Merge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6

Pull crypto update from Herbert Xu:
 "Here is the crypto update for 3.14:

   - Improved crypto_memneq helper
   - Use cyprto_memneq in arch-specific crypto code
   - Replaced orphaned DCP driver with Freescale MXS DCP driver
   - Added AVX/AVX2 version of AESNI-GCM encode and decode
   - Added AMD Cryptographic Coprocessor (CCP) driver
   - Misc fixes"

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (41 commits)
  crypto: aesni - fix build on x86 (32bit)
  crypto: mxs - Fix sparse non static symbol warning
  crypto: ccp - CCP device enabled/disabled changes
  crypto: ccp - Cleanup hash invocation calls
  crypto: ccp - Change data length declarations to u64
  crypto: ccp - Check for caller result area before using it
  crypto: ccp - Cleanup scatterlist usage
  crypto: ccp - Apply appropriate gfp_t type to memory allocations
  crypto: drivers - Sort drivers/crypto/Makefile
  ARM: mxs: dts: Enable DCP for MXS
  crypto: mxs - Add Freescale MXS DCP driver
  crypto: mxs - Remove the old DCP driver
  crypto: ahash - Fully restore ahash request before completing
  crypto: aesni - fix build on x86 (32bit)
  crypto: talitos - Remove redundant dev_set_drvdata
  crypto: ccp - Remove redundant dev_set_drvdata
  crypto: crypto4xx - Remove redundant dev_set_drvdata
  crypto: caam - simplify and harden key parsing
  crypto: omap-sham - Fix Polling mode for larger blocks
  crypto: tcrypt - Added speed tests for AEAD crypto alogrithms in tcrypt test suite
  ...

Documentation/devicetree/bindings/crypto/fsl-dcp.txt

+Freescale DCP (Data Co-Processor) found on i.MX23/i.MX28 .
+
+Required properties:
+- compatible : Should be "fsl,<soc>-dcp"
+- reg : Should contain MXS DCP registers location and length
+- interrupts : Should contain MXS DCP interrupt numbers, VMI IRQ and DCP IRQ
+               must be supplied, optionally Secure IRQ can be present, but
+	       is currently not implemented and not used.
+
+Example:
+
+dcp@80028000 {
+	compatible = "fsl,imx28-dcp", "fsl,imx23-dcp";
+	reg = <0x80028000 0x2000>;
+	interrupts = <52 53>;
+	status = "okay";
+};

MAINTAINERS

@@ -538,6 +538,13 @@ F:	drivers/tty/serial/altera_jtaguart.c
 F:	include/linux/altera_uart.h
 F:	include/linux/altera_jtaguart.h
 
+AMD CRYPTOGRAPHIC COPROCESSOR (CCP) DRIVER
+M:	Tom Lendacky <thomas.lendacky@amd.com>
+L:	linux-crypto@vger.kernel.org
+S:	Supported
+F:	drivers/crypto/ccp/
+F:	include/linux/ccp.h
+
 AMD FAM15H PROCESSOR POWER MONITORING DRIVER
 M:	Andreas Herrmann <herrmann.der.user@googlemail.com>
 L:	lm-sensors@lm-sensors.org

arch/arm/boot/dts/imx23.dtsi

@@ -337,8 +337,10 @@ dma_apbx: dma-apbx@80024000 {
 			};
 
 			dcp@80028000 {
+				compatible = "fsl,imx23-dcp";
 				reg = <0x80028000 0x2000>;
-				status = "disabled";
+				interrupts = <53 54>;
+				status = "okay";
 			};
 
 			pxp@8002a000 {

arch/arm/boot/dts/imx28.dtsi

@@ -813,9 +813,10 @@ dma_apbx: dma-apbx@80024000 {
 			};
 
 			dcp: dcp@80028000 {
+				compatible = "fsl,imx28-dcp", "fsl,imx23-dcp";
 				reg = <0x80028000 0x2000>;
 				interrupts = <52 53 54>;
-				compatible = "fsl-dcp";
+				status = "okay";
 			};
 
 			pxp: pxp@8002a000 {

arch/s390/crypto/des_s390.c

@@ -237,8 +237,8 @@ static int des3_setkey(struct crypto_tfm *tfm, const u8 *key,
 	struct s390_des_ctx *ctx = crypto_tfm_ctx(tfm);
 	u32 *flags = &tfm->crt_flags;
 
-	if (!(memcmp(key, &key[DES_KEY_SIZE], DES_KEY_SIZE) &&
-	    memcmp(&key[DES_KEY_SIZE], &key[DES_KEY_SIZE * 2],
+	if (!(crypto_memneq(key, &key[DES_KEY_SIZE], DES_KEY_SIZE) &&
+	    crypto_memneq(&key[DES_KEY_SIZE], &key[DES_KEY_SIZE * 2],
 			  DES_KEY_SIZE)) &&
 	    (*flags & CRYPTO_TFM_REQ_WEAK_KEY)) {
 		*flags |= CRYPTO_TFM_RES_WEAK_KEY;

arch/x86/crypto/Makefile

@@ -76,6 +76,7 @@ ifeq ($(avx2_supported),yes)
 endif
 
 aesni-intel-y := aesni-intel_asm.o aesni-intel_glue.o fpu.o
+aesni-intel-$(CONFIG_64BIT) += aesni-intel_avx-x86_64.o
 ghash-clmulni-intel-y := ghash-clmulni-intel_asm.o ghash-clmulni-intel_glue.o
 sha1-ssse3-y := sha1_ssse3_asm.o sha1_ssse3_glue.o
 crc32c-intel-y := crc32c-intel_glue.o

arch/x86/crypto/aesni-intel_glue.c

@@ -101,6 +101,9 @@ asmlinkage void aesni_cbc_dec(struct crypto_aes_ctx *ctx, u8 *out,
 int crypto_fpu_init(void);
 void crypto_fpu_exit(void);
 
+#define AVX_GEN2_OPTSIZE 640
+#define AVX_GEN4_OPTSIZE 4096
+
 #ifdef CONFIG_X86_64
 asmlinkage void aesni_ctr_enc(struct crypto_aes_ctx *ctx, u8 *out,
 			      const u8 *in, unsigned int len, u8 *iv);
@@ -150,6 +153,123 @@ asmlinkage void aesni_gcm_dec(void *ctx, u8 *out,
 			u8 *hash_subkey, const u8 *aad, unsigned long aad_len,
 			u8 *auth_tag, unsigned long auth_tag_len);
 
+
+#ifdef CONFIG_AS_AVX
+/*
+ * asmlinkage void aesni_gcm_precomp_avx_gen2()
+ * gcm_data *my_ctx_data, context data
+ * u8 *hash_subkey,  the Hash sub key input. Data starts on a 16-byte boundary.
+ */
+asmlinkage void aesni_gcm_precomp_avx_gen2(void *my_ctx_data, u8 *hash_subkey);
+
+asmlinkage void aesni_gcm_enc_avx_gen2(void *ctx, u8 *out,
+			const u8 *in, unsigned long plaintext_len, u8 *iv,
+			const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len);
+
+asmlinkage void aesni_gcm_dec_avx_gen2(void *ctx, u8 *out,
+			const u8 *in, unsigned long ciphertext_len, u8 *iv,
+			const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len);
+
+static void aesni_gcm_enc_avx(void *ctx, u8 *out,
+			const u8 *in, unsigned long plaintext_len, u8 *iv,
+			u8 *hash_subkey, const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len)
+{
+	if (plaintext_len < AVX_GEN2_OPTSIZE) {
+		aesni_gcm_enc(ctx, out, in, plaintext_len, iv, hash_subkey, aad,
+				aad_len, auth_tag, auth_tag_len);
+	} else {
+		aesni_gcm_precomp_avx_gen2(ctx, hash_subkey);
+		aesni_gcm_enc_avx_gen2(ctx, out, in, plaintext_len, iv, aad,
+					aad_len, auth_tag, auth_tag_len);
+	}
+}
+
+static void aesni_gcm_dec_avx(void *ctx, u8 *out,
+			const u8 *in, unsigned long ciphertext_len, u8 *iv,
+			u8 *hash_subkey, const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len)
+{
+	if (ciphertext_len < AVX_GEN2_OPTSIZE) {
+		aesni_gcm_dec(ctx, out, in, ciphertext_len, iv, hash_subkey, aad,
+				aad_len, auth_tag, auth_tag_len);
+	} else {
+		aesni_gcm_precomp_avx_gen2(ctx, hash_subkey);
+		aesni_gcm_dec_avx_gen2(ctx, out, in, ciphertext_len, iv, aad,
+					aad_len, auth_tag, auth_tag_len);
+	}
+}
+#endif
+
+#ifdef CONFIG_AS_AVX2
+/*
+ * asmlinkage void aesni_gcm_precomp_avx_gen4()
+ * gcm_data *my_ctx_data, context data
+ * u8 *hash_subkey,  the Hash sub key input. Data starts on a 16-byte boundary.
+ */
+asmlinkage void aesni_gcm_precomp_avx_gen4(void *my_ctx_data, u8 *hash_subkey);
+
+asmlinkage void aesni_gcm_enc_avx_gen4(void *ctx, u8 *out,
+			const u8 *in, unsigned long plaintext_len, u8 *iv,
+			const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len);
+
+asmlinkage void aesni_gcm_dec_avx_gen4(void *ctx, u8 *out,
+			const u8 *in, unsigned long ciphertext_len, u8 *iv,
+			const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len);
+
+static void aesni_gcm_enc_avx2(void *ctx, u8 *out,
+			const u8 *in, unsigned long plaintext_len, u8 *iv,
+			u8 *hash_subkey, const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len)
+{
+	if (plaintext_len < AVX_GEN2_OPTSIZE) {
+		aesni_gcm_enc(ctx, out, in, plaintext_len, iv, hash_subkey, aad,
+				aad_len, auth_tag, auth_tag_len);
+	} else if (plaintext_len < AVX_GEN4_OPTSIZE) {
+		aesni_gcm_precomp_avx_gen2(ctx, hash_subkey);
+		aesni_gcm_enc_avx_gen2(ctx, out, in, plaintext_len, iv, aad,
+					aad_len, auth_tag, auth_tag_len);
+	} else {
+		aesni_gcm_precomp_avx_gen4(ctx, hash_subkey);
+		aesni_gcm_enc_avx_gen4(ctx, out, in, plaintext_len, iv, aad,
+					aad_len, auth_tag, auth_tag_len);
+	}
+}
+
+static void aesni_gcm_dec_avx2(void *ctx, u8 *out,
+			const u8 *in, unsigned long ciphertext_len, u8 *iv,
+			u8 *hash_subkey, const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len)
+{
+	if (ciphertext_len < AVX_GEN2_OPTSIZE) {
+		aesni_gcm_dec(ctx, out, in, ciphertext_len, iv, hash_subkey,
+				aad, aad_len, auth_tag, auth_tag_len);
+	} else if (ciphertext_len < AVX_GEN4_OPTSIZE) {
+		aesni_gcm_precomp_avx_gen2(ctx, hash_subkey);
+		aesni_gcm_dec_avx_gen2(ctx, out, in, ciphertext_len, iv, aad,
+					aad_len, auth_tag, auth_tag_len);
+	} else {
+		aesni_gcm_precomp_avx_gen4(ctx, hash_subkey);
+		aesni_gcm_dec_avx_gen4(ctx, out, in, ciphertext_len, iv, aad,
+					aad_len, auth_tag, auth_tag_len);
+	}
+}
+#endif
+
+static void (*aesni_gcm_enc_tfm)(void *ctx, u8 *out,
+			const u8 *in, unsigned long plaintext_len, u8 *iv,
+			u8 *hash_subkey, const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len);
+
+static void (*aesni_gcm_dec_tfm)(void *ctx, u8 *out,
+			const u8 *in, unsigned long ciphertext_len, u8 *iv,
+			u8 *hash_subkey, const u8 *aad, unsigned long aad_len,
+			u8 *auth_tag, unsigned long auth_tag_len);
+
 static inline struct
 aesni_rfc4106_gcm_ctx *aesni_rfc4106_gcm_ctx_get(struct crypto_aead *tfm)
 {
@@ -915,7 +1035,7 @@ static int __driver_rfc4106_encrypt(struct aead_request *req)
 		dst = src;
 	}
 
-	aesni_gcm_enc(aes_ctx, dst, src, (unsigned long)req->cryptlen, iv,
+	aesni_gcm_enc_tfm(aes_ctx, dst, src, (unsigned long)req->cryptlen, iv,
 		ctx->hash_subkey, assoc, (unsigned long)req->assoclen, dst
 		+ ((unsigned long)req->cryptlen), auth_tag_len);
 
@@ -996,12 +1116,12 @@ static int __driver_rfc4106_decrypt(struct aead_request *req)
 		dst = src;
 	}
 
-	aesni_gcm_dec(aes_ctx, dst, src, tempCipherLen, iv,
+	aesni_gcm_dec_tfm(aes_ctx, dst, src, tempCipherLen, iv,
 		ctx->hash_subkey, assoc, (unsigned long)req->assoclen,
 		authTag, auth_tag_len);
 
 	/* Compare generated tag with passed in tag. */
-	retval = memcmp(src + tempCipherLen, authTag, auth_tag_len) ?
+	retval = crypto_memneq(src + tempCipherLen, authTag, auth_tag_len) ?
 		-EBADMSG : 0;
 
 	if (one_entry_in_sg) {
@@ -1353,6 +1473,27 @@ static int __init aesni_init(void)
 
 	if (!x86_match_cpu(aesni_cpu_id))
 		return -ENODEV;
+#ifdef CONFIG_X86_64
+#ifdef CONFIG_AS_AVX2
+	if (boot_cpu_has(X86_FEATURE_AVX2)) {
+		pr_info("AVX2 version of gcm_enc/dec engaged.\n");
+		aesni_gcm_enc_tfm = aesni_gcm_enc_avx2;
+		aesni_gcm_dec_tfm = aesni_gcm_dec_avx2;
+	} else
+#endif
+#ifdef CONFIG_AS_AVX
+	if (boot_cpu_has(X86_FEATURE_AVX)) {
+		pr_info("AVX version of gcm_enc/dec engaged.\n");
+		aesni_gcm_enc_tfm = aesni_gcm_enc_avx;
+		aesni_gcm_dec_tfm = aesni_gcm_dec_avx;
+	} else
+#endif
+	{
+		pr_info("SSE version of gcm_enc/dec engaged.\n");
+		aesni_gcm_enc_tfm = aesni_gcm_enc;
+		aesni_gcm_dec_tfm = aesni_gcm_dec;
+	}
+#endif
 
 	err = crypto_fpu_init();
 	if (err)

crypto/Makefile

@@ -2,11 +2,6 @@
 # Cryptographic API
 #
 
-# memneq MUST be built with -Os or -O0 to prevent early-return optimizations
-# that will defeat memneq's actual purpose to prevent timing attacks.
-CFLAGS_REMOVE_memneq.o := -O1 -O2 -O3
-CFLAGS_memneq.o := -Os
-
 obj-$(CONFIG_CRYPTO) += crypto.o
 crypto-y := api.o cipher.o compress.o memneq.o
 

crypto/ahash.c

@@ -213,7 +213,10 @@ static void ahash_op_unaligned_done(struct crypto_async_request *req, int err)
 
 	ahash_op_unaligned_finish(areq, err);
 
-	complete(data, err);
+	areq->base.complete = complete;
+	areq->base.data = data;
+
+	complete(&areq->base, err);
 }
 
 static int ahash_op_unaligned(struct ahash_request *req,

crypto/memneq.c

@@ -72,6 +72,7 @@ __crypto_memneq_generic(const void *a, const void *b, size_t size)
 #if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)
 	while (size >= sizeof(unsigned long)) {
 		neq |= *(unsigned long *)a ^ *(unsigned long *)b;
+		OPTIMIZER_HIDE_VAR(neq);
 		a += sizeof(unsigned long);
 		b += sizeof(unsigned long);
 		size -= sizeof(unsigned long);
@@ -79,6 +80,7 @@ __crypto_memneq_generic(const void *a, const void *b, size_t size)
 #endif /* CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS */
 	while (size > 0) {
 		neq |= *(unsigned char *)a ^ *(unsigned char *)b;
+		OPTIMIZER_HIDE_VAR(neq);
 		a += 1;
 		b += 1;
 		size -= 1;
@@ -89,33 +91,61 @@ __crypto_memneq_generic(const void *a, const void *b, size_t size)
 /* Loop-free fast-path for frequently used 16-byte size */
 static inline unsigned long __crypto_memneq_16(const void *a, const void *b)
 {
+	unsigned long neq = 0;
+
 #ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
-	if (sizeof(unsigned long) == 8)
-		return ((*(unsigned long *)(a)   ^ *(unsigned long *)(b))
-		      | (*(unsigned long *)(a+8) ^ *(unsigned long *)(b+8)));
-	else if (sizeof(unsigned int) == 4)
-		return ((*(unsigned int *)(a)    ^ *(unsigned int *)(b))
-                      | (*(unsigned int *)(a+4)  ^ *(unsigned int *)(b+4))
-		      | (*(unsigned int *)(a+8)  ^ *(unsigned int *)(b+8))
-	              | (*(unsigned int *)(a+12) ^ *(unsigned int *)(b+12)));
-	else
+	if (sizeof(unsigned long) == 8) {
+		neq |= *(unsigned long *)(a)   ^ *(unsigned long *)(b);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned long *)(a+8) ^ *(unsigned long *)(b+8);
+		OPTIMIZER_HIDE_VAR(neq);
+	} else if (sizeof(unsigned int) == 4) {
+		neq |= *(unsigned int *)(a)    ^ *(unsigned int *)(b);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned int *)(a+4)  ^ *(unsigned int *)(b+4);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned int *)(a+8)  ^ *(unsigned int *)(b+8);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned int *)(a+12) ^ *(unsigned int *)(b+12);
+		OPTIMIZER_HIDE_VAR(neq);
+	} else
 #endif /* CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS */
-		return ((*(unsigned char *)(a)    ^ *(unsigned char *)(b))
-		      | (*(unsigned char *)(a+1)  ^ *(unsigned char *)(b+1))
-		      | (*(unsigned char *)(a+2)  ^ *(unsigned char *)(b+2))
-		      | (*(unsigned char *)(a+3)  ^ *(unsigned char *)(b+3))
-		      | (*(unsigned char *)(a+4)  ^ *(unsigned char *)(b+4))
-		      | (*(unsigned char *)(a+5)  ^ *(unsigned char *)(b+5))
-		      | (*(unsigned char *)(a+6)  ^ *(unsigned char *)(b+6))
-		      | (*(unsigned char *)(a+7)  ^ *(unsigned char *)(b+7))
-		      | (*(unsigned char *)(a+8)  ^ *(unsigned char *)(b+8))
-		      | (*(unsigned char *)(a+9)  ^ *(unsigned char *)(b+9))
-		      | (*(unsigned char *)(a+10) ^ *(unsigned char *)(b+10))
-		      | (*(unsigned char *)(a+11) ^ *(unsigned char *)(b+11))
-		      | (*(unsigned char *)(a+12) ^ *(unsigned char *)(b+12))
-		      | (*(unsigned char *)(a+13) ^ *(unsigned char *)(b+13))
-		      | (*(unsigned char *)(a+14) ^ *(unsigned char *)(b+14))
-		      | (*(unsigned char *)(a+15) ^ *(unsigned char *)(b+15)));
+	{
+		neq |= *(unsigned char *)(a)    ^ *(unsigned char *)(b);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+1)  ^ *(unsigned char *)(b+1);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+2)  ^ *(unsigned char *)(b+2);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+3)  ^ *(unsigned char *)(b+3);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+4)  ^ *(unsigned char *)(b+4);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+5)  ^ *(unsigned char *)(b+5);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+6)  ^ *(unsigned char *)(b+6);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+7)  ^ *(unsigned char *)(b+7);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+8)  ^ *(unsigned char *)(b+8);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+9)  ^ *(unsigned char *)(b+9);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+10) ^ *(unsigned char *)(b+10);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+11) ^ *(unsigned char *)(b+11);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+12) ^ *(unsigned char *)(b+12);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+13) ^ *(unsigned char *)(b+13);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+14) ^ *(unsigned char *)(b+14);
+		OPTIMIZER_HIDE_VAR(neq);
+		neq |= *(unsigned char *)(a+15) ^ *(unsigned char *)(b+15);
+		OPTIMIZER_HIDE_VAR(neq);
+	}
+
+	return neq;
 }
 
 /* Compare two areas of memory without leaking timing information,

crypto/pcrypt.c

@@ -78,7 +78,7 @@ static int pcrypt_do_parallel(struct padata_priv *padata, unsigned int *cb_cpu,
 	cpu = *cb_cpu;
 
 	rcu_read_lock_bh();
-	cpumask = rcu_dereference(pcrypt->cb_cpumask);
+	cpumask = rcu_dereference_bh(pcrypt->cb_cpumask);
 	if (cpumask_test_cpu(cpu, cpumask->mask))
 			goto out;
 

crypto/tcrypt.c

@@ -137,7 +137,272 @@ static int test_cipher_cycles(struct blkcipher_desc *desc, int enc,
 	return ret;
 }
 
+static int test_aead_jiffies(struct aead_request *req, int enc,
+				int blen, int sec)
+{
+	unsigned long start, end;
+	int bcount;
+	int ret;
+
+	for (start = jiffies, end = start + sec * HZ, bcount = 0;
+	     time_before(jiffies, end); bcount++) {
+		if (enc)
+			ret = crypto_aead_encrypt(req);
+		else
+			ret = crypto_aead_decrypt(req);
+
+		if (ret)
+			return ret;
+	}
+
+	printk("%d operations in %d seconds (%ld bytes)\n",
+	       bcount, sec, (long)bcount * blen);
+	return 0;
+}
+
+static int test_aead_cycles(struct aead_request *req, int enc, int blen)
+{
+	unsigned long cycles = 0;
+	int ret = 0;
+	int i;
+
+	local_irq_disable();
+
+	/* Warm-up run. */
+	for (i = 0; i < 4; i++) {
+		if (enc)
+			ret = crypto_aead_encrypt(req);
+		else
+			ret = crypto_aead_decrypt(req);
+
+		if (ret)
+			goto out;
+	}
+
+	/* The real thing. */
+	for (i = 0; i < 8; i++) {
+		cycles_t start, end;
+
+		start = get_cycles();
+		if (enc)
+			ret = crypto_aead_encrypt(req);
+		else
+			ret = crypto_aead_decrypt(req);
+		end = get_cycles();
+
+		if (ret)
+			goto out;
+
+		cycles += end - start;
+	}
+
+out:
+	local_irq_enable();
+
+	if (ret == 0)
+		printk("1 operation in %lu cycles (%d bytes)\n",
+		       (cycles + 4) / 8, blen);
+
+	return ret;
+}
+
 static u32 block_sizes[] = { 16, 64, 256, 1024, 8192, 0 };
+static u32 aead_sizes[] = { 16, 64, 256, 512, 1024, 2048, 4096, 8192, 0 };
+
+#define XBUFSIZE 8
+#define MAX_IVLEN 32
+
+static int testmgr_alloc_buf(char *buf[XBUFSIZE])
+{
+	int i;
+
+	for (i = 0; i < XBUFSIZE; i++) {
+		buf[i] = (void *)__get_free_page(GFP_KERNEL);
+		if (!buf[i])
+			goto err_free_buf;
+	}
+
+	return 0;
+
+err_free_buf:
+	while (i-- > 0)
+		free_page((unsigned long)buf[i]);
+
+	return -ENOMEM;
+}
+
+static void testmgr_free_buf(char *buf[XBUFSIZE])
+{
+	int i;
+
+	for (i = 0; i < XBUFSIZE; i++)
+		free_page((unsigned long)buf[i]);
+}
+
+static void sg_init_aead(struct scatterlist *sg, char *xbuf[XBUFSIZE],
+			unsigned int buflen)
+{
+	int np = (buflen + PAGE_SIZE - 1)/PAGE_SIZE;
+	int k, rem;
+
+	np = (np > XBUFSIZE) ? XBUFSIZE : np;
+	rem = buflen % PAGE_SIZE;
+	if (np > XBUFSIZE) {
+		rem = PAGE_SIZE;
+		np = XBUFSIZE;
+	}
+	sg_init_table(sg, np);
+	for (k = 0; k < np; ++k) {
+		if (k == (np-1))
+			sg_set_buf(&sg[k], xbuf[k], rem);
+		else
+			sg_set_buf(&sg[k], xbuf[k], PAGE_SIZE);
+	}
+}
+
+static void test_aead_speed(const char *algo, int enc, unsigned int sec,
+			    struct aead_speed_template *template,
+			    unsigned int tcount, u8 authsize,
+			    unsigned int aad_size, u8 *keysize)
+{
+	unsigned int i, j;
+	struct crypto_aead *tfm;
+	int ret = -ENOMEM;
+	const char *key;
+	struct aead_request *req;
+	struct scatterlist *sg;
+	struct scatterlist *asg;
+	struct scatterlist *sgout;
+	const char *e;
+	void *assoc;
+	char iv[MAX_IVLEN];
+	char *xbuf[XBUFSIZE];
+	char *xoutbuf[XBUFSIZE];
+	char *axbuf[XBUFSIZE];
+	unsigned int *b_size;
+	unsigned int iv_len;
+
+	if (enc == ENCRYPT)
+		e = "encryption";
+	else
+		e = "decryption";
+
+	if (testmgr_alloc_buf(xbuf))
+		goto out_noxbuf;
+	if (testmgr_alloc_buf(axbuf))
+		goto out_noaxbuf;
+	if (testmgr_alloc_buf(xoutbuf))
+		goto out_nooutbuf;
+
+	sg = kmalloc(sizeof(*sg) * 8 * 3, GFP_KERNEL);
+	if (!sg)
+		goto out_nosg;
+	asg = &sg[8];
+	sgout = &asg[8];
+
+
+	printk(KERN_INFO "\ntesting speed of %s %s\n", algo, e);
+
+	tfm = crypto_alloc_aead(algo, 0, 0);
+
+	if (IS_ERR(tfm)) {
+		pr_err("alg: aead: Failed to load transform for %s: %ld\n", algo,
+		       PTR_ERR(tfm));
+		return;
+	}
+
+	req = aead_request_alloc(tfm, GFP_KERNEL);
+	if (!req) {
+		pr_err("alg: aead: Failed to allocate request for %s\n",
+		       algo);
+		goto out;
+	}
+
+	i = 0;
+	do {
+		b_size = aead_sizes;
+		do {
+			assoc = axbuf[0];
+
+			if (aad_size < PAGE_SIZE)
+				memset(assoc, 0xff, aad_size);
+			else {
+				pr_err("associate data length (%u) too big\n",
+					aad_size);
+				goto out_nosg;
+			}
+			sg_init_one(&asg[0], assoc, aad_size);
+
+			if ((*keysize + *b_size) > TVMEMSIZE * PAGE_SIZE) {
+				pr_err("template (%u) too big for tvmem (%lu)\n",
+				       *keysize + *b_size,
+					TVMEMSIZE * PAGE_SIZE);
+				goto out;
+			}
+
+			key = tvmem[0];
+			for (j = 0; j < tcount; j++) {
+				if (template[j].klen == *keysize) {
+					key = template[j].key;
+					break;
+				}
+			}
+			ret = crypto_aead_setkey(tfm, key, *keysize);
+			ret = crypto_aead_setauthsize(tfm, authsize);
+
+			iv_len = crypto_aead_ivsize(tfm);
+			if (iv_len)
+				memset(&iv, 0xff, iv_len);
+
+			crypto_aead_clear_flags(tfm, ~0);
+			printk(KERN_INFO "test %u (%d bit key, %d byte blocks): ",
+					i, *keysize * 8, *b_size);
+
+
+			memset(tvmem[0], 0xff, PAGE_SIZE);
+
+			if (ret) {
+				pr_err("setkey() failed flags=%x\n",
+						crypto_aead_get_flags(tfm));
+				goto out;
+			}
+
+			sg_init_aead(&sg[0], xbuf,
+				    *b_size + (enc ? authsize : 0));
+
+			sg_init_aead(&sgout[0], xoutbuf,
+				    *b_size + (enc ? authsize : 0));
+
+			aead_request_set_crypt(req, sg, sgout, *b_size, iv);
+			aead_request_set_assoc(req, asg, aad_size);
+
+			if (sec)
+				ret = test_aead_jiffies(req, enc, *b_size, sec);
+			else
+				ret = test_aead_cycles(req, enc, *b_size);
+
+			if (ret) {
+				pr_err("%s() failed return code=%d\n", e, ret);
+				break;
+			}
+			b_size++;
+			i++;
+		} while (*b_size);
+		keysize++;
+	} while (*keysize);
+
+out:
+	crypto_free_aead(tfm);
+	kfree(sg);
+out_nosg:
+	testmgr_free_buf(xoutbuf);
+out_nooutbuf:
+	testmgr_free_buf(axbuf);
+out_noaxbuf:
+	testmgr_free_buf(xbuf);
+out_noxbuf:
+	return;
+}
 
 static void test_cipher_speed(const char *algo, int enc, unsigned int sec,
 			      struct cipher_speed_template *template,
@@ -1427,6 +1692,11 @@ static int do_test(int m)
 				  speed_template_32_64);
 		break;
 
+	case 211:
+		test_aead_speed("rfc4106(gcm(aes))", ENCRYPT, sec,
+				NULL, 0, 16, 8, aead_speed_template_20);
+		break;
+
 	case 300:
 		/* fall through */
 

crypto/tcrypt.h

@@ -22,6 +22,11 @@ struct cipher_speed_template {
 	unsigned int klen;
 };
 
+struct aead_speed_template {
+	const char *key;
+	unsigned int klen;
+};
+
 struct hash_speed {
 	unsigned int blen;	/* buffer length */
 	unsigned int plen;	/* per-update length */
@@ -57,6 +62,11 @@ static u8 speed_template_32_48[] = {32, 48, 0};
 static u8 speed_template_32_48_64[] = {32, 48, 64, 0};
 static u8 speed_template_32_64[] = {32, 64, 0};
 
+/*
+ * AEAD speed tests
+ */
+static u8 aead_speed_template_20[] = {20, 0};
+
 /*
  * Digest speed tests
  */

drivers/crypto/Kconfig

@@ -289,16 +289,6 @@ config CRYPTO_DEV_SAHARA
 	  This option enables support for the SAHARA HW crypto accelerator
 	  found in some Freescale i.MX chips.
 
-config CRYPTO_DEV_DCP
-	tristate "Support for the DCP engine"
-	depends on ARCH_MXS && OF
-	select CRYPTO_BLKCIPHER
-	select CRYPTO_AES
-	select CRYPTO_CBC
-	help
-	  This options enables support for the hardware crypto-acceleration
-	  capabilities of the DCP co-processor
-
 config CRYPTO_DEV_S5P
 	tristate "Support for Samsung S5PV210 crypto accelerator"
 	depends on ARCH_S5PV210
@@ -399,4 +389,33 @@ config CRYPTO_DEV_ATMEL_SHA
 	  To compile this driver as a module, choose M here: the module
 	  will be called atmel-sha.
 
+config CRYPTO_DEV_CCP
+	bool "Support for AMD Cryptographic Coprocessor"
+	depends on X86 && PCI
+	default n
+	help
+	  The AMD Cryptographic Coprocessor provides hardware support
+	  for encryption, hashing and related operations.
+
+if CRYPTO_DEV_CCP
+	source "drivers/crypto/ccp/Kconfig"
+endif
+
+config CRYPTO_DEV_MXS_DCP
+	tristate "Support for Freescale MXS DCP"
+	depends on ARCH_MXS
+	select CRYPTO_SHA1
+	select CRYPTO_SHA256
+	select CRYPTO_CBC
+	select CRYPTO_ECB
+	select CRYPTO_AES
+	select CRYPTO_BLKCIPHER
+	select CRYPTO_ALGAPI
+	help
+	  The Freescale i.MX23/i.MX28 has SHA1/SHA256 and AES128 CBC/ECB
+	  co-processor on the die.
+
+	  To compile this driver as a module, choose M here: the module
+	  will be called mxs-dcp.
+
 endif # CRYPTO_HW

drivers/crypto/Makefile

-obj-$(CONFIG_CRYPTO_DEV_PADLOCK_AES) += padlock-aes.o
-obj-$(CONFIG_CRYPTO_DEV_PADLOCK_SHA) += padlock-sha.o
+obj-$(CONFIG_CRYPTO_DEV_ATMEL_AES) += atmel-aes.o
+obj-$(CONFIG_CRYPTO_DEV_ATMEL_SHA) += atmel-sha.o
+obj-$(CONFIG_CRYPTO_DEV_ATMEL_TDES) += atmel-tdes.o
+obj-$(CONFIG_CRYPTO_DEV_BFIN_CRC) += bfin_crc.o
+obj-$(CONFIG_CRYPTO_DEV_CCP) += ccp/
+obj-$(CONFIG_CRYPTO_DEV_FSL_CAAM) += caam/
 obj-$(CONFIG_CRYPTO_DEV_GEODE) += geode-aes.o
-obj-$(CONFIG_CRYPTO_DEV_NIAGARA2) += n2_crypto.o
-n2_crypto-y := n2_core.o n2_asm.o
 obj-$(CONFIG_CRYPTO_DEV_HIFN_795X) += hifn_795x.o
-obj-$(CONFIG_CRYPTO_DEV_MV_CESA) += mv_cesa.o
-obj-$(CONFIG_CRYPTO_DEV_TALITOS) += talitos.o
-obj-$(CONFIG_CRYPTO_DEV_FSL_CAAM) += caam/
 obj-$(CONFIG_CRYPTO_DEV_IXP4XX) += ixp4xx_crypto.o
-obj-$(CONFIG_CRYPTO_DEV_PPC4XX) += amcc/
-obj-$(CONFIG_CRYPTO_DEV_OMAP_SHAM) += omap-sham.o
+obj-$(CONFIG_CRYPTO_DEV_MV_CESA) += mv_cesa.o
+obj-$(CONFIG_CRYPTO_DEV_MXS_DCP) += mxs-dcp.o
+obj-$(CONFIG_CRYPTO_DEV_NIAGARA2) += n2_crypto.o
+n2_crypto-y := n2_core.o n2_asm.o
+obj-$(CONFIG_CRYPTO_DEV_NX) += nx/
 obj-$(CONFIG_CRYPTO_DEV_OMAP_AES) += omap-aes.o
+obj-$(CONFIG_CRYPTO_DEV_OMAP_SHAM) += omap-sham.o
+obj-$(CONFIG_CRYPTO_DEV_PADLOCK_AES) += padlock-aes.o
+obj-$(CONFIG_CRYPTO_DEV_PADLOCK_SHA) += padlock-sha.o
 obj-$(CONFIG_CRYPTO_DEV_PICOXCELL) += picoxcell_crypto.o
-obj-$(CONFIG_CRYPTO_DEV_SAHARA) += sahara.o
-obj-$(CONFIG_CRYPTO_DEV_DCP) += dcp.o
+obj-$(CONFIG_CRYPTO_DEV_PPC4XX) += amcc/
 obj-$(CONFIG_CRYPTO_DEV_S5P) += s5p-sss.o
+obj-$(CONFIG_CRYPTO_DEV_SAHARA) += sahara.o
+obj-$(CONFIG_CRYPTO_DEV_TALITOS) += talitos.o
 obj-$(CONFIG_CRYPTO_DEV_TEGRA_AES) += tegra-aes.o
 obj-$(CONFIG_CRYPTO_DEV_UX500) += ux500/
-obj-$(CONFIG_CRYPTO_DEV_BFIN_CRC) += bfin_crc.o
-obj-$(CONFIG_CRYPTO_DEV_NX) += nx/
-obj-$(CONFIG_CRYPTO_DEV_ATMEL_AES) += atmel-aes.o
-obj-$(CONFIG_CRYPTO_DEV_ATMEL_TDES) += atmel-tdes.o
-obj-$(CONFIG_CRYPTO_DEV_ATMEL_SHA) += atmel-sha.o

drivers/crypto/amcc/crypto4xx_core.c

@@ -724,7 +724,6 @@ static void crypto4xx_stop_all(struct crypto4xx_core_device *core_dev)
 	crypto4xx_destroy_pdr(core_dev->dev);
 	crypto4xx_destroy_gdr(core_dev->dev);
 	crypto4xx_destroy_sdr(core_dev->dev);
-	dev_set_drvdata(core_dev->device, NULL);
 	iounmap(core_dev->dev->ce_base);
 	kfree(core_dev->dev);
 	kfree(core_dev);

drivers/crypto/caam/caamalg.c

@@ -467,24 +467,10 @@ static int aead_setkey(struct crypto_aead *aead,
 	static const u8 mdpadlen[] = { 16, 20, 32, 32, 64, 64 };
 	struct caam_ctx *ctx = crypto_aead_ctx(aead);
 	struct device *jrdev = ctx->jrdev;
-	struct rtattr *rta = (void *)key;
-	struct crypto_authenc_key_param *param;
-	unsigned int authkeylen;
-	unsigned int enckeylen;
+	struct crypto_authenc_keys keys;
 	int ret = 0;
 
-	param = RTA_DATA(rta);
-	enckeylen = be32_to_cpu(param->enckeylen);
-
-	key += RTA_ALIGN(rta->rta_len);
-	keylen -= RTA_ALIGN(rta->rta_len);
-
-	if (keylen < enckeylen)
-		goto badkey;
-
-	authkeylen = keylen - enckeylen;
-
-	if (keylen > CAAM_MAX_KEY_SIZE)
+	if (crypto_authenc_extractkeys(&keys, key, keylen) != 0)
 		goto badkey;
 
 	/* Pick class 2 key length from algorithm submask */
@@ -492,25 +478,29 @@ static int aead_setkey(struct crypto_aead *aead,
 				      OP_ALG_ALGSEL_SHIFT] * 2;
 	ctx->split_key_pad_len = ALIGN(ctx->split_key_len, 16);
 
+	if (ctx->split_key_pad_len + keys.enckeylen > CAAM_MAX_KEY_SIZE)
+		goto badkey;
+
 #ifdef DEBUG
 	printk(KERN_ERR "keylen %d enckeylen %d authkeylen %d\n",
-	       keylen, enckeylen, authkeylen);
+	       keys.authkeylen + keys.enckeylen, keys.enckeylen,
+	       keys.authkeylen);
 	printk(KERN_ERR "split_key_len %d split_key_pad_len %d\n",
 	       ctx->split_key_len, ctx->split_key_pad_len);
 	print_hex_dump(KERN_ERR, "key in @"__stringify(__LINE__)": ",
 		       DUMP_PREFIX_ADDRESS, 16, 4, key, keylen, 1);
 #endif
 
-	ret = gen_split_aead_key(ctx, key, authkeylen);
+	ret = gen_split_aead_key(ctx, keys.authkey, keys.authkeylen);
 	if (ret) {
 		goto badkey;
 	}
 
 	/* postpend encryption key to auth split key */
-	memcpy(ctx->key + ctx->split_key_pad_len, key + authkeylen, enckeylen);
+	memcpy(ctx->key + ctx->split_key_pad_len, keys.enckey, keys.enckeylen);
 
 	ctx->key_dma = dma_map_single(jrdev, ctx->key, ctx->split_key_pad_len +
-				       enckeylen, DMA_TO_DEVICE);
+				      keys.enckeylen, DMA_TO_DEVICE);
 	if (dma_mapping_error(jrdev, ctx->key_dma)) {
 		dev_err(jrdev, "unable to map key i/o memory\n");
 		return -ENOMEM;
@@ -518,15 +508,15 @@ static int aead_setkey(struct crypto_aead *aead,
 #ifdef DEBUG
 	print_hex_dump(KERN_ERR, "ctx.key@"__stringify(__LINE__)": ",
 		       DUMP_PREFIX_ADDRESS, 16, 4, ctx->key,
-		       ctx->split_key_pad_len + enckeylen, 1);
+		       ctx->split_key_pad_len + keys.enckeylen, 1);
 #endif
 
-	ctx->enckeylen = enckeylen;
+	ctx->enckeylen = keys.enckeylen;
 
 	ret = aead_set_sh_desc(aead);
 	if (ret) {
 		dma_unmap_single(jrdev, ctx->key_dma, ctx->split_key_pad_len +
-				 enckeylen, DMA_TO_DEVICE);
+				 keys.enckeylen, DMA_TO_DEVICE);
 	}
 
 	return ret;

drivers/crypto/ccp/Kconfig

+config CRYPTO_DEV_CCP_DD
+	tristate "Cryptographic Coprocessor device driver"
+	depends on CRYPTO_DEV_CCP
+	default m
+	select HW_RANDOM
+	help
+	  Provides the interface to use the AMD Cryptographic Coprocessor
+	  which can be used to accelerate or offload encryption operations
+	  such as SHA, AES and more. If you choose 'M' here, this module
+	  will be called ccp.
+
+config CRYPTO_DEV_CCP_CRYPTO
+	tristate "Encryption and hashing acceleration support"
+	depends on CRYPTO_DEV_CCP_DD
+	default m
+	select CRYPTO_ALGAPI
+	select CRYPTO_HASH
+	select CRYPTO_BLKCIPHER
+	select CRYPTO_AUTHENC
+	help
+	  Support for using the cryptographic API with the AMD Cryptographic
+	  Coprocessor. This module supports acceleration and offload of SHA
+	  and AES algorithms.  If you choose 'M' here, this module will be
+	  called ccp_crypto.

drivers/crypto/ccp/Makefile

+obj-$(CONFIG_CRYPTO_DEV_CCP_DD) += ccp.o
+ccp-objs := ccp-dev.o ccp-ops.o
+ccp-objs += ccp-pci.o
+
+obj-$(CONFIG_CRYPTO_DEV_CCP_CRYPTO) += ccp-crypto.o
+ccp-crypto-objs := ccp-crypto-main.o \
+		   ccp-crypto-aes.o \
+		   ccp-crypto-aes-cmac.o \
+		   ccp-crypto-aes-xts.o \
+		   ccp-crypto-sha.o

drivers/crypto/ccp/ccp-crypto-aes-cmac.c

+/*
+ * AMD Cryptographic Coprocessor (CCP) AES CMAC crypto API support
+ *
+ * Copyright (C) 2013 Advanced Micro Devices, Inc.
+ *
+ * Author: Tom Lendacky <thomas.lendacky@amd.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/delay.h>
+#include <linux/scatterlist.h>
+#include <linux/crypto.h>
+#include <crypto/algapi.h>
+#include <crypto/aes.h>
+#include <crypto/hash.h>
+#include <crypto/internal/hash.h>
+#include <crypto/scatterwalk.h>
+
+#include "ccp-crypto.h"
+
+
+static int ccp_aes_cmac_complete(struct crypto_async_request *async_req,
+				 int ret)
+{
+	struct ahash_request *req = ahash_request_cast(async_req);
+	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
+	struct ccp_aes_cmac_req_ctx *rctx = ahash_request_ctx(req);
+	unsigned int digest_size = crypto_ahash_digestsize(tfm);
+
+	if (ret)
+		goto e_free;
+
+	if (rctx->hash_rem) {
+		/* Save remaining data to buffer */
+		unsigned int offset = rctx->nbytes - rctx->hash_rem;
+		scatterwalk_map_and_copy(rctx->buf, rctx->src,
+					 offset, rctx->hash_rem, 0);
+		rctx->buf_count = rctx->hash_rem;
+	} else
+		rctx->buf_count = 0;
+
+	/* Update result area if supplied */
+	if (req->result)
+		memcpy(req->result, rctx->iv, digest_size);
+
+e_free:
+	sg_free_table(&rctx->data_sg);
+
+	return ret;
+}
+
+static int ccp_do_cmac_update(struct ahash_request *req, unsigned int nbytes,
+			      unsigned int final)
+{
+	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
+	struct ccp_ctx *ctx = crypto_ahash_ctx(tfm);
+	struct ccp_aes_cmac_req_ctx *rctx = ahash_request_ctx(req);
+	struct scatterlist *sg, *cmac_key_sg = NULL;
+	unsigned int block_size =
+		crypto_tfm_alg_blocksize(crypto_ahash_tfm(tfm));
+	unsigned int need_pad, sg_count;
+	gfp_t gfp;
+	u64 len;
+	int ret;
+
+	if (!ctx->u.aes.key_len)
+		return -EINVAL;
+
+	if (nbytes)
+		rctx->null_msg = 0;
+
+	len = (u64)rctx->buf_count + (u64)nbytes;
+
+	if (!final && (len <= block_size)) {
+		scatterwalk_map_and_copy(rctx->buf + rctx->buf_count, req->src,
+					 0, nbytes, 0);
+		rctx->buf_count += nbytes;
+
+		return 0;
+	}
+
+	rctx->src = req->src;
+	rctx->nbytes = nbytes;
+
+	rctx->final = final;
+	rctx->hash_rem = final ? 0 : len & (block_size - 1);
+	rctx->hash_cnt = len - rctx->hash_rem;
+	if (!final && !rctx->hash_rem) {
+		/* CCP can't do zero length final, so keep some data around */
+		rctx->hash_cnt -= block_size;
+		rctx->hash_rem = block_size;
+	}
+
+	if (final && (rctx->null_msg || (len & (block_size - 1))))
+		need_pad = 1;
+	else
+		need_pad = 0;
+
+	sg_init_one(&rctx->iv_sg, rctx->iv, sizeof(rctx->iv));
+
+	/* Build the data scatterlist table - allocate enough entries for all
+	 * possible data pieces (buffer, input data, padding)
+	 */
+	sg_count = (nbytes) ? sg_nents(req->src) + 2 : 2;
+	gfp = req->base.flags & CRYPTO_TFM_REQ_MAY_SLEEP ?
+		GFP_KERNEL : GFP_ATOMIC;
+	ret = sg_alloc_table(&rctx->data_sg, sg_count, gfp);
+	if (ret)
+		return ret;
+
+	sg = NULL;
+	if (rctx->buf_count) {
+		sg_init_one(&rctx->buf_sg, rctx->buf, rctx->buf_count);
+		sg = ccp_crypto_sg_table_add(&rctx->data_sg, &rctx->buf_sg);
+	}
+
+	if (nbytes)
+		sg = ccp_crypto_sg_table_add(&rctx->data_sg, req->src);
+
+	if (need_pad) {
+		int pad_length = block_size - (len & (block_size - 1));
+
+		rctx->hash_cnt += pad_length;
+
+		memset(rctx->pad, 0, sizeof(rctx->pad));
+		rctx->pad[0] = 0x80;
+		sg_init_one(&rctx->pad_sg, rctx->pad, pad_length);
+		sg = ccp_crypto_sg_table_add(&rctx->data_sg, &rctx->pad_sg);
+	}
+	if (sg) {
+		sg_mark_end(sg);
+		sg = rctx->data_sg.sgl;
+	}
+
+	/* Initialize the K1/K2 scatterlist */
+	if (final)
+		cmac_key_sg = (need_pad) ? &ctx->u.aes.k2_sg
+					 : &ctx->u.aes.k1_sg;
+
+	memset(&rctx->cmd, 0, sizeof(rctx->cmd));
+	INIT_LIST_HEAD(&rctx->cmd.entry);
+	rctx->cmd.engine = CCP_ENGINE_AES;
+	rctx->cmd.u.aes.type = ctx->u.aes.type;
+	rctx->cmd.u.aes.mode = ctx->u.aes.mode;
+	rctx->cmd.u.aes.action = CCP_AES_ACTION_ENCRYPT;
+	rctx->cmd.u.aes.key = &ctx->u.aes.key_sg;
+	rctx->cmd.u.aes.key_len = ctx->u.aes.key_len;
+	rctx->cmd.u.aes.iv = &rctx->iv_sg;
+	rctx->cmd.u.aes.iv_len = AES_BLOCK_SIZE;
+	rctx->cmd.u.aes.src = sg;
+	rctx->cmd.u.aes.src_len = rctx->hash_cnt;
+	rctx->cmd.u.aes.dst = NULL;
+	rctx->cmd.u.aes.cmac_key = cmac_key_sg;
+	rctx->cmd.u.aes.cmac_key_len = ctx->u.aes.kn_len;
+	rctx->cmd.u.aes.cmac_final = final;
+
+	ret = ccp_crypto_enqueue_request(&req->base, &rctx->cmd);
+
+	return ret;
+}
+
+static int ccp_aes_cmac_init(struct ahash_request *req)
+{
+	struct ccp_aes_cmac_req_ctx *rctx = ahash_request_ctx(req);
+
+	memset(rctx, 0, sizeof(*rctx));
+
+	rctx->null_msg = 1;
+
+	return 0;
+}
+
+static int ccp_aes_cmac_update(struct ahash_request *req)
+{
+	return ccp_do_cmac_update(req, req->nbytes, 0);
+}
+
+static int ccp_aes_cmac_final(struct ahash_request *req)
+{
+	return ccp_do_cmac_update(req, 0, 1);
+}
+
+static int ccp_aes_cmac_finup(struct ahash_request *req)
+{
+	return ccp_do_cmac_update(req, req->nbytes, 1);
+}
+
+static int ccp_aes_cmac_digest(struct ahash_request *req)
+{
+	int ret;
+
+	ret = ccp_aes_cmac_init(req);
+	if (ret)
+		return ret;
+
+	return ccp_aes_cmac_finup(req);
+}
+
+static int ccp_aes_cmac_setkey(struct crypto_ahash *tfm, const u8 *key,
+			   unsigned int key_len)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(crypto_ahash_tfm(tfm));
+	struct ccp_crypto_ahash_alg *alg =
+		ccp_crypto_ahash_alg(crypto_ahash_tfm(tfm));
+	u64 k0_hi, k0_lo, k1_hi, k1_lo, k2_hi, k2_lo;
+	u64 rb_hi = 0x00, rb_lo = 0x87;
+	__be64 *gk;
+	int ret;
+
+	switch (key_len) {
+	case AES_KEYSIZE_128:
+		ctx->u.aes.type = CCP_AES_TYPE_128;
+		break;
+	case AES_KEYSIZE_192:
+		ctx->u.aes.type = CCP_AES_TYPE_192;
+		break;
+	case AES_KEYSIZE_256:
+		ctx->u.aes.type = CCP_AES_TYPE_256;
+		break;
+	default:
+		crypto_ahash_set_flags(tfm, CRYPTO_TFM_RES_BAD_KEY_LEN);
+		return -EINVAL;
+	}
+	ctx->u.aes.mode = alg->mode;
+
+	/* Set to zero until complete */
+	ctx->u.aes.key_len = 0;
+
+	/* Set the key for the AES cipher used to generate the keys */
+	ret = crypto_cipher_setkey(ctx->u.aes.tfm_cipher, key, key_len);
+	if (ret)
+		return ret;
+
+	/* Encrypt a block of zeroes - use key area in context */
+	memset(ctx->u.aes.key, 0, sizeof(ctx->u.aes.key));
+	crypto_cipher_encrypt_one(ctx->u.aes.tfm_cipher, ctx->u.aes.key,
+				  ctx->u.aes.key);
+
+	/* Generate K1 and K2 */
+	k0_hi = be64_to_cpu(*((__be64 *)ctx->u.aes.key));
+	k0_lo = be64_to_cpu(*((__be64 *)ctx->u.aes.key + 1));
+
+	k1_hi = (k0_hi << 1) | (k0_lo >> 63);
+	k1_lo = k0_lo << 1;
+	if (ctx->u.aes.key[0] & 0x80) {
+		k1_hi ^= rb_hi;
+		k1_lo ^= rb_lo;
+	}
+	gk = (__be64 *)ctx->u.aes.k1;
+	*gk = cpu_to_be64(k1_hi);
+	gk++;
+	*gk = cpu_to_be64(k1_lo);
+
+	k2_hi = (k1_hi << 1) | (k1_lo >> 63);
+	k2_lo = k1_lo << 1;
+	if (ctx->u.aes.k1[0] & 0x80) {
+		k2_hi ^= rb_hi;
+		k2_lo ^= rb_lo;
+	}
+	gk = (__be64 *)ctx->u.aes.k2;
+	*gk = cpu_to_be64(k2_hi);
+	gk++;
+	*gk = cpu_to_be64(k2_lo);
+
+	ctx->u.aes.kn_len = sizeof(ctx->u.aes.k1);
+	sg_init_one(&ctx->u.aes.k1_sg, ctx->u.aes.k1, sizeof(ctx->u.aes.k1));
+	sg_init_one(&ctx->u.aes.k2_sg, ctx->u.aes.k2, sizeof(ctx->u.aes.k2));
+
+	/* Save the supplied key */
+	memset(ctx->u.aes.key, 0, sizeof(ctx->u.aes.key));
+	memcpy(ctx->u.aes.key, key, key_len);
+	ctx->u.aes.key_len = key_len;
+	sg_init_one(&ctx->u.aes.key_sg, ctx->u.aes.key, key_len);
+
+	return ret;
+}
+
+static int ccp_aes_cmac_cra_init(struct crypto_tfm *tfm)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(tfm);
+	struct crypto_ahash *ahash = __crypto_ahash_cast(tfm);
+	struct crypto_cipher *cipher_tfm;
+
+	ctx->complete = ccp_aes_cmac_complete;
+	ctx->u.aes.key_len = 0;
+
+	crypto_ahash_set_reqsize(ahash, sizeof(struct ccp_aes_cmac_req_ctx));
+
+	cipher_tfm = crypto_alloc_cipher("aes", 0,
+			CRYPTO_ALG_ASYNC | CRYPTO_ALG_NEED_FALLBACK);
+	if (IS_ERR(cipher_tfm)) {
+		pr_warn("could not load aes cipher driver\n");
+		return PTR_ERR(cipher_tfm);
+	}
+	ctx->u.aes.tfm_cipher = cipher_tfm;
+
+	return 0;
+}
+
+static void ccp_aes_cmac_cra_exit(struct crypto_tfm *tfm)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(tfm);
+
+	if (ctx->u.aes.tfm_cipher)
+		crypto_free_cipher(ctx->u.aes.tfm_cipher);
+	ctx->u.aes.tfm_cipher = NULL;
+}
+
+int ccp_register_aes_cmac_algs(struct list_head *head)
+{
+	struct ccp_crypto_ahash_alg *ccp_alg;
+	struct ahash_alg *alg;
+	struct hash_alg_common *halg;
+	struct crypto_alg *base;
+	int ret;
+
+	ccp_alg = kzalloc(sizeof(*ccp_alg), GFP_KERNEL);
+	if (!ccp_alg)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&ccp_alg->entry);
+	ccp_alg->mode = CCP_AES_MODE_CMAC;
+
+	alg = &ccp_alg->alg;
+	alg->init = ccp_aes_cmac_init;
+	alg->update = ccp_aes_cmac_update;
+	alg->final = ccp_aes_cmac_final;
+	alg->finup = ccp_aes_cmac_finup;
+	alg->digest = ccp_aes_cmac_digest;
+	alg->setkey = ccp_aes_cmac_setkey;
+
+	halg = &alg->halg;
+	halg->digestsize = AES_BLOCK_SIZE;
+
+	base = &halg->base;
+	snprintf(base->cra_name, CRYPTO_MAX_ALG_NAME, "cmac(aes)");
+	snprintf(base->cra_driver_name, CRYPTO_MAX_ALG_NAME, "cmac-aes-ccp");
+	base->cra_flags = CRYPTO_ALG_TYPE_AHASH | CRYPTO_ALG_ASYNC |
+			  CRYPTO_ALG_KERN_DRIVER_ONLY |
+			  CRYPTO_ALG_NEED_FALLBACK;
+	base->cra_blocksize = AES_BLOCK_SIZE;
+	base->cra_ctxsize = sizeof(struct ccp_ctx);
+	base->cra_priority = CCP_CRA_PRIORITY;
+	base->cra_type = &crypto_ahash_type;
+	base->cra_init = ccp_aes_cmac_cra_init;
+	base->cra_exit = ccp_aes_cmac_cra_exit;
+	base->cra_module = THIS_MODULE;
+
+	ret = crypto_register_ahash(alg);
+	if (ret) {
+		pr_err("%s ahash algorithm registration error (%d)\n",
+			base->cra_name, ret);
+		kfree(ccp_alg);
+		return ret;
+	}
+
+	list_add(&ccp_alg->entry, head);
+
+	return 0;
+}

drivers/crypto/ccp/ccp-crypto-aes-xts.c

+/*
+ * AMD Cryptographic Coprocessor (CCP) AES XTS crypto API support
+ *
+ * Copyright (C) 2013 Advanced Micro Devices, Inc.
+ *
+ * Author: Tom Lendacky <thomas.lendacky@amd.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/delay.h>
+#include <linux/scatterlist.h>
+#include <linux/crypto.h>
+#include <crypto/algapi.h>
+#include <crypto/aes.h>
+#include <crypto/scatterwalk.h>
+
+#include "ccp-crypto.h"
+
+
+struct ccp_aes_xts_def {
+	const char *name;
+	const char *drv_name;
+};
+
+static struct ccp_aes_xts_def aes_xts_algs[] = {
+	{
+		.name		= "xts(aes)",
+		.drv_name	= "xts-aes-ccp",
+	},
+};
+
+struct ccp_unit_size_map {
+	unsigned int size;
+	u32 value;
+};
+
+static struct ccp_unit_size_map unit_size_map[] = {
+	{
+		.size	= 4096,
+		.value	= CCP_XTS_AES_UNIT_SIZE_4096,
+	},
+	{
+		.size	= 2048,
+		.value	= CCP_XTS_AES_UNIT_SIZE_2048,
+	},
+	{
+		.size	= 1024,
+		.value	= CCP_XTS_AES_UNIT_SIZE_1024,
+	},
+	{
+		.size	= 512,
+		.value	= CCP_XTS_AES_UNIT_SIZE_512,
+	},
+	{
+		.size	= 256,
+		.value	= CCP_XTS_AES_UNIT_SIZE__LAST,
+	},
+	{
+		.size	= 128,
+		.value	= CCP_XTS_AES_UNIT_SIZE__LAST,
+	},
+	{
+		.size	= 64,
+		.value	= CCP_XTS_AES_UNIT_SIZE__LAST,
+	},
+	{
+		.size	= 32,
+		.value	= CCP_XTS_AES_UNIT_SIZE__LAST,
+	},
+	{
+		.size	= 16,
+		.value	= CCP_XTS_AES_UNIT_SIZE_16,
+	},
+	{
+		.size	= 1,
+		.value	= CCP_XTS_AES_UNIT_SIZE__LAST,
+	},
+};
+
+static int ccp_aes_xts_complete(struct crypto_async_request *async_req, int ret)
+{
+	struct ablkcipher_request *req = ablkcipher_request_cast(async_req);
+	struct ccp_aes_req_ctx *rctx = ablkcipher_request_ctx(req);
+
+	if (ret)
+		return ret;
+
+	memcpy(req->info, rctx->iv, AES_BLOCK_SIZE);
+
+	return 0;
+}
+
+static int ccp_aes_xts_setkey(struct crypto_ablkcipher *tfm, const u8 *key,
+			      unsigned int key_len)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(crypto_ablkcipher_tfm(tfm));
+
+	/* Only support 128-bit AES key with a 128-bit Tweak key,
+	 * otherwise use the fallback
+	 */
+	switch (key_len) {
+	case AES_KEYSIZE_128 * 2:
+		memcpy(ctx->u.aes.key, key, key_len);
+		break;
+	}
+	ctx->u.aes.key_len = key_len / 2;
+	sg_init_one(&ctx->u.aes.key_sg, ctx->u.aes.key, key_len);
+
+	return crypto_ablkcipher_setkey(ctx->u.aes.tfm_ablkcipher, key,
+					key_len);
+}
+
+static int ccp_aes_xts_crypt(struct ablkcipher_request *req,
+			     unsigned int encrypt)
+{
+	struct crypto_tfm *tfm =
+		crypto_ablkcipher_tfm(crypto_ablkcipher_reqtfm(req));
+	struct ccp_ctx *ctx = crypto_tfm_ctx(req->base.tfm);
+	struct ccp_aes_req_ctx *rctx = ablkcipher_request_ctx(req);
+	unsigned int unit;
+	int ret;
+
+	if (!ctx->u.aes.key_len)
+		return -EINVAL;
+
+	if (req->nbytes & (AES_BLOCK_SIZE - 1))
+		return -EINVAL;
+
+	if (!req->info)
+		return -EINVAL;
+
+	for (unit = 0; unit < ARRAY_SIZE(unit_size_map); unit++)
+		if (!(req->nbytes & (unit_size_map[unit].size - 1)))
+			break;
+
+	if ((unit_size_map[unit].value == CCP_XTS_AES_UNIT_SIZE__LAST) ||
+	    (ctx->u.aes.key_len != AES_KEYSIZE_128)) {
+		/* Use the fallback to process the request for any
+		 * unsupported unit sizes or key sizes
+		 */
+		ablkcipher_request_set_tfm(req, ctx->u.aes.tfm_ablkcipher);
+		ret = (encrypt) ? crypto_ablkcipher_encrypt(req) :
+				  crypto_ablkcipher_decrypt(req);
+		ablkcipher_request_set_tfm(req, __crypto_ablkcipher_cast(tfm));
+
+		return ret;
+	}
+
+	memcpy(rctx->iv, req->info, AES_BLOCK_SIZE);
+	sg_init_one(&rctx->iv_sg, rctx->iv, AES_BLOCK_SIZE);
+
+	memset(&rctx->cmd, 0, sizeof(rctx->cmd));
+	INIT_LIST_HEAD(&rctx->cmd.entry);
+	rctx->cmd.engine = CCP_ENGINE_XTS_AES_128;
+	rctx->cmd.u.xts.action = (encrypt) ? CCP_AES_ACTION_ENCRYPT
+					   : CCP_AES_ACTION_DECRYPT;
+	rctx->cmd.u.xts.unit_size = unit_size_map[unit].value;
+	rctx->cmd.u.xts.key = &ctx->u.aes.key_sg;
+	rctx->cmd.u.xts.key_len = ctx->u.aes.key_len;
+	rctx->cmd.u.xts.iv = &rctx->iv_sg;
+	rctx->cmd.u.xts.iv_len = AES_BLOCK_SIZE;
+	rctx->cmd.u.xts.src = req->src;
+	rctx->cmd.u.xts.src_len = req->nbytes;
+	rctx->cmd.u.xts.dst = req->dst;
+
+	ret = ccp_crypto_enqueue_request(&req->base, &rctx->cmd);
+
+	return ret;
+}
+
+static int ccp_aes_xts_encrypt(struct ablkcipher_request *req)
+{
+	return ccp_aes_xts_crypt(req, 1);
+}
+
+static int ccp_aes_xts_decrypt(struct ablkcipher_request *req)
+{
+	return ccp_aes_xts_crypt(req, 0);
+}
+
+static int ccp_aes_xts_cra_init(struct crypto_tfm *tfm)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(tfm);
+	struct crypto_ablkcipher *fallback_tfm;
+
+	ctx->complete = ccp_aes_xts_complete;
+	ctx->u.aes.key_len = 0;
+
+	fallback_tfm = crypto_alloc_ablkcipher(tfm->__crt_alg->cra_name, 0,
+					       CRYPTO_ALG_ASYNC |
+					       CRYPTO_ALG_NEED_FALLBACK);
+	if (IS_ERR(fallback_tfm)) {
+		pr_warn("could not load fallback driver %s\n",
+			tfm->__crt_alg->cra_name);
+		return PTR_ERR(fallback_tfm);
+	}
+	ctx->u.aes.tfm_ablkcipher = fallback_tfm;
+
+	tfm->crt_ablkcipher.reqsize = sizeof(struct ccp_aes_req_ctx) +
+				      fallback_tfm->base.crt_ablkcipher.reqsize;
+
+	return 0;
+}
+
+static void ccp_aes_xts_cra_exit(struct crypto_tfm *tfm)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(tfm);
+
+	if (ctx->u.aes.tfm_ablkcipher)
+		crypto_free_ablkcipher(ctx->u.aes.tfm_ablkcipher);
+	ctx->u.aes.tfm_ablkcipher = NULL;
+}
+
+
+static int ccp_register_aes_xts_alg(struct list_head *head,
+				    const struct ccp_aes_xts_def *def)
+{
+	struct ccp_crypto_ablkcipher_alg *ccp_alg;
+	struct crypto_alg *alg;
+	int ret;
+
+	ccp_alg = kzalloc(sizeof(*ccp_alg), GFP_KERNEL);
+	if (!ccp_alg)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&ccp_alg->entry);
+
+	alg = &ccp_alg->alg;
+
+	snprintf(alg->cra_name, CRYPTO_MAX_ALG_NAME, "%s", def->name);
+	snprintf(alg->cra_driver_name, CRYPTO_MAX_ALG_NAME, "%s",
+		 def->drv_name);
+	alg->cra_flags = CRYPTO_ALG_TYPE_ABLKCIPHER | CRYPTO_ALG_ASYNC |
+			 CRYPTO_ALG_KERN_DRIVER_ONLY |
+			 CRYPTO_ALG_NEED_FALLBACK;
+	alg->cra_blocksize = AES_BLOCK_SIZE;
+	alg->cra_ctxsize = sizeof(struct ccp_ctx);
+	alg->cra_priority = CCP_CRA_PRIORITY;
+	alg->cra_type = &crypto_ablkcipher_type;
+	alg->cra_ablkcipher.setkey = ccp_aes_xts_setkey;
+	alg->cra_ablkcipher.encrypt = ccp_aes_xts_encrypt;
+	alg->cra_ablkcipher.decrypt = ccp_aes_xts_decrypt;
+	alg->cra_ablkcipher.min_keysize = AES_MIN_KEY_SIZE * 2;
+	alg->cra_ablkcipher.max_keysize = AES_MAX_KEY_SIZE * 2;
+	alg->cra_ablkcipher.ivsize = AES_BLOCK_SIZE;
+	alg->cra_init = ccp_aes_xts_cra_init;
+	alg->cra_exit = ccp_aes_xts_cra_exit;
+	alg->cra_module = THIS_MODULE;
+
+	ret = crypto_register_alg(alg);
+	if (ret) {
+		pr_err("%s ablkcipher algorithm registration error (%d)\n",
+			alg->cra_name, ret);
+		kfree(ccp_alg);
+		return ret;
+	}
+
+	list_add(&ccp_alg->entry, head);
+
+	return 0;
+}
+
+int ccp_register_aes_xts_algs(struct list_head *head)
+{
+	int i, ret;
+
+	for (i = 0; i < ARRAY_SIZE(aes_xts_algs); i++) {
+		ret = ccp_register_aes_xts_alg(head, &aes_xts_algs[i]);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}

drivers/crypto/ccp/ccp-crypto-aes.c

+/*
+ * AMD Cryptographic Coprocessor (CCP) AES crypto API support
+ *
+ * Copyright (C) 2013 Advanced Micro Devices, Inc.
+ *
+ * Author: Tom Lendacky <thomas.lendacky@amd.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/delay.h>
+#include <linux/scatterlist.h>
+#include <linux/crypto.h>
+#include <crypto/algapi.h>
+#include <crypto/aes.h>
+#include <crypto/ctr.h>
+#include <crypto/scatterwalk.h>
+
+#include "ccp-crypto.h"
+
+
+static int ccp_aes_complete(struct crypto_async_request *async_req, int ret)
+{
+	struct ablkcipher_request *req = ablkcipher_request_cast(async_req);
+	struct ccp_ctx *ctx = crypto_tfm_ctx(req->base.tfm);
+	struct ccp_aes_req_ctx *rctx = ablkcipher_request_ctx(req);
+
+	if (ret)
+		return ret;
+
+	if (ctx->u.aes.mode != CCP_AES_MODE_ECB)
+		memcpy(req->info, rctx->iv, AES_BLOCK_SIZE);
+
+	return 0;
+}
+
+static int ccp_aes_setkey(struct crypto_ablkcipher *tfm, const u8 *key,
+			  unsigned int key_len)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(crypto_ablkcipher_tfm(tfm));
+	struct ccp_crypto_ablkcipher_alg *alg =
+		ccp_crypto_ablkcipher_alg(crypto_ablkcipher_tfm(tfm));
+
+	switch (key_len) {
+	case AES_KEYSIZE_128:
+		ctx->u.aes.type = CCP_AES_TYPE_128;
+		break;
+	case AES_KEYSIZE_192:
+		ctx->u.aes.type = CCP_AES_TYPE_192;
+		break;
+	case AES_KEYSIZE_256:
+		ctx->u.aes.type = CCP_AES_TYPE_256;
+		break;
+	default:
+		crypto_ablkcipher_set_flags(tfm, CRYPTO_TFM_RES_BAD_KEY_LEN);
+		return -EINVAL;
+	}
+	ctx->u.aes.mode = alg->mode;
+	ctx->u.aes.key_len = key_len;
+
+	memcpy(ctx->u.aes.key, key, key_len);
+	sg_init_one(&ctx->u.aes.key_sg, ctx->u.aes.key, key_len);
+
+	return 0;
+}
+
+static int ccp_aes_crypt(struct ablkcipher_request *req, bool encrypt)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(req->base.tfm);
+	struct ccp_aes_req_ctx *rctx = ablkcipher_request_ctx(req);
+	struct scatterlist *iv_sg = NULL;
+	unsigned int iv_len = 0;
+	int ret;
+
+	if (!ctx->u.aes.key_len)
+		return -EINVAL;
+
+	if (((ctx->u.aes.mode == CCP_AES_MODE_ECB) ||
+	     (ctx->u.aes.mode == CCP_AES_MODE_CBC) ||
+	     (ctx->u.aes.mode == CCP_AES_MODE_CFB)) &&
+	    (req->nbytes & (AES_BLOCK_SIZE - 1)))
+		return -EINVAL;
+
+	if (ctx->u.aes.mode != CCP_AES_MODE_ECB) {
+		if (!req->info)
+			return -EINVAL;
+
+		memcpy(rctx->iv, req->info, AES_BLOCK_SIZE);
+		iv_sg = &rctx->iv_sg;
+		iv_len = AES_BLOCK_SIZE;
+		sg_init_one(iv_sg, rctx->iv, iv_len);
+	}
+
+	memset(&rctx->cmd, 0, sizeof(rctx->cmd));
+	INIT_LIST_HEAD(&rctx->cmd.entry);
+	rctx->cmd.engine = CCP_ENGINE_AES;
+	rctx->cmd.u.aes.type = ctx->u.aes.type;
+	rctx->cmd.u.aes.mode = ctx->u.aes.mode;
+	rctx->cmd.u.aes.action =
+		(encrypt) ? CCP_AES_ACTION_ENCRYPT : CCP_AES_ACTION_DECRYPT;
+	rctx->cmd.u.aes.key = &ctx->u.aes.key_sg;
+	rctx->cmd.u.aes.key_len = ctx->u.aes.key_len;
+	rctx->cmd.u.aes.iv = iv_sg;
+	rctx->cmd.u.aes.iv_len = iv_len;
+	rctx->cmd.u.aes.src = req->src;
+	rctx->cmd.u.aes.src_len = req->nbytes;
+	rctx->cmd.u.aes.dst = req->dst;
+
+	ret = ccp_crypto_enqueue_request(&req->base, &rctx->cmd);
+
+	return ret;
+}
+
+static int ccp_aes_encrypt(struct ablkcipher_request *req)
+{
+	return ccp_aes_crypt(req, true);
+}
+
+static int ccp_aes_decrypt(struct ablkcipher_request *req)
+{
+	return ccp_aes_crypt(req, false);
+}
+
+static int ccp_aes_cra_init(struct crypto_tfm *tfm)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(tfm);
+
+	ctx->complete = ccp_aes_complete;
+	ctx->u.aes.key_len = 0;
+
+	tfm->crt_ablkcipher.reqsize = sizeof(struct ccp_aes_req_ctx);
+
+	return 0;
+}
+
+static void ccp_aes_cra_exit(struct crypto_tfm *tfm)
+{
+}
+
+static int ccp_aes_rfc3686_complete(struct crypto_async_request *async_req,
+				    int ret)
+{
+	struct ablkcipher_request *req = ablkcipher_request_cast(async_req);
+	struct ccp_aes_req_ctx *rctx = ablkcipher_request_ctx(req);
+
+	/* Restore the original pointer */
+	req->info = rctx->rfc3686_info;
+
+	return ccp_aes_complete(async_req, ret);
+}
+
+static int ccp_aes_rfc3686_setkey(struct crypto_ablkcipher *tfm, const u8 *key,
+				  unsigned int key_len)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(crypto_ablkcipher_tfm(tfm));
+
+	if (key_len < CTR_RFC3686_NONCE_SIZE)
+		return -EINVAL;
+
+	key_len -= CTR_RFC3686_NONCE_SIZE;
+	memcpy(ctx->u.aes.nonce, key + key_len, CTR_RFC3686_NONCE_SIZE);
+
+	return ccp_aes_setkey(tfm, key, key_len);
+}
+
+static int ccp_aes_rfc3686_crypt(struct ablkcipher_request *req, bool encrypt)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(req->base.tfm);
+	struct ccp_aes_req_ctx *rctx = ablkcipher_request_ctx(req);
+	u8 *iv;
+
+	/* Initialize the CTR block */
+	iv = rctx->rfc3686_iv;
+	memcpy(iv, ctx->u.aes.nonce, CTR_RFC3686_NONCE_SIZE);
+
+	iv += CTR_RFC3686_NONCE_SIZE;
+	memcpy(iv, req->info, CTR_RFC3686_IV_SIZE);
+
+	iv += CTR_RFC3686_IV_SIZE;
+	*(__be32 *)iv = cpu_to_be32(1);
+
+	/* Point to the new IV */
+	rctx->rfc3686_info = req->info;
+	req->info = rctx->rfc3686_iv;
+
+	return ccp_aes_crypt(req, encrypt);
+}
+
+static int ccp_aes_rfc3686_encrypt(struct ablkcipher_request *req)
+{
+	return ccp_aes_rfc3686_crypt(req, true);
+}
+
+static int ccp_aes_rfc3686_decrypt(struct ablkcipher_request *req)
+{
+	return ccp_aes_rfc3686_crypt(req, false);
+}
+
+static int ccp_aes_rfc3686_cra_init(struct crypto_tfm *tfm)
+{
+	struct ccp_ctx *ctx = crypto_tfm_ctx(tfm);
+
+	ctx->complete = ccp_aes_rfc3686_complete;
+	ctx->u.aes.key_len = 0;
+
+	tfm->crt_ablkcipher.reqsize = sizeof(struct ccp_aes_req_ctx);
+
+	return 0;
+}
+
+static void ccp_aes_rfc3686_cra_exit(struct crypto_tfm *tfm)
+{
+}
+
+static struct crypto_alg ccp_aes_defaults = {
+	.cra_flags	= CRYPTO_ALG_TYPE_ABLKCIPHER |
+			  CRYPTO_ALG_ASYNC |
+			  CRYPTO_ALG_KERN_DRIVER_ONLY |
+			  CRYPTO_ALG_NEED_FALLBACK,
+	.cra_blocksize	= AES_BLOCK_SIZE,
+	.cra_ctxsize	= sizeof(struct ccp_ctx),
+	.cra_priority	= CCP_CRA_PRIORITY,
+	.cra_type	= &crypto_ablkcipher_type,
+	.cra_init	= ccp_aes_cra_init,
+	.cra_exit	= ccp_aes_cra_exit,
+	.cra_module	= THIS_MODULE,
+	.cra_ablkcipher	= {
+		.setkey		= ccp_aes_setkey,
+		.encrypt	= ccp_aes_encrypt,
+		.decrypt	= ccp_aes_decrypt,
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+	},
+};
+
+static struct crypto_alg ccp_aes_rfc3686_defaults = {
+	.cra_flags	= CRYPTO_ALG_TYPE_ABLKCIPHER |
+			   CRYPTO_ALG_ASYNC |
+			   CRYPTO_ALG_KERN_DRIVER_ONLY |
+			   CRYPTO_ALG_NEED_FALLBACK,
+	.cra_blocksize	= CTR_RFC3686_BLOCK_SIZE,
+	.cra_ctxsize	= sizeof(struct ccp_ctx),
+	.cra_priority	= CCP_CRA_PRIORITY,
+	.cra_type	= &crypto_ablkcipher_type,
+	.cra_init	= ccp_aes_rfc3686_cra_init,
+	.cra_exit	= ccp_aes_rfc3686_cra_exit,
+	.cra_module	= THIS_MODULE,
+	.cra_ablkcipher	= {
+		.setkey		= ccp_aes_rfc3686_setkey,
+		.encrypt	= ccp_aes_rfc3686_encrypt,
+		.decrypt	= ccp_aes_rfc3686_decrypt,
+		.min_keysize	= AES_MIN_KEY_SIZE + CTR_RFC3686_NONCE_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE + CTR_RFC3686_NONCE_SIZE,
+	},
+};
+
+struct ccp_aes_def {
+	enum ccp_aes_mode mode;
+	const char *name;
+	const char *driver_name;
+	unsigned int blocksize;
+	unsigned int ivsize;
+	struct crypto_alg *alg_defaults;
+};
+
+static struct ccp_aes_def aes_algs[] = {
+	{
+		.mode		= CCP_AES_MODE_ECB,
+		.name		= "ecb(aes)",
+		.driver_name	= "ecb-aes-ccp",
+		.blocksize	= AES_BLOCK_SIZE,
+		.ivsize		= 0,
+		.alg_defaults	= &ccp_aes_defaults,
+	},
+	{
+		.mode		= CCP_AES_MODE_CBC,
+		.name		= "cbc(aes)",
+		.driver_name	= "cbc-aes-ccp",
+		.blocksize	= AES_BLOCK_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.alg_defaults	= &ccp_aes_defaults,
+	},
+	{
+		.mode		= CCP_AES_MODE_CFB,
+		.name		= "cfb(aes)",
+		.driver_name	= "cfb-aes-ccp",
+		.blocksize	= AES_BLOCK_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.alg_defaults	= &ccp_aes_defaults,
+	},
+	{
+		.mode		= CCP_AES_MODE_OFB,
+		.name		= "ofb(aes)",
+		.driver_name	= "ofb-aes-ccp",
+		.blocksize	= 1,
+		.ivsize		= AES_BLOCK_SIZE,
+		.alg_defaults	= &ccp_aes_defaults,
+	},
+	{
+		.mode		= CCP_AES_MODE_CTR,
+		.name		= "ctr(aes)",
+		.driver_name	= "ctr-aes-ccp",
+		.blocksize	= 1,
+		.ivsize		= AES_BLOCK_SIZE,
+		.alg_defaults	= &ccp_aes_defaults,
+	},
+	{
+		.mode		= CCP_AES_MODE_CTR,
+		.name		= "rfc3686(ctr(aes))",
+		.driver_name	= "rfc3686-ctr-aes-ccp",
+		.blocksize	= 1,
+		.ivsize		= CTR_RFC3686_IV_SIZE,
+		.alg_defaults	= &ccp_aes_rfc3686_defaults,
+	},
+};
+
+static int ccp_register_aes_alg(struct list_head *head,
+				const struct ccp_aes_def *def)
+{
+	struct ccp_crypto_ablkcipher_alg *ccp_alg;
+	struct crypto_alg *alg;
+	int ret;
+
+	ccp_alg = kzalloc(sizeof(*ccp_alg), GFP_KERNEL);
+	if (!ccp_alg)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&ccp_alg->entry);
+
+	ccp_alg->mode = def->mode;
+
+	/* Copy the defaults and override as necessary */
+	alg = &ccp_alg->alg;
+	*alg = *def->alg_defaults;
+	snprintf(alg->cra_name, CRYPTO_MAX_ALG_NAME, "%s", def->name);
+	snprintf(alg->cra_driver_name, CRYPTO_MAX_ALG_NAME, "%s",
+		 def->driver_name);
+	alg->cra_blocksize = def->blocksize;
+	alg->cra_ablkcipher.ivsize = def->ivsize;
+
+	ret = crypto_register_alg(alg);
+	if (ret) {
+		pr_err("%s ablkcipher algorithm registration error (%d)\n",
+			alg->cra_name, ret);
+		kfree(ccp_alg);
+		return ret;
+	}
+
+	list_add(&ccp_alg->entry, head);
+
+	return 0;
+}
+
+int ccp_register_aes_algs(struct list_head *head)
+{
+	int i, ret;
+
+	for (i = 0; i < ARRAY_SIZE(aes_algs); i++) {
+		ret = ccp_register_aes_alg(head, &aes_algs[i]);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}

drivers/crypto/ccp/ccp-crypto.h

+/*
+ * AMD Cryptographic Coprocessor (CCP) crypto API support
+ *
+ * Copyright (C) 2013 Advanced Micro Devices, Inc.
+ *
+ * Author: Tom Lendacky <thomas.lendacky@amd.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef __CCP_CRYPTO_H__
+#define __CCP_CRYPTO_H__
+
+
+#include <linux/list.h>
+#include <linux/wait.h>
+#include <linux/pci.h>
+#include <linux/ccp.h>
+#include <linux/crypto.h>
+#include <crypto/algapi.h>
+#include <crypto/aes.h>
+#include <crypto/ctr.h>
+#include <crypto/hash.h>
+#include <crypto/sha.h>
+
+
+#define CCP_CRA_PRIORITY	300
+
+struct ccp_crypto_ablkcipher_alg {
+	struct list_head entry;
+
+	u32 mode;
+
+	struct crypto_alg alg;
+};
+
+struct ccp_crypto_ahash_alg {
+	struct list_head entry;
+
+	const __be32 *init;
+	u32 type;
+	u32 mode;
+
+	/* Child algorithm used for HMAC, CMAC, etc */
+	char child_alg[CRYPTO_MAX_ALG_NAME];
+
+	struct ahash_alg alg;
+};
+
+static inline struct ccp_crypto_ablkcipher_alg *
+	ccp_crypto_ablkcipher_alg(struct crypto_tfm *tfm)
+{
+	struct crypto_alg *alg = tfm->__crt_alg;
+
+	return container_of(alg, struct ccp_crypto_ablkcipher_alg, alg);
+}
+
+static inline struct ccp_crypto_ahash_alg *
+	ccp_crypto_ahash_alg(struct crypto_tfm *tfm)
+{
+	struct crypto_alg *alg = tfm->__crt_alg;
+	struct ahash_alg *ahash_alg;
+
+	ahash_alg = container_of(alg, struct ahash_alg, halg.base);
+
+	return container_of(ahash_alg, struct ccp_crypto_ahash_alg, alg);
+}
+
+
+/***** AES related defines *****/
+struct ccp_aes_ctx {
+	/* Fallback cipher for XTS with unsupported unit sizes */
+	struct crypto_ablkcipher *tfm_ablkcipher;
+
+	/* Cipher used to generate CMAC K1/K2 keys */
+	struct crypto_cipher *tfm_cipher;
+
+	enum ccp_engine engine;
+	enum ccp_aes_type type;
+	enum ccp_aes_mode mode;
+
+	struct scatterlist key_sg;
+	unsigned int key_len;
+	u8 key[AES_MAX_KEY_SIZE];
+
+	u8 nonce[CTR_RFC3686_NONCE_SIZE];
+
+	/* CMAC key structures */
+	struct scatterlist k1_sg;
+	struct scatterlist k2_sg;
+	unsigned int kn_len;
+	u8 k1[AES_BLOCK_SIZE];
+	u8 k2[AES_BLOCK_SIZE];
+};
+
+struct ccp_aes_req_ctx {
+	struct scatterlist iv_sg;
+	u8 iv[AES_BLOCK_SIZE];
+
+	/* Fields used for RFC3686 requests */
+	u8 *rfc3686_info;
+	u8 rfc3686_iv[AES_BLOCK_SIZE];
+
+	struct ccp_cmd cmd;
+};
+
+struct ccp_aes_cmac_req_ctx {
+	unsigned int null_msg;
+	unsigned int final;
+
+	struct scatterlist *src;
+	unsigned int nbytes;
+
+	u64 hash_cnt;
+	unsigned int hash_rem;
+
+	struct sg_table data_sg;
+
+	struct scatterlist iv_sg;
+	u8 iv[AES_BLOCK_SIZE];
+
+	struct scatterlist buf_sg;
+	unsigned int buf_count;
+	u8 buf[AES_BLOCK_SIZE];
+
+	struct scatterlist pad_sg;
+	unsigned int pad_count;
+	u8 pad[AES_BLOCK_SIZE];
+
+	struct ccp_cmd cmd;
+};
+
+/***** SHA related defines *****/
+#define MAX_SHA_CONTEXT_SIZE	SHA256_DIGEST_SIZE
+#define MAX_SHA_BLOCK_SIZE	SHA256_BLOCK_SIZE
+
+struct ccp_sha_ctx {
+	unsigned int key_len;
+	u8 key[MAX_SHA_BLOCK_SIZE];
+	u8 ipad[MAX_SHA_BLOCK_SIZE];
+	u8 opad[MAX_SHA_BLOCK_SIZE];
+	struct crypto_ahash *hmac_tfm;
+};
+
+struct ccp_sha_req_ctx {
+	enum ccp_sha_type type;
+
+	u64 msg_bits;
+
+	unsigned int first;
+	unsigned int final;
+
+	struct scatterlist *src;
+	unsigned int nbytes;
+
+	u64 hash_cnt;
+	unsigned int hash_rem;
+
+	struct sg_table data_sg;
+
+	struct scatterlist ctx_sg;
+	u8 ctx[MAX_SHA_CONTEXT_SIZE];
+
+	struct scatterlist buf_sg;
+	unsigned int buf_count;
+	u8 buf[MAX_SHA_BLOCK_SIZE];
+
+	/* HMAC support field */
+	struct scatterlist pad_sg;
+
+	/* CCP driver command */
+	struct ccp_cmd cmd;
+};
+
+/***** Common Context Structure *****/
+struct ccp_ctx {
+	int (*complete)(struct crypto_async_request *req, int ret);
+
+	union {
+		struct ccp_aes_ctx aes;
+		struct ccp_sha_ctx sha;
+	} u;
+};
+
+int ccp_crypto_enqueue_request(struct crypto_async_request *req,
+			       struct ccp_cmd *cmd);
+struct scatterlist *ccp_crypto_sg_table_add(struct sg_table *table,
+					    struct scatterlist *sg_add);
+
+int ccp_register_aes_algs(struct list_head *head);
+int ccp_register_aes_cmac_algs(struct list_head *head);
+int ccp_register_aes_xts_algs(struct list_head *head);
+int ccp_register_sha_algs(struct list_head *head);
+
+#endif

drivers/crypto/ccp/ccp-dev.h

+/*
+ * AMD Cryptographic Coprocessor (CCP) driver
+ *
+ * Copyright (C) 2013 Advanced Micro Devices, Inc.
+ *
+ * Author: Tom Lendacky <thomas.lendacky@amd.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef __CCP_DEV_H__
+#define __CCP_DEV_H__
+
+#include <linux/device.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include <linux/mutex.h>
+#include <linux/list.h>
+#include <linux/wait.h>
+#include <linux/dmapool.h>
+#include <linux/hw_random.h>
+
+
+#define IO_OFFSET			0x20000
+
+#define MAX_DMAPOOL_NAME_LEN		32
+
+#define MAX_HW_QUEUES			5
+#define MAX_CMD_QLEN			100
+
+#define TRNG_RETRIES			10
+
+
+/****** Register Mappings ******/
+#define Q_MASK_REG			0x000
+#define TRNG_OUT_REG			0x00c
+#define IRQ_MASK_REG			0x040
+#define IRQ_STATUS_REG			0x200
+
+#define DEL_CMD_Q_JOB			0x124
+#define DEL_Q_ACTIVE			0x00000200
+#define DEL_Q_ID_SHIFT			6
+
+#define CMD_REQ0			0x180
+#define CMD_REQ_INCR			0x04
+
+#define CMD_Q_STATUS_BASE		0x210
+#define CMD_Q_INT_STATUS_BASE		0x214
+#define CMD_Q_STATUS_INCR		0x20
+
+#define CMD_Q_CACHE			0x228
+#define CMD_Q_CACHE_INC			0x20
+
+#define CMD_Q_ERROR(__qs)		((__qs) & 0x0000003f);
+#define CMD_Q_DEPTH(__qs)		(((__qs) >> 12) & 0x0000000f);
+
+/****** REQ0 Related Values ******/
+#define REQ0_WAIT_FOR_WRITE		0x00000004
+#define REQ0_INT_ON_COMPLETE		0x00000002
+#define REQ0_STOP_ON_COMPLETE		0x00000001
+
+#define REQ0_CMD_Q_SHIFT		9
+#define REQ0_JOBID_SHIFT		3
+
+/****** REQ1 Related Values ******/
+#define REQ1_PROTECT_SHIFT		27
+#define REQ1_ENGINE_SHIFT		23
+#define REQ1_KEY_KSB_SHIFT		2
+
+#define REQ1_EOM			0x00000002
+#define REQ1_INIT			0x00000001
+
+/* AES Related Values */
+#define REQ1_AES_TYPE_SHIFT		21
+#define REQ1_AES_MODE_SHIFT		18
+#define REQ1_AES_ACTION_SHIFT		17
+#define REQ1_AES_CFB_SIZE_SHIFT		10
+
+/* XTS-AES Related Values */
+#define REQ1_XTS_AES_SIZE_SHIFT		10
+
+/* SHA Related Values */
+#define REQ1_SHA_TYPE_SHIFT		21
+
+/* RSA Related Values */
+#define REQ1_RSA_MOD_SIZE_SHIFT		10
+
+/* Pass-Through Related Values */
+#define REQ1_PT_BW_SHIFT		12
+#define REQ1_PT_BS_SHIFT		10
+
+/* ECC Related Values */
+#define REQ1_ECC_AFFINE_CONVERT		0x00200000
+#define REQ1_ECC_FUNCTION_SHIFT		18
+
+/****** REQ4 Related Values ******/
+#define REQ4_KSB_SHIFT			18
+#define REQ4_MEMTYPE_SHIFT		16
+
+/****** REQ6 Related Values ******/
+#define REQ6_MEMTYPE_SHIFT		16
+
+
+/****** Key Storage Block ******/
+#define KSB_START			77
+#define KSB_END				127
+#define KSB_COUNT			(KSB_END - KSB_START + 1)
+#define CCP_KSB_BITS			256
+#define CCP_KSB_BYTES			32
+
+#define CCP_JOBID_MASK			0x0000003f
+
+#define CCP_DMAPOOL_MAX_SIZE		64
+#define CCP_DMAPOOL_ALIGN		(1 << 5)
+
+#define CCP_REVERSE_BUF_SIZE		64
+
+#define CCP_AES_KEY_KSB_COUNT		1
+#define CCP_AES_CTX_KSB_COUNT		1
+
+#define CCP_XTS_AES_KEY_KSB_COUNT	1
+#define CCP_XTS_AES_CTX_KSB_COUNT	1
+
+#define CCP_SHA_KSB_COUNT		1
+
+#define CCP_RSA_MAX_WIDTH		4096
+
+#define CCP_PASSTHRU_BLOCKSIZE		256
+#define CCP_PASSTHRU_MASKSIZE		32
+#define CCP_PASSTHRU_KSB_COUNT		1
+
+#define CCP_ECC_MODULUS_BYTES		48      /* 384-bits */
+#define CCP_ECC_MAX_OPERANDS		6
+#define CCP_ECC_MAX_OUTPUTS		3
+#define CCP_ECC_SRC_BUF_SIZE		448
+#define CCP_ECC_DST_BUF_SIZE		192
+#define CCP_ECC_OPERAND_SIZE		64
+#define CCP_ECC_OUTPUT_SIZE		64
+#define CCP_ECC_RESULT_OFFSET		60
+#define CCP_ECC_RESULT_SUCCESS		0x0001
+
+
+struct ccp_device;
+struct ccp_cmd;
+
+struct ccp_cmd_queue {
+	struct ccp_device *ccp;
+
+	/* Queue identifier */
+	u32 id;
+
+	/* Queue dma pool */
+	struct dma_pool *dma_pool;
+
+	/* Queue reserved KSB regions */
+	u32 ksb_key;
+	u32 ksb_ctx;
+
+	/* Queue processing thread */
+	struct task_struct *kthread;
+	unsigned int active;
+	unsigned int suspended;
+
+	/* Number of free command slots available */
+	unsigned int free_slots;
+
+	/* Interrupt masks */
+	u32 int_ok;
+	u32 int_err;
+
+	/* Register addresses for queue */
+	void __iomem *reg_status;
+	void __iomem *reg_int_status;
+
+	/* Status values from job */
+	u32 int_status;
+	u32 q_status;
+	u32 q_int_status;
+	u32 cmd_error;
+
+	/* Interrupt wait queue */
+	wait_queue_head_t int_queue;
+	unsigned int int_rcvd;
+} ____cacheline_aligned;
+
+struct ccp_device {
+	struct device *dev;
+
+	/*
+	 * Bus specific device information
+	 */
+	void *dev_specific;
+	int (*get_irq)(struct ccp_device *ccp);
+	void (*free_irq)(struct ccp_device *ccp);
+
+	/*
+	 * I/O area used for device communication. The register mapping
+	 * starts at an offset into the mapped bar.
+	 *   The CMD_REQx registers and the Delete_Cmd_Queue_Job register
+	 *   need to be protected while a command queue thread is accessing
+	 *   them.
+	 */
+	struct mutex req_mutex ____cacheline_aligned;
+	void __iomem *io_map;
+	void __iomem *io_regs;
+
+	/*
+	 * Master lists that all cmds are queued on. Because there can be
+	 * more than one CCP command queue that can process a cmd a separate
+	 * backlog list is neeeded so that the backlog completion call
+	 * completes before the cmd is available for execution.
+	 */
+	spinlock_t cmd_lock ____cacheline_aligned;
+	unsigned int cmd_count;
+	struct list_head cmd;
+	struct list_head backlog;
+
+	/*
+	 * The command queues. These represent the queues available on the
+	 * CCP that are available for processing cmds
+	 */
+	struct ccp_cmd_queue cmd_q[MAX_HW_QUEUES];
+	unsigned int cmd_q_count;
+
+	/*
+	 * Support for the CCP True RNG
+	 */
+	struct hwrng hwrng;
+	unsigned int hwrng_retries;
+
+	/*
+	 * A counter used to generate job-ids for cmds submitted to the CCP
+	 */
+	atomic_t current_id ____cacheline_aligned;
+
+	/*
+	 * The CCP uses key storage blocks (KSB) to maintain context for certain
+	 * operations. To prevent multiple cmds from using the same KSB range
+	 * a command queue reserves a KSB range for the duration of the cmd.
+	 * Each queue, will however, reserve 2 KSB blocks for operations that
+	 * only require single KSB entries (eg. AES context/iv and key) in order
+	 * to avoid allocation contention.  This will reserve at most 10 KSB
+	 * entries, leaving 40 KSB entries available for dynamic allocation.
+	 */
+	struct mutex ksb_mutex ____cacheline_aligned;
+	DECLARE_BITMAP(ksb, KSB_COUNT);
+	wait_queue_head_t ksb_queue;
+	unsigned int ksb_avail;
+	unsigned int ksb_count;
+	u32 ksb_start;
+
+	/* Suspend support */
+	unsigned int suspending;
+	wait_queue_head_t suspend_queue;
+};
+
+
+int ccp_pci_init(void);
+void ccp_pci_exit(void);
+
+struct ccp_device *ccp_alloc_struct(struct device *dev);
+int ccp_init(struct ccp_device *ccp);
+void ccp_destroy(struct ccp_device *ccp);
+bool ccp_queues_suspended(struct ccp_device *ccp);
+
+irqreturn_t ccp_irq_handler(int irq, void *data);
+
+int ccp_run_cmd(struct ccp_cmd_queue *cmd_q, struct ccp_cmd *cmd);
+
+#endif

drivers/crypto/ccp/ccp-pci.c

+/*
+ * AMD Cryptographic Coprocessor (CCP) driver
+ *
+ * Copyright (C) 2013 Advanced Micro Devices, Inc.
+ *
+ * Author: Tom Lendacky <thomas.lendacky@amd.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/pci.h>
+#include <linux/pci_ids.h>
+#include <linux/kthread.h>
+#include <linux/sched.h>
+#include <linux/interrupt.h>
+#include <linux/spinlock.h>
+#include <linux/delay.h>
+#include <linux/ccp.h>
+
+#include "ccp-dev.h"
+
+#define IO_BAR				2
+#define MSIX_VECTORS			2
+
+struct ccp_msix {
+	u32 vector;
+	char name[16];
+};
+
+struct ccp_pci {
+	int msix_count;
+	struct ccp_msix msix[MSIX_VECTORS];
+};
+
+static int ccp_get_msix_irqs(struct ccp_device *ccp)
+{
+	struct ccp_pci *ccp_pci = ccp->dev_specific;
+	struct device *dev = ccp->dev;
+	struct pci_dev *pdev = container_of(dev, struct pci_dev, dev);
+	struct msix_entry msix_entry[MSIX_VECTORS];
+	unsigned int name_len = sizeof(ccp_pci->msix[0].name) - 1;
+	int v, ret;
+
+	for (v = 0; v < ARRAY_SIZE(msix_entry); v++)
+		msix_entry[v].entry = v;
+
+	while ((ret = pci_enable_msix(pdev, msix_entry, v)) > 0)
+		v = ret;
+	if (ret)
+		return ret;
+
+	ccp_pci->msix_count = v;
+	for (v = 0; v < ccp_pci->msix_count; v++) {
+		/* Set the interrupt names and request the irqs */
+		snprintf(ccp_pci->msix[v].name, name_len, "ccp-%u", v);
+		ccp_pci->msix[v].vector = msix_entry[v].vector;
+		ret = request_irq(ccp_pci->msix[v].vector, ccp_irq_handler,
+				  0, ccp_pci->msix[v].name, dev);
+		if (ret) {
+			dev_notice(dev, "unable to allocate MSI-X IRQ (%d)\n",
+				   ret);
+			goto e_irq;
+		}
+	}
+
+	return 0;
+
+e_irq:
+	while (v--)
+		free_irq(ccp_pci->msix[v].vector, dev);
+
+	pci_disable_msix(pdev);
+
+	ccp_pci->msix_count = 0;
+
+	return ret;
+}
+
+static int ccp_get_msi_irq(struct ccp_device *ccp)
+{
+	struct device *dev = ccp->dev;
+	struct pci_dev *pdev = container_of(dev, struct pci_dev, dev);
+	int ret;
+
+	ret = pci_enable_msi(pdev);
+	if (ret)
+		return ret;
+
+	ret = request_irq(pdev->irq, ccp_irq_handler, 0, "ccp", dev);
+	if (ret) {
+		dev_notice(dev, "unable to allocate MSI IRQ (%d)\n", ret);
+		goto e_msi;
+	}
+
+	return 0;
+
+e_msi:
+	pci_disable_msi(pdev);
+
+	return ret;
+}
+
+static int ccp_get_irqs(struct ccp_device *ccp)
+{
+	struct device *dev = ccp->dev;
+	int ret;
+
+	ret = ccp_get_msix_irqs(ccp);
+	if (!ret)
+		return 0;
+
+	/* Couldn't get MSI-X vectors, try MSI */
+	dev_notice(dev, "could not enable MSI-X (%d), trying MSI\n", ret);
+	ret = ccp_get_msi_irq(ccp);
+	if (!ret)
+		return 0;
+
+	/* Couldn't get MSI interrupt */
+	dev_notice(dev, "could not enable MSI (%d)\n", ret);
+
+	return ret;
+}
+
+static void ccp_free_irqs(struct ccp_device *ccp)
+{
+	struct ccp_pci *ccp_pci = ccp->dev_specific;
+	struct device *dev = ccp->dev;
+	struct pci_dev *pdev = container_of(dev, struct pci_dev, dev);
+
+	if (ccp_pci->msix_count) {
+		while (ccp_pci->msix_count--)
+			free_irq(ccp_pci->msix[ccp_pci->msix_count].vector,
+				 dev);
+		pci_disable_msix(pdev);
+	} else {
+		free_irq(pdev->irq, dev);
+		pci_disable_msi(pdev);
+	}
+}
+
+static int ccp_find_mmio_area(struct ccp_device *ccp)
+{
+	struct device *dev = ccp->dev;
+	struct pci_dev *pdev = container_of(dev, struct pci_dev, dev);
+	resource_size_t io_len;
+	unsigned long io_flags;
+	int bar;
+
+	io_flags = pci_resource_flags(pdev, IO_BAR);
+	io_len = pci_resource_len(pdev, IO_BAR);
+	if ((io_flags & IORESOURCE_MEM) && (io_len >= (IO_OFFSET + 0x800)))
+		return IO_BAR;
+
+	for (bar = 0; bar < PCI_STD_RESOURCE_END; bar++) {
+		io_flags = pci_resource_flags(pdev, bar);
+		io_len = pci_resource_len(pdev, bar);
+		if ((io_flags & IORESOURCE_MEM) &&
+		    (io_len >= (IO_OFFSET + 0x800)))
+			return bar;
+	}
+
+	return -EIO;
+}
+
+static int ccp_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
+{
+	struct ccp_device *ccp;
+	struct ccp_pci *ccp_pci;
+	struct device *dev = &pdev->dev;
+	unsigned int bar;
+	int ret;
+
+	ret = -ENOMEM;
+	ccp = ccp_alloc_struct(dev);
+	if (!ccp)
+		goto e_err;
+
+	ccp_pci = kzalloc(sizeof(*ccp_pci), GFP_KERNEL);
+	if (!ccp_pci) {
+		ret = -ENOMEM;
+		goto e_free1;
+	}
+	ccp->dev_specific = ccp_pci;
+	ccp->get_irq = ccp_get_irqs;
+	ccp->free_irq = ccp_free_irqs;
+
+	ret = pci_request_regions(pdev, "ccp");
+	if (ret) {
+		dev_err(dev, "pci_request_regions failed (%d)\n", ret);
+		goto e_free2;
+	}
+
+	ret = pci_enable_device(pdev);
+	if (ret) {
+		dev_err(dev, "pci_enable_device failed (%d)\n", ret);
+		goto e_regions;
+	}
+
+	pci_set_master(pdev);
+
+	ret = ccp_find_mmio_area(ccp);
+	if (ret < 0)
+		goto e_device;
+	bar = ret;
+
+	ret = -EIO;
+	ccp->io_map = pci_iomap(pdev, bar, 0);
+	if (ccp->io_map == NULL) {
+		dev_err(dev, "pci_iomap failed\n");
+		goto e_device;
+	}
+	ccp->io_regs = ccp->io_map + IO_OFFSET;
+
+	ret = dma_set_mask(dev, DMA_BIT_MASK(48));
+	if (ret == 0) {
+		ret = dma_set_coherent_mask(dev, DMA_BIT_MASK(48));
+		if (ret) {
+			dev_err(dev,
+				"pci_set_consistent_dma_mask failed (%d)\n",
+				ret);
+			goto e_bar0;
+		}
+	} else {
+		ret = dma_set_mask(dev, DMA_BIT_MASK(32));
+		if (ret) {
+			dev_err(dev, "pci_set_dma_mask failed (%d)\n", ret);
+			goto e_bar0;
+		}
+	}
+
+	dev_set_drvdata(dev, ccp);
+
+	ret = ccp_init(ccp);
+	if (ret)
+		goto e_bar0;
+
+	dev_notice(dev, "enabled\n");
+
+	return 0;
+
+e_bar0:
+	pci_iounmap(pdev, ccp->io_map);
+
+e_device:
+	pci_disable_device(pdev);
+
+e_regions:
+	pci_release_regions(pdev);
+
+e_free2:
+	kfree(ccp_pci);
+
+e_free1:
+	kfree(ccp);
+
+e_err:
+	dev_notice(dev, "initialization failed\n");
+	return ret;
+}
+
+static void ccp_pci_remove(struct pci_dev *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct ccp_device *ccp = dev_get_drvdata(dev);
+
+	if (!ccp)
+		return;
+
+	ccp_destroy(ccp);
+
+	pci_iounmap(pdev, ccp->io_map);
+
+	pci_disable_device(pdev);
+
+	pci_release_regions(pdev);
+
+	kfree(ccp);
+
+	dev_notice(dev, "disabled\n");
+}
+
+#ifdef CONFIG_PM
+static int ccp_pci_suspend(struct pci_dev *pdev, pm_message_t state)
+{
+	struct device *dev = &pdev->dev;
+	struct ccp_device *ccp = dev_get_drvdata(dev);
+	unsigned long flags;
+	unsigned int i;
+
+	spin_lock_irqsave(&ccp->cmd_lock, flags);
+
+	ccp->suspending = 1;
+
+	/* Wake all the queue kthreads to prepare for suspend */
+	for (i = 0; i < ccp->cmd_q_count; i++)
+		wake_up_process(ccp->cmd_q[i].kthread);
+
+	spin_unlock_irqrestore(&ccp->cmd_lock, flags);
+
+	/* Wait for all queue kthreads to say they're done */
+	while (!ccp_queues_suspended(ccp))
+		wait_event_interruptible(ccp->suspend_queue,
+					 ccp_queues_suspended(ccp));
+
+	return 0;
+}
+
+static int ccp_pci_resume(struct pci_dev *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct ccp_device *ccp = dev_get_drvdata(dev);
+	unsigned long flags;
+	unsigned int i;
+
+	spin_lock_irqsave(&ccp->cmd_lock, flags);
+
+	ccp->suspending = 0;
+
+	/* Wake up all the kthreads */
+	for (i = 0; i < ccp->cmd_q_count; i++) {
+		ccp->cmd_q[i].suspended = 0;
+		wake_up_process(ccp->cmd_q[i].kthread);
+	}
+
+	spin_unlock_irqrestore(&ccp->cmd_lock, flags);
+
+	return 0;
+}
+#endif
+
+static DEFINE_PCI_DEVICE_TABLE(ccp_pci_table) = {
+	{ PCI_VDEVICE(AMD, 0x1537), },
+	/* Last entry must be zero */
+	{ 0, }
+};
+MODULE_DEVICE_TABLE(pci, ccp_pci_table);
+
+static struct pci_driver ccp_pci_driver = {
+	.name = "AMD Cryptographic Coprocessor",
+	.id_table = ccp_pci_table,
+	.probe = ccp_pci_probe,
+	.remove = ccp_pci_remove,
+#ifdef CONFIG_PM
+	.suspend = ccp_pci_suspend,
+	.resume = ccp_pci_resume,
+#endif
+};
+
+int ccp_pci_init(void)
+{
+	return pci_register_driver(&ccp_pci_driver);
+}
+
+void ccp_pci_exit(void)
+{
+	pci_unregister_driver(&ccp_pci_driver);
+}

drivers/crypto/omap-aes.c

@@ -784,6 +784,7 @@ static int omap_aes_ctr_decrypt(struct ablkcipher_request *req)
 static int omap_aes_cra_init(struct crypto_tfm *tfm)
 {
 	struct omap_aes_dev *dd = NULL;
+	int err;
 
 	/* Find AES device, currently picks the first device */
 	spin_lock_bh(&list_lock);
@@ -792,7 +793,13 @@ static int omap_aes_cra_init(struct crypto_tfm *tfm)
 	}
 	spin_unlock_bh(&list_lock);
 
-	pm_runtime_get_sync(dd->dev);
+	err = pm_runtime_get_sync(dd->dev);
+	if (err < 0) {
+		dev_err(dd->dev, "%s: failed to get_sync(%d)\n",
+			__func__, err);
+		return err;
+	}
+
 	tfm->crt_ablkcipher.reqsize = sizeof(struct omap_aes_reqctx);
 
 	return 0;
@@ -1182,7 +1189,12 @@ static int omap_aes_probe(struct platform_device *pdev)
 	dd->phys_base = res.start;
 
 	pm_runtime_enable(dev);
-	pm_runtime_get_sync(dev);
+	err = pm_runtime_get_sync(dev);
+	if (err < 0) {
+		dev_err(dev, "%s: failed to get_sync(%d)\n",
+			__func__, err);
+		goto err_res;
+	}
 
 	omap_aes_dma_stop(dd);
 

drivers/crypto/omap-sham.c

@@ -789,10 +789,13 @@ static int omap_sham_update_cpu(struct omap_sham_dev *dd)
 	dev_dbg(dd->dev, "cpu: bufcnt: %u, digcnt: %d, final: %d\n",
 		ctx->bufcnt, ctx->digcnt, final);
 
+	if (final || (ctx->bufcnt == ctx->buflen && ctx->total)) {
 		bufcnt = ctx->bufcnt;
 		ctx->bufcnt = 0;
-
 		return omap_sham_xmit_cpu(dd, ctx->buffer, bufcnt, final);
+	}
+
+	return 0;
 }
 
 static int omap_sham_update_dma_stop(struct omap_sham_dev *dd)
@@ -1103,6 +1106,9 @@ static int omap_sham_update(struct ahash_request *req)
 		return 0;
 	}
 
+	if (dd->polling_mode)
+		ctx->flags |= BIT(FLAGS_CPU);
+
 	return omap_sham_enqueue(req, OP_UPDATE);
 }
 
@@ -1970,6 +1976,7 @@ static int omap_sham_probe(struct platform_device *pdev)
 			crypto_unregister_ahash(
 					&dd->pdata->algs_info[i].algs_list[j]);
 	pm_runtime_disable(dev);
+	if (dd->dma_lch)
 		dma_release_channel(dd->dma_lch);
 data_err:
 	dev_err(dev, "initialization failed.\n");
@@ -1994,6 +2001,8 @@ static int omap_sham_remove(struct platform_device *pdev)
 					&dd->pdata->algs_info[i].algs_list[j]);
 	tasklet_kill(&dd->done_task);
 	pm_runtime_disable(&pdev->dev);
+
+	if (dd->dma_lch)
 		dma_release_channel(dd->dma_lch);
 
 	return 0;

drivers/crypto/talitos.c

@@ -338,20 +338,29 @@ DEF_TALITOS_DONE(ch1_3, TALITOS_ISR_CH_1_3_DONE)
 static u32 current_desc_hdr(struct device *dev, int ch)
 {
 	struct talitos_private *priv = dev_get_drvdata(dev);
-	int tail = priv->chan[ch].tail;
+	int tail, iter;
 	dma_addr_t cur_desc;
 
-	cur_desc = in_be32(priv->chan[ch].reg + TALITOS_CDPR_LO);
+	cur_desc = ((u64)in_be32(priv->chan[ch].reg + TALITOS_CDPR)) << 32;
+	cur_desc |= in_be32(priv->chan[ch].reg + TALITOS_CDPR_LO);
 
-	while (priv->chan[ch].fifo[tail].dma_desc != cur_desc) {
-		tail = (tail + 1) & (priv->fifo_len - 1);
-		if (tail == priv->chan[ch].tail) {
+	if (!cur_desc) {
+		dev_err(dev, "CDPR is NULL, giving up search for offending descriptor\n");
+		return 0;
+	}
+
+	tail = priv->chan[ch].tail;
+
+	iter = tail;
+	while (priv->chan[ch].fifo[iter].dma_desc != cur_desc) {
+		iter = (iter + 1) & (priv->fifo_len - 1);
+		if (iter == tail) {
 			dev_err(dev, "couldn't locate current descriptor\n");
 			return 0;
 		}
 	}
 
-	return priv->chan[ch].fifo[tail].desc->hdr;
+	return priv->chan[ch].fifo[iter].desc->hdr;
 }
 
 /*
@@ -2486,8 +2495,6 @@ static int talitos_remove(struct platform_device *ofdev)
 
 	iounmap(priv->reg);
 
-	dev_set_drvdata(dev, NULL);
-
 	kfree(priv);
 
 	return 0;

include/linux/compiler-gcc.h

@@ -37,6 +37,9 @@
     __asm__ ("" : "=r"(__ptr) : "0"(ptr));		\
     (typeof(ptr)) (__ptr + (off)); })
 
+/* Make the optimizer believe the variable can be manipulated arbitrarily. */
+#define OPTIMIZER_HIDE_VAR(var) __asm__ ("" : "=r" (var) : "0" (var))
+
 #ifdef __CHECKER__
 #define __must_be_array(arr) 0
 #else

include/linux/compiler-intel.h

@@ -15,6 +15,7 @@
  */
 #undef barrier
 #undef RELOC_HIDE
+#undef OPTIMIZER_HIDE_VAR
 
 #define barrier() __memory_barrier()
 
@@ -23,6 +24,12 @@
      __ptr = (unsigned long) (ptr);				\
     (typeof(ptr)) (__ptr + (off)); })
 
+/* This should act as an optimization barrier on var.
+ * Given that this compiler does not have inline assembly, a compiler barrier
+ * is the best we can do.
+ */
+#define OPTIMIZER_HIDE_VAR(var) barrier()
+
 /* Intel ECC compiler doesn't support __builtin_types_compatible_p() */
 #define __must_be_array(a) 0
 

include/linux/compiler.h

@@ -170,6 +170,10 @@ void ftrace_likely_update(struct ftrace_branch_data *f, int val, int expect);
     (typeof(ptr)) (__ptr + (off)); })
 #endif
 
+#ifndef OPTIMIZER_HIDE_VAR
+#define OPTIMIZER_HIDE_VAR(var) barrier()
+#endif
+
 /* Not-quite-unique ID. */
 #ifndef __UNIQUE_ID
 # define __UNIQUE_ID(prefix) __PASTE(__PASTE(__UNIQUE_ID_, prefix), __LINE__)

kernel/padata.c

@@ -112,7 +112,7 @@ int padata_do_parallel(struct padata_instance *pinst,
 
 	rcu_read_lock_bh();
 
-	pd = rcu_dereference(pinst->pd);
+	pd = rcu_dereference_bh(pinst->pd);
 
 	err = -EINVAL;
 	if (!(pinst->flags & PADATA_INIT) || pinst->flags & PADATA_INVALID)