This patchset adds the 'pcrypt' parallel crypto template. With this template it
is possible to process the crypto requests of a transform in parallel without
getting request reorder. This is in particular interesting for IPsec.
The parallel crypto template is based on a generic parallelization/serialization
method. This method uses the remote softirq invocation infrastructure for
parallelization and serialization. With this method data objects can be
processed in parallel, starting at some given point.
After doing some expensive operations in parallel, it is possible to serialize
again. The parallelized data objects return after serialization in the order as
they were before the parallelization. In the case of IPsec, this makes it
possible to run the expensive parts in parallel without getting packet
reordering.
Changes from v4:
- Use the dynamic percpu allocator
- Drop of the obsolete eseqiv changes (eseqiv is the default IV generator
for blockcipher algorithms on smp machines now).
Changes from v3:
- The generic aead wrapper is dropped.
- tcrypt is extended to test algorithms by name. So it is possible to
instantiate pcrypt by doing e.g.:
modprobe tcrypt alg="pcrypt(authenc(hmac(sha1),cbc(aes)))" type=3
Changes from v2:
- The xfrm netlink configuration code is dropped,
this will be an extra patchset.
- Add generic aead wrapper interface to be able to wrap an aead algorithm
with an arbitrary crypto template.
- Convert pcrypt to use the generic aead wrapper.
- Add support for aead algorithms to eseqiv.
- Add support for the pcrypt aead wrapper to authenc. It's now possible to
choose for pcrypt as the default authenc wrapper with a module parameter.
- Patchset applies to linux-2.6 git current.
Changes from v1:
- cpu_chainiv is dropped, pcrypt uses eseqiv as it's IV generator now.
- Add a xfrm netlink message to be able to choose for pcrypt from userspace.
- Use pcrypt just if it is selected from userspace.
Steffen
This patch introduces an interface to process data objects
in parallel. On request it is possible to serialize again.
The parallelized objects return after serialization in the
same order as they were before the parallelization.
Signed-off-by: Steffen Klassert <[email protected]>
---
include/linux/interrupt.h | 3 +-
include/linux/padata.h | 121 +++++++++++
kernel/Makefile | 2 +-
kernel/padata.c | 519 +++++++++++++++++++++++++++++++++++++++++++++
kernel/softirq.c | 2 +-
5 files changed, 644 insertions(+), 3 deletions(-)
create mode 100644 include/linux/padata.h
create mode 100644 kernel/padata.c
diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index 35e7df1..b648cca 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -347,7 +347,8 @@ enum
TASKLET_SOFTIRQ,
SCHED_SOFTIRQ,
HRTIMER_SOFTIRQ,
- RCU_SOFTIRQ, /* Preferable RCU should always be the last softirq */
+ PADATA_SOFTIRQ,
+ RCU_SOFTIRQ, /* Preferable RCU should always be the last softirq */
NR_SOFTIRQS
};
diff --git a/include/linux/padata.h b/include/linux/padata.h
new file mode 100644
index 0000000..a81161d
--- /dev/null
+++ b/include/linux/padata.h
@@ -0,0 +1,121 @@
+/*
+ * padata.h - header for the padata parallelization interface
+ *
+ * Copyright (C) 2008, 2009 secunet Security Networks AG
+ * Copyright (C) 2008, 2009 Steffen Klassert <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#ifndef PADATA_H
+#define PADATA_H
+
+#include <linux/interrupt.h>
+#include <linux/smp.h>
+#include <linux/list.h>
+
+enum
+{
+ NO_PADATA=0,
+ AEAD_ENC_PADATA,
+ AEAD_DEC_PADATA,
+ NR_PADATA
+};
+
+struct padata_priv {
+ struct list_head list;
+ struct call_single_data csd;
+ int cb_cpu;
+ int seq_nr;
+ unsigned int nr;
+ int info;
+ void (*parallel)(struct padata_priv *padata);
+ void (*serial)(struct padata_priv *padata);
+};
+
+struct padata_queue {
+ struct list_head list;
+ atomic_t num_obj;
+ int cpu_index;
+ spinlock_t lock;
+};
+
+struct parallel_data {
+ struct work_struct work;
+ struct padata_queue *queue;
+ atomic_t seq_nr;
+ atomic_t queued_objects;
+ cpumask_var_t cpumask;
+ cpumask_var_t new_cpumask;
+ u8 flags;
+#define PADATA_INIT 1
+#define PADATA_FLUSH_HARD 2
+#define PADATA_RESET_IN_PROGRESS 4
+ spinlock_t lock;
+};
+
+#ifdef CONFIG_USE_GENERIC_SMP_HELPERS
+extern void __init padata_init(unsigned int nr, const struct cpumask *cpumask);
+extern void padata_dont_wait(unsigned int nr, struct padata_priv *padata);
+extern int padata_do_parallel(unsigned int softirq_nr, unsigned int nr,
+ struct padata_priv *padata, int cb_cpu);
+extern int padata_do_serial(unsigned int nr, struct padata_priv *padata);
+extern int padata_cpumask_weight(unsigned int nr);
+extern int padata_index_to_cpu(unsigned int nr, int cpu_index);
+extern void padata_set_cpumask(unsigned int nr, cpumask_var_t cpumask);
+extern void padata_add_cpu(unsigned int nr, int cpu);
+extern void padata_remove_cpu(unsigned int nr, int cpu);
+extern void padata_start(unsigned int nr);
+extern void padata_stop(unsigned int nr);
+#else
+static inline void padata_init(unsigned int nr, const struct cpumask *cpumask)
+{
+}
+static inline void padata_dont_wait(unsigned int nr, struct padata_priv *padata)
+{
+}
+static inline int padata_do_parallel(unsigned int softirq_nr, unsigned int nr,
+ struct padata_priv *padata, int cb_cpu)
+{
+ return 0;
+}
+static inline int padata_do_serial(unsigned int nr, struct padata_priv *padata)
+{
+ return 0;
+}
+static inline int padata_cpumask_weight(unsigned int nr)
+{
+ return 0;
+}
+static inline int padata_index_to_cpu(unsigned int nr, int cpu_index)
+{
+ return -ENOSYS;
+}
+static inline void padata_set_cpumask(unsigned int nr, cpumask_var_t cpumask)
+{
+}
+static inline padata_add_cpu(unsigned int nr, int cpu)
+{
+}
+static inline padata_remove_cpu(unsigned int nr, int cpu)
+{
+}
+static inline padata_start(unsigned int nr)
+{
+}
+static inline padata_stop(unsigned int nr)
+{
+}
+#endif
+#endif
diff --git a/kernel/Makefile b/kernel/Makefile
index 2093a69..6ff7f1a 100644
--- a/kernel/Makefile
+++ b/kernel/Makefile
@@ -41,7 +41,7 @@ obj-$(CONFIG_RT_MUTEXES) += rtmutex.o
obj-$(CONFIG_DEBUG_RT_MUTEXES) += rtmutex-debug.o
obj-$(CONFIG_RT_MUTEX_TESTER) += rtmutex-tester.o
obj-$(CONFIG_GENERIC_ISA_DMA) += dma.o
-obj-$(CONFIG_USE_GENERIC_SMP_HELPERS) += smp.o
+obj-$(CONFIG_USE_GENERIC_SMP_HELPERS) += smp.o padata.o
ifneq ($(CONFIG_SMP),y)
obj-y += up.o
endif
diff --git a/kernel/padata.c b/kernel/padata.c
new file mode 100644
index 0000000..3523f25
--- /dev/null
+++ b/kernel/padata.c
@@ -0,0 +1,519 @@
+/*
+ * padata.c - generic interface to process data streams in parallel
+ *
+ * Copyright (C) 2008, 2009 secunet Security Networks AG
+ * Copyright (C) 2008, 2009 Steffen Klassert <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <linux/module.h>
+#include <linux/cpumask.h>
+#include <linux/err.h>
+#include <linux/padata.h>
+
+#define MAX_SEQ_NR INT_MAX - NR_CPUS
+
+static struct parallel_data padata_vec[NR_PADATA];
+static struct padata_priv *padata_get_next(struct parallel_data *par_data);
+
+static void padata_flush_hard(struct parallel_data *par_data)
+{
+ int cpu;
+ struct padata_priv *padata;
+ struct padata_queue *queue;
+
+ for_each_cpu(cpu, par_data->cpumask) {
+ queue = per_cpu_ptr(par_data->queue, cpu);
+
+ while(!list_empty(&queue->list)) {
+ padata = list_entry(queue->list.next,
+ struct padata_priv, list);
+
+ spin_lock(&queue->lock);
+ list_del_init(&padata->list);
+ spin_unlock(&queue->lock);
+
+ atomic_dec(&par_data->queued_objects);
+ padata->serial(padata);
+ }
+ }
+}
+
+static void padata_flush_order(struct parallel_data *par_data)
+{
+ struct padata_priv *padata;
+
+ while (1) {
+ padata = padata_get_next(par_data);
+
+ if (padata && !IS_ERR(padata))
+ padata->serial(padata);
+ else
+ break;
+ }
+
+ padata_flush_hard(par_data);
+}
+
+static void padata_reset_work(struct work_struct *work)
+{
+ int cpu, cpu_index;
+ struct padata_queue *queue;
+ struct parallel_data *par_data;
+
+ par_data = container_of(work, struct parallel_data, work);
+
+ if (par_data->flags & (PADATA_INIT|PADATA_RESET_IN_PROGRESS))
+ return;
+
+ spin_lock_bh(&par_data->lock);
+ par_data->flags |= PADATA_RESET_IN_PROGRESS;
+
+ if (!(par_data->flags & PADATA_FLUSH_HARD))
+ padata_flush_order(par_data);
+ else
+ padata_flush_hard(par_data);
+
+ cpu_index = 0;
+
+ cpumask_copy(par_data->cpumask, par_data->new_cpumask);
+
+ for_each_cpu(cpu, par_data->cpumask) {
+ queue = per_cpu_ptr(par_data->queue, cpu);
+
+ atomic_set(&queue->num_obj, 0);
+ queue->cpu_index = cpu_index;
+ cpu_index++;
+ }
+ spin_unlock_bh(&par_data->lock);
+
+ atomic_set(&par_data->seq_nr, -1);
+ par_data->flags &= ~PADATA_RESET_IN_PROGRESS;
+ par_data->flags |= PADATA_INIT;
+}
+
+static struct padata_priv *padata_get_next(struct parallel_data *par_data)
+{
+ int cpu, num_cpus, empty;
+ int seq_nr, calc_seq_nr, next_nr;
+ struct padata_queue *queue, *next_queue;
+ struct padata_priv *padata;
+
+ empty = 0;
+ next_nr = -1;
+ next_queue = NULL;
+
+ num_cpus = cpumask_weight(par_data->cpumask);
+
+ for_each_cpu(cpu, par_data->cpumask) {
+ queue = per_cpu_ptr(par_data->queue, cpu);
+
+ /*
+ * Calculate the seq_nr of the object that should be
+ * next in this queue.
+ */
+ calc_seq_nr = (atomic_read(&queue->num_obj) * num_cpus)
+ + queue->cpu_index;
+
+ if (!list_empty(&queue->list)) {
+ padata = list_entry(queue->list.next,
+ struct padata_priv, list);
+
+ seq_nr = padata->seq_nr;
+
+ if (unlikely(calc_seq_nr != seq_nr)) {
+ par_data->flags &= ~PADATA_INIT;
+ par_data->flags |= PADATA_FLUSH_HARD;
+ padata = NULL;
+ goto out;
+ }
+ } else {
+ seq_nr = calc_seq_nr;
+ empty++;
+ }
+
+ if (next_nr < 0 || seq_nr < next_nr) {
+ next_nr = seq_nr;
+ next_queue = queue;
+ }
+ }
+
+ padata = NULL;
+
+ if (empty == num_cpus)
+ goto out;
+
+ if (!list_empty(&next_queue->list)) {
+ padata = list_entry(next_queue->list.next,
+ struct padata_priv, list);
+
+ spin_lock(&next_queue->lock);
+ list_del_init(&padata->list);
+ spin_unlock(&next_queue->lock);
+
+ atomic_dec(&par_data->queued_objects);
+ atomic_inc(&next_queue->num_obj);
+
+ goto out;
+ }
+
+ if (next_nr % num_cpus == next_queue->cpu_index) {
+ padata = ERR_PTR(-ENODATA);
+ goto out;
+ }
+
+ padata = ERR_PTR(-EINPROGRESS);
+out:
+ return padata;
+}
+
+static void padata_action(struct softirq_action *h)
+{
+ struct list_head *cpu_list, local_list;
+
+ cpu_list = &__get_cpu_var(softirq_work_list[PADATA_SOFTIRQ]);
+
+ local_irq_disable();
+ list_replace_init(cpu_list, &local_list);
+ local_irq_enable();
+
+ while (!list_empty(&local_list)) {
+ struct padata_priv *padata;
+
+ padata = list_entry(local_list.next,
+ struct padata_priv, csd.list);
+
+ list_del_init(&padata->csd.list);
+
+ padata->serial(padata);
+ }
+}
+
+static int padata_cpu_hash(unsigned int nr, struct padata_priv *padata)
+{
+ int this_cpu, cpu_index;
+
+ this_cpu = smp_processor_id();
+
+ if (padata->nr != 0)
+ return this_cpu;
+
+ if (!(padata_vec[nr].flags & PADATA_INIT))
+ return this_cpu;
+
+ padata->seq_nr = atomic_inc_return(&padata_vec[nr].seq_nr);
+
+ if (padata->seq_nr > MAX_SEQ_NR) {
+ padata_vec[nr].flags &= ~PADATA_INIT;
+ padata->seq_nr = 0;
+ schedule_work(&padata_vec[nr].work);
+ return this_cpu;
+ }
+
+ padata->nr = nr;
+
+ /*
+ * Hash the sequence numbers to the cpus by taking
+ * seq_nr mod. number of cpus in use.
+ */
+ cpu_index = padata->seq_nr % cpumask_weight(padata_vec[nr].cpumask);
+
+ return padata_index_to_cpu(nr, cpu_index);
+}
+
+/*
+ * padata_dont_wait - must be called if an object that runs in parallel will
+ * not be serialized with padata_do_serial
+ *
+ * @nr: number of the padata instance
+ * @padata: object that will not be seen by padata_do_serial
+ */
+void padata_dont_wait(unsigned int nr, struct padata_priv *padata)
+{
+ struct padata_queue *queue;
+
+ if (!(padata_vec[nr].flags & PADATA_INIT))
+ return;
+
+ if (padata->nr == 0 || padata->nr != nr)
+ return;
+
+ queue = per_cpu_ptr(padata_vec[nr].queue, smp_processor_id());
+ atomic_inc(&queue->num_obj);
+
+ padata->nr = 0;
+ padata->seq_nr = 0;
+}
+EXPORT_SYMBOL(padata_dont_wait);
+
+/*
+ * padata_do_parallel - padata parallelization function
+ *
+ * @softirq_nr: number of the softirq that will do the parallelization
+ * @nr: number of the padata instance
+ * @padata: object to be parallelized
+ * @cb_cpu: cpu number on which the serialization callback function will run
+ */
+int padata_do_parallel(unsigned int softirq_nr, unsigned int nr,
+ struct padata_priv *padata, int cb_cpu)
+{
+ int target_cpu;
+
+ padata->cb_cpu = cb_cpu;
+
+ local_bh_disable();
+ target_cpu = padata_cpu_hash(nr, padata);
+ local_bh_enable();
+
+ send_remote_softirq(&padata->csd, target_cpu, softirq_nr);
+
+ return 1;
+}
+EXPORT_SYMBOL(padata_do_parallel);
+
+/*
+ * padata_do_serial - padata serialization function
+ *
+ * @nr: number of the padata instance
+ * @padata: object to be serialized
+ *
+ * returns 1 if the serialization callback function will be called
+ * from padata, 0 else
+ */
+int padata_do_serial(unsigned int nr, struct padata_priv *padata)
+{
+ int cpu;
+ struct padata_queue *reorder_queue;
+
+ if (!(padata_vec[nr].flags & PADATA_INIT))
+ return 0;
+
+ if (padata->nr != nr || padata->nr == 0) {
+ padata->serial(padata);
+ return 1;
+ }
+
+ cpu = smp_processor_id();
+
+ reorder_queue = per_cpu_ptr(padata_vec[nr].queue, cpu);
+
+ spin_lock(&reorder_queue->lock);
+ list_add_tail(&padata->list, &reorder_queue->list);
+ spin_unlock(&reorder_queue->lock);
+
+ atomic_inc(&padata_vec[nr].queued_objects);
+
+try_again:
+ if (!spin_trylock(&padata_vec[nr].lock))
+ goto out;
+
+ while(1) {
+ padata = padata_get_next(&padata_vec[nr]);
+
+ if (!padata || PTR_ERR(padata) == -EINPROGRESS)
+ break;
+ if (PTR_ERR(padata) == -ENODATA) {
+ spin_unlock(&padata_vec[nr].lock);
+ goto out;
+ }
+
+ send_remote_softirq(&padata->csd, padata->cb_cpu,
+ PADATA_SOFTIRQ);
+ }
+
+ if (unlikely(!(padata_vec[nr].flags & PADATA_INIT))) {
+ spin_unlock(&padata_vec[nr].lock);
+ goto reset_out;
+ }
+
+ spin_unlock(&padata_vec[nr].lock);
+
+ if (atomic_read(&padata_vec[nr].queued_objects))
+ goto try_again;
+
+out:
+ return 1;
+reset_out:
+ schedule_work(&padata_vec[nr].work);
+ return 1;
+}
+EXPORT_SYMBOL(padata_do_serial);
+
+/*
+ * padata_cpumask_weight - get the number of cpus that are actually in use
+ *
+ * @nr: number of the padata instance
+ */
+int padata_cpumask_weight(unsigned int nr)
+{
+ return cpumask_weight(padata_vec[nr].cpumask);
+}
+EXPORT_SYMBOL(padata_cpumask_weight);
+
+/*
+ * padata_index_to_cpu - get the cpu for a given cpu index
+ *
+ * @nr: number of the padata instance
+ * @cpu_index: index of the cpu in question
+ *
+ * The range of cpu_index is 0 <= cpu_index < padata_cpumask_weight(),
+ * so padata_cpumask_weight must be called before padata_index_to_cpu.
+ */
+int padata_index_to_cpu(unsigned int nr, int cpu_index)
+{
+ int cpu, target_cpu;
+
+ target_cpu = cpumask_first(padata_vec[nr].cpumask);
+ for (cpu = 0; cpu < cpu_index; cpu++)
+ target_cpu = cpumask_next(target_cpu, padata_vec[nr].cpumask);
+
+ return target_cpu;
+}
+EXPORT_SYMBOL(padata_index_to_cpu);
+
+/*
+ * padata_set_cpumask - set the cpumask that padata uses
+ *
+ * @nr: number of the padata instance
+ * @cpumask: the cpumask to use
+ */
+void padata_set_cpumask(unsigned int nr, cpumask_var_t cpumask)
+{
+ cpumask_copy(padata_vec[nr].new_cpumask, cpumask);
+ padata_vec[nr].flags &= ~PADATA_INIT;
+ padata_vec[nr].flags |= PADATA_FLUSH_HARD;
+
+ schedule_work(&padata_vec[nr].work);
+}
+EXPORT_SYMBOL(padata_set_cpumask);
+
+/*
+ * padata_add_cpu - add a cpu to the padata cpumask
+ *
+ * @nr: number of the padata instance
+ * @cpu: cpu to remove
+ */
+void padata_add_cpu(unsigned int nr, int cpu)
+{
+ cpumask_set_cpu(cpu, padata_vec[nr].cpumask);
+ padata_set_cpumask(nr, padata_vec[nr].cpumask);
+}
+EXPORT_SYMBOL(padata_add_cpu);
+
+/*
+ * padata_remove_cpu - remove a cpu from the padata cpumask
+ *
+ * @nr: number of the padata instance
+ * @cpu: cpu to remove
+ */
+void padata_remove_cpu(unsigned int nr, int cpu)
+{
+ cpumask_set_cpu(cpu, padata_vec[nr].cpumask);
+ padata_set_cpumask(nr, padata_vec[nr].cpumask);
+}
+EXPORT_SYMBOL(padata_remove_cpu);
+
+/*
+ * padata_start - start the parallel processing
+ *
+ * @nr: number of the padata instance
+ */
+void padata_start(unsigned int nr)
+{
+ if (padata_vec[nr].flags & PADATA_INIT)
+ return;
+
+ schedule_work(&padata_vec[nr].work);
+}
+EXPORT_SYMBOL(padata_start);
+
+/*
+ * padata_stop - stop the parallel processing
+ *
+ * @nr: number of the padata instance
+ */
+void padata_stop(unsigned int nr)
+{
+ padata_vec[nr].flags &= ~PADATA_INIT;
+}
+EXPORT_SYMBOL(padata_stop);
+
+/*
+ * padata_init - initialize a padata instance
+ *
+ * @nr: number of the padata instance
+ * @cpumask: cpumask that padata uses for parallelization
+ */
+void __init padata_init(unsigned int nr, const struct cpumask *cpumask)
+{
+ int cpu, cpu_index;
+ struct padata_queue *percpu_queue, *queue;
+
+ percpu_queue = alloc_percpu(struct padata_queue);
+
+ if (!percpu_queue) {
+ printk("padata_init: Failed to alloc the serialization"
+ "queues for padata nr %d, exiting!\n", nr);
+ return;
+ }
+
+ if (!alloc_cpumask_var(&padata_vec[nr].cpumask, GFP_KERNEL))
+ goto err_free;
+
+ if (!alloc_cpumask_var(&padata_vec[nr].new_cpumask, GFP_KERNEL))
+ goto err_free_mask;
+
+ cpu_index = 0;
+
+ for_each_possible_cpu(cpu) {
+ queue = per_cpu_ptr(percpu_queue, cpu);
+
+ if (cpumask_test_cpu(cpu, cpumask)) {
+ queue->cpu_index = cpu_index;
+ cpu_index++;
+ }
+
+ INIT_LIST_HEAD(&queue->list);
+ spin_lock_init(&queue->lock);
+ atomic_set(&queue->num_obj, 0);
+ }
+
+ INIT_WORK(&padata_vec[nr].work, padata_reset_work);
+
+ cpumask_copy(padata_vec[nr].cpumask, cpumask);
+ cpumask_copy(padata_vec[nr].new_cpumask, cpumask);
+
+ atomic_set(&padata_vec[nr].seq_nr, -1);
+ atomic_set(&padata_vec[nr].queued_objects, 0);
+ padata_vec[nr].queue = percpu_queue;
+ padata_vec[nr].flags = 0;
+ spin_lock_init(&padata_vec[nr].lock);
+
+ return;
+
+err_free_mask:
+ free_cpumask_var(padata_vec[nr].new_cpumask);
+
+err_free:
+ free_percpu(percpu_queue);
+}
+
+static int __init padata_initcall(void)
+{
+ open_softirq(PADATA_SOFTIRQ, padata_action);
+
+ return 0;
+}
+subsys_initcall(padata_initcall);
diff --git a/kernel/softirq.c b/kernel/softirq.c
index eb5e131..769ffc4 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -58,7 +58,7 @@ static DEFINE_PER_CPU(struct task_struct *, ksoftirqd);
char *softirq_to_name[NR_SOFTIRQS] = {
"HI", "TIMER", "NET_TX", "NET_RX", "BLOCK",
- "TASKLET", "SCHED", "HRTIMER", "RCU"
+ "TASKLET", "SCHED", "HRTIMER", "PADATA", "RCU"
};
/*
--
1.5.4.2
This patch adds a parallel crypto template that takes a crypto
algorithm and converts it to process the crypto transforms in
parallel. For the moment only aead is supported.
Signed-off-by: Steffen Klassert <[email protected]>
---
crypto/Kconfig | 13 ++
crypto/Makefile | 2 +
crypto/pcrypt.c | 409 +++++++++++++++++++++++++++++++++++++++++++++
crypto/pcrypt_core.c | 106 ++++++++++++
include/crypto/pcrypt.h | 51 ++++++
include/linux/interrupt.h | 2 +
6 files changed, 583 insertions(+), 0 deletions(-)
create mode 100644 crypto/pcrypt.c
create mode 100644 crypto/pcrypt_core.c
create mode 100644 include/crypto/pcrypt.h
diff --git a/crypto/Kconfig b/crypto/Kconfig
index 7623442..dbb1ec0 100644
--- a/crypto/Kconfig
+++ b/crypto/Kconfig
@@ -114,6 +114,19 @@ config CRYPTO_NULL
help
These are 'Null' algorithms, used by IPsec, which do nothing.
+config CRYPTO_PCRYPT_CORE
+ bool
+
+config CRYPTO_PCRYPT
+ tristate "Parallel crypto engine (EXPERIMENTAL)"
+ depends on USE_GENERIC_SMP_HELPERS && EXPERIMENTAL
+ select CRYPTO_MANAGER
+ select CRYPTO_PCRYPT_CORE
+ select CRYPTO_AEAD
+ help
+ This converts an arbitrary crypto algorithm into a parallel
+ algorithm that is executed in a softirq.
+
config CRYPTO_WORKQUEUE
tristate
diff --git a/crypto/Makefile b/crypto/Makefile
index c2ca721..48345da 100644
--- a/crypto/Makefile
+++ b/crypto/Makefile
@@ -55,6 +55,8 @@ obj-$(CONFIG_CRYPTO_XTS) += xts.o
obj-$(CONFIG_CRYPTO_CTR) += ctr.o
obj-$(CONFIG_CRYPTO_GCM) += gcm.o
obj-$(CONFIG_CRYPTO_CCM) += ccm.o
+obj-$(CONFIG_CRYPTO_PCRYPT_CORE) += pcrypt_core.o
+obj-$(CONFIG_CRYPTO_PCRYPT) += pcrypt.o
obj-$(CONFIG_CRYPTO_CRYPTD) += cryptd.o
obj-$(CONFIG_CRYPTO_DES) += des_generic.o
obj-$(CONFIG_CRYPTO_FCRYPT) += fcrypt.o
diff --git a/crypto/pcrypt.c b/crypto/pcrypt.c
new file mode 100644
index 0000000..26b6ec1
--- /dev/null
+++ b/crypto/pcrypt.c
@@ -0,0 +1,409 @@
+/*
+ * pcrypt - Parallel crypto wrapper.
+ *
+ * Copyright (C) 2009 secunet Security Networks AG
+ * Copyright (C) 2009 Steffen Klassert <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <crypto/algapi.h>
+#include <crypto/internal/aead.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <crypto/pcrypt.h>
+
+struct pcrypt_instance_ctx {
+ struct crypto_spawn spawn;
+ unsigned int tfm_count;
+};
+
+struct pcrypt_aead_ctx {
+ struct crypto_aead *child;
+ unsigned int tfm_nr;
+};
+
+static int pcrypt_do_parallel(struct padata_priv *padata, unsigned int tfm_nr,
+ unsigned int softirq, unsigned int padata_nr)
+{
+ unsigned int cpu_index, num_cpus, cb_cpu;
+
+ num_cpus = padata_cpumask_weight(padata_nr);
+ if (!num_cpus)
+ return 0;
+
+ cpu_index = tfm_nr % num_cpus;
+
+ cb_cpu = padata_index_to_cpu(padata_nr, cpu_index);
+
+ return padata_do_parallel(softirq, padata_nr, padata, cb_cpu);
+}
+
+static int pcrypt_aead_setkey(struct crypto_aead *parent,
+ const u8 *key, unsigned int keylen)
+{
+ struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(parent);
+
+ return crypto_aead_setkey(ctx->child, key, keylen);
+}
+
+static int pcrypt_aead_setauthsize(struct crypto_aead *parent,
+ unsigned int authsize)
+{
+ struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(parent);
+
+ return crypto_aead_setauthsize(ctx->child, authsize);
+}
+
+static void pcrypt_aead_serial(struct padata_priv *padata)
+{
+ struct pcrypt_request *preq = pcrypt_padata_request(padata);
+ struct aead_request *req = pcrypt_request_ctx(preq);
+
+ aead_request_complete(req->base.data, padata->info);
+}
+
+static void pcrypt_aead_giv_serial(struct padata_priv *padata)
+{
+ struct pcrypt_request *preq = pcrypt_padata_request(padata);
+ struct aead_givcrypt_request *req = pcrypt_request_ctx(preq);
+
+ aead_request_complete(req->areq.base.data, padata->info);
+}
+
+static void pcrypt_aead_done(struct crypto_async_request *areq, int err)
+{
+ struct aead_request *req = areq->data;
+ struct pcrypt_request *preq = aead_request_ctx(req);
+ struct padata_priv *padata = pcrypt_request_padata(preq);
+
+ padata->info = err;
+ req->base.flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
+
+ local_bh_disable();
+ if (padata_do_serial(padata->nr, padata))
+ goto out;
+
+ aead_request_complete(req, padata->info);
+
+out:
+ local_bh_enable();
+}
+
+static void pcrypt_aead_enc(struct padata_priv *padata)
+{
+ struct pcrypt_request *preq = pcrypt_padata_request(padata);
+ struct aead_request *req = pcrypt_request_ctx(preq);
+
+ padata->info = crypto_aead_encrypt(req);
+
+ if (padata->info)
+ return;
+
+ if (padata_do_serial(AEAD_ENC_PADATA, padata))
+ return;
+
+ aead_request_complete(req->base.data, padata->info);
+}
+
+static int pcrypt_aead_encrypt(struct aead_request *req)
+{
+ int err;
+ struct pcrypt_request *preq = aead_request_ctx(req);
+ struct aead_request *creq = pcrypt_request_ctx(preq);
+ struct padata_priv *padata = pcrypt_request_padata(preq);
+ struct crypto_aead *aead = crypto_aead_reqtfm(req);
+ struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(aead);
+ u32 flags = aead_request_flags(req);
+
+ memset(padata, 0, sizeof(struct padata_priv));
+
+ padata->parallel = pcrypt_aead_enc;
+ padata->serial = pcrypt_aead_serial;
+
+ aead_request_set_tfm(creq, ctx->child);
+ aead_request_set_callback(creq, flags & ~CRYPTO_TFM_REQ_MAY_SLEEP,
+ pcrypt_aead_done, req);
+ aead_request_set_crypt(creq, req->src, req->dst,
+ req->cryptlen, req->iv);
+ aead_request_set_assoc(creq, req->assoc, req->assoclen);
+
+ if (pcrypt_do_parallel(padata, ctx->tfm_nr, AEAD_ENC_SOFTIRQ,
+ AEAD_ENC_PADATA))
+ err = -EINPROGRESS;
+ else
+ err = crypto_aead_encrypt(creq);
+
+ return err;
+}
+
+static void pcrypt_aead_dec(struct padata_priv *padata)
+{
+ struct pcrypt_request *preq = pcrypt_padata_request(padata);
+ struct aead_request *req = pcrypt_request_ctx(preq);
+
+ padata->info = crypto_aead_decrypt(req);
+
+ if (padata->info)
+ return;
+
+ if (padata_do_serial(AEAD_DEC_PADATA, padata))
+ return;
+
+ aead_request_complete(req->base.data, padata->info);
+}
+
+static int pcrypt_aead_decrypt(struct aead_request *req)
+{
+ int err;
+ struct pcrypt_request *preq = aead_request_ctx(req);
+ struct aead_request *creq = pcrypt_request_ctx(preq);
+ struct padata_priv *padata = pcrypt_request_padata(preq);
+ struct crypto_aead *aead = crypto_aead_reqtfm(req);
+ struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(aead);
+ u32 flags = aead_request_flags(req);
+
+ memset(padata, 0, sizeof(struct padata_priv));
+
+ padata->parallel = pcrypt_aead_dec;
+ padata->serial = pcrypt_aead_serial;
+
+ aead_request_set_tfm(creq, ctx->child);
+ aead_request_set_callback(creq, flags & ~CRYPTO_TFM_REQ_MAY_SLEEP,
+ pcrypt_aead_done, req);
+ aead_request_set_crypt(creq, req->src, req->dst,
+ req->cryptlen, req->iv);
+ aead_request_set_assoc(creq, req->assoc, req->assoclen);
+
+ if (pcrypt_do_parallel(padata, ctx->tfm_nr, AEAD_DEC_SOFTIRQ,
+ AEAD_DEC_PADATA))
+ err = -EINPROGRESS;
+ else
+ err = crypto_aead_decrypt(creq);
+
+ return err;
+}
+
+static void pcrypt_aead_givenc(struct padata_priv *padata)
+{
+ struct pcrypt_request *preq = pcrypt_padata_request(padata);
+ struct aead_givcrypt_request *req = pcrypt_request_ctx(preq);
+
+ padata->info = crypto_aead_givencrypt(req);
+
+ if (padata->info)
+ return;
+
+ if (padata_do_serial(AEAD_ENC_PADATA, padata))
+ return;
+
+ aead_request_complete(req->areq.base.data, padata->info);
+}
+
+static int pcrypt_aead_givencrypt(struct aead_givcrypt_request *req)
+{
+ int err;
+ struct aead_request *areq = &req->areq;
+ struct pcrypt_request *preq = aead_request_ctx(areq);
+ struct aead_givcrypt_request *creq = pcrypt_request_ctx(preq);
+ struct padata_priv *padata = pcrypt_request_padata(preq);
+ struct crypto_aead *aead = aead_givcrypt_reqtfm(req);
+ struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(aead);
+ u32 flags = aead_request_flags(areq);
+
+ memset(padata, 0, sizeof(struct padata_priv));
+
+ padata->parallel = pcrypt_aead_givenc;
+ padata->serial = pcrypt_aead_giv_serial;
+
+ aead_givcrypt_set_tfm(creq, ctx->child);
+ aead_givcrypt_set_callback(creq, flags & ~CRYPTO_TFM_REQ_MAY_SLEEP,
+ pcrypt_aead_done, areq);
+ aead_givcrypt_set_crypt(creq, areq->src, areq->dst,
+ areq->cryptlen, areq->iv);
+ aead_givcrypt_set_assoc(creq, areq->assoc, areq->assoclen);
+ aead_givcrypt_set_giv(creq, req->giv, req->seq);
+
+
+ if (pcrypt_do_parallel(padata, ctx->tfm_nr, AEAD_ENC_SOFTIRQ,
+ AEAD_ENC_PADATA))
+ err = -EINPROGRESS;
+ else
+ err = crypto_aead_givencrypt(creq);
+
+ return err;
+}
+
+static int pcrypt_aead_init_tfm(struct crypto_tfm *tfm)
+{
+ struct crypto_instance *inst = crypto_tfm_alg_instance(tfm);
+ struct pcrypt_instance_ctx *ictx = crypto_instance_ctx(inst);
+ struct pcrypt_aead_ctx *ctx = crypto_tfm_ctx(tfm);
+ struct crypto_aead *cipher;
+
+ ictx->tfm_count++;
+ ctx->tfm_nr = ictx->tfm_count;
+
+ cipher = crypto_spawn_aead(crypto_instance_ctx(inst));
+
+ if (IS_ERR(cipher))
+ return PTR_ERR(cipher);
+
+ ctx->child = cipher;
+ tfm->crt_aead.reqsize = sizeof(struct pcrypt_request)
+ + sizeof(struct aead_givcrypt_request)
+ + crypto_aead_reqsize(cipher);
+
+ return 0;
+}
+
+static void pcrypt_aead_exit_tfm(struct crypto_tfm *tfm)
+{
+ struct pcrypt_aead_ctx *ctx = crypto_tfm_ctx(tfm);
+
+ crypto_free_aead(ctx->child);
+}
+
+static struct crypto_instance *pcrypt_alloc_instance(struct crypto_alg *alg)
+{
+ struct crypto_instance *inst;
+ struct pcrypt_instance_ctx *ctx;
+ int err;
+
+ inst = kzalloc(sizeof(*inst) + sizeof(*ctx), GFP_KERNEL);
+ if (!inst) {
+ inst = ERR_PTR(-ENOMEM);
+ goto out;
+ }
+
+ err = -ENAMETOOLONG;
+ if (snprintf(inst->alg.cra_driver_name, CRYPTO_MAX_ALG_NAME,
+ "pcrypt(%s)", alg->cra_driver_name) >= CRYPTO_MAX_ALG_NAME)
+ goto out_free_inst;
+
+ memcpy(inst->alg.cra_name, alg->cra_name, CRYPTO_MAX_ALG_NAME);
+
+ ctx = crypto_instance_ctx(inst);
+ err = crypto_init_spawn(&ctx->spawn, alg, inst,
+ CRYPTO_ALG_TYPE_MASK);
+ if (err)
+ goto out_free_inst;
+
+ inst->alg.cra_priority = alg->cra_priority + 100;
+ inst->alg.cra_blocksize = alg->cra_blocksize;
+ inst->alg.cra_alignmask = alg->cra_alignmask;
+
+out:
+ return inst;
+
+out_free_inst:
+ kfree(inst);
+ inst = ERR_PTR(err);
+ goto out;
+}
+
+static struct crypto_instance *pcrypt_alloc_aead(struct rtattr **tb)
+{
+ struct crypto_instance *inst;
+ struct crypto_alg *alg;
+ struct crypto_attr_type *algt;
+
+ algt = crypto_get_attr_type(tb);
+
+ alg = crypto_get_attr_alg(tb, algt->type,
+ (algt->mask & CRYPTO_ALG_TYPE_MASK));
+ if (IS_ERR(alg))
+ return ERR_CAST(alg);
+
+ inst = pcrypt_alloc_instance(alg);
+ if (IS_ERR(inst))
+ goto out_put_alg;
+
+ inst->alg.cra_flags = CRYPTO_ALG_TYPE_AEAD | CRYPTO_ALG_ASYNC;
+ inst->alg.cra_type = &crypto_aead_type;
+
+ inst->alg.cra_aead.ivsize = alg->cra_aead.ivsize;
+ inst->alg.cra_aead.geniv = alg->cra_aead.geniv;
+ inst->alg.cra_aead.maxauthsize = alg->cra_aead.maxauthsize;
+
+ inst->alg.cra_ctxsize = sizeof(struct pcrypt_aead_ctx);
+
+ inst->alg.cra_init = pcrypt_aead_init_tfm;
+ inst->alg.cra_exit = pcrypt_aead_exit_tfm;
+
+ inst->alg.cra_aead.setkey = pcrypt_aead_setkey;
+ inst->alg.cra_aead.setauthsize = pcrypt_aead_setauthsize;
+ inst->alg.cra_aead.encrypt = pcrypt_aead_encrypt;
+ inst->alg.cra_aead.decrypt = pcrypt_aead_decrypt;
+ inst->alg.cra_aead.givencrypt = pcrypt_aead_givencrypt;
+
+out_put_alg:
+ crypto_mod_put(alg);
+ return inst;
+}
+
+static struct crypto_instance *pcrypt_alloc(struct rtattr **tb)
+{
+ struct crypto_attr_type *algt;
+
+ algt = crypto_get_attr_type(tb);
+ if (IS_ERR(algt))
+ return ERR_CAST(algt);
+
+ switch (algt->type & algt->mask & CRYPTO_ALG_TYPE_MASK) {
+ case CRYPTO_ALG_TYPE_AEAD:
+ return pcrypt_alloc_aead(tb);
+ }
+
+ return ERR_PTR(-EINVAL);
+}
+
+static void pcrypt_free(struct crypto_instance *inst)
+{
+ struct pcrypt_instance_ctx *ctx = crypto_instance_ctx(inst);
+
+ crypto_drop_spawn(&ctx->spawn);
+ kfree(inst);
+}
+
+static struct crypto_template pcrypt_tmpl = {
+ .name = "pcrypt",
+ .alloc = pcrypt_alloc,
+ .free = pcrypt_free,
+ .module = THIS_MODULE,
+};
+
+static int __init pcrypt_init(void)
+{
+ padata_start(AEAD_ENC_PADATA);
+ padata_start(AEAD_DEC_PADATA);
+
+ return crypto_register_template(&pcrypt_tmpl);
+}
+
+static void __exit pcrypt_exit(void)
+{
+ padata_stop(AEAD_ENC_PADATA);
+ padata_stop(AEAD_DEC_PADATA);
+
+ crypto_unregister_template(&pcrypt_tmpl);
+}
+
+module_init(pcrypt_init);
+module_exit(pcrypt_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Parallel crypto engine");
diff --git a/crypto/pcrypt_core.c b/crypto/pcrypt_core.c
new file mode 100644
index 0000000..065d0a3
--- /dev/null
+++ b/crypto/pcrypt_core.c
@@ -0,0 +1,106 @@
+/*
+ * pcrypt_core.c - Core functions for the pcrypt crypto parallelization
+ *
+ * Copyright (C) 2009 secunet Security Networks AG
+ * Copyright (C) 2009 Steffen Klassert <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <linux/interrupt.h>
+#include <linux/cpu.h>
+#include <linux/err.h>
+#include <linux/module.h>
+#include <crypto/pcrypt.h>
+
+static void aead_enc_action(struct softirq_action *h)
+{
+ struct list_head *cpu_list, local_list;
+
+ cpu_list = &__get_cpu_var(softirq_work_list[AEAD_ENC_SOFTIRQ]);
+
+ local_irq_disable();
+ list_replace_init(cpu_list, &local_list);
+ local_irq_enable();
+
+ while (!list_empty(&local_list)) {
+ struct padata_priv *padata;
+
+ padata = list_entry(local_list.next, struct padata_priv,
+ csd.list);
+
+ list_del_init(&padata->csd.list);
+
+ padata->parallel(padata);
+ }
+}
+
+static void aead_dec_action(struct softirq_action *h)
+{
+ struct list_head *cpu_list, local_list;
+
+ cpu_list = &__get_cpu_var(softirq_work_list[AEAD_DEC_SOFTIRQ]);
+
+ local_irq_disable();
+ list_replace_init(cpu_list, &local_list);
+ local_irq_enable();
+
+ while (!list_empty(&local_list)) {
+ struct padata_priv *padata;
+
+ padata = list_entry(local_list.next, struct padata_priv,
+ csd.list);
+
+ list_del_init(&padata->csd.list);
+
+ padata->parallel(padata);
+ }
+}
+
+static int __devinit pcrypt_cpu_callback(struct notifier_block *nfb,
+ unsigned long action, void *hcpu)
+{
+ int cpu = (unsigned long)hcpu;
+
+ switch (action) {
+ case CPU_ONLINE:
+ case CPU_ONLINE_FROZEN:
+ padata_add_cpu(AEAD_ENC_PADATA, cpu);
+ padata_add_cpu(AEAD_DEC_PADATA, cpu);
+ break;
+
+ case CPU_DEAD:
+ case CPU_DEAD_FROZEN:
+ padata_remove_cpu(AEAD_ENC_PADATA, cpu);
+ padata_remove_cpu(AEAD_DEC_PADATA, cpu);
+
+ break;
+ }
+
+ return NOTIFY_OK;
+}
+
+static int __init pcrypt_init_padata(void)
+{
+ open_softirq(AEAD_ENC_SOFTIRQ, aead_enc_action);
+ open_softirq(AEAD_DEC_SOFTIRQ, aead_dec_action);
+
+ padata_init(AEAD_ENC_PADATA, cpu_online_mask);
+ padata_init(AEAD_DEC_PADATA, cpu_online_mask);
+
+ hotcpu_notifier(pcrypt_cpu_callback, 0);
+
+ return 0;
+}
+subsys_initcall(pcrypt_init_padata);
diff --git a/include/crypto/pcrypt.h b/include/crypto/pcrypt.h
new file mode 100644
index 0000000..d7d8bd8
--- /dev/null
+++ b/include/crypto/pcrypt.h
@@ -0,0 +1,51 @@
+/*
+ * pcrypt - Parallel crypto engine.
+ *
+ * Copyright (C) 2009 secunet Security Networks AG
+ * Copyright (C) 2009 Steffen Klassert <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#ifndef _CRYPTO_PCRYPT_H
+#define _CRYPTO_PCRYPT_H
+
+#include <linux/crypto.h>
+#include <linux/kernel.h>
+#include <linux/padata.h>
+
+struct pcrypt_request {
+ struct padata_priv padata;
+ void *data;
+ void *__ctx[] CRYPTO_MINALIGN_ATTR;
+};
+
+static inline void *pcrypt_request_ctx(struct pcrypt_request *req)
+{
+ return req->__ctx;
+}
+
+static inline
+struct padata_priv *pcrypt_request_padata(struct pcrypt_request *req)
+{
+ return &req->padata;
+}
+
+static inline
+struct pcrypt_request *pcrypt_padata_request(struct padata_priv *padata)
+{
+ return container_of(padata, struct pcrypt_request, padata);
+}
+
+#endif
diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index b648cca..41ef14b 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -347,6 +347,8 @@ enum
TASKLET_SOFTIRQ,
SCHED_SOFTIRQ,
HRTIMER_SOFTIRQ,
+ AEAD_ENC_SOFTIRQ,
+ AEAD_DEC_SOFTIRQ,
PADATA_SOFTIRQ,
RCU_SOFTIRQ, /* Preferable RCU should always be the last softirq */
--
1.5.4.2
On Mon, Aug 31, 2009 at 11:12:49AM +0200, Steffen Klassert wrote:
> This patch introduces an interface to process data objects
> in parallel. On request it is possible to serialize again.
> The parallelized objects return after serialization in the
> same order as they were before the parallelization.
>
> Signed-off-by: Steffen Klassert <[email protected]>
Acked-by: Herbert Xu <[email protected]>
Looks good to me. Dave, do you want to pick this one up?
Thanks,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
On Mon, Aug 31, 2009 at 11:14:42AM +0200, Steffen Klassert wrote:
> This patch adds a parallel crypto template that takes a crypto
> algorithm and converts it to process the crypto transforms in
> parallel. For the moment only aead is supported.
>
> Signed-off-by: Steffen Klassert <[email protected]>
I'll pick this up once we get the first patch through.
Thanks,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
On Sat, Sep 19, 2009 at 04:19:59PM -0700, Herbert Xu wrote:
> On Mon, Aug 31, 2009 at 11:12:49AM +0200, Steffen Klassert wrote:
> > This patch introduces an interface to process data objects
> > in parallel. On request it is possible to serialize again.
> > The parallelized objects return after serialization in the
> > same order as they were before the parallelization.
> >
> > Signed-off-by: Steffen Klassert <[email protected]>
>
> Acked-by: Herbert Xu <[email protected]>
>
> Looks good to me. Dave, do you want to pick this one up?
>
I plan to send a rebased patchset tomorrow,
so who wants to take this one?
From: Steffen Klassert <[email protected]>
Date: Wed, 7 Oct 2009 16:22:41 +0200
> On Sat, Sep 19, 2009 at 04:19:59PM -0700, Herbert Xu wrote:
>> On Mon, Aug 31, 2009 at 11:12:49AM +0200, Steffen Klassert wrote:
>> > This patch introduces an interface to process data objects
>> > in parallel. On request it is possible to serialize again.
>> > The parallelized objects return after serialization in the
>> > same order as they were before the parallelization.
>> >
>> > Signed-off-by: Steffen Klassert <[email protected]>
>>
>> Acked-by: Herbert Xu <[email protected]>
>>
>> Looks good to me. Dave, do you want to pick this one up?
>>
>
> I plan to send a rebased patchset tomorrow,
> so who wants to take this one?
I'll take the padata change, as I agreed with Herbert to do.
But the rest goes to Herbert.