diff --git a/Documentation/padata.txt b/Documentation/padata.txt
index 269d7d0..473ebf2 100644
--- a/Documentation/padata.txt
+++ b/Documentation/padata.txt
@@ -1,5 +1,5 @@
 The padata parallel execution mechanism
-Last updated for 2.6.34
+Last updated for 2.6.36
 
 Padata is a mechanism by which the kernel can farm work out to be done in
 parallel on multiple CPUs while retaining the ordering of tasks.  It was
@@ -13,31 +13,86 @@
 
     #include <linux/padata.h>
 
-    struct padata_instance *padata_alloc(const struct cpumask *cpumask,
-				         struct workqueue_struct *wq);
+    struct padata_instance *padata_alloc(struct workqueue_struct *wq,
+					 const struct cpumask *pcpumask,
+					 const struct cpumask *cbcpumask);
 
-The cpumask describes which processors will be used to execute work
-submitted to this instance.  The workqueue wq is where the work will
-actually be done; it should be a multithreaded queue, naturally.
+The pcpumask describes which processors will be used to execute work
+submitted to this instance in parallel. The cbcpumask defines which
+processors are allowed to use as the serialization callback processor.
+The workqueue wq is where the work will actually be done; it should be
+a multithreaded queue, naturally.
+
+To allocate a padata instance with the cpu_possible_mask for both
+cpumasks this helper function can be used:
+
+    struct padata_instance *padata_alloc_possible(struct workqueue_struct *wq);
+
+Note: Padata maintains two kinds of cpumasks internally. The user supplied
+cpumasks, submitted by padata_alloc/padata_alloc_possible and the 'usable'
+cpumasks. The usable cpumasks are always the subset of active cpus in the
+user supplied cpumasks, these are the cpumasks padata actually use. So
+it is legal to supply a cpumask to padata that contains offline cpus.
+Once a offline cpu in the user supplied cpumask comes online, padata
+is going to use it.
 
 There are functions for enabling and disabling the instance:
 
-    void padata_start(struct padata_instance *pinst);
+    int padata_start(struct padata_instance *pinst);
     void padata_stop(struct padata_instance *pinst);
 
-These functions literally do nothing beyond setting or clearing the
-"padata_start() was called" flag; if that flag is not set, other functions
-will refuse to work.
+These functions are setting or clearing the "PADATA_INIT" flag;
+if that flag is not set, other functions will refuse to work.
+padata_start returns zero on success (flag set) or -EINVAL if the
+padata cpumask contains no active cpu (flag not set).
+padata_stop clears the flag and blocks until the padata instance
+is unused.
 
 The list of CPUs to be used can be adjusted with these functions:
 
-    int padata_set_cpumask(struct padata_instance *pinst,
+    int padata_set_cpumasks(struct padata_instance *pinst,
+			    cpumask_var_t pcpumask,
+			    cpumask_var_t cbcpumask);
+    int padata_set_cpumask(struct padata_instance *pinst, int cpumask_type,
 			   cpumask_var_t cpumask);
-    int padata_add_cpu(struct padata_instance *pinst, int cpu);
-    int padata_remove_cpu(struct padata_instance *pinst, int cpu);
+    int padata_add_cpu(struct padata_instance *pinst, int cpu, int mask);
+    int padata_remove_cpu(struct padata_instance *pinst, int cpu, int mask);
 
-Changing the CPU mask has the look of an expensive operation, though, so it
-probably should not be done with great frequency.
+Changing the CPU masks are expensive operations, though, so it should not be
+done with great frequency.
+
+It's possible to change both cpumasks of a padata instance with
+padata_set_cpumasks by specifying the cpumasks for parallel execution (pcpumask)
+and for the serial callback function (cbcpumask). padata_set_cpumask is to
+change just one of the cpumasks. Here cpumask_type is one of PADATA_CPU_SERIAL,
+PADATA_CPU_PARALLEL and cpumask specifies the new cpumask to use.
+To simply add or remove one cpu from a certain cpumask the functions
+padata_add_cpu/padata_remove_cpu are used. cpu specifies the cpu to add or
+remove and mask is one of PADATA_CPU_SERIAL, PADATA_CPU_PARALLEL.
+
+If a user is interested in padata cpumask changes, he can register to
+the padata cpumask change notifier:
+
+    int padata_register_cpumask_notifier(struct padata_instance *pinst,
+					 struct notifier_block *nblock);
+
+To unregister from that notifier:
+
+    int padata_unregister_cpumask_notifier(struct padata_instance *pinst,
+					   struct notifier_block *nblock);
+
+The padata cpumask change notifier notifies about changes of the usable
+cpumasks, i.e. the subset of active cpus in the user supplied cpumask.
+
+Padata calls the notifier chain with:
+
+    blocking_notifier_call_chain(&pinst->cpumask_change_notifier,
+				 notification_mask,
+				 &pd_new->cpumask);
+
+Here cpumask_change_notifier is registered notifier, notification_mask
+is one of PADATA_CPU_SERIAL, PADATA_CPU_PARALLEL and cpumask is a pointer
+to a struct padata_cpumask that contains the new cpumask informations.
 
 Actually submitting work to the padata instance requires the creation of a
 padata_priv structure:
@@ -50,7 +105,7 @@
 
 This structure will almost certainly be embedded within some larger
 structure specific to the work to be done.  Most its fields are private to
-padata, but the structure should be zeroed at initialization time, and the
+padata, but the structure should be zeroed at initialisation time, and the
 parallel() and serial() functions should be provided.  Those functions will
 be called in the process of getting the work done as we will see
 momentarily.
@@ -63,12 +118,10 @@
 The pinst and padata structures must be set up as described above; cb_cpu
 specifies which CPU will be used for the final callback when the work is
 done; it must be in the current instance's CPU mask.  The return value from
-padata_do_parallel() is a little strange; zero is an error return
-indicating that the caller forgot the padata_start() formalities.  -EBUSY
-means that somebody, somewhere else is messing with the instance's CPU
-mask, while -EINVAL is a complaint about cb_cpu not being in that CPU mask.
-If all goes well, this function will return -EINPROGRESS, indicating that
-the work is in progress.
+padata_do_parallel() is zero on success, indicating that the work is in
+progress. -EBUSY means that somebody, somewhere else is messing with the
+instance's CPU mask, while -EINVAL is a complaint about cb_cpu not being
+in that CPU mask or about a not running instance.
 
 Each task submitted to padata_do_parallel() will, in turn, be passed to
 exactly one call to the above-mentioned parallel() function, on one CPU, so
diff --git a/arch/s390/crypto/Makefile b/arch/s390/crypto/Makefile
index 6a1157f..1cf81d7 100644
--- a/arch/s390/crypto/Makefile
+++ b/arch/s390/crypto/Makefile
@@ -5,6 +5,6 @@
 obj-$(CONFIG_CRYPTO_SHA1_S390) += sha1_s390.o sha_common.o
 obj-$(CONFIG_CRYPTO_SHA256_S390) += sha256_s390.o sha_common.o
 obj-$(CONFIG_CRYPTO_SHA512_S390) += sha512_s390.o sha_common.o
-obj-$(CONFIG_CRYPTO_DES_S390) += des_s390.o des_check_key.o
+obj-$(CONFIG_CRYPTO_DES_S390) += des_s390.o
 obj-$(CONFIG_CRYPTO_AES_S390) += aes_s390.o
 obj-$(CONFIG_S390_PRNG) += prng.o
diff --git a/arch/s390/crypto/crypto_des.h b/arch/s390/crypto/crypto_des.h
index c964b64..6210457 100644
--- a/arch/s390/crypto/crypto_des.h
+++ b/arch/s390/crypto/crypto_des.h
@@ -15,4 +15,4 @@
 
 extern int crypto_des_check_key(const u8*, unsigned int, u32*);
 
-#endif //__CRYPTO_DES_H__
+#endif /*__CRYPTO_DES_H__*/
diff --git a/arch/s390/crypto/des_s390.c b/arch/s390/crypto/des_s390.c
index 2bc479a..cc54201 100644
--- a/arch/s390/crypto/des_s390.c
+++ b/arch/s390/crypto/des_s390.c
@@ -14,32 +14,21 @@
  *
  */
 
-#include <crypto/algapi.h>
 #include <linux/init.h>
 #include <linux/module.h>
+#include <linux/crypto.h>
+#include <crypto/algapi.h>
+#include <crypto/des.h>
 
 #include "crypt_s390.h"
-#include "crypto_des.h"
-
-#define DES_BLOCK_SIZE 8
-#define DES_KEY_SIZE 8
-
-#define DES3_128_KEY_SIZE	(2 * DES_KEY_SIZE)
-#define DES3_128_BLOCK_SIZE	DES_BLOCK_SIZE
 
 #define DES3_192_KEY_SIZE	(3 * DES_KEY_SIZE)
-#define DES3_192_BLOCK_SIZE	DES_BLOCK_SIZE
 
 struct crypt_s390_des_ctx {
 	u8 iv[DES_BLOCK_SIZE];
 	u8 key[DES_KEY_SIZE];
 };
 
-struct crypt_s390_des3_128_ctx {
-	u8 iv[DES_BLOCK_SIZE];
-	u8 key[DES3_128_KEY_SIZE];
-};
-
 struct crypt_s390_des3_192_ctx {
 	u8 iv[DES_BLOCK_SIZE];
 	u8 key[DES3_192_KEY_SIZE];
@@ -50,13 +39,16 @@
 {
 	struct crypt_s390_des_ctx *dctx = crypto_tfm_ctx(tfm);
 	u32 *flags = &tfm->crt_flags;
-	int ret;
+	u32 tmp[DES_EXPKEY_WORDS];
 
-	/* test if key is valid (not a weak key) */
-	ret = crypto_des_check_key(key, keylen, flags);
-	if (ret == 0)
-		memcpy(dctx->key, key, keylen);
-	return ret;
+	/* check for weak keys */
+	if (!des_ekey(tmp, key) && (*flags & CRYPTO_TFM_REQ_WEAK_KEY)) {
+		*flags |= CRYPTO_TFM_RES_WEAK_KEY;
+		return -EINVAL;
+	}
+
+	memcpy(dctx->key, key, keylen);
+	return 0;
 }
 
 static void des_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in)
@@ -237,165 +229,6 @@
  *   complementation keys.  Any weakness is obviated by the use of
  *   multiple keys.
  *
- *   However, if the two  independent 64-bit keys are equal,
- *   then the DES3 operation is simply the same as DES.
- *   Implementers MUST reject keys that exhibit this property.
- *
- */
-static int des3_128_setkey(struct crypto_tfm *tfm, const u8 *key,
-			   unsigned int keylen)
-{
-	int i, ret;
-	struct crypt_s390_des3_128_ctx *dctx = crypto_tfm_ctx(tfm);
-	const u8 *temp_key = key;
-	u32 *flags = &tfm->crt_flags;
-
-	if (!(memcmp(key, &key[DES_KEY_SIZE], DES_KEY_SIZE)) &&
-	    (*flags & CRYPTO_TFM_REQ_WEAK_KEY)) {
-		*flags |= CRYPTO_TFM_RES_WEAK_KEY;
-		return -EINVAL;
-	}
-	for (i = 0; i < 2; i++, temp_key += DES_KEY_SIZE) {
-		ret = crypto_des_check_key(temp_key, DES_KEY_SIZE, flags);
-		if (ret < 0)
-			return ret;
-	}
-	memcpy(dctx->key, key, keylen);
-	return 0;
-}
-
-static void des3_128_encrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
-{
-	struct crypt_s390_des3_128_ctx *dctx = crypto_tfm_ctx(tfm);
-
-	crypt_s390_km(KM_TDEA_128_ENCRYPT, dctx->key, dst, (void*)src,
-		      DES3_128_BLOCK_SIZE);
-}
-
-static void des3_128_decrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
-{
-	struct crypt_s390_des3_128_ctx *dctx = crypto_tfm_ctx(tfm);
-
-	crypt_s390_km(KM_TDEA_128_DECRYPT, dctx->key, dst, (void*)src,
-		      DES3_128_BLOCK_SIZE);
-}
-
-static struct crypto_alg des3_128_alg = {
-	.cra_name		=	"des3_ede128",
-	.cra_driver_name	=	"des3_ede128-s390",
-	.cra_priority		=	CRYPT_S390_PRIORITY,
-	.cra_flags		=	CRYPTO_ALG_TYPE_CIPHER,
-	.cra_blocksize		=	DES3_128_BLOCK_SIZE,
-	.cra_ctxsize		=	sizeof(struct crypt_s390_des3_128_ctx),
-	.cra_module		=	THIS_MODULE,
-	.cra_list		=	LIST_HEAD_INIT(des3_128_alg.cra_list),
-	.cra_u			=	{
-		.cipher = {
-			.cia_min_keysize	=	DES3_128_KEY_SIZE,
-			.cia_max_keysize	=	DES3_128_KEY_SIZE,
-			.cia_setkey		=	des3_128_setkey,
-			.cia_encrypt		=	des3_128_encrypt,
-			.cia_decrypt		=	des3_128_decrypt,
-		}
-	}
-};
-
-static int ecb_des3_128_encrypt(struct blkcipher_desc *desc,
-				struct scatterlist *dst,
-				struct scatterlist *src, unsigned int nbytes)
-{
-	struct crypt_s390_des3_128_ctx *sctx = crypto_blkcipher_ctx(desc->tfm);
-	struct blkcipher_walk walk;
-
-	blkcipher_walk_init(&walk, dst, src, nbytes);
-	return ecb_desall_crypt(desc, KM_TDEA_128_ENCRYPT, sctx->key, &walk);
-}
-
-static int ecb_des3_128_decrypt(struct blkcipher_desc *desc,
-				struct scatterlist *dst,
-				struct scatterlist *src, unsigned int nbytes)
-{
-	struct crypt_s390_des3_128_ctx *sctx = crypto_blkcipher_ctx(desc->tfm);
-	struct blkcipher_walk walk;
-
-	blkcipher_walk_init(&walk, dst, src, nbytes);
-	return ecb_desall_crypt(desc, KM_TDEA_128_DECRYPT, sctx->key, &walk);
-}
-
-static struct crypto_alg ecb_des3_128_alg = {
-	.cra_name		=	"ecb(des3_ede128)",
-	.cra_driver_name	=	"ecb-des3_ede128-s390",
-	.cra_priority		=	CRYPT_S390_COMPOSITE_PRIORITY,
-	.cra_flags		=	CRYPTO_ALG_TYPE_BLKCIPHER,
-	.cra_blocksize		=	DES3_128_BLOCK_SIZE,
-	.cra_ctxsize		=	sizeof(struct crypt_s390_des3_128_ctx),
-	.cra_type		=	&crypto_blkcipher_type,
-	.cra_module		=	THIS_MODULE,
-	.cra_list		=	LIST_HEAD_INIT(
-						ecb_des3_128_alg.cra_list),
-	.cra_u			=	{
-		.blkcipher = {
-			.min_keysize		=	DES3_128_KEY_SIZE,
-			.max_keysize		=	DES3_128_KEY_SIZE,
-			.setkey			=	des3_128_setkey,
-			.encrypt		=	ecb_des3_128_encrypt,
-			.decrypt		=	ecb_des3_128_decrypt,
-		}
-	}
-};
-
-static int cbc_des3_128_encrypt(struct blkcipher_desc *desc,
-				struct scatterlist *dst,
-				struct scatterlist *src, unsigned int nbytes)
-{
-	struct crypt_s390_des3_128_ctx *sctx = crypto_blkcipher_ctx(desc->tfm);
-	struct blkcipher_walk walk;
-
-	blkcipher_walk_init(&walk, dst, src, nbytes);
-	return cbc_desall_crypt(desc, KMC_TDEA_128_ENCRYPT, sctx->iv, &walk);
-}
-
-static int cbc_des3_128_decrypt(struct blkcipher_desc *desc,
-				struct scatterlist *dst,
-				struct scatterlist *src, unsigned int nbytes)
-{
-	struct crypt_s390_des3_128_ctx *sctx = crypto_blkcipher_ctx(desc->tfm);
-	struct blkcipher_walk walk;
-
-	blkcipher_walk_init(&walk, dst, src, nbytes);
-	return cbc_desall_crypt(desc, KMC_TDEA_128_DECRYPT, sctx->iv, &walk);
-}
-
-static struct crypto_alg cbc_des3_128_alg = {
-	.cra_name		=	"cbc(des3_ede128)",
-	.cra_driver_name	=	"cbc-des3_ede128-s390",
-	.cra_priority		=	CRYPT_S390_COMPOSITE_PRIORITY,
-	.cra_flags		=	CRYPTO_ALG_TYPE_BLKCIPHER,
-	.cra_blocksize		=	DES3_128_BLOCK_SIZE,
-	.cra_ctxsize		=	sizeof(struct crypt_s390_des3_128_ctx),
-	.cra_type		=	&crypto_blkcipher_type,
-	.cra_module		=	THIS_MODULE,
-	.cra_list		=	LIST_HEAD_INIT(
-						cbc_des3_128_alg.cra_list),
-	.cra_u			=	{
-		.blkcipher = {
-			.min_keysize		=	DES3_128_KEY_SIZE,
-			.max_keysize		=	DES3_128_KEY_SIZE,
-			.ivsize			=	DES3_128_BLOCK_SIZE,
-			.setkey			=	des3_128_setkey,
-			.encrypt		=	cbc_des3_128_encrypt,
-			.decrypt		=	cbc_des3_128_decrypt,
-		}
-	}
-};
-
-/*
- * RFC2451:
- *
- *   For DES-EDE3, there is no known need to reject weak or
- *   complementation keys.  Any weakness is obviated by the use of
- *   multiple keys.
- *
  *   However, if the first two or last two independent 64-bit keys are
  *   equal (k1 == k2 or k2 == k3), then the DES3 operation is simply the
  *   same as DES.  Implementers MUST reject keys that exhibit this
@@ -405,9 +238,7 @@
 static int des3_192_setkey(struct crypto_tfm *tfm, const u8 *key,
 			   unsigned int keylen)
 {
-	int i, ret;
 	struct crypt_s390_des3_192_ctx *dctx = crypto_tfm_ctx(tfm);
-	const u8 *temp_key = key;
 	u32 *flags = &tfm->crt_flags;
 
 	if (!(memcmp(key, &key[DES_KEY_SIZE], DES_KEY_SIZE) &&
@@ -417,11 +248,6 @@
 		*flags |= CRYPTO_TFM_RES_WEAK_KEY;
 		return -EINVAL;
 	}
-	for (i = 0; i < 3; i++, temp_key += DES_KEY_SIZE) {
-		ret = crypto_des_check_key(temp_key, DES_KEY_SIZE, flags);
-		if (ret < 0)
-			return ret;
-	}
 	memcpy(dctx->key, key, keylen);
 	return 0;
 }
@@ -431,7 +257,7 @@
 	struct crypt_s390_des3_192_ctx *dctx = crypto_tfm_ctx(tfm);
 
 	crypt_s390_km(KM_TDEA_192_ENCRYPT, dctx->key, dst, (void*)src,
-		      DES3_192_BLOCK_SIZE);
+		      DES_BLOCK_SIZE);
 }
 
 static void des3_192_decrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
@@ -439,7 +265,7 @@
 	struct crypt_s390_des3_192_ctx *dctx = crypto_tfm_ctx(tfm);
 
 	crypt_s390_km(KM_TDEA_192_DECRYPT, dctx->key, dst, (void*)src,
-		      DES3_192_BLOCK_SIZE);
+		      DES_BLOCK_SIZE);
 }
 
 static struct crypto_alg des3_192_alg = {
@@ -447,7 +273,7 @@
 	.cra_driver_name	=	"des3_ede-s390",
 	.cra_priority		=	CRYPT_S390_PRIORITY,
 	.cra_flags		=	CRYPTO_ALG_TYPE_CIPHER,
-	.cra_blocksize		=	DES3_192_BLOCK_SIZE,
+	.cra_blocksize		=	DES_BLOCK_SIZE,
 	.cra_ctxsize		=	sizeof(struct crypt_s390_des3_192_ctx),
 	.cra_module		=	THIS_MODULE,
 	.cra_list		=	LIST_HEAD_INIT(des3_192_alg.cra_list),
@@ -489,7 +315,7 @@
 	.cra_driver_name	=	"ecb-des3_ede-s390",
 	.cra_priority		=	CRYPT_S390_COMPOSITE_PRIORITY,
 	.cra_flags		=	CRYPTO_ALG_TYPE_BLKCIPHER,
-	.cra_blocksize		=	DES3_192_BLOCK_SIZE,
+	.cra_blocksize		=	DES_BLOCK_SIZE,
 	.cra_ctxsize		=	sizeof(struct crypt_s390_des3_192_ctx),
 	.cra_type		=	&crypto_blkcipher_type,
 	.cra_module		=	THIS_MODULE,
@@ -533,7 +359,7 @@
 	.cra_driver_name	=	"cbc-des3_ede-s390",
 	.cra_priority		=	CRYPT_S390_COMPOSITE_PRIORITY,
 	.cra_flags		=	CRYPTO_ALG_TYPE_BLKCIPHER,
-	.cra_blocksize		=	DES3_192_BLOCK_SIZE,
+	.cra_blocksize		=	DES_BLOCK_SIZE,
 	.cra_ctxsize		=	sizeof(struct crypt_s390_des3_192_ctx),
 	.cra_type		=	&crypto_blkcipher_type,
 	.cra_module		=	THIS_MODULE,
@@ -543,7 +369,7 @@
 		.blkcipher = {
 			.min_keysize		=	DES3_192_KEY_SIZE,
 			.max_keysize		=	DES3_192_KEY_SIZE,
-			.ivsize			=	DES3_192_BLOCK_SIZE,
+			.ivsize			=	DES_BLOCK_SIZE,
 			.setkey			=	des3_192_setkey,
 			.encrypt		=	cbc_des3_192_encrypt,
 			.decrypt		=	cbc_des3_192_decrypt,
@@ -553,10 +379,9 @@
 
 static int des_s390_init(void)
 {
-	int ret = 0;
+	int ret;
 
 	if (!crypt_s390_func_available(KM_DEA_ENCRYPT) ||
-	    !crypt_s390_func_available(KM_TDEA_128_ENCRYPT) ||
 	    !crypt_s390_func_available(KM_TDEA_192_ENCRYPT))
 		return -EOPNOTSUPP;
 
@@ -569,17 +394,6 @@
 	ret = crypto_register_alg(&cbc_des_alg);
 	if (ret)
 		goto cbc_des_err;
-
-	ret = crypto_register_alg(&des3_128_alg);
-	if (ret)
-		goto des3_128_err;
-	ret = crypto_register_alg(&ecb_des3_128_alg);
-	if (ret)
-		goto ecb_des3_128_err;
-	ret = crypto_register_alg(&cbc_des3_128_alg);
-	if (ret)
-		goto cbc_des3_128_err;
-
 	ret = crypto_register_alg(&des3_192_alg);
 	if (ret)
 		goto des3_192_err;
@@ -589,7 +403,6 @@
 	ret = crypto_register_alg(&cbc_des3_192_alg);
 	if (ret)
 		goto cbc_des3_192_err;
-
 out:
 	return ret;
 
@@ -598,12 +411,6 @@
 ecb_des3_192_err:
 	crypto_unregister_alg(&des3_192_alg);
 des3_192_err:
-	crypto_unregister_alg(&cbc_des3_128_alg);
-cbc_des3_128_err:
-	crypto_unregister_alg(&ecb_des3_128_alg);
-ecb_des3_128_err:
-	crypto_unregister_alg(&des3_128_alg);
-des3_128_err:
 	crypto_unregister_alg(&cbc_des_alg);
 cbc_des_err:
 	crypto_unregister_alg(&ecb_des_alg);
@@ -613,21 +420,18 @@
 	goto out;
 }
 
-static void __exit des_s390_fini(void)
+static void __exit des_s390_exit(void)
 {
 	crypto_unregister_alg(&cbc_des3_192_alg);
 	crypto_unregister_alg(&ecb_des3_192_alg);
 	crypto_unregister_alg(&des3_192_alg);
-	crypto_unregister_alg(&cbc_des3_128_alg);
-	crypto_unregister_alg(&ecb_des3_128_alg);
-	crypto_unregister_alg(&des3_128_alg);
 	crypto_unregister_alg(&cbc_des_alg);
 	crypto_unregister_alg(&ecb_des_alg);
 	crypto_unregister_alg(&des_alg);
 }
 
 module_init(des_s390_init);
-module_exit(des_s390_fini);
+module_exit(des_s390_exit);
 
 MODULE_ALIAS("des");
 MODULE_ALIAS("des3_ede");
diff --git a/crypto/Kconfig b/crypto/Kconfig
index 9d9434f..1cd497d 100644
--- a/crypto/Kconfig
+++ b/crypto/Kconfig
@@ -80,6 +80,11 @@
 
 config CRYPTO_PCOMP
 	tristate
+	select CRYPTO_PCOMP2
+	select CRYPTO_ALGAPI
+
+config CRYPTO_PCOMP2
+	tristate
 	select CRYPTO_ALGAPI2
 
 config CRYPTO_MANAGER
@@ -94,7 +99,15 @@
 	select CRYPTO_AEAD2
 	select CRYPTO_HASH2
 	select CRYPTO_BLKCIPHER2
-	select CRYPTO_PCOMP
+	select CRYPTO_PCOMP2
+
+config CRYPTO_MANAGER_TESTS
+	bool "Run algolithms' self-tests"
+	default y
+	depends on CRYPTO_MANAGER2
+	help
+	  Run cryptomanager's tests for the new crypto algorithms being
+	  registered.
 
 config CRYPTO_GF128MUL
 	tristate "GF(2^128) multiplication functions (EXPERIMENTAL)"
diff --git a/crypto/Makefile b/crypto/Makefile
index d7e6441..423b7de 100644
--- a/crypto/Makefile
+++ b/crypto/Makefile
@@ -26,7 +26,7 @@
 crypto_hash-objs += shash.o
 obj-$(CONFIG_CRYPTO_HASH2) += crypto_hash.o
 
-obj-$(CONFIG_CRYPTO_PCOMP) += pcompress.o
+obj-$(CONFIG_CRYPTO_PCOMP2) += pcompress.o
 
 cryptomgr-objs := algboss.o testmgr.o
 
@@ -61,7 +61,7 @@
 obj-$(CONFIG_CRYPTO_DES) += des_generic.o
 obj-$(CONFIG_CRYPTO_FCRYPT) += fcrypt.o
 obj-$(CONFIG_CRYPTO_BLOWFISH) += blowfish.o
-obj-$(CONFIG_CRYPTO_TWOFISH) += twofish.o
+obj-$(CONFIG_CRYPTO_TWOFISH) += twofish_generic.o
 obj-$(CONFIG_CRYPTO_TWOFISH_COMMON) += twofish_common.o
 obj-$(CONFIG_CRYPTO_SERPENT) += serpent.o
 obj-$(CONFIG_CRYPTO_AES) += aes_generic.o
diff --git a/crypto/algboss.c b/crypto/algboss.c
index c3c196b..40bd391 100644
--- a/crypto/algboss.c
+++ b/crypto/algboss.c
@@ -206,6 +206,7 @@
 	return NOTIFY_OK;
 }
 
+#ifdef CONFIG_CRYPTO_MANAGER_TESTS
 static int cryptomgr_test(void *data)
 {
 	struct crypto_test_param *param = data;
@@ -266,6 +267,7 @@
 err:
 	return NOTIFY_OK;
 }
+#endif /* CONFIG_CRYPTO_MANAGER_TESTS */
 
 static int cryptomgr_notify(struct notifier_block *this, unsigned long msg,
 			    void *data)
@@ -273,8 +275,10 @@
 	switch (msg) {
 	case CRYPTO_MSG_ALG_REQUEST:
 		return cryptomgr_schedule_probe(data);
+#ifdef CONFIG_CRYPTO_MANAGER_TESTS
 	case CRYPTO_MSG_ALG_REGISTER:
 		return cryptomgr_schedule_test(data);
+#endif
 	}
 
 	return NOTIFY_DONE;
diff --git a/crypto/authenc.c b/crypto/authenc.c
index b9884ee..a5a22cf 100644
--- a/crypto/authenc.c
+++ b/crypto/authenc.c
@@ -616,7 +616,7 @@
 	auth = ahash_attr_alg(tb[1], CRYPTO_ALG_TYPE_HASH,
 			       CRYPTO_ALG_TYPE_AHASH_MASK);
 	if (IS_ERR(auth))
-		return ERR_PTR(PTR_ERR(auth));
+		return ERR_CAST(auth);
 
 	auth_base = &auth->base;
 
diff --git a/crypto/ctr.c b/crypto/ctr.c
index 6c3bfab..4ca7222 100644
--- a/crypto/ctr.c
+++ b/crypto/ctr.c
@@ -185,7 +185,7 @@
 	alg = crypto_attr_alg(tb[1], CRYPTO_ALG_TYPE_CIPHER,
 				  CRYPTO_ALG_TYPE_MASK);
 	if (IS_ERR(alg))
-		return ERR_PTR(PTR_ERR(alg));
+		return ERR_CAST(alg);
 
 	/* Block size must be >= 4 bytes. */
 	err = -EINVAL;
diff --git a/crypto/pcrypt.c b/crypto/pcrypt.c
index 247178c..de30782 100644
--- a/crypto/pcrypt.c
+++ b/crypto/pcrypt.c
@@ -24,12 +24,40 @@
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/slab.h>
+#include <linux/notifier.h>
+#include <linux/kobject.h>
+#include <linux/cpu.h>
 #include <crypto/pcrypt.h>
 
-static struct padata_instance *pcrypt_enc_padata;
-static struct padata_instance *pcrypt_dec_padata;
-static struct workqueue_struct *encwq;
-static struct workqueue_struct *decwq;
+struct padata_pcrypt {
+	struct padata_instance *pinst;
+	struct workqueue_struct *wq;
+
+	/*
+	 * Cpumask for callback CPUs. It should be
+	 * equal to serial cpumask of corresponding padata instance,
+	 * so it is updated when padata notifies us about serial
+	 * cpumask change.
+	 *
+	 * cb_cpumask is protected by RCU. This fact prevents us from
+	 * using cpumask_var_t directly because the actual type of
+	 * cpumsak_var_t depends on kernel configuration(particularly on
+	 * CONFIG_CPUMASK_OFFSTACK macro). Depending on the configuration
+	 * cpumask_var_t may be either a pointer to the struct cpumask
+	 * or a variable allocated on the stack. Thus we can not safely use
+	 * cpumask_var_t with RCU operations such as rcu_assign_pointer or
+	 * rcu_dereference. So cpumask_var_t is wrapped with struct
+	 * pcrypt_cpumask which makes possible to use it with RCU.
+	 */
+	struct pcrypt_cpumask {
+		cpumask_var_t mask;
+	} *cb_cpumask;
+	struct notifier_block nblock;
+};
+
+static struct padata_pcrypt pencrypt;
+static struct padata_pcrypt pdecrypt;
+static struct kset           *pcrypt_kset;
 
 struct pcrypt_instance_ctx {
 	struct crypto_spawn spawn;
@@ -42,25 +70,32 @@
 };
 
 static int pcrypt_do_parallel(struct padata_priv *padata, unsigned int *cb_cpu,
-			      struct padata_instance *pinst)
+			      struct padata_pcrypt *pcrypt)
 {
 	unsigned int cpu_index, cpu, i;
+	struct pcrypt_cpumask *cpumask;
 
 	cpu = *cb_cpu;
 
-	if (cpumask_test_cpu(cpu, cpu_active_mask))
+	rcu_read_lock_bh();
+	cpumask = rcu_dereference(pcrypt->cb_cpumask);
+	if (cpumask_test_cpu(cpu, cpumask->mask))
 			goto out;
 
-	cpu_index = cpu % cpumask_weight(cpu_active_mask);
+	if (!cpumask_weight(cpumask->mask))
+			goto out;
 
-	cpu = cpumask_first(cpu_active_mask);
+	cpu_index = cpu % cpumask_weight(cpumask->mask);
+
+	cpu = cpumask_first(cpumask->mask);
 	for (i = 0; i < cpu_index; i++)
-		cpu = cpumask_next(cpu, cpu_active_mask);
+		cpu = cpumask_next(cpu, cpumask->mask);
 
 	*cb_cpu = cpu;
 
 out:
-	return padata_do_parallel(pinst, padata, cpu);
+	rcu_read_unlock_bh();
+	return padata_do_parallel(pcrypt->pinst, padata, cpu);
 }
 
 static int pcrypt_aead_setkey(struct crypto_aead *parent,
@@ -142,11 +177,9 @@
 			       req->cryptlen, req->iv);
 	aead_request_set_assoc(creq, req->assoc, req->assoclen);
 
-	err = pcrypt_do_parallel(padata, &ctx->cb_cpu, pcrypt_enc_padata);
-	if (err)
-		return err;
-	else
-		err = crypto_aead_encrypt(creq);
+	err = pcrypt_do_parallel(padata, &ctx->cb_cpu, &pencrypt);
+	if (!err)
+		return -EINPROGRESS;
 
 	return err;
 }
@@ -186,11 +219,9 @@
 			       req->cryptlen, req->iv);
 	aead_request_set_assoc(creq, req->assoc, req->assoclen);
 
-	err = pcrypt_do_parallel(padata, &ctx->cb_cpu, pcrypt_dec_padata);
-	if (err)
-		return err;
-	else
-		err = crypto_aead_decrypt(creq);
+	err = pcrypt_do_parallel(padata, &ctx->cb_cpu, &pdecrypt);
+	if (!err)
+		return -EINPROGRESS;
 
 	return err;
 }
@@ -232,11 +263,9 @@
 	aead_givcrypt_set_assoc(creq, areq->assoc, areq->assoclen);
 	aead_givcrypt_set_giv(creq, req->giv, req->seq);
 
-	err = pcrypt_do_parallel(padata, &ctx->cb_cpu, pcrypt_enc_padata);
-	if (err)
-		return err;
-	else
-		err = crypto_aead_givencrypt(creq);
+	err = pcrypt_do_parallel(padata, &ctx->cb_cpu, &pencrypt);
+	if (!err)
+		return -EINPROGRESS;
 
 	return err;
 }
@@ -376,6 +405,115 @@
 	kfree(inst);
 }
 
+static int pcrypt_cpumask_change_notify(struct notifier_block *self,
+					unsigned long val, void *data)
+{
+	struct padata_pcrypt *pcrypt;
+	struct pcrypt_cpumask *new_mask, *old_mask;
+	struct padata_cpumask *cpumask = (struct padata_cpumask *)data;
+
+	if (!(val & PADATA_CPU_SERIAL))
+		return 0;
+
+	pcrypt = container_of(self, struct padata_pcrypt, nblock);
+	new_mask = kmalloc(sizeof(*new_mask), GFP_KERNEL);
+	if (!new_mask)
+		return -ENOMEM;
+	if (!alloc_cpumask_var(&new_mask->mask, GFP_KERNEL)) {
+		kfree(new_mask);
+		return -ENOMEM;
+	}
+
+	old_mask = pcrypt->cb_cpumask;
+
+	cpumask_copy(new_mask->mask, cpumask->cbcpu);
+	rcu_assign_pointer(pcrypt->cb_cpumask, new_mask);
+	synchronize_rcu_bh();
+
+	free_cpumask_var(old_mask->mask);
+	kfree(old_mask);
+	return 0;
+}
+
+static int pcrypt_sysfs_add(struct padata_instance *pinst, const char *name)
+{
+	int ret;
+
+	pinst->kobj.kset = pcrypt_kset;
+	ret = kobject_add(&pinst->kobj, NULL, name);
+	if (!ret)
+		kobject_uevent(&pinst->kobj, KOBJ_ADD);
+
+	return ret;
+}
+
+static int pcrypt_init_padata(struct padata_pcrypt *pcrypt,
+			      const char *name)
+{
+	int ret = -ENOMEM;
+	struct pcrypt_cpumask *mask;
+
+	get_online_cpus();
+
+	pcrypt->wq = create_workqueue(name);
+	if (!pcrypt->wq)
+		goto err;
+
+	pcrypt->pinst = padata_alloc_possible(pcrypt->wq);
+	if (!pcrypt->pinst)
+		goto err_destroy_workqueue;
+
+	mask = kmalloc(sizeof(*mask), GFP_KERNEL);
+	if (!mask)
+		goto err_free_padata;
+	if (!alloc_cpumask_var(&mask->mask, GFP_KERNEL)) {
+		kfree(mask);
+		goto err_free_padata;
+	}
+
+	cpumask_and(mask->mask, cpu_possible_mask, cpu_active_mask);
+	rcu_assign_pointer(pcrypt->cb_cpumask, mask);
+
+	pcrypt->nblock.notifier_call = pcrypt_cpumask_change_notify;
+	ret = padata_register_cpumask_notifier(pcrypt->pinst, &pcrypt->nblock);
+	if (ret)
+		goto err_free_cpumask;
+
+	ret = pcrypt_sysfs_add(pcrypt->pinst, name);
+	if (ret)
+		goto err_unregister_notifier;
+
+	put_online_cpus();
+
+	return ret;
+
+err_unregister_notifier:
+	padata_unregister_cpumask_notifier(pcrypt->pinst, &pcrypt->nblock);
+err_free_cpumask:
+	free_cpumask_var(mask->mask);
+	kfree(mask);
+err_free_padata:
+	padata_free(pcrypt->pinst);
+err_destroy_workqueue:
+	destroy_workqueue(pcrypt->wq);
+err:
+	put_online_cpus();
+
+	return ret;
+}
+
+static void pcrypt_fini_padata(struct padata_pcrypt *pcrypt)
+{
+	kobject_put(&pcrypt->pinst->kobj);
+	free_cpumask_var(pcrypt->cb_cpumask->mask);
+	kfree(pcrypt->cb_cpumask);
+
+	padata_stop(pcrypt->pinst);
+	padata_unregister_cpumask_notifier(pcrypt->pinst, &pcrypt->nblock);
+	destroy_workqueue(pcrypt->wq);
+	padata_free(pcrypt->pinst);
+}
+
 static struct crypto_template pcrypt_tmpl = {
 	.name = "pcrypt",
 	.alloc = pcrypt_alloc,
@@ -385,52 +523,39 @@
 
 static int __init pcrypt_init(void)
 {
-	encwq = create_workqueue("pencrypt");
-	if (!encwq)
+	int err = -ENOMEM;
+
+	pcrypt_kset = kset_create_and_add("pcrypt", NULL, kernel_kobj);
+	if (!pcrypt_kset)
 		goto err;
 
-	decwq = create_workqueue("pdecrypt");
-	if (!decwq)
-		goto err_destroy_encwq;
+	err = pcrypt_init_padata(&pencrypt, "pencrypt");
+	if (err)
+		goto err_unreg_kset;
 
+	err = pcrypt_init_padata(&pdecrypt, "pdecrypt");
+	if (err)
+		goto err_deinit_pencrypt;
 
-	pcrypt_enc_padata = padata_alloc(cpu_possible_mask, encwq);
-	if (!pcrypt_enc_padata)
-		goto err_destroy_decwq;
-
-	pcrypt_dec_padata = padata_alloc(cpu_possible_mask, decwq);
-	if (!pcrypt_dec_padata)
-		goto err_free_padata;
-
-	padata_start(pcrypt_enc_padata);
-	padata_start(pcrypt_dec_padata);
+	padata_start(pencrypt.pinst);
+	padata_start(pdecrypt.pinst);
 
 	return crypto_register_template(&pcrypt_tmpl);
 
-err_free_padata:
-	padata_free(pcrypt_enc_padata);
-
-err_destroy_decwq:
-	destroy_workqueue(decwq);
-
-err_destroy_encwq:
-	destroy_workqueue(encwq);
-
+err_deinit_pencrypt:
+	pcrypt_fini_padata(&pencrypt);
+err_unreg_kset:
+	kset_unregister(pcrypt_kset);
 err:
-	return -ENOMEM;
+	return err;
 }
 
 static void __exit pcrypt_exit(void)
 {
-	padata_stop(pcrypt_enc_padata);
-	padata_stop(pcrypt_dec_padata);
+	pcrypt_fini_padata(&pencrypt);
+	pcrypt_fini_padata(&pdecrypt);
 
-	destroy_workqueue(encwq);
-	destroy_workqueue(decwq);
-
-	padata_free(pcrypt_enc_padata);
-	padata_free(pcrypt_dec_padata);
-
+	kset_unregister(pcrypt_kset);
 	crypto_unregister_template(&pcrypt_tmpl);
 }
 
diff --git a/crypto/testmgr.c b/crypto/testmgr.c
index 5c8aaa0..abd980c 100644
--- a/crypto/testmgr.c
+++ b/crypto/testmgr.c
@@ -22,6 +22,17 @@
 #include <crypto/rng.h>
 
 #include "internal.h"
+
+#ifndef CONFIG_CRYPTO_MANAGER_TESTS
+
+/* a perfect nop */
+int alg_test(const char *driver, const char *alg, u32 type, u32 mask)
+{
+	return 0;
+}
+
+#else
+
 #include "testmgr.h"
 
 /*
@@ -2530,4 +2541,7 @@
 non_fips_alg:
 	return -EINVAL;
 }
+
+#endif /* CONFIG_CRYPTO_MANAGER_TESTS */
+
 EXPORT_SYMBOL_GPL(alg_test);
diff --git a/crypto/twofish.c b/crypto/twofish_generic.c
similarity index 99%
rename from crypto/twofish.c
rename to crypto/twofish_generic.c
index dfcda23..1f07b84 100644
--- a/crypto/twofish.c
+++ b/crypto/twofish_generic.c
@@ -212,3 +212,4 @@
 
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION ("Twofish Cipher Algorithm");
+MODULE_ALIAS("twofish");
diff --git a/crypto/xts.c b/crypto/xts.c
index d87b0f3..555ecaa 100644
--- a/crypto/xts.c
+++ b/crypto/xts.c
@@ -224,7 +224,7 @@
 	alg = crypto_get_attr_alg(tb, CRYPTO_ALG_TYPE_CIPHER,
 				  CRYPTO_ALG_TYPE_MASK);
 	if (IS_ERR(alg))
-		return ERR_PTR(PTR_ERR(alg));
+		return ERR_CAST(alg);
 
 	inst = crypto_alloc_instance("xts", alg);
 	if (IS_ERR(inst))
diff --git a/drivers/char/hw_random/n2-drv.c b/drivers/char/hw_random/n2-drv.c
index 0f9cbf1..101d5f2 100644
--- a/drivers/char/hw_random/n2-drv.c
+++ b/drivers/char/hw_random/n2-drv.c
@@ -387,7 +387,7 @@
 
 static int n2rng_data_read(struct hwrng *rng, u32 *data)
 {
-	struct n2rng *np = (struct n2rng *) rng->priv;
+	struct n2rng *np = rng->priv;
 	unsigned long ra = __pa(&np->test_data);
 	int len;
 
diff --git a/drivers/char/random.c b/drivers/char/random.c
index 8d85587..caef35a 100644
--- a/drivers/char/random.c
+++ b/drivers/char/random.c
@@ -407,8 +407,8 @@
 	struct poolinfo *poolinfo;
 	__u32 *pool;
 	const char *name;
-	int limit;
 	struct entropy_store *pull;
+	int limit;
 
 	/* read-write data: */
 	spinlock_t lock;
diff --git a/drivers/crypto/geode-aes.c b/drivers/crypto/geode-aes.c
index 09389dd..219d09c 100644
--- a/drivers/crypto/geode-aes.c
+++ b/drivers/crypto/geode-aes.c
@@ -573,7 +573,7 @@
 }
 
 static struct pci_device_id geode_aes_tbl[] = {
-	{ PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_LX_AES, PCI_ANY_ID, PCI_ANY_ID} ,
+	{ PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_LX_AES), } ,
 	{ 0, }
 };
 
diff --git a/drivers/crypto/hifn_795x.c b/drivers/crypto/hifn_795x.c
index 16fce3a..e449ac5 100644
--- a/drivers/crypto/hifn_795x.c
+++ b/drivers/crypto/hifn_795x.c
@@ -2018,7 +2018,6 @@
 {
 	unsigned long flags;
 	struct crypto_async_request *async_req;
-	struct hifn_context *ctx;
 	struct ablkcipher_request *req;
 	struct hifn_dma *dma = (struct hifn_dma *)dev->desc_virt;
 	int i;
@@ -2035,7 +2034,6 @@
 
 	spin_lock_irqsave(&dev->lock, flags);
 	while ((async_req = crypto_dequeue_request(&dev->queue))) {
-		ctx = crypto_tfm_ctx(async_req->tfm);
 		req = container_of(async_req, struct ablkcipher_request, base);
 		spin_unlock_irqrestore(&dev->lock, flags);
 
@@ -2139,7 +2137,6 @@
 static int hifn_process_queue(struct hifn_device *dev)
 {
 	struct crypto_async_request *async_req, *backlog;
-	struct hifn_context *ctx;
 	struct ablkcipher_request *req;
 	unsigned long flags;
 	int err = 0;
@@ -2156,7 +2153,6 @@
 		if (backlog)
 			backlog->complete(backlog, -EINPROGRESS);
 
-		ctx = crypto_tfm_ctx(async_req->tfm);
 		req = container_of(async_req, struct ablkcipher_request, base);
 
 		err = hifn_handle_req(req);
diff --git a/drivers/crypto/mv_cesa.c b/drivers/crypto/mv_cesa.c
index e095422..7d279e5 100644
--- a/drivers/crypto/mv_cesa.c
+++ b/drivers/crypto/mv_cesa.c
@@ -1055,20 +1055,20 @@
 	cp->queue_th = kthread_run(queue_manag, cp, "mv_crypto");
 	if (IS_ERR(cp->queue_th)) {
 		ret = PTR_ERR(cp->queue_th);
-		goto err_thread;
+		goto err_unmap_sram;
 	}
 
 	ret = request_irq(irq, crypto_int, IRQF_DISABLED, dev_name(&pdev->dev),
 			cp);
 	if (ret)
-		goto err_unmap_sram;
+		goto err_thread;
 
 	writel(SEC_INT_ACCEL0_DONE, cpg->reg + SEC_ACCEL_INT_MASK);
 	writel(SEC_CFG_STOP_DIG_ERR, cpg->reg + SEC_ACCEL_CFG);
 
 	ret = crypto_register_alg(&mv_aes_alg_ecb);
 	if (ret)
-		goto err_reg;
+		goto err_irq;
 
 	ret = crypto_register_alg(&mv_aes_alg_cbc);
 	if (ret)
@@ -1091,9 +1091,9 @@
 	return 0;
 err_unreg_ecb:
 	crypto_unregister_alg(&mv_aes_alg_ecb);
-err_thread:
+err_irq:
 	free_irq(irq, cp);
-err_reg:
+err_thread:
 	kthread_stop(cp->queue_th);
 err_unmap_sram:
 	iounmap(cp->sram);
diff --git a/drivers/crypto/n2_core.c b/drivers/crypto/n2_core.c
index 23163fd..b99c38f 100644
--- a/drivers/crypto/n2_core.c
+++ b/drivers/crypto/n2_core.c
@@ -239,21 +239,57 @@
 }
 #endif
 
-struct n2_base_ctx {
-	struct list_head		list;
+struct n2_ahash_alg {
+	struct list_head	entry;
+	const char		*hash_zero;
+	const u32		*hash_init;
+	u8			hw_op_hashsz;
+	u8			digest_size;
+	u8			auth_type;
+	u8			hmac_type;
+	struct ahash_alg	alg;
 };
 
-static void n2_base_ctx_init(struct n2_base_ctx *ctx)
+static inline struct n2_ahash_alg *n2_ahash_alg(struct crypto_tfm *tfm)
 {
-	INIT_LIST_HEAD(&ctx->list);
+	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 n2_ahash_alg, alg);
+}
+
+struct n2_hmac_alg {
+	const char		*child_alg;
+	struct n2_ahash_alg	derived;
+};
+
+static inline struct n2_hmac_alg *n2_hmac_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 n2_hmac_alg, derived.alg);
 }
 
 struct n2_hash_ctx {
-	struct n2_base_ctx		base;
-
 	struct crypto_ahash		*fallback_tfm;
 };
 
+#define N2_HASH_KEY_MAX			32 /* HW limit for all HMAC requests */
+
+struct n2_hmac_ctx {
+	struct n2_hash_ctx		base;
+
+	struct crypto_shash		*child_shash;
+
+	int				hash_key_len;
+	unsigned char			hash_key[N2_HASH_KEY_MAX];
+};
+
 struct n2_hash_req_ctx {
 	union {
 		struct md5_state	md5;
@@ -261,9 +297,6 @@
 		struct sha256_state	sha256;
 	} u;
 
-	unsigned char			hash_key[64];
-	unsigned char			keyed_zero_hash[32];
-
 	struct ahash_request		fallback_req;
 };
 
@@ -356,6 +389,94 @@
 	crypto_free_ahash(ctx->fallback_tfm);
 }
 
+static int n2_hmac_cra_init(struct crypto_tfm *tfm)
+{
+	const char *fallback_driver_name = tfm->__crt_alg->cra_name;
+	struct crypto_ahash *ahash = __crypto_ahash_cast(tfm);
+	struct n2_hmac_ctx *ctx = crypto_ahash_ctx(ahash);
+	struct n2_hmac_alg *n2alg = n2_hmac_alg(tfm);
+	struct crypto_ahash *fallback_tfm;
+	struct crypto_shash *child_shash;
+	int err;
+
+	fallback_tfm = crypto_alloc_ahash(fallback_driver_name, 0,
+					  CRYPTO_ALG_NEED_FALLBACK);
+	if (IS_ERR(fallback_tfm)) {
+		pr_warning("Fallback driver '%s' could not be loaded!\n",
+			   fallback_driver_name);
+		err = PTR_ERR(fallback_tfm);
+		goto out;
+	}
+
+	child_shash = crypto_alloc_shash(n2alg->child_alg, 0, 0);
+	if (IS_ERR(child_shash)) {
+		pr_warning("Child shash '%s' could not be loaded!\n",
+			   n2alg->child_alg);
+		err = PTR_ERR(child_shash);
+		goto out_free_fallback;
+	}
+
+	crypto_ahash_set_reqsize(ahash, (sizeof(struct n2_hash_req_ctx) +
+					 crypto_ahash_reqsize(fallback_tfm)));
+
+	ctx->child_shash = child_shash;
+	ctx->base.fallback_tfm = fallback_tfm;
+	return 0;
+
+out_free_fallback:
+	crypto_free_ahash(fallback_tfm);
+
+out:
+	return err;
+}
+
+static void n2_hmac_cra_exit(struct crypto_tfm *tfm)
+{
+	struct crypto_ahash *ahash = __crypto_ahash_cast(tfm);
+	struct n2_hmac_ctx *ctx = crypto_ahash_ctx(ahash);
+
+	crypto_free_ahash(ctx->base.fallback_tfm);
+	crypto_free_shash(ctx->child_shash);
+}
+
+static int n2_hmac_async_setkey(struct crypto_ahash *tfm, const u8 *key,
+				unsigned int keylen)
+{
+	struct n2_hmac_ctx *ctx = crypto_ahash_ctx(tfm);
+	struct crypto_shash *child_shash = ctx->child_shash;
+	struct crypto_ahash *fallback_tfm;
+	struct {
+		struct shash_desc shash;
+		char ctx[crypto_shash_descsize(child_shash)];
+	} desc;
+	int err, bs, ds;
+
+	fallback_tfm = ctx->base.fallback_tfm;
+	err = crypto_ahash_setkey(fallback_tfm, key, keylen);
+	if (err)
+		return err;
+
+	desc.shash.tfm = child_shash;
+	desc.shash.flags = crypto_ahash_get_flags(tfm) &
+		CRYPTO_TFM_REQ_MAY_SLEEP;
+
+	bs = crypto_shash_blocksize(child_shash);
+	ds = crypto_shash_digestsize(child_shash);
+	BUG_ON(ds > N2_HASH_KEY_MAX);
+	if (keylen > bs) {
+		err = crypto_shash_digest(&desc.shash, key, keylen,
+					  ctx->hash_key);
+		if (err)
+			return err;
+		keylen = ds;
+	} else if (keylen <= N2_HASH_KEY_MAX)
+		memcpy(ctx->hash_key, key, keylen);
+
+	ctx->hash_key_len = keylen;
+
+	return err;
+}
+
 static unsigned long wait_for_tail(struct spu_queue *qp)
 {
 	unsigned long head, hv_ret;
@@ -385,12 +506,12 @@
 	return hv_ret;
 }
 
-static int n2_hash_async_digest(struct ahash_request *req,
-				unsigned int auth_type, unsigned int digest_size,
-				unsigned int result_size, void *hash_loc)
+static int n2_do_async_digest(struct ahash_request *req,
+			      unsigned int auth_type, unsigned int digest_size,
+			      unsigned int result_size, void *hash_loc,
+			      unsigned long auth_key, unsigned int auth_key_len)
 {
 	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
-	struct n2_hash_ctx *ctx = crypto_ahash_ctx(tfm);
 	struct cwq_initial_entry *ent;
 	struct crypto_hash_walk walk;
 	struct spu_queue *qp;
@@ -403,6 +524,7 @@
 	 */
 	if (unlikely(req->nbytes > (1 << 16))) {
 		struct n2_hash_req_ctx *rctx = ahash_request_ctx(req);
+		struct n2_hash_ctx *ctx = crypto_ahash_ctx(tfm);
 
 		ahash_request_set_tfm(&rctx->fallback_req, ctx->fallback_tfm);
 		rctx->fallback_req.base.flags =
@@ -414,8 +536,6 @@
 		return crypto_ahash_digest(&rctx->fallback_req);
 	}
 
-	n2_base_ctx_init(&ctx->base);
-
 	nbytes = crypto_hash_walk_first(req, &walk);
 
 	cpu = get_cpu();
@@ -430,13 +550,13 @@
 	 */
 	ent = qp->q + qp->tail;
 
-	ent->control = control_word_base(nbytes, 0, 0,
+	ent->control = control_word_base(nbytes, auth_key_len, 0,
 					 auth_type, digest_size,
 					 false, true, false, false,
 					 OPCODE_INPLACE_BIT |
 					 OPCODE_AUTH_MAC);
 	ent->src_addr = __pa(walk.data);
-	ent->auth_key_addr = 0UL;
+	ent->auth_key_addr = auth_key;
 	ent->auth_iv_addr = __pa(hash_loc);
 	ent->final_auth_state_addr = 0UL;
 	ent->enc_key_addr = 0UL;
@@ -475,114 +595,55 @@
 	return err;
 }
 
-static int n2_md5_async_digest(struct ahash_request *req)
+static int n2_hash_async_digest(struct ahash_request *req)
 {
+	struct n2_ahash_alg *n2alg = n2_ahash_alg(req->base.tfm);
 	struct n2_hash_req_ctx *rctx = ahash_request_ctx(req);
-	struct md5_state *m = &rctx->u.md5;
+	int ds;
 
+	ds = n2alg->digest_size;
 	if (unlikely(req->nbytes == 0)) {
-		static const char md5_zero[MD5_DIGEST_SIZE] = {
-			0xd4, 0x1d, 0x8c, 0xd9, 0x8f, 0x00, 0xb2, 0x04,
-			0xe9, 0x80, 0x09, 0x98, 0xec, 0xf8, 0x42, 0x7e,
-		};
-
-		memcpy(req->result, md5_zero, MD5_DIGEST_SIZE);
+		memcpy(req->result, n2alg->hash_zero, ds);
 		return 0;
 	}
-	m->hash[0] = cpu_to_le32(0x67452301);
-	m->hash[1] = cpu_to_le32(0xefcdab89);
-	m->hash[2] = cpu_to_le32(0x98badcfe);
-	m->hash[3] = cpu_to_le32(0x10325476);
+	memcpy(&rctx->u, n2alg->hash_init, n2alg->hw_op_hashsz);
 
-	return n2_hash_async_digest(req, AUTH_TYPE_MD5,
-				    MD5_DIGEST_SIZE, MD5_DIGEST_SIZE,
-				    m->hash);
+	return n2_do_async_digest(req, n2alg->auth_type,
+				  n2alg->hw_op_hashsz, ds,
+				  &rctx->u, 0UL, 0);
 }
 
-static int n2_sha1_async_digest(struct ahash_request *req)
+static int n2_hmac_async_digest(struct ahash_request *req)
 {
+	struct n2_hmac_alg *n2alg = n2_hmac_alg(req->base.tfm);
 	struct n2_hash_req_ctx *rctx = ahash_request_ctx(req);
-	struct sha1_state *s = &rctx->u.sha1;
+	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
+	struct n2_hmac_ctx *ctx = crypto_ahash_ctx(tfm);
+	int ds;
 
-	if (unlikely(req->nbytes == 0)) {
-		static const char sha1_zero[SHA1_DIGEST_SIZE] = {
-			0xda, 0x39, 0xa3, 0xee, 0x5e, 0x6b, 0x4b, 0x0d, 0x32,
-			0x55, 0xbf, 0xef, 0x95, 0x60, 0x18, 0x90, 0xaf, 0xd8,
-			0x07, 0x09
-		};
+	ds = n2alg->derived.digest_size;
+	if (unlikely(req->nbytes == 0) ||
+	    unlikely(ctx->hash_key_len > N2_HASH_KEY_MAX)) {
+		struct n2_hash_req_ctx *rctx = ahash_request_ctx(req);
+		struct n2_hash_ctx *ctx = crypto_ahash_ctx(tfm);
 
-		memcpy(req->result, sha1_zero, SHA1_DIGEST_SIZE);
-		return 0;
+		ahash_request_set_tfm(&rctx->fallback_req, ctx->fallback_tfm);
+		rctx->fallback_req.base.flags =
+			req->base.flags & CRYPTO_TFM_REQ_MAY_SLEEP;
+		rctx->fallback_req.nbytes = req->nbytes;
+		rctx->fallback_req.src = req->src;
+		rctx->fallback_req.result = req->result;
+
+		return crypto_ahash_digest(&rctx->fallback_req);
 	}
-	s->state[0] = SHA1_H0;
-	s->state[1] = SHA1_H1;
-	s->state[2] = SHA1_H2;
-	s->state[3] = SHA1_H3;
-	s->state[4] = SHA1_H4;
+	memcpy(&rctx->u, n2alg->derived.hash_init,
+	       n2alg->derived.hw_op_hashsz);
 
-	return n2_hash_async_digest(req, AUTH_TYPE_SHA1,
-				    SHA1_DIGEST_SIZE, SHA1_DIGEST_SIZE,
-				    s->state);
-}
-
-static int n2_sha256_async_digest(struct ahash_request *req)
-{
-	struct n2_hash_req_ctx *rctx = ahash_request_ctx(req);
-	struct sha256_state *s = &rctx->u.sha256;
-
-	if (req->nbytes == 0) {
-		static const char sha256_zero[SHA256_DIGEST_SIZE] = {
-			0xe3, 0xb0, 0xc4, 0x42, 0x98, 0xfc, 0x1c, 0x14, 0x9a,
-			0xfb, 0xf4, 0xc8, 0x99, 0x6f, 0xb9, 0x24, 0x27, 0xae,
-			0x41, 0xe4, 0x64, 0x9b, 0x93, 0x4c, 0xa4, 0x95, 0x99,
-			0x1b, 0x78, 0x52, 0xb8, 0x55
-		};
-
-		memcpy(req->result, sha256_zero, SHA256_DIGEST_SIZE);
-		return 0;
-	}
-	s->state[0] = SHA256_H0;
-	s->state[1] = SHA256_H1;
-	s->state[2] = SHA256_H2;
-	s->state[3] = SHA256_H3;
-	s->state[4] = SHA256_H4;
-	s->state[5] = SHA256_H5;
-	s->state[6] = SHA256_H6;
-	s->state[7] = SHA256_H7;
-
-	return n2_hash_async_digest(req, AUTH_TYPE_SHA256,
-				    SHA256_DIGEST_SIZE, SHA256_DIGEST_SIZE,
-				    s->state);
-}
-
-static int n2_sha224_async_digest(struct ahash_request *req)
-{
-	struct n2_hash_req_ctx *rctx = ahash_request_ctx(req);
-	struct sha256_state *s = &rctx->u.sha256;
-
-	if (req->nbytes == 0) {
-		static const char sha224_zero[SHA224_DIGEST_SIZE] = {
-			0xd1, 0x4a, 0x02, 0x8c, 0x2a, 0x3a, 0x2b, 0xc9, 0x47,
-			0x61, 0x02, 0xbb, 0x28, 0x82, 0x34, 0xc4, 0x15, 0xa2,
-			0xb0, 0x1f, 0x82, 0x8e, 0xa6, 0x2a, 0xc5, 0xb3, 0xe4,
-			0x2f
-		};
-
-		memcpy(req->result, sha224_zero, SHA224_DIGEST_SIZE);
-		return 0;
-	}
-	s->state[0] = SHA224_H0;
-	s->state[1] = SHA224_H1;
-	s->state[2] = SHA224_H2;
-	s->state[3] = SHA224_H3;
-	s->state[4] = SHA224_H4;
-	s->state[5] = SHA224_H5;
-	s->state[6] = SHA224_H6;
-	s->state[7] = SHA224_H7;
-
-	return n2_hash_async_digest(req, AUTH_TYPE_SHA256,
-				    SHA256_DIGEST_SIZE, SHA224_DIGEST_SIZE,
-				    s->state);
+	return n2_do_async_digest(req, n2alg->derived.hmac_type,
+				  n2alg->derived.hw_op_hashsz, ds,
+				  &rctx->u,
+				  __pa(&ctx->hash_key),
+				  ctx->hash_key_len);
 }
 
 struct n2_cipher_context {
@@ -1209,35 +1270,92 @@
 
 struct n2_hash_tmpl {
 	const char	*name;
-	int		(*digest)(struct ahash_request *req);
+	const char	*hash_zero;
+	const u32	*hash_init;
+	u8		hw_op_hashsz;
 	u8		digest_size;
 	u8		block_size;
+	u8		auth_type;
+	u8		hmac_type;
 };
+
+static const char md5_zero[MD5_DIGEST_SIZE] = {
+	0xd4, 0x1d, 0x8c, 0xd9, 0x8f, 0x00, 0xb2, 0x04,
+	0xe9, 0x80, 0x09, 0x98, 0xec, 0xf8, 0x42, 0x7e,
+};
+static const u32 md5_init[MD5_HASH_WORDS] = {
+	cpu_to_le32(0x67452301),
+	cpu_to_le32(0xefcdab89),
+	cpu_to_le32(0x98badcfe),
+	cpu_to_le32(0x10325476),
+};
+static const char sha1_zero[SHA1_DIGEST_SIZE] = {
+	0xda, 0x39, 0xa3, 0xee, 0x5e, 0x6b, 0x4b, 0x0d, 0x32,
+	0x55, 0xbf, 0xef, 0x95, 0x60, 0x18, 0x90, 0xaf, 0xd8,
+	0x07, 0x09
+};
+static const u32 sha1_init[SHA1_DIGEST_SIZE / 4] = {
+	SHA1_H0, SHA1_H1, SHA1_H2, SHA1_H3, SHA1_H4,
+};
+static const char sha256_zero[SHA256_DIGEST_SIZE] = {
+	0xe3, 0xb0, 0xc4, 0x42, 0x98, 0xfc, 0x1c, 0x14, 0x9a,
+	0xfb, 0xf4, 0xc8, 0x99, 0x6f, 0xb9, 0x24, 0x27, 0xae,
+	0x41, 0xe4, 0x64, 0x9b, 0x93, 0x4c, 0xa4, 0x95, 0x99,
+	0x1b, 0x78, 0x52, 0xb8, 0x55
+};
+static const u32 sha256_init[SHA256_DIGEST_SIZE / 4] = {
+	SHA256_H0, SHA256_H1, SHA256_H2, SHA256_H3,
+	SHA256_H4, SHA256_H5, SHA256_H6, SHA256_H7,
+};
+static const char sha224_zero[SHA224_DIGEST_SIZE] = {
+	0xd1, 0x4a, 0x02, 0x8c, 0x2a, 0x3a, 0x2b, 0xc9, 0x47,
+	0x61, 0x02, 0xbb, 0x28, 0x82, 0x34, 0xc4, 0x15, 0xa2,
+	0xb0, 0x1f, 0x82, 0x8e, 0xa6, 0x2a, 0xc5, 0xb3, 0xe4,
+	0x2f
+};
+static const u32 sha224_init[SHA256_DIGEST_SIZE / 4] = {
+	SHA224_H0, SHA224_H1, SHA224_H2, SHA224_H3,
+	SHA224_H4, SHA224_H5, SHA224_H6, SHA224_H7,
+};
+
 static const struct n2_hash_tmpl hash_tmpls[] = {
 	{ .name		= "md5",
-	  .digest	= n2_md5_async_digest,
+	  .hash_zero	= md5_zero,
+	  .hash_init	= md5_init,
+	  .auth_type	= AUTH_TYPE_MD5,
+	  .hmac_type	= AUTH_TYPE_HMAC_MD5,
+	  .hw_op_hashsz	= MD5_DIGEST_SIZE,
 	  .digest_size	= MD5_DIGEST_SIZE,
 	  .block_size	= MD5_HMAC_BLOCK_SIZE },
 	{ .name		= "sha1",
-	  .digest	= n2_sha1_async_digest,
+	  .hash_zero	= sha1_zero,
+	  .hash_init	= sha1_init,
+	  .auth_type	= AUTH_TYPE_SHA1,
+	  .hmac_type	= AUTH_TYPE_HMAC_SHA1,
+	  .hw_op_hashsz	= SHA1_DIGEST_SIZE,
 	  .digest_size	= SHA1_DIGEST_SIZE,
 	  .block_size	= SHA1_BLOCK_SIZE },
 	{ .name		= "sha256",
-	  .digest	= n2_sha256_async_digest,
+	  .hash_zero	= sha256_zero,
+	  .hash_init	= sha256_init,
+	  .auth_type	= AUTH_TYPE_SHA256,
+	  .hmac_type	= AUTH_TYPE_HMAC_SHA256,
+	  .hw_op_hashsz	= SHA256_DIGEST_SIZE,
 	  .digest_size	= SHA256_DIGEST_SIZE,
 	  .block_size	= SHA256_BLOCK_SIZE },
 	{ .name		= "sha224",
-	  .digest	= n2_sha224_async_digest,
+	  .hash_zero	= sha224_zero,
+	  .hash_init	= sha224_init,
+	  .auth_type	= AUTH_TYPE_SHA256,
+	  .hmac_type	= AUTH_TYPE_RESERVED,
+	  .hw_op_hashsz	= SHA256_DIGEST_SIZE,
 	  .digest_size	= SHA224_DIGEST_SIZE,
 	  .block_size	= SHA224_BLOCK_SIZE },
 };
 #define NUM_HASH_TMPLS ARRAY_SIZE(hash_tmpls)
 
-struct n2_ahash_alg {
-	struct list_head	entry;
-	struct ahash_alg	alg;
-};
 static LIST_HEAD(ahash_algs);
+static LIST_HEAD(hmac_algs);
 
 static int algs_registered;
 
@@ -1245,12 +1363,18 @@
 {
 	struct n2_cipher_alg *cipher, *cipher_tmp;
 	struct n2_ahash_alg *alg, *alg_tmp;
+	struct n2_hmac_alg *hmac, *hmac_tmp;
 
 	list_for_each_entry_safe(cipher, cipher_tmp, &cipher_algs, entry) {
 		crypto_unregister_alg(&cipher->alg);
 		list_del(&cipher->entry);
 		kfree(cipher);
 	}
+	list_for_each_entry_safe(hmac, hmac_tmp, &hmac_algs, derived.entry) {
+		crypto_unregister_ahash(&hmac->derived.alg);
+		list_del(&hmac->derived.entry);
+		kfree(hmac);
+	}
 	list_for_each_entry_safe(alg, alg_tmp, &ahash_algs, entry) {
 		crypto_unregister_ahash(&alg->alg);
 		list_del(&alg->entry);
@@ -1290,8 +1414,49 @@
 	list_add(&p->entry, &cipher_algs);
 	err = crypto_register_alg(alg);
 	if (err) {
+		pr_err("%s alg registration failed\n", alg->cra_name);
 		list_del(&p->entry);
 		kfree(p);
+	} else {
+		pr_info("%s alg registered\n", alg->cra_name);
+	}
+	return err;
+}
+
+static int __devinit __n2_register_one_hmac(struct n2_ahash_alg *n2ahash)
+{
+	struct n2_hmac_alg *p = kzalloc(sizeof(*p), GFP_KERNEL);
+	struct ahash_alg *ahash;
+	struct crypto_alg *base;
+	int err;
+
+	if (!p)
+		return -ENOMEM;
+
+	p->child_alg = n2ahash->alg.halg.base.cra_name;
+	memcpy(&p->derived, n2ahash, sizeof(struct n2_ahash_alg));
+	INIT_LIST_HEAD(&p->derived.entry);
+
+	ahash = &p->derived.alg;
+	ahash->digest = n2_hmac_async_digest;
+	ahash->setkey = n2_hmac_async_setkey;
+
+	base = &ahash->halg.base;
+	snprintf(base->cra_name, CRYPTO_MAX_ALG_NAME, "hmac(%s)", p->child_alg);
+	snprintf(base->cra_driver_name, CRYPTO_MAX_ALG_NAME, "hmac-%s-n2", p->child_alg);
+
+	base->cra_ctxsize = sizeof(struct n2_hmac_ctx);
+	base->cra_init = n2_hmac_cra_init;
+	base->cra_exit = n2_hmac_cra_exit;
+
+	list_add(&p->derived.entry, &hmac_algs);
+	err = crypto_register_ahash(ahash);
+	if (err) {
+		pr_err("%s alg registration failed\n", base->cra_name);
+		list_del(&p->derived.entry);
+		kfree(p);
+	} else {
+		pr_info("%s alg registered\n", base->cra_name);
 	}
 	return err;
 }
@@ -1307,12 +1472,19 @@
 	if (!p)
 		return -ENOMEM;
 
+	p->hash_zero = tmpl->hash_zero;
+	p->hash_init = tmpl->hash_init;
+	p->auth_type = tmpl->auth_type;
+	p->hmac_type = tmpl->hmac_type;
+	p->hw_op_hashsz = tmpl->hw_op_hashsz;
+	p->digest_size = tmpl->digest_size;
+
 	ahash = &p->alg;
 	ahash->init = n2_hash_async_init;
 	ahash->update = n2_hash_async_update;
 	ahash->final = n2_hash_async_final;
 	ahash->finup = n2_hash_async_finup;
-	ahash->digest = tmpl->digest;
+	ahash->digest = n2_hash_async_digest;
 
 	halg = &ahash->halg;
 	halg->digestsize = tmpl->digest_size;
@@ -1331,9 +1503,14 @@
 	list_add(&p->entry, &ahash_algs);
 	err = crypto_register_ahash(ahash);
 	if (err) {
+		pr_err("%s alg registration failed\n", base->cra_name);
 		list_del(&p->entry);
 		kfree(p);
+	} else {
+		pr_info("%s alg registered\n", base->cra_name);
 	}
+	if (!err && p->hmac_type != AUTH_TYPE_RESERVED)
+		err = __n2_register_one_hmac(p);
 	return err;
 }
 
diff --git a/drivers/crypto/omap-sham.c b/drivers/crypto/omap-sham.c
index 8b03433..7d14856 100644
--- a/drivers/crypto/omap-sham.c
+++ b/drivers/crypto/omap-sham.c
@@ -15,7 +15,6 @@
 
 #define pr_fmt(fmt) "%s: " fmt, __func__
 
-#include <linux/version.h>
 #include <linux/err.h>
 #include <linux/device.h>
 #include <linux/module.h>
diff --git a/drivers/crypto/talitos.c b/drivers/crypto/talitos.c
index bd78acf..97f4af1 100644
--- a/drivers/crypto/talitos.c
+++ b/drivers/crypto/talitos.c
@@ -720,7 +720,6 @@
 #define TALITOS_MDEU_MAX_CONTEXT_SIZE	TALITOS_MDEU_CONTEXT_SIZE_SHA384_SHA512
 
 struct talitos_ahash_req_ctx {
-	u64 count;
 	u32 hw_context[TALITOS_MDEU_MAX_CONTEXT_SIZE / sizeof(u32)];
 	unsigned int hw_context_size;
 	u8 buf[HASH_MAX_BLOCK_SIZE];
@@ -729,6 +728,7 @@
 	unsigned int first;
 	unsigned int last;
 	unsigned int to_hash_later;
+	u64 nbuf;
 	struct scatterlist bufsl[2];
 	struct scatterlist *psrc;
 };
@@ -1613,6 +1613,7 @@
 	if (!req_ctx->last && req_ctx->to_hash_later) {
 		/* Position any partial block for next update/final/finup */
 		memcpy(req_ctx->buf, req_ctx->bufnext, req_ctx->to_hash_later);
+		req_ctx->nbuf = req_ctx->to_hash_later;
 	}
 	common_nonsnoop_hash_unmap(dev, edesc, areq);
 
@@ -1728,7 +1729,7 @@
 	struct talitos_ahash_req_ctx *req_ctx = ahash_request_ctx(areq);
 
 	/* Initialize the context */
-	req_ctx->count = 0;
+	req_ctx->nbuf = 0;
 	req_ctx->first = 1; /* first indicates h/w must init its context */
 	req_ctx->swinit = 0; /* assume h/w init of context */
 	req_ctx->hw_context_size =
@@ -1776,52 +1777,54 @@
 			crypto_tfm_alg_blocksize(crypto_ahash_tfm(tfm));
 	unsigned int nbytes_to_hash;
 	unsigned int to_hash_later;
-	unsigned int index;
+	unsigned int nsg;
 	int chained;
 
-	index = req_ctx->count & (blocksize - 1);
-	req_ctx->count += nbytes;
-
-	if (!req_ctx->last && (index + nbytes) < blocksize) {
-		/* Buffer the partial block */
+	if (!req_ctx->last && (nbytes + req_ctx->nbuf <= blocksize)) {
+		/* Buffer up to one whole block */
 		sg_copy_to_buffer(areq->src,
 				  sg_count(areq->src, nbytes, &chained),
-				  req_ctx->buf + index, nbytes);
+				  req_ctx->buf + req_ctx->nbuf, nbytes);
+		req_ctx->nbuf += nbytes;
 		return 0;
 	}
 
-	if (index) {
-		/* partial block from previous update; chain it in. */
-		sg_init_table(req_ctx->bufsl, (nbytes) ? 2 : 1);
-		sg_set_buf(req_ctx->bufsl, req_ctx->buf, index);
-		if (nbytes)
-			scatterwalk_sg_chain(req_ctx->bufsl, 2,
-					     areq->src);
+	/* At least (blocksize + 1) bytes are available to hash */
+	nbytes_to_hash = nbytes + req_ctx->nbuf;
+	to_hash_later = nbytes_to_hash & (blocksize - 1);
+
+	if (req_ctx->last)
+		to_hash_later = 0;
+	else if (to_hash_later)
+		/* There is a partial block. Hash the full block(s) now */
+		nbytes_to_hash -= to_hash_later;
+	else {
+		/* Keep one block buffered */
+		nbytes_to_hash -= blocksize;
+		to_hash_later = blocksize;
+	}
+
+	/* Chain in any previously buffered data */
+	if (req_ctx->nbuf) {
+		nsg = (req_ctx->nbuf < nbytes_to_hash) ? 2 : 1;
+		sg_init_table(req_ctx->bufsl, nsg);
+		sg_set_buf(req_ctx->bufsl, req_ctx->buf, req_ctx->nbuf);
+		if (nsg > 1)
+			scatterwalk_sg_chain(req_ctx->bufsl, 2, areq->src);
 		req_ctx->psrc = req_ctx->bufsl;
-	} else {
+	} else
 		req_ctx->psrc = areq->src;
-	}
-	nbytes_to_hash =  index + nbytes;
-	if (!req_ctx->last) {
-		to_hash_later = (nbytes_to_hash & (blocksize - 1));
-		if (to_hash_later) {
-			int nents;
-			/* Must copy to_hash_later bytes from the end
-			 * to bufnext (a partial block) for later.
-			 */
-			nents = sg_count(areq->src, nbytes, &chained);
-			sg_copy_end_to_buffer(areq->src, nents,
-					      req_ctx->bufnext,
-					      to_hash_later,
-					      nbytes - to_hash_later);
 
-			/* Adjust count for what will be hashed now */
-			nbytes_to_hash -= to_hash_later;
-		}
-		req_ctx->to_hash_later = to_hash_later;
+	if (to_hash_later) {
+		int nents = sg_count(areq->src, nbytes, &chained);
+		sg_copy_end_to_buffer(areq->src, nents,
+				      req_ctx->bufnext,
+				      to_hash_later,
+				      nbytes - to_hash_later);
 	}
+	req_ctx->to_hash_later = to_hash_later;
 
-	/* allocate extended descriptor */
+	/* Allocate extended descriptor */
 	edesc = ahash_edesc_alloc(areq, nbytes_to_hash);
 	if (IS_ERR(edesc))
 		return PTR_ERR(edesc);
diff --git a/include/linux/padata.h b/include/linux/padata.h
index 8d84062..bdcd1e9 100644
--- a/include/linux/padata.h
+++ b/include/linux/padata.h
@@ -25,6 +25,11 @@
 #include <linux/spinlock.h>
 #include <linux/list.h>
 #include <linux/timer.h>
+#include <linux/notifier.h>
+#include <linux/kobject.h>
+
+#define PADATA_CPU_SERIAL   0x01
+#define PADATA_CPU_PARALLEL 0x02
 
 /**
  * struct padata_priv -  Embedded to the users data structure.
@@ -59,7 +64,20 @@
 };
 
 /**
- * struct padata_queue - The percpu padata queues.
+* struct padata_serial_queue - The percpu padata serial queue
+*
+* @serial: List to wait for serialization after reordering.
+* @work: work struct for serialization.
+* @pd: Backpointer to the internal control structure.
+*/
+struct padata_serial_queue {
+       struct padata_list    serial;
+       struct work_struct    work;
+       struct parallel_data *pd;
+};
+
+/**
+ * struct padata_parallel_queue - The percpu padata parallel queue
  *
  * @parallel: List to wait for parallelization.
  * @reorder: List to wait for reordering after parallel processing.
@@ -67,18 +85,28 @@
  * @pwork: work struct for parallelization.
  * @swork: work struct for serialization.
  * @pd: Backpointer to the internal control structure.
+ * @work: work struct for parallelization.
  * @num_obj: Number of objects that are processed by this cpu.
  * @cpu_index: Index of the cpu.
  */
-struct padata_queue {
-	struct padata_list	parallel;
-	struct padata_list	reorder;
-	struct padata_list	serial;
-	struct work_struct	pwork;
-	struct work_struct	swork;
-	struct parallel_data    *pd;
-	atomic_t		num_obj;
-	int			cpu_index;
+struct padata_parallel_queue {
+       struct padata_list    parallel;
+       struct padata_list    reorder;
+       struct parallel_data *pd;
+       struct work_struct    work;
+       atomic_t              num_obj;
+       int                   cpu_index;
+};
+
+/**
+ * struct padata_cpumask - The cpumasks for the parallel/serial workers
+ *
+ * @pcpu: cpumask for the parallel workers.
+ * @cbcpu: cpumask for the serial (callback) workers.
+ */
+struct padata_cpumask {
+	cpumask_var_t	pcpu;
+	cpumask_var_t	cbcpu;
 };
 
 /**
@@ -86,25 +114,29 @@
  * that depends on the cpumask in use.
  *
  * @pinst: padata instance.
- * @queue: percpu padata queues.
+ * @pqueue: percpu padata queues used for parallelization.
+ * @squeue: percpu padata queues used for serialuzation.
  * @seq_nr: The sequence number that will be attached to the next object.
  * @reorder_objects: Number of objects waiting in the reorder queues.
  * @refcnt: Number of objects holding a reference on this parallel_data.
  * @max_seq_nr:  Maximal used sequence number.
- * @cpumask: cpumask in use.
+ * @cpumask: The cpumasks in use for parallel and serial workers.
  * @lock: Reorder lock.
+ * @processed: Number of already processed objects.
  * @timer: Reorder timer.
  */
 struct parallel_data {
-	struct padata_instance	*pinst;
-	struct padata_queue	*queue;
-	atomic_t		seq_nr;
-	atomic_t		reorder_objects;
-	atomic_t                refcnt;
-	unsigned int		max_seq_nr;
-	cpumask_var_t		cpumask;
-	spinlock_t              lock;
-	struct timer_list       timer;
+	struct padata_instance		*pinst;
+	struct padata_parallel_queue	*pqueue;
+	struct padata_serial_queue	*squeue;
+	atomic_t			seq_nr;
+	atomic_t			reorder_objects;
+	atomic_t			refcnt;
+	unsigned int			max_seq_nr;
+	struct padata_cpumask		cpumask;
+	spinlock_t                      lock ____cacheline_aligned;
+	unsigned int			processed;
+	struct timer_list		timer;
 };
 
 /**
@@ -113,31 +145,48 @@
  * @cpu_notifier: cpu hotplug notifier.
  * @wq: The workqueue in use.
  * @pd: The internal control structure.
- * @cpumask: User supplied cpumask.
+ * @cpumask: User supplied cpumasks for parallel and serial works.
+ * @cpumask_change_notifier: Notifiers chain for user-defined notify
+ *            callbacks that will be called when either @pcpu or @cbcpu
+ *            or both cpumasks change.
+ * @kobj: padata instance kernel object.
  * @lock: padata instance lock.
  * @flags: padata flags.
  */
 struct padata_instance {
-	struct notifier_block   cpu_notifier;
-	struct workqueue_struct *wq;
-	struct parallel_data	*pd;
-	cpumask_var_t           cpumask;
-	struct mutex		lock;
-	u8			flags;
-#define	PADATA_INIT		1
-#define	PADATA_RESET		2
+	struct notifier_block		 cpu_notifier;
+	struct workqueue_struct		*wq;
+	struct parallel_data		*pd;
+	struct padata_cpumask		cpumask;
+	struct blocking_notifier_head	 cpumask_change_notifier;
+	struct kobject                   kobj;
+	struct mutex			 lock;
+	u8				 flags;
+#define	PADATA_INIT	1
+#define	PADATA_RESET	2
+#define	PADATA_INVALID	4
 };
 
-extern struct padata_instance *padata_alloc(const struct cpumask *cpumask,
-					    struct workqueue_struct *wq);
+extern struct padata_instance *padata_alloc_possible(
+					struct workqueue_struct *wq);
+extern struct padata_instance *padata_alloc(struct workqueue_struct *wq,
+					    const struct cpumask *pcpumask,
+					    const struct cpumask *cbcpumask);
 extern void padata_free(struct padata_instance *pinst);
 extern int padata_do_parallel(struct padata_instance *pinst,
 			      struct padata_priv *padata, int cb_cpu);
 extern void padata_do_serial(struct padata_priv *padata);
-extern int padata_set_cpumask(struct padata_instance *pinst,
+extern int padata_set_cpumask(struct padata_instance *pinst, int cpumask_type,
 			      cpumask_var_t cpumask);
-extern int padata_add_cpu(struct padata_instance *pinst, int cpu);
-extern int padata_remove_cpu(struct padata_instance *pinst, int cpu);
-extern void padata_start(struct padata_instance *pinst);
+extern int padata_set_cpumasks(struct padata_instance *pinst,
+			       cpumask_var_t pcpumask,
+			       cpumask_var_t cbcpumask);
+extern int padata_add_cpu(struct padata_instance *pinst, int cpu, int mask);
+extern int padata_remove_cpu(struct padata_instance *pinst, int cpu, int mask);
+extern int padata_start(struct padata_instance *pinst);
 extern void padata_stop(struct padata_instance *pinst);
+extern int padata_register_cpumask_notifier(struct padata_instance *pinst,
+					    struct notifier_block *nblock);
+extern int padata_unregister_cpumask_notifier(struct padata_instance *pinst,
+					      struct notifier_block *nblock);
 #endif
diff --git a/kernel/padata.c b/kernel/padata.c
index fdd8ae6..7510194 100644
--- a/kernel/padata.c
+++ b/kernel/padata.c
@@ -26,18 +26,19 @@
 #include <linux/mutex.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
+#include <linux/sysfs.h>
 #include <linux/rcupdate.h>
 
-#define MAX_SEQ_NR INT_MAX - NR_CPUS
+#define MAX_SEQ_NR (INT_MAX - NR_CPUS)
 #define MAX_OBJ_NUM 1000
 
 static int padata_index_to_cpu(struct parallel_data *pd, int cpu_index)
 {
 	int cpu, target_cpu;
 
-	target_cpu = cpumask_first(pd->cpumask);
+	target_cpu = cpumask_first(pd->cpumask.pcpu);
 	for (cpu = 0; cpu < cpu_index; cpu++)
-		target_cpu = cpumask_next(target_cpu, pd->cpumask);
+		target_cpu = cpumask_next(target_cpu, pd->cpumask.pcpu);
 
 	return target_cpu;
 }
@@ -53,26 +54,27 @@
 	 * Hash the sequence numbers to the cpus by taking
 	 * seq_nr mod. number of cpus in use.
 	 */
-	cpu_index =  padata->seq_nr % cpumask_weight(pd->cpumask);
+	cpu_index =  padata->seq_nr % cpumask_weight(pd->cpumask.pcpu);
 
 	return padata_index_to_cpu(pd, cpu_index);
 }
 
-static void padata_parallel_worker(struct work_struct *work)
+static void padata_parallel_worker(struct work_struct *parallel_work)
 {
-	struct padata_queue *queue;
+	struct padata_parallel_queue *pqueue;
 	struct parallel_data *pd;
 	struct padata_instance *pinst;
 	LIST_HEAD(local_list);
 
 	local_bh_disable();
-	queue = container_of(work, struct padata_queue, pwork);
-	pd = queue->pd;
+	pqueue = container_of(parallel_work,
+			      struct padata_parallel_queue, work);
+	pd = pqueue->pd;
 	pinst = pd->pinst;
 
-	spin_lock(&queue->parallel.lock);
-	list_replace_init(&queue->parallel.list, &local_list);
-	spin_unlock(&queue->parallel.lock);
+	spin_lock(&pqueue->parallel.lock);
+	list_replace_init(&pqueue->parallel.list, &local_list);
+	spin_unlock(&pqueue->parallel.lock);
 
 	while (!list_empty(&local_list)) {
 		struct padata_priv *padata;
@@ -94,7 +96,7 @@
  * @pinst: padata instance
  * @padata: object to be parallelized
  * @cb_cpu: cpu the serialization callback function will run on,
- *          must be in the cpumask of padata.
+ *          must be in the serial cpumask of padata(i.e. cpumask.cbcpu).
  *
  * The parallelization callback function will run with BHs off.
  * Note: Every object which is parallelized by padata_do_parallel
@@ -104,15 +106,18 @@
 		       struct padata_priv *padata, int cb_cpu)
 {
 	int target_cpu, err;
-	struct padata_queue *queue;
+	struct padata_parallel_queue *queue;
 	struct parallel_data *pd;
 
 	rcu_read_lock_bh();
 
 	pd = rcu_dereference(pinst->pd);
 
-	err = 0;
-	if (!(pinst->flags & PADATA_INIT))
+	err = -EINVAL;
+	if (!(pinst->flags & PADATA_INIT) || pinst->flags & PADATA_INVALID)
+		goto out;
+
+	if (!cpumask_test_cpu(cb_cpu, pd->cpumask.cbcpu))
 		goto out;
 
 	err =  -EBUSY;
@@ -122,11 +127,7 @@
 	if (atomic_read(&pd->refcnt) >= MAX_OBJ_NUM)
 		goto out;
 
-	err = -EINVAL;
-	if (!cpumask_test_cpu(cb_cpu, pd->cpumask))
-		goto out;
-
-	err = -EINPROGRESS;
+	err = 0;
 	atomic_inc(&pd->refcnt);
 	padata->pd = pd;
 	padata->cb_cpu = cb_cpu;
@@ -137,13 +138,13 @@
 	padata->seq_nr = atomic_inc_return(&pd->seq_nr);
 
 	target_cpu = padata_cpu_hash(padata);
-	queue = per_cpu_ptr(pd->queue, target_cpu);
+	queue = per_cpu_ptr(pd->pqueue, target_cpu);
 
 	spin_lock(&queue->parallel.lock);
 	list_add_tail(&padata->list, &queue->parallel.list);
 	spin_unlock(&queue->parallel.lock);
 
-	queue_work_on(target_cpu, pinst->wq, &queue->pwork);
+	queue_work_on(target_cpu, pinst->wq, &queue->work);
 
 out:
 	rcu_read_unlock_bh();
@@ -171,84 +172,52 @@
  */
 static struct padata_priv *padata_get_next(struct parallel_data *pd)
 {
-	int cpu, num_cpus, empty, calc_seq_nr;
-	int seq_nr, next_nr, overrun, next_overrun;
-	struct padata_queue *queue, *next_queue;
+	int cpu, num_cpus;
+	int next_nr, next_index;
+	struct padata_parallel_queue *queue, *next_queue;
 	struct padata_priv *padata;
 	struct padata_list *reorder;
 
-	empty = 0;
-	next_nr = -1;
-	next_overrun = 0;
-	next_queue = NULL;
+	num_cpus = cpumask_weight(pd->cpumask.pcpu);
 
-	num_cpus = cpumask_weight(pd->cpumask);
+	/*
+	 * Calculate the percpu reorder queue and the sequence
+	 * number of the next object.
+	 */
+	next_nr = pd->processed;
+	next_index = next_nr % num_cpus;
+	cpu = padata_index_to_cpu(pd, next_index);
+	next_queue = per_cpu_ptr(pd->pqueue, cpu);
 
-	for_each_cpu(cpu, pd->cpumask) {
-		queue = per_cpu_ptr(pd->queue, cpu);
-		reorder = &queue->reorder;
-
-		/*
-		 * Calculate the seq_nr of the object that should be
-		 * next in this reorder queue.
-		 */
-		overrun = 0;
-		calc_seq_nr = (atomic_read(&queue->num_obj) * num_cpus)
-			       + queue->cpu_index;
-
-		if (unlikely(calc_seq_nr > pd->max_seq_nr)) {
-			calc_seq_nr = calc_seq_nr - pd->max_seq_nr - 1;
-			overrun = 1;
-		}
-
-		if (!list_empty(&reorder->list)) {
-			padata = list_entry(reorder->list.next,
-					    struct padata_priv, list);
-
-			seq_nr  = padata->seq_nr;
-			BUG_ON(calc_seq_nr != seq_nr);
-		} else {
-			seq_nr = calc_seq_nr;
-			empty++;
-		}
-
-		if (next_nr < 0 || seq_nr < next_nr
-		    || (next_overrun && !overrun)) {
-			next_nr = seq_nr;
-			next_overrun = overrun;
-			next_queue = queue;
-		}
+	if (unlikely(next_nr > pd->max_seq_nr)) {
+		next_nr = next_nr - pd->max_seq_nr - 1;
+		next_index = next_nr % num_cpus;
+		cpu = padata_index_to_cpu(pd, next_index);
+		next_queue = per_cpu_ptr(pd->pqueue, cpu);
+		pd->processed = 0;
 	}
 
 	padata = NULL;
 
-	if (empty == num_cpus)
-		goto out;
-
 	reorder = &next_queue->reorder;
 
 	if (!list_empty(&reorder->list)) {
 		padata = list_entry(reorder->list.next,
 				    struct padata_priv, list);
 
-		if (unlikely(next_overrun)) {
-			for_each_cpu(cpu, pd->cpumask) {
-				queue = per_cpu_ptr(pd->queue, cpu);
-				atomic_set(&queue->num_obj, 0);
-			}
-		}
+		BUG_ON(next_nr != padata->seq_nr);
 
 		spin_lock(&reorder->lock);
 		list_del_init(&padata->list);
 		atomic_dec(&pd->reorder_objects);
 		spin_unlock(&reorder->lock);
 
-		atomic_inc(&next_queue->num_obj);
+		pd->processed++;
 
 		goto out;
 	}
 
-	queue = per_cpu_ptr(pd->queue, smp_processor_id());
+	queue = per_cpu_ptr(pd->pqueue, smp_processor_id());
 	if (queue->cpu_index == next_queue->cpu_index) {
 		padata = ERR_PTR(-ENODATA);
 		goto out;
@@ -262,7 +231,7 @@
 static void padata_reorder(struct parallel_data *pd)
 {
 	struct padata_priv *padata;
-	struct padata_queue *queue;
+	struct padata_serial_queue *squeue;
 	struct padata_instance *pinst = pd->pinst;
 
 	/*
@@ -301,13 +270,13 @@
 			return;
 		}
 
-		queue = per_cpu_ptr(pd->queue, padata->cb_cpu);
+		squeue = per_cpu_ptr(pd->squeue, padata->cb_cpu);
 
-		spin_lock(&queue->serial.lock);
-		list_add_tail(&padata->list, &queue->serial.list);
-		spin_unlock(&queue->serial.lock);
+		spin_lock(&squeue->serial.lock);
+		list_add_tail(&padata->list, &squeue->serial.list);
+		spin_unlock(&squeue->serial.lock);
 
-		queue_work_on(padata->cb_cpu, pinst->wq, &queue->swork);
+		queue_work_on(padata->cb_cpu, pinst->wq, &squeue->work);
 	}
 
 	spin_unlock_bh(&pd->lock);
@@ -333,19 +302,19 @@
 	padata_reorder(pd);
 }
 
-static void padata_serial_worker(struct work_struct *work)
+static void padata_serial_worker(struct work_struct *serial_work)
 {
-	struct padata_queue *queue;
+	struct padata_serial_queue *squeue;
 	struct parallel_data *pd;
 	LIST_HEAD(local_list);
 
 	local_bh_disable();
-	queue = container_of(work, struct padata_queue, swork);
-	pd = queue->pd;
+	squeue = container_of(serial_work, struct padata_serial_queue, work);
+	pd = squeue->pd;
 
-	spin_lock(&queue->serial.lock);
-	list_replace_init(&queue->serial.list, &local_list);
-	spin_unlock(&queue->serial.lock);
+	spin_lock(&squeue->serial.lock);
+	list_replace_init(&squeue->serial.list, &local_list);
+	spin_unlock(&squeue->serial.lock);
 
 	while (!list_empty(&local_list)) {
 		struct padata_priv *padata;
@@ -372,18 +341,18 @@
 void padata_do_serial(struct padata_priv *padata)
 {
 	int cpu;
-	struct padata_queue *queue;
+	struct padata_parallel_queue *pqueue;
 	struct parallel_data *pd;
 
 	pd = padata->pd;
 
 	cpu = get_cpu();
-	queue = per_cpu_ptr(pd->queue, cpu);
+	pqueue = per_cpu_ptr(pd->pqueue, cpu);
 
-	spin_lock(&queue->reorder.lock);
+	spin_lock(&pqueue->reorder.lock);
 	atomic_inc(&pd->reorder_objects);
-	list_add_tail(&padata->list, &queue->reorder.list);
-	spin_unlock(&queue->reorder.lock);
+	list_add_tail(&padata->list, &pqueue->reorder.list);
+	spin_unlock(&pqueue->reorder.lock);
 
 	put_cpu();
 
@@ -391,52 +360,89 @@
 }
 EXPORT_SYMBOL(padata_do_serial);
 
-/* Allocate and initialize the internal cpumask dependend resources. */
-static struct parallel_data *padata_alloc_pd(struct padata_instance *pinst,
-					     const struct cpumask *cpumask)
+static int padata_setup_cpumasks(struct parallel_data *pd,
+				 const struct cpumask *pcpumask,
+				 const struct cpumask *cbcpumask)
 {
-	int cpu, cpu_index, num_cpus;
-	struct padata_queue *queue;
-	struct parallel_data *pd;
+	if (!alloc_cpumask_var(&pd->cpumask.pcpu, GFP_KERNEL))
+		return -ENOMEM;
+
+	cpumask_and(pd->cpumask.pcpu, pcpumask, cpu_active_mask);
+	if (!alloc_cpumask_var(&pd->cpumask.cbcpu, GFP_KERNEL)) {
+		free_cpumask_var(pd->cpumask.cbcpu);
+		return -ENOMEM;
+	}
+
+	cpumask_and(pd->cpumask.cbcpu, cbcpumask, cpu_active_mask);
+	return 0;
+}
+
+static void __padata_list_init(struct padata_list *pd_list)
+{
+	INIT_LIST_HEAD(&pd_list->list);
+	spin_lock_init(&pd_list->lock);
+}
+
+/* Initialize all percpu queues used by serial workers */
+static void padata_init_squeues(struct parallel_data *pd)
+{
+	int cpu;
+	struct padata_serial_queue *squeue;
+
+	for_each_cpu(cpu, pd->cpumask.cbcpu) {
+		squeue = per_cpu_ptr(pd->squeue, cpu);
+		squeue->pd = pd;
+		__padata_list_init(&squeue->serial);
+		INIT_WORK(&squeue->work, padata_serial_worker);
+	}
+}
+
+/* Initialize all percpu queues used by parallel workers */
+static void padata_init_pqueues(struct parallel_data *pd)
+{
+	int cpu_index, num_cpus, cpu;
+	struct padata_parallel_queue *pqueue;
 
 	cpu_index = 0;
+	for_each_cpu(cpu, pd->cpumask.pcpu) {
+		pqueue = per_cpu_ptr(pd->pqueue, cpu);
+		pqueue->pd = pd;
+		pqueue->cpu_index = cpu_index;
+		cpu_index++;
+
+		__padata_list_init(&pqueue->reorder);
+		__padata_list_init(&pqueue->parallel);
+		INIT_WORK(&pqueue->work, padata_parallel_worker);
+		atomic_set(&pqueue->num_obj, 0);
+	}
+
+	num_cpus = cpumask_weight(pd->cpumask.pcpu);
+	pd->max_seq_nr = num_cpus ? (MAX_SEQ_NR / num_cpus) * num_cpus - 1 : 0;
+}
+
+/* Allocate and initialize the internal cpumask dependend resources. */
+static struct parallel_data *padata_alloc_pd(struct padata_instance *pinst,
+					     const struct cpumask *pcpumask,
+					     const struct cpumask *cbcpumask)
+{
+	struct parallel_data *pd;
 
 	pd = kzalloc(sizeof(struct parallel_data), GFP_KERNEL);
 	if (!pd)
 		goto err;
 
-	pd->queue = alloc_percpu(struct padata_queue);
-	if (!pd->queue)
+	pd->pqueue = alloc_percpu(struct padata_parallel_queue);
+	if (!pd->pqueue)
 		goto err_free_pd;
 
-	if (!alloc_cpumask_var(&pd->cpumask, GFP_KERNEL))
-		goto err_free_queue;
+	pd->squeue = alloc_percpu(struct padata_serial_queue);
+	if (!pd->squeue)
+		goto err_free_pqueue;
+	if (padata_setup_cpumasks(pd, pcpumask, cbcpumask) < 0)
+		goto err_free_squeue;
 
-	cpumask_and(pd->cpumask, cpumask, cpu_active_mask);
-
-	for_each_cpu(cpu, pd->cpumask) {
-		queue = per_cpu_ptr(pd->queue, cpu);
-
-		queue->pd = pd;
-
-		queue->cpu_index = cpu_index;
-		cpu_index++;
-
-		INIT_LIST_HEAD(&queue->reorder.list);
-		INIT_LIST_HEAD(&queue->parallel.list);
-		INIT_LIST_HEAD(&queue->serial.list);
-		spin_lock_init(&queue->reorder.lock);
-		spin_lock_init(&queue->parallel.lock);
-		spin_lock_init(&queue->serial.lock);
-
-		INIT_WORK(&queue->pwork, padata_parallel_worker);
-		INIT_WORK(&queue->swork, padata_serial_worker);
-		atomic_set(&queue->num_obj, 0);
-	}
-
-	num_cpus = cpumask_weight(pd->cpumask);
-	pd->max_seq_nr = (MAX_SEQ_NR / num_cpus) * num_cpus - 1;
-
+	padata_init_pqueues(pd);
+	padata_init_squeues(pd);
 	setup_timer(&pd->timer, padata_reorder_timer, (unsigned long)pd);
 	atomic_set(&pd->seq_nr, -1);
 	atomic_set(&pd->reorder_objects, 0);
@@ -446,8 +452,10 @@
 
 	return pd;
 
-err_free_queue:
-	free_percpu(pd->queue);
+err_free_squeue:
+	free_percpu(pd->squeue);
+err_free_pqueue:
+	free_percpu(pd->pqueue);
 err_free_pd:
 	kfree(pd);
 err:
@@ -456,8 +464,10 @@
 
 static void padata_free_pd(struct parallel_data *pd)
 {
-	free_cpumask_var(pd->cpumask);
-	free_percpu(pd->queue);
+	free_cpumask_var(pd->cpumask.pcpu);
+	free_cpumask_var(pd->cpumask.cbcpu);
+	free_percpu(pd->pqueue);
+	free_percpu(pd->squeue);
 	kfree(pd);
 }
 
@@ -465,11 +475,12 @@
 static void padata_flush_queues(struct parallel_data *pd)
 {
 	int cpu;
-	struct padata_queue *queue;
+	struct padata_parallel_queue *pqueue;
+	struct padata_serial_queue *squeue;
 
-	for_each_cpu(cpu, pd->cpumask) {
-		queue = per_cpu_ptr(pd->queue, cpu);
-		flush_work(&queue->pwork);
+	for_each_cpu(cpu, pd->cpumask.pcpu) {
+		pqueue = per_cpu_ptr(pd->pqueue, cpu);
+		flush_work(&pqueue->work);
 	}
 
 	del_timer_sync(&pd->timer);
@@ -477,19 +488,39 @@
 	if (atomic_read(&pd->reorder_objects))
 		padata_reorder(pd);
 
-	for_each_cpu(cpu, pd->cpumask) {
-		queue = per_cpu_ptr(pd->queue, cpu);
-		flush_work(&queue->swork);
+	for_each_cpu(cpu, pd->cpumask.cbcpu) {
+		squeue = per_cpu_ptr(pd->squeue, cpu);
+		flush_work(&squeue->work);
 	}
 
 	BUG_ON(atomic_read(&pd->refcnt) != 0);
 }
 
+static void __padata_start(struct padata_instance *pinst)
+{
+	pinst->flags |= PADATA_INIT;
+}
+
+static void __padata_stop(struct padata_instance *pinst)
+{
+	if (!(pinst->flags & PADATA_INIT))
+		return;
+
+	pinst->flags &= ~PADATA_INIT;
+
+	synchronize_rcu();
+
+	get_online_cpus();
+	padata_flush_queues(pinst->pd);
+	put_online_cpus();
+}
+
 /* Replace the internal control stucture with a new one. */
 static void padata_replace(struct padata_instance *pinst,
 			   struct parallel_data *pd_new)
 {
 	struct parallel_data *pd_old = pinst->pd;
+	int notification_mask = 0;
 
 	pinst->flags |= PADATA_RESET;
 
@@ -497,41 +528,162 @@
 
 	synchronize_rcu();
 
+	if (!cpumask_equal(pd_old->cpumask.pcpu, pd_new->cpumask.pcpu))
+		notification_mask |= PADATA_CPU_PARALLEL;
+	if (!cpumask_equal(pd_old->cpumask.cbcpu, pd_new->cpumask.cbcpu))
+		notification_mask |= PADATA_CPU_SERIAL;
+
 	padata_flush_queues(pd_old);
 	padata_free_pd(pd_old);
 
+	if (notification_mask)
+		blocking_notifier_call_chain(&pinst->cpumask_change_notifier,
+					     notification_mask,
+					     &pd_new->cpumask);
+
 	pinst->flags &= ~PADATA_RESET;
 }
 
 /**
- * padata_set_cpumask - set the cpumask that padata should use
+ * padata_register_cpumask_notifier - Registers a notifier that will be called
+ *                             if either pcpu or cbcpu or both cpumasks change.
  *
- * @pinst: padata instance
- * @cpumask: the cpumask to use
+ * @pinst: A poineter to padata instance
+ * @nblock: A pointer to notifier block.
  */
-int padata_set_cpumask(struct padata_instance *pinst,
-			cpumask_var_t cpumask)
+int padata_register_cpumask_notifier(struct padata_instance *pinst,
+				     struct notifier_block *nblock)
 {
-	struct parallel_data *pd;
-	int err = 0;
+	return blocking_notifier_chain_register(&pinst->cpumask_change_notifier,
+						nblock);
+}
+EXPORT_SYMBOL(padata_register_cpumask_notifier);
 
-	mutex_lock(&pinst->lock);
+/**
+ * padata_unregister_cpumask_notifier - Unregisters cpumask notifier
+ *        registered earlier  using padata_register_cpumask_notifier
+ *
+ * @pinst: A pointer to data instance.
+ * @nlock: A pointer to notifier block.
+ */
+int padata_unregister_cpumask_notifier(struct padata_instance *pinst,
+				       struct notifier_block *nblock)
+{
+	return blocking_notifier_chain_unregister(
+		&pinst->cpumask_change_notifier,
+		nblock);
+}
+EXPORT_SYMBOL(padata_unregister_cpumask_notifier);
 
-	get_online_cpus();
 
-	pd = padata_alloc_pd(pinst, cpumask);
-	if (!pd) {
-		err = -ENOMEM;
-		goto out;
+/* If cpumask contains no active cpu, we mark the instance as invalid. */
+static bool padata_validate_cpumask(struct padata_instance *pinst,
+				    const struct cpumask *cpumask)
+{
+	if (!cpumask_intersects(cpumask, cpu_active_mask)) {
+		pinst->flags |= PADATA_INVALID;
+		return false;
 	}
 
-	cpumask_copy(pinst->cpumask, cpumask);
+	pinst->flags &= ~PADATA_INVALID;
+	return true;
+}
+
+static int __padata_set_cpumasks(struct padata_instance *pinst,
+				 cpumask_var_t pcpumask,
+				 cpumask_var_t cbcpumask)
+{
+	int valid;
+	struct parallel_data *pd;
+
+	valid = padata_validate_cpumask(pinst, pcpumask);
+	if (!valid) {
+		__padata_stop(pinst);
+		goto out_replace;
+	}
+
+	valid = padata_validate_cpumask(pinst, cbcpumask);
+	if (!valid)
+		__padata_stop(pinst);
+
+out_replace:
+	pd = padata_alloc_pd(pinst, pcpumask, cbcpumask);
+	if (!pd)
+		return -ENOMEM;
+
+	cpumask_copy(pinst->cpumask.pcpu, pcpumask);
+	cpumask_copy(pinst->cpumask.cbcpu, cbcpumask);
 
 	padata_replace(pinst, pd);
 
+	if (valid)
+		__padata_start(pinst);
+
+	return 0;
+}
+
+/**
+ * padata_set_cpumasks - Set both parallel and serial cpumasks. The first
+ *                       one is used by parallel workers and the second one
+ *                       by the wokers doing serialization.
+ *
+ * @pinst: padata instance
+ * @pcpumask: the cpumask to use for parallel workers
+ * @cbcpumask: the cpumsak to use for serial workers
+ */
+int padata_set_cpumasks(struct padata_instance *pinst, cpumask_var_t pcpumask,
+			cpumask_var_t cbcpumask)
+{
+	int err;
+
+	mutex_lock(&pinst->lock);
+	get_online_cpus();
+
+	err = __padata_set_cpumasks(pinst, pcpumask, cbcpumask);
+
+	put_online_cpus();
+	mutex_unlock(&pinst->lock);
+
+	return err;
+
+}
+EXPORT_SYMBOL(padata_set_cpumasks);
+
+/**
+ * padata_set_cpumask: Sets specified by @cpumask_type cpumask to the value
+ *                     equivalent to @cpumask.
+ *
+ * @pinst: padata instance
+ * @cpumask_type: PADATA_CPU_SERIAL or PADATA_CPU_PARALLEL corresponding
+ *                to parallel and serial cpumasks respectively.
+ * @cpumask: the cpumask to use
+ */
+int padata_set_cpumask(struct padata_instance *pinst, int cpumask_type,
+		       cpumask_var_t cpumask)
+{
+	struct cpumask *serial_mask, *parallel_mask;
+	int err = -EINVAL;
+
+	mutex_lock(&pinst->lock);
+	get_online_cpus();
+
+	switch (cpumask_type) {
+	case PADATA_CPU_PARALLEL:
+		serial_mask = pinst->cpumask.cbcpu;
+		parallel_mask = cpumask;
+		break;
+	case PADATA_CPU_SERIAL:
+		parallel_mask = pinst->cpumask.pcpu;
+		serial_mask = cpumask;
+		break;
+	default:
+		 goto out;
+	}
+
+	err =  __padata_set_cpumasks(pinst, parallel_mask, serial_mask);
+
 out:
 	put_online_cpus();
-
 	mutex_unlock(&pinst->lock);
 
 	return err;
@@ -543,30 +695,48 @@
 	struct parallel_data *pd;
 
 	if (cpumask_test_cpu(cpu, cpu_active_mask)) {
-		pd = padata_alloc_pd(pinst, pinst->cpumask);
+		pd = padata_alloc_pd(pinst, pinst->cpumask.pcpu,
+				     pinst->cpumask.cbcpu);
 		if (!pd)
 			return -ENOMEM;
 
 		padata_replace(pinst, pd);
+
+		if (padata_validate_cpumask(pinst, pinst->cpumask.pcpu) &&
+		    padata_validate_cpumask(pinst, pinst->cpumask.cbcpu))
+			__padata_start(pinst);
 	}
 
 	return 0;
 }
 
-/**
- * padata_add_cpu - add a cpu to the padata cpumask
+ /**
+ * padata_add_cpu - add a cpu to one or both(parallel and serial)
+ *                  padata cpumasks.
  *
  * @pinst: padata instance
  * @cpu: cpu to add
+ * @mask: bitmask of flags specifying to which cpumask @cpu shuld be added.
+ *        The @mask may be any combination of the following flags:
+ *          PADATA_CPU_SERIAL   - serial cpumask
+ *          PADATA_CPU_PARALLEL - parallel cpumask
  */
-int padata_add_cpu(struct padata_instance *pinst, int cpu)
+
+int padata_add_cpu(struct padata_instance *pinst, int cpu, int mask)
 {
 	int err;
 
+	if (!(mask & (PADATA_CPU_SERIAL | PADATA_CPU_PARALLEL)))
+		return -EINVAL;
+
 	mutex_lock(&pinst->lock);
 
 	get_online_cpus();
-	cpumask_set_cpu(cpu, pinst->cpumask);
+	if (mask & PADATA_CPU_SERIAL)
+		cpumask_set_cpu(cpu, pinst->cpumask.cbcpu);
+	if (mask & PADATA_CPU_PARALLEL)
+		cpumask_set_cpu(cpu, pinst->cpumask.pcpu);
+
 	err = __padata_add_cpu(pinst, cpu);
 	put_online_cpus();
 
@@ -578,10 +748,16 @@
 
 static int __padata_remove_cpu(struct padata_instance *pinst, int cpu)
 {
-	struct parallel_data *pd;
+	struct parallel_data *pd = NULL;
 
 	if (cpumask_test_cpu(cpu, cpu_online_mask)) {
-		pd = padata_alloc_pd(pinst, pinst->cpumask);
+
+		if (!padata_validate_cpumask(pinst, pinst->cpumask.pcpu) ||
+		    !padata_validate_cpumask(pinst, pinst->cpumask.cbcpu))
+			__padata_stop(pinst);
+
+		pd = padata_alloc_pd(pinst, pinst->cpumask.pcpu,
+				     pinst->cpumask.cbcpu);
 		if (!pd)
 			return -ENOMEM;
 
@@ -591,20 +767,32 @@
 	return 0;
 }
 
-/**
- * padata_remove_cpu - remove a cpu from the padata cpumask
+ /**
+ * padata_remove_cpu - remove a cpu from the one or both(serial and paralell)
+ *                     padata cpumasks.
  *
  * @pinst: padata instance
  * @cpu: cpu to remove
+ * @mask: bitmask specifying from which cpumask @cpu should be removed
+ *        The @mask may be any combination of the following flags:
+ *          PADATA_CPU_SERIAL   - serial cpumask
+ *          PADATA_CPU_PARALLEL - parallel cpumask
  */
-int padata_remove_cpu(struct padata_instance *pinst, int cpu)
+int padata_remove_cpu(struct padata_instance *pinst, int cpu, int mask)
 {
 	int err;
 
+	if (!(mask & (PADATA_CPU_SERIAL | PADATA_CPU_PARALLEL)))
+		return -EINVAL;
+
 	mutex_lock(&pinst->lock);
 
 	get_online_cpus();
-	cpumask_clear_cpu(cpu, pinst->cpumask);
+	if (mask & PADATA_CPU_SERIAL)
+		cpumask_clear_cpu(cpu, pinst->cpumask.cbcpu);
+	if (mask & PADATA_CPU_PARALLEL)
+		cpumask_clear_cpu(cpu, pinst->cpumask.pcpu);
+
 	err = __padata_remove_cpu(pinst, cpu);
 	put_online_cpus();
 
@@ -619,11 +807,20 @@
  *
  * @pinst: padata instance to start
  */
-void padata_start(struct padata_instance *pinst)
+int padata_start(struct padata_instance *pinst)
 {
+	int err = 0;
+
 	mutex_lock(&pinst->lock);
-	pinst->flags |= PADATA_INIT;
+
+	if (pinst->flags & PADATA_INVALID)
+		err =-EINVAL;
+
+	 __padata_start(pinst);
+
 	mutex_unlock(&pinst->lock);
+
+	return err;
 }
 EXPORT_SYMBOL(padata_start);
 
@@ -635,12 +832,20 @@
 void padata_stop(struct padata_instance *pinst)
 {
 	mutex_lock(&pinst->lock);
-	pinst->flags &= ~PADATA_INIT;
+	__padata_stop(pinst);
 	mutex_unlock(&pinst->lock);
 }
 EXPORT_SYMBOL(padata_stop);
 
 #ifdef CONFIG_HOTPLUG_CPU
+
+static inline int pinst_has_cpu(struct padata_instance *pinst, int cpu)
+{
+	return cpumask_test_cpu(cpu, pinst->cpumask.pcpu) ||
+		cpumask_test_cpu(cpu, pinst->cpumask.cbcpu);
+}
+
+
 static int padata_cpu_callback(struct notifier_block *nfb,
 			       unsigned long action, void *hcpu)
 {
@@ -653,7 +858,7 @@
 	switch (action) {
 	case CPU_ONLINE:
 	case CPU_ONLINE_FROZEN:
-		if (!cpumask_test_cpu(cpu, pinst->cpumask))
+		if (!pinst_has_cpu(pinst, cpu))
 			break;
 		mutex_lock(&pinst->lock);
 		err = __padata_add_cpu(pinst, cpu);
@@ -664,7 +869,7 @@
 
 	case CPU_DOWN_PREPARE:
 	case CPU_DOWN_PREPARE_FROZEN:
-		if (!cpumask_test_cpu(cpu, pinst->cpumask))
+		if (!pinst_has_cpu(pinst, cpu))
 			break;
 		mutex_lock(&pinst->lock);
 		err = __padata_remove_cpu(pinst, cpu);
@@ -675,7 +880,7 @@
 
 	case CPU_UP_CANCELED:
 	case CPU_UP_CANCELED_FROZEN:
-		if (!cpumask_test_cpu(cpu, pinst->cpumask))
+		if (!pinst_has_cpu(pinst, cpu))
 			break;
 		mutex_lock(&pinst->lock);
 		__padata_remove_cpu(pinst, cpu);
@@ -683,7 +888,7 @@
 
 	case CPU_DOWN_FAILED:
 	case CPU_DOWN_FAILED_FROZEN:
-		if (!cpumask_test_cpu(cpu, pinst->cpumask))
+		if (!pinst_has_cpu(pinst, cpu))
 			break;
 		mutex_lock(&pinst->lock);
 		__padata_add_cpu(pinst, cpu);
@@ -694,36 +899,202 @@
 }
 #endif
 
-/**
- * padata_alloc - allocate and initialize a padata instance
- *
- * @cpumask: cpumask that padata uses for parallelization
- * @wq: workqueue to use for the allocated padata instance
+static void __padata_free(struct padata_instance *pinst)
+{
+#ifdef CONFIG_HOTPLUG_CPU
+	unregister_hotcpu_notifier(&pinst->cpu_notifier);
+#endif
+
+	padata_stop(pinst);
+	padata_free_pd(pinst->pd);
+	free_cpumask_var(pinst->cpumask.pcpu);
+	free_cpumask_var(pinst->cpumask.cbcpu);
+	kfree(pinst);
+}
+
+#define kobj2pinst(_kobj)					\
+	container_of(_kobj, struct padata_instance, kobj)
+#define attr2pentry(_attr)					\
+	container_of(_attr, struct padata_sysfs_entry, attr)
+
+static void padata_sysfs_release(struct kobject *kobj)
+{
+	struct padata_instance *pinst = kobj2pinst(kobj);
+	__padata_free(pinst);
+}
+
+struct padata_sysfs_entry {
+	struct attribute attr;
+	ssize_t (*show)(struct padata_instance *, struct attribute *, char *);
+	ssize_t (*store)(struct padata_instance *, struct attribute *,
+			 const char *, size_t);
+};
+
+static ssize_t show_cpumask(struct padata_instance *pinst,
+			    struct attribute *attr,  char *buf)
+{
+	struct cpumask *cpumask;
+	ssize_t len;
+
+	mutex_lock(&pinst->lock);
+	if (!strcmp(attr->name, "serial_cpumask"))
+		cpumask = pinst->cpumask.cbcpu;
+	else
+		cpumask = pinst->cpumask.pcpu;
+
+	len = bitmap_scnprintf(buf, PAGE_SIZE, cpumask_bits(cpumask),
+			       nr_cpu_ids);
+	if (PAGE_SIZE - len < 2)
+		len = -EINVAL;
+	else
+		len += sprintf(buf + len, "\n");
+
+	mutex_unlock(&pinst->lock);
+	return len;
+}
+
+static ssize_t store_cpumask(struct padata_instance *pinst,
+			     struct attribute *attr,
+			     const char *buf, size_t count)
+{
+	cpumask_var_t new_cpumask;
+	ssize_t ret;
+	int mask_type;
+
+	if (!alloc_cpumask_var(&new_cpumask, GFP_KERNEL))
+		return -ENOMEM;
+
+	ret = bitmap_parse(buf, count, cpumask_bits(new_cpumask),
+			   nr_cpumask_bits);
+	if (ret < 0)
+		goto out;
+
+	mask_type = !strcmp(attr->name, "serial_cpumask") ?
+		PADATA_CPU_SERIAL : PADATA_CPU_PARALLEL;
+	ret = padata_set_cpumask(pinst, mask_type, new_cpumask);
+	if (!ret)
+		ret = count;
+
+out:
+	free_cpumask_var(new_cpumask);
+	return ret;
+}
+
+#define PADATA_ATTR_RW(_name, _show_name, _store_name)		\
+	static struct padata_sysfs_entry _name##_attr =		\
+		__ATTR(_name, 0644, _show_name, _store_name)
+#define PADATA_ATTR_RO(_name, _show_name)		\
+	static struct padata_sysfs_entry _name##_attr = \
+		__ATTR(_name, 0400, _show_name, NULL)
+
+PADATA_ATTR_RW(serial_cpumask, show_cpumask, store_cpumask);
+PADATA_ATTR_RW(parallel_cpumask, show_cpumask, store_cpumask);
+
+/*
+ * Padata sysfs provides the following objects:
+ * serial_cpumask   [RW] - cpumask for serial workers
+ * parallel_cpumask [RW] - cpumask for parallel workers
  */
-struct padata_instance *padata_alloc(const struct cpumask *cpumask,
-				     struct workqueue_struct *wq)
+static struct attribute *padata_default_attrs[] = {
+	&serial_cpumask_attr.attr,
+	&parallel_cpumask_attr.attr,
+	NULL,
+};
+
+static ssize_t padata_sysfs_show(struct kobject *kobj,
+				 struct attribute *attr, char *buf)
 {
 	struct padata_instance *pinst;
-	struct parallel_data *pd;
+	struct padata_sysfs_entry *pentry;
+	ssize_t ret = -EIO;
+
+	pinst = kobj2pinst(kobj);
+	pentry = attr2pentry(attr);
+	if (pentry->show)
+		ret = pentry->show(pinst, attr, buf);
+
+	return ret;
+}
+
+static ssize_t padata_sysfs_store(struct kobject *kobj, struct attribute *attr,
+				  const char *buf, size_t count)
+{
+	struct padata_instance *pinst;
+	struct padata_sysfs_entry *pentry;
+	ssize_t ret = -EIO;
+
+	pinst = kobj2pinst(kobj);
+	pentry = attr2pentry(attr);
+	if (pentry->show)
+		ret = pentry->store(pinst, attr, buf, count);
+
+	return ret;
+}
+
+static const struct sysfs_ops padata_sysfs_ops = {
+	.show = padata_sysfs_show,
+	.store = padata_sysfs_store,
+};
+
+static struct kobj_type padata_attr_type = {
+	.sysfs_ops = &padata_sysfs_ops,
+	.default_attrs = padata_default_attrs,
+	.release = padata_sysfs_release,
+};
+
+/**
+ * padata_alloc_possible - Allocate and initialize padata instance.
+ *                         Use the cpu_possible_mask for serial and
+ *                         parallel workers.
+ *
+ * @wq: workqueue to use for the allocated padata instance
+ */
+struct padata_instance *padata_alloc_possible(struct workqueue_struct *wq)
+{
+	return padata_alloc(wq, cpu_possible_mask, cpu_possible_mask);
+}
+EXPORT_SYMBOL(padata_alloc_possible);
+
+/**
+ * padata_alloc - allocate and initialize a padata instance and specify
+ *                cpumasks for serial and parallel workers.
+ *
+ * @wq: workqueue to use for the allocated padata instance
+ * @pcpumask: cpumask that will be used for padata parallelization
+ * @cbcpumask: cpumask that will be used for padata serialization
+ */
+struct padata_instance *padata_alloc(struct workqueue_struct *wq,
+				     const struct cpumask *pcpumask,
+				     const struct cpumask *cbcpumask)
+{
+	struct padata_instance *pinst;
+	struct parallel_data *pd = NULL;
 
 	pinst = kzalloc(sizeof(struct padata_instance), GFP_KERNEL);
 	if (!pinst)
 		goto err;
 
 	get_online_cpus();
-
-	pd = padata_alloc_pd(pinst, cpumask);
-	if (!pd)
+	if (!alloc_cpumask_var(&pinst->cpumask.pcpu, GFP_KERNEL))
 		goto err_free_inst;
+	if (!alloc_cpumask_var(&pinst->cpumask.cbcpu, GFP_KERNEL)) {
+		free_cpumask_var(pinst->cpumask.pcpu);
+		goto err_free_inst;
+	}
+	if (!padata_validate_cpumask(pinst, pcpumask) ||
+	    !padata_validate_cpumask(pinst, cbcpumask))
+		goto err_free_masks;
 
-	if (!alloc_cpumask_var(&pinst->cpumask, GFP_KERNEL))
-		goto err_free_pd;
+	pd = padata_alloc_pd(pinst, pcpumask, cbcpumask);
+	if (!pd)
+		goto err_free_masks;
 
 	rcu_assign_pointer(pinst->pd, pd);
 
 	pinst->wq = wq;
 
-	cpumask_copy(pinst->cpumask, cpumask);
+	cpumask_copy(pinst->cpumask.pcpu, pcpumask);
+	cpumask_copy(pinst->cpumask.cbcpu, cbcpumask);
 
 	pinst->flags = 0;
 
@@ -735,12 +1106,15 @@
 
 	put_online_cpus();
 
+	BLOCKING_INIT_NOTIFIER_HEAD(&pinst->cpumask_change_notifier);
+	kobject_init(&pinst->kobj, &padata_attr_type);
 	mutex_init(&pinst->lock);
 
 	return pinst;
 
-err_free_pd:
-	padata_free_pd(pd);
+err_free_masks:
+	free_cpumask_var(pinst->cpumask.pcpu);
+	free_cpumask_var(pinst->cpumask.cbcpu);
 err_free_inst:
 	kfree(pinst);
 	put_online_cpus();
@@ -756,19 +1130,6 @@
  */
 void padata_free(struct padata_instance *pinst)
 {
-	padata_stop(pinst);
-
-	synchronize_rcu();
-
-#ifdef CONFIG_HOTPLUG_CPU
-	unregister_hotcpu_notifier(&pinst->cpu_notifier);
-#endif
-	get_online_cpus();
-	padata_flush_queues(pinst->pd);
-	put_online_cpus();
-
-	padata_free_pd(pinst->pd);
-	free_cpumask_var(pinst->cpumask);
-	kfree(pinst);
+	kobject_put(&pinst->kobj);
 }
 EXPORT_SYMBOL(padata_free);