): 2.72x faster
test  1 (   64 byte blocks,   16 bytes per update,   4 updates): 2.45x faster
test  2 (   64 byte blocks,   64 bytes per update,   1 updates): 3.65x faster
test  3 (  256 byte blocks,   16 bytes per update,  16 updates): 2.18x faster
test  4 (  256 byte blocks,   64 bytes per update,   4 updates): 3.74x faster
test  5 (  256 byte blocks,  256 bytes per update,   1 updates): 5.72x faster
test  6 ( 1024 byte blocks,   16 bytes per update,  64 updates): 2.08x faster
test  7 ( 1024 byte blocks,  256 bytes per update,   4 updates): 6.54x faster
test  8 ( 1024 byte blocks, 1024 bytes per update,   1 updates): 8.19x faster
test  9 ( 2048 byte blocks,   16 bytes per update, 128 updates): 2.06x faster
test 10 ( 2048 byte blocks,  256 bytes per update,   8 updates): 6.77x faster
test 11 ( 2048 byte blocks, 1024 bytes per update,   2 updates): 8.56x faster
test 12 ( 2048 byte blocks, 2048 bytes per update,   1 updates): 9.01x faster
test 13 ( 4096 byte blocks,   16 bytes per update, 256 updates): 2.05x faster
test 14 ( 4096 byte blocks,  256 bytes per update,  16 updates): 6.89x faster
test 15 ( 4096 byte blocks, 1024 bytes per update,   4 updates): 8.82x faster
test 16 ( 4096 byte blocks, 4096 bytes per update,   1 updates): 9.50x faster
test 17 ( 8192 byte blocks,   16 bytes per update, 512 updates): 2.04x faster
test 18 ( 8192 byte blocks,  256 bytes per update,  32 updates): 6.96x faster
test 19 ( 8192 byte blocks, 1024 bytes per update,   8 updates): 8.95x faster
test 20 ( 8192 byte blocks, 4096 bytes per update,   2 updates): 9.66x faster
test 21 ( 8192 byte blocks, 8192 bytes per update,   1 updates): 9.67x faster

	SHA512:

test  0 (   16 byte blocks,   16 bytes per update,   1 updates): 3.19x faster
test  1 (   64 byte blocks,   16 bytes per update,   4 updates): 2.18x faster
test  2 (   64 byte blocks,   64 bytes per update,   1 updates): 3.19x faster
test  3 (  256 byte blocks,   16 bytes per update,  16 updates): 2.12x faster
test  4 (  256 byte blocks,   64 bytes per update,   4 updates): 3.54x faster
test  5 (  256 byte blocks,  256 bytes per update,   1 updates): 5.16x faster
test  6 ( 1024 byte blocks,   16 bytes per update,  64 updates): 1.92x faster
test  7 ( 1024 byte blocks,  256 bytes per update,   4 updates): 5.80x faster
test  8 ( 1024 byte blocks, 1024 bytes per update,   1 updates): 8.07x faster
test  9 ( 2048 byte blocks,   16 bytes per update, 128 updates): 1.88x faster
test 10 ( 2048 byte blocks,  256 bytes per update,   8 updates): 6.00x faster
test 11 ( 2048 byte blocks, 1024 bytes per update,   2 updates): 8.64x faster
test 12 ( 2048 byte blocks, 2048 bytes per update,   1 updates): 9.40x faster
test 13 ( 4096 byte blocks,   16 bytes per update, 256 updates): 1.86x faster
test 14 ( 4096 byte blocks,  256 bytes per update,  16 updates): 6.12x faster
test 15 ( 4096 byte blocks, 1024 bytes per update,   4 updates): 9.03x faster
test 16 ( 4096 byte blocks, 4096 bytes per update,   1 updates): 10.31x faster
test 17 ( 8192 byte blocks,   16 bytes per update, 512 updates): 1.85x faster
test 18 ( 8192 byte blocks,  256 bytes per update,  32 updates): 6.18x faster
test 19 ( 8192 byte blocks, 1024 bytes per update,   8 updates): 9.26x faster
test 20 ( 8192 byte blocks, 4096 bytes per update,   2 updates): 10.64x faster
test 21 ( 8192 byte blocks, 8192 bytes per update,   1 updates): 10.65x faster

A.

Aaro Koskinen (7):
  crypto: octeon - don't disable bottom half in octeon-md5
  crypto: octeon - always disable preemption when using crypto engine
  crypto: octeon - add instruction definitions for SHA1/256/512
  crypto: octeon - add SHA1 module
  crypto: octeon - add SHA256 module
  crypto: octeon - add SHA512 module
  crypto: octeon - enable OCTEON SHA1/256/512 module selection

 arch/mips/cavium-octeon/crypto/Makefile        |   5 +-
 arch/mips/cavium-octeon/crypto/octeon-crypto.c |   4 +-
 arch/mips/cavium-octeon/crypto/octeon-crypto.h |  83 +++++++-
 arch/mips/cavium-octeon/crypto/octeon-md5.c    |   8 -
 arch/mips/cavium-octeon/crypto/octeon-sha1.c   | 241 +++++++++++++++++++++
 arch/mips/cavium-octeon/crypto/octeon-sha256.c | 280 +++++++++++++++++++++++++
 arch/mips/cavium-octeon/crypto/octeon-sha512.c | 277 ++++++++++++++++++++++++
 crypto/Kconfig                                 |  27 +++
 8 files changed, 911 insertions(+), 14 deletions(-)
 create mode 100644 arch/mips/cavium-octeon/crypto/octeon-sha1.c
 create mode 100644 arch/mips/cavium-octeon/crypto/octeon-sha256.c
 create mode 100644 arch/mips/cavium-octeon/crypto/octeon-sha512.c

-- 
2.2.0

[PATCH 1/7] crypto: octeon - don't disable bottom half in octeon-md5

[Aaro Koskinen]

Don't disable bottom half while the crypto engine is in use, as it
should be unnecessary: All kernel crypto engine usage is wrapped with
crypto engine state save/restore, so if we get interrupted by softirq
that uses crypto they should save and restore our context.

This actually fixes an issue when running OCTEON MD5 with interrupts
disabled (tcrypt mode=302). There's a WARNING because the module is
trying to enable the bottom half with irqs disabled:

[   52.656610] ------------[ cut here ]------------
[   52.661439] WARNING: CPU: 1 PID: 428 at /home/aaro/git/linux/kernel/softirq.c:150 __local_bh_enable_ip+0x9c/0xd8()
[   52.671780] Modules linked in: tcrypt(+)
[...]
[   52.763539] [<ffffffff8114082c>] warn_slowpath_common+0x94/0xd8
[   52.769465] [<ffffffff81144614>] __local_bh_enable_ip+0x9c/0xd8
[   52.775390] [<ffffffff81119574>] octeon_md5_final+0x12c/0x1e8
[   52.781144] [<ffffffff81337050>] shash_compat_digest+0xd0/0x1b0

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
---
 arch/mips/cavium-octeon/crypto/octeon-md5.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/arch/mips/cavium-octeon/crypto/octeon-md5.c b/arch/mips/cavium-octeon/crypto/octeon-md5.c
index b909881..3dd8845 100644
--- a/arch/mips/cavium-octeon/crypto/octeon-md5.c
+++ b/arch/mips/cavium-octeon/crypto/octeon-md5.c
@@ -97,7 +97,6 @@ static int octeon_md5_update(struct shash_desc *desc, const u8 *data,
 	memcpy((char *)mctx->block + (sizeof(mctx->block) - avail), data,
 	       avail);
 
-	local_bh_disable();
 	preempt_disable();
 	flags = octeon_crypto_enable(&state);
 	octeon_md5_store_hash(mctx);
@@ -115,7 +114,6 @@ static int octeon_md5_update(struct shash_desc *desc, const u8 *data,
 	octeon_md5_read_hash(mctx);
 	octeon_crypto_disable(&state, flags);
 	preempt_enable();
-	local_bh_enable();
 
 	memcpy(mctx->block, data, len);
 
@@ -133,7 +131,6 @@ static int octeon_md5_final(struct shash_desc *desc, u8 *out)
 
 	*p++ = 0x80;
 
-	local_bh_disable();
 	preempt_disable();
 	flags = octeon_crypto_enable(&state);
 	octeon_md5_store_hash(mctx);
@@ -153,7 +150,6 @@ static int octeon_md5_final(struct shash_desc *desc, u8 *out)
 	octeon_md5_read_hash(mctx);
 	octeon_crypto_disable(&state, flags);
 	preempt_enable();
-	local_bh_enable();
 
 	memcpy(out, mctx->hash, sizeof(mctx->hash));
 	memset(mctx, 0, sizeof(*mctx));
-- 
2.2.0

[PATCH 2/7] crypto: octeon - always disable preemption when using crypto engine

[Aaro Koskinen]

Always disable preemption on behalf of the drivers when crypto engine
is taken into use. This will simplify the usage.

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
---
 arch/mips/cavium-octeon/crypto/octeon-crypto.c | 4 +++-
 arch/mips/cavium-octeon/crypto/octeon-md5.c    | 4 ----
 2 files changed, 3 insertions(+), 5 deletions(-)

diff --git a/arch/mips/cavium-octeon/crypto/octeon-crypto.c b/arch/mips/cavium-octeon/crypto/octeon-crypto.c
index 7c82ff4..f66bd1a 100644
--- a/arch/mips/cavium-octeon/crypto/octeon-crypto.c
+++ b/arch/mips/cavium-octeon/crypto/octeon-crypto.c
@@ -17,7 +17,7 @@
  * crypto operations in calls to octeon_crypto_enable/disable in order to make
  * sure the state of COP2 isn't corrupted if userspace is also performing
  * hardware crypto operations. Allocate the state parameter on the stack.
- * Preemption must be disabled to prevent context switches.
+ * Returns with preemption disabled.
  *
  * @state: Pointer to state structure to store current COP2 state in.
  *
@@ -28,6 +28,7 @@ unsigned long octeon_crypto_enable(struct octeon_cop2_state *state)
 	int status;
 	unsigned long flags;
 
+	preempt_disable();
 	local_irq_save(flags);
 	status = read_c0_status();
 	write_c0_status(status | ST0_CU2);
@@ -62,5 +63,6 @@ void octeon_crypto_disable(struct octeon_cop2_state *state,
 	else
 		write_c0_status(read_c0_status() & ~ST0_CU2);
 	local_irq_restore(flags);
+	preempt_enable();
 }
 EXPORT_SYMBOL_GPL(octeon_crypto_disable);
diff --git a/arch/mips/cavium-octeon/crypto/octeon-md5.c b/arch/mips/cavium-octeon/crypto/octeon-md5.c
index 3dd8845..12dccdb 100644
--- a/arch/mips/cavium-octeon/crypto/octeon-md5.c
+++ b/arch/mips/cavium-octeon/crypto/octeon-md5.c
@@ -97,7 +97,6 @@ static int octeon_md5_update(struct shash_desc *desc, const u8 *data,
 	memcpy((char *)mctx->block + (sizeof(mctx->block) - avail), data,
 	       avail);
 
-	preempt_disable();
 	flags = octeon_crypto_enable(&state);
 	octeon_md5_store_hash(mctx);
 
@@ -113,7 +112,6 @@ static int octeon_md5_update(struct shash_desc *desc, const u8 *data,
 
 	octeon_md5_read_hash(mctx);
 	octeon_crypto_disable(&state, flags);
-	preempt_enable();
 
 	memcpy(mctx->block, data, len);
 
@@ -131,7 +129,6 @@ static int octeon_md5_final(struct shash_desc *desc, u8 *out)
 
 	*p++ = 0x80;
 
-	preempt_disable();
 	flags = octeon_crypto_enable(&state);
 	octeon_md5_store_hash(mctx);
 
@@ -149,7 +146,6 @@ static int octeon_md5_final(struct shash_desc *desc, u8 *out)
 
 	octeon_md5_read_hash(mctx);
 	octeon_crypto_disable(&state, flags);
-	preempt_enable();
 
 	memcpy(out, mctx->hash, sizeof(mctx->hash));
 	memset(mctx, 0, sizeof(*mctx));
-- 
2.2.0

[PATCH 3/7] crypto: octeon - add instruction definitions for SHA1/256/512

[Aaro Koskinen]

Add instruction definitions for SHA1/256/512.

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
---
 arch/mips/cavium-octeon/crypto/octeon-crypto.h | 83 ++++++++++++++++++++++++--
 1 file changed, 79 insertions(+), 4 deletions(-)

diff --git a/arch/mips/cavium-octeon/crypto/octeon-crypto.h b/arch/mips/cavium-octeon/crypto/octeon-crypto.h
index e2a4aec..3550725 100644
--- a/arch/mips/cavium-octeon/crypto/octeon-crypto.h
+++ b/arch/mips/cavium-octeon/crypto/octeon-crypto.h
@@ -5,7 +5,8 @@
  *
  * Copyright (C) 2012-2013 Cavium Inc., All Rights Reserved.
  *
- * MD5 instruction definitions added by Aaro Koskinen <aaro.koskinen@iki.fi>.
+ * MD5/SHA1/SHA256/SHA512 instruction definitions added by
+ * Aaro Koskinen <aaro.koskinen@iki.fi>.
  *
  */
 #ifndef __LINUX_OCTEON_CRYPTO_H
@@ -21,11 +22,11 @@ extern void octeon_crypto_disable(struct octeon_cop2_state *state,
 				  unsigned long flags);
 
 /*
- * Macros needed to implement MD5:
+ * Macros needed to implement MD5/SHA1/SHA256:
  */
 
 /*
- * The index can be 0-1.
+ * The index can be 0-1 (MD5) or 0-2 (SHA1), 0-3 (SHA256).
  */
 #define write_octeon_64bit_hash_dword(value, index)	\
 do {							\
@@ -36,7 +37,7 @@ do {							\
 } while (0)
 
 /*
- * The index can be 0-1.
+ * The index can be 0-1 (MD5) or 0-2 (SHA1), 0-3 (SHA256).
  */
 #define read_octeon_64bit_hash_dword(index)		\
 ({							\
@@ -72,4 +73,78 @@ do {							\
 	: [rt] "d" (value));				\
 } while (0)
 
+/*
+ * The value is the final block dword (64-bit).
+ */
+#define octeon_sha1_start(value)			\
+do {							\
+	__asm__ __volatile__ (				\
+	"dmtc2 %[rt],0x4057"				\
+	:						\
+	: [rt] "d" (value));				\
+} while (0)
+
+/*
+ * The value is the final block dword (64-bit).
+ */
+#define octeon_sha256_start(value)			\
+do {							\
+	__asm__ __volatile__ (				\
+	"dmtc2 %[rt],0x404f"				\
+	:						\
+	: [rt] "d" (value));				\
+} while (0)
+
+/*
+ * Macros needed to implement SHA512:
+ */
+
+/*
+ * The index can be 0-7.
+ */
+#define write_octeon_64bit_hash_sha512(value, index)	\
+do {							\
+	__asm__ __volatile__ (				\
+	"dmtc2 %[rt],0x0250+" STR(index)		\
+	:						\
+	: [rt] "d" (value));				\
+} while (0)
+
+/*
+ * The index can be 0-7.
+ */
+#define read_octeon_64bit_hash_sha512(index)		\
+({							\
+	u64 __value;					\
+							\
+	__asm__ __volatile__ (				\
+	"dmfc2 %[rt],0x0250+" STR(index)		\
+	: [rt] "=d" (__value)				\
+	: );						\
+							\
+	__value;					\
+})
+
+/*
+ * The index can be 0-14.
+ */
+#define write_octeon_64bit_block_sha512(value, index)	\
+do {							\
+	__asm__ __volatile__ (				\
+	"dmtc2 %[rt],0x0240+" STR(index)		\
+	:						\
+	: [rt] "d" (value));				\
+} while (0)
+
+/*
+ * The value is the final block word (64-bit).
+ */
+#define octeon_sha512_start(value)			\
+do {							\
+	__asm__ __volatile__ (				\
+	"dmtc2 %[rt],0x424f"				\
+	:						\
+	: [rt] "d" (value));				\
+} while (0)
+
 #endif /* __LINUX_OCTEON_CRYPTO_H */
-- 
2.2.0

[PATCH 4/7] crypto: octeon - add SHA1 module

[Aaro Koskinen]

Add OCTEON SHA1 module.

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
---
 arch/mips/cavium-octeon/crypto/Makefile      |   3 +-
 arch/mips/cavium-octeon/crypto/octeon-sha1.c | 241 +++++++++++++++++++++++++++
 2 files changed, 243 insertions(+), 1 deletion(-)
 create mode 100644 arch/mips/cavium-octeon/crypto/octeon-sha1.c

diff --git a/arch/mips/cavium-octeon/crypto/Makefile b/arch/mips/cavium-octeon/crypto/Makefile
index a74f76d..3f671d6 100644
--- a/arch/mips/cavium-octeon/crypto/Makefile
+++ b/arch/mips/cavium-octeon/crypto/Makefile
@@ -4,4 +4,5 @@
 
 obj-y += octeon-crypto.o
 
-obj-$(CONFIG_CRYPTO_MD5_OCTEON) += octeon-md5.o
+obj-$(CONFIG_CRYPTO_MD5_OCTEON)		+= octeon-md5.o
+obj-$(CONFIG_CRYPTO_SHA1_OCTEON)	+= octeon-sha1.o
diff --git a/arch/mips/cavium-octeon/crypto/octeon-sha1.c b/arch/mips/cavium-octeon/crypto/octeon-sha1.c
new file mode 100644
index 0000000..2b74b5b
--- /dev/null
+++ b/arch/mips/cavium-octeon/crypto/octeon-sha1.c
@@ -0,0 +1,241 @@
+/*
+ * Cryptographic API.
+ *
+ * SHA1 Secure Hash Algorithm.
+ *
+ * Adapted for OCTEON by Aaro Koskinen <aaro.koskinen@iki.fi>.
+ *
+ * Based on crypto/sha1_generic.c, which is:
+ *
+ * Copyright (c) Alan Smithee.
+ * Copyright (c) Andrew McDonald <andrew@mcdonald.org.uk>
+ * Copyright (c) Jean-Francois Dive <jef@linuxbe.org>
+ *
+ * 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 the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ */
+
+#include <linux/mm.h>
+#include <crypto/sha.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/module.h>
+#include <asm/byteorder.h>
+#include <asm/octeon/octeon.h>
+#include <crypto/internal/hash.h>
+
+#include "octeon-crypto.h"
+
+/*
+ * We pass everything as 64-bit. OCTEON can handle misaligned data.
+ */
+
+static void octeon_sha1_store_hash(struct sha1_state *sctx)
+{
+	u64 *hash = (u64 *)sctx->state;
+	union {
+		u32 word[2];
+		u64 dword;
+	} hash_tail = { { sctx->state[4], } };
+
+	write_octeon_64bit_hash_dword(hash[0], 0);
+	write_octeon_64bit_hash_dword(hash[1], 1);
+	write_octeon_64bit_hash_dword(hash_tail.dword, 2);
+	memzero_explicit(&hash_tail.word[0], sizeof(hash_tail.word[0]));
+}
+
+static void octeon_sha1_read_hash(struct sha1_state *sctx)
+{
+	u64 *hash = (u64 *)sctx->state;
+	union {
+		u32 word[2];
+		u64 dword;
+	} hash_tail;
+
+	hash[0]		= read_octeon_64bit_hash_dword(0);
+	hash[1]		= read_octeon_64bit_hash_dword(1);
+	hash_tail.dword	= read_octeon_64bit_hash_dword(2);
+	sctx->state[4]	= hash_tail.word[0];
+	memzero_explicit(&hash_tail.dword, sizeof(hash_tail.dword));
+}
+
+static void octeon_sha1_transform(const void *_block)
+{
+	const u64 *block = _block;
+
+	write_octeon_64bit_block_dword(block[0], 0);
+	write_octeon_64bit_block_dword(block[1], 1);
+	write_octeon_64bit_block_dword(block[2], 2);
+	write_octeon_64bit_block_dword(block[3], 3);
+	write_octeon_64bit_block_dword(block[4], 4);
+	write_octeon_64bit_block_dword(block[5], 5);
+	write_octeon_64bit_block_dword(block[6], 6);
+	octeon_sha1_start(block[7]);
+}
+
+static int octeon_sha1_init(struct shash_desc *desc)
+{
+	struct sha1_state *sctx = shash_desc_ctx(desc);
+
+	sctx->state[0] = SHA1_H0;
+	sctx->state[1] = SHA1_H1;
+	sctx->state[2] = SHA1_H2;
+	sctx->state[3] = SHA1_H3;
+	sctx->state[4] = SHA1_H4;
+	sctx->count = 0;
+
+	return 0;
+}
+
+static void __octeon_sha1_update(struct sha1_state *sctx, const u8 *data,
+				 unsigned int len)
+{
+	unsigned int partial;
+	unsigned int done;
+	const u8 *src;
+
+	partial = sctx->count % SHA1_BLOCK_SIZE;
+	sctx->count += len;
+	done = 0;
+	src = data;
+
+	if ((partial + len) >= SHA1_BLOCK_SIZE) {
+		if (partial) {
+			done = -partial;
+			memcpy(sctx->buffer + partial, data,
+			       done + SHA1_BLOCK_SIZE);
+			src = sctx->buffer;
+		}
+
+		do {
+			octeon_sha1_transform(src);
+			done += SHA1_BLOCK_SIZE;
+			src = data + done;
+		} while (done + SHA1_BLOCK_SIZE <= len);
+
+		partial = 0;
+	}
+	memcpy(sctx->buffer + partial, src, len - done);
+}
+
+static int octeon_sha1_update(struct shash_desc *desc, const u8 *data,
+			unsigned int len)
+{
+	struct sha1_state *sctx = shash_desc_ctx(desc);
+	struct octeon_cop2_state state;
+	unsigned long flags;
+
+	/*
+	 * Small updates never reach the crypto engine, so the generic sha1 is
+	 * faster because of the heavyweight octeon_crypto_enable() /
+	 * octeon_crypto_disable().
+	 */
+	if ((sctx->count % SHA1_BLOCK_SIZE) + len < SHA1_BLOCK_SIZE)
+		return crypto_sha1_update(desc, data, len);
+
+	flags = octeon_crypto_enable(&state);
+	octeon_sha1_store_hash(sctx);
+
+	__octeon_sha1_update(sctx, data, len);
+
+	octeon_sha1_read_hash(sctx);
+	octeon_crypto_disable(&state, flags);
+
+	return 0;
+}
+
+static int octeon_sha1_final(struct shash_desc *desc, u8 *out)
+{
+	struct sha1_state *sctx = shash_desc_ctx(desc);
+	static const u8 padding[64] = { 0x80, };
+	struct octeon_cop2_state state;
+	__be32 *dst = (__be32 *)out;
+	unsigned int pad_len;
+	unsigned long flags;
+	unsigned int index;
+	__be64 bits;
+	int i;
+
+	/* Save number of bits. */
+	bits = cpu_to_be64(sctx->count << 3);
+
+	/* Pad out to 56 mod 64. */
+	index = sctx->count & 0x3f;
+	pad_len = (index < 56) ? (56 - index) : ((64+56) - index);
+
+	flags = octeon_crypto_enable(&state);
+	octeon_sha1_store_hash(sctx);
+
+	__octeon_sha1_update(sctx, padding, pad_len);
+
+	/* Append length (before padding). */
+	__octeon_sha1_update(sctx, (const u8 *)&bits, sizeof(bits));
+
+	octeon_sha1_read_hash(sctx);
+	octeon_crypto_disable(&state, flags);
+
+	/* Store state in digest */
+	for (i = 0; i < 5; i++)
+		dst[i] = cpu_to_be32(sctx->state[i]);
+
+	/* Zeroize sensitive information. */
+	memset(sctx, 0, sizeof(*sctx));
+
+	return 0;
+}
+
+static int octeon_sha1_export(struct shash_desc *desc, void *out)
+{
+	struct sha1_state *sctx = shash_desc_ctx(desc);
+
+	memcpy(out, sctx, sizeof(*sctx));
+	return 0;
+}
+
+static int octeon_sha1_import(struct shash_desc *desc, const void *in)
+{
+	struct sha1_state *sctx = shash_desc_ctx(desc);
+
+	memcpy(sctx, in, sizeof(*sctx));
+	return 0;
+}
+
+static struct shash_alg octeon_sha1_alg = {
+	.digestsize	=	SHA1_DIGEST_SIZE,
+	.init		=	octeon_sha1_init,
+	.update		=	octeon_sha1_update,
+	.final		=	octeon_sha1_final,
+	.export		=	octeon_sha1_export,
+	.import		=	octeon_sha1_import,
+	.descsize	=	sizeof(struct sha1_state),
+	.statesize	=	sizeof(struct sha1_state),
+	.base		=	{
+		.cra_name	=	"sha1",
+		.cra_driver_name=	"octeon-sha1",
+		.cra_priority	=	OCTEON_CR_OPCODE_PRIORITY,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA1_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+};
+
+static int __init octeon_sha1_mod_init(void)
+{
+	if (!octeon_has_crypto())
+		return -ENOTSUPP;
+	return crypto_register_shash(&octeon_sha1_alg);
+}
+
+static void __exit octeon_sha1_mod_fini(void)
+{
+	crypto_unregister_shash(&octeon_sha1_alg);
+}
+
+module_init(octeon_sha1_mod_init);
+module_exit(octeon_sha1_mod_fini);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("SHA1 Secure Hash Algorithm (OCTEON)");
+MODULE_AUTHOR("Aaro Koskinen <aaro.koskinen@iki.fi>");
-- 
2.2.0

[PATCH 5/7] crypto: octeon - add SHA256 module

[Aaro Koskinen]

Add OCTEON SHA256 module.

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
---
 arch/mips/cavium-octeon/crypto/Makefile        |   1 +
 arch/mips/cavium-octeon/crypto/octeon-sha256.c | 280 +++++++++++++++++++++++++
 2 files changed, 281 insertions(+)
 create mode 100644 arch/mips/cavium-octeon/crypto/octeon-sha256.c

diff --git a/arch/mips/cavium-octeon/crypto/Makefile b/arch/mips/cavium-octeon/crypto/Makefile
index 3f671d6..47806a5 100644
--- a/arch/mips/cavium-octeon/crypto/Makefile
+++ b/arch/mips/cavium-octeon/crypto/Makefile
@@ -6,3 +6,4 @@ obj-y += octeon-crypto.o
 
 obj-$(CONFIG_CRYPTO_MD5_OCTEON)		+= octeon-md5.o
 obj-$(CONFIG_CRYPTO_SHA1_OCTEON)	+= octeon-sha1.o
+obj-$(CONFIG_CRYPTO_SHA256_OCTEON)	+= octeon-sha256.o
diff --git a/arch/mips/cavium-octeon/crypto/octeon-sha256.c b/arch/mips/cavium-octeon/crypto/octeon-sha256.c
new file mode 100644
index 0000000..97e96fe
--- /dev/null
+++ b/arch/mips/cavium-octeon/crypto/octeon-sha256.c
@@ -0,0 +1,280 @@
+/*
+ * Cryptographic API.
+ *
+ * SHA-224 and SHA-256 Secure Hash Algorithm.
+ *
+ * Adapted for OCTEON by Aaro Koskinen <aaro.koskinen@iki.fi>.
+ *
+ * Based on crypto/sha256_generic.c, which is:
+ *
+ * Copyright (c) Jean-Luc Cooke <jlcooke@certainkey.com>
+ * Copyright (c) Andrew McDonald <andrew@mcdonald.org.uk>
+ * Copyright (c) 2002 James Morris <jmorris@intercode.com.au>
+ * SHA224 Support Copyright 2007 Intel Corporation <jonathan.lynch@intel.com>
+ *
+ * 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 the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ */
+
+#include <linux/mm.h>
+#include <crypto/sha.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/module.h>
+#include <asm/byteorder.h>
+#include <asm/octeon/octeon.h>
+#include <crypto/internal/hash.h>
+
+#include "octeon-crypto.h"
+
+/*
+ * We pass everything as 64-bit. OCTEON can handle misaligned data.
+ */
+
+static void octeon_sha256_store_hash(struct sha256_state *sctx)
+{
+	u64 *hash = (u64 *)sctx->state;
+
+	write_octeon_64bit_hash_dword(hash[0], 0);
+	write_octeon_64bit_hash_dword(hash[1], 1);
+	write_octeon_64bit_hash_dword(hash[2], 2);
+	write_octeon_64bit_hash_dword(hash[3], 3);
+}
+
+static void octeon_sha256_read_hash(struct sha256_state *sctx)
+{
+	u64 *hash = (u64 *)sctx->state;
+
+	hash[0] = read_octeon_64bit_hash_dword(0);
+	hash[1] = read_octeon_64bit_hash_dword(1);
+	hash[2] = read_octeon_64bit_hash_dword(2);
+	hash[3] = read_octeon_64bit_hash_dword(3);
+}
+
+static void octeon_sha256_transform(const void *_block)
+{
+	const u64 *block = _block;
+
+	write_octeon_64bit_block_dword(block[0], 0);
+	write_octeon_64bit_block_dword(block[1], 1);
+	write_octeon_64bit_block_dword(block[2], 2);
+	write_octeon_64bit_block_dword(block[3], 3);
+	write_octeon_64bit_block_dword(block[4], 4);
+	write_octeon_64bit_block_dword(block[5], 5);
+	write_octeon_64bit_block_dword(block[6], 6);
+	octeon_sha256_start(block[7]);
+}
+
+static int octeon_sha224_init(struct shash_desc *desc)
+{
+	struct sha256_state *sctx = shash_desc_ctx(desc);
+
+	sctx->state[0] = SHA224_H0;
+	sctx->state[1] = SHA224_H1;
+	sctx->state[2] = SHA224_H2;
+	sctx->state[3] = SHA224_H3;
+	sctx->state[4] = SHA224_H4;
+	sctx->state[5] = SHA224_H5;
+	sctx->state[6] = SHA224_H6;
+	sctx->state[7] = SHA224_H7;
+	sctx->count = 0;
+
+	return 0;
+}
+
+static int octeon_sha256_init(struct shash_desc *desc)
+{
+	struct sha256_state *sctx = shash_desc_ctx(desc);
+
+	sctx->state[0] = SHA256_H0;
+	sctx->state[1] = SHA256_H1;
+	sctx->state[2] = SHA256_H2;
+	sctx->state[3] = SHA256_H3;
+	sctx->state[4] = SHA256_H4;
+	sctx->state[5] = SHA256_H5;
+	sctx->state[6] = SHA256_H6;
+	sctx->state[7] = SHA256_H7;
+	sctx->count = 0;
+
+	return 0;
+}
+
+static void __octeon_sha256_update(struct sha256_state *sctx, const u8 *data,
+				   unsigned int len)
+{
+	unsigned int partial;
+	unsigned int done;
+	const u8 *src;
+
+	partial = sctx->count % SHA256_BLOCK_SIZE;
+	sctx->count += len;
+	done = 0;
+	src = data;
+
+	if ((partial + len) >= SHA256_BLOCK_SIZE) {
+		if (partial) {
+			done = -partial;
+			memcpy(sctx->buf + partial, data,
+			       done + SHA256_BLOCK_SIZE);
+			src = sctx->buf;
+		}
+
+		do {
+			octeon_sha256_transform(src);
+			done += SHA256_BLOCK_SIZE;
+			src = data + done;
+		} while (done + SHA256_BLOCK_SIZE <= len);
+
+		partial = 0;
+	}
+	memcpy(sctx->buf + partial, src, len - done);
+}
+
+static int octeon_sha256_update(struct shash_desc *desc, const u8 *data,
+				unsigned int len)
+{
+	struct sha256_state *sctx = shash_desc_ctx(desc);
+	struct octeon_cop2_state state;
+	unsigned long flags;
+
+	/*
+	 * Small updates never reach the crypto engine, so the generic sha256 is
+	 * faster because of the heavyweight octeon_crypto_enable() /
+	 * octeon_crypto_disable().
+	 */
+	if ((sctx->count % SHA256_BLOCK_SIZE) + len < SHA256_BLOCK_SIZE)
+		return crypto_sha256_update(desc, data, len);
+
+	flags = octeon_crypto_enable(&state);
+	octeon_sha256_store_hash(sctx);
+
+	__octeon_sha256_update(sctx, data, len);
+
+	octeon_sha256_read_hash(sctx);
+	octeon_crypto_disable(&state, flags);
+
+	return 0;
+}
+
+static int octeon_sha256_final(struct shash_desc *desc, u8 *out)
+{
+	struct sha256_state *sctx = shash_desc_ctx(desc);
+	static const u8 padding[64] = { 0x80, };
+	struct octeon_cop2_state state;
+	__be32 *dst = (__be32 *)out;
+	unsigned int pad_len;
+	unsigned long flags;
+	unsigned int index;
+	__be64 bits;
+	int i;
+
+	/* Save number of bits. */
+	bits = cpu_to_be64(sctx->count << 3);
+
+	/* Pad out to 56 mod 64. */
+	index = sctx->count & 0x3f;
+	pad_len = (index < 56) ? (56 - index) : ((64+56) - index);
+
+	flags = octeon_crypto_enable(&state);
+	octeon_sha256_store_hash(sctx);
+
+	__octeon_sha256_update(sctx, padding, pad_len);
+
+	/* Append length (before padding). */
+	__octeon_sha256_update(sctx, (const u8 *)&bits, sizeof(bits));
+
+	octeon_sha256_read_hash(sctx);
+	octeon_crypto_disable(&state, flags);
+
+	/* Store state in digest */
+	for (i = 0; i < 8; i++)
+		dst[i] = cpu_to_be32(sctx->state[i]);
+
+	/* Zeroize sensitive information. */
+	memset(sctx, 0, sizeof(*sctx));
+
+	return 0;
+}
+
+static int octeon_sha224_final(struct shash_desc *desc, u8 *hash)
+{
+	u8 D[SHA256_DIGEST_SIZE];
+
+	octeon_sha256_final(desc, D);
+
+	memcpy(hash, D, SHA224_DIGEST_SIZE);
+	memzero_explicit(D, SHA256_DIGEST_SIZE);
+
+	return 0;
+}
+
+static int octeon_sha256_export(struct shash_desc *desc, void *out)
+{
+	struct sha256_state *sctx = shash_desc_ctx(desc);
+
+	memcpy(out, sctx, sizeof(*sctx));
+	return 0;
+}
+
+static int octeon_sha256_import(struct shash_desc *desc, const void *in)
+{
+	struct sha256_state *sctx = shash_desc_ctx(desc);
+
+	memcpy(sctx, in, sizeof(*sctx));
+	return 0;
+}
+
+static struct shash_alg octeon_sha256_algs[2] = { {
+	.digestsize	=	SHA256_DIGEST_SIZE,
+	.init		=	octeon_sha256_init,
+	.update		=	octeon_sha256_update,
+	.final		=	octeon_sha256_final,
+	.export		=	octeon_sha256_export,
+	.import		=	octeon_sha256_import,
+	.descsize	=	sizeof(struct sha256_state),
+	.statesize	=	sizeof(struct sha256_state),
+	.base		=	{
+		.cra_name	=	"sha256",
+		.cra_driver_name=	"octeon-sha256",
+		.cra_priority	=	OCTEON_CR_OPCODE_PRIORITY,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA256_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+}, {
+	.digestsize	=	SHA224_DIGEST_SIZE,
+	.init		=	octeon_sha224_init,
+	.update		=	octeon_sha256_update,
+	.final		=	octeon_sha224_final,
+	.descsize	=	sizeof(struct sha256_state),
+	.base		=	{
+		.cra_name	=	"sha224",
+		.cra_driver_name=	"octeon-sha224",
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA224_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+} };
+
+static int __init octeon_sha256_mod_init(void)
+{
+	if (!octeon_has_crypto())
+		return -ENOTSUPP;
+	return crypto_register_shashes(octeon_sha256_algs,
+				       ARRAY_SIZE(octeon_sha256_algs));
+}
+
+static void __exit octeon_sha256_mod_fini(void)
+{
+	crypto_unregister_shashes(octeon_sha256_algs,
+				  ARRAY_SIZE(octeon_sha256_algs));
+}
+
+module_init(octeon_sha256_mod_init);
+module_exit(octeon_sha256_mod_fini);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("SHA-224 and SHA-256 Secure Hash Algorithm (OCTEON)");
+MODULE_AUTHOR("Aaro Koskinen <aaro.koskinen@iki.fi>");
-- 
2.2.0

[PATCH 6/7] crypto: octeon - add SHA512 module

[Aaro Koskinen]

Add OCTEON SHA512 module.

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
---
 arch/mips/cavium-octeon/crypto/Makefile        |   1 +
 arch/mips/cavium-octeon/crypto/octeon-sha512.c | 277 +++++++++++++++++++++++++
 2 files changed, 278 insertions(+)
 create mode 100644 arch/mips/cavium-octeon/crypto/octeon-sha512.c

diff --git a/arch/mips/cavium-octeon/crypto/Makefile b/arch/mips/cavium-octeon/crypto/Makefile
index 47806a5..f7aa9d5 100644
--- a/arch/mips/cavium-octeon/crypto/Makefile
+++ b/arch/mips/cavium-octeon/crypto/Makefile
@@ -7,3 +7,4 @@ obj-y += octeon-crypto.o
 obj-$(CONFIG_CRYPTO_MD5_OCTEON)		+= octeon-md5.o
 obj-$(CONFIG_CRYPTO_SHA1_OCTEON)	+= octeon-sha1.o
 obj-$(CONFIG_CRYPTO_SHA256_OCTEON)	+= octeon-sha256.o
+obj-$(CONFIG_CRYPTO_SHA512_OCTEON)	+= octeon-sha512.o
diff --git a/arch/mips/cavium-octeon/crypto/octeon-sha512.c b/arch/mips/cavium-octeon/crypto/octeon-sha512.c
new file mode 100644
index 0000000..d5fb3c6
--- /dev/null
+++ b/arch/mips/cavium-octeon/crypto/octeon-sha512.c
@@ -0,0 +1,277 @@
+/*
+ * Cryptographic API.
+ *
+ * SHA-512 and SHA-384 Secure Hash Algorithm.
+ *
+ * Adapted for OCTEON by Aaro Koskinen <aaro.koskinen@iki.fi>.
+ *
+ * Based on crypto/sha512_generic.c, which is:
+ *
+ * Copyright (c) Jean-Luc Cooke <jlcooke@certainkey.com>
+ * Copyright (c) Andrew McDonald <andrew@mcdonald.org.uk>
+ * Copyright (c) 2003 Kyle McMartin <kyle@debian.org>
+ *
+ * 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 the
+ * Free Software Foundation; either version 2, or (at your option) any
+ * later version.
+ */
+
+#include <linux/mm.h>
+#include <crypto/sha.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/module.h>
+#include <asm/byteorder.h>
+#include <asm/octeon/octeon.h>
+#include <crypto/internal/hash.h>
+
+#include "octeon-crypto.h"
+
+/*
+ * We pass everything as 64-bit. OCTEON can handle misaligned data.
+ */
+
+static void octeon_sha512_store_hash(struct sha512_state *sctx)
+{
+	write_octeon_64bit_hash_sha512(sctx->state[0], 0);
+	write_octeon_64bit_hash_sha512(sctx->state[1], 1);
+	write_octeon_64bit_hash_sha512(sctx->state[2], 2);
+	write_octeon_64bit_hash_sha512(sctx->state[3], 3);
+	write_octeon_64bit_hash_sha512(sctx->state[4], 4);
+	write_octeon_64bit_hash_sha512(sctx->state[5], 5);
+	write_octeon_64bit_hash_sha512(sctx->state[6], 6);
+	write_octeon_64bit_hash_sha512(sctx->state[7], 7);
+}
+
+static void octeon_sha512_read_hash(struct sha512_state *sctx)
+{
+	sctx->state[0] = read_octeon_64bit_hash_sha512(0);
+	sctx->state[1] = read_octeon_64bit_hash_sha512(1);
+	sctx->state[2] = read_octeon_64bit_hash_sha512(2);
+	sctx->state[3] = read_octeon_64bit_hash_sha512(3);
+	sctx->state[4] = read_octeon_64bit_hash_sha512(4);
+	sctx->state[5] = read_octeon_64bit_hash_sha512(5);
+	sctx->state[6] = read_octeon_64bit_hash_sha512(6);
+	sctx->state[7] = read_octeon_64bit_hash_sha512(7);
+}
+
+static void octeon_sha512_transform(const void *_block)
+{
+	const u64 *block = _block;
+
+	write_octeon_64bit_block_sha512(block[0], 0);
+	write_octeon_64bit_block_sha512(block[1], 1);
+	write_octeon_64bit_block_sha512(block[2], 2);
+	write_octeon_64bit_block_sha512(block[3], 3);
+	write_octeon_64bit_block_sha512(block[4], 4);
+	write_octeon_64bit_block_sha512(block[5], 5);
+	write_octeon_64bit_block_sha512(block[6], 6);
+	write_octeon_64bit_block_sha512(block[7], 7);
+	write_octeon_64bit_block_sha512(block[8], 8);
+	write_octeon_64bit_block_sha512(block[9], 9);
+	write_octeon_64bit_block_sha512(block[10], 10);
+	write_octeon_64bit_block_sha512(block[11], 11);
+	write_octeon_64bit_block_sha512(block[12], 12);
+	write_octeon_64bit_block_sha512(block[13], 13);
+	write_octeon_64bit_block_sha512(block[14], 14);
+	octeon_sha512_start(block[15]);
+}
+
+static int octeon_sha512_init(struct shash_desc *desc)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+
+	sctx->state[0] = SHA512_H0;
+	sctx->state[1] = SHA512_H1;
+	sctx->state[2] = SHA512_H2;
+	sctx->state[3] = SHA512_H3;
+	sctx->state[4] = SHA512_H4;
+	sctx->state[5] = SHA512_H5;
+	sctx->state[6] = SHA512_H6;
+	sctx->state[7] = SHA512_H7;
+	sctx->count[0] = sctx->count[1] = 0;
+
+	return 0;
+}
+
+static int octeon_sha384_init(struct shash_desc *desc)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+
+	sctx->state[0] = SHA384_H0;
+	sctx->state[1] = SHA384_H1;
+	sctx->state[2] = SHA384_H2;
+	sctx->state[3] = SHA384_H3;
+	sctx->state[4] = SHA384_H4;
+	sctx->state[5] = SHA384_H5;
+	sctx->state[6] = SHA384_H6;
+	sctx->state[7] = SHA384_H7;
+	sctx->count[0] = sctx->count[1] = 0;
+
+	return 0;
+}
+
+static void __octeon_sha512_update(struct sha512_state *sctx, const u8 *data,
+				   unsigned int len)
+{
+	unsigned int part_len;
+	unsigned int index;
+	unsigned int i;
+
+	/* Compute number of bytes mod 128. */
+	index = sctx->count[0] % SHA512_BLOCK_SIZE;
+
+	/* Update number of bytes. */
+	if ((sctx->count[0] += len) < len)
+		sctx->count[1]++;
+
+	part_len = SHA512_BLOCK_SIZE - index;
+
+	/* Transform as many times as possible. */
+	if (len >= part_len) {
+		memcpy(&sctx->buf[index], data, part_len);
+		octeon_sha512_transform(sctx->buf);
+
+		for (i = part_len; i + SHA512_BLOCK_SIZE <= len;
+			i += SHA512_BLOCK_SIZE)
+			octeon_sha512_transform(&data[i]);
+
+		index = 0;
+	} else {
+		i = 0;
+	}
+
+	/* Buffer remaining input. */
+	memcpy(&sctx->buf[index], &data[i], len - i);
+}
+
+static int octeon_sha512_update(struct shash_desc *desc, const u8 *data,
+				unsigned int len)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+	struct octeon_cop2_state state;
+	unsigned long flags;
+
+	/*
+	 * Small updates never reach the crypto engine, so the generic sha512 is
+	 * faster because of the heavyweight octeon_crypto_enable() /
+	 * octeon_crypto_disable().
+	 */
+	if ((sctx->count[0] % SHA512_BLOCK_SIZE) + len < SHA512_BLOCK_SIZE)
+		return crypto_sha512_update(desc, data, len);
+
+	flags = octeon_crypto_enable(&state);
+	octeon_sha512_store_hash(sctx);
+
+	__octeon_sha512_update(sctx, data, len);
+
+	octeon_sha512_read_hash(sctx);
+	octeon_crypto_disable(&state, flags);
+
+	return 0;
+}
+
+static int octeon_sha512_final(struct shash_desc *desc, u8 *hash)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+	static u8 padding[128] = { 0x80, };
+	struct octeon_cop2_state state;
+	__be64 *dst = (__be64 *)hash;
+	unsigned int pad_len;
+	unsigned long flags;
+	unsigned int index;
+	__be64 bits[2];
+	int i;
+
+	/* Save number of bits. */
+	bits[1] = cpu_to_be64(sctx->count[0] << 3);
+	bits[0] = cpu_to_be64(sctx->count[1] << 3 | sctx->count[0] >> 61);
+
+	/* Pad out to 112 mod 128. */
+	index = sctx->count[0] & 0x7f;
+	pad_len = (index < 112) ? (112 - index) : ((128+112) - index);
+
+	flags = octeon_crypto_enable(&state);
+	octeon_sha512_store_hash(sctx);
+
+	__octeon_sha512_update(sctx, padding, pad_len);
+
+	/* Append length (before padding). */
+	__octeon_sha512_update(sctx, (const u8 *)bits, sizeof(bits));
+
+	octeon_sha512_read_hash(sctx);
+	octeon_crypto_disable(&state, flags);
+
+	/* Store state in digest. */
+	for (i = 0; i < 8; i++)
+		dst[i] = cpu_to_be64(sctx->state[i]);
+
+	/* Zeroize sensitive information. */
+	memset(sctx, 0, sizeof(struct sha512_state));
+
+	return 0;
+}
+
+static int octeon_sha384_final(struct shash_desc *desc, u8 *hash)
+{
+	u8 D[64];
+
+	octeon_sha512_final(desc, D);
+
+	memcpy(hash, D, 48);
+	memzero_explicit(D, 64);
+
+	return 0;
+}
+
+static struct shash_alg octeon_sha512_algs[2] = { {
+	.digestsize	=	SHA512_DIGEST_SIZE,
+	.init		=	octeon_sha512_init,
+	.update		=	octeon_sha512_update,
+	.final		=	octeon_sha512_final,
+	.descsize	=	sizeof(struct sha512_state),
+	.base		=	{
+		.cra_name	=	"sha512",
+		.cra_driver_name=	"octeon-sha512",
+		.cra_priority	=	OCTEON_CR_OPCODE_PRIORITY,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA512_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+}, {
+	.digestsize	=	SHA384_DIGEST_SIZE,
+	.init		=	octeon_sha384_init,
+	.update		=	octeon_sha512_update,
+	.final		=	octeon_sha384_final,
+	.descsize	=	sizeof(struct sha512_state),
+	.base		=	{
+		.cra_name	=	"sha384",
+		.cra_driver_name=	"octeon-sha384",
+		.cra_priority	=	OCTEON_CR_OPCODE_PRIORITY,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA384_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+} };
+
+static int __init octeon_sha512_mod_init(void)
+{
+	if (!octeon_has_crypto())
+		return -ENOTSUPP;
+	return crypto_register_shashes(octeon_sha512_algs,
+				       ARRAY_SIZE(octeon_sha512_algs));
+}
+
+static void __exit octeon_sha512_mod_fini(void)
+{
+	crypto_unregister_shashes(octeon_sha512_algs,
+				  ARRAY_SIZE(octeon_sha512_algs));
+}
+
+module_init(octeon_sha512_mod_init);
+module_exit(octeon_sha512_mod_fini);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("SHA-512 and SHA-384 Secure Hash Algorithms (OCTEON)");
+MODULE_AUTHOR("Aaro Koskinen <aaro.koskinen@iki.fi>");
-- 
2.2.0

[PATCH 7/7] crypto: octeon - enable OCTEON SHA1/256/512 module selection

[Aaro Koskinen]

Enable user to select OCTEON SHA1/256/512 modules.

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
---
 crypto/Kconfig | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

diff --git a/crypto/Kconfig b/crypto/Kconfig
index 50f4da4..38b2315 100644
--- a/crypto/Kconfig
+++ b/crypto/Kconfig
@@ -546,6 +546,15 @@ config CRYPTO_SHA512_SSSE3
 	  Extensions version 1 (AVX1), or Advanced Vector Extensions
 	  version 2 (AVX2) instructions, when available.
 
+config CRYPTO_SHA1_OCTEON
+	tristate "SHA1 digest algorithm (OCTEON)"
+	depends on CPU_CAVIUM_OCTEON
+	select CRYPTO_SHA1
+	select CRYPTO_HASH
+	help
+	  SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2) implemented
+	  using OCTEON crypto instructions, when available.
+
 config CRYPTO_SHA1_SPARC64
 	tristate "SHA1 digest algorithm (SPARC64)"
 	depends on SPARC64
@@ -610,6 +619,15 @@ config CRYPTO_SHA256
 	  This code also includes SHA-224, a 224 bit hash with 112 bits
 	  of security against collision attacks.
 
+config CRYPTO_SHA256_OCTEON
+	tristate "SHA224 and SHA256 digest algorithm (OCTEON)"
+	depends on CPU_CAVIUM_OCTEON
+	select CRYPTO_SHA256
+	select CRYPTO_HASH
+	help
+	  SHA-256 secure hash standard (DFIPS 180-2) implemented
+	  using OCTEON crypto instructions, when available.
+
 config CRYPTO_SHA256_SPARC64
 	tristate "SHA224 and SHA256 digest algorithm (SPARC64)"
 	depends on SPARC64
@@ -631,6 +649,15 @@ config CRYPTO_SHA512
 	  This code also includes SHA-384, a 384 bit hash with 192 bits
 	  of security against collision attacks.
 
+config CRYPTO_SHA512_OCTEON
+	tristate "SHA384 and SHA512 digest algorithms (OCTEON)"
+	depends on CPU_CAVIUM_OCTEON
+	select CRYPTO_SHA512
+	select CRYPTO_HASH
+	help
+	  SHA-512 secure hash standard (DFIPS 180-2) implemented
+	  using OCTEON crypto instructions, when available.
+
 config CRYPTO_SHA512_SPARC64
 	tristate "SHA384 and SHA512 digest algorithm (SPARC64)"
 	depends on SPARC64
-- 
2.2.0

Re: [PATCH 0/7] crypto: OCTEON MD5 bugfix + SHA modules

[Herbert Xu]

On Sun, Mar 08, 2015 at 10:07:40PM +0200, Aaro Koskinen wrote:
> Hi,
> 
> The first patch is a bug fix for OCTEON MD5 aimed for 4.0-rc cycle.

Please send such bug fixes in a separate series in future.
For this one in particular it does not appear to be critical
enough to go in straight away so I have pushed it into cryptodev
along with the rest.

> The remaining patches add SHA1/SHA256/SHA512 modules. I have tested
> these on the following OCTEON boards and CPUs with 4.0-rc2:
> 
> 	D-Link DSR-1000N:	CN5010p1.1-500-SCP
> 	EdgeRouter Lite:	CN5020p1.1-500-SCP
> 	EdgeRouter Pro:		CN6120p1.1-1000-NSP
> 
> All selftests are passing. With tcrypt, I get the following numbers
> on speed compared to the generic modules:

All applied.  Thanks!
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt