[PATCH v5 0/3] crypto: engine - support for parallel and batch requests

[PATCH v5 0/3] crypto: engine - support for parallel and batch requests

[Iuliana Prodan]

Added support for executing multiple, independent or not, requests
for crypto engine based on a retry mechanism. If hardware was unable
to execute a backlog request, enqueue it back in front of crypto-engine
queue, to keep the order of requests.

Now do_one_request() returns:
>= 0: hardware executed the request successfully;
< 0: this is the old error path. If hardware has support for retry
mechanism, the backlog request is put back in front of crypto-engine
queue. For backwards compatibility, if the retry support is not available,
the crypto-engine will work as before. 
Only MAY_BACKLOG requests are enqueued back into crypto-engine's queue,
since the others can be dropped.

If hardware supports batch requests, crypto-engine can handle this use-case
through do_batch_requests callback.

Since, these new features, cannot be supported by all hardware,
the crypto-engine framework is backward compatible:
- by using the crypto_engine_alloc_init function, to initialize
crypto-engine, the new callback is NULL and retry mechanism is
disabled, so crypto-engine will work as before these changes;
- to support multiple requests, in parallel, retry_support variable
must be set on true, in driver. 
- to support batch requests, do_batch_requests callback must be
implemented in driver, to execute a batch of requests. The link
between the requests, is expected to be done in driver, in
do_one_request(). 

---
Changes since V4:
- added, in algapi a function to add a request in front of queue;
- added a retry mechanism: if hardware is unable to execute
a backlog request, enqueue it back in front of crypto-engine
queue, to keep the order of requests.

Changes since V3:
- removed can_enqueue_hardware callback and added a start-stop
mechanism based on the on the return value of do_one_request().

Changes since V2:
- readded cur_req in crypto-engine, to keep, the exact behavior as before
these changes, if can_enqueue_more is not implemented: send requests
to hardware, _one-by-one_, on crypto_pump_requests, and complete it,
on crypto_finalize_request, and so on.
- do_batch_requests is available only with can_enqueue_more.

Changes since V1:
- changed the name of can_enqueue_hardware callback to can_enqueue_more, and
the argument of this callback to crypto_engine structure (for cases when more
than ore crypto-engine is used).
- added a new patch with support for batch requests.

Changes since V0 (RFC):
- removed max_no_req and no_req, as the number of request that can be
processed in parallel;
- added a new callback, can_enqueue_more, to check whether the hardware
can process a new request.


Iuliana Prodan (3):
  crypto: algapi - create function to add request in front of queue
  crypto: engine - support for parallel requests based on retry
    mechanism
  crypto: engine - support for batch requests

 crypto/algapi.c         |  11 +++
 crypto/crypto_engine.c  | 165 ++++++++++++++++++++++++++++++++--------
 include/crypto/algapi.h |   2 +
 include/crypto/engine.h |  15 +++-
 4 files changed, 161 insertions(+), 32 deletions(-)

-- 
2.17.1

[PATCH v5 1/3] crypto: algapi - create function to add request in front of queue

[Iuliana Prodan]

Add crypto_enqueue_request_head function that enqueues a
MAY_BACKLOG request in front of queue.
This will be used in crypto-engine, on error path. In case a request
was not executed by hardware, enqueue it back in front of queue (to
keep the order of requests).

Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
---
 crypto/algapi.c         | 11 +++++++++++
 include/crypto/algapi.h |  2 ++
 2 files changed, 13 insertions(+)

diff --git a/crypto/algapi.c b/crypto/algapi.c
index 69605e21af92..a8732e8e7843 100644
--- a/crypto/algapi.c
+++ b/crypto/algapi.c
@@ -904,6 +904,17 @@ int crypto_enqueue_request(struct crypto_queue *queue,
 }
 EXPORT_SYMBOL_GPL(crypto_enqueue_request);
 
+void crypto_enqueue_request_head(struct crypto_queue *queue,
+				 struct crypto_async_request *request)
+{
+	/* only backlog requests can be enqueued */
+	BUG_ON(!(request->flags & CRYPTO_TFM_REQ_MAY_BACKLOG));
+
+	queue->qlen++;
+	list_add(&request->list, &queue->list);
+}
+EXPORT_SYMBOL_GPL(crypto_enqueue_request_head);
+
 struct crypto_async_request *crypto_dequeue_request(struct crypto_queue *queue)
 {
 	struct list_head *request;
diff --git a/include/crypto/algapi.h b/include/crypto/algapi.h
index e115f9215ed5..00a9cf98debe 100644
--- a/include/crypto/algapi.h
+++ b/include/crypto/algapi.h
@@ -125,6 +125,8 @@ int crypto_inst_setname(struct crypto_instance *inst, const char *name,
 void crypto_init_queue(struct crypto_queue *queue, unsigned int max_qlen);
 int crypto_enqueue_request(struct crypto_queue *queue,
 			   struct crypto_async_request *request);
+void crypto_enqueue_request_head(struct crypto_queue *queue,
+				 struct crypto_async_request *request);
 struct crypto_async_request *crypto_dequeue_request(struct crypto_queue *queue);
 static inline unsigned int crypto_queue_len(struct crypto_queue *queue)
 {
-- 
2.17.1

[PATCH v5 2/3] crypto: engine - support for parallel requests based on retry mechanism

[Iuliana Prodan]

Added support for executing multiple requests, in parallel,
for crypto engine based on a retry mechanism.
If hardware was unable to execute a backlog request, enqueue it
back in front of crypto-engine queue, to keep the order
of requests.

A new variable is added, retry_support (this is to keep the
backward compatibility of crypto-engine) , which keeps track
whether the hardware has support for retry mechanism and,
also, if can run multiple requests.

If do_one_request() returns:
>= 0: hardware executed the request successfully;
< 0: this is the old error path. If hardware has support for retry
mechanism, the request is put back in front of crypto-engine queue.
For backwards compatibility, if the retry support is not available,
the crypto-engine will work as before.
Only MAY_BACKLOG requests are enqueued back into
crypto-engine's queue, since the others can be dropped.

The new crypto_engine_alloc_init_and_set function, initializes
crypto-engine, sets the maximum size for crypto-engine software
queue (not hardcoded anymore) and the retry_support variable
is set, by default, to false.
On crypto_pump_requests(), if do_one_request() returns >= 0,
a new request is send to hardware, until there is no space in
hardware and do_one_request() returns < 0.

By default, retry_support is false and crypto-engine will
work as before - will send requests to hardware,
one-by-one, on crypto_pump_requests(), and complete it, on
crypto_finalize_request(), and so on.

To support multiple requests, in each driver, retry_support
must be set on true, and if do_one_request() returns an error
the request must not be freed, since it will be enqueued back
into crypto-engine's queue.

When all drivers, that use crypto-engine now, will be updated for
retry mechanism, the retry_support variable can be removed.

Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
---
 crypto/crypto_engine.c  | 140 +++++++++++++++++++++++++++++++---------
 include/crypto/engine.h |  10 ++-
 2 files changed, 118 insertions(+), 32 deletions(-)

diff --git a/crypto/crypto_engine.c b/crypto/crypto_engine.c
index eb029ff1e05a..2e2b879d31bb 100644
--- a/crypto/crypto_engine.c
+++ b/crypto/crypto_engine.c
@@ -22,32 +22,36 @@
  * @err: error number
  */
 static void crypto_finalize_request(struct crypto_engine *engine,
-			     struct crypto_async_request *req, int err)
+				    struct crypto_async_request *req, int err)
 {
 	unsigned long flags;
-	bool finalize_cur_req = false;
+	bool finalize_req = false;
 	int ret;
 	struct crypto_engine_ctx *enginectx;
 
-	spin_lock_irqsave(&engine->queue_lock, flags);
-	if (engine->cur_req == req)
-		finalize_cur_req = true;
-	spin_unlock_irqrestore(&engine->queue_lock, flags);
+	/*
+	 * If hardware cannot enqueue more requests
+	 * and retry mechanism is not supported
+	 * make sure we are completing the current request
+	 */
+	if (!engine->retry_support) {
+		spin_lock_irqsave(&engine->queue_lock, flags);
+		if (engine->cur_req == req) {
+			finalize_req = true;
+			engine->cur_req = NULL;
+		}
+		spin_unlock_irqrestore(&engine->queue_lock, flags);
+	}
 
-	if (finalize_cur_req) {
+	if (finalize_req || engine->retry_support) {
 		enginectx = crypto_tfm_ctx(req->tfm);
-		if (engine->cur_req_prepared &&
+		if (enginectx->op.prepare_request &&
 		    enginectx->op.unprepare_request) {
 			ret = enginectx->op.unprepare_request(engine, req);
 			if (ret)
 				dev_err(engine->dev, "failed to unprepare request\n");
 		}
-		spin_lock_irqsave(&engine->queue_lock, flags);
-		engine->cur_req = NULL;
-		engine->cur_req_prepared = false;
-		spin_unlock_irqrestore(&engine->queue_lock, flags);
 	}
-
 	req->complete(req, err);
 
 	kthread_queue_work(engine->kworker, &engine->pump_requests);
@@ -74,7 +78,7 @@ static void crypto_pump_requests(struct crypto_engine *engine,
 	spin_lock_irqsave(&engine->queue_lock, flags);
 
 	/* Make sure we are not already running a request */
-	if (engine->cur_req)
+	if (!engine->retry_support && engine->cur_req)
 		goto out;
 
 	/* If another context is idling then defer */
@@ -108,13 +112,21 @@ static void crypto_pump_requests(struct crypto_engine *engine,
 		goto out;
 	}
 
+start_request:
 	/* Get the fist request from the engine queue to handle */
 	backlog = crypto_get_backlog(&engine->queue);
 	async_req = crypto_dequeue_request(&engine->queue);
 	if (!async_req)
 		goto out;
 
-	engine->cur_req = async_req;
+	/*
+	 * If hardware doesn't support the retry mechanism,
+	 * keep track of the request we are processing now.
+	 * We'll need it on completion (crypto_finalize_request).
+	 */
+	if (!engine->retry_support)
+		engine->cur_req = async_req;
+
 	if (backlog)
 		backlog->complete(backlog, -EINPROGRESS);
 
@@ -130,7 +142,7 @@ static void crypto_pump_requests(struct crypto_engine *engine,
 		ret = engine->prepare_crypt_hardware(engine);
 		if (ret) {
 			dev_err(engine->dev, "failed to prepare crypt hardware\n");
-			goto req_err;
+			goto req_err_2;
 		}
 	}
 
@@ -141,28 +153,75 @@ static void crypto_pump_requests(struct crypto_engine *engine,
 		if (ret) {
 			dev_err(engine->dev, "failed to prepare request: %d\n",
 				ret);
-			goto req_err;
+			goto req_err_2;
 		}
-		engine->cur_req_prepared = true;
 	}
 	if (!enginectx->op.do_one_request) {
 		dev_err(engine->dev, "failed to do request\n");
 		ret = -EINVAL;
-		goto req_err;
+		goto req_err_1;
 	}
+
 	ret = enginectx->op.do_one_request(engine, async_req);
-	if (ret) {
-		dev_err(engine->dev, "Failed to do one request from queue: %d\n", ret);
-		goto req_err;
+
+	/* Request unsuccessfully executed by hardware */
+	if (ret < 0) {
+		/* Non-backlog requests can be dropped */
+		if (!engine->retry_support ||
+		    !(async_req->flags & CRYPTO_TFM_REQ_MAY_BACKLOG)) {
+			dev_err(engine->dev,
+				"Failed to do one request from queue: %d\n",
+				ret);
+			goto req_err_1;
+		}
+		/*
+		 * If retry mechanism is supported,
+		 * unprepare current request and
+		 * enqueue it back into crypto-engine queue.
+		 */
+		if (enginectx->op.unprepare_request) {
+			ret = enginectx->op.unprepare_request(engine,
+							      async_req);
+			if (ret)
+				dev_err(engine->dev,
+					"failed to unprepare request\n");
+		}
+
+		spin_lock_irqsave(&engine->queue_lock, flags);
+		/*
+		 * If hardware was unable to execute request, enqueue it
+		 * back in front of crypto-engine queue, to keep the order
+		 * of requests.
+		 */
+		crypto_enqueue_request_head(&engine->queue, async_req);
+
+		kthread_queue_work(engine->kworker, &engine->pump_requests);
+		goto out;
 	}
-	return;
 
-req_err:
-	crypto_finalize_request(engine, async_req, ret);
+	goto retry;
+
+req_err_1:
+	if (enginectx->op.unprepare_request) {
+		ret = enginectx->op.unprepare_request(engine, async_req);
+		if (ret)
+			dev_err(engine->dev, "failed to unprepare request\n");
+	}
+
+req_err_2:
+	async_req->complete(async_req, ret);
+
+retry:
+	/* If retry mechanism is supported, send new requests to engine */
+	if (engine->retry_support) {
+		spin_lock_irqsave(&engine->queue_lock, flags);
+		goto start_request;
+	}
 	return;
 
 out:
 	spin_unlock_irqrestore(&engine->queue_lock, flags);
+	return;
 }
 
 static void crypto_pump_work(struct kthread_work *work)
@@ -386,15 +445,20 @@ int crypto_engine_stop(struct crypto_engine *engine)
 EXPORT_SYMBOL_GPL(crypto_engine_stop);
 
 /**
- * crypto_engine_alloc_init - allocate crypto hardware engine structure and
- * initialize it.
+ * crypto_engine_alloc_init_and_set - allocate crypto hardware engine structure
+ * and initialize it by setting the maximum number of entries in the software
+ * crypto-engine queue.
  * @dev: the device attached with one hardware engine
+ * @retry_support: whether hardware has support for retry mechanism
  * @rt: whether this queue is set to run as a realtime task
+ * @qlen: maximum size of the crypto-engine queue
  *
  * This must be called from context that can sleep.
  * Return: the crypto engine structure on success, else NULL.
  */
-struct crypto_engine *crypto_engine_alloc_init(struct device *dev, bool rt)
+struct crypto_engine *crypto_engine_alloc_init_and_set(struct device *dev,
+						       bool retry_support,
+						       bool rt, int qlen)
 {
 	struct sched_param param = { .sched_priority = MAX_RT_PRIO / 2 };
 	struct crypto_engine *engine;
@@ -411,12 +475,12 @@ struct crypto_engine *crypto_engine_alloc_init(struct device *dev, bool rt)
 	engine->running = false;
 	engine->busy = false;
 	engine->idling = false;
-	engine->cur_req_prepared = false;
+	engine->retry_support = retry_support;
 	engine->priv_data = dev;
 	snprintf(engine->name, sizeof(engine->name),
 		 "%s-engine", dev_name(dev));
 
-	crypto_init_queue(&engine->queue, CRYPTO_ENGINE_MAX_QLEN);
+	crypto_init_queue(&engine->queue, qlen);
 	spin_lock_init(&engine->queue_lock);
 
 	engine->kworker = kthread_create_worker(0, "%s", engine->name);
@@ -433,6 +497,22 @@ struct crypto_engine *crypto_engine_alloc_init(struct device *dev, bool rt)
 
 	return engine;
 }
+EXPORT_SYMBOL_GPL(crypto_engine_alloc_init_and_set);
+
+/**
+ * crypto_engine_alloc_init - allocate crypto hardware engine structure and
+ * initialize it.
+ * @dev: the device attached with one hardware engine
+ * @rt: whether this queue is set to run as a realtime task
+ *
+ * This must be called from context that can sleep.
+ * Return: the crypto engine structure on success, else NULL.
+ */
+struct crypto_engine *crypto_engine_alloc_init(struct device *dev, bool rt)
+{
+	return crypto_engine_alloc_init_and_set(dev, false, rt,
+						CRYPTO_ENGINE_MAX_QLEN);
+}
 EXPORT_SYMBOL_GPL(crypto_engine_alloc_init);
 
 /**
diff --git a/include/crypto/engine.h b/include/crypto/engine.h
index e29cd67f93c7..b92d7ff1528a 100644
--- a/include/crypto/engine.h
+++ b/include/crypto/engine.h
@@ -24,7 +24,9 @@
  * @idling: the engine is entering idle state
  * @busy: request pump is busy
  * @running: the engine is on working
- * @cur_req_prepared: current request is prepared
+ * @retry_support: indication that the hardware allows re-execution
+ * of a failed backlog request
+ * crypto-engine, in head position to keep order
  * @list: link with the global crypto engine list
  * @queue_lock: spinlock to syncronise access to request queue
  * @queue: the crypto queue of the engine
@@ -45,7 +47,8 @@ struct crypto_engine {
 	bool			idling;
 	bool			busy;
 	bool			running;
-	bool			cur_req_prepared;
+
+	bool			retry_support;
 
 	struct list_head	list;
 	spinlock_t		queue_lock;
@@ -102,6 +105,9 @@ void crypto_finalize_skcipher_request(struct crypto_engine *engine,
 int crypto_engine_start(struct crypto_engine *engine);
 int crypto_engine_stop(struct crypto_engine *engine);
 struct crypto_engine *crypto_engine_alloc_init(struct device *dev, bool rt);
+struct crypto_engine *crypto_engine_alloc_init_and_set(struct device *dev,
+						       bool retry_support,
+						       bool rt, int qlen);
 int crypto_engine_exit(struct crypto_engine *engine);
 
 #endif /* _CRYPTO_ENGINE_H */
-- 
2.17.1

[PATCH v5 3/3] crypto: engine - support for batch requests

[Iuliana Prodan]

Added support for batch requests, per crypto engine.
A new callback is added, do_batch_requests, which executes a
batch of requests. This has the crypto_engine structure as argument
(for cases when more than one crypto-engine is used).
The crypto_engine_alloc_init_and_set function, initializes
crypto-engine, but also, sets the do_batch_requests callback.
On crypto_pump_requests, if do_batch_requests callback is
implemented in a driver, this will be executed. The link between
the requests will be done in driver, if possible.
do_batch_requests is available only if the hardware has support
for multiple request.

Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
---
 crypto/crypto_engine.c  | 27 ++++++++++++++++++++++++++-
 include/crypto/engine.h |  5 +++++
 2 files changed, 31 insertions(+), 1 deletion(-)

diff --git a/crypto/crypto_engine.c b/crypto/crypto_engine.c
index 2e2b879d31bb..217eae3e8037 100644
--- a/crypto/crypto_engine.c
+++ b/crypto/crypto_engine.c
@@ -221,6 +221,18 @@ static void crypto_pump_requests(struct crypto_engine *engine,
 
 out:
 	spin_unlock_irqrestore(&engine->queue_lock, flags);
+
+	/*
+	 * Batch requests is possible only if
+	 * hardware can enqueue multiple requests
+	 */
+	if (engine->do_batch_requests) {
+		ret = engine->do_batch_requests(engine);
+		if (ret)
+			dev_err(engine->dev, "failed to do batch requests: %d\n",
+				ret);
+	}
+
 	return;
 }
 
@@ -450,6 +462,12 @@ EXPORT_SYMBOL_GPL(crypto_engine_stop);
  * crypto-engine queue.
  * @dev: the device attached with one hardware engine
  * @retry_support: whether hardware has support for retry mechanism
+ * @cbk_do_batch: pointer to a callback function to be invoked when executing a
+ *                a batch of requests.
+ *                This has the form:
+ *                callback(struct crypto_engine *engine)
+ *                where:
+ *                @engine: the crypto engine structure.
  * @rt: whether this queue is set to run as a realtime task
  * @qlen: maximum size of the crypto-engine queue
  *
@@ -458,6 +476,7 @@ EXPORT_SYMBOL_GPL(crypto_engine_stop);
  */
 struct crypto_engine *crypto_engine_alloc_init_and_set(struct device *dev,
 						       bool retry_support,
+						       int (*cbk_do_batch)(struct crypto_engine *engine),
 						       bool rt, int qlen)
 {
 	struct sched_param param = { .sched_priority = MAX_RT_PRIO / 2 };
@@ -477,6 +496,12 @@ struct crypto_engine *crypto_engine_alloc_init_and_set(struct device *dev,
 	engine->idling = false;
 	engine->retry_support = retry_support;
 	engine->priv_data = dev;
+	/*
+	 * Batch requests is possible only if
+	 * hardware has support for retry mechanism.
+	 */
+	engine->do_batch_requests = retry_support ? cbk_do_batch : NULL;
+
 	snprintf(engine->name, sizeof(engine->name),
 		 "%s-engine", dev_name(dev));
 
@@ -510,7 +535,7 @@ EXPORT_SYMBOL_GPL(crypto_engine_alloc_init_and_set);
  */
 struct crypto_engine *crypto_engine_alloc_init(struct device *dev, bool rt)
 {
-	return crypto_engine_alloc_init_and_set(dev, false, rt,
+	return crypto_engine_alloc_init_and_set(dev, false, NULL, rt,
 						CRYPTO_ENGINE_MAX_QLEN);
 }
 EXPORT_SYMBOL_GPL(crypto_engine_alloc_init);
diff --git a/include/crypto/engine.h b/include/crypto/engine.h
index b92d7ff1528a..3f06e40d063a 100644
--- a/include/crypto/engine.h
+++ b/include/crypto/engine.h
@@ -37,6 +37,8 @@
  * @unprepare_crypt_hardware: there are currently no more requests on the
  * queue so the subsystem notifies the driver that it may relax the
  * hardware by issuing this call
+ * @do_batch_requests: execute a batch of requests. Depends on multiple
+ * requests support.
  * @kworker: kthread worker struct for request pump
  * @pump_requests: work struct for scheduling work to the request pump
  * @priv_data: the engine private data
@@ -59,6 +61,8 @@ struct crypto_engine {
 
 	int (*prepare_crypt_hardware)(struct crypto_engine *engine);
 	int (*unprepare_crypt_hardware)(struct crypto_engine *engine);
+	int (*do_batch_requests)(struct crypto_engine *engine);
+
 
 	struct kthread_worker           *kworker;
 	struct kthread_work             pump_requests;
@@ -107,6 +111,7 @@ int crypto_engine_stop(struct crypto_engine *engine);
 struct crypto_engine *crypto_engine_alloc_init(struct device *dev, bool rt);
 struct crypto_engine *crypto_engine_alloc_init_and_set(struct device *dev,
 						       bool retry_support,
+						       int (*cbk_do_batch)(struct crypto_engine *engine),
 						       bool rt, int qlen);
 int crypto_engine_exit(struct crypto_engine *engine);
 
-- 
2.17.1

Re: [PATCH v5 2/3] crypto: engine - support for parallel requests based on retry mechanism

[Herbert Xu]

On Wed, Apr 15, 2020 at 11:26:14PM +0300, Iuliana Prodan wrote:
> Added support for executing multiple requests, in parallel,
> for crypto engine based on a retry mechanism.
> If hardware was unable to execute a backlog request, enqueue it
> back in front of crypto-engine queue, to keep the order
> of requests.
> 
> A new variable is added, retry_support (this is to keep the
> backward compatibility of crypto-engine) , which keeps track
> whether the hardware has support for retry mechanism and,
> also, if can run multiple requests.
> 
> If do_one_request() returns:
> >= 0: hardware executed the request successfully;
> < 0: this is the old error path. If hardware has support for retry
> mechanism, the request is put back in front of crypto-engine queue.
> For backwards compatibility, if the retry support is not available,
> the crypto-engine will work as before.
> Only MAY_BACKLOG requests are enqueued back into
> crypto-engine's queue, since the others can be dropped.

This looks a lot nicer!

However, I do have one little issue with the error case.  I think
we should not lump all errors together.  For queueing errors, we
should requeue regardless of MAY_BACKLOG.  After all, we don't
want to have random packet loss just becayse the queue was full.

For other errors (e.g., a kmalloc error), we should requeue the
MAY_BACKLOG requests and drop everythin else.

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

Re: [PATCH v5 2/3] crypto: engine - support for parallel requests based on retry mechanism

[Iuliana Prodan]

On 4/23/2020 2:47 PM, Herbert Xu wrote:
> On Wed, Apr 15, 2020 at 11:26:14PM +0300, Iuliana Prodan wrote:
>> Added support for executing multiple requests, in parallel,
>> for crypto engine based on a retry mechanism.
>> If hardware was unable to execute a backlog request, enqueue it
>> back in front of crypto-engine queue, to keep the order
>> of requests.
>>
>> A new variable is added, retry_support (this is to keep the
>> backward compatibility of crypto-engine) , which keeps track
>> whether the hardware has support for retry mechanism and,
>> also, if can run multiple requests.
>>
>> If do_one_request() returns:
>>> = 0: hardware executed the request successfully;
>> < 0: this is the old error path. If hardware has support for retry
>> mechanism, the request is put back in front of crypto-engine queue.
>> For backwards compatibility, if the retry support is not available,
>> the crypto-engine will work as before.
>> Only MAY_BACKLOG requests are enqueued back into
>> crypto-engine's queue, since the others can be dropped.
> 
> This looks a lot nicer!
> 
> However, I do have one little issue with the error case.  I think
> we should not lump all errors together.  For queueing errors, we
> should requeue regardless of MAY_BACKLOG.  After all, we don't
> want to have random packet loss just becayse the queue was full.
> 
> For other errors (e.g., a kmalloc error), we should requeue the
> MAY_BACKLOG requests and drop everythin else.
> 

Do you refer to the error codes returned by do_one_request (which sends 
the req to hw for execution)?

If this returns:
1. 0, success;
2. -ENOSPC, I'll requeue it regardless of MAY_BACKLOG;
3. any other error (-EIO, -EINVAL, -ENOMEM, etc), I'll requeue only 
MAY_BACKLOG requests.

Thanks,
Iulia

Re: [PATCH v5 2/3] crypto: engine - support for parallel requests based on retry mechanism

[Herbert Xu]

On Fri, Apr 24, 2020 at 02:34:57PM +0000, Iuliana Prodan wrote:
>
> Do you refer to the error codes returned by do_one_request (which sends 
> the req to hw for execution)?
> 
> If this returns:
> 1. 0, success;
> 2. -ENOSPC, I'll requeue it regardless of MAY_BACKLOG;
> 3. any other error (-EIO, -EINVAL, -ENOMEM, etc), I'll requeue only 
> MAY_BACKLOG requests.

Yes that sounds perfect.

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