[dpdk-dev] [PATCH v2 00/22] vhost: generic vhost API

[Yuanhan Liu]

This patchset makes DPDK vhost library be generic enough, so that we could
build other vhost-user drivers on top of it. For example, SPDK (Storage
Performance Development Kit) is trying to enable vhost-user SCSI.

The basic idea is, let DPDK vhost be a vhost-user agent. It stores all
the info about the virtio device (i.e. vring address, negotiated features,
etc) and let the specific vhost-user driver to fetch them (by the API
provided by DPDK vhost lib). With those info being provided, the vhost-user
driver then could get/put vring entries, thus, it could exchange data
between the guest and host.

The last patch demonstrates how to use these new APIs to implement a
very simple vhost-user net driver, without any fancy features enabled.


Major API/ABI Changes summary
=

- some renames
  * "struct virtio_net_device_ops" ==> "struct vhost_device_ops"
  * "rte_virtio_net.h"  ==> "rte_vhost.h"

- driver related APIs are bond with the socket file
  * rte_vhost_driver_set_features(socket_file, features);
  * rte_vhost_driver_get_features(socket_file, features);
  * rte_vhost_driver_enable_features(socket_file, features)
  * rte_vhost_driver_disable_features(socket_file, features)
  * rte_vhost_driver_callback_register(socket_file, notify_ops);
  * rte_vhost_driver_start(socket_file);
This function replaces rte_vhost_driver_session_start(). Check patch
18 for more information.

- new APIs to fetch guest and vring info
  * rte_vhost_get_mem_table(vid, mem);
  * rte_vhost_get_negotiated_features(vid);
  * rte_vhost_get_vhost_vring(vid, vring_idx, vring);

- new exported structures 
  * struct rte_vhost_vring
  * struct rte_vhost_mem_region
  * struct rte_vhost_memory

- a new device ops callback: features_changed().


Change log
==

v2: - rebase
- updated release note
- updated API comments
- renamed rte_vhost_get_vhost_memory to rte_vhost_get_mem_table

- added a new device callback: features_changed(), bascially for live
  migration support
- introduced rte_vhost_driver_start() to start a specific driver
- misc fixes


Some design choices
===

While making this patchset, I met quite few design choices and here are
two of them, with the issue and the reason I made such choices provided.
Please let me know if you have any comments (or better ideas).

Export public structures or not
---

I made an ABI refactor last time (v16.07): move all the structures
internally and let applications use a "vid" to reference the internal
struct. With that, I hope we could never worry about the annoying ABI
issues.

It works great (and as expected) since then, as far as we only support
virito-net, as far as we can handle all the descs inside vhost lib. It
becomes problematic when a user wants to implement a vhost-user driver
somewhere. For example, it needs do the GPA to VVA translation. Without
any structs exported, some functions like gpa_to_vva() can't be inlined.
Calling it would be costly, especially it's a function we have to invoke
for processing each vring desc.

For that reason, the guest memory regions are exported. With that, the
gpa_to_vva could be inlined.

  
Add helper functions to fetch/update descs or not
-

I intended to do it like this way: introduce one function to get @count
of descs from a specific vring and another one to update the used descs.
It's something like
rte_vhost_vring_get_descs(vid, vring_idx, count, offset, iov, descs);
rte_vhost_vring_update_used_descs(vid, vring_idx, count, offset, descs);

With that, vhost-user driver programmer's task would be easier, as he/she
doesn't have to parse the descs any more (such as to handle indirect desc).

But judging that virtio 1.1 is just emerged and it proposes a completely
ring layout, and most importantly, the vring desc structure is also changed,
I'd like to hold the introducation of such two functions. Otherwise, it's
very likely the two will be invalid when virtio 1.1 is out. Though I think
it may could be addressed with a care design, something like making the IOV
generic enough:

struct rte_vhost_iov {
uint64_tgpa;
uint64_tvva;
uint64_tlen;
};

Instead, I go with the other way: introduce few APIs to export all the vring
infos (vring size, vring addr, callfd, etc), and let the vhost-user driver
read and update the descs. Those info could be passed to vhost-user driver
by introducing one API for each, but for saving few APIs and reducing few
calls for the programmer, I packed few key fields into a new structure, so
that it can be fetched with one call:
struct rte_vhost_vring {
struct vring_desc   *desc;
struct vring_avail  *avail;
struct vring_used   *used;
uint64_tlog_guest_addr;
   
 

[dpdk-dev] [PATCH v2 01/22] vhost: introduce driver features related APIs

[Yuanhan Liu]

Introduce few APIs to set/get/enable/disable driver features.

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---
 doc/guides/prog_guide/vhost_lib.rst| 10 +++-
 lib/librte_vhost/rte_vhost_version.map |  4 ++
 lib/librte_vhost/rte_virtio_net.h  | 51 +++
 lib/librte_vhost/socket.c  | 90 ++
 4 files changed, 153 insertions(+), 2 deletions(-)

diff --git a/doc/guides/prog_guide/vhost_lib.rst 
b/doc/guides/prog_guide/vhost_lib.rst
index f0862e6..6a4d206 100644
--- a/doc/guides/prog_guide/vhost_lib.rst
+++ b/doc/guides/prog_guide/vhost_lib.rst
@@ -53,7 +53,7 @@ vhost library should be able to:
 Vhost API Overview
 --
 
-The following is an overview of the Vhost API functions:
+The following is an overview of some key Vhost API functions:
 
 * ``rte_vhost_driver_register(path, flags)``
 
@@ -110,6 +110,12 @@ The following is an overview of the Vhost API functions:
   of those segments, thus the fewer the segments, the quicker we will get
   the mapping. NOTE: we may speed it by using tree searching in future.
 
+* ``rte_vhost_driver_set_features(path, features)``
+
+  This function sets the feature bits the vhost-user driver supports. The
+  vhost-user driver could be vhost-user net, yet it could be something else,
+  say, vhost-user SCSI.
+
 * ``rte_vhost_driver_session_start()``
 
   This function starts the vhost session loop to handle vhost messages. It
@@ -145,7 +151,7 @@ The following is an overview of the Vhost API functions:
 
   Receives (dequeues) ``count`` packets from guest, and stored them at 
``pkts``.
 
-* ``rte_vhost_feature_disable/rte_vhost_feature_enable(feature_mask)``
+* ``rte_vhost_driver_disable/enable_features(path, features))``
 
   This function disables/enables some features. For example, it can be used to
   disable mergeable buffers and TSO features, which both are enabled by
diff --git a/lib/librte_vhost/rte_vhost_version.map 
b/lib/librte_vhost/rte_vhost_version.map
index 30b4671..ca6259c 100644
--- a/lib/librte_vhost/rte_vhost_version.map
+++ b/lib/librte_vhost/rte_vhost_version.map
@@ -34,6 +34,10 @@ DPDK_16.07 {
 DPDK_17.05 {
global:
 
+   rte_vhost_driver_disable_features;
+   rte_vhost_driver_enable_features;
+   rte_vhost_driver_get_features;
+   rte_vhost_driver_set_features;
rte_vhost_get_mtu;
 
 } DPDK_16.07;
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 56829aa..3daf35c 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -94,6 +94,57 @@ struct virtio_net_device_ops {
 /* Unregister vhost driver. This is only meaningful to vhost user. */
 int rte_vhost_driver_unregister(const char *path);
 
+/**
+ * Set the feature bits the vhost-user driver supports.
+ *
+ * @param path
+ *  The vhost-user socket file path
+ * @return
+ *  0 on success, -1 on failure
+ */
+int rte_vhost_driver_set_features(const char *path, uint64_t features);
+
+/**
+ * Enable vhost-user driver features.
+ *
+ * Note that
+ * - the param @features should be a subset of the feature bits provided
+ *   by rte_vhost_driver_set_features().
+ * - it must be invoked before vhost-user negotiation starts.
+ *
+ * @param path
+ *  The vhost-user socket file path
+ * @param features
+ *  Features to enable
+ * @return
+ *  0 on success, -1 on failure
+ */
+int rte_vhost_driver_enable_features(const char *path, uint64_t features);
+
+/**
+ * Disable vhost-user driver features.
+ *
+ * The two notes at rte_vhost_driver_enable_features() also apply here.
+ *
+ * @param path
+ *  The vhost-user socket file path
+ * @param features
+ *  Features to disable
+ * @return
+ *  0 on success, -1 on failure
+ */
+int rte_vhost_driver_disable_features(const char *path, uint64_t features);
+
+/**
+ * Get the final feature bits for feature negotiation.
+ *
+ * @param path
+ *  The vhost-user socket file path
+ * @return
+ *  Feature bits on success, 0 on failure
+ */
+uint64_t rte_vhost_driver_get_features(const char *path);
+
 /* Register callbacks. */
 int rte_vhost_driver_callback_register(struct virtio_net_device_ops const * 
const);
 /* Start vhost driver session blocking loop. */
diff --git a/lib/librte_vhost/socket.c b/lib/librte_vhost/socket.c
index 6a30a31..bdccd96 100644
--- a/lib/librte_vhost/socket.c
+++ b/lib/librte_vhost/socket.c
@@ -63,6 +63,16 @@ struct vhost_user_socket {
bool is_server;
bool reconnect;
bool dequeue_zero_copy;
+
+   /*
+* The "supported_features" indicates the feature bits the
+* vhost driver supports. The "features" indicates the feature
+* bits after the rte_vhost_driver_features_disable/enable().
+* It is also the final feature bits used for vhost-user
+* features negotiation.
+*/
+   uint64_t supported_features;
+   uint64_t features;
 };
 
 struct vhost_user_connection {
@@ -475,6 +485,86 @@ stru

[dpdk-dev] [PATCH v2 03/22] vhost: use new APIs to handle features

[Yuanhan Liu]

Signed-off-by: Yuanhan Liu 
---
 examples/tep_termination/main.c|  4 +++-
 examples/vhost/main.c  | 43 +-
 lib/librte_vhost/rte_vhost_version.map |  3 ---
 lib/librte_vhost/rte_virtio_net.h  | 13 --
 lib/librte_vhost/socket.c  | 23 +-
 lib/librte_vhost/vhost.c   | 42 -
 lib/librte_vhost/vhost.h   | 21 +
 lib/librte_vhost/vhost_user.c  |  8 +++
 8 files changed, 77 insertions(+), 80 deletions(-)

diff --git a/examples/tep_termination/main.c b/examples/tep_termination/main.c
index 20dafdb..8097dcd 100644
--- a/examples/tep_termination/main.c
+++ b/examples/tep_termination/main.c
@@ -1248,12 +1248,14 @@ static inline void __attribute__((always_inline))
rte_eal_remote_launch(switch_worker,
mbuf_pool, lcore_id);
}
-   rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_MRG_RXBUF);
 
ret = rte_vhost_driver_register((char *)&dev_basename, 0);
if (ret != 0)
rte_exit(EXIT_FAILURE, "failed to register vhost driver.\n");
 
+   rte_vhost_driver_disable_features(dev_basename,
+   1ULL << VIRTIO_NET_F_MRG_RXBUF);
+
rte_vhost_driver_callback_register(&virtio_net_device_ops);
 
rte_vhost_driver_session_start();
diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index 4789947..972a6a8 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -328,16 +328,6 @@ struct mbuf_table {
 
if (port >= rte_eth_dev_count()) return -1;
 
-   if (enable_tx_csum == 0)
-   rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_CSUM);
-
-   if (enable_tso == 0) {
-   rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_HOST_TSO4);
-   rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_HOST_TSO6);
-   rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_GUEST_TSO4);
-   rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_GUEST_TSO6);
-   }
-
rx_rings = (uint16_t)dev_info.max_rx_queues;
/* Configure ethernet device. */
retval = rte_eth_dev_configure(port, rx_rings, tx_rings, &port_conf);
@@ -531,7 +521,6 @@ struct mbuf_table {
vmdq_conf_default.rx_adv_conf.vmdq_rx_conf.rx_mode =
ETH_VMDQ_ACCEPT_BROADCAST |
ETH_VMDQ_ACCEPT_MULTICAST;
-   rte_vhost_feature_enable(1ULL << VIRTIO_NET_F_CTRL_RX);
 
break;
 
@@ -1509,9 +1498,6 @@ static inline void __attribute__((always_inline))
RTE_LCORE_FOREACH_SLAVE(lcore_id)
rte_eal_remote_launch(switch_worker, NULL, lcore_id);
 
-   if (mergeable == 0)
-   rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_MRG_RXBUF);
-
if (client_mode)
flags |= RTE_VHOST_USER_CLIENT;
 
@@ -1520,13 +1506,38 @@ static inline void __attribute__((always_inline))
 
/* Register vhost user driver to handle vhost messages. */
for (i = 0; i < nb_sockets; i++) {
-   ret = rte_vhost_driver_register
-   (socket_files + i * PATH_MAX, flags);
+   char *file = socket_files + i * PATH_MAX;
+   ret = rte_vhost_driver_register(file, flags);
if (ret != 0) {
unregister_drivers(i);
rte_exit(EXIT_FAILURE,
"vhost driver register failure.\n");
}
+   if (mergeable == 0) {
+   rte_vhost_driver_disable_features(file,
+   1ULL << VIRTIO_NET_F_MRG_RXBUF);
+   }
+
+   if (enable_tx_csum == 0) {
+   rte_vhost_driver_disable_features(file,
+   1ULL << VIRTIO_NET_F_CSUM);
+   }
+
+   if (enable_tso == 0) {
+   rte_vhost_driver_disable_features(file,
+   1ULL << VIRTIO_NET_F_HOST_TSO4);
+   rte_vhost_driver_disable_features(file,
+   1ULL << VIRTIO_NET_F_HOST_TSO6);
+   rte_vhost_driver_disable_features(file,
+   1ULL << VIRTIO_NET_F_GUEST_TSO4);
+   rte_vhost_driver_disable_features(file,
+   1ULL << VIRTIO_NET_F_GUEST_TSO6);
+   }
+
+   if (promiscuous) {
+   rte_vhost_driver_enable_features(file,
+   1ULL << VIRTIO_NET_F_CTRL_RX);
+   }
}
 
rte_vhost_driver_callback_register(&virtio_net_device_ops);
diff --git a/lib/librte_vhost/rte_vhost_version.map 
b/lib/librte_vhost/rte_vhost_version.map
index ca6259c..1150017 100644
--- a/lib/librte_vhost/rte_vhost_version.ma

[dpdk-dev] [PATCH v2 02/22] net/vhost: remove feature related APIs

[Yuanhan Liu]

The rte_eth_vhost_feature_disable/enable/get APIs are just a wrapper of
rte_vhost_feature_disable/enable/get. However, the later are going to
be refactored; it's going to take an extra parameter (socket_file path),
to let it be per-device.

Instead of changing those vhost-pmd APIs to adapt to the new vhost APIs,
we could simply remove them, and let vdev to serve this purpose. After
all, vdev options is better for disabling/enabling some features.

Signed-off-by: Yuanhan Liu 
Acked-by: Maxime Coquelin 
---

v2: - write more informative commit log on why they are removed.
- update release note
---
 doc/guides/rel_notes/release_17_05.rst  |  7 +++
 drivers/net/vhost/rte_eth_vhost.c   | 25 
 drivers/net/vhost/rte_eth_vhost.h   | 30 -
 drivers/net/vhost/rte_pmd_vhost_version.map |  3 ---
 4 files changed, 7 insertions(+), 58 deletions(-)

diff --git a/doc/guides/rel_notes/release_17_05.rst 
b/doc/guides/rel_notes/release_17_05.rst
index bb64428..4e405b1 100644
--- a/doc/guides/rel_notes/release_17_05.rst
+++ b/doc/guides/rel_notes/release_17_05.rst
@@ -129,6 +129,13 @@ API Changes
 * The LPM ``next_hop`` field is extended from 8 bits to 21 bits for IPv6
   while keeping ABI compatibility.
 
+   * The following vhost-pmd APIs are removed
+
+ * ``rte_eth_vhost_feature_disable``
+ * ``rte_eth_vhost_feature_enable``
+ * ``rte_eth_vhost_feature_get``
+
+
 ABI Changes
 ---
 
diff --git a/drivers/net/vhost/rte_eth_vhost.c 
b/drivers/net/vhost/rte_eth_vhost.c
index a4435da..83063c2 100644
--- a/drivers/net/vhost/rte_eth_vhost.c
+++ b/drivers/net/vhost/rte_eth_vhost.c
@@ -965,31 +965,6 @@ struct vhost_xstats_name_off {
return 0;
 }
 
-/**
- * Disable features in feature_mask. Returns 0 on success.
- */
-int
-rte_eth_vhost_feature_disable(uint64_t feature_mask)
-{
-   return rte_vhost_feature_disable(feature_mask);
-}
-
-/**
- * Enable features in feature_mask. Returns 0 on success.
- */
-int
-rte_eth_vhost_feature_enable(uint64_t feature_mask)
-{
-   return rte_vhost_feature_enable(feature_mask);
-}
-
-/* Returns currently supported vhost features */
-uint64_t
-rte_eth_vhost_feature_get(void)
-{
-   return rte_vhost_feature_get();
-}
-
 static const struct eth_dev_ops ops = {
.dev_start = eth_dev_start,
.dev_stop = eth_dev_stop,
diff --git a/drivers/net/vhost/rte_eth_vhost.h 
b/drivers/net/vhost/rte_eth_vhost.h
index 7c98b1a..ea4bce4 100644
--- a/drivers/net/vhost/rte_eth_vhost.h
+++ b/drivers/net/vhost/rte_eth_vhost.h
@@ -43,36 +43,6 @@
 
 #include 
 
-/**
- * Disable features in feature_mask.
- *
- * @param feature_mask
- *  Vhost features defined in "linux/virtio_net.h".
- * @return
- *  - On success, zero.
- *  - On failure, a negative value.
- */
-int rte_eth_vhost_feature_disable(uint64_t feature_mask);
-
-/**
- * Enable features in feature_mask.
- *
- * @param feature_mask
- *  Vhost features defined in "linux/virtio_net.h".
- * @return
- *  - On success, zero.
- *  - On failure, a negative value.
- */
-int rte_eth_vhost_feature_enable(uint64_t feature_mask);
-
-/**
- * Returns currently supported vhost features.
- *
- * @return
- *  Vhost features defined in "linux/virtio_net.h".
- */
-uint64_t rte_eth_vhost_feature_get(void);
-
 /*
  * Event description.
  */
diff --git a/drivers/net/vhost/rte_pmd_vhost_version.map 
b/drivers/net/vhost/rte_pmd_vhost_version.map
index 3d44083..695db85 100644
--- a/drivers/net/vhost/rte_pmd_vhost_version.map
+++ b/drivers/net/vhost/rte_pmd_vhost_version.map
@@ -1,9 +1,6 @@
 DPDK_16.04 {
global:
 
-   rte_eth_vhost_feature_disable;
-   rte_eth_vhost_feature_enable;
-   rte_eth_vhost_feature_get;
rte_eth_vhost_get_queue_event;
 
local: *;
-- 
1.9.0

[dpdk-dev] [PATCH v2 04/22] vhost: make notify ops per vhost driver

[Yuanhan Liu]

Assume there is an application both support vhost-user net and
vhost-user scsi, the callback should be different. Making notify
ops per vhost driver allow application define different set of
callbacks for different driver.

Signed-off-by: Yuanhan Liu 
---

v2: - check the return value of callback_register and callback_get
- update release note
---
 doc/guides/prog_guide/vhost_lib.rst|  2 +-
 doc/guides/rel_notes/release_17_05.rst |  3 +++
 drivers/net/vhost/rte_eth_vhost.c  | 20 +++-
 examples/tep_termination/main.c|  7 ++-
 examples/vhost/main.c  |  9 +++--
 lib/librte_vhost/rte_virtio_net.h  |  3 ++-
 lib/librte_vhost/socket.c  | 32 
 lib/librte_vhost/vhost.c   | 16 +---
 lib/librte_vhost/vhost.h   |  5 -
 lib/librte_vhost/vhost_user.c  | 22 --
 10 files changed, 83 insertions(+), 36 deletions(-)

diff --git a/doc/guides/prog_guide/vhost_lib.rst 
b/doc/guides/prog_guide/vhost_lib.rst
index 6a4d206..40f3b3b 100644
--- a/doc/guides/prog_guide/vhost_lib.rst
+++ b/doc/guides/prog_guide/vhost_lib.rst
@@ -122,7 +122,7 @@ The following is an overview of some key Vhost API 
functions:
   starts an infinite loop, therefore it should be called in a dedicated
   thread.
 
-* ``rte_vhost_driver_callback_register(virtio_net_device_ops)``
+* ``rte_vhost_driver_callback_register(path, virtio_net_device_ops)``
 
   This function registers a set of callbacks, to let DPDK applications take
   the appropriate action when some events happen. The following events are
diff --git a/doc/guides/rel_notes/release_17_05.rst 
b/doc/guides/rel_notes/release_17_05.rst
index 4e405b1..dfa636d 100644
--- a/doc/guides/rel_notes/release_17_05.rst
+++ b/doc/guides/rel_notes/release_17_05.rst
@@ -135,6 +135,9 @@ API Changes
  * ``rte_eth_vhost_feature_enable``
  * ``rte_eth_vhost_feature_get``
 
+   * The vhost API ``rte_vhost_driver_callback_register(ops)`` takes one
+ more argument: ``rte_vhost_driver_callback_register(path, ops)``.
+
 
 ABI Changes
 ---
diff --git a/drivers/net/vhost/rte_eth_vhost.c 
b/drivers/net/vhost/rte_eth_vhost.c
index 83063c2..f6ad616 100644
--- a/drivers/net/vhost/rte_eth_vhost.c
+++ b/drivers/net/vhost/rte_eth_vhost.c
@@ -669,6 +669,12 @@ struct vhost_xstats_name_off {
return 0;
 }
 
+static struct virtio_net_device_ops vhost_ops = {
+   .new_device  = new_device,
+   .destroy_device  = destroy_device,
+   .vring_state_changed = vring_state_changed,
+};
+
 int
 rte_eth_vhost_get_queue_event(uint8_t port_id,
struct rte_eth_vhost_queue_event *event)
@@ -738,15 +744,6 @@ struct vhost_xstats_name_off {
 static void *
 vhost_driver_session(void *param __rte_unused)
 {
-   static struct virtio_net_device_ops vhost_ops;
-
-   /* set vhost arguments */
-   vhost_ops.new_device = new_device;
-   vhost_ops.destroy_device = destroy_device;
-   vhost_ops.vring_state_changed = vring_state_changed;
-   if (rte_vhost_driver_callback_register(&vhost_ops) < 0)
-   RTE_LOG(ERR, PMD, "Can't register callbacks\n");
-
/* start event handling */
rte_vhost_driver_session_start();
 
@@ -1079,6 +1076,11 @@ struct vhost_xstats_name_off {
if (rte_vhost_driver_register(iface_name, flags))
goto error;
 
+   if (rte_vhost_driver_callback_register(iface_name, &vhost_ops) < 0) {
+   RTE_LOG(ERR, PMD, "Can't register callbacks\n");
+   goto error;
+   }
+
/* We need only one message handling thread */
if (rte_atomic16_add_return(&nb_started_ports, 1) == 1) {
if (vhost_driver_session_start())
diff --git a/examples/tep_termination/main.c b/examples/tep_termination/main.c
index 8097dcd..18b977e 100644
--- a/examples/tep_termination/main.c
+++ b/examples/tep_termination/main.c
@@ -1256,7 +1256,12 @@ static inline void __attribute__((always_inline))
rte_vhost_driver_disable_features(dev_basename,
1ULL << VIRTIO_NET_F_MRG_RXBUF);
 
-   rte_vhost_driver_callback_register(&virtio_net_device_ops);
+   ret = rte_vhost_driver_callback_register(dev_basename,
+   &virtio_net_device_ops);
+   if (ret != 0) {
+   rte_exit(EXIT_FAILURE,
+   "failed to register vhost driver callbacks.\n");
+   }
 
rte_vhost_driver_session_start();
 
diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index 972a6a8..72a9d69 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -1538,9 +1538,14 @@ static inline void __attribute__((always_inline))
rte_vhost_driver_enable_features(file,
1ULL << VIRTIO_NET_F_CTRL_RX);
}
-   }
 
-   rte_vhost_driver_callback_register(&virtio_net_device_ops);
+   ret = rte_vh

[dpdk-dev] [PATCH v2 06/22] vhost: introduce API to fetch negotiated features

[Yuanhan Liu]

Signed-off-by: Yuanhan Liu 
---
 lib/librte_vhost/rte_vhost_version.map |  2 ++
 lib/librte_vhost/rte_virtio_net.h  | 10 ++
 lib/librte_vhost/vhost.c   | 12 
 3 files changed, 24 insertions(+)

diff --git a/lib/librte_vhost/rte_vhost_version.map 
b/lib/librte_vhost/rte_vhost_version.map
index 664a5f3..e21b788 100644
--- a/lib/librte_vhost/rte_vhost_version.map
+++ b/lib/librte_vhost/rte_vhost_version.map
@@ -37,5 +37,7 @@ DPDK_17.05 {
rte_vhost_driver_set_features;
rte_vhost_get_mem_table;
rte_vhost_get_mtu;
+   rte_vhost_get_negotiated_features
+   rte_vhost_get_vhost_memory;
 
 } DPDK_16.07;
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 5a91f97..57e57e3 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -154,6 +154,16 @@ struct virtio_net_device_ops {
  */
 uint64_t rte_vhost_driver_get_features(const char *path);
 
+/**
+ * Get the feature bits after negotiation
+ *
+ * @param vid
+ *  Vhost device ID
+ * @return
+ *  Negotiated feature bits on success, 0 on failure
+ */
+uint64_t rte_vhost_get_negotiated_features(int vid);
+
 /* Register callbacks. */
 int rte_vhost_driver_callback_register(const char *path,
struct virtio_net_device_ops const * const ops);
diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c
index 2b41652..08dccfb 100644
--- a/lib/librte_vhost/vhost.c
+++ b/lib/librte_vhost/vhost.c
@@ -359,6 +359,18 @@ struct virtio_net *
return 0;
 }
 
+uint64_t
+rte_vhost_get_negotiated_features(int vid)
+{
+   struct virtio_net *dev;
+
+   dev = get_device(vid);
+   if (!dev)
+   return 0;
+
+   return dev->features;
+}
+
 int
 rte_vhost_get_mem_table(int vid, struct rte_vhost_memory **mem)
 {
-- 
1.9.0

[dpdk-dev] [PATCH v2 05/22] vhost: export guest memory regions

[Yuanhan Liu]

Some vhost-user driver may need this info to setup its own page tables
for GPA (guest physical addr) to HPA (host physical addr) translation.
SPDK (Storage Performance Development Kit) is one example.

Besides, by exporting this memory info, we could also export the
gpa_to_vva() as an inline function, which helps for performance.
Otherwise, it has to be referenced indirectly by a "vid".

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---

v2: add API comments
---
 lib/librte_vhost/rte_vhost_version.map |  1 +
 lib/librte_vhost/rte_virtio_net.h  | 38 ++
 lib/librte_vhost/vhost.c   | 23 
 lib/librte_vhost/vhost.h   | 28 ++---
 lib/librte_vhost/vhost_user.c  | 12 +--
 5 files changed, 70 insertions(+), 32 deletions(-)

diff --git a/lib/librte_vhost/rte_vhost_version.map 
b/lib/librte_vhost/rte_vhost_version.map
index 1150017..664a5f3 100644
--- a/lib/librte_vhost/rte_vhost_version.map
+++ b/lib/librte_vhost/rte_vhost_version.map
@@ -35,6 +35,7 @@ DPDK_17.05 {
rte_vhost_driver_enable_features;
rte_vhost_driver_get_features;
rte_vhost_driver_set_features;
+   rte_vhost_get_mem_table;
rte_vhost_get_mtu;
 
 } DPDK_16.07;
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 67bd125..5a91f97 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -59,6 +59,28 @@
 enum {VIRTIO_RXQ, VIRTIO_TXQ, VIRTIO_QNUM};
 
 /**
+ * Information relating to memory regions including offsets to
+ * addresses in QEMUs memory file.
+ */
+struct rte_vhost_mem_region {
+   uint64_t guest_phys_addr;
+   uint64_t guest_user_addr;
+   uint64_t host_user_addr;
+   uint64_t size;
+   void *mmap_addr;
+   uint64_t mmap_size;
+   int fd;
+};
+
+/**
+ * Memory structure includes region and mapping information.
+ */
+struct rte_vhost_memory {
+   uint32_t nregions;
+   struct rte_vhost_mem_region regions[0];
+};
+
+/**
  * Device and vring operations.
  */
 struct virtio_net_device_ops {
@@ -244,4 +266,20 @@ uint16_t rte_vhost_enqueue_burst(int vid, uint16_t 
queue_id,
 uint16_t rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
struct rte_mempool *mbuf_pool, struct rte_mbuf **pkts, uint16_t count);
 
+/**
+ * Get guest mem table: a list of memory regions.
+ *
+ * An rte_vhost_vhost_memory object will be allocated internaly, to hold the
+ * guest memory regions. Application should free it at destroy_device()
+ * callback.
+ *
+ * @param vid
+ *  vhost device ID
+ * @param mem
+ *  To store the returned mem regions
+ * @return
+ *  0 on success, -1 on failure
+ */
+int rte_vhost_get_mem_table(int vid, struct rte_vhost_memory **mem);
+
 #endif /* _VIRTIO_NET_H_ */
diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c
index 7d7bb3c..2b41652 100644
--- a/lib/librte_vhost/vhost.c
+++ b/lib/librte_vhost/vhost.c
@@ -359,6 +359,29 @@ struct virtio_net *
return 0;
 }
 
+int
+rte_vhost_get_mem_table(int vid, struct rte_vhost_memory **mem)
+{
+   struct virtio_net *dev;
+   struct rte_vhost_memory *m;
+   size_t size;
+
+   dev = get_device(vid);
+   if (!dev)
+   return -1;
+
+   size = dev->mem->nregions * sizeof(struct rte_vhost_mem_region);
+   m = malloc(size);
+   if (!m)
+   return -1;
+
+   m->nregions = dev->mem->nregions;
+   memcpy(m->regions, dev->mem->regions, size);
+   *mem = m;
+
+   return 0;
+}
+
 uint16_t
 rte_vhost_avail_entries(int vid, uint16_t queue_id)
 {
diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
index 6186216..6d1986a 100644
--- a/lib/librte_vhost/vhost.h
+++ b/lib/librte_vhost/vhost.h
@@ -166,7 +166,7 @@ struct guest_page {
  */
 struct virtio_net {
/* Frontend (QEMU) memory and memory region information */
-   struct virtio_memory*mem;
+   struct rte_vhost_memory *mem;
uint64_tfeatures;
uint64_tprotocol_features;
int vid;
@@ -192,30 +192,6 @@ struct virtio_net {
struct guest_page   *guest_pages;
 } __rte_cache_aligned;
 
-/**
- * Information relating to memory regions including offsets to
- * addresses in QEMUs memory file.
- */
-struct virtio_memory_region {
-   uint64_t guest_phys_addr;
-   uint64_t guest_user_addr;
-   uint64_t host_user_addr;
-   uint64_t size;
-   void *mmap_addr;
-   uint64_t mmap_size;
-   int fd;
-};
-
-
-/**
- * Memory structure includes region and mapping information.
- */
-struct virtio_memory {
-   uint32_t nregions;
-   struct virtio_memory_region regions[0];
-};
-
-
 /* Macros for printing using RTE_LOG */
 #define RTE_LOGTYPE_VHOST_CONFIG RTE_LOGTYPE_USER1
 #define RTE_LOGTYPE_VHOST_DATA   RTE_LOGTYPE_USER1
@@ -255,7 +231,7 @@ struct virtio_memory {
 static inline 

[dpdk-dev] [PATCH v2 08/22] vhost: export API to translate gpa to vva

[Yuanhan Liu]

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---
 lib/librte_vhost/rte_vhost_version.map |  1 +
 lib/librte_vhost/rte_virtio_net.h  | 28 
 lib/librte_vhost/vhost.h   | 19 ---
 lib/librte_vhost/virtio_net.c  | 23 +--
 4 files changed, 42 insertions(+), 29 deletions(-)

diff --git a/lib/librte_vhost/rte_vhost_version.map 
b/lib/librte_vhost/rte_vhost_version.map
index 28664a4..7df7af6 100644
--- a/lib/librte_vhost/rte_vhost_version.map
+++ b/lib/librte_vhost/rte_vhost_version.map
@@ -40,5 +40,6 @@ DPDK_17.05 {
rte_vhost_get_negotiated_features
rte_vhost_get_vhost_memory;
rte_vhost_get_vhost_vring;
+   rte_vhost_gpa_to_vva;
 
 } DPDK_16.07;
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 5142337..36674bb 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -103,6 +103,34 @@ struct virtio_net_device_ops {
void *reserved[5]; /**< Reserved for future extension */
 };
 
+/**
+ * Convert guest physical address to host virtual address
+ *
+ * @param mem
+ *  the guest memory regions
+ * @param gpa
+ *  the guest physical address for querying
+ * @return
+ *  the host virtual address on success, 0 on failure
+ */
+static inline uint64_t __attribute__((always_inline))
+rte_vhost_gpa_to_vva(struct rte_vhost_memory *mem, uint64_t gpa)
+{
+   struct rte_vhost_mem_region *reg;
+   uint32_t i;
+
+   for (i = 0; i < mem->nregions; i++) {
+   reg = &mem->regions[i];
+   if (gpa >= reg->guest_phys_addr &&
+   gpa <  reg->guest_phys_addr + reg->size) {
+   return gpa - reg->guest_phys_addr +
+  reg->host_user_addr;
+   }
+   }
+
+   return 0;
+}
+
 int rte_vhost_enable_guest_notification(int vid, uint16_t queue_id, int 
enable);
 
 /**
diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
index 68ca197..b5c5046 100644
--- a/lib/librte_vhost/vhost.h
+++ b/lib/librte_vhost/vhost.h
@@ -229,25 +229,6 @@ struct virtio_net {
 #define MAX_VHOST_DEVICE   1024
 extern struct virtio_net *vhost_devices[MAX_VHOST_DEVICE];
 
-/* Convert guest physical Address to host virtual address */
-static inline uint64_t __attribute__((always_inline))
-gpa_to_vva(struct virtio_net *dev, uint64_t gpa)
-{
-   struct rte_vhost_mem_region *reg;
-   uint32_t i;
-
-   for (i = 0; i < dev->mem->nregions; i++) {
-   reg = &dev->mem->regions[i];
-   if (gpa >= reg->guest_phys_addr &&
-   gpa <  reg->guest_phys_addr + reg->size) {
-   return gpa - reg->guest_phys_addr +
-  reg->host_user_addr;
-   }
-   }
-
-   return 0;
-}
-
 /* Convert guest physical address to host physical address */
 static inline phys_addr_t __attribute__((always_inline))
 gpa_to_hpa(struct virtio_net *dev, uint64_t gpa, uint64_t size)
diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c
index 337470d..6b9b4c3 100644
--- a/lib/librte_vhost/virtio_net.c
+++ b/lib/librte_vhost/virtio_net.c
@@ -199,7 +199,7 @@ static inline int __attribute__((always_inline))
uint16_t nr_desc = 1;
 
desc = &descs[desc_idx];
-   desc_addr = gpa_to_vva(dev, desc->addr);
+   desc_addr = rte_vhost_gpa_to_vva(dev->mem, desc->addr);
/*
 * Checking of 'desc_addr' placed outside of 'unlikely' macro to avoid
 * performance issue with some versions of gcc (4.8.4 and 5.3.0) which
@@ -239,7 +239,7 @@ static inline int __attribute__((always_inline))
return -1;
 
desc = &descs[desc->next];
-   desc_addr = gpa_to_vva(dev, desc->addr);
+   desc_addr = rte_vhost_gpa_to_vva(dev->mem, desc->addr);
if (unlikely(!desc_addr))
return -1;
 
@@ -323,7 +323,8 @@ static inline uint32_t __attribute__((always_inline))
int err;
 
if (vq->desc[desc_idx].flags & VRING_DESC_F_INDIRECT) {
-   descs = (struct vring_desc *)(uintptr_t)gpa_to_vva(dev,
+   descs = (struct vring_desc *)(uintptr_t)
+   rte_vhost_gpa_to_vva(dev->mem,
vq->desc[desc_idx].addr);
if (unlikely(!descs)) {
count = i;
@@ -383,7 +384,7 @@ static inline int __attribute__((always_inline))
 
if (vq->desc[idx].flags & VRING_DESC_F_INDIRECT) {
descs = (struct vring_desc *)(uintptr_t)
-   gpa_to_vva(dev, vq->desc[idx].addr);
+   rte_vhost_gpa_to_vva(dev->mem, vq->desc[idx].addr);
if (unlikely(!descs))
 

[dpdk-dev] [PATCH v2 07/22] vhost: export vhost vring info

[Yuanhan Liu]

Signed-off-by: Yuanhan Liu 
---

v2: - fix off-by-one check
- add API comments
---
 lib/librte_vhost/rte_vhost_version.map |  1 +
 lib/librte_vhost/rte_virtio_net.h  | 26 ++
 lib/librte_vhost/vhost.c   | 30 ++
 lib/librte_vhost/vhost.h   |  2 ++
 4 files changed, 59 insertions(+)

diff --git a/lib/librte_vhost/rte_vhost_version.map 
b/lib/librte_vhost/rte_vhost_version.map
index e21b788..28664a4 100644
--- a/lib/librte_vhost/rte_vhost_version.map
+++ b/lib/librte_vhost/rte_vhost_version.map
@@ -39,5 +39,6 @@ DPDK_17.05 {
rte_vhost_get_mtu;
rte_vhost_get_negotiated_features
rte_vhost_get_vhost_memory;
+   rte_vhost_get_vhost_vring;
 
 } DPDK_16.07;
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 57e57e3..5142337 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -80,6 +80,17 @@ struct rte_vhost_memory {
struct rte_vhost_mem_region regions[0];
 };
 
+struct rte_vhost_vring {
+   struct vring_desc   *desc;
+   struct vring_avail  *avail;
+   struct vring_used   *used;
+   uint64_tlog_guest_addr;
+
+   int callfd;
+   int kickfd;
+   uint16_tsize;
+};
+
 /**
  * Device and vring operations.
  */
@@ -292,4 +303,19 @@ uint16_t rte_vhost_dequeue_burst(int vid, uint16_t 
queue_id,
  */
 int rte_vhost_get_mem_table(int vid, struct rte_vhost_memory **mem);
 
+/**
+ * Get guest vring info, including the vring address, vring size, etc.
+ *
+ * @param vid
+ *  vhost device ID
+ * @param vring_idx
+ *  vring index
+ * @param vring
+ *  the structure to hold the requested vring info
+ * @return
+ *  0 on success, -1 on failure
+ */
+int rte_vhost_get_vhost_vring(int vid, uint16_t vring_idx,
+ struct rte_vhost_vring *vring);
+
 #endif /* _VIRTIO_NET_H_ */
diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c
index 08dccfb..6fe613b 100644
--- a/lib/librte_vhost/vhost.c
+++ b/lib/librte_vhost/vhost.c
@@ -394,6 +394,36 @@ struct virtio_net *
return 0;
 }
 
+int
+rte_vhost_get_vhost_vring(int vid, uint16_t vring_idx,
+ struct rte_vhost_vring *vring)
+{
+   struct virtio_net *dev;
+   struct vhost_virtqueue *vq;
+
+   dev = get_device(vid);
+   if (!dev)
+   return -1;
+
+   if (vring_idx >= VHOST_MAX_VRING)
+   return -1;
+
+   vq = dev->virtqueue[vring_idx];
+   if (!vq)
+   return -1;
+
+   vring->desc  = vq->desc;
+   vring->avail = vq->avail;
+   vring->used  = vq->used;
+   vring->log_guest_addr  = vq->log_guest_addr;
+
+   vring->callfd  = vq->callfd;
+   vring->kickfd  = vq->kickfd;
+   vring->size= vq->size;
+
+   return 0;
+}
+
 uint16_t
 rte_vhost_avail_entries(int vid, uint16_t queue_id)
 {
diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
index 6d1986a..68ca197 100644
--- a/lib/librte_vhost/vhost.h
+++ b/lib/librte_vhost/vhost.h
@@ -120,6 +120,8 @@ struct vhost_virtqueue {
 #ifndef VIRTIO_NET_F_MQ
  #define VIRTIO_NET_F_MQ   22
 #endif
+
+#define VHOST_MAX_VRING0x100
 #define VHOST_MAX_QUEUE_PAIRS  0x80
 
 #ifndef VIRTIO_NET_F_MTU
-- 
1.9.0

[dpdk-dev] [PATCH v2 09/22] vhost: turn queue pair to vring

[Yuanhan Liu]

The queue pair is very virtio-net specific, other devices don't have
such concept. To make it generic, we should log the number of vrings
instead of the number of queue pairs.

This patch just does a simple convert, a later patch would export the
number of vrings to applications.

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---

v2: enable all vrings by unconditionally
---
 lib/librte_vhost/vhost.c  | 80 +++
 lib/librte_vhost/vhost.h  |  4 +--
 lib/librte_vhost/vhost_user.c | 28 +--
 lib/librte_vhost/virtio_net.c | 10 +++---
 4 files changed, 42 insertions(+), 80 deletions(-)

diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c
index 6fe613b..70477c6 100644
--- a/lib/librte_vhost/vhost.c
+++ b/lib/librte_vhost/vhost.c
@@ -84,10 +84,8 @@ struct virtio_net *
 
vhost_backend_cleanup(dev);
 
-   for (i = 0; i < dev->virt_qp_nb; i++) {
-   cleanup_vq(dev->virtqueue[i * VIRTIO_QNUM + VIRTIO_RXQ], 
destroy);
-   cleanup_vq(dev->virtqueue[i * VIRTIO_QNUM + VIRTIO_TXQ], 
destroy);
-   }
+   for (i = 0; i < dev->nr_vring; i++)
+   cleanup_vq(dev->virtqueue[i], destroy);
 }
 
 /*
@@ -97,24 +95,21 @@ struct virtio_net *
 free_device(struct virtio_net *dev)
 {
uint32_t i;
-   struct vhost_virtqueue *rxq, *txq;
+   struct vhost_virtqueue *vq;
 
-   for (i = 0; i < dev->virt_qp_nb; i++) {
-   rxq = dev->virtqueue[i * VIRTIO_QNUM + VIRTIO_RXQ];
-   txq = dev->virtqueue[i * VIRTIO_QNUM + VIRTIO_TXQ];
+   for (i = 0; i < dev->nr_vring; i++) {
+   vq = dev->virtqueue[i];
 
-   rte_free(rxq->shadow_used_ring);
-   rte_free(txq->shadow_used_ring);
+   rte_free(vq->shadow_used_ring);
 
-   /* rxq and txq are allocated together as queue-pair */
-   rte_free(rxq);
+   rte_free(vq);
}
 
rte_free(dev);
 }
 
 static void
-init_vring_queue(struct vhost_virtqueue *vq, int qp_idx)
+init_vring_queue(struct vhost_virtqueue *vq)
 {
memset(vq, 0, sizeof(struct vhost_virtqueue));
 
@@ -124,69 +119,48 @@ struct virtio_net *
/* Backends are set to -1 indicating an inactive device. */
vq->backend = -1;
 
-   /* always set the default vq pair to enabled */
-   if (qp_idx == 0)
-   vq->enabled = 1;
+   /*
+* always set the vq to enabled; this is to keep compatibility
+* with the old QEMU, whereas there is no SET_VRING_ENABLE message.
+*/
+   vq->enabled = 1;
 
TAILQ_INIT(&vq->zmbuf_list);
 }
 
 static void
-init_vring_queue_pair(struct virtio_net *dev, uint32_t qp_idx)
-{
-   uint32_t base_idx = qp_idx * VIRTIO_QNUM;
-
-   init_vring_queue(dev->virtqueue[base_idx + VIRTIO_RXQ], qp_idx);
-   init_vring_queue(dev->virtqueue[base_idx + VIRTIO_TXQ], qp_idx);
-}
-
-static void
-reset_vring_queue(struct vhost_virtqueue *vq, int qp_idx)
+reset_vring_queue(struct vhost_virtqueue *vq)
 {
int callfd;
 
callfd = vq->callfd;
-   init_vring_queue(vq, qp_idx);
+   init_vring_queue(vq);
vq->callfd = callfd;
 }
 
-static void
-reset_vring_queue_pair(struct virtio_net *dev, uint32_t qp_idx)
-{
-   uint32_t base_idx = qp_idx * VIRTIO_QNUM;
-
-   reset_vring_queue(dev->virtqueue[base_idx + VIRTIO_RXQ], qp_idx);
-   reset_vring_queue(dev->virtqueue[base_idx + VIRTIO_TXQ], qp_idx);
-}
-
 int
-alloc_vring_queue_pair(struct virtio_net *dev, uint32_t qp_idx)
+alloc_vring_queue(struct virtio_net *dev, uint32_t vring_idx)
 {
-   struct vhost_virtqueue *virtqueue = NULL;
-   uint32_t virt_rx_q_idx = qp_idx * VIRTIO_QNUM + VIRTIO_RXQ;
-   uint32_t virt_tx_q_idx = qp_idx * VIRTIO_QNUM + VIRTIO_TXQ;
+   struct vhost_virtqueue *vq;
 
-   virtqueue = rte_malloc(NULL,
-  sizeof(struct vhost_virtqueue) * VIRTIO_QNUM, 0);
-   if (virtqueue == NULL) {
+   vq = rte_malloc(NULL, sizeof(struct vhost_virtqueue), 0);
+   if (vq == NULL) {
RTE_LOG(ERR, VHOST_CONFIG,
-   "Failed to allocate memory for virt qp:%d.\n", qp_idx);
+   "Failed to allocate memory for vring:%u.\n", vring_idx);
return -1;
}
 
-   dev->virtqueue[virt_rx_q_idx] = virtqueue;
-   dev->virtqueue[virt_tx_q_idx] = virtqueue + VIRTIO_TXQ;
-
-   init_vring_queue_pair(dev, qp_idx);
+   dev->virtqueue[vring_idx] = vq;
+   init_vring_queue(vq);
 
-   dev->virt_qp_nb += 1;
+   dev->nr_vring += 1;
 
return 0;
 }
 
 /*
  * Reset some variables in device structure, while keeping few
- * others untouched, such as vid, ifname, virt_qp_nb: they
+ * others untouched, such as vid, ifname, nr_vring: they
  * should be same unless the device is removed.
  */
 void
@@ -198,8 +172,8 @@ struct virtio_net *
dev->protocol_features = 0;
 

[dpdk-dev] [PATCH v2 10/22] vhost: export the number of vrings

[Yuanhan Liu]

We used to use rte_vhost_get_queue_num() for telling how many vrings.
However, the return value is the number of "queue pairs", which is
very virtio-net specific. To make it generic, we should return the
number of vrings instead, and let the driver do the proper translation.
Say, virtio-net driver could turn it to the number of queue pairs by
dividing 2.

Meanwhile, mark rte_vhost_get_queue_num as deprecated.

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---
v2: - update release note
---
 doc/guides/rel_notes/release_17_05.rst |  3 +++
 drivers/net/vhost/rte_eth_vhost.c  |  2 +-
 lib/librte_vhost/rte_vhost_version.map |  1 +
 lib/librte_vhost/rte_virtio_net.h  | 17 +
 lib/librte_vhost/vhost.c   | 11 +++
 5 files changed, 33 insertions(+), 1 deletion(-)

diff --git a/doc/guides/rel_notes/release_17_05.rst 
b/doc/guides/rel_notes/release_17_05.rst
index dfa636d..eca9451 100644
--- a/doc/guides/rel_notes/release_17_05.rst
+++ b/doc/guides/rel_notes/release_17_05.rst
@@ -138,6 +138,9 @@ API Changes
* The vhost API ``rte_vhost_driver_callback_register(ops)`` takes one
  more argument: ``rte_vhost_driver_callback_register(path, ops)``.
 
+   * The vhost API ``rte_vhost_get_queue_num`` is deprecated, instead,
+ ``rte_vhost_get_vring_num`` should be used.
+
 
 ABI Changes
 ---
diff --git a/drivers/net/vhost/rte_eth_vhost.c 
b/drivers/net/vhost/rte_eth_vhost.c
index f6ad616..dc583e4 100644
--- a/drivers/net/vhost/rte_eth_vhost.c
+++ b/drivers/net/vhost/rte_eth_vhost.c
@@ -569,7 +569,7 @@ struct vhost_xstats_name_off {
vq->port = eth_dev->data->port_id;
}
 
-   for (i = 0; i < rte_vhost_get_queue_num(vid) * VIRTIO_QNUM; i++)
+   for (i = 0; i < rte_vhost_get_vring_num(vid); i++)
rte_vhost_enable_guest_notification(vid, i, 0);
 
rte_vhost_get_mtu(vid, ð_dev->data->mtu);
diff --git a/lib/librte_vhost/rte_vhost_version.map 
b/lib/librte_vhost/rte_vhost_version.map
index 7df7af6..ff62c39 100644
--- a/lib/librte_vhost/rte_vhost_version.map
+++ b/lib/librte_vhost/rte_vhost_version.map
@@ -40,6 +40,7 @@ DPDK_17.05 {
rte_vhost_get_negotiated_features
rte_vhost_get_vhost_memory;
rte_vhost_get_vhost_vring;
+   rte_vhost_get_vring_num;
rte_vhost_gpa_to_vva;
 
 } DPDK_16.07;
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 36674bb..f700d2f 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -237,17 +237,34 @@ int rte_vhost_driver_callback_register(const char *path,
 int rte_vhost_get_numa_node(int vid);
 
 /**
+ * @deprecated
  * Get the number of queues the device supports.
  *
+ * Note this function is deprecated, as it returns a queue pair number,
+ * which is virtio-net specific. Instead, rte_vhost_get_vring_num should
+ * be used.
+ *
  * @param vid
  *  virtio-net device ID
  *
  * @return
  *  The number of queues, 0 on failure
  */
+__rte_deprecated
 uint32_t rte_vhost_get_queue_num(int vid);
 
 /**
+ * Get the number of vrings the device supports.
+ *
+ * @param vid
+ *  vhost device ID
+ *
+ * @return
+ *  The number of vrings, 0 on failure
+ */
+uint16_t rte_vhost_get_vring_num(int vid);
+
+/**
  * Get the virtio net device's ifname, which is the vhost-user socket
  * file path.
  *
diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c
index 70477c6..74ae3b2 100644
--- a/lib/librte_vhost/vhost.c
+++ b/lib/librte_vhost/vhost.c
@@ -317,6 +317,17 @@ struct virtio_net *
return dev->nr_vring / 2;
 }
 
+uint16_t
+rte_vhost_get_vring_num(int vid)
+{
+   struct virtio_net *dev = get_device(vid);
+
+   if (dev == NULL)
+   return 0;
+
+   return dev->nr_vring;
+}
+
 int
 rte_vhost_get_ifname(int vid, char *buf, size_t len)
 {
-- 
1.9.0

[dpdk-dev] [PATCH v2 14/22] vhost: rename device ops struct

[Yuanhan Liu]

rename "virtio_net_device_ops" to "vhost_device_ops", to not let it
be virtio-net specific.

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---
 doc/guides/prog_guide/vhost_lib.rst| 2 +-
 doc/guides/rel_notes/release_17_05.rst | 3 +++
 drivers/net/vhost/rte_eth_vhost.c  | 2 +-
 examples/tep_termination/main.c| 2 +-
 examples/vhost/main.c  | 2 +-
 lib/librte_vhost/Makefile  | 2 +-
 lib/librte_vhost/rte_virtio_net.h  | 4 ++--
 lib/librte_vhost/socket.c  | 6 +++---
 lib/librte_vhost/vhost.h   | 4 ++--
 9 files changed, 15 insertions(+), 12 deletions(-)

diff --git a/doc/guides/prog_guide/vhost_lib.rst 
b/doc/guides/prog_guide/vhost_lib.rst
index 40f3b3b..e6e34f3 100644
--- a/doc/guides/prog_guide/vhost_lib.rst
+++ b/doc/guides/prog_guide/vhost_lib.rst
@@ -122,7 +122,7 @@ The following is an overview of some key Vhost API 
functions:
   starts an infinite loop, therefore it should be called in a dedicated
   thread.
 
-* ``rte_vhost_driver_callback_register(path, virtio_net_device_ops)``
+* ``rte_vhost_driver_callback_register(path, vhost_device_ops)``
 
   This function registers a set of callbacks, to let DPDK applications take
   the appropriate action when some events happen. The following events are
diff --git a/doc/guides/rel_notes/release_17_05.rst 
b/doc/guides/rel_notes/release_17_05.rst
index 2b56e80..2efe292 100644
--- a/doc/guides/rel_notes/release_17_05.rst
+++ b/doc/guides/rel_notes/release_17_05.rst
@@ -154,6 +154,9 @@ API Changes
  * ``linux/if.h``
  * ``rte_ether.h``
 
+   * The vhost struct ``virtio_net_device_ops`` is renamed to
+ ``vhost_device_ops``
+
 
 ABI Changes
 ---
diff --git a/drivers/net/vhost/rte_eth_vhost.c 
b/drivers/net/vhost/rte_eth_vhost.c
index 891ee70..97a765f 100644
--- a/drivers/net/vhost/rte_eth_vhost.c
+++ b/drivers/net/vhost/rte_eth_vhost.c
@@ -671,7 +671,7 @@ struct vhost_xstats_name_off {
return 0;
 }
 
-static struct virtio_net_device_ops vhost_ops = {
+static struct vhost_device_ops vhost_ops = {
.new_device  = new_device,
.destroy_device  = destroy_device,
.vring_state_changed = vring_state_changed,
diff --git a/examples/tep_termination/main.c b/examples/tep_termination/main.c
index 18b977e..738f2d2 100644
--- a/examples/tep_termination/main.c
+++ b/examples/tep_termination/main.c
@@ -1081,7 +1081,7 @@ static inline void __attribute__((always_inline))
  * These callback allow devices to be added to the data core when configuration
  * has been fully complete.
  */
-static const struct virtio_net_device_ops virtio_net_device_ops = {
+static const struct vhost_device_ops virtio_net_device_ops = {
.new_device =  new_device,
.destroy_device = destroy_device,
 };
diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index 72a9d69..4395306 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -1270,7 +1270,7 @@ static inline void __attribute__((always_inline))
  * These callback allow devices to be added to the data core when configuration
  * has been fully complete.
  */
-static const struct virtio_net_device_ops virtio_net_device_ops =
+static const struct vhost_device_ops virtio_net_device_ops =
 {
.new_device =  new_device,
.destroy_device = destroy_device,
diff --git a/lib/librte_vhost/Makefile b/lib/librte_vhost/Makefile
index 415ffc6..5cf4e93 100644
--- a/lib/librte_vhost/Makefile
+++ b/lib/librte_vhost/Makefile
@@ -36,7 +36,7 @@ LIB = librte_vhost.a
 
 EXPORT_MAP := rte_vhost_version.map
 
-LIBABIVER := 3
+LIBABIVER := 4
 
 CFLAGS += $(WERROR_FLAGS) -I$(SRCDIR) -O3 -D_FILE_OFFSET_BITS=64
 CFLAGS += -I vhost_user
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 0063949..26ac35f 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -87,7 +87,7 @@ struct rte_vhost_vring {
 /**
  * Device and vring operations.
  */
-struct virtio_net_device_ops {
+struct vhost_device_ops {
int (*new_device)(int vid); /**< Add device. */
void (*destroy_device)(int vid);/**< Remove device. */
 
@@ -198,7 +198,7 @@ static inline uint64_t __attribute__((always_inline))
 
 /* Register callbacks. */
 int rte_vhost_driver_callback_register(const char *path,
-   struct virtio_net_device_ops const * const ops);
+   struct vhost_device_ops const * const ops);
 /* Start vhost driver session blocking loop. */
 int rte_vhost_driver_session_start(void);
 
diff --git a/lib/librte_vhost/socket.c b/lib/librte_vhost/socket.c
index 8431511..31b868d 100644
--- a/lib/librte_vhost/socket.c
+++ b/lib/librte_vhost/socket.c
@@ -74,7 +74,7 @@ struct vhost_user_socket {
uint64_t supported_features;
uint64_t features;
 
-   struct virtio_net_device_ops const *notify_ops;
+   struct vhost_device_ops const *notify_ops;
 };
 
 struct vhost_user_connection {
@@ -725,7 +725,7 @@ 

[dpdk-dev] [PATCH v2 11/22] vhost: move the device ready check at proper place

[Yuanhan Liu]

Currently, we check vq->desc, vq->kickfd and vq->callfd to know whether
a virtio device is ready or not. However, we only do it when handling
SET_VRING_KICK message, which could be wrong if a vhost-user frontend
send SET_VRING_KICK first and SET_VRING_CALL later.

To work for all possible vhost-user frontend implementations, we could
move the ready check at the end of vhost-user message handler.

Meanwhile, since we do the check more often than before, the "virtio
not ready" message is dropped, to not flood the screen.

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---
 lib/librte_vhost/vhost_user.c | 40 ++--
 1 file changed, 18 insertions(+), 22 deletions(-)

diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
index f841c9b..7f93f27 100644
--- a/lib/librte_vhost/vhost_user.c
+++ b/lib/librte_vhost/vhost_user.c
@@ -609,14 +609,14 @@
struct vhost_virtqueue *vq;
uint32_t i;
 
+   if (dev->nr_vring == 0)
+   return 0;
+
for (i = 0; i < dev->nr_vring; i++) {
vq = dev->virtqueue[i];
 
-   if (!vq_is_ready(vq)) {
-   RTE_LOG(INFO, VHOST_CONFIG,
-   "virtio is not ready for processing.\n");
+   if (!vq_is_ready(vq))
return 0;
-   }
}
 
RTE_LOG(INFO, VHOST_CONFIG,
@@ -645,10 +645,6 @@
vq->callfd = file.fd;
 }
 
-/*
- *  In vhost-user, when we receive kick message, will test whether virtio
- *  device is ready for packet processing.
- */
 static void
 vhost_user_set_vring_kick(struct virtio_net *dev, struct VhostUserMsg *pmsg)
 {
@@ -667,20 +663,6 @@
if (vq->kickfd >= 0)
close(vq->kickfd);
vq->kickfd = file.fd;
-
-   if (virtio_is_ready(dev)) {
-   dev->flags |= VIRTIO_DEV_READY;
-
-   if (!(dev->flags & VIRTIO_DEV_RUNNING)) {
-   if (dev->dequeue_zero_copy) {
-   RTE_LOG(INFO, VHOST_CONFIG,
-   "dequeue zero copy is 
enabled\n");
-   }
-
-   if (dev->notify_ops->new_device(dev->vid) == 0)
-   dev->flags |= VIRTIO_DEV_RUNNING;
-   }
-   }
 }
 
 static void
@@ -1102,5 +1084,19 @@
send_vhost_message(fd, &msg);
}
 
+   if (!(dev->flags & VIRTIO_DEV_RUNNING) && virtio_is_ready(dev)) {
+   dev->flags |= VIRTIO_DEV_READY;
+
+   if (!(dev->flags & VIRTIO_DEV_RUNNING)) {
+   if (dev->dequeue_zero_copy) {
+   RTE_LOG(INFO, VHOST_CONFIG,
+   "dequeue zero copy is 
enabled\n");
+   }
+
+   if (dev->notify_ops->new_device(dev->vid) == 0)
+   dev->flags |= VIRTIO_DEV_RUNNING;
+   }
+   }
+
return 0;
 }
-- 
1.9.0

[dpdk-dev] [PATCH v2 12/22] vhost: drop the Rx and Tx queue macro

[Yuanhan Liu]

They are virtio-net specific and should be defined inside the virtio-net
driver.

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---

v2: - update release note
---
 doc/guides/rel_notes/release_17_05.rst | 6 ++
 drivers/net/vhost/rte_eth_vhost.c  | 2 ++
 examples/tep_termination/main.h| 2 ++
 examples/vhost/main.h  | 2 ++
 lib/librte_vhost/rte_virtio_net.h  | 3 ---
 5 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/doc/guides/rel_notes/release_17_05.rst 
b/doc/guides/rel_notes/release_17_05.rst
index eca9451..55bf136 100644
--- a/doc/guides/rel_notes/release_17_05.rst
+++ b/doc/guides/rel_notes/release_17_05.rst
@@ -141,6 +141,12 @@ API Changes
* The vhost API ``rte_vhost_get_queue_num`` is deprecated, instead,
  ``rte_vhost_get_vring_num`` should be used.
 
+   * Few macros are removed in ``rte_virtio_net.h``
+
+ * ``VIRTIO_RXQ``
+ * ``VIRTIO_TXQ``
+ * ``VIRTIO_QNUM``
+
 
 ABI Changes
 ---
diff --git a/drivers/net/vhost/rte_eth_vhost.c 
b/drivers/net/vhost/rte_eth_vhost.c
index dc583e4..891ee70 100644
--- a/drivers/net/vhost/rte_eth_vhost.c
+++ b/drivers/net/vhost/rte_eth_vhost.c
@@ -45,6 +45,8 @@
 
 #include "rte_eth_vhost.h"
 
+enum {VIRTIO_RXQ, VIRTIO_TXQ, VIRTIO_QNUM};
+
 #define ETH_VHOST_IFACE_ARG"iface"
 #define ETH_VHOST_QUEUES_ARG   "queues"
 #define ETH_VHOST_CLIENT_ARG   "client"
diff --git a/examples/tep_termination/main.h b/examples/tep_termination/main.h
index c0ea766..8ed817d 100644
--- a/examples/tep_termination/main.h
+++ b/examples/tep_termination/main.h
@@ -54,6 +54,8 @@
 /* Max number of devices. Limited by the application. */
 #define MAX_DEVICES 64
 
+enum {VIRTIO_RXQ, VIRTIO_TXQ, VIRTIO_QNUM};
+
 /* Per-device statistics struct */
 struct device_statistics {
uint64_t tx_total;
diff --git a/examples/vhost/main.h b/examples/vhost/main.h
index 6bb42e8..7a3d251 100644
--- a/examples/vhost/main.h
+++ b/examples/vhost/main.h
@@ -41,6 +41,8 @@
 #define RTE_LOGTYPE_VHOST_DATA   RTE_LOGTYPE_USER2
 #define RTE_LOGTYPE_VHOST_PORT   RTE_LOGTYPE_USER3
 
+enum {VIRTIO_RXQ, VIRTIO_TXQ, VIRTIO_QNUM};
+
 struct device_statistics {
uint64_ttx;
uint64_ttx_total;
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index f700d2f..1ae1920 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -55,9 +55,6 @@
 #define RTE_VHOST_USER_NO_RECONNECT(1ULL << 1)
 #define RTE_VHOST_USER_DEQUEUE_ZERO_COPY   (1ULL << 2)
 
-/* Enum for virtqueue management. */
-enum {VIRTIO_RXQ, VIRTIO_TXQ, VIRTIO_QNUM};
-
 /**
  * Information relating to memory regions including offsets to
  * addresses in QEMUs memory file.
-- 
1.9.0

[dpdk-dev] [PATCH v2 13/22] vhost: do not include net specific headers

[Yuanhan Liu]

Include it internally, at vhost.h.

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---

v2: - update release note
---
 doc/guides/rel_notes/release_17_05.rst | 7 +++
 examples/vhost/main.h  | 2 ++
 lib/librte_vhost/rte_virtio_net.h  | 4 
 lib/librte_vhost/vhost.h   | 4 
 4 files changed, 13 insertions(+), 4 deletions(-)

diff --git a/doc/guides/rel_notes/release_17_05.rst 
b/doc/guides/rel_notes/release_17_05.rst
index 55bf136..2b56e80 100644
--- a/doc/guides/rel_notes/release_17_05.rst
+++ b/doc/guides/rel_notes/release_17_05.rst
@@ -147,6 +147,13 @@ API Changes
  * ``VIRTIO_TXQ``
  * ``VIRTIO_QNUM``
 
+   * Few net specific header files are removed in ``rte_virtio_net.h``
+
+ * ``linux/virtio_net.h``
+ * ``sys/socket.h``
+ * ``linux/if.h``
+ * ``rte_ether.h``
+
 
 ABI Changes
 ---
diff --git a/examples/vhost/main.h b/examples/vhost/main.h
index 7a3d251..ddcd858 100644
--- a/examples/vhost/main.h
+++ b/examples/vhost/main.h
@@ -36,6 +36,8 @@
 
 #include 
 
+#include 
+
 /* Macros for printing using RTE_LOG */
 #define RTE_LOGTYPE_VHOST_CONFIG RTE_LOGTYPE_USER1
 #define RTE_LOGTYPE_VHOST_DATA   RTE_LOGTYPE_USER2
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 1ae1920..0063949 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -42,14 +42,10 @@
 #include 
 #include 
 #include 
-#include 
 #include 
-#include 
-#include 
 
 #include 
 #include 
-#include 
 
 #define RTE_VHOST_USER_CLIENT  (1ULL << 0)
 #define RTE_VHOST_USER_NO_RECONNECT(1ULL << 1)
diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
index 84e379a..672098b 100644
--- a/lib/librte_vhost/vhost.h
+++ b/lib/librte_vhost/vhost.h
@@ -39,8 +39,12 @@
 #include 
 #include 
 #include 
+#include 
+#include 
+#include 
 
 #include 
+#include 
 
 #include "rte_virtio_net.h"
 
-- 
1.9.0

[dpdk-dev] [PATCH v2 16/22] vhost: add features changed callback

[Yuanhan Liu]

Features could be changed after the feature negotiation. For example,
VHOST_F_LOG_ALL will be set/cleared at the start/end of live migration,
respecitively. Thus, we need a new callback to inform the application
on such change.

Signed-off-by: Yuanhan Liu 
---
 doc/guides/prog_guide/vhost_lib.rst |  6 ++
 lib/librte_vhost/rte_virtio_net.h   | 10 +-
 lib/librte_vhost/vhost_user.c   |  5 +
 3 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/doc/guides/prog_guide/vhost_lib.rst 
b/doc/guides/prog_guide/vhost_lib.rst
index 88f0591..a4fb1f1 100644
--- a/doc/guides/prog_guide/vhost_lib.rst
+++ b/doc/guides/prog_guide/vhost_lib.rst
@@ -143,6 +143,12 @@ The following is an overview of some key Vhost API 
functions:
 This callback is invoked when a specific queue's state is changed, for
 example to enabled or disabled.
 
+  * ``features_changed(int vid, uint64_t features)``
+
+This callback is invoked when the features is changed. For example,
+``VHOST_F_LOG_ALL`` will be set/cleared at the start/end of live
+migration, respectively.
+
 * ``rte_vhost_enqueue_burst(vid, queue_id, pkts, count)``
 
   Transmits (enqueues) ``count`` packets from host to guest.
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 845d0fd..4256927 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -93,7 +93,15 @@ struct vhost_device_ops {
 
int (*vring_state_changed)(int vid, uint16_t queue_id, int enable); 
/**< triggered when a vring is enabled or disabled */
 
-   void *reserved[5]; /**< Reserved for future extension */
+   /**
+* Features could be changed after the feature negotiation.
+* For example, VHOST_F_LOG_ALL will be set/cleared at the
+* start/end of live migration, respectively. This callback
+* is used to inform the application on such change.
+*/
+   int (*features_changed)(int vid, uint64_t features);
+
+   void *reserved[4]; /**< Reserved for future extension */
 };
 
 /**
diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
index 7f93f27..40cc973 100644
--- a/lib/librte_vhost/vhost_user.c
+++ b/lib/librte_vhost/vhost_user.c
@@ -161,6 +161,11 @@
if (features & ~rte_vhost_driver_get_features(dev->ifname))
return -1;
 
+   if ((dev->flags & VIRTIO_DEV_RUNNING) && dev->features != features) {
+   if (dev->notify_ops->features_changed)
+   dev->notify_ops->features_changed(dev->vid, features);
+   }
+
dev->features = features;
if (dev->features &
((1 << VIRTIO_NET_F_MRG_RXBUF) | (1ULL << VIRTIO_F_VERSION_1))) 
{
-- 
1.9.0

[dpdk-dev] [PATCH v2 15/22] vhost: rename virtio-net to vhost

[Yuanhan Liu]

Rename "virtio-net" to "vhost" in the API comments and vhost prog guide.

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---
 doc/guides/prog_guide/vhost_lib.rst |  6 +++---
 lib/librte_vhost/rte_virtio_net.h   | 14 +++---
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/doc/guides/prog_guide/vhost_lib.rst 
b/doc/guides/prog_guide/vhost_lib.rst
index e6e34f3..88f0591 100644
--- a/doc/guides/prog_guide/vhost_lib.rst
+++ b/doc/guides/prog_guide/vhost_lib.rst
@@ -130,12 +130,12 @@ The following is an overview of some key Vhost API 
functions:
 
   * ``new_device(int vid)``
 
-This callback is invoked when a virtio net device becomes ready. ``vid``
-is the virtio net device ID.
+This callback is invoked when a virtio device becomes ready. ``vid``
+is the vhost device ID.
 
   * ``destroy_device(int vid)``
 
-This callback is invoked when a virtio net device shuts down (or when the
+This callback is invoked when a virtio device shuts down (or when the
 vhost connection is broken).
 
   * ``vring_state_changed(int vid, uint16_t queue_id, int enable)``
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 26ac35f..845d0fd 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -222,7 +222,7 @@ int rte_vhost_driver_callback_register(const char *path,
  * is allocated.
  *
  * @param vid
- *  virtio-net device ID
+ *  vhost device ID
  *
  * @return
  *  The numa node, -1 on failure
@@ -234,11 +234,11 @@ int rte_vhost_driver_callback_register(const char *path,
  * Get the number of queues the device supports.
  *
  * Note this function is deprecated, as it returns a queue pair number,
- * which is virtio-net specific. Instead, rte_vhost_get_vring_num should
+ * which is vhost specific. Instead, rte_vhost_get_vring_num should
  * be used.
  *
  * @param vid
- *  virtio-net device ID
+ *  vhost device ID
  *
  * @return
  *  The number of queues, 0 on failure
@@ -262,7 +262,7 @@ int rte_vhost_driver_callback_register(const char *path,
  * file path.
  *
  * @param vid
- *  virtio-net device ID
+ *  vhost device ID
  * @param buf
  *  The buffer to stored the queried ifname
  * @param len
@@ -277,7 +277,7 @@ int rte_vhost_driver_callback_register(const char *path,
  * Get how many avail entries are left in the queue
  *
  * @param vid
- *  virtio-net device ID
+ *  vhost device ID
  * @param queue_id
  *  virtio queue index
  *
@@ -292,7 +292,7 @@ int rte_vhost_driver_callback_register(const char *path,
  * count is returned to indicate the number of packets that were succesfully
  * added to the RX queue.
  * @param vid
- *  virtio-net device ID
+ *  vhost device ID
  * @param queue_id
  *  virtio queue index in mq case
  * @param pkts
@@ -310,7 +310,7 @@ uint16_t rte_vhost_enqueue_burst(int vid, uint16_t queue_id,
  * construct host mbufs, copies guest buffer content to host mbufs and
  * store them in pkts to be processed.
  * @param vid
- *  virtio-net device
+ *  vhost device ID
  * @param queue_id
  *  virtio queue index in mq case
  * @param mbuf_pool
-- 
1.9.0

[dpdk-dev] [PATCH v2 19/22] vhost: rename header file

[Yuanhan Liu]

Rename "rte_virtio_net.h" to "rte_vhost.h", to not let it be virtio
net specific.

Signed-off-by: Yuanhan Liu 
Reviewed-by: Maxime Coquelin 
---
 doc/guides/rel_notes/deprecation.rst   |   9 -
 doc/guides/rel_notes/release_17_05.rst |   3 +
 drivers/net/vhost/rte_eth_vhost.c  |   2 +-
 drivers/net/vhost/rte_eth_vhost.h  |   2 +-
 examples/tep_termination/main.c|   2 +-
 examples/tep_termination/vxlan_setup.c |   2 +-
 examples/vhost/main.c  |   2 +-
 lib/librte_vhost/Makefile  |   2 +-
 lib/librte_vhost/rte_vhost.h   | 421 +
 lib/librte_vhost/rte_virtio_net.h  | 421 -
 lib/librte_vhost/vhost.c   |   2 +-
 lib/librte_vhost/vhost.h   |   2 +-
 lib/librte_vhost/vhost_user.h  |   2 +-
 lib/librte_vhost/virtio_net.c  |   2 +-
 14 files changed, 434 insertions(+), 440 deletions(-)
 create mode 100644 lib/librte_vhost/rte_vhost.h
 delete mode 100644 lib/librte_vhost/rte_virtio_net.h

diff --git a/doc/guides/rel_notes/deprecation.rst 
b/doc/guides/rel_notes/deprecation.rst
index d6544ed..9708b39 100644
--- a/doc/guides/rel_notes/deprecation.rst
+++ b/doc/guides/rel_notes/deprecation.rst
@@ -95,15 +95,6 @@ Deprecation Notices
   Target release for removal of the legacy API will be defined once most
   PMDs have switched to rte_flow.
 
-* vhost: API/ABI changes are planned for 17.05, for making DPDK vhost library
-  generic enough so that applications can build different vhost-user drivers
-  (instead of vhost-user net only) on top of that.
-  Specifically, ``virtio_net_device_ops`` will be renamed to 
``vhost_device_ops``.
-  Correspondingly, some API's parameter need be changed. Few more functions 
also
-  need be reworked to let it be device aware. For example, different virtio 
device
-  has different feature set, meaning functions like 
``rte_vhost_feature_disable``
-  need be changed. Last, file rte_virtio_net.h will be renamed to rte_vhost.h.
-
 * ABI changes are planned for 17.05 in the ``rte_cryptodev_ops`` structure.
   A pointer to a rte_cryptodev_config structure will be added to the
   function prototype ``cryptodev_configure_t``, as a new parameter.
diff --git a/doc/guides/rel_notes/release_17_05.rst 
b/doc/guides/rel_notes/release_17_05.rst
index 8f06fc4..c053fff 100644
--- a/doc/guides/rel_notes/release_17_05.rst
+++ b/doc/guides/rel_notes/release_17_05.rst
@@ -165,6 +165,9 @@ API Changes
* The vhost API ``rte_vhost_driver_session_start`` is removed. Instead,
  ``rte_vhost_driver_start`` should be used.
 
+   * The vhost public header file ``rte_virtio_net.h`` is renamed to
+ ``rte_vhost.h``
+
 
 ABI Changes
 ---
diff --git a/drivers/net/vhost/rte_eth_vhost.c 
b/drivers/net/vhost/rte_eth_vhost.c
index e6c0758..32e774b 100644
--- a/drivers/net/vhost/rte_eth_vhost.c
+++ b/drivers/net/vhost/rte_eth_vhost.c
@@ -40,7 +40,7 @@
 #include 
 #include 
 #include 
-#include 
+#include 
 #include 
 
 #include "rte_eth_vhost.h"
diff --git a/drivers/net/vhost/rte_eth_vhost.h 
b/drivers/net/vhost/rte_eth_vhost.h
index ea4bce4..39ca771 100644
--- a/drivers/net/vhost/rte_eth_vhost.h
+++ b/drivers/net/vhost/rte_eth_vhost.h
@@ -41,7 +41,7 @@
 #include 
 #include 
 
-#include 
+#include 
 
 /*
  * Event description.
diff --git a/examples/tep_termination/main.c b/examples/tep_termination/main.c
index 24c62cd..cd6e3f1 100644
--- a/examples/tep_termination/main.c
+++ b/examples/tep_termination/main.c
@@ -49,7 +49,7 @@
 #include 
 #include 
 #include 
-#include 
+#include 
 
 #include "main.h"
 #include "vxlan.h"
diff --git a/examples/tep_termination/vxlan_setup.c 
b/examples/tep_termination/vxlan_setup.c
index 8f1f15b..87de74d 100644
--- a/examples/tep_termination/vxlan_setup.c
+++ b/examples/tep_termination/vxlan_setup.c
@@ -49,7 +49,7 @@
 #include 
 
 #include "main.h"
-#include "rte_virtio_net.h"
+#include "rte_vhost.h"
 #include "vxlan.h"
 #include "vxlan_setup.h"
 
diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index 64b3eea..08b82f6 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -49,7 +49,7 @@
 #include 
 #include 
 #include 
-#include 
+#include 
 #include 
 #include 
 
diff --git a/lib/librte_vhost/Makefile b/lib/librte_vhost/Makefile
index 5cf4e93..4847069 100644
--- a/lib/librte_vhost/Makefile
+++ b/lib/librte_vhost/Makefile
@@ -51,7 +51,7 @@ SRCS-$(CONFIG_RTE_LIBRTE_VHOST) := fd_man.c socket.c vhost.c 
vhost_user.c \
   virtio_net.c
 
 # install includes
-SYMLINK-$(CONFIG_RTE_LIBRTE_VHOST)-include += rte_virtio_net.h
+SYMLINK-$(CONFIG_RTE_LIBRTE_VHOST)-include += rte_vhost.h
 
 # dependencies
 DEPDIRS-$(CONFIG_RTE_LIBRTE_VHOST) += lib/librte_eal
diff --git a/lib/librte_vhost/rte_vhost.h b/lib/librte_vhost/rte_vhost.h
new file mode 100644
index 000..d4ee210
--- /dev/null
+++ b/lib/librte_vhost/rte_vhost.h
@@ -0,0 +1,421 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 20

[dpdk-dev] [PATCH v2 20/22] vhost: workaround the build dependency on mbuf header

[Yuanhan Liu]

rte_mbuf struct is something more likely will be used only in vhost-user
net driver, while we have made vhost-user generic enough that it can
be used for implementing other drivers (such as vhost-user SCSI), they
have also include . Otherwise, the build will be broken.

We could workaround it by using forward declaration, so that other
non-net drivers won't need include .

Signed-off-by: Yuanhan Liu 
---
 lib/librte_vhost/rte_vhost.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/lib/librte_vhost/rte_vhost.h b/lib/librte_vhost/rte_vhost.h
index d4ee210..bc1a958 100644
--- a/lib/librte_vhost/rte_vhost.h
+++ b/lib/librte_vhost/rte_vhost.h
@@ -348,6 +348,8 @@ int rte_vhost_driver_callback_register(const char *path,
  */
 uint16_t rte_vhost_avail_entries(int vid, uint16_t queue_id);
 
+struct rte_mbuf;
+struct rte_mempool;
 /**
  * This function adds buffers to the virtio devices RX virtqueue. Buffers can
  * be received from the physical port or from another virtual device. A packet
-- 
1.9.0

[dpdk-dev] [PATCH v2 17/22] vhost: export APIs for live migration support

[Yuanhan Liu]

Export few APIs for the vhost-user driver to log the guest memory writes,
which is a must for live migration support.

This patch basically moves vhost_log_write() and vhost_log_used_vring()
into vhost.h and then add an wrapper (the public API) to them.

Signed-off-by: Yuanhan Liu 
---
 lib/librte_vhost/rte_vhost_version.map |  2 ++
 lib/librte_vhost/rte_virtio_net.h  | 43 ++
 lib/librte_vhost/vhost.c   | 31 
 lib/librte_vhost/vhost.h   | 38 ++
 lib/librte_vhost/virtio_net.c  | 36 
 5 files changed, 114 insertions(+), 36 deletions(-)

diff --git a/lib/librte_vhost/rte_vhost_version.map 
b/lib/librte_vhost/rte_vhost_version.map
index ff62c39..70c28f7 100644
--- a/lib/librte_vhost/rte_vhost_version.map
+++ b/lib/librte_vhost/rte_vhost_version.map
@@ -42,5 +42,7 @@ DPDK_17.05 {
rte_vhost_get_vhost_vring;
rte_vhost_get_vring_num;
rte_vhost_gpa_to_vva;
+   rte_vhost_log_used_vring;
+   rte_vhost_log_write;
 
 } DPDK_16.07;
diff --git a/lib/librte_vhost/rte_virtio_net.h 
b/lib/librte_vhost/rte_virtio_net.h
index 4256927..11b204d 100644
--- a/lib/librte_vhost/rte_virtio_net.h
+++ b/lib/librte_vhost/rte_virtio_net.h
@@ -132,6 +132,49 @@ static inline uint64_t __attribute__((always_inline))
return 0;
 }
 
+#define RTE_VHOST_NEED_LOG(features)   ((features) & (1ULL << VHOST_F_LOG_ALL))
+
+/**
+ * Log the memory write start with given address.
+ *
+ * This function only need be invoked when the live migration starts.
+ * Therefore, we won't need call it at all in the most of time. For
+ * making the performance impact be minimum, it's suggested to do a
+ * check before calling it:
+ *
+ *if (unlikely(RTE_VHOST_NEED_LOG(features)))
+ *rte_vhost_log_write(vid, addr, len);
+ *
+ * @param vid
+ *  vhost device ID
+ * @param addr
+ *  the starting address for write
+ * @param len
+ *  the length to write
+ */
+void rte_vhost_log_write(int vid, uint64_t addr, uint64_t len);
+
+/**
+ * Log the used ring update start at given offset.
+ *
+ * Same as rte_vhost_log_write, it's suggested to do a check before
+ * calling it:
+ *
+ *if (unlikely(RTE_VHOST_NEED_LOG(features)))
+ *rte_vhost_log_used_vring(vid, vring_idx, offset, len);
+ *
+ * @param vid
+ *  vhost device ID
+ * @param vring_idx
+ *  the vring index
+ * @param offset
+ *  the offset inside the used ring
+ * @param len
+ *  the length to write
+ */
+void rte_vhost_log_used_vring(int vid, uint16_t vring_idx,
+ uint64_t offset, uint64_t len);
+
 int rte_vhost_enable_guest_notification(int vid, uint16_t queue_id, int 
enable);
 
 /**
diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c
index 74ae3b2..8be5b6a 100644
--- a/lib/librte_vhost/vhost.c
+++ b/lib/librte_vhost/vhost.c
@@ -443,3 +443,34 @@ struct virtio_net *
dev->virtqueue[queue_id]->used->flags = VRING_USED_F_NO_NOTIFY;
return 0;
 }
+
+void
+rte_vhost_log_write(int vid, uint64_t addr, uint64_t len)
+{
+   struct virtio_net *dev = get_device(vid);
+
+   if (dev == NULL)
+   return;
+
+   vhost_log_write(dev, addr, len);
+}
+
+void
+rte_vhost_log_used_vring(int vid, uint16_t vring_idx,
+uint64_t offset, uint64_t len)
+{
+   struct virtio_net *dev;
+   struct vhost_virtqueue *vq;
+
+   dev = get_device(vid);
+   if (dev == NULL)
+   return;
+
+   if (vring_idx >= VHOST_MAX_VRING)
+   return;
+   vq = dev->virtqueue[vring_idx];
+   if (!vq)
+   return;
+
+   vhost_log_used_vring(dev, vq, offset, len);
+}
diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
index 225ff2e..a199ee6 100644
--- a/lib/librte_vhost/vhost.h
+++ b/lib/librte_vhost/vhost.h
@@ -198,6 +198,44 @@ struct virtio_net {
struct guest_page   *guest_pages;
 } __rte_cache_aligned;
 
+
+#define VHOST_LOG_PAGE 4096
+
+static inline void __attribute__((always_inline))
+vhost_log_page(uint8_t *log_base, uint64_t page)
+{
+   log_base[page / 8] |= 1 << (page % 8);
+}
+
+static inline void __attribute__((always_inline))
+vhost_log_write(struct virtio_net *dev, uint64_t addr, uint64_t len)
+{
+   uint64_t page;
+
+   if (likely(((dev->features & (1ULL << VHOST_F_LOG_ALL)) == 0) ||
+  !dev->log_base || !len))
+   return;
+
+   if (unlikely(dev->log_size <= ((addr + len - 1) / VHOST_LOG_PAGE / 8)))
+   return;
+
+   /* To make sure guest memory updates are committed before logging */
+   rte_smp_wmb();
+
+   page = addr / VHOST_LOG_PAGE;
+   while (page * VHOST_LOG_PAGE < addr + len) {
+   vhost_log_page((uint8_t *)(uintptr_t)dev->log_base, page);
+   page += 1;
+   }
+}
+
+static inline void __attribute__((always_inline))
+vhost_log_used_

[dpdk-dev] [PATCH v2 18/22] vhost: introduce API to start a specific driver

[Yuanhan Liu]

We used to use rte_vhost_driver_session_start() to trigger the vhost-user
session. It takes no argument, thus it's a global trigger. And it could
be problematic.

The issue is, currently, rte_vhost_driver_register(path, flags) actually
tries to put it into the session loop (by fdset_add). However, it needs
a set of APIs to set a vhost-user driver properly:
  * rte_vhost_driver_register(path, flags);
  * rte_vhost_driver_set_features(path, features);
  * rte_vhost_driver_callback_register(path, vhost_device_ops);

If a new vhost-user driver is registered after the trigger (think OVS-DPDK
that could add a port dynamically from cmdline), the current code will
effectively starts the session for the new driver just after the first
API rte_vhost_driver_register() is invoked, leaving later calls taking
no effect at all.

To handle the case properly, this patch introduce a new API,
rte_vhost_driver_start(path), to trigger a specific vhost-user driver.
To do that, the rte_vhost_driver_register(path, flags) is simplified
to create the socket only and let rte_vhost_driver_start(path) to
actually put it into the session loop.

Meanwhile, the rte_vhost_driver_session_start is removed: we could hide
the session thread internally (create the thread if it has not been
created). This would also simplify the application.

NOTE: the API order in prog guide is slightly adjusted for showing the
correct invoke order.

Signed-off-by: Yuanhan Liu 
---
 doc/guides/prog_guide/vhost_lib.rst| 24 +--
 doc/guides/rel_notes/release_17_05.rst |  8 
 drivers/net/vhost/rte_eth_vhost.c  | 50 ++---
 examples/tep_termination/main.c|  8 +++-
 examples/vhost/main.c  |  9 +++-
 lib/librte_vhost/fd_man.c  |  9 ++--
 lib/librte_vhost/fd_man.h  |  2 +-
 lib/librte_vhost/rte_vhost_version.map |  2 +-
 lib/librte_vhost/rte_virtio_net.h  | 15 ++-
 lib/librte_vhost/socket.c  | 79 +++---
 10 files changed, 104 insertions(+), 102 deletions(-)

diff --git a/doc/guides/prog_guide/vhost_lib.rst 
b/doc/guides/prog_guide/vhost_lib.rst
index a4fb1f1..5979290 100644
--- a/doc/guides/prog_guide/vhost_lib.rst
+++ b/doc/guides/prog_guide/vhost_lib.rst
@@ -116,12 +116,6 @@ The following is an overview of some key Vhost API 
functions:
   vhost-user driver could be vhost-user net, yet it could be something else,
   say, vhost-user SCSI.
 
-* ``rte_vhost_driver_session_start()``
-
-  This function starts the vhost session loop to handle vhost messages. It
-  starts an infinite loop, therefore it should be called in a dedicated
-  thread.
-
 * ``rte_vhost_driver_callback_register(path, vhost_device_ops)``
 
   This function registers a set of callbacks, to let DPDK applications take
@@ -149,6 +143,17 @@ The following is an overview of some key Vhost API 
functions:
 ``VHOST_F_LOG_ALL`` will be set/cleared at the start/end of live
 migration, respectively.
 
+* ``rte_vhost_driver_disable/enable_features(path, features))``
+
+  This function disables/enables some features. For example, it can be used to
+  disable mergeable buffers and TSO features, which both are enabled by
+  default.
+
+* ``rte_vhost_driver_start(path)``
+
+  This function triggers the vhost-user negotiation. It should be invoked at
+  the end of initializing a vhost-user driver.
+
 * ``rte_vhost_enqueue_burst(vid, queue_id, pkts, count)``
 
   Transmits (enqueues) ``count`` packets from host to guest.
@@ -157,13 +162,6 @@ The following is an overview of some key Vhost API 
functions:
 
   Receives (dequeues) ``count`` packets from guest, and stored them at 
``pkts``.
 
-* ``rte_vhost_driver_disable/enable_features(path, features))``
-
-  This function disables/enables some features. For example, it can be used to
-  disable mergeable buffers and TSO features, which both are enabled by
-  default.
-
-
 Vhost-user Implementations
 --
 
diff --git a/doc/guides/rel_notes/release_17_05.rst 
b/doc/guides/rel_notes/release_17_05.rst
index 2efe292..8f06fc4 100644
--- a/doc/guides/rel_notes/release_17_05.rst
+++ b/doc/guides/rel_notes/release_17_05.rst
@@ -57,6 +57,11 @@ New Features
   * Enable Vhost PMD's MTU get feature.
   * Get max MTU value from host in Virtio PMD
 
+* **Made the vhost lib be a generic vhost-user lib.**
+
+  Now it could be used to implement any other vhost-user drivers, such
+  as, vhost-user SCSI.
+
 
 Resolved Issues
 ---
@@ -157,6 +162,9 @@ API Changes
* The vhost struct ``virtio_net_device_ops`` is renamed to
  ``vhost_device_ops``
 
+   * The vhost API ``rte_vhost_driver_session_start`` is removed. Instead,
+ ``rte_vhost_driver_start`` should be used.
+
 
 ABI Changes
 ---
diff --git a/drivers/net/vhost/rte_eth_vhost.c 
b/drivers/net/vhost/rte_eth_vhost.c
index 97a765f..e6c0758 100644
--- a/drivers/net/vhost/rte_eth_vhost.c
+++ b/drivers/net/vhost/rte_eth_vhost.c
@@ -127,9 +127,6 @@ stru

[dpdk-dev] [PATCH v2 21/22] vhost: do not destroy device on repeat mem table message

[Yuanhan Liu]

It doesn't make any sense to invoke destroy_device() callback at
while handling SET_MEM_TABLE message.

>From the vhost-user spec, it's the GET_VRING_BASE message indicates
the end of a vhost device: the destroy_device() should be invoked
from there (luckily, we already did that).

Signed-off-by: Yuanhan Liu 
---
 lib/librte_vhost/vhost_user.c | 6 --
 1 file changed, 6 deletions(-)

diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
index 40cc973..079c55e 100644
--- a/lib/librte_vhost/vhost_user.c
+++ b/lib/librte_vhost/vhost_user.c
@@ -496,12 +496,6 @@
uint32_t i;
int fd;
 
-   /* Remove from the data plane. */
-   if (dev->flags & VIRTIO_DEV_RUNNING) {
-   dev->flags &= ~VIRTIO_DEV_RUNNING;
-   dev->notify_ops->destroy_device(dev->vid);
-   }
-
if (dev->mem) {
free_mem_region(dev);
rte_free(dev->mem);
-- 
1.9.0

[dpdk-dev] [PATCH v2 22/22] examples/vhost: demonstrate the new generic vhost APIs

[Yuanhan Liu]

Now DPDK vhost lib has been generic enough, that it can be used to
implement any vhost-user drivers.

For example, this patch implements a very simple vhost-user net driver,
mainly for demonstrating how to use those generic vhost APIs.

And when the --builtin-net-driver option is used, the example virtio-net
driver code will be invoked, instead of the one provided from the vhost
library.

Signed-off-by: Yuanhan Liu 
---
 examples/vhost/Makefile |   2 +-
 examples/vhost/main.c   |  37 +++-
 examples/vhost/main.h   |  29 +++-
 examples/vhost/virtio_net.c | 405 
 4 files changed, 466 insertions(+), 7 deletions(-)
 create mode 100644 examples/vhost/virtio_net.c

diff --git a/examples/vhost/Makefile b/examples/vhost/Makefile
index e95c68a..af7be99 100644
--- a/examples/vhost/Makefile
+++ b/examples/vhost/Makefile
@@ -48,7 +48,7 @@ else
 APP = vhost-switch
 
 # all source are stored in SRCS-y
-SRCS-y := main.c
+SRCS-y := main.c virtio_net.c
 
 CFLAGS += -O2 -D_FILE_OFFSET_BITS=64
 CFLAGS += $(WERROR_FLAGS)
diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index 08b82f6..022438e 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -1,7 +1,7 @@
 /*-
  *   BSD LICENSE
  *
- *   Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2010-2017 Intel Corporation. All rights reserved.
  *   All rights reserved.
  *
  *   Redistribution and use in source and binary forms, with or without
@@ -65,7 +65,6 @@
 #define MBUF_CACHE_SIZE128
 #define MBUF_DATA_SIZE RTE_MBUF_DEFAULT_BUF_SIZE
 
-#define MAX_PKT_BURST 32   /* Max burst size for RX/TX */
 #define BURST_TX_DRAIN_US 100  /* TX drain every ~100us */
 
 #define BURST_RX_WAIT_US 15/* Defines how long we wait between retries on 
RX */
@@ -129,6 +128,8 @@
 static int client_mode;
 static int dequeue_zero_copy;
 
+static int builtin_net_driver;
+
 /* Specify timeout (in useconds) between retries on RX. */
 static uint32_t burst_rx_delay_time = BURST_RX_WAIT_US;
 /* Specify the number of retries on RX. */
@@ -499,6 +500,7 @@ struct mbuf_table {
{"tso", required_argument, NULL, 0},
{"client", no_argument, &client_mode, 1},
{"dequeue-zero-copy", no_argument, &dequeue_zero_copy, 1},
+   {"builtin-net-driver", no_argument, &builtin_net_driver, 1},
{NULL, 0, 0, 0},
};
 
@@ -795,7 +797,12 @@ static inline void __attribute__((always_inline))
 {
uint16_t ret;
 
-   ret = rte_vhost_enqueue_burst(dst_vdev->vid, VIRTIO_RXQ, &m, 1);
+   if (builtin_net_driver) {
+   ret = vs_enqueue_pkts(dst_vdev, VIRTIO_RXQ, &m, 1);
+   } else {
+   ret = rte_vhost_enqueue_burst(dst_vdev->vid, VIRTIO_RXQ, &m, 1);
+   }
+
if (enable_stats) {
rte_atomic64_inc(&dst_vdev->stats.rx_total_atomic);
rte_atomic64_add(&dst_vdev->stats.rx_atomic, ret);
@@ -1066,8 +1073,13 @@ static inline void __attribute__((always_inline))
}
}
 
-   enqueue_count = rte_vhost_enqueue_burst(vdev->vid, VIRTIO_RXQ,
+   if (builtin_net_driver) {
+   enqueue_count = vs_enqueue_pkts(vdev, VIRTIO_RXQ,
pkts, rx_count);
+   } else {
+   enqueue_count = rte_vhost_enqueue_burst(vdev->vid, VIRTIO_RXQ,
+   pkts, rx_count);
+   }
if (enable_stats) {
rte_atomic64_add(&vdev->stats.rx_total_atomic, rx_count);
rte_atomic64_add(&vdev->stats.rx_atomic, enqueue_count);
@@ -1083,8 +1095,13 @@ static inline void __attribute__((always_inline))
uint16_t count;
uint16_t i;
 
-   count = rte_vhost_dequeue_burst(vdev->vid, VIRTIO_TXQ, mbuf_pool,
+   if (builtin_net_driver) {
+   count = vs_dequeue_pkts(vdev, VIRTIO_TXQ, mbuf_pool,
pkts, MAX_PKT_BURST);
+   } else {
+   count = rte_vhost_dequeue_burst(vdev->vid, VIRTIO_TXQ,
+   mbuf_pool, pkts, MAX_PKT_BURST);
+   }
 
/* setup VMDq for the first packet */
if (unlikely(vdev->ready == DEVICE_MAC_LEARNING) && count) {
@@ -1187,6 +1204,9 @@ static inline void __attribute__((always_inline))
rte_pause();
}
 
+   if (builtin_net_driver)
+   vs_vhost_net_remove(vdev);
+
TAILQ_REMOVE(&lcore_info[vdev->coreid].vdev_list, vdev,
 lcore_vdev_entry);
TAILQ_REMOVE(&vhost_dev_list, vdev, global_vdev_entry);
@@ -1235,6 +1255,9 @@ static inline void __attribute__((always_inline))
}
vdev->vid = vid;
 
+   if (builtin_net_driver)
+   vs_vhost_net_setup(vdev);
+
TAILQ_INSERT_TAIL(&vhost_dev_list, vdev, global_vdev_entry);
vdev->vmdq_rx_q = vid * queues_per_pool + vmdq_queue_

[dpdk-dev] [PATCH v2 1/2] cryptodev: add api for attach-detach session with queue pair

[akhil.goyal]

From: Akhil Goyal 

HW based crypto drivers may only support limited number of
sessions per queue pair. This requires support for attaching
sessions to specific queue pair.  New APIs  are introduced to
attach/detach a session with/from a particular queue pair.
These are optional APIs.

Application can call attach API after creating a session
and can call detach API before deleting a session.

Application needs to check if max_nb_sessions_per_qp > 0,
then it should call the attach API.

max_nb_sessions_per_qp = 0 means infinite sessions per qp

Signed-off-by: Akhil Goyal 
---
 lib/librte_cryptodev/rte_cryptodev.c   | 47 ++
 lib/librte_cryptodev/rte_cryptodev.h   | 34 +++
 lib/librte_cryptodev/rte_cryptodev_pmd.h   | 29 
 lib/librte_cryptodev/rte_cryptodev_version.map |  2 ++
 4 files changed, 112 insertions(+)

diff --git a/lib/librte_cryptodev/rte_cryptodev.c 
b/lib/librte_cryptodev/rte_cryptodev.c
index 0ac23ed..eb72f03 100644
--- a/lib/librte_cryptodev/rte_cryptodev.c
+++ b/lib/librte_cryptodev/rte_cryptodev.c
@@ -1393,6 +1393,53 @@ rte_cryptodev_sym_session_create(uint8_t dev_id,
return sess;
 }
 
+int
+rte_cryptodev_queue_pair_attach_sym_session(uint16_t qp_id,
+   struct rte_cryptodev_sym_session *sess)
+{
+   struct rte_cryptodev *dev;
+
+   if (!rte_cryptodev_pmd_is_valid_dev(sess->dev_id)) {
+   CDEV_LOG_ERR("Invalid dev_id=%d", sess->dev_id);
+   return -EINVAL;
+   }
+
+   dev = &rte_crypto_devices[sess->dev_id];
+
+   /* The API is optional, not returning error if driver do not suuport */
+   RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->qp_attach_session, 0);
+   if (dev->dev_ops->qp_attach_session(dev, qp_id, sess->_private)) {
+   CDEV_LOG_ERR("dev_id %d failed to attach qp: %d with session",
+   sess->dev_id, qp_id);
+   return -EPERM;
+   }
+
+   return 0;
+}
+
+int
+rte_cryptodev_queue_pair_detach_sym_session(uint16_t qp_id,
+   struct rte_cryptodev_sym_session *sess)
+{
+   struct rte_cryptodev *dev;
+
+   if (!rte_cryptodev_pmd_is_valid_dev(sess->dev_id)) {
+   CDEV_LOG_ERR("Invalid dev_id=%d", sess->dev_id);
+   return -EINVAL;
+   }
+
+   dev = &rte_crypto_devices[sess->dev_id];
+
+   /* The API is optional, not returning error if driver do not suuport */
+   RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->qp_detach_session, 0);
+   if (dev->dev_ops->qp_detach_session(dev, qp_id, sess->_private)) {
+   CDEV_LOG_ERR("dev_id %d failed to detach qp: %d from session",
+   sess->dev_id, qp_id);
+   return -EPERM;
+   }
+
+   return 0;
+}
 struct rte_cryptodev_sym_session *
 rte_cryptodev_sym_session_free(uint8_t dev_id,
struct rte_cryptodev_sym_session *sess)
diff --git a/lib/librte_cryptodev/rte_cryptodev.h 
b/lib/librte_cryptodev/rte_cryptodev.h
index d61a43e..05e5b4e 100644
--- a/lib/librte_cryptodev/rte_cryptodev.h
+++ b/lib/librte_cryptodev/rte_cryptodev.h
@@ -332,6 +332,10 @@ struct rte_cryptodev_info {
struct {
unsigned max_nb_sessions;
/**< Maximum number of sessions supported by device. */
+   unsigned max_nb_sessions_per_qp;
+   /**< Maximum number of sessions per queue pair.
+* Default 0 for infinite sessions
+*/
} sym;
 };
 
@@ -915,6 +919,36 @@ extern struct rte_cryptodev_sym_session *
 rte_cryptodev_sym_session_free(uint8_t dev_id,
struct rte_cryptodev_sym_session *session);
 
+/**
+ * Attach queue pair with sym session.
+ *
+ * @param  qp_id   Queue pair to which session will be attached.
+ * @param  session Session pointer previously allocated by
+ * *rte_cryptodev_sym_session_create*.
+ *
+ * @return
+ *  - On success, zero.
+ *  - On failure, a negative value.
+ */
+extern int
+rte_cryptodev_queue_pair_attach_sym_session(uint16_t qp_id,
+   struct rte_cryptodev_sym_session *session);
+
+/**
+ * Detach queue pair with sym session.
+ *
+ * @param  qp_id   Queue pair to which session is attached.
+ * @param  session Session pointer previously allocated by
+ * *rte_cryptodev_sym_session_create*.
+ *
+ * @return
+ *  - On success, zero.
+ *  - On failure, a negative value.
+ */
+extern int
+rte_cryptodev_queue_pair_detach_sym_session(uint16_t qp_id,
+   struct rte_cryptodev_sym_session *session);
+
 
 #ifdef __cplusplus
 }
diff --git a/lib/librte_cryptodev/rte_cryptodev_pmd.h 
b/lib/librte_cryptodev/rte_cryptodev_pmd.h
index 1a417e2..df92817 100644
--- a/lib/librte_cryptodev/rte_cryptodev_pmd.h
+++ b/lib/librte_cryptodev/rte_cryptodev_pmd.h
@@ -381,6 +381,31 @@ typedef void * (*cryptodev_sym_configure_session

[dpdk-dev] [PATCH v2 2/2] examples/ipsec-secgw: attach session-qp

[akhil.goyal]

From: Akhil Goyal 

adding support for attaching session to queue pairs.
This is required as underlying crypto driver may only
support limited number of sessions per queue pair
if max_nb_sessions_per_qp > 0, session should be
attached to a particular qp.

Signed-off-by: Akhil Goyal 
---
 examples/ipsec-secgw/ipsec.c | 12 
 1 file changed, 12 insertions(+)

diff --git a/examples/ipsec-secgw/ipsec.c b/examples/ipsec-secgw/ipsec.c
index 144f0aa..b35b30f 100644
--- a/examples/ipsec-secgw/ipsec.c
+++ b/examples/ipsec-secgw/ipsec.c
@@ -47,6 +47,7 @@
 static inline int
 create_session(struct ipsec_ctx *ipsec_ctx __rte_unused, struct ipsec_sa *sa)
 {
+   struct rte_cryptodev_info cdev_info;
unsigned long cdev_id_qp = 0;
int32_t ret;
struct cdev_key key = { 0 };
@@ -73,6 +74,17 @@ create_session(struct ipsec_ctx *ipsec_ctx __rte_unused, 
struct ipsec_sa *sa)
sa->crypto_session = rte_cryptodev_sym_session_create(
ipsec_ctx->tbl[cdev_id_qp].id, sa->xforms);
 
+   rte_cryptodev_info_get(ipsec_ctx->tbl[cdev_id_qp].id, &cdev_info);
+   if (cdev_info.sym.max_nb_sessions_per_qp > 0) {
+   ret = rte_cryptodev_queue_pair_attach_sym_session(
+   ipsec_ctx->tbl[cdev_id_qp].qp,
+   sa->crypto_session);
+   if (ret < 0) {
+   RTE_LOG(ERR, IPSEC, "Session cannot be attached"
+   " to qp %u ", ipsec_ctx->tbl[cdev_id_qp].qp);
+   return -1;
+   }
+   }
sa->cdev_id_qp = cdev_id_qp;
 
return 0;
-- 
2.9.3

[dpdk-dev] [PATCH v7 1/4] net/tap: move private elements to external header

[Pascal Mazon]

In the next patch, access to struct pmd_internals will be necessary in
tap_flow.c to store the flows.

Signed-off-by: Pascal Mazon 
Acked-by: Olga Shern 
Acked-by: Keith Wiles 
---
 drivers/net/tap/Makefile  |  1 +
 drivers/net/tap/rte_eth_tap.c | 36 ++---
 drivers/net/tap/rte_eth_tap.h | 75 +++
 3 files changed, 78 insertions(+), 34 deletions(-)
 create mode 100644 drivers/net/tap/rte_eth_tap.h

diff --git a/drivers/net/tap/Makefile b/drivers/net/tap/Makefile
index ddf87232d335..fa4658bd1e75 100644
--- a/drivers/net/tap/Makefile
+++ b/drivers/net/tap/Makefile
@@ -40,6 +40,7 @@ EXPORT_MAP := rte_pmd_tap_version.map
 LIBABIVER := 1
 
 CFLAGS += -O3
+CFLAGS += -I$(SRCDIR)
 CFLAGS += $(WERROR_FLAGS)
 
 #
diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index f8d9cc7dc3b2..6bb63e5ec873 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c
@@ -55,6 +55,8 @@
 #include 
 #include 
 
+#include 
+
 /* Linux based path to the TUN device */
 #define TUN_TAP_DEV_PATH"/dev/net/tun"
 #define DEFAULT_TAP_NAME"dtap"
@@ -87,40 +89,6 @@ static struct rte_eth_link pmd_link = {
.link_autoneg = ETH_LINK_SPEED_AUTONEG
 };
 
-struct pkt_stats {
-   uint64_t opackets;  /* Number of output packets */
-   uint64_t ipackets;  /* Number of input packets */
-   uint64_t obytes;/* Number of bytes on output */
-   uint64_t ibytes;/* Number of bytes on input */
-   uint64_t errs;  /* Number of error packets */
-};
-
-struct rx_queue {
-   struct rte_mempool *mp; /* Mempool for RX packets */
-   uint32_t trigger_seen;  /* Last seen Rx trigger value */
-   uint16_t in_port;   /* Port ID */
-   int fd;
-
-   struct pkt_stats stats; /* Stats for this RX queue */
-};
-
-struct tx_queue {
-   int fd;
-   struct pkt_stats stats; /* Stats for this TX queue */
-};
-
-struct pmd_internals {
-   char name[RTE_ETH_NAME_MAX_LEN];/* Internal Tap device name */
-   uint16_t nb_queues; /* Number of queues supported */
-   struct ether_addr eth_addr; /* Mac address of the device port */
-
-   int if_index;   /* IF_INDEX for the port */
-   int ioctl_sock; /* socket for ioctl calls */
-
-   struct rx_queue rxq[RTE_PMD_TAP_MAX_QUEUES];/* List of RX queues */
-   struct tx_queue txq[RTE_PMD_TAP_MAX_QUEUES];/* List of TX queues */
-};
-
 static void
 tap_trigger_cb(int sig __rte_unused)
 {
diff --git a/drivers/net/tap/rte_eth_tap.h b/drivers/net/tap/rte_eth_tap.h
new file mode 100644
index ..aafdef1faa99
--- /dev/null
+++ b/drivers/net/tap/rte_eth_tap.h
@@ -0,0 +1,75 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright 2017 6WIND S.A.
+ *   Copyright 2017 Mellanox.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of 6WIND S.A. nor the names of its
+ *   contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _RTE_ETH_TAP_H_
+#define _RTE_ETH_TAP_H_
+
+#include 
+
+#include 
+#include 
+
+#define RTE_PMD_TAP_MAX_QUEUES 16
+
+struct pkt_stats {
+   uint64_t opackets;  /* Number of output packets */
+   uint64_t ipackets;  /* Number of input packets */
+   uint64_t obytes;/* Number of bytes on output */
+   uint64_t ibytes;/* Number of bytes on input */
+   uint64_t 

[dpdk-dev] [PATCH v7 3/4] net/tap: add netlink back-end for flow API

[Pascal Mazon]

Each kernel netdevice may have queueing disciplines set for it, which
determine how to handle the packet (mostly on egress). That's part of
the TC (Traffic Control) mechanism.

Through TC, it is possible to set filter rules that match specific
packets, and act according to what is in the rule. This is a perfect
candidate to implement the flow API for the tap PMD, as it has an
associated kernel netdevice automatically.

Each flow API rule will be translated into its TC counterpart.

To leverage TC, it is necessary to communicate with the kernel using
netlink. This patch introduces a library to help that communication.

Inside netlink.c, functions are generic for any netlink messaging.
Inside tcmsgs.c, functions are specific to deal with TC rules.

Signed-off-by: Pascal Mazon 
Acked-by: Olga Shern 
Acked-by: Keith Wiles 
---
 drivers/net/tap/Makefile  |   2 +
 drivers/net/tap/tap_netlink.c | 367 
 drivers/net/tap/tap_netlink.h |  69 
 drivers/net/tap/tap_tcmsgs.c  | 378 ++
 drivers/net/tap/tap_tcmsgs.h  |  63 +++
 5 files changed, 879 insertions(+)
 create mode 100644 drivers/net/tap/tap_netlink.c
 create mode 100644 drivers/net/tap/tap_netlink.h
 create mode 100644 drivers/net/tap/tap_tcmsgs.c
 create mode 100644 drivers/net/tap/tap_tcmsgs.h

diff --git a/drivers/net/tap/Makefile b/drivers/net/tap/Makefile
index 45c67de8e970..3a33b560d3b5 100644
--- a/drivers/net/tap/Makefile
+++ b/drivers/net/tap/Makefile
@@ -48,6 +48,8 @@ CFLAGS += $(WERROR_FLAGS)
 #
 SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += rte_eth_tap.c
 SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += tap_flow.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += tap_netlink.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += tap_tcmsgs.c
 
 # this lib depends upon:
 DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += lib/librte_eal
diff --git a/drivers/net/tap/tap_netlink.c b/drivers/net/tap/tap_netlink.c
new file mode 100644
index ..9710e41a7801
--- /dev/null
+++ b/drivers/net/tap/tap_netlink.c
@@ -0,0 +1,367 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright 2017 6WIND S.A.
+ *   Copyright 2017 Mellanox.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of 6WIND S.A. nor the names of its
+ *   contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include 
+#include 
+#include 
+
+/* Must be quite large to support dumping a huge list of QDISC or filters. */
+#define BUF_SIZE (32 * 1024) /* Size of the buffer to receive kernel messages 
*/
+#define SNDBUF_SIZE 32768 /* Send buffer size for the netlink socket */
+#define RCVBUF_SIZE 32768 /* Receive buffer size for the netlink socket */
+
+struct nested_tail {
+   struct rtattr *tail;
+   struct nested_tail *prev;
+};
+
+/**
+ * Initialize a netlink socket for communicating with the kernel.
+ *
+ * @return
+ *   netlink socket file descriptor on success, -1 otherwise.
+ */
+int
+nl_init(void)
+{
+   int fd, sndbuf_size = SNDBUF_SIZE, rcvbuf_size = RCVBUF_SIZE;
+   struct sockaddr_nl local = { .nl_family = AF_NETLINK };
+
+   fd = socket(AF_NETLINK, SOCK_RAW | SOCK_CLOEXEC, NETLINK_ROUTE);
+   if (fd < 0) {
+   RTE_LOG(ERR, PMD, "Unable to create a netlink socket\n");
+   return -1;
+   }
+   if (setsockopt(fd, SOL_SOCKET, SO_SNDBUF, &sndbuf_size, sizeof(int))) {
+   RTE_LOG(ERR, PMD, "Unable to set socket buffer send size\n");
+   return -1;
+   }
+   if (setsockopt(fd, SOL_S

[dpdk-dev] [PATCH v7 0/4] net/tap: support flow API

[Pascal Mazon]

This series adds support for the flow API in tap PMD.

It enables filtering specific packets incoming on the tap netdevice, to
process only desired ones. Under the hood, it uses kernel TC (traffic
control), which takes place very early in the stack, and supports most
common pattern items and actions defined in the flow API.

v7 changes:
   - provide more details in doc/guides/nics/tap.rst

v6 changes:
  - fix compilation issue on i686 (wrong cast for rte flow handle)

v5 changes:
  - rebase after adrien's patches on Tx poll and Rx signaling
  - better spaces for comments in rte_eth_tap.h

v4 changes:
  - rebase on top of "net/tap: add additional management ops" series
  - fix a few netlink doxygen comments
  - rename tap.h -> rte_eth_tap.h
  - flush flow rules only when applicable

v3 changes:
  - vlan patterns enabled depending on running kernel (4.9+)
  - update doc/guides/nics/tap.rst for Flow API support
  - rebase on top of "net/tap: add additional management ops" series

v2 changes:
  - support compilation on kernels < 4.2 (where flower support appeared)
  - set whitespaces in tap.h
  - remove unnecessary goto

Pascal Mazon (4):
  net/tap: move private elements to external header
  net/tap: add preliminary support for rte_flow
  net/tap: add netlink back-end for flow API
  net/tap: add basic flow API patterns and actions

 doc/guides/nics/features/tap.ini |1 +
 doc/guides/nics/tap.rst  |   45 ++
 drivers/net/tap/Makefile |   44 ++
 drivers/net/tap/rte_eth_tap.c|  101 ++--
 drivers/net/tap/rte_eth_tap.h|   80 +++
 drivers/net/tap/tap_flow.c   | 1070 ++
 drivers/net/tap/tap_flow.h   |   58 +++
 drivers/net/tap/tap_netlink.c|  367 +
 drivers/net/tap/tap_netlink.h|   69 +++
 drivers/net/tap/tap_tcmsgs.c |  378 ++
 drivers/net/tap/tap_tcmsgs.h |   63 +++
 11 files changed, 2241 insertions(+), 35 deletions(-)
 create mode 100644 drivers/net/tap/rte_eth_tap.h
 create mode 100644 drivers/net/tap/tap_flow.c
 create mode 100644 drivers/net/tap/tap_flow.h
 create mode 100644 drivers/net/tap/tap_netlink.c
 create mode 100644 drivers/net/tap/tap_netlink.h
 create mode 100644 drivers/net/tap/tap_tcmsgs.c
 create mode 100644 drivers/net/tap/tap_tcmsgs.h

-- 
2.12.0.306.g4a9b9b3

[dpdk-dev] [PATCH v7 2/4] net/tap: add preliminary support for rte_flow

[Pascal Mazon]

The flow API provides the ability to classify packets received by a tap
netdevice.

This patch only implements skeleton functions for flow API support, no
patterns are supported yet.

Signed-off-by: Pascal Mazon 
Acked-by: Olga Shern 
Acked-by: Keith Wiles 
---
 doc/guides/nics/features/tap.ini |   1 +
 drivers/net/tap/Makefile |   1 +
 drivers/net/tap/rte_eth_tap.c|   6 ++
 drivers/net/tap/rte_eth_tap.h|   2 +
 drivers/net/tap/tap_flow.c   | 185 +++
 drivers/net/tap/tap_flow.h   |  46 ++
 6 files changed, 241 insertions(+)
 create mode 100644 drivers/net/tap/tap_flow.c
 create mode 100644 drivers/net/tap/tap_flow.h

diff --git a/doc/guides/nics/features/tap.ini b/doc/guides/nics/features/tap.ini
index a51712dce066..9d73f61cca3b 100644
--- a/doc/guides/nics/features/tap.ini
+++ b/doc/guides/nics/features/tap.ini
@@ -9,6 +9,7 @@ Jumbo frame  = Y
 Promiscuous mode = Y
 Allmulticast mode= Y
 Basic stats  = Y
+Flow API = Y
 MTU update   = Y
 Multicast MAC filter = Y
 Speed capabilities   = Y
diff --git a/drivers/net/tap/Makefile b/drivers/net/tap/Makefile
index fa4658bd1e75..45c67de8e970 100644
--- a/drivers/net/tap/Makefile
+++ b/drivers/net/tap/Makefile
@@ -47,6 +47,7 @@ CFLAGS += $(WERROR_FLAGS)
 # all source are stored in SRCS-y
 #
 SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += rte_eth_tap.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += tap_flow.c
 
 # this lib depends upon:
 DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += lib/librte_eal
diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index 6bb63e5ec873..9127c739a214 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c
@@ -56,6 +56,7 @@
 #include 
 
 #include 
+#include 
 
 /* Linux based path to the TUN device */
 #define TUN_TAP_DEV_PATH"/dev/net/tun"
@@ -482,6 +483,7 @@ tap_dev_close(struct rte_eth_dev *dev __rte_unused)
struct pmd_internals *internals = dev->data->dev_private;
 
tap_link_set_down(dev);
+   tap_flow_flush(dev, NULL);
 
for (i = 0; i < internals->nb_queues; i++) {
if (internals->rxq[i].fd != -1)
@@ -806,6 +808,7 @@ static const struct eth_dev_ops ops = {
.stats_get  = tap_stats_get,
.stats_reset= tap_stats_reset,
.dev_supported_ptypes_get = tap_dev_supported_ptypes_get,
+   .filter_ctrl= tap_dev_filter_ctrl,
 };
 
 static int
@@ -877,6 +880,8 @@ eth_dev_tap_create(const char *name, char *tap_name)
pmd->txq[i].fd = -1;
}
 
+   LIST_INIT(&pmd->flows);
+
return 0;
 
 error_exit:
@@ -990,6 +995,7 @@ rte_pmd_tap_remove(const char *name)
return 0;
 
internals = eth_dev->data->dev_private;
+   tap_flow_flush(eth_dev, NULL);
for (i = 0; i < internals->nb_queues; i++)
if (internals->rxq[i].fd != -1)
close(internals->rxq[i].fd);
diff --git a/drivers/net/tap/rte_eth_tap.h b/drivers/net/tap/rte_eth_tap.h
index aafdef1faa99..bf8226736627 100644
--- a/drivers/net/tap/rte_eth_tap.h
+++ b/drivers/net/tap/rte_eth_tap.h
@@ -34,6 +34,7 @@
 #ifndef _RTE_ETH_TAP_H_
 #define _RTE_ETH_TAP_H_
 
+#include 
 #include 
 
 #include 
@@ -68,6 +69,7 @@ struct pmd_internals {
struct ether_addr eth_addr;   /* Mac address of the device port */
int if_index; /* IF_INDEX for the port */
int ioctl_sock;   /* socket for ioctl calls */
+   LIST_HEAD(tap_flows, rte_flow) flows;/* rte_flow rules */
struct rx_queue rxq[RTE_PMD_TAP_MAX_QUEUES]; /* List of RX queues */
struct tx_queue txq[RTE_PMD_TAP_MAX_QUEUES]; /* List of TX queues */
 };
diff --git a/drivers/net/tap/tap_flow.c b/drivers/net/tap/tap_flow.c
new file mode 100644
index ..c32ed382d745
--- /dev/null
+++ b/drivers/net/tap/tap_flow.c
@@ -0,0 +1,185 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright 2017 6WIND S.A.
+ *   Copyright 2017 Mellanox.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of 6WIND S.A. nor the names of its
+ *   contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, 

[dpdk-dev] [PATCH v7 4/4] net/tap: add basic flow API patterns and actions

[Pascal Mazon]

Supported flow rules are now mapped to TC rules on the tap netdevice.
The netlink message used for creating the TC rule is stored in struct
rte_flow. That way, by simply changing a metadata in it, we can require
for the rule deletion without further parsing.

Supported items:
- eth: src and dst (with variable masks), and eth_type (0x mask).
- vlan: vid, pcp, tpid, but not eid.
- ipv4/6: src and dst (with variable masks), and ip_proto (0x mask).
- udp/tcp: src and dst port (0x) mask.

Supported actions:
- DROP
- QUEUE
- PASSTHRU

It is generally not possible to provide a "last" item. However, if the
"last" item, once masked, is identical to the masked spec, then it is
supported.

Only IPv4/6 and MAC addresses can use a variable mask. All other
items need a full mask (exact match).

Support for VLAN requires kernel headers >= 4.9, checked using
auto-config.sh.

Signed-off-by: Pascal Mazon 
Acked-by: Olga Shern 
Acked-by: Keith Wiles 
---
 doc/guides/nics/tap.rst   |  45 +++
 drivers/net/tap/Makefile  |  40 ++
 drivers/net/tap/rte_eth_tap.c |  61 ++-
 drivers/net/tap/rte_eth_tap.h |   3 +
 drivers/net/tap/tap_flow.c| 911 +-
 drivers/net/tap/tap_flow.h|  12 +
 6 files changed, 1057 insertions(+), 15 deletions(-)

diff --git a/doc/guides/nics/tap.rst b/doc/guides/nics/tap.rst
index c4f207be3b47..4986e47e9f57 100644
--- a/doc/guides/nics/tap.rst
+++ b/doc/guides/nics/tap.rst
@@ -82,6 +82,51 @@ can utilize that stack to handle the network protocols. Plus 
you would be able
 to address the interface using an IP address assigned to the internal
 interface.
 
+Flow API support
+
+
+The tap PMD supports major flow API pattern items and actions, when running on
+linux kernels above 4.2 ("Flower" classifier required). Supported items:
+
+- eth: src and dst (with variable masks), and eth_type (0x mask).
+- vlan: vid, pcp, tpid, but not eid. (requires kernel 4.9)
+- ipv4/6: src and dst (with variable masks), and ip_proto (0x mask).
+- udp/tcp: src and dst port (0x) mask.
+
+Supported actions:
+
+- DROP
+- QUEUE
+- PASSTHRU
+
+It is generally not possible to provide a "last" item. However, if the "last"
+item, once masked, is identical to the masked spec, then it is supported.
+
+Only IPv4/6 and MAC addresses can use a variable mask. All other items need a
+full mask (exact match).
+
+As rules are translated to TC, it is possible to show them with something 
like::
+
+   tc -s filter show dev tap1 parent 1:
+
+Examples of testpmd flow rules
+~~
+
+Drop packets for destination IP 192.168.0.1::
+
+   testpmd> flow create 0 priority 1 ingress pattern eth / ipv4 dst is 1.1.1.1 
\
+/ end actions drop / end
+
+Ensure packets from a given MAC address are received on a queue 2::
+
+   testpmd> flow create 0 priority 2 ingress pattern eth src is 
06:05:04:03:02:01 \
+/ end actions queue index 2 / end
+
+Drop UDP packets in vlan 3::
+
+   testpmd> flow create 0 priority 3 ingress pattern eth / vlan vid is 3 / \
+ipv4 proto is 17 / end actions drop / end
+
 Example
 ---
 
diff --git a/drivers/net/tap/Makefile b/drivers/net/tap/Makefile
index 3a33b560d3b5..c42a680555e9 100644
--- a/drivers/net/tap/Makefile
+++ b/drivers/net/tap/Makefile
@@ -41,6 +41,7 @@ LIBABIVER := 1
 
 CFLAGS += -O3
 CFLAGS += -I$(SRCDIR)
+CFLAGS += -I.
 CFLAGS += $(WERROR_FLAGS)
 
 #
@@ -58,5 +59,44 @@ DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += lib/librte_mempool
 DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += lib/librte_ether
 DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += lib/librte_kvargs
 DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += lib/librte_net
+DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += lib/librte_hash
 
 include $(RTE_SDK)/mk/rte.lib.mk
+
+# Generate and clean-up tap_autoconf.h.
+
+export CC CFLAGS CPPFLAGS EXTRA_CFLAGS EXTRA_CPPFLAGS
+export AUTO_CONFIG_CFLAGS = -Wno-error
+
+ifndef V
+AUTOCONF_OUTPUT := >/dev/null
+endif
+
+tap_autoconf.h.new: FORCE
+
+tap_autoconf.h.new: $(RTE_SDK)/buildtools/auto-config-h.sh
+   $Q $(RM) -f -- '$@'
+   $Q sh -- '$<' '$@' \
+   HAVE_TC_FLOWER \
+   linux/pkt_cls.h \
+   enum TCA_FLOWER_UNSPEC \
+   $(AUTOCONF_OUTPUT)
+   $Q sh -- '$<' '$@' \
+   HAVE_TC_VLAN_ID \
+   linux/pkt_cls.h \
+   enum TCA_FLOWER_KEY_VLAN_PRIO \
+   $(AUTOCONF_OUTPUT)
+
+# Create tap_autoconf.h or update it in case it differs from the new one.
+
+tap_autoconf.h: tap_autoconf.h.new
+   $Q [ -f '$@' ] && \
+   cmp '$<' '$@' $(AUTOCONF_OUTPUT) || \
+   mv '$<' '$@'
+
+$(SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP):.c=.o): tap_autoconf.h
+
+clean_tap: FORCE
+   $Q rm -f -- tap_autoconf.h tap_autoconf.h.new
+
+clean: clean_tap
diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index 9127c739a214..c711b36c3222 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/driv

[dpdk-dev] [PATCH v6 0/1] net/tap: remote netdevice traffic capture

[Pascal Mazon]

This patchset adds the special "remote" feature to the tap PMD, that
actually enables capturing traffic from another netdevice. This is
especially useful to get packets into the DPDK app, when the remote
netdevice has no DPDK support.

The "remote" feature requires flow API support as flow rules will be
configured on the remote netdevice for redirection, using the same
mechanism.

This series applies on top of:

  [PATCH 0/4] net/tap: support flow API

v6 changes:
  - update documentation with use case

v5 changes:
  - rebase after adrien's patches on Tx poll and Rx signaling

v4 changes:
  - rebase on top of updated "net/tap: support flow API"
  - use only a single patch now as MTU, MAC and flags can be easily managed
with tap_ioctl()

v3 changes:
  - memset(0) for remote_iface in rte_pmd_tap_probe()
  - use snprintf instead of strncpy to correctly handle terminating \0

v2 changes:
  - rebase on top of updated "net/tap: support flow API"
  - fix implicit flow flush when closing the netdevices

Pascal Mazon (1):
  net/tap: add remote netdevice traffic capture

 doc/guides/nics/tap.rst   |  16 ++
 drivers/net/tap/rte_eth_tap.c | 101 +-
 drivers/net/tap/rte_eth_tap.h |   4 +
 drivers/net/tap/tap_flow.c| 451 --
 drivers/net/tap/tap_flow.h|  24 +++
 5 files changed, 574 insertions(+), 22 deletions(-)

-- 
2.12.0.306.g4a9b9b3

Re: [dpdk-dev] [PATCH] doc: reformat crypto drivers overview

[Mcnamara, John]

> -Original Message-
> From: De Lara Guarch, Pablo
> Sent: Wednesday, March 22, 2017 4:15 PM
> To: Mcnamara, John 
> Cc: Trahe, Fiona ; dev@dpdk.org; De Lara Guarch,
> Pablo 
> Subject: [PATCH] doc: reformat crypto drivers overview
> 
> Follow the approach in the network devices overview, for the feature
> matrix, so it improves readibility and maintainability.
> 


There are pep8 warnings that should be fixed. Mainly for long lines:

$ pep8 doc/guides/conf.py   
doc/guides/conf.py:318:80: E501 line too long (109 > 79 characters)
doc/guides/conf.py:320:80: E501 line too long (105 > 79 characters)
doc/guides/conf.py:321:80: E501 line too long (81 > 79 characters)
doc/guides/conf.py:322:80: E501 line too long (108 > 79 characters)
doc/guides/conf.py:324:80: E501 line too long (122 > 79 characters)
doc/guides/conf.py:326:80: E501 line too long (102 > 79 characters)

Apart from that it is a good change and a nice generalization of the
table building approach.

With the above fixed:

Acked-by: John McNamara 

[dpdk-dev] [PATCH v6 1/1] net/tap: add remote netdevice traffic capture

[Pascal Mazon]

By default, a tap netdevice is of no use when not fed by a separate
process. The ability to automatically feed it from another netdevice
allows applications to capture any kind of traffic normally destined to
the kernel stack.

This patch implements this ability through a new optional "remote"
parameter.

Packets matching filtering rules created with the flow API are matched
on the remote device and redirected to the tap PMD, where the relevant
action will be performed.

Signed-off-by: Pascal Mazon 
Acked-by: Olga Shern 
Acked-by: Keith Wiles 
---
 doc/guides/nics/tap.rst   |  16 ++
 drivers/net/tap/rte_eth_tap.c | 101 +-
 drivers/net/tap/rte_eth_tap.h |   4 +
 drivers/net/tap/tap_flow.c| 451 --
 drivers/net/tap/tap_flow.h|  24 +++
 5 files changed, 574 insertions(+), 22 deletions(-)

diff --git a/doc/guides/nics/tap.rst b/doc/guides/nics/tap.rst
index 4986e47e9f57..5c5ba5357bba 100644
--- a/doc/guides/nics/tap.rst
+++ b/doc/guides/nics/tap.rst
@@ -58,6 +58,22 @@ needed, but the interface does not enforce that speed, for 
example::
 
--vdev=net_tap0,iface=foo0,speed=25000
 
+It is possible to specify a remote netdevice to capture packets from by adding
+``remote=foo1``, for example::
+
+   --vdev=net_tap,iface=tap0,remote=foo1
+
+If a ``remote`` is set, the tap MAC address will be set to match the remote one
+just after netdevice creation. Using TC rules, traffic from the remote 
netdevice
+will be redirected to the tap. If the tap is in promiscuous mode, then all
+packets will be redirected. In allmulti mode, all multicast packets will be
+redirected.
+
+Using the remote feature is especially useful for capturing traffic from a
+netdevice that has no support in the DPDK. It is possible to add explicit
+rte_flow rules on the tap PMD to capture specific traffic (see next section for
+examples).
+
 After the DPDK application is started you can send and receive packets on the
 interface using the standard rx_burst/tx_burst APIs in DPDK. From the host
 point of view you can use any host tool like tcpdump, Wireshark, ping, Pktgen
diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index c711b36c3222..69fa282ca176 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c
@@ -67,6 +67,7 @@
 
 #define ETH_TAP_IFACE_ARG   "iface"
 #define ETH_TAP_SPEED_ARG   "speed"
+#define ETH_TAP_REMOTE_ARG  "remote"
 
 #ifdef IFF_MULTI_QUEUE
 #define RTE_PMD_TAP_MAX_QUEUES 16
@@ -82,6 +83,7 @@ static struct rte_vdev_driver pmd_tap_drv;
 static const char *valid_arguments[] = {
ETH_TAP_IFACE_ARG,
ETH_TAP_SPEED_ARG,
+   ETH_TAP_REMOTE_ARG,
NULL
 };
 
@@ -237,10 +239,43 @@ tun_alloc(struct pmd_internals *pmd, uint16_t qid)
pmd->name);
return fd;
}
+   if (pmd->remote_if_index) {
+   /*
+* Flush usually returns negative value because it tries
+* to delete every QDISC (and on a running device, one
+* QDISC at least is needed). Ignore negative return
+* value.
+*/
+   qdisc_flush(pmd->nlsk_fd, pmd->remote_if_index);
+   if (qdisc_create_ingress(pmd->nlsk_fd,
+pmd->remote_if_index) < 0)
+   goto remote_fail;
+   LIST_INIT(&pmd->implicit_flows);
+   if (tap_flow_implicit_create(
+   pmd, TAP_REMOTE_LOCAL_MAC) < 0)
+   goto remote_fail;
+   if (tap_flow_implicit_create(
+   pmd, TAP_REMOTE_BROADCAST) < 0)
+   goto remote_fail;
+   if (tap_flow_implicit_create(
+   pmd, TAP_REMOTE_BROADCASTV6) < 0)
+   goto remote_fail;
+   if (tap_flow_implicit_create(
+   pmd, TAP_REMOTE_TX) < 0)
+   goto remote_fail;
+   }
}
 
return fd;
 
+remote_fail:
+   RTE_LOG(ERR, PMD,
+   "Could not set up remote flow rules for %s: remote disabled.\n",
+   pmd->name);
+   pmd->remote_if_index = 0;
+   tap_flow_implicit_flush(pmd, NULL);
+   return fd;
+
 error:
if (fd > 0)
close(fd);
@@ -336,8 +371,17 @@ tap_ioctl(struct pmd_internals *pmd, unsigned long request,
  struct ifreq *ifr, int set)
 {
short req_flags = ifr->ifr_flags;
+   int remote = !!pmd->remote_if_index;
 
-   snprintf(ifr->ifr_name, IFNAMSIZ, "%s", pmd->name);
+   /*
+* If there is a remote netdevice, apply ioctl on it, then apply it on
+* the tap ne

[dpdk-dev] [PATCH] net/sfc: fix incorrect rebase during development

[Andrew Rybchenko]

Patch which factors out libefx-based Tx datapath was incorrectly
rebased during development over patch which removes
RTE_LIBRTE_SFC_EFX_TSO config option.

Fixes: c5a7da9b5d10 ("net/sfc: factor out libefx-based Tx datapath")

Signed-off-by: Andrew Rybchenko 
---
Please, squash the patch into c5a7da9b5d10.

 drivers/net/sfc/sfc_tx.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/net/sfc/sfc_tx.c b/drivers/net/sfc/sfc_tx.c
index e7ec802..d1a064e 100644
--- a/drivers/net/sfc/sfc_tx.c
+++ b/drivers/net/sfc/sfc_tx.c
@@ -668,7 +668,6 @@ sfc_efx_xmit_pkts(void *tx_queue, struct rte_mbuf 
**tx_pkts, uint16_t nb_pkts)
 */
pkt_descs += sfc_efx_tx_maybe_insert_tag(txq, m_seg, &pend);
 
-#ifdef RTE_LIBRTE_SFC_EFX_TSO
if (m_seg->ol_flags & PKT_TX_TCP_SEG) {
/*
 * We expect correct 'pkt->l[2, 3, 4]_len' values
@@ -706,7 +705,6 @@ sfc_efx_xmit_pkts(void *tx_queue, struct rte_mbuf 
**tx_pkts, uint16_t nb_pkts)
 * as for the usual non-TSO path
 */
}
-#endif /* RTE_LIBRTE_SFC_EFX_TSO */
 
for (; m_seg != NULL; m_seg = m_seg->next) {
efsys_dma_addr_tnext_frag;
-- 
2.9.3

[dpdk-dev] [PATCH v3 0/5] pipeline personalization profile support

[Beilei Xing]

Due to limited resources of X*710 (parser and analyzer configuration
tables, number of packet classification types, number of packet types,
filters configuration tables, etc.), it's impossible to simultaneously
support all protocols/filters required for different parts on network.
To enable protocols/filters extensions for X*710, new Admin Command
for loading user defined configurations is added.
PPP is a format of extend configuration for X*710, it allows user to
load user defined configuration to X*710.

List of possible use cases for extended X*710 configuration using
profiles could include following:
Configuring Analyzer/Parser to support new protocols, e.g.
- IP L2TPv3 tunneling protocol
- IPSec ESP/AH protocols
- MPLSoGRE, MPLSoUDP tunnels
- GTP-C/GTP-U tunnels
New PCTYPEs for offloading packet classification to X*710. e.g.
- new IP Protocol in addition to TCP/UDP/SCTP
- new TCP/UDP subtypes, like TCP SYN, TCP FIN
- new PCTYPE for tunneled packets like GTP-C, GTP-U
New PTYPEs for packets identification, e.g.
- MAC, MPLS, IP4, UDP
- MAC, MPLS, MPLS, IP6, TCP
Fixes for NVM configuration, e.g.
- list of enabled stat counters to improve throughput
- parser/analyzer configuration for some corner cases

Beilei Xing (5):
  net/i40e: add pipeline personalization profile processing
  app/testpmd: add command for loading a profile
  net/i40e: add get all loaded profiles
  app/testpmd: add command for getting loaded profiles
  doc: add pipeline personalization profile support for i40e

 app/test-pmd/cmdline.c| 163 +
 app/test-pmd/config.c |  67 +
 app/test-pmd/testpmd.h|  25 
 doc/guides/rel_notes/release_17_05.rst|   4 +
 drivers/net/i40e/i40e_ethdev.c| 228 ++
 drivers/net/i40e/i40e_ethdev.h|   5 +
 drivers/net/i40e/rte_pmd_i40e.h   |  32 +
 drivers/net/i40e/rte_pmd_i40e_version.map |   7 +
 8 files changed, 531 insertions(+)

-- 
2.5.5

[dpdk-dev] [PATCH v3 2/5] app/testpmd: add command for loading a profile

[Beilei Xing]

This patch is to add testpmd CLI for loading a pipeline
personalization profile.

Signed-off-by: Beilei Xing 
---
 app/test-pmd/cmdline.c  | 73 +
 app/test-pmd/config.c   | 67 +
 app/test-pmd/testpmd.h  |  3 ++
 drivers/net/i40e/rte_pmd_i40e.h |  1 -
 4 files changed, 143 insertions(+), 1 deletion(-)

diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index 6e0625d..5d6dd73 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -579,6 +579,9 @@ static void cmd_help_long_parsed(void *parsed_result,
"E-tag set filter del e-tag-id (value) port (port_id)\n"
"Delete an E-tag forwarding filter on a port\n\n"
 
+   "add ppp (port_id) (profile_path)\n"
+   "Load a profile package on a port\n\n"
+
, list_pkt_forwarding_modes()
);
}
@@ -12402,6 +12405,75 @@ cmdline_parse_inst_t cmd_set_vf_vlan_tag = {
},
 };
 
+/* Load Pipeline Personalization Profile */
+struct cmd_add_ppp_result {
+   cmdline_fixed_string_t add;
+   cmdline_fixed_string_t ppp;
+   uint8_t port_id;
+   char filepath[];
+};
+
+cmdline_parse_token_string_t cmd_add_ppp_add =
+   TOKEN_STRING_INITIALIZER(struct cmd_add_ppp_result, add, "add");
+cmdline_parse_token_string_t cmd_add_ppp_ppp =
+   TOKEN_STRING_INITIALIZER(struct cmd_add_ppp_result, ppp, "ppp");
+cmdline_parse_token_num_t cmd_add_ppp_port_id =
+   TOKEN_NUM_INITIALIZER(struct cmd_add_ppp_result, port_id, UINT8);
+cmdline_parse_token_string_t cmd_add_ppp_filepath =
+   TOKEN_STRING_INITIALIZER(struct cmd_add_ppp_result, filepath, NULL);
+
+static void
+cmd_add_ppp_parsed(
+   void *parsed_result,
+   __attribute__((unused)) struct cmdline *cl,
+   __attribute__((unused)) void *data)
+{
+   struct cmd_add_ppp_result *res = parsed_result;
+   uint8_t *buff;
+   uint32_t size;
+   int ret = -ENOTSUP;
+
+   if (res->port_id > nb_ports) {
+   printf("Invalid port, range is [0, %d]\n", nb_ports - 1);
+   return;
+   }
+
+   if (!all_ports_stopped()) {
+   printf("Please stop all ports first\n");
+   return;
+   }
+
+   buff = open_ppp_package_file(res->filepath, &size);
+   if (!buff)
+   return;
+
+#ifdef RTE_LIBRTE_I40E_PMD
+   if (ret == -ENOTSUP) {
+   ret = rte_pmd_i40e_process_ppp_package(res->port_id,
+  buff, size, 1);
+   if (ret < 0)
+   printf("Failed to load profile.\n");
+   else if (ret > 0)
+   printf("Profile has already existed.\n");
+   }
+#endif
+
+   close_ppp_package_file(buff);
+}
+
+cmdline_parse_inst_t cmd_add_ppp = {
+   .f = cmd_add_ppp_parsed,
+   .data = NULL,
+   .help_str = "add/remove ppp  ",
+   .tokens = {
+   (void *)&cmd_add_ppp_add,
+   (void *)&cmd_add_ppp_ppp,
+   (void *)&cmd_add_ppp_port_id,
+   (void *)&cmd_add_ppp_filepath,
+   NULL,
+   },
+};
+
 /* 

 */
 
 /* list of instructions */
@@ -12577,6 +12649,7 @@ cmdline_parse_ctx_t main_ctx[] = {
(cmdline_parse_inst_t *)&cmd_set_vf_allmulti,
(cmdline_parse_inst_t *)&cmd_set_vf_broadcast,
(cmdline_parse_inst_t *)&cmd_set_vf_vlan_tag,
+   (cmdline_parse_inst_t *)&cmd_add_ppp,
NULL,
 };
 
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index 80491fc..eb3d572 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -3245,3 +3245,70 @@ port_dcb_info_display(uint8_t port_id)
printf("\t%4d", dcb_info.tc_queue.tc_txq[0][i].nb_queue);
printf("\n");
 }
+
+uint8_t *
+open_ppp_package_file(const char *file_path, uint32_t *size)
+{
+   FILE *fh = fopen(file_path, "rb");
+   uint32_t pkg_size;
+   uint8_t *buf = NULL;
+   int ret = 0;
+
+   if (size)
+   *size = 0;
+
+   if (fh == NULL) {
+   printf("%s: Failed to open %s\n", __func__, file_path);
+   return buf;
+   }
+
+   ret = fseek(fh, 0, SEEK_END);
+   if (ret < 0) {
+   fclose(fh);
+   printf("%s: File operations failed\n", __func__);
+   return buf;
+   }
+
+   pkg_size = ftell(fh);
+
+   buf = (uint8_t *)malloc(pkg_size);
+   if (!buf) {
+   fclose(fh);
+   printf("%s: Failed to malloc memory\n", __func__);
+   return buf;
+   }
+
+   ret = fseek(fh, 0, SEEK_SET);
+   if (ret < 0) {
+   fclose(fh);
+   printf("%s: File seek operation failed\n", __func__);
+   close_ppp_packag

[dpdk-dev] [PATCH v3 4/5] app/testpmd: add command for getting loaded profiles

[Beilei Xing]

This patch is to add testpmd CLI for getting all loaded profiles.

Signed-off-by: Beilei Xing 
---
 app/test-pmd/cmdline.c | 89 ++
 app/test-pmd/testpmd.h | 22 +
 2 files changed, 111 insertions(+)

diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index 5d6dd73..9ed17a5 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -215,6 +215,9 @@ static void cmd_help_long_parsed(void *parsed_result,
 
"read txd (port_id) (queue_id) (txd_id)\n"
"Display a TX descriptor of a port TX queue.\n\n"
+
+   "get ppp list (port_id)\n"
+   "Get ppp profile info list\n\n"
);
}
 
@@ -12474,6 +12477,91 @@ cmdline_parse_inst_t cmd_add_ppp = {
},
 };
 
+/* Get Pipeline Personalization Profile list*/
+#define PROFILE_INFO_SIZE 48
+#define MAX_PROFILE_NUM 16
+
+struct cmd_get_ppp_list_result {
+   cmdline_fixed_string_t get;
+   cmdline_fixed_string_t ppp;
+   cmdline_fixed_string_t list;
+   uint8_t port_id;
+};
+
+cmdline_parse_token_string_t cmd_get_ppp_list_get =
+   TOKEN_STRING_INITIALIZER(struct cmd_get_ppp_list_result, get, "get");
+cmdline_parse_token_string_t cmd_get_ppp_list_ppp =
+   TOKEN_STRING_INITIALIZER(struct cmd_get_ppp_list_result, ppp, "ppp");
+cmdline_parse_token_string_t cmd_get_ppp_list_list =
+   TOKEN_STRING_INITIALIZER(struct cmd_get_ppp_list_result, list, "list");
+cmdline_parse_token_num_t cmd_get_ppp_list_port_id =
+   TOKEN_NUM_INITIALIZER(struct cmd_get_ppp_list_result, port_id, UINT8);
+
+static void
+cmd_get_ppp_list_parsed(
+   void *parsed_result,
+   __attribute__((unused)) struct cmdline *cl,
+   __attribute__((unused)) void *data)
+{
+   struct cmd_get_ppp_list_result *res = parsed_result;
+   struct profile_list *p_list;
+   struct profile_info *p_info;
+   uint32_t p_num;
+   uint32_t size;
+   uint32_t i;
+   int ret = -ENOTSUP;
+
+   if (res->port_id > nb_ports) {
+   printf("Invalid port, range is [0, %d]\n", nb_ports - 1);
+   return;
+   }
+
+   size = PROFILE_INFO_SIZE * MAX_PROFILE_NUM + 4;
+   p_list = (struct profile_list *)malloc(size);
+   if (!p_list)
+   printf("%s: Failed to malloc buffer\n", __func__);
+
+#ifdef RTE_LIBRTE_I40E_PMD
+   if (ret == -ENOTSUP) {
+   ret = rte_pmd_i40e_get_ppp_list(res->port_id,
+   (uint8_t *)p_list, size);
+   if (ret < 0)
+   printf("Failed to get ppp list\n");
+   }
+#endif
+
+   if (!ret) {
+   p_num = p_list->p_count;
+   printf("Profile number is: %d\n\n", p_num);
+
+   for (i = 0; i < p_num; i++) {
+   p_info = &p_list->p_info[i];
+   printf("Profile %d:\n", i);
+   printf("Track id: 0x%x\n", p_info->track_id);
+   printf("Version:  %d.%d.%d.%d \n",
+  p_info->version.major,
+  p_info->version.minor,
+  p_info->version.update,
+  p_info->version.draft);
+   printf("Profile name: %s\n\n", p_info->name);
+   }
+   }
+
+   free(p_list);
+}
+
+cmdline_parse_inst_t cmd_get_ppp_list = {
+   .f = cmd_get_ppp_list_parsed,
+   .data = NULL,
+   .help_str = "get ppp list ",
+   .tokens = {
+   (void *)&cmd_get_ppp_list_get,
+   (void *)&cmd_get_ppp_list_ppp,
+   (void *)&cmd_get_ppp_list_list,
+   (void *)&cmd_get_ppp_list_port_id,
+   NULL,
+   },
+};
 /* 

 */
 
 /* list of instructions */
@@ -12650,6 +12738,7 @@ cmdline_parse_ctx_t main_ctx[] = {
(cmdline_parse_inst_t *)&cmd_set_vf_broadcast,
(cmdline_parse_inst_t *)&cmd_set_vf_vlan_tag,
(cmdline_parse_inst_t *)&cmd_add_ppp,
+   (cmdline_parse_inst_t *)&cmd_get_ppp_list,
NULL,
 };
 
diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h
index 6bcd9c4..b9132fa 100644
--- a/app/test-pmd/testpmd.h
+++ b/app/test-pmd/testpmd.h
@@ -97,6 +97,28 @@ struct pkt_burst_stats {
 };
 #endif
 
+#ifdef RTE_LIBRTE_I40E_PMD
+#define PPP_NAME_SIZE 32
+struct ppp_version {
+   uint8_t major;
+   uint8_t minor;
+   uint8_t update;
+   uint8_t draft;
+};
+
+struct profile_info {
+   uint32_t track_id;
+   struct ppp_version version;
+   uint8_t reserved[8];
+   uint8_t name[PPP_NAME_SIZE];
+};
+
+struct profile_list {
+   uint32_t p_count;
+   struct profile_info p_info[1];
+};
+#endif
+
 /**
  * The data structure associated with a forwarding stream between a receive
  * por

[dpdk-dev] [PATCH v3 3/5] net/i40e: add get all loaded profiles

[Beilei Xing]

This patch is to add get all loaded profiles function.

Signed-off-by: Beilei Xing 
---
 drivers/net/i40e/i40e_ethdev.c| 28 
 drivers/net/i40e/rte_pmd_i40e.h   | 12 
 drivers/net/i40e/rte_pmd_i40e_version.map |  1 +
 3 files changed, 41 insertions(+)

diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index 7aff9a3..9c0441b 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -11260,6 +11260,9 @@ rte_pmd_i40e_reset_vf_stats(uint8_t port,
return 0;
 }
 
+#define I40E_PROFILE_INFO_SIZE 48
+#define I40E_MAX_PROFILE_NUM 16
+
 static void
 i40e_generate_profile_info_sec(char *name, struct i40e_ppp_version *version,
   uint32_t track_id, uint8_t *profile_info_sec,
@@ -11459,3 +11462,28 @@ rte_pmd_i40e_process_ppp_package(uint8_t port, uint8_t 
*buff,
rte_free(profile_info_sec);
return status;
 }
+
+int
+rte_pmd_i40e_get_ppp_list(uint8_t port, uint8_t *buff, uint32_t size)
+{
+   struct rte_eth_dev *dev;
+   struct i40e_hw *hw;
+   enum i40e_status_code status = I40E_SUCCESS;
+
+   RTE_ETH_VALID_PORTID_OR_ERR_RET(port, -ENODEV);
+
+   dev = &rte_eth_devices[port];
+
+   if (!is_device_supported(dev, &rte_i40e_pmd))
+   return -ENOTSUP;
+
+   if (size < (I40E_PROFILE_INFO_SIZE * I40E_MAX_PROFILE_NUM + 4))
+   return -EINVAL;
+
+   hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+
+   status = i40e_aq_get_ppp_list(hw, (void *)buff,
+ size, 0, NULL);
+
+   return status;
+}
diff --git a/drivers/net/i40e/rte_pmd_i40e.h b/drivers/net/i40e/rte_pmd_i40e.h
index 29e5c61..6d9488f 100644
--- a/drivers/net/i40e/rte_pmd_i40e.h
+++ b/drivers/net/i40e/rte_pmd_i40e.h
@@ -352,4 +352,16 @@ int rte_pmd_i40e_reset_vf_stats(uint8_t port,
  */
 int rte_pmd_i40e_process_ppp_package(uint8_t port, uint8_t *buff,
 uint32_t size, bool add);
+
+/**
+ * rte_pmd_i40e_get_ppp_list - Get loaded profile list
+ * @port: port id
+ * @buff: buffer for response
+ * @size: buffer size
+ * @return
+ *   - (0) if successful.
+ *   - (-ENODEV) if *port* invalid.
+ *   - (-EINVAL) if bad parameter.
+ */
+int rte_pmd_i40e_get_ppp_list(uint8_t port, uint8_t *buff, uint32_t size);
 #endif /* _PMD_I40E_H_ */
diff --git a/drivers/net/i40e/rte_pmd_i40e_version.map 
b/drivers/net/i40e/rte_pmd_i40e_version.map
index 01c4a90..9808893 100644
--- a/drivers/net/i40e/rte_pmd_i40e_version.map
+++ b/drivers/net/i40e/rte_pmd_i40e_version.map
@@ -27,4 +27,5 @@ DPDK_17.05 {
global:
 
rte_pmd_i40e_process_ppp_package;
+   rte_pmd_i40e_get_ppp_list;
 };
-- 
2.5.5

[dpdk-dev] [PATCH v3 1/5] net/i40e: add pipeline personalization profile processing

[Beilei Xing]

Add support for adding or removing a pipeline personalization
profile package.

Signed-off-by: Beilei Xing 
---
 app/test-pmd/cmdline.c|   1 +
 drivers/net/i40e/i40e_ethdev.c| 200 ++
 drivers/net/i40e/i40e_ethdev.h|   5 +
 drivers/net/i40e/rte_pmd_i40e.h   |  21 
 drivers/net/i40e/rte_pmd_i40e_version.map |   6 +
 5 files changed, 233 insertions(+)

diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index 47f935d..6e0625d 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -37,6 +37,7 @@
 #include 
 #include 
 #include 
+#include 
 #include 
 #include 
 #include 
diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index 3702214..7aff9a3 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -11259,3 +11259,203 @@ rte_pmd_i40e_reset_vf_stats(uint8_t port,
 
return 0;
 }
+
+static void
+i40e_generate_profile_info_sec(char *name, struct i40e_ppp_version *version,
+  uint32_t track_id, uint8_t *profile_info_sec,
+  bool add)
+{
+   struct i40e_profile_section_header *sec = NULL;
+   struct i40e_profile_info *pinfo;
+
+   sec = (struct i40e_profile_section_header *)profile_info_sec;
+   sec->tbl_size = 1;
+   sec->data_end = sizeof(struct i40e_profile_section_header) +
+   sizeof(struct i40e_profile_info);
+   sec->section.type = SECTION_TYPE_INFO;
+   sec->section.offset = sizeof(struct i40e_profile_section_header);
+   sec->section.size = sizeof(struct i40e_profile_info);
+   pinfo = (struct i40e_profile_info *)(profile_info_sec +
+sec->section.offset);
+   pinfo->track_id = track_id;
+   memcpy(pinfo->name, name, I40E_PPP_NAME_SIZE);
+   memcpy(&pinfo->version, version, sizeof(struct i40e_ppp_version));
+   if (add)
+   pinfo->op = I40E_PPP_ADD_TRACKID;
+   else
+   pinfo->op = I40E_PPP_REMOVE_TRACKID;
+}
+
+static enum i40e_status_code
+i40e_add_rm_profile_info(struct i40e_hw *hw, uint8_t *profile_info_sec)
+{
+   enum i40e_status_code status = I40E_SUCCESS;
+   struct i40e_profile_section_header *sec;
+   uint32_t track_id;
+   uint32_t offset = 0, info = 0;
+
+   sec = (struct i40e_profile_section_header *)profile_info_sec;
+   track_id = ((struct i40e_profile_info *)(profile_info_sec +
+sec->section.offset))->track_id;
+
+   status = i40e_aq_write_ppp(hw, (void *)sec, sec->data_end,
+  track_id, &offset, &info, NULL);
+   if (status)
+   PMD_DRV_LOG(ERR, "Failed to add/remove profile info: "
+   "offset %d, info %d",
+   offset, info);
+
+   return status;
+}
+
+#define I40E_PROFILE_INFO_SIZE 48
+#define I40E_MAX_PROFILE_NUM 16
+
+/* Check if the profile info exists */
+static int
+i40e_check_profile_info(uint8_t port, uint8_t *profile_info_sec)
+{
+   struct rte_eth_dev *dev = &rte_eth_devices[port];
+   struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+   uint8_t *buff;
+   struct i40e_profile_list *p_list;
+   struct i40e_profile_info *pinfo, *p;
+   int ret;
+
+   buff = rte_zmalloc("i40e_pinfo_list",
+  (I40E_PROFILE_INFO_SIZE * I40E_MAX_PROFILE_NUM + 4),
+  0);
+   if (!buff) {
+   PMD_DRV_LOG(ERR, "failed to allocate memory");
+   return -1;
+   }
+
+   ret = i40e_aq_get_ppp_list(hw, (void *)buff,
+ (I40E_PROFILE_INFO_SIZE * I40E_MAX_PROFILE_NUM + 4),
+ 0, NULL);
+   if (ret) {
+   PMD_DRV_LOG(ERR, "Failed to get profile info list.");
+   rte_free(buff);
+   return -1;
+   }
+   p_list = (struct i40e_profile_list *)buff;
+   pinfo = (struct i40e_profile_info *)(profile_info_sec +
+sizeof(struct i40e_profile_section_header));
+   for (uint32_t i = 0; i < p_list->p_count; i++) {
+   p = &p_list->p_info[i];
+   if ((pinfo->track_id == p->track_id) &&
+   !memcmp(&pinfo->version, &p->version,
+   sizeof(struct i40e_ppp_version)) &&
+   !memcmp(&pinfo->name, &p->name,
+   I40E_PPP_NAME_SIZE)) {
+   PMD_DRV_LOG(INFO, "Profile exists.");
+   rte_free(buff);
+   return 1;
+   }
+   }
+
+   rte_free(buff);
+   return 0;
+}
+
+int
+rte_pmd_i40e_process_ppp_package(uint8_t port, uint8_t *buff,
+uint32_t size, bool add)
+{
+   struct rte_eth_dev *dev;
+   struct i40e_hw *hw;
+   struct i40e_package_header *pk

[dpdk-dev] [PATCH v3 5/5] doc: add pipeline personalization profile support for i40e

[Beilei Xing]

Signed-off-by: Beilei Xing 
---
 doc/guides/rel_notes/release_17_05.rst | 4 
 1 file changed, 4 insertions(+)

diff --git a/doc/guides/rel_notes/release_17_05.rst 
b/doc/guides/rel_notes/release_17_05.rst
index 3e48224..f17e03b 100644
--- a/doc/guides/rel_notes/release_17_05.rst
+++ b/doc/guides/rel_notes/release_17_05.rst
@@ -85,6 +85,10 @@ New Features
   Data Ring, ability to register memory regions.
 
 
+* **Added pipeline personalization profile support for i40e.**
+
+  * Added loading pipeline personalization profile to i40e FW.
+
 Resolved Issues
 ---
 
-- 
2.5.5

Re: [dpdk-dev] [PATCH 0/7] net/i40e: base code update

[Ferruh Yigit]

On 3/23/2017 1:26 AM, Wu, Jingjing wrote:
> 
> 
>> -Original Message-
>> From: Yigit, Ferruh
>> Sent: Wednesday, March 22, 2017 10:29 PM
>> To: Wu, Jingjing ; dev@dpdk.org
>> Cc: Zhang, Helin ; Xing, Beilei 
>> 
>> Subject: Re: [dpdk-dev] [PATCH 0/7] net/i40e: base code update
>>
>> On 3/22/2017 9:24 AM, Jingjing Wu wrote:
>>> Base code update with main changes:
>>>  - new AQ commands for Pipeline Personalization Profile
>>>  - new AQ commands for cloud filter
>>>  - reduce wait time for adminq command completion
>>>
>>> Jingjing Wu (7):
>>>   net/i40e/base: define bit for HW ATR evict control
>>>   net/i40e/base: control register read/write on X722
>>>   net/i40e/base: fix potential out of bound array access
>>>   net/i40e/base: new AQ commands for ppp
>>>   net/i40e/base: reduce wait time for adminq command
>>>   net/i40e/base: add VF offload flags
>>>   net/i40e/base: new AQ commands for cloud filter
>>
>> Series applied to dpdk-next-net/master, thanks.
>>
>> Not for this patch but future ones, what do you think adding a note to keep
>> record of the updated base code version? Do you think does it help tracing
>> share code updates?
> 
> Agree. I will send doc update for it.
> Add one readme in i40e/base folder just like ixgbe did. What do you think?

I think that is good to have.

> 
> Thanks
> Jingjing
> 

Re: [dpdk-dev] [PATCH 01/39] eventdev: update PMD dequeue timeout conversion callback

[Jerin Jacob]

On Wed, Mar 15, 2017 at 05:27:15PM +, Van Haaren, Harry wrote:
> > From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> > Sent: Friday, March 3, 2017 5:28 PM
> > To: dev@dpdk.org
> > Subject: [dpdk-dev] [PATCH 01/39] eventdev: update PMD dequeue timeout 
> > conversion
> > callback
> > 
> > eventdev driver may return error on dequeue timeout tick conversion.
> > Change the pmd callback interface to address the same.
> > 
> > Signed-off-by: Jerin Jacob 
> 
> 
> Given this is not really related to octeontx PMD, I suggest applying to tree 
> and dropping from octeontx patchset.
> 
> Acked-by: Harry van Haaren 

Applied to dpdk-next-eventdev/master. Thanks.

> 

Re: [dpdk-dev] [PATCH 02/39] app/test: fix eventdev reconfigure test

[Jerin Jacob]

On Wed, Mar 15, 2017 at 05:28:35PM +, Van Haaren, Harry wrote:
> > From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> > Sent: Friday, March 3, 2017 5:28 PM
> > To: dev@dpdk.org
> > Subject: [dpdk-dev] [PATCH 02/39] app/test: fix eventdev reconfigure test
> > 
> > Minimum value of nb_event_ports and/or nb_event_queues
> > should be one before reconfiguring the event device.
> > 
> > Fixes: 4c9a26e419a7 ("app/test: unit test case for eventdev APIs")
> > 
> > Signed-off-by: Jerin Jacob 
> 
> Acked-by: Harry van Haaren 

Applied to dpdk-next-eventdev/master. Thanks.

> 
> 

Re: [dpdk-dev] [PATCH v2] eventdev: remove default queue overriding

[Jerin Jacob]

On Tue, Mar 21, 2017 at 01:51:45PM +0530, Jerin Jacob wrote:
> On Fri, Mar 10, 2017 at 03:19:15PM +, Harry van Haaren wrote:
> > PMDs that only do a specific type of scheduling cannot provide
> > CFG_ALL_TYPES, so the Eventdev infrastructure should not demand
> > that every PMD supports CFG_ALL_TYPES.
> > 
> > By not overriding the default configuration of the queue as
> > suggested by the PMD, the eventdev_common unit tests can pass
> > on all PMDs, regardless of their capabilities.
> > 
> > RTE_EVENT_QUEUE_CFG_DEFAULT is no longer used by the eventdev layer
> > it can be removed now. Applications should use CFG_ALL_TYPES
> > if they require enqueue of all types a queue, or specify which
> > type of queue they require.
> > 
> > The CFG_DEFAULT value is changed to CFG_ALL_TYPES in event/skeleton,
> > to not break the compile.
> > 
> > A capability flag is added that indicates if the underlying PMD
> > supports creating queues of ALL_TYPES.
> > 
> > Signed-off-by: Harry van Haaren 
> 
> I think It is reasonable to have this capability if SW PMD can not
> support it for performance reasons. The only downside is, In application
> side there will be changes in the fast path.I think, The reasonable
> trade-off between performance and portability to keep packet processing
> functions as common and keep the pipeline advancement logic as different
> main loop based on capability.
> 
> Two reasons why CFG_ALL_TYPES important for HW
> - Event queue is the precious resource and it is very limited and it
>   consumes power and internal resources like SRAM
> - The use case like flow based event pipelining will not have constraint
>   on which event queues it enqueues to
> 
> I think We can add this capability flag now and once we have performance
> and latency test cases for eventdev then we can check is there any scope
> for improvement in SW PMD.With that note,
> 
> Acked-by: Jerin Jacob 

Applied to dpdk-next-eventdev/master. Thanks.


> 
> > 
> > ---
> > 
> > v2:
> > - added capability flag to indicate if PMD supports ALL_TYPES
> > 
> > ---
> >  drivers/event/skeleton/skeleton_eventdev.c |  2 +-
> >  lib/librte_eventdev/rte_eventdev.c |  1 -
> >  lib/librte_eventdev/rte_eventdev.h | 13 +++--
> >  3 files changed, 8 insertions(+), 8 deletions(-)
> > 
> > diff --git a/drivers/event/skeleton/skeleton_eventdev.c 
> > b/drivers/event/skeleton/skeleton_eventdev.c
> > index dee0faf..308e28e 100644
> > --- a/drivers/event/skeleton/skeleton_eventdev.c
> > +++ b/drivers/event/skeleton/skeleton_eventdev.c
> > @@ -196,7 +196,7 @@ skeleton_eventdev_queue_def_conf(struct rte_eventdev 
> > *dev, uint8_t queue_id,
> >  
> > queue_conf->nb_atomic_flows = (1ULL << 20);
> > queue_conf->nb_atomic_order_sequences = (1ULL << 20);
> > -   queue_conf->event_queue_cfg = RTE_EVENT_QUEUE_CFG_DEFAULT;
> > +   queue_conf->event_queue_cfg = RTE_EVENT_QUEUE_CFG_ALL_TYPES;
> > queue_conf->priority = RTE_EVENT_DEV_PRIORITY_NORMAL;
> >  }
> >  
> > diff --git a/lib/librte_eventdev/rte_eventdev.c 
> > b/lib/librte_eventdev/rte_eventdev.c
> > index 68bfc3b..c32a776 100644
> > --- a/lib/librte_eventdev/rte_eventdev.c
> > +++ b/lib/librte_eventdev/rte_eventdev.c
> > @@ -593,7 +593,6 @@ rte_event_queue_setup(uint8_t dev_id, uint8_t queue_id,
> > RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->queue_def_conf,
> > -ENOTSUP);
> > (*dev->dev_ops->queue_def_conf)(dev, queue_id, &def_conf);
> > -   def_conf.event_queue_cfg = RTE_EVENT_QUEUE_CFG_DEFAULT;
> > queue_conf = &def_conf;
> > }
> >  
> > diff --git a/lib/librte_eventdev/rte_eventdev.h 
> > b/lib/librte_eventdev/rte_eventdev.h
> > index 7073987..4c73a82 100644
> > --- a/lib/librte_eventdev/rte_eventdev.h
> > +++ b/lib/librte_eventdev/rte_eventdev.h
> > @@ -271,6 +271,13 @@ struct rte_mbuf; /* we just use mbuf pointers; no need 
> > to include rte_mbuf.h */
> >   *
> >   * @see rte_event_schedule(), rte_event_dequeue_burst()
> >   */
> > +#define RTE_EVENT_DEV_CAP_QUEUE_ALL_TYPES (1ULL << 3)
> > +/**< Event device is capable of enqueuing events of any type to any queue.
> > + * If this capability is not set, the queue only supports events of the
> > + *  *RTE_EVENT_QUEUE_CFG_* type that it was created with.
> > + *
> > + * @see RTE_EVENT_QUEUE_CFG_* values
> > + */
> >  
> >  /* Event device priority levels */
> >  #define RTE_EVENT_DEV_PRIORITY_HIGHEST   0
> > @@ -471,12 +478,6 @@ rte_event_dev_configure(uint8_t dev_id,
> >  /* Event queue specific APIs */
> >  
> >  /* Event queue configuration bitmap flags */
> > -#define RTE_EVENT_QUEUE_CFG_DEFAULT(0)
> > -/**< Default value of *event_queue_cfg* when rte_event_queue_setup() 
> > invoked
> > - * with queue_conf == NULL
> > - *
> > - * @see rte_event_queue_setup()
> > - */
> >  #define RTE_EVENT_QUEUE_CFG_TYPE_MASK  (3ULL << 0)
> >  /**< Mask for event queue schedule type configuration request */
> >  #define RTE_

Re: [dpdk-dev] [PATCH v4 02/17] app/test: eventdev link all queues before start

[Jerin Jacob]

On Mon, Mar 20, 2017 at 10:16:04AM +0530, Jerin Jacob wrote:
> On Fri, Mar 10, 2017 at 07:43:17PM +, Harry van Haaren wrote:
> > The software eventdev can lock-up if not all queues are
> > linked to a port. For this reason, the software evendev
> > fails to start if queues are not linked to anything.
> > 
> > This commit creates dummy links from all queues to port
> > 0 in the eventdev setup function and start/stop test,
> > which would otherwise fail due to unlinked queues.
> > 
> > Signed-off-by: Harry van Haaren 
> 
> Acked-by: Jerin Jacob 

Applied to dpdk-next-eventdev/master. Thanks.

> 
> > ---
> >  app/test/test_eventdev.c | 8 
> >  1 file changed, 8 insertions(+)
> > 
> > diff --git a/app/test/test_eventdev.c b/app/test/test_eventdev.c
> > index 042a446..324ef5a 100644
> > --- a/app/test/test_eventdev.c
> > +++ b/app/test/test_eventdev.c
> > @@ -543,6 +543,10 @@ test_eventdev_start_stop(void)
> > TEST_ASSERT_SUCCESS(ret, "Failed to setup port%d", i);
> > }
> >  
> > +   ret = rte_event_port_link(TEST_DEV_ID, 0, NULL, NULL, 0);
> > +   TEST_ASSERT(ret == rte_event_queue_count(TEST_DEV_ID),
> > +   "Failed to link port, device %d", TEST_DEV_ID);
> > +
> > ret = rte_event_dev_start(TEST_DEV_ID);
> > TEST_ASSERT_SUCCESS(ret, "Failed to start device%d", TEST_DEV_ID);
> >  
> > @@ -569,6 +573,10 @@ eventdev_setup_device(void)
> > TEST_ASSERT_SUCCESS(ret, "Failed to setup port%d", i);
> > }
> >  
> > +   ret = rte_event_port_link(TEST_DEV_ID, 0, NULL, NULL, 0);
> > +   TEST_ASSERT(ret == rte_event_queue_count(TEST_DEV_ID),
> > +   "Failed to link port, device %d", TEST_DEV_ID);
> > +
> > ret = rte_event_dev_start(TEST_DEV_ID);
> > TEST_ASSERT_SUCCESS(ret, "Failed to start device%d", TEST_DEV_ID);
> >  
> > -- 
> > 2.7.4
> > 

Re: [dpdk-dev] [PATCH v4 04/17] eventdev: add APIs for extended stats

[Jerin Jacob]

On Fri, Mar 17, 2017 at 05:52:28PM +0530, Jerin Jacob wrote:
> On Fri, Mar 10, 2017 at 07:43:19PM +, Harry van Haaren wrote:
> > From: Bruce Richardson 
> > 
> > Add in APIs for extended stats so that eventdev implementations can report
> > out information on their internal state. The APIs are based on, but not
> > identical to, the equivalent ethdev functions.
> > 
> > Signed-off-by: Bruce Richardson 
> > Signed-off-by: Harry van Haaren 
> 
> Acked-by: Jerin Jacob 

Applied to dpdk-next-eventdev/master. Thanks.

> 
> > 
> > ---
> > 
> > v3 -> v4:
> > - xstats API reset allows selection of device/port/queue to reset
> > ---
> >  lib/librte_eventdev/rte_eventdev.c   |  83 
> >  lib/librte_eventdev/rte_eventdev.h   | 142 
> > +++
> >  lib/librte_eventdev/rte_eventdev_pmd.h   |  74 ++
> >  lib/librte_eventdev/rte_eventdev_version.map |   4 +
> >  4 files changed, 303 insertions(+)
> > 
> > diff --git a/lib/librte_eventdev/rte_eventdev.c 
> > b/lib/librte_eventdev/rte_eventdev.c
> > index 68bfc3b..0280ef0 100644
> > --- a/lib/librte_eventdev/rte_eventdev.c
> > +++ b/lib/librte_eventdev/rte_eventdev.c
> > @@ -920,6 +920,89 @@ rte_event_dev_dump(uint8_t dev_id, FILE *f)
> >  
> >  }
> >  
> > +static int
> > +xstats_get_count(uint8_t dev_id, enum rte_event_dev_xstats_mode mode,
> > +   uint8_t queue_port_id)
> > +{
> > +   struct rte_eventdev *dev = &rte_eventdevs[dev_id];
> > +   if (dev->dev_ops->xstats_get_names != NULL)
> > +   return (*dev->dev_ops->xstats_get_names)(dev, mode,
> > +   queue_port_id,
> > +   NULL, NULL, 0);
> > +   return 0;
> > +}
> > +
> > +int
> > +rte_event_dev_xstats_names_get(uint8_t dev_id,
> > +   enum rte_event_dev_xstats_mode mode, uint8_t queue_port_id,
> > +   struct rte_event_dev_xstats_name *xstats_names,
> > +   unsigned int *ids, unsigned int size)
> > +{
> > +   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -ENODEV);
> > +   const int cnt_expected_entries = xstats_get_count(dev_id, mode,
> > + queue_port_id);
> > +   if (xstats_names == NULL || cnt_expected_entries < 0 ||
> > +   (int)size < cnt_expected_entries)
> > +   return cnt_expected_entries;
> > +
> > +   /* dev_id checked above */
> > +   const struct rte_eventdev *dev = &rte_eventdevs[dev_id];
> > +
> > +   if (dev->dev_ops->xstats_get_names != NULL)
> > +   return (*dev->dev_ops->xstats_get_names)(dev, mode,
> > +   queue_port_id, xstats_names, ids, size);
> > +
> > +   return -ENOTSUP;
> > +}
> > +
> > +/* retrieve eventdev extended statistics */
> > +int
> > +rte_event_dev_xstats_get(uint8_t dev_id, enum rte_event_dev_xstats_mode 
> > mode,
> > +   uint8_t queue_port_id, const unsigned int ids[],
> > +   uint64_t values[], unsigned int n)
> > +{
> > +   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -ENODEV);
> > +   const struct rte_eventdev *dev = &rte_eventdevs[dev_id];
> > +
> > +   /* implemented by the driver */
> > +   if (dev->dev_ops->xstats_get != NULL)
> > +   return (*dev->dev_ops->xstats_get)(dev, mode, queue_port_id,
> > +   ids, values, n);
> > +   return -ENOTSUP;
> > +}
> > +
> > +uint64_t
> > +rte_event_dev_xstats_by_name_get(uint8_t dev_id, const char *name,
> > +   unsigned int *id)
> > +{
> > +   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, 0);
> > +   const struct rte_eventdev *dev = &rte_eventdevs[dev_id];
> > +   unsigned int temp = -1;
> > +
> > +   if (id != NULL)
> > +   *id = (unsigned int)-1;
> > +   else
> > +   id = &temp; /* ensure driver never gets a NULL value */
> > +
> > +   /* implemented by driver */
> > +   if (dev->dev_ops->xstats_get_by_name != NULL)
> > +   return (*dev->dev_ops->xstats_get_by_name)(dev, name, id);
> > +   return -ENOTSUP;
> > +}
> > +
> > +int rte_event_dev_xstats_reset(uint8_t dev_id,
> > +   enum rte_event_dev_xstats_mode mode, int16_t queue_port_id,
> > +   const uint32_t ids[], uint32_t nb_ids)
> > +{
> > +   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
> > +   struct rte_eventdev *dev = &rte_eventdevs[dev_id];
> > +
> > +   if (dev->dev_ops->xstats_reset != NULL)
> > +   return (*dev->dev_ops->xstats_reset)(dev, mode, queue_port_id,
> > +   ids, nb_ids);
> > +   return -ENOTSUP;
> > +}
> > +
> >  int
> >  rte_event_dev_start(uint8_t dev_id)
> >  {
> > diff --git a/lib/librte_eventdev/rte_eventdev.h 
> > b/lib/librte_eventdev/rte_eventdev.h
> > index 4876353..d462047 100644
> > --- a/lib/librte_eventdev/rte_eventdev.h
> > +++ b/lib/librte_eventdev/rte_eventdev.h
> > @@ -1405,6 +1405,148 @@ rte_event_port_links_get(uint8_t dev_id, uint8_t 
> > port_id,
> >  int
> >  rte_event_dev_dum

Re: [dpdk-dev] [PATCH 0/5] add device removal event

[Gaetan Rivet]

Hi all,

Following the discussion that took place during the tech board meeting,
it seems that some clarification and some additional discussion is
needed about this series.

As of now, what I plan to do is to:

1. Rebase this series upon the bus framework [1][2][3].
2. Write further documentation to clarify the meaning of this event and
further identify the best approach to implement it.

What I expect from this is to spawn a discussion about the best
way to achieve a true hot-plug feature in DPDK.

Any thoughts, comments regarding this?
Regards,

[1] http://dpdk.org/ml/archives/dev/2017-January/054165.html
[2] http://dpdk.org/ml/archives/dev/2017-March/059423.html
[3] http://dpdk.org/ml/archives/dev/2017-March/059376.html

On Fri, Mar 3, 2017 at 4:40 PM, Gaetan Rivet  wrote:
>
> This new event represents the sudden removal of a device from its bus.
> The underlying resources exposed by the bus are expected not to be available
> anymore. The application should thus be able to react and possibly clean up
> related resources that it reserved for the removed device.
>
> This event is different from the current hotplug API available in the DPDK
> for two reasons:
>
> 1. It is a reactive design: the application reacts to a device that has been
>removed instead of removing a device from its pool.
>
> 2. The event itself is going further than the current detaching of a device
>from a DPDK application. If the bus is a hardware one, it is expected of 
> the
>underlying resources to not be available anymore.
>
> This series adds a new event type to ethdev and implements it in mlx4.
> Testpmd is also updated to report all asynchronous ethdev events including 
> this
> one for testing purposes and as a practical usage example.
>
> This series depends on the series titled
> [PATCH 1/2] net/mlx4: split the definitions to the header file
>
> Gaetan Rivet (5):
>   ethdev: introduce device removal event
>   net/mlx4: device removal event support
>   app/testpmd: generic event handler
>   app/testpmd: request link status interrupt
>   app/testpmd: request device removal interrupt
>
>  app/test-pmd/parameters.c   |   8 +
>  app/test-pmd/testpmd.c  | 103 ++
>  app/test-pmd/testpmd.h  |   2 +
>  doc/guides/nics/features/default.ini|   1 +
>  doc/guides/nics/features/mlx4.ini   |   1 +
>  doc/guides/prog_guide/env_abstraction_layer.rst |  21 +-
>  drivers/net/mlx4/mlx4.c | 258 
> 
>  drivers/net/mlx4/mlx4.h |   1 +
>  lib/librte_eal/common/include/rte_pci.h |   2 +
>  lib/librte_ether/rte_ethdev.c   |  13 +-
>  lib/librte_ether/rte_ethdev.h   |   9 +-
>  11 files changed, 375 insertions(+), 44 deletions(-)
>
> --
> 2.1.4
>

Re: [dpdk-dev] [PATCH v2 2/2] examples/ipsec-secgw: attach session-qp

[Sergio Gonzalez Monroy]

That was simpler than I thought.

For some reason I understood that the device will support thousands of 
queues and single session per queue which would have needed more app 
changes.


On 23/03/2017 08:06, akhil.go...@nxp.com wrote:

From: Akhil Goyal 

adding support for attaching session to queue pairs.
This is required as underlying crypto driver may only
support limited number of sessions per queue pair
if max_nb_sessions_per_qp > 0, session should be
attached to a particular qp.

Signed-off-by: Akhil Goyal 
---
  examples/ipsec-secgw/ipsec.c | 12 
  1 file changed, 12 insertions(+)

diff --git a/examples/ipsec-secgw/ipsec.c b/examples/ipsec-secgw/ipsec.c
index 144f0aa..b35b30f 100644
--- a/examples/ipsec-secgw/ipsec.c
+++ b/examples/ipsec-secgw/ipsec.c
@@ -47,6 +47,7 @@
  static inline int
  create_session(struct ipsec_ctx *ipsec_ctx __rte_unused, struct ipsec_sa *sa)
  {
+   struct rte_cryptodev_info cdev_info;
unsigned long cdev_id_qp = 0;
int32_t ret;
struct cdev_key key = { 0 };
@@ -73,6 +74,17 @@ create_session(struct ipsec_ctx *ipsec_ctx __rte_unused, 
struct ipsec_sa *sa)
sa->crypto_session = rte_cryptodev_sym_session_create(
ipsec_ctx->tbl[cdev_id_qp].id, sa->xforms);
  
+	rte_cryptodev_info_get(ipsec_ctx->tbl[cdev_id_qp].id, &cdev_info);

+   if (cdev_info.sym.max_nb_sessions_per_qp > 0) {
+   ret = rte_cryptodev_queue_pair_attach_sym_session(
+   ipsec_ctx->tbl[cdev_id_qp].qp,
+   sa->crypto_session);
+   if (ret < 0) {
+   RTE_LOG(ERR, IPSEC, "Session cannot be attached"
+   " to qp %u ", ipsec_ctx->tbl[cdev_id_qp].qp);


Guideline is to keep error strings in single line to facilitate grep.
Other than that:

Acked-by: Sergio Gonzalez Monroy 


+   return -1;
+   }
+   }
sa->cdev_id_qp = cdev_id_qp;
  
  	return 0;

Re: [dpdk-dev] [PATCH] net/sfc: fix incorrect rebase during development

[Ferruh Yigit]

On 3/23/2017 8:24 AM, Andrew Rybchenko wrote:
> Patch which factors out libefx-based Tx datapath was incorrectly
> rebased during development over patch which removes
> RTE_LIBRTE_SFC_EFX_TSO config option.
> 
> Fixes: c5a7da9b5d10 ("net/sfc: factor out libefx-based Tx datapath")
> 
> Signed-off-by: Andrew Rybchenko 
> ---
> Please, squash the patch into c5a7da9b5d10.

On next-net, squashed on top of c5a7da9b5d10, thanks.

Re: [dpdk-dev] [PATCH v2 2/2] examples/ipsec-secgw: attach session-qp

[Akhil Goyal]

On 3/23/2017 4:00 PM, Sergio Gonzalez Monroy wrote:

That was simpler than I thought.

For some reason I understood that the device will support thousands of
queues and single session per queue which would have needed more app
changes.

On 23/03/2017 08:06, akhil.go...@nxp.com wrote:

From: Akhil Goyal 

adding support for attaching session to queue pairs.
This is required as underlying crypto driver may only
support limited number of sessions per queue pair
if max_nb_sessions_per_qp > 0, session should be
attached to a particular qp.

Signed-off-by: Akhil Goyal 
---
  examples/ipsec-secgw/ipsec.c | 12 
  1 file changed, 12 insertions(+)

diff --git a/examples/ipsec-secgw/ipsec.c b/examples/ipsec-secgw/ipsec.c
index 144f0aa..b35b30f 100644
--- a/examples/ipsec-secgw/ipsec.c
+++ b/examples/ipsec-secgw/ipsec.c
@@ -47,6 +47,7 @@
  static inline int
  create_session(struct ipsec_ctx *ipsec_ctx __rte_unused, struct
ipsec_sa *sa)
  {
+struct rte_cryptodev_info cdev_info;
  unsigned long cdev_id_qp = 0;
  int32_t ret;
  struct cdev_key key = { 0 };
@@ -73,6 +74,17 @@ create_session(struct ipsec_ctx *ipsec_ctx
__rte_unused, struct ipsec_sa *sa)
  sa->crypto_session = rte_cryptodev_sym_session_create(
  ipsec_ctx->tbl[cdev_id_qp].id, sa->xforms);
  +rte_cryptodev_info_get(ipsec_ctx->tbl[cdev_id_qp].id, &cdev_info);
+if (cdev_info.sym.max_nb_sessions_per_qp > 0) {
+ret = rte_cryptodev_queue_pair_attach_sym_session(
+ipsec_ctx->tbl[cdev_id_qp].qp,
+sa->crypto_session);
+if (ret < 0) {
+RTE_LOG(ERR, IPSEC, "Session cannot be attached"
+" to qp %u ", ipsec_ctx->tbl[cdev_id_qp].qp);


Guideline is to keep error strings in single line to facilitate grep.
Other than that:

Acked-by: Sergio Gonzalez Monroy 


+return -1;
+}
+}
  sa->cdev_id_qp = cdev_id_qp;
return 0;





Hi Sergio,

Similar error string is mentioned above my change also in the same 
function. I deliberately did that as per the error strings in the file.


Thanks,
Akhil

Re: [dpdk-dev] [PATCH v7 1/3] ethdev: new API to free consumed buffers in Tx ring

[Olivier MATZ]

Hi Billy,

On Wed, 15 Mar 2017 14:02:24 -0400, Billy McFall  wrote:
> Add a new API to force free consumed buffers on Tx ring. API will return
> the number of packets freed (0-n) or error code if feature not supported
> (-ENOTSUP) or input invalid (-ENODEV).
> 
> Signed-off-by: Billy McFall 
> ---
>  doc/guides/conf.py  |  7 +--
>  doc/guides/nics/features/default.ini|  4 +++-
>  doc/guides/prog_guide/poll_mode_drv.rst | 28 
>  doc/guides/rel_notes/release_17_05.rst  |  7 ++-
>  lib/librte_ether/rte_ethdev.c   | 14 ++
>  lib/librte_ether/rte_ethdev.h   | 31 +++
>  6 files changed, 87 insertions(+), 4 deletions(-)
> 

[...]

> --- a/doc/guides/prog_guide/poll_mode_drv.rst
> +++ b/doc/guides/prog_guide/poll_mode_drv.rst
> @@ -249,6 +249,34 @@ One descriptor in the TX ring is used as a sentinel to 
> avoid a hardware race con
>  
>  When configuring for DCB operation, at port initialization, both the 
> number of transmit queues and the number of receive queues must be set to 128.
>  
> +Free Tx mbuf on Demand
> +~~
> +
> +Many of the drivers don't release the mbuf back to the mempool, or local 
> cache, immediately after the packet has been
> +transmitted.
> +Instead, they leave the mbuf in their Tx ring and either perform a bulk 
> release when the ``tx_rs_thresh`` has been
> +crossed or free the mbuf when a slot in the Tx ring is needed.
> +
> +An application can request the driver to release used mbufs with the 
> ``rte_eth_tx_done_cleanup()`` API.
> +This API requests the driver to release mbufs that are no longer in use, 
> independent of whether or not the
> +``tx_rs_thresh`` has been crossed.
> +There are two scenarios when an application may want the mbuf released 
> immediately:
> +
> +* When a given packet needs to be sent to multiple destination interfaces 
> (either for Layer 2 flooding or Layer 3
> +  multi-cast).
> +  One option is to make a copy of the packet or a copy of the header portion 
> that needs to be manipulated.
> +  A second option is to transmit the packet and then poll the 
> ``rte_eth_tx_done_cleanup()`` API until the reference
> +  count on the packet is decremented.
> +  Then the same packet can be transmitted to the next destination interface.

By reading this paragraph, it's not so clear to me that the packet
that will be transmitted on all interfaces will be different from
one port to another.

Maybe it could be reworded to insist on that?


> +
> +* If an application is designed to make multiple runs, like a packet 
> generator, and one run has completed.
> +  The application may want to reset to a clean state.

I'd reword into:

Some applications are designed to make multiple runs, like a packet generator.
Between each run, the application may want to reset to a clean state.

What do you mean by "clean state"? All mbufs returned into the mempools?
Why would a packet generator need that? For performance?

Also, do we want to ensure that all packets are actually transmitted?
Can we do that with this API or should we use another API like
rte_eth_tx_descriptor_status() [1] ?

[1] http://dpdk.org/dev/patchwork/patch/21549/


Thanks,
Olivier

Re: [dpdk-dev] [PATCH v2 2/2] examples/ipsec-secgw: attach session-qp

[Sergio Gonzalez Monroy]

On 23/03/2017 10:38, Akhil Goyal wrote:

On 3/23/2017 4:00 PM, Sergio Gonzalez Monroy wrote:

That was simpler than I thought.

For some reason I understood that the device will support thousands of
queues and single session per queue which would have needed more app
changes.

On 23/03/2017 08:06, akhil.go...@nxp.com wrote:

From: Akhil Goyal 

adding support for attaching session to queue pairs.
This is required as underlying crypto driver may only
support limited number of sessions per queue pair
if max_nb_sessions_per_qp > 0, session should be
attached to a particular qp.

Signed-off-by: Akhil Goyal 
---
  examples/ipsec-secgw/ipsec.c | 12 
  1 file changed, 12 insertions(+)

diff --git a/examples/ipsec-secgw/ipsec.c 
b/examples/ipsec-secgw/ipsec.c

index 144f0aa..b35b30f 100644
--- a/examples/ipsec-secgw/ipsec.c
+++ b/examples/ipsec-secgw/ipsec.c
@@ -47,6 +47,7 @@
  static inline int
  create_session(struct ipsec_ctx *ipsec_ctx __rte_unused, struct
ipsec_sa *sa)
  {
+struct rte_cryptodev_info cdev_info;
  unsigned long cdev_id_qp = 0;
  int32_t ret;
  struct cdev_key key = { 0 };
@@ -73,6 +74,17 @@ create_session(struct ipsec_ctx *ipsec_ctx
__rte_unused, struct ipsec_sa *sa)
  sa->crypto_session = rte_cryptodev_sym_session_create(
  ipsec_ctx->tbl[cdev_id_qp].id, sa->xforms);
  + rte_cryptodev_info_get(ipsec_ctx->tbl[cdev_id_qp].id, &cdev_info);
+if (cdev_info.sym.max_nb_sessions_per_qp > 0) {
+ret = rte_cryptodev_queue_pair_attach_sym_session(
+ipsec_ctx->tbl[cdev_id_qp].qp,
+sa->crypto_session);
+if (ret < 0) {
+RTE_LOG(ERR, IPSEC, "Session cannot be attached"
+" to qp %u ", ipsec_ctx->tbl[cdev_id_qp].qp);


Guideline is to keep error strings in single line to facilitate grep.
Other than that:

Acked-by: Sergio Gonzalez Monroy 


+return -1;
+}
+}
  sa->cdev_id_qp = cdev_id_qp;
return 0;





Hi Sergio,

Similar error string is mentioned above my change also in the same 
function. I deliberately did that as per the error strings in the file.


Thanks,
Akhil



That means that we need to update those strings too, but new code should 
try to keep error string in single line.


Thanks,
Sergio

[dpdk-dev] [PATCH v2 0/4] Rework tunnel filter functions

[Beilei Xing]

1. Rework tunnel filter functions to align with the new
   added cloud filer command buffer structure.
2. Support tunnel filter to VF for consistent filter API.

v2 changes:
 Remove replace cloud filter function as it's in share code.

Beilei Xing (4):
  net/i40e: rework tunnel filter functions
  net/i40e: change tunnel filter function name
  net/i40e: support tunnel filter to VF
  net/i40e: refine consistent tunnel filter

 drivers/net/i40e/i40e_ethdev.c | 245 +++--
 drivers/net/i40e/i40e_ethdev.h |  47 
 drivers/net/i40e/i40e_flow.c   | 142 +---
 3 files changed, 336 insertions(+), 98 deletions(-)

-- 
2.5.5

[dpdk-dev] [PATCH v2 1/4] net/i40e: rework tunnel filter functions

[Beilei Xing]

Rework tunnel filter functions to align with the
new command buffer for add/remove cloud filter.

Signed-off-by: Beilei Xing 
---
 drivers/net/i40e/i40e_ethdev.c | 100 +++--
 drivers/net/i40e/i40e_ethdev.h |   1 +
 drivers/net/i40e/i40e_flow.c   |  28 +++-
 3 files changed, 86 insertions(+), 43 deletions(-)

diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index 9c0441b..5e6cc59 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -410,7 +410,7 @@ static int i40e_sw_ethertype_filter_insert(struct i40e_pf 
*pf,
   struct i40e_ethertype_filter *filter);
 
 static int i40e_tunnel_filter_convert(
-   struct i40e_aqc_add_remove_cloud_filters_element_data *cld_filter,
+   struct i40e_aqc_add_rm_cloud_filt_elem_ext *cld_filter,
struct i40e_tunnel_filter *tunnel_filter);
 static int i40e_sw_tunnel_filter_insert(struct i40e_pf *pf,
struct i40e_tunnel_filter *tunnel_filter);
@@ -6717,24 +6717,27 @@ i40e_dev_get_filter_type(uint16_t filter_type, uint16_t 
*flag)
 
 /* Convert tunnel filter structure */
 static int
-i40e_tunnel_filter_convert(struct 
i40e_aqc_add_remove_cloud_filters_element_data
-  *cld_filter,
-  struct i40e_tunnel_filter *tunnel_filter)
+i40e_tunnel_filter_convert(
+   struct i40e_aqc_add_rm_cloud_filt_elem_ext *cld_filter,
+   struct i40e_tunnel_filter *tunnel_filter)
 {
-   ether_addr_copy((struct ether_addr *)&cld_filter->outer_mac,
+   ether_addr_copy((struct ether_addr *)&cld_filter->element.outer_mac,
(struct ether_addr *)&tunnel_filter->input.outer_mac);
-   ether_addr_copy((struct ether_addr *)&cld_filter->inner_mac,
+   ether_addr_copy((struct ether_addr *)&cld_filter->element.inner_mac,
(struct ether_addr *)&tunnel_filter->input.inner_mac);
-   tunnel_filter->input.inner_vlan = cld_filter->inner_vlan;
-   if ((rte_le_to_cpu_16(cld_filter->flags) &
+   tunnel_filter->input.inner_vlan = cld_filter->element.inner_vlan;
+   if ((rte_le_to_cpu_16(cld_filter->element.flags) &
 I40E_AQC_ADD_CLOUD_FLAGS_IPV6) ==
I40E_AQC_ADD_CLOUD_FLAGS_IPV6)
tunnel_filter->input.ip_type = I40E_TUNNEL_IPTYPE_IPV6;
else
tunnel_filter->input.ip_type = I40E_TUNNEL_IPTYPE_IPV4;
-   tunnel_filter->input.flags = cld_filter->flags;
-   tunnel_filter->input.tenant_id = cld_filter->tenant_id;
-   tunnel_filter->queue = cld_filter->queue_number;
+   tunnel_filter->input.flags = cld_filter->element.flags;
+   tunnel_filter->input.tenant_id = cld_filter->element.tenant_id;
+   tunnel_filter->queue = cld_filter->element.queue_number;
+   rte_memcpy(tunnel_filter->input.general_fields,
+  cld_filter->general_fields,
+  sizeof(cld_filter->general_fields));
 
return 0;
 }
@@ -6813,15 +6816,16 @@ i40e_dev_tunnel_filter_set(struct i40e_pf *pf,
int val, ret = 0;
struct i40e_hw *hw = I40E_PF_TO_HW(pf);
struct i40e_vsi *vsi = pf->main_vsi;
-   struct i40e_aqc_add_remove_cloud_filters_element_data  *cld_filter;
-   struct i40e_aqc_add_remove_cloud_filters_element_data  *pfilter;
+   struct i40e_aqc_add_rm_cloud_filt_elem_ext *cld_filter;
+   struct i40e_aqc_add_rm_cloud_filt_elem_ext *pfilter;
struct i40e_tunnel_rule *tunnel_rule = &pf->tunnel;
struct i40e_tunnel_filter *tunnel, *node;
struct i40e_tunnel_filter check_filter; /* Check if filter exists */
+   bool big_buffer = 0;
 
cld_filter = rte_zmalloc("tunnel_filter",
-   sizeof(struct i40e_aqc_add_remove_cloud_filters_element_data),
-   0);
+sizeof(struct i40e_aqc_add_rm_cloud_filt_elem_ext),
+   0);
 
if (NULL == cld_filter) {
PMD_DRV_LOG(ERR, "Failed to alloc memory.");
@@ -6829,24 +6833,28 @@ i40e_dev_tunnel_filter_set(struct i40e_pf *pf,
}
pfilter = cld_filter;
 
-   ether_addr_copy(&tunnel_filter->outer_mac, (struct 
ether_addr*)&pfilter->outer_mac);
-   ether_addr_copy(&tunnel_filter->inner_mac, (struct 
ether_addr*)&pfilter->inner_mac);
+   ether_addr_copy(&tunnel_filter->outer_mac,
+   (struct ether_addr *)&pfilter->element.outer_mac);
+   ether_addr_copy(&tunnel_filter->inner_mac,
+   (struct ether_addr *)&pfilter->element.inner_mac);
 
-   pfilter->inner_vlan = rte_cpu_to_le_16(tunnel_filter->inner_vlan);
+   pfilter->element.inner_vlan =
+   rte_cpu_to_le_16(tunnel_filter->inner_vlan);
if (tunnel_filter->ip_type == RTE_TUNNEL_IPTYPE_IPV4) {
ip_type = I40E_AQC_ADD_CLOUD_FLAGS_IPV4;
ipv4_addr = rte_be_to_cpu_32(tunnel_filter->ip_addr.ipv4_addr);
-

[dpdk-dev] [PATCH v2 4/4] net/i40e: refine consistent tunnel filter

[Beilei Xing]

Add i40e_tunnel_type enumeration type to refine consistent
tunnel filter, it will be esay to add new tunnel type for
i40e.

Signed-off-by: Beilei Xing 
---
 drivers/net/i40e/i40e_ethdev.c |  8 
 drivers/net/i40e/i40e_ethdev.h | 18 --
 drivers/net/i40e/i40e_flow.c   |  6 +++---
 3 files changed, 23 insertions(+), 9 deletions(-)

diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index 2b3d41b..506b957 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -6973,7 +6973,7 @@ i40e_dev_consistent_tunnel_filter_set(struct i40e_pf *pf,
 
pfilter->element.inner_vlan =
rte_cpu_to_le_16(tunnel_filter->inner_vlan);
-   if (tunnel_filter->ip_type == RTE_TUNNEL_IPTYPE_IPV4) {
+   if (tunnel_filter->ip_type == I40E_TUNNEL_IPTYPE_IPV4) {
ip_type = I40E_AQC_ADD_CLOUD_FLAGS_IPV4;
ipv4_addr = rte_be_to_cpu_32(tunnel_filter->ip_addr.ipv4_addr);
rte_memcpy(&pfilter->element.ipaddr.v4.data,
@@ -6993,13 +6993,13 @@ i40e_dev_consistent_tunnel_filter_set(struct i40e_pf 
*pf,
 
/* check tunneled type */
switch (tunnel_filter->tunnel_type) {
-   case RTE_TUNNEL_TYPE_VXLAN:
+   case I40E_TUNNEL_TYPE_VXLAN:
tun_type = I40E_AQC_ADD_CLOUD_TNL_TYPE_VXLAN;
break;
-   case RTE_TUNNEL_TYPE_NVGRE:
+   case I40E_TUNNEL_TYPE_NVGRE:
tun_type = I40E_AQC_ADD_CLOUD_TNL_TYPE_NVGRE_OMAC;
break;
-   case RTE_TUNNEL_TYPE_IP_IN_GRE:
+   case I40E_TUNNEL_TYPE_IP_IN_GRE:
tun_type = I40E_AQC_ADD_CLOUD_TNL_TYPE_IP;
break;
default:
diff --git a/drivers/net/i40e/i40e_ethdev.h b/drivers/net/i40e/i40e_ethdev.h
index d976f7a..7d86ca2 100644
--- a/drivers/net/i40e/i40e_ethdev.h
+++ b/drivers/net/i40e/i40e_ethdev.h
@@ -532,6 +532,20 @@ struct i40e_tunnel_rule {
 };
 
 /**
+ * Tunnel type.
+ */
+enum i40e_tunnel_type {
+   I40E_TUNNEL_TYPE_NONE = 0,
+   I40E_TUNNEL_TYPE_VXLAN,
+   I40E_TUNNEL_TYPE_GENEVE,
+   I40E_TUNNEL_TYPE_TEREDO,
+   I40E_TUNNEL_TYPE_NVGRE,
+   I40E_TUNNEL_TYPE_IP_IN_GRE,
+   I40E_L2_TUNNEL_TYPE_E_TAG,
+   I40E_TUNNEL_TYPE_MAX,
+};
+
+/**
  * Tunneling Packet filter configuration.
  */
 struct i40e_tunnel_filter_conf {
@@ -539,7 +553,7 @@ struct i40e_tunnel_filter_conf {
struct ether_addr inner_mac;/**< Inner MAC address to match. */
uint16_t inner_vlan;/**< Inner VLAN to match. */
uint32_t outer_vlan;/**< Outer VLAN to match */
-   enum rte_tunnel_iptype ip_type; /**< IP address type. */
+   enum i40e_tunnel_iptype ip_type; /**< IP address type. */
/**
 * Outer destination IP address to match if ETH_TUNNEL_FILTER_OIP
 * is set in filter_type, or inner destination IP address to match
@@ -551,7 +565,7 @@ struct i40e_tunnel_filter_conf {
} ip_addr;
/** Flags from ETH_TUNNEL_FILTER_XX - see above. */
uint16_t filter_type;
-   enum rte_eth_tunnel_type tunnel_type; /**< Tunnel Type. */
+   enum i40e_tunnel_type tunnel_type; /**< Tunnel Type. */
uint32_t tenant_id; /**< Tenant ID to match. VNI, GRE key... */
uint16_t queue_id;  /**< Queue assigned to if match. */
uint8_t is_to_vf;   /**< 0 - to PF, 1 - to VF */
diff --git a/drivers/net/i40e/i40e_flow.c b/drivers/net/i40e/i40e_flow.c
index 449299a..ffc14a9 100644
--- a/drivers/net/i40e/i40e_flow.c
+++ b/drivers/net/i40e/i40e_flow.c
@@ -1324,7 +1324,7 @@ i40e_flow_parse_vxlan_pattern(__rte_unused struct 
rte_eth_dev *dev,
}
break;
case RTE_FLOW_ITEM_TYPE_IPV4:
-   filter->ip_type = RTE_TUNNEL_IPTYPE_IPV4;
+   filter->ip_type = I40E_TUNNEL_IPTYPE_IPV4;
/* IPv4 is used to describe protocol,
 * spec and mask should be NULL.
 */
@@ -1337,7 +1337,7 @@ i40e_flow_parse_vxlan_pattern(__rte_unused struct 
rte_eth_dev *dev,
}
break;
case RTE_FLOW_ITEM_TYPE_IPV6:
-   filter->ip_type = RTE_TUNNEL_IPTYPE_IPV6;
+   filter->ip_type = I40E_TUNNEL_IPTYPE_IPV6;
/* IPv6 is used to describe protocol,
 * spec and mask should be NULL.
 */
@@ -1480,7 +1480,7 @@ i40e_flow_parse_vxlan_pattern(__rte_unused struct 
rte_eth_dev *dev,
return -rte_errno;
}
 
-   filter->tunnel_type = RTE_TUNNEL_TYPE_VXLAN;
+   filter->tunnel_type = I40E_TUNNEL_TYPE_VXLAN;
 
return 0;
 }
-- 
2.5.5

[dpdk-dev] [PATCH v2 3/4] net/i40e: support tunnel filter to VF

[Beilei Xing]

Previously, only tunnel filter to PF is supported.
This patch adds i40e_dev_consistent_tunnel_filter_set
function for consistent filter API to support tunnel
filter to VF.

Signed-off-by: Beilei Xing 
---
 drivers/net/i40e/i40e_ethdev.c | 145 +
 drivers/net/i40e/i40e_ethdev.h |  32 +
 drivers/net/i40e/i40e_flow.c   |  50 +-
 3 files changed, 212 insertions(+), 15 deletions(-)

diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index 5e6cc59..2b3d41b 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -6935,6 +6935,151 @@ i40e_dev_tunnel_filter_set(struct i40e_pf *pf,
return ret;
 }
 
+int
+i40e_dev_consistent_tunnel_filter_set(struct i40e_pf *pf,
+ struct i40e_tunnel_filter_conf *tunnel_filter,
+ uint8_t add)
+{
+   uint16_t ip_type;
+   uint32_t ipv4_addr;
+   uint8_t i, tun_type = 0;
+   /* internal variable to convert ipv6 byte order */
+   uint32_t convert_ipv6[4];
+   int val, ret = 0;
+   struct i40e_pf_vf *vf = NULL;
+   struct i40e_hw *hw = I40E_PF_TO_HW(pf);
+   struct i40e_vsi *vsi;
+   struct i40e_aqc_add_rm_cloud_filt_elem_ext *cld_filter;
+   struct i40e_aqc_add_rm_cloud_filt_elem_ext *pfilter;
+   struct i40e_tunnel_rule *tunnel_rule = &pf->tunnel;
+   struct i40e_tunnel_filter *tunnel, *node;
+   struct i40e_tunnel_filter check_filter; /* Check if filter exists */
+   bool big_buffer = 0;
+
+   cld_filter = rte_zmalloc("tunnel_filter",
+sizeof(struct i40e_aqc_add_rm_cloud_filt_elem_ext),
+0);
+
+   if (cld_filter == NULL) {
+   PMD_DRV_LOG(ERR, "Failed to alloc memory.");
+   return -EINVAL;
+   }
+   pfilter = cld_filter;
+
+   ether_addr_copy(&tunnel_filter->outer_mac,
+   (struct ether_addr *)&pfilter->element.outer_mac);
+   ether_addr_copy(&tunnel_filter->inner_mac,
+   (struct ether_addr *)&pfilter->element.inner_mac);
+
+   pfilter->element.inner_vlan =
+   rte_cpu_to_le_16(tunnel_filter->inner_vlan);
+   if (tunnel_filter->ip_type == RTE_TUNNEL_IPTYPE_IPV4) {
+   ip_type = I40E_AQC_ADD_CLOUD_FLAGS_IPV4;
+   ipv4_addr = rte_be_to_cpu_32(tunnel_filter->ip_addr.ipv4_addr);
+   rte_memcpy(&pfilter->element.ipaddr.v4.data,
+   &rte_cpu_to_le_32(ipv4_addr),
+   sizeof(pfilter->element.ipaddr.v4.data));
+   } else {
+   ip_type = I40E_AQC_ADD_CLOUD_FLAGS_IPV6;
+   for (i = 0; i < 4; i++) {
+   convert_ipv6[i] =
+   rte_cpu_to_le_32(rte_be_to_cpu_32(
+tunnel_filter->ip_addr.ipv6_addr[i]));
+   }
+   rte_memcpy(&pfilter->element.ipaddr.v6.data,
+  &convert_ipv6,
+  sizeof(pfilter->element.ipaddr.v6.data));
+   }
+
+   /* check tunneled type */
+   switch (tunnel_filter->tunnel_type) {
+   case RTE_TUNNEL_TYPE_VXLAN:
+   tun_type = I40E_AQC_ADD_CLOUD_TNL_TYPE_VXLAN;
+   break;
+   case RTE_TUNNEL_TYPE_NVGRE:
+   tun_type = I40E_AQC_ADD_CLOUD_TNL_TYPE_NVGRE_OMAC;
+   break;
+   case RTE_TUNNEL_TYPE_IP_IN_GRE:
+   tun_type = I40E_AQC_ADD_CLOUD_TNL_TYPE_IP;
+   break;
+   default:
+   /* Other tunnel types is not supported. */
+   PMD_DRV_LOG(ERR, "tunnel type is not supported.");
+   rte_free(cld_filter);
+   return -EINVAL;
+   }
+
+   val = i40e_dev_get_filter_type(tunnel_filter->filter_type,
+  &pfilter->element.flags);
+   if (val < 0) {
+   rte_free(cld_filter);
+   return -EINVAL;
+   }
+
+   pfilter->element.flags |= rte_cpu_to_le_16(
+   I40E_AQC_ADD_CLOUD_FLAGS_TO_QUEUE |
+   ip_type | (tun_type << I40E_AQC_ADD_CLOUD_TNL_TYPE_SHIFT));
+   pfilter->element.tenant_id = rte_cpu_to_le_32(tunnel_filter->tenant_id);
+   pfilter->element.queue_number =
+   rte_cpu_to_le_16(tunnel_filter->queue_id);
+
+   if (!tunnel_filter->is_to_vf)
+   vsi = pf->main_vsi;
+   else {
+   if (tunnel_filter->vf_id >= pf->vf_num) {
+   PMD_DRV_LOG(ERR, "Invalid argument.");
+   return -EINVAL;
+   }
+   vf = &pf->vfs[tunnel_filter->vf_id];
+   vsi = vf->vsi;
+   }
+
+   /* Check if there is the filter in SW list */
+   memset(&check_filter, 0, sizeof(check_filter));
+   i40e_tunnel_filter_convert(cld_filter, &check_filter);
+   node = i40e_sw_tunnel_filter_lookup(tunnel_rule, &c

[dpdk-dev] [PATCH v2 2/4] net/i40e: change tunnel filter function name

[Beilei Xing]

Change tunnel filter function name to VXLAN filter
function, prepare for other tunnel filter function.

Signed-off-by: Beilei Xing 
---
 drivers/net/i40e/i40e_flow.c | 58 
 1 file changed, 21 insertions(+), 37 deletions(-)

diff --git a/drivers/net/i40e/i40e_flow.c b/drivers/net/i40e/i40e_flow.c
index 5513c3c..412cd22 100644
--- a/drivers/net/i40e/i40e_flow.c
+++ b/drivers/net/i40e/i40e_flow.c
@@ -90,10 +90,6 @@ static int i40e_flow_parse_fdir_action(struct rte_eth_dev 
*dev,
   const struct rte_flow_action *actions,
   struct rte_flow_error *error,
   struct rte_eth_fdir_filter *filter);
-static int i40e_flow_parse_tunnel_pattern(__rte_unused struct rte_eth_dev *dev,
- const struct rte_flow_item *pattern,
- struct rte_flow_error *error,
- struct rte_eth_tunnel_filter_conf *filter);
 static int i40e_flow_parse_tunnel_action(struct rte_eth_dev *dev,
 const struct rte_flow_action *actions,
 struct rte_flow_error *error,
@@ -112,12 +108,12 @@ static int i40e_flow_parse_fdir_filter(struct rte_eth_dev 
*dev,
   const struct rte_flow_action actions[],
   struct rte_flow_error *error,
   union i40e_filter_t *filter);
-static int i40e_flow_parse_tunnel_filter(struct rte_eth_dev *dev,
-const struct rte_flow_attr *attr,
-const struct rte_flow_item pattern[],
-const struct rte_flow_action actions[],
-struct rte_flow_error *error,
-union i40e_filter_t *filter);
+static int i40e_flow_parse_vxlan_filter(struct rte_eth_dev *dev,
+   const struct rte_flow_attr *attr,
+   const struct rte_flow_item pattern[],
+   const struct rte_flow_action actions[],
+   struct rte_flow_error *error,
+   union i40e_filter_t *filter);
 static int i40e_flow_destroy_ethertype_filter(struct i40e_pf *pf,
  struct i40e_ethertype_filter *filter);
 static int i40e_flow_destroy_tunnel_filter(struct i40e_pf *pf,
@@ -302,11 +298,11 @@ static struct i40e_valid_pattern 
i40e_supported_patterns[] = {
{ pattern_fdir_ipv6_tcp_ext, i40e_flow_parse_fdir_filter },
{ pattern_fdir_ipv6_sctp, i40e_flow_parse_fdir_filter },
{ pattern_fdir_ipv6_sctp_ext, i40e_flow_parse_fdir_filter },
-   /* tunnel */
-   { pattern_vxlan_1, i40e_flow_parse_tunnel_filter },
-   { pattern_vxlan_2, i40e_flow_parse_tunnel_filter },
-   { pattern_vxlan_3, i40e_flow_parse_tunnel_filter },
-   { pattern_vxlan_4, i40e_flow_parse_tunnel_filter },
+   /* VXLAN */
+   { pattern_vxlan_1, i40e_flow_parse_vxlan_filter },
+   { pattern_vxlan_2, i40e_flow_parse_vxlan_filter },
+   { pattern_vxlan_3, i40e_flow_parse_vxlan_filter },
+   { pattern_vxlan_4, i40e_flow_parse_vxlan_filter },
 };
 
 #define NEXT_ITEM_OF_ACTION(act, actions, index)\
@@ -1205,7 +1201,8 @@ i40e_check_tenant_id_mask(const uint8_t *mask)
  *filled with 0.
  */
 static int
-i40e_flow_parse_vxlan_pattern(const struct rte_flow_item *pattern,
+i40e_flow_parse_vxlan_pattern(__rte_unused struct rte_eth_dev *dev,
+ const struct rte_flow_item *pattern,
  struct rte_flow_error *error,
  struct rte_eth_tunnel_filter_conf *filter)
 {
@@ -1469,32 +1466,19 @@ i40e_flow_parse_vxlan_pattern(const struct 
rte_flow_item *pattern,
 }
 
 static int
-i40e_flow_parse_tunnel_pattern(__rte_unused struct rte_eth_dev *dev,
-  const struct rte_flow_item *pattern,
-  struct rte_flow_error *error,
-  struct rte_eth_tunnel_filter_conf *filter)
-{
-   int ret;
-
-   ret = i40e_flow_parse_vxlan_pattern(pattern, error, filter);
-
-   return ret;
-}
-
-static int
-i40e_flow_parse_tunnel_filter(struct rte_eth_dev *dev,
- const struct rte_flow_attr *attr,
- const struct rte_flow_item pattern[],
- const struct rte_flow_action actions[],
- struct rte_flow_error *error,
- union i40e_filter_t *filter)
+i40e_flow_parse_vxlan_filter(struct rte_eth_dev *dev,
+const struct rte_flow_attr *attr,
+   

[dpdk-dev] Minutes of tech-board meeting 2017-03-20

[Ananyev, Konstantin]

Hi,

Below are the minutes of the last DPDK technical board meeting that occurred at 
2017-03-20.

Please note that meetings are open to all to attend.
The next meeting is planned for the first week of April, any topics to be 
referred
to the tech board for discussion at that meeting should be emailed to 
techbo...@dpdk.org
by March 31th at the latest.

Member attendees:

- Bruce Richardson
- Hemant Agrawal
- Jan Blunck
- Jerin Jacob
- Konstantin Ananyev
- Olivier Matz
- Stephen Hemminger
- Thomas Monjalon
- Yuanhan Liu

1. General Meeting Admin


 Jerin Jacob volunteered to chair the next meeting.

 No particular date was agreed for next meeting (proposed 3/5th of April 3PM 
UTC).

 Jerin to follow on mail to select a date/time and send a reminder with agenda 
a few days in advance.

2. AVP PMD
---

TB has no major concerns about AVP PMD to be up-streamed into dpdk.org.
TB recommends  to provide a good documentation introduction for the AVP PMD and 
its scope.
It should be clearly stated  that right now AVP PMD is limited to use with 
WindRiver hypervisor
and for most cases virtio is still a preferred virtual device to use with QEMU.

3. Failsafe PMD
---

TB recommendation:
 - For 17.05 try to implement hotplug support at EAL level
-  For 17.08 reconsider failsafe mode.

4. PRGDEV
-

TB opinion is that it is not ready for inclusion right now.
The exact requirements should be clearly described, so any member can 
participate
in the discussion and propose other ways to achieve the same goal.
TB also recommends to include some good use case(s) into the proposal.

5. Traffic management (Hierarchical QoS scheduler)
---

Include into next-tree with Cristian Dumitrescu as a maintainer .
Target release to merge into dpdk.org mainline: 17.08.
Gating condition: 1+ HW device support.
 
6. Divergence between DPDK/Linux PF/VF implementations


Discussion postponed until next TB meeting.

[dpdk-dev] [PATCH v2 0/3] add support for MPLS tunnel filter

[Beilei Xing]

This patchset adds support for MPLSoGRE and MPLSoUDP
tunnel filters.
I40e NICs can't recongnize MPLS tunnel packets by
default, so need to load a profile to FW first, then
MPLS tunnel packets can be recongnized with packet
type 61 and 63.
It depends on pipeline personalization profile support
and changes of tunnel filter function.

Beilei Xing (3):
  app/testpmd: add support for MPLS and GRE items
  net/i40e: add MPLS parsing function
  net/i40e: enable tunnel filter for MPLS

 app/test-pmd/cmdline_flow.c |  46 +
 app/test-pmd/config.c   |   2 +
 doc/guides/prog_guide/rte_flow.rst  |  20 +-
 doc/guides/testpmd_app_ug/testpmd_funcs.rst |   8 +
 drivers/net/i40e/i40e_ethdev.c  |  45 +++-
 drivers/net/i40e/i40e_ethdev.h  |  12 ++
 drivers/net/i40e/i40e_flow.c| 307 
 lib/librte_ether/rte_flow.h |  45 
 8 files changed, 478 insertions(+), 7 deletions(-)

-- 
2.5.5

[dpdk-dev] [PATCH v2 1/3] app/testpmd: add support for MPLS and GRE items

[Beilei Xing]

This patch adds sopport for MPLS and GRE items to generic filter
API.

Signed-off-by: Beilei Xing 
---
 app/test-pmd/cmdline_flow.c | 46 +
 app/test-pmd/config.c   |  2 ++
 doc/guides/prog_guide/rte_flow.rst  | 20 +++--
 doc/guides/testpmd_app_ug/testpmd_funcs.rst |  8 +
 lib/librte_ether/rte_flow.h | 45 
 5 files changed, 119 insertions(+), 2 deletions(-)

diff --git a/app/test-pmd/cmdline_flow.c b/app/test-pmd/cmdline_flow.c
index ff98690..ff1e47c 100644
--- a/app/test-pmd/cmdline_flow.c
+++ b/app/test-pmd/cmdline_flow.c
@@ -159,6 +159,10 @@ enum index {
ITEM_SCTP_CKSUM,
ITEM_VXLAN,
ITEM_VXLAN_VNI,
+   ITEM_MPLS,
+   ITEM_MPLS_LABEL,
+   ITEM_GRE,
+   ITEM_GRE_PROTO,
 
/* Validate/create actions. */
ACTIONS,
@@ -432,6 +436,8 @@ static const enum index next_item[] = {
ITEM_TCP,
ITEM_SCTP,
ITEM_VXLAN,
+   ITEM_MPLS,
+   ITEM_GRE,
ZERO,
 };
 
@@ -538,6 +544,18 @@ static const enum index item_vxlan[] = {
ZERO,
 };
 
+static const enum index item_mpls[] = {
+   ITEM_MPLS_LABEL,
+   ITEM_NEXT,
+   ZERO,
+};
+
+static const enum index item_gre[] = {
+   ITEM_GRE_PROTO,
+   ITEM_NEXT,
+   ZERO,
+};
+
 static const enum index next_action[] = {
ACTION_END,
ACTION_VOID,
@@ -1279,6 +1297,34 @@ static const struct token token_list[] = {
.next = NEXT(item_vxlan, NEXT_ENTRY(UNSIGNED), item_param),
.args = ARGS(ARGS_ENTRY_HTON(struct rte_flow_item_vxlan, vni)),
},
+   [ITEM_MPLS] = {
+   .name = "mpls",
+   .help = "match MPLS header",
+   .priv = PRIV_ITEM(MPLS, sizeof(struct rte_flow_item_mpls)),
+   .next = NEXT(item_mpls),
+   .call = parse_vc,
+   },
+   [ITEM_MPLS_LABEL] = {
+   .name = "label",
+   .help = "MPLS label",
+   .next = NEXT(item_mpls, NEXT_ENTRY(UNSIGNED), item_param),
+   .args = ARGS(ARGS_ENTRY_HTON(struct rte_flow_item_mpls,
+label_tc_s_ttl)),
+   },
+   [ITEM_GRE] = {
+   .name = "gre",
+   .help = "match GRE header",
+   .priv = PRIV_ITEM(GRE, sizeof(struct rte_flow_item_gre)),
+   .next = NEXT(item_gre),
+   .call = parse_vc,
+   },
+   [ITEM_GRE_PROTO] = {
+   .name = "protocol",
+   .help = "GRE protocol type",
+   .next = NEXT(item_gre, NEXT_ENTRY(UNSIGNED), item_param),
+   .args = ARGS(ARGS_ENTRY_HTON(struct rte_flow_item_gre,
+protocol)),
+   },
/* Validate/create actions. */
[ACTIONS] = {
.name = "actions",
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index eb3d572..90d153e 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -963,6 +963,8 @@ static const struct {
MK_FLOW_ITEM(TCP, sizeof(struct rte_flow_item_tcp)),
MK_FLOW_ITEM(SCTP, sizeof(struct rte_flow_item_sctp)),
MK_FLOW_ITEM(VXLAN, sizeof(struct rte_flow_item_vxlan)),
+   MK_FLOW_ITEM(MPLS, sizeof(struct rte_flow_item_mpls)),
+   MK_FLOW_ITEM(GRE, sizeof(struct rte_flow_item_gre)),
 };
 
 /** Compute storage space needed by item specification. */
diff --git a/doc/guides/prog_guide/rte_flow.rst 
b/doc/guides/prog_guide/rte_flow.rst
index 3bf8871..45897cd 100644
--- a/doc/guides/prog_guide/rte_flow.rst
+++ b/doc/guides/prog_guide/rte_flow.rst
@@ -187,8 +187,8 @@ Pattern item
 Pattern items fall in two categories:
 
 - Matching protocol headers and packet data (ANY, RAW, ETH, VLAN, IPV4,
-  IPV6, ICMP, UDP, TCP, SCTP, VXLAN and so on), usually associated with a
-  specification structure.
+  IPV6, ICMP, UDP, TCP, SCTP, VXLAN, MPLS, GRE and so on), usually
+  associated with a specification structure.
 
 - Matching meta-data or affecting pattern processing (END, VOID, INVERT, PF,
   VF, PORT and so on), often without a specification structure.
@@ -863,6 +863,22 @@ Matches a VXLAN header (RFC 7348).
 - ``rsvd1``: reserved, normally 0x00.
 - Default ``mask`` matches VNI only.
 
+Item: ``MPLS``
+^^
+
+Matches a MPLS header.
+
+- ``label_tc_s_ttl``: Lable, TC, Bottom of Stack and TTL.
+- Default ``mask`` matches label only.
+
+Item: ``GRE``
+^^
+
+Matches a GRE header.
+
+- ``c_rsvd0_ver``: Checksum, reserved 0 and version.
+- ``protocol``: Protocol type.
+
 Actions
 ~~~
 
diff --git a/doc/guides/testpmd_app_ug/testpmd_funcs.rst 
b/doc/guides/testpmd_app_ug/testpmd_funcs.rst
index bdc6a14..f9fa5d6 100644
--- a/doc/guides/testpmd_app_ug/testpmd_funcs.rst
+++ b/doc/guides/testpmd_app_ug/testpmd_funcs.rst
@@ -2453,6 +2453,14 @@ This section lists supported pattern items and their 
attribu

[dpdk-dev] [PATCH v2 2/3] net/i40e: add MPLS parsing function

[Beilei Xing]

This patch add MPLS parsing function to support
MPLS filtering.

Signed-off-by: Beilei Xing 
---
 drivers/net/i40e/i40e_ethdev.h |   5 +
 drivers/net/i40e/i40e_flow.c   | 217 +
 2 files changed, 222 insertions(+)

diff --git a/drivers/net/i40e/i40e_ethdev.h b/drivers/net/i40e/i40e_ethdev.h
index 7d86ca2..df345b1 100644
--- a/drivers/net/i40e/i40e_ethdev.h
+++ b/drivers/net/i40e/i40e_ethdev.h
@@ -500,6 +500,9 @@ struct i40e_ethertype_rule {
 /* Tunnel filter number HW supports */
 #define I40E_MAX_TUNNEL_FILTER_NUM 400
 
+#define I40E_AQC_ADD_CLOUD_TNL_TYPE_MPLSoUDP 0x11
+#define I40E_AQC_ADD_CLOUD_TNL_TYPE_MPLSoGRE 0x12
+
 enum i40e_tunnel_iptype {
I40E_TUNNEL_IPTYPE_IPV4,
I40E_TUNNEL_IPTYPE_IPV6,
@@ -542,6 +545,8 @@ enum i40e_tunnel_type {
I40E_TUNNEL_TYPE_NVGRE,
I40E_TUNNEL_TYPE_IP_IN_GRE,
I40E_L2_TUNNEL_TYPE_E_TAG,
+   I40E_TUNNEL_TYPE_MPLSoUDP,
+   I40E_TUNNEL_TYPE_MPLSoGRE,
I40E_TUNNEL_TYPE_MAX,
 };
 
diff --git a/drivers/net/i40e/i40e_flow.c b/drivers/net/i40e/i40e_flow.c
index ffc14a9..eefb15b 100644
--- a/drivers/net/i40e/i40e_flow.c
+++ b/drivers/net/i40e/i40e_flow.c
@@ -57,6 +57,7 @@
 #define I40E_IPV6_FRAG_HEADER  44
 #define I40E_TENANT_ARRAY_NUM  3
 #define I40E_TCI_MASK  0x
+#define I40E_MPLS_LABEL_MASK   0xF
 
 static int i40e_flow_validate(struct rte_eth_dev *dev,
  const struct rte_flow_attr *attr,
@@ -114,6 +115,12 @@ static int i40e_flow_parse_vxlan_filter(struct rte_eth_dev 
*dev,
const struct rte_flow_action actions[],
struct rte_flow_error *error,
union i40e_filter_t *filter);
+static int i40e_flow_parse_mpls_filter(struct rte_eth_dev *dev,
+  const struct rte_flow_attr *attr,
+  const struct rte_flow_item pattern[],
+  const struct rte_flow_action actions[],
+  struct rte_flow_error *error,
+  union i40e_filter_t *filter);
 static int i40e_flow_destroy_ethertype_filter(struct i40e_pf *pf,
  struct i40e_ethertype_filter *filter);
 static int i40e_flow_destroy_tunnel_filter(struct i40e_pf *pf,
@@ -278,6 +285,39 @@ static enum rte_flow_item_type pattern_vxlan_4[] = {
RTE_FLOW_ITEM_TYPE_END,
 };
 
+/* Pattern matched MPLS */
+static enum rte_flow_item_type pattern_mpls_1[] = {
+   RTE_FLOW_ITEM_TYPE_ETH,
+   RTE_FLOW_ITEM_TYPE_IPV4,
+   RTE_FLOW_ITEM_TYPE_UDP,
+   RTE_FLOW_ITEM_TYPE_MPLS,
+   RTE_FLOW_ITEM_TYPE_END,
+};
+
+static enum rte_flow_item_type pattern_mpls_2[] = {
+   RTE_FLOW_ITEM_TYPE_ETH,
+   RTE_FLOW_ITEM_TYPE_IPV6,
+   RTE_FLOW_ITEM_TYPE_UDP,
+   RTE_FLOW_ITEM_TYPE_MPLS,
+   RTE_FLOW_ITEM_TYPE_END,
+};
+
+static enum rte_flow_item_type pattern_mpls_3[] = {
+   RTE_FLOW_ITEM_TYPE_ETH,
+   RTE_FLOW_ITEM_TYPE_IPV4,
+   RTE_FLOW_ITEM_TYPE_GRE,
+   RTE_FLOW_ITEM_TYPE_MPLS,
+   RTE_FLOW_ITEM_TYPE_END,
+};
+
+static enum rte_flow_item_type pattern_mpls_4[] = {
+   RTE_FLOW_ITEM_TYPE_ETH,
+   RTE_FLOW_ITEM_TYPE_IPV6,
+   RTE_FLOW_ITEM_TYPE_GRE,
+   RTE_FLOW_ITEM_TYPE_MPLS,
+   RTE_FLOW_ITEM_TYPE_END,
+};
+
 static struct i40e_valid_pattern i40e_supported_patterns[] = {
/* Ethertype */
{ pattern_ethertype, i40e_flow_parse_ethertype_filter },
@@ -303,6 +343,11 @@ static struct i40e_valid_pattern i40e_supported_patterns[] 
= {
{ pattern_vxlan_2, i40e_flow_parse_vxlan_filter },
{ pattern_vxlan_3, i40e_flow_parse_vxlan_filter },
{ pattern_vxlan_4, i40e_flow_parse_vxlan_filter },
+   /* MPLSoUDP & MPLSoGRE */
+   { pattern_mpls_1, i40e_flow_parse_mpls_filter },
+   { pattern_mpls_2, i40e_flow_parse_mpls_filter },
+   { pattern_mpls_3, i40e_flow_parse_mpls_filter },
+   { pattern_mpls_4, i40e_flow_parse_mpls_filter },
 };
 
 #define NEXT_ITEM_OF_ACTION(act, actions, index)\
@@ -1515,6 +1560,166 @@ i40e_flow_parse_vxlan_filter(struct rte_eth_dev *dev,
return ret;
 }
 
+/* 1. Last in item should be NULL as range is not supported.
+ * 2. Supported filter types: MPLS label.
+ * 3. Mask of fields which need to be matched should be
+ *filled with 1.
+ * 4. Mask of fields which needn't to be matched should be
+ *filled with 0.
+ */
+static int
+i40e_flow_parse_mpls_pattern(__rte_unused struct rte_eth_dev *dev,
+const struct rte_flow_item *pattern,
+struct rte_flow_error *error,
+struct i40e_tunnel_filter_conf *filter)
+{
+   const struct rte_flow_item *item = pattern;
+   const struct rte_flow_item_mpls *mpls_spec;
+   const struct rte_flow_item_mpl

[dpdk-dev] [PATCH v2 3/3] net/i40e: enable tunnel filter for MPLS

[Beilei Xing]

This patch enables MPLS tunnel filter by replacing
inner_mac filter.
This configuration will be set when adding MPLSoUDP
and MPLSoGRE filter rules, and it will be invalid
only by NIC core reset.

Signed-off-by: Beilei Xing 
---
 drivers/net/i40e/i40e_ethdev.c |  45 +++--
 drivers/net/i40e/i40e_ethdev.h |  11 -
 drivers/net/i40e/i40e_flow.c   | 110 +
 3 files changed, 149 insertions(+), 17 deletions(-)

diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index 506b957..5e6e2f5 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -6954,6 +6954,7 @@ i40e_dev_consistent_tunnel_filter_set(struct i40e_pf *pf,
struct i40e_tunnel_rule *tunnel_rule = &pf->tunnel;
struct i40e_tunnel_filter *tunnel, *node;
struct i40e_tunnel_filter check_filter; /* Check if filter exists */
+   uint32_t teid_le;
bool big_buffer = 0;
 
cld_filter = rte_zmalloc("tunnel_filter",
@@ -7002,6 +7003,32 @@ i40e_dev_consistent_tunnel_filter_set(struct i40e_pf *pf,
case I40E_TUNNEL_TYPE_IP_IN_GRE:
tun_type = I40E_AQC_ADD_CLOUD_TNL_TYPE_IP;
break;
+   case I40E_TUNNEL_TYPE_MPLSoUDP:
+   i40e_replace_mpls_l1_filter(pf);
+   i40e_replace_mpls_cloud_filter(pf);
+   teid_le = rte_cpu_to_le_32(tunnel_filter->tenant_id);
+   pfilter->general_fields[I40E_AQC_ADD_CLOUD_FV_FLU_0X11_WORD0] =
+   teid_le >> 4;
+   pfilter->general_fields[I40E_AQC_ADD_CLOUD_FV_FLU_0X11_WORD1] =
+   (teid_le & 0xF) << 12;
+   pfilter->general_fields[I40E_AQC_ADD_CLOUD_FV_FLU_0X11_WORD2] =
+   0x40;
+   big_buffer = 1;
+   tun_type = I40E_AQC_ADD_CLOUD_TNL_TYPE_MPLSoUDP;
+   break;
+   case I40E_TUNNEL_TYPE_MPLSoGRE:
+   i40e_replace_mpls_l1_filter(pf);
+   i40e_replace_mpls_cloud_filter(pf);
+   teid_le = rte_cpu_to_le_32(tunnel_filter->tenant_id);
+   pfilter->general_fields[I40E_AQC_ADD_CLOUD_FV_FLU_0X11_WORD0] =
+   teid_le >> 4;
+   pfilter->general_fields[I40E_AQC_ADD_CLOUD_FV_FLU_0X11_WORD1] =
+   (teid_le & 0xF) << 12;
+   pfilter->general_fields[I40E_AQC_ADD_CLOUD_FV_FLU_0X11_WORD2] =
+   0x0;
+   big_buffer = 1;
+   tun_type = I40E_AQC_ADD_CLOUD_TNL_TYPE_MPLSoGRE;
+   break;
default:
/* Other tunnel types is not supported. */
PMD_DRV_LOG(ERR, "tunnel type is not supported.");
@@ -7009,11 +7036,19 @@ i40e_dev_consistent_tunnel_filter_set(struct i40e_pf 
*pf,
return -EINVAL;
}
 
-   val = i40e_dev_get_filter_type(tunnel_filter->filter_type,
-  &pfilter->element.flags);
-   if (val < 0) {
-   rte_free(cld_filter);
-   return -EINVAL;
+   if (tunnel_filter->tunnel_type == I40E_TUNNEL_TYPE_MPLSoUDP)
+   pfilter->element.flags =
+   I40E_AQC_ADD_CLOUD_FILTER_TEID_MPLSoUDP;
+   else if (tunnel_filter->tunnel_type == I40E_TUNNEL_TYPE_MPLSoGRE)
+   pfilter->element.flags =
+   I40E_AQC_ADD_CLOUD_FILTER_TEID_MPLSoGRE;
+   else {
+   val = i40e_dev_get_filter_type(tunnel_filter->filter_type,
+   &pfilter->element.flags);
+   if (val < 0) {
+   rte_free(cld_filter);
+   return -EINVAL;
+   }
}
 
pfilter->element.flags |= rte_cpu_to_le_16(
diff --git a/drivers/net/i40e/i40e_ethdev.h b/drivers/net/i40e/i40e_ethdev.h
index df345b1..934c679 100644
--- a/drivers/net/i40e/i40e_ethdev.h
+++ b/drivers/net/i40e/i40e_ethdev.h
@@ -500,8 +500,13 @@ struct i40e_ethertype_rule {
 /* Tunnel filter number HW supports */
 #define I40E_MAX_TUNNEL_FILTER_NUM 400
 
-#define I40E_AQC_ADD_CLOUD_TNL_TYPE_MPLSoUDP 0x11
-#define I40E_AQC_ADD_CLOUD_TNL_TYPE_MPLSoGRE 0x12
+#define I40E_AQC_REPLACE_CLOUD_CMD_INPUT_FV_TEID_WORD0 44
+#define I40E_AQC_REPLACE_CLOUD_CMD_INPUT_FV_TEID_WORD1 45
+#define I40E_AQC_ADD_CLOUD_TNL_TYPE_MPLSoUDP 8
+#define I40E_AQC_ADD_CLOUD_TNL_TYPE_MPLSoGRE 9
+#define I40E_AQC_ADD_CLOUD_FILTER_TEID_MPLSoUDP 0x11
+#define I40E_AQC_ADD_CLOUD_FILTER_TEID_MPLSoGRE 0x12
+#define I40E_AQC_ADD_L1_FILTER_TEID_MPLS 0x11
 
 enum i40e_tunnel_iptype {
I40E_TUNNEL_IPTYPE_IPV4,
@@ -863,6 +868,8 @@ int i40e_dev_consistent_tunnel_filter_set(struct i40e_pf 
*pf,
  struct i40e_tunnel_filter_conf *tunnel_filter,
  uint8_t add);
 int i40e_fdir_flush(struct rte_eth_dev *dev);
+enum i40e_status_code i40e_replace_mpls_l1_filter(struct i40e_pf *pf);
+enum i40e_status

[dpdk-dev] [PATCH v2 2/3] crypto/scheduler: enable packet size based scheduling mode

[Fan Zhang]

This patch enables the packet size based scheduling mode in scheduler PMD.

Signed-off-by: Fan Zhang 
---
 drivers/crypto/scheduler/Makefile  | 1 +
 drivers/crypto/scheduler/rte_cryptodev_scheduler.c | 7 +++
 drivers/crypto/scheduler/rte_cryptodev_scheduler.h | 3 +++
 3 files changed, 11 insertions(+)

diff --git a/drivers/crypto/scheduler/Makefile 
b/drivers/crypto/scheduler/Makefile
index 0cce6f2..93e135f 100644
--- a/drivers/crypto/scheduler/Makefile
+++ b/drivers/crypto/scheduler/Makefile
@@ -54,6 +54,7 @@ SRCS-$(CONFIG_RTE_LIBRTE_PMD_CRYPTO_SCHEDULER) += 
scheduler_pmd.c
 SRCS-$(CONFIG_RTE_LIBRTE_PMD_CRYPTO_SCHEDULER) += scheduler_pmd_ops.c
 SRCS-$(CONFIG_RTE_LIBRTE_PMD_CRYPTO_SCHEDULER) += rte_cryptodev_scheduler.c
 SRCS-$(CONFIG_RTE_LIBRTE_PMD_CRYPTO_SCHEDULER) += scheduler_roundrobin.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_CRYPTO_SCHEDULER) += scheduler_pkt_size_distr.c
 
 # library dependencies
 DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_CRYPTO_SCHEDULER) += lib/librte_cryptodev
diff --git a/drivers/crypto/scheduler/rte_cryptodev_scheduler.c 
b/drivers/crypto/scheduler/rte_cryptodev_scheduler.c
index 11e8143..eab1906 100644
--- a/drivers/crypto/scheduler/rte_cryptodev_scheduler.c
+++ b/drivers/crypto/scheduler/rte_cryptodev_scheduler.c
@@ -336,6 +336,13 @@ rte_crpytodev_scheduler_mode_set(uint8_t scheduler_id,
return -1;
}
break;
+   case CDEV_SCHED_MODE_PKT_SIZE_DISTR:
+   if (rte_cryptodev_scheduler_load_user_scheduler(scheduler_id,
+   pkt_size_based_distr_scheduler) < 0) {
+   CS_LOG_ERR("Failed to load scheduler");
+   return -1;
+   }
+   break;
default:
CS_LOG_ERR("Not yet supported");
return -ENOTSUP;
diff --git a/drivers/crypto/scheduler/rte_cryptodev_scheduler.h 
b/drivers/crypto/scheduler/rte_cryptodev_scheduler.h
index 7ef44e7..a1d4c14 100644
--- a/drivers/crypto/scheduler/rte_cryptodev_scheduler.h
+++ b/drivers/crypto/scheduler/rte_cryptodev_scheduler.h
@@ -47,6 +47,8 @@ enum rte_cryptodev_scheduler_mode {
CDEV_SCHED_MODE_NOT_SET = 0,
CDEV_SCHED_MODE_USERDEFINED,
CDEV_SCHED_MODE_ROUNDROBIN,
+   /** packet-size based distribution mode */
+   CDEV_SCHED_MODE_PKT_SIZE_DISTR,
 
CDEV_SCHED_MODE_COUNT /* number of modes */
 };
@@ -158,6 +160,7 @@ struct rte_cryptodev_scheduler {
 };
 
 extern struct rte_cryptodev_scheduler *roundrobin_scheduler;
+extern struct rte_cryptodev_scheduler *pkt_size_based_distr_scheduler;
 
 #ifdef __cplusplus
 }
-- 
2.7.4

[dpdk-dev] [PATCH v2 3/3] doc: update cryptodev scheduler PMD documentation

[Fan Zhang]

This patch updates packet size based scheduling mode description.

Signed-off-by: Fan Zhang 
---
 doc/guides/cryptodevs/scheduler.rst | 14 ++
 1 file changed, 14 insertions(+)

diff --git a/doc/guides/cryptodevs/scheduler.rst 
b/doc/guides/cryptodevs/scheduler.rst
index 70fb62e..51f2a78 100644
--- a/doc/guides/cryptodevs/scheduler.rst
+++ b/doc/guides/cryptodevs/scheduler.rst
@@ -126,3 +126,17 @@ operation:
among its slaves in a round-robin manner. This mode may help to fill
the throughput gap between the physical core and the existing cryptodevs
to increase the overall performance.
+
+*   **CDEV_SCHED_MODE_PKT_SIZE_DISTR:**
+
+   Packet-size based distribution mode, which works with 2 slaves, primary
+   slave and secondary slave, and distribute the enqueued crypto ops to them
+   based on their data lengths. A crypto op will be distributed to the primary
+   slave if its data length equals or bigger than the designated threshold,
+   otherwise it will be handled by the secondary slave.
+
+   A typical usecase of this mode is with QAT cryptodev as primary and a
+   software cryptodev as secondary slave. This may help the application being
+   capable of processing extra crypto workload than what the sole QAT cryptodev
+   can handle, by making use of the available CPU cycles to deal with lesser
+   crypto workloads.
-- 
2.7.4

[dpdk-dev] [PATCH v2 1/3] crypto/scheduler: add packet size based mode code

[Fan Zhang]

This patch adds the packet size based distribution mode main source
file.

Signed-off-by: Fan Zhang 
---
 .../crypto/scheduler/scheduler_pkt_size_distr.c| 427 +
 1 file changed, 427 insertions(+)
 create mode 100644 drivers/crypto/scheduler/scheduler_pkt_size_distr.c

diff --git a/drivers/crypto/scheduler/scheduler_pkt_size_distr.c 
b/drivers/crypto/scheduler/scheduler_pkt_size_distr.c
new file mode 100644
index 000..d1e8b7c
--- /dev/null
+++ b/drivers/crypto/scheduler/scheduler_pkt_size_distr.c
@@ -0,0 +1,427 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2017 Intel Corporation. All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of Intel Corporation nor the names of its
+ *   contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include 
+#include 
+
+#include "rte_cryptodev_scheduler_operations.h"
+#include "scheduler_pmd_private.h"
+
+#define PKT_SIZE_THRESHOLD (0xff80)
+#define SLAVE_IDX_SWITCH_MASK  (0x01)
+#define PRIMARY_SLAVE_IDX  0
+#define SECONDARY_SLAVE_IDX1
+#define NB_PKT_SIZE_SLAVES 2
+
+struct psd_scheduler_ctx {
+   uint16_t threshold;
+};
+
+struct psd_scheduler_qp_ctx {
+   struct scheduler_slave primary_slave;
+   struct scheduler_slave secondary_slave;
+   uint16_t threshold;
+   uint8_t deq_idx;
+} __rte_cache_aligned;
+
+/** scheduling operation variables' wrapping */
+struct psd_schedule_op {
+   uint16_t slave_idx;
+   int pos;
+   int pos_shift;
+};
+
+static uint16_t
+schedule_enqueue(void *qp, struct rte_crypto_op **ops, uint16_t nb_ops)
+{
+   struct psd_scheduler_qp_ctx *qp_ctx =
+   ((struct scheduler_qp_ctx *)qp)->private_qp_ctx;
+   struct rte_crypto_op *sched_ops[nb_ops];
+   struct rte_cryptodev_sym_session *sessions[nb_ops];
+   struct scheduler_session *sess;
+   struct psd_schedule_op enq_ops[NB_PKT_SIZE_SLAVES] = {
+   {PRIMARY_SLAVE_IDX, 0, 1},
+   {SECONDARY_SLAVE_IDX, (int)(nb_ops - 1), -1}
+   };
+   struct psd_schedule_op *p_enq_op;
+   uint16_t i, processed_ops = 0, processed_ops2 = 0, nb_ops_to_enq;
+   uint32_t job_len;
+
+   if (unlikely(nb_ops == 0))
+   return 0;
+
+   for (i = 0; i < nb_ops && i < 4; i++) {
+   rte_prefetch0(ops[i]->sym);
+   rte_prefetch0(ops[i]->sym->session);
+   }
+
+   for (i = 0; (i < (nb_ops - 8)) && (nb_ops > 8); i += 4) {
+   rte_prefetch0(ops[i + 4]->sym);
+   rte_prefetch0(ops[i + 4]->sym->session);
+   rte_prefetch0(ops[i + 5]->sym);
+   rte_prefetch0(ops[i + 5]->sym->session);
+   rte_prefetch0(ops[i + 6]->sym);
+   rte_prefetch0(ops[i + 6]->sym->session);
+   rte_prefetch0(ops[i + 7]->sym);
+   rte_prefetch0(ops[i + 7]->sym->session);
+
+   sess = (struct scheduler_session *)
+   ops[i]->sym->session->_private;
+   job_len = ops[i]->sym->cipher.data.length;
+   job_len += (ops[i]->sym->auth.data.length == 0) *
+   ops[i]->sym->auth.data.length;
+   /* decide the target op based on the job length */
+   p_enq_op = &enq_ops[!(job_len & qp_ctx->threshold)];
+   sched_ops[p_enq_op->pos] = ops[i];
+   sessions[p_enq_op->pos] = ops[i]->sym->session;
+   ops[i

[dpdk-dev] [PATCH v2 0/3] crypto/scheduler: add packet-base scheduling mode

[Fan Zhang]

Packet-size based distribution mode, which works with 2 slaves, primary
slave and secondary slave, and distribute the enqueued crypto ops to them
based on their data lengths. A crypto op will be distributed to the primary
slave if its data length equals or bigger than the designated threshold,
otherwise it will be handled by the secondary slave.

Fan Zhang (3):
  crypto/scheduler: add packet size based mode code
  crypto/scheduler: enable packet size based scheduling mode
  doc: update cryptodev scheduler PMD documentation

 doc/guides/cryptodevs/scheduler.rst|  14 +
 drivers/crypto/scheduler/Makefile  |   1 +
 drivers/crypto/scheduler/rte_cryptodev_scheduler.c |   7 +
 drivers/crypto/scheduler/rte_cryptodev_scheduler.h |   3 +
 .../crypto/scheduler/scheduler_pkt_size_distr.c| 427 +
 5 files changed, 452 insertions(+)
 create mode 100644 drivers/crypto/scheduler/scheduler_pkt_size_distr.c

-- 
2.7.4

[dpdk-dev] [PATCH v5 00/14] Wind River Systems AVP PMD

[Allain Legacy]

This patch series submits an initial version of the AVP PMD from Wind River
Systems.  The series includes shared header files, driver implementation,
and changes to documentation files in support of this new driver.  The AVP
driver is a shared memory based device.  It is intended to be used as a PMD
within a virtual machine running on a Wind River virtualization platform.
See: http://www.windriver.com/products/titanium-cloud/

It enables optimized packet throughput without requiring any packet
processing in qemu. This allowed us to provide our customers with a
significant performance increase for both DPDK and non-DPDK applications
in the VM.   Since our AVP implementation supports VM live-migration it
is viewed as a better alternative to PCI passthrough or PCI SRIOV since
neither of those support VM live-migration without manual intervention
or significant performance penalties.

Since the initial implementation of AVP devices, vhost-user has become part
of the qemu offering with a significant performance increase over the
original virtio implementation.  However, vhost-user still does not achieve
the level of performance that the AVP device can provide to our customers
for DPDK based guests.

A number of our customers have requested that we upstream the driver to
dpdk.org.

v2:
* Fixed coding style violations that slipped in accidentally because of an
  out of date checkpatch.pl from an older kernel.

v3:
* Updated 17.05 release notes to add a section for this new PMD
* Added additional info to the AVP nic guide document to clarify the
  benefit of using AVP over virtio.
* Fixed spelling error in debug log missed by local checkpatch.pl version
* Split the transmit patch to separate the stats functions as they
  accidentally got squashed in the last patchset.
* Fixed debug log strings so that they exceed 80 characters rather than
  span multiple lines.
* Renamed RTE_AVP_* defines that were in avp_ethdev.h to be AVP_* instead
* Replaced usage of RTE_WRITE32 and RTE_READ32 with rte_write32_relaxed
  and rte_read32_relaxed.
* Declared rte_pci_id table as const

v4:
* Split our interrupt handlers to a separate patch and moved to the end
  of the series.
* Removed memset() from stats_get API
* Removed usage of RTE_AVP_ALIGNMENT
* Removed unnecessary parentheses in rte_avp_common.h
* Removed unneeded "goto unlock" where there are no statements in between
  the goto and the end of the function.
* Re-tested with pktgen and found that rte_eth_tx_burst() is being called
  with 0 packets even before starting traffic which resulted in
  incrementing oerrors; fixed in transmit patch.

v5:
* Updated documentation to remove references to ivshmem as it lead to
  confusion about whether AVP is exactly like ivshmem or simply based on
  how ivshmem exports memory to a VM via a PCI device.
* Restructured first set of patches to condense them down to a base patch
  with the files needed to apply subsequent patches.
* Removed static prototypes from init/uninit functions in avp_ethdev.c
* Moved MAC addresses init to the device initialization patch because it
  is setup by the avp_dev_create() function.
* Split the changes to the avp.ini features file so that features are
  marked as enabled in the patch that actually enables them.

Allain Legacy (14):
  drivers/net: adds AVP PMD base files
  net/avp: public header files
  net/avp: debug log macros
  net/avp: driver registration
  net/avp: device initialization
  net/avp: device configuration
  net/avp: queue setup and release
  net/avp: packet receive functions
  net/avp: packet transmit functions
  net/avp: device statistics operations
  net/avp: device promiscuous functions
  net/avp: device start and stop operations
  net/avp: migration interrupt handling
  doc: adds information related to the AVP PMD

 MAINTAINERS  |6 +
 config/common_base   |9 +
 config/common_linuxapp   |1 +
 config/defconfig_i686-native-linuxapp-gcc|5 +
 config/defconfig_i686-native-linuxapp-icc|5 +
 config/defconfig_x86_x32-native-linuxapp-gcc |5 +
 doc/guides/nics/avp.rst  |  107 ++
 doc/guides/nics/features/avp.ini |   16 +
 doc/guides/nics/index.rst|1 +
 doc/guides/rel_notes/release_17_05.rst   |5 +
 drivers/net/Makefile |1 +
 drivers/net/avp/Makefile |   61 +
 drivers/net/avp/avp_ethdev.c | 2294 ++
 drivers/net/avp/avp_logs.h   |   59 +
 drivers/net/avp/rte_avp_common.h |  416 +
 drivers/net/avp/rte_avp_fifo.h   |  157 ++
 drivers/net/avp/rte_pmd_avp_version.map  |4 +
 mk/rte.app.mk|1 +
 18 files changed, 3153 insertions(+)
 create mode 100644 doc/guides/nics/avp.rst
 create mode 100644 doc/guides/nics/features/avp.ini
 create mode 100644 drivers/net/avp/Mak

[dpdk-dev] [PATCH v5 03/14] net/avp: debug log macros

[Allain Legacy]

Adds a header file with log macros for the AVP PMD

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 config/common_base |  4 
 drivers/net/avp/avp_logs.h | 59 ++
 2 files changed, 63 insertions(+)
 create mode 100644 drivers/net/avp/avp_logs.h

diff --git a/config/common_base b/config/common_base
index a4ad577..5beedc8 100644
--- a/config/common_base
+++ b/config/common_base
@@ -356,6 +356,10 @@ CONFIG_RTE_LIBRTE_PMD_AF_PACKET=n
 # Compile WRS accelerated virtual port (AVP) guest PMD driver
 #
 CONFIG_RTE_LIBRTE_AVP_PMD=n
+CONFIG_RTE_LIBRTE_AVP_DEBUG_RX=n
+CONFIG_RTE_LIBRTE_AVP_DEBUG_TX=n
+CONFIG_RTE_LIBRTE_AVP_DEBUG_DRIVER=y
+CONFIG_RTE_LIBRTE_AVP_DEBUG_BUFFERS=n
 
 #
 # Compile the TAP PMD
diff --git a/drivers/net/avp/avp_logs.h b/drivers/net/avp/avp_logs.h
new file mode 100644
index 000..252cab7
--- /dev/null
+++ b/drivers/net/avp/avp_logs.h
@@ -0,0 +1,59 @@
+/*
+ * BSD LICENSE
+ *
+ * Copyright (c) 2013-2015, Wind River Systems, Inc.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1) Redistributions of source code must retain the above copyright notice,
+ * this list of conditions and the following disclaimer.
+ *
+ * 2) Redistributions in binary form must reproduce the above copyright notice,
+ * this list of conditions and the following disclaimer in the documentation
+ * and/or other materials provided with the distribution.
+ *
+ * 3) Neither the name of Wind River Systems nor the names of its contributors
+ * may be used to endorse or promote products derived from this software
+ * without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _AVP_LOGS_H_
+#define _AVP_LOGS_H_
+
+#include 
+
+#ifdef RTE_LIBRTE_AVP_DEBUG_RX
+#define PMD_RX_LOG(level, fmt, args...) \
+   RTE_LOG(level, PMD, "%s() rx: " fmt, __func__, ## args)
+#else
+#define PMD_RX_LOG(level, fmt, args...) do { } while (0)
+#endif
+
+#ifdef RTE_LIBRTE_AVP_DEBUG_TX
+#define PMD_TX_LOG(level, fmt, args...) \
+   RTE_LOG(level, PMD, "%s() tx: " fmt, __func__, ## args)
+#else
+#define PMD_TX_LOG(level, fmt, args...) do { } while (0)
+#endif
+
+#ifdef RTE_LIBRTE_AVP_DEBUG_DRIVER
+#define PMD_DRV_LOG(level, fmt, args...) \
+   RTE_LOG(level, PMD, "%s(): " fmt, __func__, ## args)
+#else
+#define PMD_DRV_LOG(level, fmt, args...) do { } while (0)
+#endif
+
+#endif /* _AVP_LOGS_H_ */
-- 
1.8.3.1

[dpdk-dev] [PATCH v5 04/14] net/avp: driver registration

[Allain Legacy]

Adds the initial framework for registering the driver against the support
PCI device identifiers.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 config/common_linuxapp   |   1 +
 config/defconfig_i686-native-linuxapp-gcc|   5 +
 config/defconfig_i686-native-linuxapp-icc|   5 +
 config/defconfig_x86_x32-native-linuxapp-gcc |   5 +
 drivers/net/avp/Makefile |   8 ++
 drivers/net/avp/avp_ethdev.c | 205 +++
 mk/rte.app.mk|   1 +
 7 files changed, 230 insertions(+)
 create mode 100644 drivers/net/avp/avp_ethdev.c

diff --git a/config/common_linuxapp b/config/common_linuxapp
index 00ebaac..8690a00 100644
--- a/config/common_linuxapp
+++ b/config/common_linuxapp
@@ -43,6 +43,7 @@ CONFIG_RTE_LIBRTE_VHOST=y
 CONFIG_RTE_LIBRTE_PMD_VHOST=y
 CONFIG_RTE_LIBRTE_PMD_AF_PACKET=y
 CONFIG_RTE_LIBRTE_PMD_TAP=y
+CONFIG_RTE_LIBRTE_AVP_PMD=y
 CONFIG_RTE_LIBRTE_NFP_PMD=y
 CONFIG_RTE_LIBRTE_POWER=y
 CONFIG_RTE_VIRTIO_USER=y
diff --git a/config/defconfig_i686-native-linuxapp-gcc 
b/config/defconfig_i686-native-linuxapp-gcc
index 745c401..9847bdb 100644
--- a/config/defconfig_i686-native-linuxapp-gcc
+++ b/config/defconfig_i686-native-linuxapp-gcc
@@ -75,3 +75,8 @@ CONFIG_RTE_LIBRTE_PMD_KASUMI=n
 # ZUC PMD is not supported on 32-bit
 #
 CONFIG_RTE_LIBRTE_PMD_ZUC=n
+
+#
+# AVP PMD is not supported on 32-bit
+#
+CONFIG_RTE_LIBRTE_AVP_PMD=n
diff --git a/config/defconfig_i686-native-linuxapp-icc 
b/config/defconfig_i686-native-linuxapp-icc
index 50a3008..269e88e 100644
--- a/config/defconfig_i686-native-linuxapp-icc
+++ b/config/defconfig_i686-native-linuxapp-icc
@@ -75,3 +75,8 @@ CONFIG_RTE_LIBRTE_PMD_KASUMI=n
 # ZUC PMD is not supported on 32-bit
 #
 CONFIG_RTE_LIBRTE_PMD_ZUC=n
+
+#
+# AVP PMD is not supported on 32-bit
+#
+CONFIG_RTE_LIBRTE_AVP_PMD=n
diff --git a/config/defconfig_x86_x32-native-linuxapp-gcc 
b/config/defconfig_x86_x32-native-linuxapp-gcc
index 3e55c5c..19573cb 100644
--- a/config/defconfig_x86_x32-native-linuxapp-gcc
+++ b/config/defconfig_x86_x32-native-linuxapp-gcc
@@ -50,3 +50,8 @@ CONFIG_RTE_LIBRTE_KNI=n
 # Solarflare PMD is not supported on 32-bit
 #
 CONFIG_RTE_LIBRTE_SFC_EFX_PMD=n
+
+#
+# AVP PMD is not supported on 32-bit
+#
+CONFIG_RTE_LIBRTE_AVP_PMD=n
diff --git a/drivers/net/avp/Makefile b/drivers/net/avp/Makefile
index 68a0fa5..9cf0449 100644
--- a/drivers/net/avp/Makefile
+++ b/drivers/net/avp/Makefile
@@ -49,4 +49,12 @@ LIBABIVER := 1
 SYMLINK-$(CONFIG_RTE_LIBRTE_AVP_PMD)-include += rte_avp_common.h
 SYMLINK-$(CONFIG_RTE_LIBRTE_AVP_PMD)-include += rte_avp_fifo.h
 
+#
+# all source files are stored in SRCS-y
+#
+SRCS-$(CONFIG_RTE_LIBRTE_AVP_PMD) += avp_ethdev.c
+
+# this lib depends upon:
+DEPDIRS-$(CONFIG_RTE_LIBRTE_AVP_PMD) += lib/librte_eal lib/librte_ether
+
 include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
new file mode 100644
index 000..f5fa453
--- /dev/null
+++ b/drivers/net/avp/avp_ethdev.c
@@ -0,0 +1,205 @@
+/*
+ *   BSD LICENSE
+ *
+ * Copyright (c) 2013-2017, Wind River Systems, Inc.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1) Redistributions of source code must retain the above copyright notice,
+ * this list of conditions and the following disclaimer.
+ *
+ * 2) Redistributions in binary form must reproduce the above copyright notice,
+ * this list of conditions and the following disclaimer in the documentation
+ * and/or other materials provided with the distribution.
+ *
+ * 3) Neither the name of Wind River Systems nor the names of its contributors
+ * may be used to endorse or promote products derived from this software
+ * without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include "rte_avp_common.h"
+#include "rte_avp_fifo.h"
+
+#include "avp_logs.h"
+

[dpdk-dev] [PATCH v5 02/14] net/avp: public header files

[Allain Legacy]

Adds public/exported header files for the AVP PMD.  The AVP device is a
shared memory based device.  The structures and constants that define the
method of operation of the device must be visible by both the PMD and the
host DPDK application.  They must not change without proper version
controls and updates to both the hypervisor DPDK application and the PMD.

The hypervisor DPDK application is a Wind River Systems proprietary
virtual switch.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 drivers/net/avp/Makefile |   5 +
 drivers/net/avp/rte_avp_common.h | 416 +++
 drivers/net/avp/rte_avp_fifo.h   | 157 +++
 3 files changed, 578 insertions(+)
 create mode 100644 drivers/net/avp/rte_avp_common.h
 create mode 100644 drivers/net/avp/rte_avp_fifo.h

diff --git a/drivers/net/avp/Makefile b/drivers/net/avp/Makefile
index c6e03d5..68a0fa5 100644
--- a/drivers/net/avp/Makefile
+++ b/drivers/net/avp/Makefile
@@ -44,4 +44,9 @@ EXPORT_MAP := rte_pmd_avp_version.map
 
 LIBABIVER := 1
 
+# install public header files to enable compilation of the hypervisor level
+# dpdk application
+SYMLINK-$(CONFIG_RTE_LIBRTE_AVP_PMD)-include += rte_avp_common.h
+SYMLINK-$(CONFIG_RTE_LIBRTE_AVP_PMD)-include += rte_avp_fifo.h
+
 include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/net/avp/rte_avp_common.h b/drivers/net/avp/rte_avp_common.h
new file mode 100644
index 000..31d763e
--- /dev/null
+++ b/drivers/net/avp/rte_avp_common.h
@@ -0,0 +1,416 @@
+/*-
+ *   This file is provided under a dual BSD/LGPLv2 license.  When using or
+ *   redistributing this file, you may do so under either license.
+ *
+ *   GNU LESSER GENERAL PUBLIC LICENSE
+ *
+ *   Copyright(c) 2010-2013 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2014-2017 Wind River Systems, Inc. All rights reserved.
+ *
+ *   This program is free software; you can redistribute it and/or modify
+ *   it under the terms of version 2.1 of the GNU Lesser General Public License
+ *   as published by the Free Software Foundation.
+ *
+ *   This program is distributed in the hope that it will be useful, but
+ *   WITHOUT ANY WARRANTY; without even the implied warranty of
+ *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ *   Lesser General Public License for more details.
+ *
+ *   Contact Information:
+ *   Wind River Systems, Inc.
+ *
+ *
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2010-2013 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2014-2017 Wind River Systems, Inc. All rights reserved.
+ *   All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ *   * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ *   * Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ *THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+#ifndef _RTE_AVP_COMMON_H_
+#define _RTE_AVP_COMMON_H_
+
+#ifdef __KERNEL__
+#include 
+#endif
+
+/**
+ * AVP name is part of network device name.
+ */
+#define RTE_AVP_NAMESIZE 32
+
+/**
+ * AVP alias is a user-defined value used for lookups from secondary
+ * processes.  Typically, this is a UUID.
+ */
+#define RTE_AVP_ALIASSIZE 128
+
+/*
+ * Request id.
+ */
+enum rte_avp_req_id {
+   RTE_AVP_REQ_UNKNOWN = 0,
+   RTE_AVP_REQ_CHANGE_MTU,
+   RTE_AVP_REQ_CFG_NETWORK_IF,
+   RTE_AVP_REQ_CFG_DEVICE,
+   RTE_AVP_REQ_SHUTDOWN_DEVICE,
+   RTE_AVP_REQ_MAX,
+};
+
+/**@{ AVP device driver types */
+#define RTE_AVP_DRIVER_TYPE_UNKNOWN 0
+#define RTE_AVP_DRIVER_TYPE_DPDK 1
+#define RTE_AVP_DRIVER_TYPE_KERNEL 2
+#define RTE_AVP_DRIVER_TYPE_QEMU 3
+/**@} */
+
+/**@{ A

[dpdk-dev] [PATCH v5 06/14] net/avp: device configuration

[Allain Legacy]

Adds support for "dev_configure" operations to allow an application to
configure the device.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 doc/guides/nics/features/avp.ini |   4 +
 drivers/net/avp/avp_ethdev.c | 241 +++
 2 files changed, 245 insertions(+)

diff --git a/doc/guides/nics/features/avp.ini b/doc/guides/nics/features/avp.ini
index 4353929..45a2185 100644
--- a/doc/guides/nics/features/avp.ini
+++ b/doc/guides/nics/features/avp.ini
@@ -4,3 +4,7 @@
 ; Refer to default.ini for the full list of available PMD features.
 ;
 [Features]
+Link status  = Y
+VLAN offload = Y
+Linux UIO= Y
+x86-64   = Y
diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
index e937fb52..d9ad3f1 100644
--- a/drivers/net/avp/avp_ethdev.c
+++ b/drivers/net/avp/avp_ethdev.c
@@ -61,6 +61,13 @@
 
 
 
+static int avp_dev_configure(struct rte_eth_dev *dev);
+static void avp_dev_info_get(struct rte_eth_dev *dev,
+struct rte_eth_dev_info *dev_info);
+static void avp_vlan_offload_set(struct rte_eth_dev *dev, int mask);
+static int avp_dev_link_update(struct rte_eth_dev *dev,
+  __rte_unused int wait_to_complete);
+
 #define AVP_DEV_TO_PCI(eth_dev) RTE_DEV_TO_PCI((eth_dev)->device)
 
 
@@ -98,6 +105,15 @@
},
 };
 
+/*
+ * dev_ops for avp, bare necessities for basic operation
+ */
+static const struct eth_dev_ops avp_eth_dev_ops = {
+   .dev_configure   = avp_dev_configure,
+   .dev_infos_get   = avp_dev_info_get,
+   .vlan_offload_set= avp_vlan_offload_set,
+   .link_update = avp_dev_link_update,
+};
 
 /**@{ AVP device flags */
 #define AVP_F_PROMISC (1 << 1)
@@ -183,6 +199,91 @@ struct avp_queue {
uint64_t errors;
 };
 
+/* send a request and wait for a response
+ *
+ * @warning must be called while holding the avp->lock spinlock.
+ */
+static int
+avp_dev_process_request(struct avp_dev *avp, struct rte_avp_request *request)
+{
+   unsigned int retry = AVP_MAX_REQUEST_RETRY;
+   void *resp_addr = NULL;
+   unsigned int count;
+   int ret;
+
+   PMD_DRV_LOG(DEBUG, "Sending request %u to host\n", request->req_id);
+
+   request->result = -ENOTSUP;
+
+   /* Discard any stale responses before starting a new request */
+   while (avp_fifo_get(avp->resp_q, (void **)&resp_addr, 1))
+   PMD_DRV_LOG(DEBUG, "Discarding stale response\n");
+
+   rte_memcpy(avp->sync_addr, request, sizeof(*request));
+   count = avp_fifo_put(avp->req_q, &avp->host_sync_addr, 1);
+   if (count < 1) {
+   PMD_DRV_LOG(ERR, "Cannot send request %u to host\n",
+   request->req_id);
+   ret = -EBUSY;
+   goto done;
+   }
+
+   while (retry--) {
+   /* wait for a response */
+   usleep(AVP_REQUEST_DELAY_USECS);
+
+   count = avp_fifo_count(avp->resp_q);
+   if (count >= 1) {
+   /* response received */
+   break;
+   }
+
+   if ((count < 1) && (retry == 0)) {
+   PMD_DRV_LOG(ERR, "Timeout while waiting for a response 
for %u\n",
+   request->req_id);
+   ret = -ETIME;
+   goto done;
+   }
+   }
+
+   /* retrieve the response */
+   count = avp_fifo_get(avp->resp_q, (void **)&resp_addr, 1);
+   if ((count != 1) || (resp_addr != avp->host_sync_addr)) {
+   PMD_DRV_LOG(ERR, "Invalid response from host, count=%u resp=%p 
host_sync_addr=%p\n",
+   count, resp_addr, avp->host_sync_addr);
+   ret = -ENODATA;
+   goto done;
+   }
+
+   /* copy to user buffer */
+   rte_memcpy(request, avp->sync_addr, sizeof(*request));
+   ret = 0;
+
+   PMD_DRV_LOG(DEBUG, "Result %d received for request %u\n",
+   request->result, request->req_id);
+
+done:
+   return ret;
+}
+
+static int
+avp_dev_ctrl_set_config(struct rte_eth_dev *eth_dev,
+   struct rte_avp_device_config *config)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   struct rte_avp_request request;
+   int ret;
+
+   /* setup a configure request */
+   memset(&request, 0, sizeof(request));
+   request.req_id = RTE_AVP_REQ_CFG_DEVICE;
+   memcpy(&request.config, config, sizeof(request.config));
+
+   ret = avp_dev_process_request(avp, &request);
+
+   return ret == 0 ? request.result : ret;
+}
+
 /* translate from host physical address to guest virtual address */
 static void *
 avp_dev_translate_address(struct rte_eth_dev *eth_dev,
@@ -298,6 +399,38 @@ struct avp_queue {
return 0;
 }
 
+static void
+_avp_set_queue_counts(struct rte_eth_dev *eth_dev)
+{
+   

[dpdk-dev] [PATCH v5 07/14] net/avp: queue setup and release

[Allain Legacy]

Adds queue management operations so that an appliation can setup and
release the transmit and receive queues.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 drivers/net/avp/avp_ethdev.c | 180 ++-
 1 file changed, 179 insertions(+), 1 deletion(-)

diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
index d9ad3f1..ecc581f 100644
--- a/drivers/net/avp/avp_ethdev.c
+++ b/drivers/net/avp/avp_ethdev.c
@@ -67,7 +67,21 @@ static void avp_dev_info_get(struct rte_eth_dev *dev,
 static void avp_vlan_offload_set(struct rte_eth_dev *dev, int mask);
 static int avp_dev_link_update(struct rte_eth_dev *dev,
   __rte_unused int wait_to_complete);
-
+static int avp_dev_rx_queue_setup(struct rte_eth_dev *dev,
+ uint16_t rx_queue_id,
+ uint16_t nb_rx_desc,
+ unsigned int socket_id,
+ const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *pool);
+
+static int avp_dev_tx_queue_setup(struct rte_eth_dev *dev,
+ uint16_t tx_queue_id,
+ uint16_t nb_tx_desc,
+ unsigned int socket_id,
+ const struct rte_eth_txconf *tx_conf);
+
+static void avp_dev_rx_queue_release(void *rxq);
+static void avp_dev_tx_queue_release(void *txq);
 #define AVP_DEV_TO_PCI(eth_dev) RTE_DEV_TO_PCI((eth_dev)->device)
 
 
@@ -113,6 +127,10 @@ static int avp_dev_link_update(struct rte_eth_dev *dev,
.dev_infos_get   = avp_dev_info_get,
.vlan_offload_set= avp_vlan_offload_set,
.link_update = avp_dev_link_update,
+   .rx_queue_setup  = avp_dev_rx_queue_setup,
+   .rx_queue_release= avp_dev_rx_queue_release,
+   .tx_queue_setup  = avp_dev_tx_queue_setup,
+   .tx_queue_release= avp_dev_tx_queue_release,
 };
 
 /**@{ AVP device flags */
@@ -400,6 +418,42 @@ struct avp_queue {
 }
 
 static void
+_avp_set_rx_queue_mappings(struct rte_eth_dev *eth_dev, uint16_t rx_queue_id)
+{
+   struct avp_dev *avp =
+   AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   struct avp_queue *rxq;
+   uint16_t queue_count;
+   uint16_t remainder;
+
+   rxq = (struct avp_queue *)eth_dev->data->rx_queues[rx_queue_id];
+
+   /*
+* Must map all AVP fifos as evenly as possible between the configured
+* device queues.  Each device queue will service a subset of the AVP
+* fifos. If there is an odd number of device queues the first set of
+* device queues will get the extra AVP fifos.
+*/
+   queue_count = avp->num_rx_queues / eth_dev->data->nb_rx_queues;
+   remainder = avp->num_rx_queues % eth_dev->data->nb_rx_queues;
+   if (rx_queue_id < remainder) {
+   /* these queues must service one extra FIFO */
+   rxq->queue_base = rx_queue_id * (queue_count + 1);
+   rxq->queue_limit = rxq->queue_base + (queue_count + 1) - 1;
+   } else {
+   /* these queues service the regular number of FIFO */
+   rxq->queue_base = ((remainder * (queue_count + 1)) +
+  ((rx_queue_id - remainder) * queue_count));
+   rxq->queue_limit = rxq->queue_base + queue_count - 1;
+   }
+
+   PMD_DRV_LOG(DEBUG, "rxq %u at %p base %u limit %u\n",
+   rx_queue_id, rxq, rxq->queue_base, rxq->queue_limit);
+
+   rxq->queue_id = rxq->queue_base;
+}
+
+static void
 _avp_set_queue_counts(struct rte_eth_dev *eth_dev)
 {
struct rte_pci_device *pci_dev = AVP_DEV_TO_PCI(eth_dev);
@@ -650,6 +704,130 @@ struct avp_queue {
 
 
 static int
+avp_dev_rx_queue_setup(struct rte_eth_dev *eth_dev,
+  uint16_t rx_queue_id,
+  uint16_t nb_rx_desc,
+  unsigned int socket_id,
+  const struct rte_eth_rxconf *rx_conf,
+  struct rte_mempool *pool)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   struct rte_pktmbuf_pool_private *mbp_priv;
+   struct avp_queue *rxq;
+
+   if (rx_queue_id >= eth_dev->data->nb_rx_queues) {
+   PMD_DRV_LOG(ERR, "RX queue id is out of range: rx_queue_id=%u, 
nb_rx_queues=%u\n",
+   rx_queue_id, eth_dev->data->nb_rx_queues);
+   return -EINVAL;
+   }
+
+   /* Save mbuf pool pointer */
+   avp->pool = pool;
+
+   /* Save the local mbuf size */
+   mbp_priv = rte_mempool_get_priv(pool);
+   avp->guest_mbuf_size = (uint16_t)(mbp_priv->mbuf_data_room_size);
+   avp->guest_mbuf_size -= RTE_PKTMBUF_HEADROOM;
+
+   PMD_DRV_LOG(DEBUG, "AVP max_rx_pkt_len=(%u,%u) mbuf_size=(%u,%u)\n",
+   avp->max_rx_p

[dpdk-dev] [PATCH v5 05/14] net/avp: device initialization

[Allain Legacy]

Adds support for initialization newly probed AVP PCI devices.  Initial
queue translations are setup in preparation for device configuration.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 drivers/net/avp/avp_ethdev.c | 315 +++
 1 file changed, 315 insertions(+)

diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
index f5fa453..e937fb52 100644
--- a/drivers/net/avp/avp_ethdev.c
+++ b/drivers/net/avp/avp_ethdev.c
@@ -52,6 +52,7 @@
 #include 
 #include 
 #include 
+#include 
 
 #include "rte_avp_common.h"
 #include "rte_avp_fifo.h"
@@ -98,6 +99,15 @@
 };
 
 
+/**@{ AVP device flags */
+#define AVP_F_PROMISC (1 << 1)
+#define AVP_F_CONFIGURED (1 << 2)
+#define AVP_F_LINKUP (1 << 3)
+/**@} */
+
+/* Ethernet device validation marker */
+#define AVP_ETHDEV_MAGIC 0x92972862
+
 /*
  * Defines the AVP device attributes which are attached to an RTE ethernet
  * device
@@ -142,18 +152,292 @@ struct avp_adapter {
struct avp_dev avp;
 } __rte_cache_aligned;
 
+
+/* 32-bit MMIO register write */
+#define AVP_WRITE32(_value, _addr) rte_write32_relaxed((_value), (_addr))
+
+/* 32-bit MMIO register read */
+#define AVP_READ32(_addr) rte_read32_relaxed((_addr))
+
 /* Macro to cast the ethernet device private data to a AVP object */
 #define AVP_DEV_PRIVATE_TO_HW(adapter) \
(&((struct avp_adapter *)adapter)->avp)
 
 /*
+ * Defines the structure of a AVP device queue for the purpose of handling the
+ * receive and transmit burst callback functions
+ */
+struct avp_queue {
+   struct rte_eth_dev_data *dev_data;
+   /**< Backpointer to ethernet device data */
+   struct avp_dev *avp; /**< Backpointer to AVP device */
+   uint16_t queue_id;
+   /**< Queue identifier used for indexing current queue */
+   uint16_t queue_base;
+   /**< Base queue identifier for queue servicing */
+   uint16_t queue_limit;
+   /**< Maximum queue identifier for queue servicing */
+
+   uint64_t packets;
+   uint64_t bytes;
+   uint64_t errors;
+};
+
+/* translate from host physical address to guest virtual address */
+static void *
+avp_dev_translate_address(struct rte_eth_dev *eth_dev,
+ phys_addr_t host_phys_addr)
+{
+   struct rte_pci_device *pci_dev = AVP_DEV_TO_PCI(eth_dev);
+   struct rte_mem_resource *resource;
+   struct rte_avp_memmap_info *info;
+   struct rte_avp_memmap *map;
+   off_t offset;
+   void *addr;
+   unsigned int i;
+
+   addr = pci_dev->mem_resource[RTE_AVP_PCI_MEMORY_BAR].addr;
+   resource = &pci_dev->mem_resource[RTE_AVP_PCI_MEMMAP_BAR];
+   info = (struct rte_avp_memmap_info *)resource->addr;
+
+   offset = 0;
+   for (i = 0; i < info->nb_maps; i++) {
+   /* search all segments looking for a matching address */
+   map = &info->maps[i];
+
+   if ((host_phys_addr >= map->phys_addr) &&
+   (host_phys_addr < (map->phys_addr + map->length))) {
+   /* address is within this segment */
+   offset += (host_phys_addr - map->phys_addr);
+   addr = RTE_PTR_ADD(addr, offset);
+
+   PMD_DRV_LOG(DEBUG, "Translating host physical 0x%" 
PRIx64 " to guest virtual 0x%p\n",
+   host_phys_addr, addr);
+
+   return addr;
+   }
+   offset += map->length;
+   }
+
+   return NULL;
+}
+
+/* verify that the incoming device version is compatible with our version */
+static int
+avp_dev_version_check(uint32_t version)
+{
+   uint32_t driver = RTE_AVP_STRIP_MINOR_VERSION(AVP_DPDK_DRIVER_VERSION);
+   uint32_t device = RTE_AVP_STRIP_MINOR_VERSION(version);
+
+   if (device <= driver) {
+   /* the host driver version is less than or equal to ours */
+   return 0;
+   }
+
+   return 1;
+}
+
+/* verify that memory regions have expected version and validation markers */
+static int
+avp_dev_check_regions(struct rte_eth_dev *eth_dev)
+{
+   struct rte_pci_device *pci_dev = AVP_DEV_TO_PCI(eth_dev);
+   struct rte_avp_memmap_info *memmap;
+   struct rte_avp_device_info *info;
+   struct rte_mem_resource *resource;
+   unsigned int i;
+
+   /* Dump resource info for debug */
+   for (i = 0; i < PCI_MAX_RESOURCE; i++) {
+   resource = &pci_dev->mem_resource[i];
+   if ((resource->phys_addr == 0) || (resource->len == 0))
+   continue;
+
+   PMD_DRV_LOG(DEBUG, "resource[%u]: phys=0x%" PRIx64 " len=%" 
PRIu64 " addr=%p\n",
+   i, resource->phys_addr,
+   resource->len, resource->addr);
+
+   switch (i) {
+   case RTE_AVP_PCI_MEMMAP_BAR:
+   memmap = (struct rte_avp_memmap_info *)resource->addr;
+   i

[dpdk-dev] [PATCH v5 01/14] drivers/net: adds AVP PMD base files

[Allain Legacy]

This commit introduces the AVP PMD file structure without adding any actual
driver functionality.  Functional blocks will be added in later patches.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 MAINTAINERS |  5 
 config/common_base  |  5 
 doc/guides/nics/features/avp.ini|  6 +
 drivers/net/Makefile|  1 +
 drivers/net/avp/Makefile| 47 +
 drivers/net/avp/rte_pmd_avp_version.map |  4 +++
 6 files changed, 68 insertions(+)
 create mode 100644 doc/guides/nics/features/avp.ini
 create mode 100644 drivers/net/avp/Makefile
 create mode 100644 drivers/net/avp/rte_pmd_avp_version.map

diff --git a/MAINTAINERS b/MAINTAINERS
index 39bc78e..4e9aa00 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -416,6 +416,11 @@ Null Networking PMD
 M: Tetsuya Mukawa 
 F: drivers/net/null/
 
+Wind River AVP PMD
+M: Allain Legacy 
+M: Matt Peters 
+F: drivers/net/avp
+
 
 Crypto Drivers
 --
diff --git a/config/common_base b/config/common_base
index 37aa1e1..a4ad577 100644
--- a/config/common_base
+++ b/config/common_base
@@ -353,6 +353,11 @@ CONFIG_RTE_LIBRTE_QEDE_FW=""
 CONFIG_RTE_LIBRTE_PMD_AF_PACKET=n
 
 #
+# Compile WRS accelerated virtual port (AVP) guest PMD driver
+#
+CONFIG_RTE_LIBRTE_AVP_PMD=n
+
+#
 # Compile the TAP PMD
 # It is enabled by default for Linux only.
 #
diff --git a/doc/guides/nics/features/avp.ini b/doc/guides/nics/features/avp.ini
new file mode 100644
index 000..4353929
--- /dev/null
+++ b/doc/guides/nics/features/avp.ini
@@ -0,0 +1,6 @@
+;
+; Supported features of the 'AVP' network poll mode driver.
+;
+; Refer to default.ini for the full list of available PMD features.
+;
+[Features]
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index a16f25e..52b9297 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -32,6 +32,7 @@
 include $(RTE_SDK)/mk/rte.vars.mk
 
 DIRS-$(CONFIG_RTE_LIBRTE_PMD_AF_PACKET) += af_packet
+DIRS-$(CONFIG_RTE_LIBRTE_AVP_PMD) += avp
 DIRS-$(CONFIG_RTE_LIBRTE_BNX2X_PMD) += bnx2x
 DIRS-$(CONFIG_RTE_LIBRTE_PMD_BOND) += bonding
 DIRS-$(CONFIG_RTE_LIBRTE_CXGBE_PMD) += cxgbe
diff --git a/drivers/net/avp/Makefile b/drivers/net/avp/Makefile
new file mode 100644
index 000..c6e03d5
--- /dev/null
+++ b/drivers/net/avp/Makefile
@@ -0,0 +1,47 @@
+#   BSD LICENSE
+#
+#   Copyright(c) 2013-2017, Wind River Systems, Inc. All rights reserved.
+#   All rights reserved.
+#
+#   Redistribution and use in source and binary forms, with or without
+#   modification, are permitted provided that the following conditions
+#   are met:
+#
+# * Redistributions of source code must retain the above copyright
+#   notice, this list of conditions and the following disclaimer.
+# * Redistributions in binary form must reproduce the above copyright
+#   notice, this list of conditions and the following disclaimer in
+#   the documentation and/or other materials provided with the
+#   distribution.
+# * Neither the name of Wind River Systems nor the names of its
+#   contributors may be used to endorse or promote products derived
+#   from this software without specific prior written permission.
+#
+#   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+#   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+#   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+#   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+#   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+#   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+#   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+#   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+#   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+#   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+#   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+#
+# library name
+#
+LIB = librte_pmd_avp.a
+
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS) -I$(SRCDIR)
+
+EXPORT_MAP := rte_pmd_avp_version.map
+
+LIBABIVER := 1
+
+include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/net/avp/rte_pmd_avp_version.map 
b/drivers/net/avp/rte_pmd_avp_version.map
new file mode 100644
index 000..af8f3f4
--- /dev/null
+++ b/drivers/net/avp/rte_pmd_avp_version.map
@@ -0,0 +1,4 @@
+DPDK_17.05 {
+
+local: *;
+};
-- 
1.8.3.1

[dpdk-dev] [PATCH v5 10/14] net/avp: device statistics operations

[Allain Legacy]

Adds device functions to query and reset statistics.

Signed-off-by: Allain Legacy 
---
 doc/guides/nics/features/avp.ini |  2 ++
 drivers/net/avp/avp_ethdev.c | 67 
 2 files changed, 69 insertions(+)

diff --git a/doc/guides/nics/features/avp.ini b/doc/guides/nics/features/avp.ini
index e748ea8..0de761c 100644
--- a/doc/guides/nics/features/avp.ini
+++ b/doc/guides/nics/features/avp.ini
@@ -9,5 +9,7 @@ Jumbo frame  = Y
 Scattered Rx = Y
 Unicast MAC filter   = Y
 VLAN offload = Y
+Basic stats  = Y
+Stats per queue  = Y
 Linux UIO= Y
 x86-64   = Y
diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
index 07efd42..f24c6a8 100644
--- a/drivers/net/avp/avp_ethdev.c
+++ b/drivers/net/avp/avp_ethdev.c
@@ -99,6 +99,10 @@ static uint16_t avp_xmit_pkts(void *tx_queue,
 static void avp_dev_rx_queue_release(void *rxq);
 static void avp_dev_tx_queue_release(void *txq);
 
+static void avp_dev_stats_get(struct rte_eth_dev *dev,
+ struct rte_eth_stats *stats);
+static void avp_dev_stats_reset(struct rte_eth_dev *dev);
+
 
 #define AVP_DEV_TO_PCI(eth_dev) RTE_DEV_TO_PCI((eth_dev)->device)
 
@@ -146,6 +150,8 @@ static uint16_t avp_xmit_pkts(void *tx_queue,
.dev_configure   = avp_dev_configure,
.dev_infos_get   = avp_dev_info_get,
.vlan_offload_set= avp_vlan_offload_set,
+   .stats_get   = avp_dev_stats_get,
+   .stats_reset = avp_dev_stats_reset,
.link_update = avp_dev_link_update,
.rx_queue_setup  = avp_dev_rx_queue_setup,
.rx_queue_release= avp_dev_rx_queue_release,
@@ -1720,6 +1726,67 @@ struct avp_queue {
}
 }
 
+static void
+avp_dev_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   unsigned int i;
+
+   for (i = 0; i < avp->num_rx_queues; i++) {
+   struct avp_queue *rxq = avp->dev_data->rx_queues[i];
+
+   if (rxq) {
+   stats->ipackets += rxq->packets;
+   stats->ibytes += rxq->bytes;
+   stats->ierrors += rxq->errors;
+
+   stats->q_ipackets[i] += rxq->packets;
+   stats->q_ibytes[i] += rxq->bytes;
+   stats->q_errors[i] += rxq->errors;
+   }
+   }
+
+   for (i = 0; i < avp->num_tx_queues; i++) {
+   struct avp_queue *txq = avp->dev_data->tx_queues[i];
+
+   if (txq) {
+   stats->opackets += txq->packets;
+   stats->obytes += txq->bytes;
+   stats->oerrors += txq->errors;
+
+   stats->q_opackets[i] += txq->packets;
+   stats->q_obytes[i] += txq->bytes;
+   stats->q_errors[i] += txq->errors;
+   }
+   }
+}
+
+static void
+avp_dev_stats_reset(struct rte_eth_dev *eth_dev)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   unsigned int i;
+
+   for (i = 0; i < avp->num_rx_queues; i++) {
+   struct avp_queue *rxq = avp->dev_data->rx_queues[i];
+
+   if (rxq) {
+   rxq->bytes = 0;
+   rxq->packets = 0;
+   rxq->errors = 0;
+   }
+   }
+
+   for (i = 0; i < avp->num_tx_queues; i++) {
+   struct avp_queue *txq = avp->dev_data->tx_queues[i];
+
+   if (txq) {
+   txq->bytes = 0;
+   txq->packets = 0;
+   txq->errors = 0;
+   }
+   }
+}
 
 RTE_PMD_REGISTER_PCI(rte_avp, rte_avp_pmd.pci_drv);
 RTE_PMD_REGISTER_PCI_TABLE(rte_avp, pci_id_avp_map);
-- 
1.8.3.1

[dpdk-dev] [PATCH v5 09/14] net/avp: packet transmit functions

[Allain Legacy]

Adds support for packet transmit functions so that an application can send
packets to the host application via an AVP device queue.  Both the simple
and scattered functions are supported.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 drivers/net/avp/avp_ethdev.c | 335 +++
 1 file changed, 335 insertions(+)

diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
index 7524e1a..07efd42 100644
--- a/drivers/net/avp/avp_ethdev.c
+++ b/drivers/net/avp/avp_ethdev.c
@@ -87,12 +87,24 @@ static uint16_t avp_recv_scattered_pkts(void *rx_queue,
 static uint16_t avp_recv_pkts(void *rx_queue,
  struct rte_mbuf **rx_pkts,
  uint16_t nb_pkts);
+
+static uint16_t avp_xmit_scattered_pkts(void *tx_queue,
+   struct rte_mbuf **tx_pkts,
+   uint16_t nb_pkts);
+
+static uint16_t avp_xmit_pkts(void *tx_queue,
+ struct rte_mbuf **tx_pkts,
+ uint16_t nb_pkts);
+
 static void avp_dev_rx_queue_release(void *rxq);
 static void avp_dev_tx_queue_release(void *txq);
+
+
 #define AVP_DEV_TO_PCI(eth_dev) RTE_DEV_TO_PCI((eth_dev)->device)
 
 
 #define AVP_MAX_RX_BURST 64
+#define AVP_MAX_TX_BURST 64
 #define AVP_MAX_MAC_ADDRS 1
 #define AVP_MIN_RX_BUFSIZE ETHER_MIN_LEN
 
@@ -646,6 +658,7 @@ struct avp_queue {
pci_dev = AVP_DEV_TO_PCI(eth_dev);
eth_dev->dev_ops = &avp_eth_dev_ops;
eth_dev->rx_pkt_burst = &avp_recv_pkts;
+   eth_dev->tx_pkt_burst = &avp_xmit_pkts;
 
if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
/*
@@ -657,6 +670,7 @@ struct avp_queue {
if (eth_dev->data->scattered_rx) {
PMD_DRV_LOG(NOTICE, "AVP device configured for chained 
mbufs\n");
eth_dev->rx_pkt_burst = avp_recv_scattered_pkts;
+   eth_dev->tx_pkt_burst = avp_xmit_scattered_pkts;
}
return 0;
}
@@ -788,6 +802,7 @@ struct avp_queue {
PMD_DRV_LOG(NOTICE, "AVP device configured for chained 
mbufs\n");
eth_dev->data->scattered_rx = 1;
eth_dev->rx_pkt_burst = avp_recv_scattered_pkts;
+   eth_dev->tx_pkt_burst = avp_xmit_scattered_pkts;
}
}
 
@@ -1250,6 +1265,326 @@ struct avp_queue {
return count;
 }
 
+/*
+ * Copy a chained mbuf to a set of host buffers.  This function assumes that
+ * there are sufficient destination buffers to contain the entire source
+ * packet.
+ */
+static inline uint16_t
+avp_dev_copy_to_buffers(struct avp_dev *avp,
+   struct rte_mbuf *mbuf,
+   struct rte_avp_desc **buffers,
+   unsigned int count)
+{
+   struct rte_avp_desc *previous_buf = NULL;
+   struct rte_avp_desc *first_buf = NULL;
+   struct rte_avp_desc *pkt_buf;
+   struct rte_avp_desc *buf;
+   size_t total_length;
+   struct rte_mbuf *m;
+   size_t copy_length;
+   size_t src_offset;
+   char *pkt_data;
+   unsigned int i;
+
+   __rte_mbuf_sanity_check(mbuf, 1);
+
+   m = mbuf;
+   src_offset = 0;
+   total_length = rte_pktmbuf_pkt_len(m);
+   for (i = 0; (i < count) && (m != NULL); i++) {
+   /* fill each destination buffer */
+   buf = buffers[i];
+
+   if (i < count - 1) {
+   /* prefetch next entry while processing this one */
+   pkt_buf = avp_dev_translate_buffer(avp, buffers[i + 1]);
+   rte_prefetch0(pkt_buf);
+   }
+
+   /* Adjust pointers for guest addressing */
+   pkt_buf = avp_dev_translate_buffer(avp, buf);
+   pkt_data = avp_dev_translate_buffer(avp, pkt_buf->data);
+
+   /* setup the buffer chain */
+   if (previous_buf != NULL)
+   previous_buf->next = buf;
+   else
+   first_buf = pkt_buf;
+
+   previous_buf = pkt_buf;
+
+   do {
+   /*
+* copy as many source mbuf segments as will fit in the
+* destination buffer.
+*/
+   copy_length = RTE_MIN((avp->host_mbuf_size -
+  pkt_buf->data_len),
+ (rte_pktmbuf_data_len(m) -
+  src_offset));
+   rte_memcpy(RTE_PTR_ADD(pkt_data, pkt_buf->data_len),
+  RTE_PTR_ADD(rte_pktmbuf_mtod(m, void *),
+  src_offset),
+  copy_length);
+   pkt_

[dpdk-dev] [PATCH v5 11/14] net/avp: device promiscuous functions

[Allain Legacy]

Adds support for setting and clearing promiscuous mode on an AVP device.
When enabled the _mac_filter function will allow packets destined to any
MAC address to be processed by the receive functions.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 doc/guides/nics/features/avp.ini |  1 +
 drivers/net/avp/avp_ethdev.c | 28 
 2 files changed, 29 insertions(+)

diff --git a/doc/guides/nics/features/avp.ini b/doc/guides/nics/features/avp.ini
index 0de761c..ceb6993 100644
--- a/doc/guides/nics/features/avp.ini
+++ b/doc/guides/nics/features/avp.ini
@@ -7,6 +7,7 @@
 Link status  = Y
 Jumbo frame  = Y
 Scattered Rx = Y
+Promiscuous mode = Y
 Unicast MAC filter   = Y
 VLAN offload = Y
 Basic stats  = Y
diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
index f24c6a8..d008e36 100644
--- a/drivers/net/avp/avp_ethdev.c
+++ b/drivers/net/avp/avp_ethdev.c
@@ -67,6 +67,9 @@ static void avp_dev_info_get(struct rte_eth_dev *dev,
 static void avp_vlan_offload_set(struct rte_eth_dev *dev, int mask);
 static int avp_dev_link_update(struct rte_eth_dev *dev,
   __rte_unused int wait_to_complete);
+static void avp_dev_promiscuous_enable(struct rte_eth_dev *dev);
+static void avp_dev_promiscuous_disable(struct rte_eth_dev *dev);
+
 static int avp_dev_rx_queue_setup(struct rte_eth_dev *dev,
  uint16_t rx_queue_id,
  uint16_t nb_rx_desc,
@@ -153,6 +156,8 @@ static void avp_dev_stats_get(struct rte_eth_dev *dev,
.stats_get   = avp_dev_stats_get,
.stats_reset = avp_dev_stats_reset,
.link_update = avp_dev_link_update,
+   .promiscuous_enable  = avp_dev_promiscuous_enable,
+   .promiscuous_disable = avp_dev_promiscuous_disable,
.rx_queue_setup  = avp_dev_rx_queue_setup,
.rx_queue_release= avp_dev_rx_queue_release,
.tx_queue_setup  = avp_dev_tx_queue_setup,
@@ -1679,6 +1684,29 @@ struct avp_queue {
return -1;
 }
 
+static void
+avp_dev_promiscuous_enable(struct rte_eth_dev *eth_dev)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+
+   if ((avp->flags & AVP_F_PROMISC) == 0) {
+   avp->flags |= AVP_F_PROMISC;
+   PMD_DRV_LOG(DEBUG, "Promiscuous mode enabled on %u\n",
+   eth_dev->data->port_id);
+   }
+}
+
+static void
+avp_dev_promiscuous_disable(struct rte_eth_dev *eth_dev)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+
+   if ((avp->flags & AVP_F_PROMISC) != 0) {
+   avp->flags &= ~AVP_F_PROMISC;
+   PMD_DRV_LOG(DEBUG, "Promiscuous mode disabled on %u\n",
+   eth_dev->data->port_id);
+   }
+}
 
 static void
 avp_dev_info_get(struct rte_eth_dev *eth_dev,
-- 
1.8.3.1

[dpdk-dev] [PATCH v5 13/14] net/avp: migration interrupt handling

[Allain Legacy]

This commit introduces changes required to support VM live-migration.  This
is done by registering and responding to interrupts coming from the host to
signal that the memory is about to be made invalid and replaced with a new
memory zone on the destination compute node.

Enabling and disabling of the interrupts are maintained outside of the
start/stop functions because they must be enabled for the lifetime of the
device.  This is so that host interrupts are serviced and acked even in
cases where the app may have stopped the device.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 drivers/net/avp/avp_ethdev.c | 372 +++
 1 file changed, 372 insertions(+)

diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
index 9824190..e166867 100644
--- a/drivers/net/avp/avp_ethdev.c
+++ b/drivers/net/avp/avp_ethdev.c
@@ -48,6 +48,7 @@
 #include 
 #include 
 #include 
+#include 
 #include 
 #include 
 #include 
@@ -60,6 +61,8 @@
 #include "avp_logs.h"
 
 
+static int avp_dev_create(struct rte_pci_device *pci_dev,
+ struct rte_eth_dev *eth_dev);
 
 static int avp_dev_configure(struct rte_eth_dev *dev);
 static int avp_dev_start(struct rte_eth_dev *dev);
@@ -174,6 +177,7 @@ static void avp_dev_stats_get(struct rte_eth_dev *dev,
 #define AVP_F_PROMISC (1 << 1)
 #define AVP_F_CONFIGURED (1 << 2)
 #define AVP_F_LINKUP (1 << 3)
+#define AVP_F_DETACHED (1 << 4)
 /**@} */
 
 /* Ethernet device validation marker */
@@ -209,6 +213,9 @@ struct avp_dev {
struct rte_avp_fifo *free_q[RTE_AVP_MAX_QUEUES];
/**< To be freed mbufs queue */
 
+   /* mutual exclusion over the 'flag' and 'resp_q/req_q' fields */
+   rte_spinlock_t lock;
+
/* For request & response */
struct rte_avp_fifo *req_q; /**< Request queue */
struct rte_avp_fifo *resp_q; /**< Response queue */
@@ -496,6 +503,46 @@ struct avp_queue {
return 0;
 }
 
+static int
+avp_dev_detach(struct rte_eth_dev *eth_dev)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   int ret;
+
+   PMD_DRV_LOG(NOTICE, "Detaching port %u from AVP device 0x%" PRIx64 "\n",
+   eth_dev->data->port_id, avp->device_id);
+
+   rte_spinlock_lock(&avp->lock);
+
+   if (avp->flags & AVP_F_DETACHED) {
+   PMD_DRV_LOG(NOTICE, "port %u already detached\n",
+   eth_dev->data->port_id);
+   ret = 0;
+   goto unlock;
+   }
+
+   /* shutdown the device first so the host stops sending us packets. */
+   ret = avp_dev_ctrl_shutdown(eth_dev);
+   if (ret < 0) {
+   PMD_DRV_LOG(ERR, "Failed to send/recv shutdown to host, 
ret=%d\n",
+   ret);
+   avp->flags &= ~AVP_F_DETACHED;
+   goto unlock;
+   }
+
+   avp->flags |= AVP_F_DETACHED;
+   rte_wmb();
+
+   /* wait for queues to acknowledge the presence of the detach flag */
+   rte_delay_ms(1);
+
+   ret = 0;
+
+unlock:
+   rte_spinlock_unlock(&avp->lock);
+   return ret;
+}
+
 static void
 _avp_set_rx_queue_mappings(struct rte_eth_dev *eth_dev, uint16_t rx_queue_id)
 {
@@ -564,6 +611,240 @@ struct avp_queue {
avp->num_tx_queues, avp->num_rx_queues);
 }
 
+static int
+avp_dev_attach(struct rte_eth_dev *eth_dev)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   struct rte_avp_device_config config;
+   unsigned int i;
+   int ret;
+
+   PMD_DRV_LOG(NOTICE, "Attaching port %u to AVP device 0x%" PRIx64 "\n",
+   eth_dev->data->port_id, avp->device_id);
+
+   rte_spinlock_lock(&avp->lock);
+
+   if (!(avp->flags & AVP_F_DETACHED)) {
+   PMD_DRV_LOG(NOTICE, "port %u already attached\n",
+   eth_dev->data->port_id);
+   ret = 0;
+   goto unlock;
+   }
+
+   /*
+* make sure that the detached flag is set prior to reconfiguring the
+* queues.
+*/
+   avp->flags |= AVP_F_DETACHED;
+   rte_wmb();
+
+   /*
+* re-run the device create utility which will parse the new host info
+* and setup the AVP device queue pointers.
+*/
+   ret = avp_dev_create(AVP_DEV_TO_PCI(eth_dev), eth_dev);
+   if (ret < 0) {
+   PMD_DRV_LOG(ERR, "Failed to re-create AVP device, ret=%d\n",
+   ret);
+   goto unlock;
+   }
+
+   if (avp->flags & AVP_F_CONFIGURED) {
+   /*
+* Update the receive queue mapping to handle cases where the
+* source and destination hosts have different queue
+* requirements.  As long as the DETACHED flag is asserted the
+* queue table should not be referenced so it should be safe to
+* update it.
+*/
+ 

[dpdk-dev] [PATCH v5 12/14] net/avp: device start and stop operations

[Allain Legacy]

Adds support for device start and stop functions.  This allows an
application to control the administrative state of an AVP device.  Stopping
the device will notify the host application to stop sending packets on that
device's receive queues.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 drivers/net/avp/avp_ethdev.c | 102 +++
 1 file changed, 102 insertions(+)

diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
index d008e36..9824190 100644
--- a/drivers/net/avp/avp_ethdev.c
+++ b/drivers/net/avp/avp_ethdev.c
@@ -62,6 +62,9 @@
 
 
 static int avp_dev_configure(struct rte_eth_dev *dev);
+static int avp_dev_start(struct rte_eth_dev *dev);
+static void avp_dev_stop(struct rte_eth_dev *dev);
+static void avp_dev_close(struct rte_eth_dev *dev);
 static void avp_dev_info_get(struct rte_eth_dev *dev,
 struct rte_eth_dev_info *dev_info);
 static void avp_vlan_offload_set(struct rte_eth_dev *dev, int mask);
@@ -151,6 +154,9 @@ static void avp_dev_stats_get(struct rte_eth_dev *dev,
  */
 static const struct eth_dev_ops avp_eth_dev_ops = {
.dev_configure   = avp_dev_configure,
+   .dev_start   = avp_dev_start,
+   .dev_stop= avp_dev_stop,
+   .dev_close   = avp_dev_close,
.dev_infos_get   = avp_dev_info_get,
.vlan_offload_set= avp_vlan_offload_set,
.stats_get   = avp_dev_stats_get,
@@ -316,6 +322,23 @@ struct avp_queue {
 }
 
 static int
+avp_dev_ctrl_set_link_state(struct rte_eth_dev *eth_dev, unsigned int state)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   struct rte_avp_request request;
+   int ret;
+
+   /* setup a link state change request */
+   memset(&request, 0, sizeof(request));
+   request.req_id = RTE_AVP_REQ_CFG_NETWORK_IF;
+   request.if_up = state;
+
+   ret = avp_dev_process_request(avp, &request);
+
+   return ret == 0 ? request.result : ret;
+}
+
+static int
 avp_dev_ctrl_set_config(struct rte_eth_dev *eth_dev,
struct rte_avp_device_config *config)
 {
@@ -333,6 +356,22 @@ struct avp_queue {
return ret == 0 ? request.result : ret;
 }
 
+static int
+avp_dev_ctrl_shutdown(struct rte_eth_dev *eth_dev)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   struct rte_avp_request request;
+   int ret;
+
+   /* setup a shutdown request */
+   memset(&request, 0, sizeof(request));
+   request.req_id = RTE_AVP_REQ_SHUTDOWN_DEVICE;
+
+   ret = avp_dev_process_request(avp, &request);
+
+   return ret == 0 ? request.result : ret;
+}
+
 /* translate from host mbuf virtual address to guest virtual address */
 static inline void *
 avp_dev_translate_buffer(struct avp_dev *avp, void *host_mbuf_address)
@@ -1669,6 +1708,69 @@ struct avp_queue {
return ret;
 }
 
+static int
+avp_dev_start(struct rte_eth_dev *eth_dev)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   int ret;
+
+   /* disable features that we do not support */
+   eth_dev->data->dev_conf.rxmode.hw_ip_checksum = 0;
+   eth_dev->data->dev_conf.rxmode.hw_vlan_filter = 0;
+   eth_dev->data->dev_conf.rxmode.hw_vlan_extend = 0;
+   eth_dev->data->dev_conf.rxmode.hw_strip_crc = 0;
+
+   /* update link state */
+   ret = avp_dev_ctrl_set_link_state(eth_dev, 1);
+   if (ret < 0) {
+   PMD_DRV_LOG(ERR, "Link state change failed by host, ret=%d\n",
+   ret);
+   goto unlock;
+   }
+
+   /* remember current link state */
+   avp->flags |= AVP_F_LINKUP;
+
+   ret = 0;
+
+unlock:
+   return ret;
+}
+
+static void
+avp_dev_stop(struct rte_eth_dev *eth_dev)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   int ret;
+
+   avp->flags &= ~AVP_F_LINKUP;
+
+   /* update link state */
+   ret = avp_dev_ctrl_set_link_state(eth_dev, 0);
+   if (ret < 0) {
+   PMD_DRV_LOG(ERR, "Link state change failed by host, ret=%d\n",
+   ret);
+   }
+}
+
+static void
+avp_dev_close(struct rte_eth_dev *eth_dev)
+{
+   struct avp_dev *avp = AVP_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
+   int ret;
+
+   /* remember current link state */
+   avp->flags &= ~AVP_F_LINKUP;
+   avp->flags &= ~AVP_F_CONFIGURED;
+
+   /* update device state */
+   ret = avp_dev_ctrl_shutdown(eth_dev);
+   if (ret < 0) {
+   PMD_DRV_LOG(ERR, "Device shutdown failed by host, ret=%d\n",
+   ret);
+   /* continue */
+   }
+}
 
 static int
 avp_dev_link_update(struct rte_eth_dev *eth_dev,
-- 
1.8.3.1

[dpdk-dev] [PATCH v5 14/14] doc: adds information related to the AVP PMD

[Allain Legacy]

Updates the documentation and feature lists for the AVP PMD device.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
Acked-by: John McNamara 
---
 MAINTAINERS|   1 +
 doc/guides/nics/avp.rst| 107 +
 doc/guides/nics/index.rst  |   1 +
 doc/guides/rel_notes/release_17_05.rst |   5 ++
 4 files changed, 114 insertions(+)
 create mode 100644 doc/guides/nics/avp.rst

diff --git a/MAINTAINERS b/MAINTAINERS
index 4e9aa00..ec27449 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -420,6 +420,7 @@ Wind River AVP PMD
 M: Allain Legacy 
 M: Matt Peters 
 F: drivers/net/avp
+F: doc/guides/nics/avp.rst
 
 
 Crypto Drivers
diff --git a/doc/guides/nics/avp.rst b/doc/guides/nics/avp.rst
new file mode 100644
index 000..c301f65
--- /dev/null
+++ b/doc/guides/nics/avp.rst
@@ -0,0 +1,107 @@
+..  BSD LICENSE
+Copyright(c) 2017 Wind River Systems, Inc. rights reserved.
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions
+are met:
+
+* Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+* Redistributions in binary form must reproduce the above copyright
+notice, this list of conditions and the following disclaimer in
+the documentation and/or other materials provided with the
+distribution.
+* Neither the name of Intel Corporation nor the names of its
+contributors may be used to endorse or promote products derived
+from this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+AVP Poll Mode Driver
+=
+
+The Accelerated Virtual Port (AVP) device is a shared memory based device
+available on the `virtualization platforms 
`_
+from Wind River Systems.
+
+It enables optimized packet throughput without requiring any packet processing
+in qemu. This provides our customers with a significant performance increase
+for DPDK applications in the VM.  Since our AVP implementation supports VM
+live-migration it is viewed as a better alternative to PCI passthrough or PCI
+SRIOV since neither of those support VM live-migration without manual
+intervention or significant performance penalties.
+
+The driver binds to PCI devices that are exported by the hypervisor DPDK
+application via a shared memory mechanism.  It supports a subset of the full
+Ethernet device API which enables the application to use the typical
+configuration and packet transfer functions.
+
+The definition of the device structure and configuration options are defined in
+rte_avp_common.h and rte_avp_fifo.h public header files.  These two header
+files are made available as part of the PMD implementation in order to share
+the device definitions between the guest implementation (i.e., the PMD) and the
+host implementation (i.e., the hypervisor DPDK vswitch application).
+
+
+Features and Limitations of the AVP PMD
+---
+
+The AVP PMD driver provides the following functionality.
+
+*   Receive and transmit of both simple and chained mbuf packets,
+
+*   Chained mbufs may include up to 5 chained segments,
+
+*   Up to 8 receive and transmit queues per device,
+
+*   Only a single MAC address is supported,
+
+*   The MAC address cannot be modified,
+
+*   The maximum receive packet length is 9238 bytes,
+
+*   VLAN header stripping and inserting,
+
+*   Promiscuous mode
+
+*   VM live-migration
+
+*   PCI hotplug insertion and removal
+
+
+Prerequisites
+-
+
+The following prerequisites apply:
+
+*   A virtual machine running in a Wind River Systems virtualization
+environment and configured with at least one neutron port defined with a
+vif-model set to "avp".
+
+
+Launching a VM with an AVP type network attachment
+--
+
+The following example will launch a VM with three network attachments.  The
+first attachment will have a default 

[dpdk-dev] [PATCH v5 08/14] net/avp: packet receive functions

[Allain Legacy]

Adds function required for receiving packets from the host application via
AVP device queues.  Both the simple and scattered functions are supported.

Signed-off-by: Allain Legacy 
Signed-off-by: Matt Peters 
---
 doc/guides/nics/features/avp.ini |   3 +
 drivers/net/avp/Makefile |   1 +
 drivers/net/avp/avp_ethdev.c | 451 +++
 3 files changed, 455 insertions(+)

diff --git a/doc/guides/nics/features/avp.ini b/doc/guides/nics/features/avp.ini
index 45a2185..e748ea8 100644
--- a/doc/guides/nics/features/avp.ini
+++ b/doc/guides/nics/features/avp.ini
@@ -5,6 +5,9 @@
 ;
 [Features]
 Link status  = Y
+Jumbo frame  = Y
+Scattered Rx = Y
+Unicast MAC filter   = Y
 VLAN offload = Y
 Linux UIO= Y
 x86-64   = Y
diff --git a/drivers/net/avp/Makefile b/drivers/net/avp/Makefile
index 9cf0449..3013cd1 100644
--- a/drivers/net/avp/Makefile
+++ b/drivers/net/avp/Makefile
@@ -56,5 +56,6 @@ SRCS-$(CONFIG_RTE_LIBRTE_AVP_PMD) += avp_ethdev.c
 
 # this lib depends upon:
 DEPDIRS-$(CONFIG_RTE_LIBRTE_AVP_PMD) += lib/librte_eal lib/librte_ether
+DEPDIRS-$(CONFIG_RTE_LIBRTE_AVP_PMD) += lib/librte_mempool lib/librte_mbuf
 
 include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
index ecc581f..7524e1a 100644
--- a/drivers/net/avp/avp_ethdev.c
+++ b/drivers/net/avp/avp_ethdev.c
@@ -80,11 +80,19 @@ static int avp_dev_tx_queue_setup(struct rte_eth_dev *dev,
  unsigned int socket_id,
  const struct rte_eth_txconf *tx_conf);
 
+static uint16_t avp_recv_scattered_pkts(void *rx_queue,
+   struct rte_mbuf **rx_pkts,
+   uint16_t nb_pkts);
+
+static uint16_t avp_recv_pkts(void *rx_queue,
+ struct rte_mbuf **rx_pkts,
+ uint16_t nb_pkts);
 static void avp_dev_rx_queue_release(void *rxq);
 static void avp_dev_tx_queue_release(void *txq);
 #define AVP_DEV_TO_PCI(eth_dev) RTE_DEV_TO_PCI((eth_dev)->device)
 
 
+#define AVP_MAX_RX_BURST 64
 #define AVP_MAX_MAC_ADDRS 1
 #define AVP_MIN_RX_BUFSIZE ETHER_MIN_LEN
 
@@ -302,6 +310,15 @@ struct avp_queue {
return ret == 0 ? request.result : ret;
 }
 
+/* translate from host mbuf virtual address to guest virtual address */
+static inline void *
+avp_dev_translate_buffer(struct avp_dev *avp, void *host_mbuf_address)
+{
+   return RTE_PTR_ADD(RTE_PTR_SUB(host_mbuf_address,
+  (uintptr_t)avp->host_mbuf_addr),
+  (uintptr_t)avp->mbuf_addr);
+}
+
 /* translate from host physical address to guest virtual address */
 static void *
 avp_dev_translate_address(struct rte_eth_dev *eth_dev,
@@ -628,6 +645,7 @@ struct avp_queue {
 
pci_dev = AVP_DEV_TO_PCI(eth_dev);
eth_dev->dev_ops = &avp_eth_dev_ops;
+   eth_dev->rx_pkt_burst = &avp_recv_pkts;
 
if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
/*
@@ -636,6 +654,10 @@ struct avp_queue {
 * be mapped to the same virtual address so all pointers should
 * be valid.
 */
+   if (eth_dev->data->scattered_rx) {
+   PMD_DRV_LOG(NOTICE, "AVP device configured for chained 
mbufs\n");
+   eth_dev->rx_pkt_burst = avp_recv_scattered_pkts;
+   }
return 0;
}
 
@@ -704,6 +726,38 @@ struct avp_queue {
 
 
 static int
+avp_dev_enable_scattered(struct rte_eth_dev *eth_dev,
+struct avp_dev *avp)
+{
+   unsigned int max_rx_pkt_len;
+
+   max_rx_pkt_len = eth_dev->data->dev_conf.rxmode.max_rx_pkt_len;
+
+   if ((max_rx_pkt_len > avp->guest_mbuf_size) ||
+   (max_rx_pkt_len > avp->host_mbuf_size)) {
+   /*
+* If the guest MTU is greater than either the host or guest
+* buffers then chained mbufs have to be enabled in the TX
+* direction.  It is assumed that the application will not need
+* to send packets larger than their max_rx_pkt_len (MRU).
+*/
+   return 1;
+   }
+
+   if ((avp->max_rx_pkt_len > avp->guest_mbuf_size) ||
+   (avp->max_rx_pkt_len > avp->host_mbuf_size)) {
+   /*
+* If the host MRU is greater than its own mbuf size or the
+* guest mbuf size then chained mbufs have to be enabled in the
+* RX direction.
+*/
+   return 1;
+   }
+
+   return 0;
+}
+
+static int
 avp_dev_rx_queue_setup(struct rte_eth_dev *eth_dev,
   uint16_t rx_queue_id,
   uint16_t nb_rx_desc,
@@ -729,6 +783,14 @@ struct avp_queue {
avp->guest_mbuf_size = (uint16_t)(mbp_priv->mbuf_data_room_size);
avp->guest

Re: [dpdk-dev] [PATCH v4 1/7] net/ark: PMD for Atomic Rules Arkville driver stub

[Ferruh Yigit]

On 3/23/2017 1:03 AM, Ed Czeck wrote:
> Enable Arkville on supported configurations
> Add overview documentation
> Minimum driver support for valid compile
> Arkville PMD is not supported on ARM or PowerPC at this time

Can you please send new version of patchset as reply to previous
version, using --in-reply-to ?
This helps to see all patches in same mail tread and helps seeing
previous comments, also in mail archive all versions stays in same place.

> 
> v4:
> * Address issues report from review
> * Add internal comments on driver arg
> * provide a bare-biones dev init to avoid compiler warnings
> 
> v3:
> * Split large patch into several smaller ones
> 
> Signed-off-by: Ed Czeck 
> ---

<...>

> diff --git a/config/defconfig_ppc_64-power8-linuxapp-gcc 
> b/config/defconfig_ppc_64-power8-linuxapp-gcc
> index 35f7fb6..89bc396 100644
> --- a/config/defconfig_ppc_64-power8-linuxapp-gcc
> +++ b/config/defconfig_ppc_64-power8-linuxapp-gcc
> @@ -48,6 +48,7 @@ CONFIG_RTE_LIBRTE_EAL_VMWARE_TSC_MAP_SUPPORT=n
>  
>  # Note: Initially, all of the PMD drivers compilation are turned off on Power
>  # Will turn on them only after the successful testing on Power
> +CONFIG_RTE_LIBRTE_ARK_PMD=n

Is it not tested or known that it is not supported?

>  CONFIG_RTE_LIBRTE_IXGBE_PMD=n
>  CONFIG_RTE_LIBRTE_I40E_PMD=n
>  CONFIG_RTE_LIBRTE_VIRTIO_PMD=y

<...>

> +Configuration Information
> +-
> +
> +**DPDK Configuration Parameters**
> +
> +  The following configuration options are available for the ARK PMD:
> +
> +   * **CONFIG_RTE_LIBRTE_ARK_PMD** (default y): Enables or disables inclusion
> + of the ARK PMD driver in the DPDK compilation.
> +
> +   * **CONFIG_RTE_LIBRTE_ARK_PAD_TX** (default y):  When enabled TX
> + packets are padded to 60 bytes to support downstream MACS.
> +
> +   * **CONFIG_RTE_LIBRTE_ARK_DEBUG_RX** (default n): Enables or disables 
> debug
> + logging and internal checking of RX ingress logic within the ARK PMD 
> driver.
> +
> +   * **CONFIG_RTE_LIBRTE_ARK_DEBUG_TX** (default n): Enables or disables 
> debug
> + logging and internal checking of TX egress logic within the ARK PMD 
> driver.
> +
> +   * **CONFIG_RTE_LIBRTE_ARK_DEBUG_STATS** (default n): Enables or disables 
> debug
> + logging of detailed packet and performance statistics gathered in
> + the PMD and FPGA.
> +
> +   * **CONFIG_RTE_LIBRTE_ARK_DEBUG_TRACE** (default n): Enables or disables 
> debug
> + logging of detailed PMD events and status.
> +
> +

Can you also please document the device arguments in this file?

Device Parameters
-

"Pkt_gen"
"Pkt_chkr"
"Pkt_dir"

Also perhaps you can include these device arguments in below "Usage
Example" section of this document.

It can be good to document expected format/dataset for these arguments.

> +Building DPDK
> +-
> +
> +See the :ref:`DPDK Getting Started Guide for Linux ` for
> +instructions on how to build DPDK.
> +
> +By default the ARK PMD library will be built into the DPDK library.
> +
> +For configuring and using UIO and VFIO frameworks, please also refer 
> :ref:`the
> +documentation that comes with DPDK suite `.
> +

<...>

> diff --git a/doc/guides/nics/features/ark.ini 
> b/doc/guides/nics/features/ark.ini
> new file mode 100644
> index 000..dc8a0e2
> --- /dev/null
> +++ b/doc/guides/nics/features/ark.ini
> @@ -0,0 +1,15 @@
> +;
> +; Supported features of the 'ark' poll mode driver.
> +;
> +; Refer to default.ini for the full list of available PMD features.
> +;
> +[Features]
> +Queue start/stop = Y
> +Jumbo frame  = Y
> +Scattered Rx = Y
> +Basic stats  = Y
> +Stats per queue  = Y
> +FW version   = Y

Features can be added with the patch that adds functionality. I believe
above features not supported with current patch.

> +Linux UIO= Y
> +x86-64   = Y
> +Usage doc= Y

<...>

> +#
> +# all source are stored in SRCS-y
> +#
> +SRCS-y += ark_ethdev.c

Please use SRCS-y only in comment, for actual usage please prefer:
SRCS-$(CONFIG_RTE_LIBRTE_ARK_PMD)

> +
> +
> +# this lib depends upon:
> +DEPDIRS-y += lib/librte_mbuf

DEPDIRS-$(CONFIG_RTE_LIBRTE_ARK_PMD)

> +DEPDIRS-y += lib/librte_ether

<...>

> +#define ARK_TRACE_ON(fmt, ...) \
> + PMD_DRV_LOG(ERR, fmt, ##__VA_ARGS__)
> +
> +#define ARK_TRACE_OFF(fmt, ...) \
> + do {if (0) PMD_DRV_LOG(ERR, fmt, ##__VA_ARGS__); } while (0)

why not just "do { } while(0)" ?

> +
> +/* Debug macro for reporting Packet stats */
> +#ifdef RTE_LIBRTE_ARK_DEBUG_STATS
> +#define ARK_DEBUG_STATS(fmt, ...) ARK_TRACE_ON(fmt, ##__VA_ARGS__)

This is debug option but prints only "ERR" level log, shouldn't this be
DEBUG.
Also there are dpdk wide log level option, helps optimizing out some
code, if you use only ERR type, you won't be benefiting from it.

> +#else
> +#define ARK_DEBUG_STATS(fmt, ...)  ARK_TRACE_OFF(fmt, ##__VA_ARGS__)
> +#endif
> +
> +/* Debug macro for tracing full behavior*/
>

Re: [dpdk-dev] [PATCH v4 2/7] net/ark: HW API part 1 of 3

[Ferruh Yigit]

On 3/23/2017 1:03 AM, Ed Czeck wrote:
> Provide C-level interface for Arkville's internal HW resources
> mpu, pktdir, and rqp modules

For patch title, instead of using " HW API part x of y", can you please
describe the content of the patch, like:

" net/ark: HW API for MPU, pktdir, and RQP modules"

> 
> Signed-off-by: Ed Czeck 
<...>

[dpdk-dev] [PATCH v2] doc: reformat crypto drivers overview

[Pablo de Lara]

Follow the approach in the network devices overview,
for the feature matrix, so it improves readibility
and maintainability.

Signed-off-by: Pablo de Lara 
Acked-by: John McNamara 
---

Changes in v2:
- Added missing algorithms
- Fixed PEP8 errors

 .gitignore   |   4 +
 doc/guides/conf.py   |  62 --
 doc/guides/cryptodevs/features/aesni_gcm.ini |  27 +++
 doc/guides/cryptodevs/features/aesni_mb.ini  |  41 
 doc/guides/cryptodevs/features/armv8.ini |  28 +++
 doc/guides/cryptodevs/features/default.ini   |  70 +++
 doc/guides/cryptodevs/features/kasumi.ini|  24 +++
 doc/guides/cryptodevs/features/null.ini  |  25 +++
 doc/guides/cryptodevs/features/openssl.ini   |  47 +
 doc/guides/cryptodevs/features/qat.ini   |  50 +
 doc/guides/cryptodevs/features/snow3g.ini|  24 +++
 doc/guides/cryptodevs/features/zuc.ini   |  24 +++
 doc/guides/cryptodevs/overview.rst   | 291 +--
 13 files changed, 636 insertions(+), 81 deletions(-)
 create mode 100644 doc/guides/cryptodevs/features/aesni_gcm.ini
 create mode 100644 doc/guides/cryptodevs/features/aesni_mb.ini
 create mode 100644 doc/guides/cryptodevs/features/armv8.ini
 create mode 100644 doc/guides/cryptodevs/features/default.ini
 create mode 100644 doc/guides/cryptodevs/features/kasumi.ini
 create mode 100644 doc/guides/cryptodevs/features/null.ini
 create mode 100644 doc/guides/cryptodevs/features/openssl.ini
 create mode 100644 doc/guides/cryptodevs/features/qat.ini
 create mode 100644 doc/guides/cryptodevs/features/snow3g.ini
 create mode 100644 doc/guides/cryptodevs/features/zuc.ini

diff --git a/.gitignore b/.gitignore
index a722abe..c733967 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1 +1,5 @@
 doc/guides/nics/overview_table.txt
+doc/guides/cryptodevs/overview_feature_table.txt
+doc/guides/cryptodevs/overview_cipher_table.txt
+doc/guides/cryptodevs/overview_auth_table.txt
+doc/guides/cryptodevs/overview_aead_table.txt
diff --git a/doc/guides/conf.py b/doc/guides/conf.py
index 34c62de..a7f6c99 100644
--- a/doc/guides/conf.py
+++ b/doc/guides/conf.py
@@ -175,25 +175,24 @@ def process_numref(app, doctree, from_docname):
 node.replace_self(newnode)
 
 
-def generate_nic_overview_table(output_filename):
+def generate_overview_table(output_filename, section, table_name, title):
 """
-Function to generate the NIC Overview Table from the ini files that define
-the features for each NIC.
+Function to generate the Overview Table from the ini files that define
+the features for each driver.
 
 The default features for the table and their order is defined by the
 'default.ini' file.
 
 """
-# Default worning string.
-warning = 'Warning generate_nic_overview_table()'
+# Default warning string.
+warning = 'Warning generate_crypto_overview_table()'
 
 # Get the default features and order from the 'default.ini' file.
 ini_path = path_join(dirname(output_filename), 'features')
 config = configparser.ConfigParser()
 config.optionxform = str
 config.read(path_join(ini_path, 'default.ini'))
-default_section = 'Features'
-default_features = config.items(default_section)
+default_features = config.items(section)
 
 # Create a dict of the valid features to validate the other ini files.
 valid_features = {}
@@ -203,7 +202,7 @@ def generate_nic_overview_table(output_filename):
 valid_features[key] = ' '
 max_feature_length = max(max_feature_length, len(key))
 
-# Get a list of NIC ini files, excluding 'default.ini'.
+# Get a list of driver ini files, excluding 'default.ini'.
 ini_files = [basename(file) for file in listdir(ini_path)
  if file.endswith('.ini') and file != 'default.ini']
 ini_files.sort()
@@ -223,7 +222,7 @@ def generate_nic_overview_table(output_filename):
 
 header_names.append(name)
 
-# Create a dict of the defined features for each NIC from the ini files.
+# Create a dict of the defined features for each crypto from the ini files.
 ini_data = {}
 for ini_filename in ini_files:
 config = configparser.ConfigParser()
@@ -234,14 +233,14 @@ def generate_nic_overview_table(output_filename):
 ini_data[ini_filename] = valid_features.copy()
 
 # Check for a valid ini section.
-if not config.has_section(default_section):
+if not config.has_section(section):
 print("{}: File '{}' has no [{}] secton".format(warning,
 ini_filename,
-default_section))
+section))
 continue
 
 # Check for valid features names.
-for name, value in config.items(default_section):
+for name, value in config.items(section):
 if na

Re: [dpdk-dev] [PATCH 22/39] app/test: octeontx eventdev unit test infrastructure

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 22/39] app/test: octeontx eventdev unit test 
> infrastructure
> 
> add test setup and teardown routines.
> 
> Signed-off-by: Jerin Jacob 

1) I think the autotest_data.py file needs to be updated to run the test 
command too, event/sw example at top of patch
http://dpdk.org/dev/patchwork/patch/21694/

2) This requires a rebase against latest head, app/test has moved to test/test/.
I will ignore this app/test to test/test move for the remaining patches of the 
series.

With fixes above,

Acked-by: Harry van Haaren 


> ---
>  app/test/Makefile |  5 +-
>  app/test/test_eventdev_octeontx.c | 98 
> +++
>  2 files changed, 102 insertions(+), 1 deletion(-)
>  create mode 100644 app/test/test_eventdev_octeontx.c
> 
> diff --git a/app/test/Makefile b/app/test/Makefile
> index a426548..b45a1d3 100644
> --- a/app/test/Makefile
> +++ b/app/test/Makefile
> @@ -197,7 +197,10 @@ SRCS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += 
> test_cryptodev_blockcipher.c
>  SRCS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += test_cryptodev_perf.c
>  SRCS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += test_cryptodev.c
> 
> -SRCS-$(CONFIG_RTE_LIBRTE_EVENTDEV) += test_eventdev.c
> +ifeq ($(CONFIG_RTE_LIBRTE_EVENTDEV),y)
> +SRCS-y += test_eventdev.c
> +SRCS-$(CONFIG_RTE_LIBRTE_PMD_OCTEONTX_SSOVF) += test_eventdev_octeontx.c
> +endif
> 
>  SRCS-$(CONFIG_RTE_LIBRTE_KVARGS) += test_kvargs.c
> 
> diff --git a/app/test/test_eventdev_octeontx.c 
> b/app/test/test_eventdev_octeontx.c
> new file mode 100644
> index 000..9744961
> --- /dev/null
> +++ b/app/test/test_eventdev_octeontx.c
> @@ -0,0 +1,98 @@
> +/*-
> + *   BSD LICENSE
> + *
> + *   Copyright(c) 2017 Cavium networks. All rights reserved.
> + *
> + *   Redistribution and use in source and binary forms, with or without
> + *   modification, are permitted provided that the following conditions
> + *   are met:
> + *
> + ** Redistributions of source code must retain the above copyright
> + *  notice, this list of conditions and the following disclaimer.
> + ** Redistributions in binary form must reproduce the above copyright
> + *  notice, this list of conditions and the following disclaimer in
> + *  the documentation and/or other materials provided with the
> + *  distribution.
> + ** Neither the name of Cavium networks nor the names of its
> + *  contributors may be used to endorse or promote products derived
> + *  from this software without specific prior written permission.
> + *
> + *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> + *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> + *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
> + *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> + *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
> + *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
> + *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
> + *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
> + *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
> + *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
> + *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
> + */
> +
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +#include 
> +
> +#include "test.h"
> +
> +#define NUM_PACKETS (1 << 18)
> +#define MAX_EVENTS  (16 * 1024)
> +
> +static int evdev;
> +
> +static int
> +testsuite_setup(void)
> +{
> + const char *eventdev_name = "event_octeontx";
> +
> + evdev = rte_event_dev_get_dev_id(eventdev_name);
> + if (evdev < 0) {
> + printf("%d: Eventdev %s not found - creating.\n",
> + __LINE__, eventdev_name);
> + if (rte_eal_vdev_init(eventdev_name, NULL) < 0) {
> + printf("Error creating eventdev\n");
> + return TEST_FAILED;
> + }
> + evdev = rte_event_dev_get_dev_id(eventdev_name);
> + if (evdev < 0) {
> + printf("Error finding newly created eventdev\n");
> + return TEST_FAILED;
> + }
> + }
> +
> + return TEST_SUCCESS;
> +}
> +
> +static void
> +testsuite_teardown(void)
> +{
> + rte_event_dev_close(evdev);
> +}
> +
> +
> +static struct unit_test_suite eventdev_octeontx_testsuite  = {
> + .suite_name = "eventdev octe

Re: [dpdk-dev] [PATCH 23/39] app/test: octeontx unit test case setup and teardown

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 23/39] app/test: octeontx unit test case setup and 
> teardown
> 
> Each test case expected to run as standalone.
> On setup, configure the device in requested mode and start the device.
> On tear down, close the device and free the allocated resources
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 25/39] app/test: octeontx simple event enqueue and dequeue test

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 25/39] app/test: octeontx simple event enqueue and 
> dequeue test
> 
> Added unit test case to verify simple event enqueue and dequeue
> operation with different schedule types
> 
> Signed-off-by: Jerin Jacob 


Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 24/39] app/test: octeontx unit test case helper functions

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 24/39] app/test: octeontx unit test case helper 
> functions
> 
> Add helper functions to generate, inject, consume and validate the events.
> 
> Signed-off-by: Jerin Jacob 


Acked-by: Harry van Haaren 



> +static inline void
> +update_event_and_validation_attr(struct rte_mbuf *m, struct rte_event *ev,
> + uint32_t flow_id, uint8_t event_type,
> + uint8_t sub_event_type, uint8_t sched_type,
> + uint8_t queue, uint8_t port)
> +{
> + struct event_attr *attr;
> +
> + /* Store the event attributes in mbuf for future reference */
> + attr = rte_pktmbuf_mtod(m, struct event_attr *);



> +static inline int
> +validate_event(struct rte_event *ev)
> +{
> + struct event_attr *attr;
> +
> + attr = rte_pktmbuf_mtod(ev->mbuf, struct event_attr *);
> + TEST_ASSERT_EQUAL(attr->flow_id, ev->flow_id,
> + "flow_id mismatch enq=%d deq =%d",
> + attr->flow_id, ev->flow_id);
> + TEST_ASSERT_EQUAL(attr->event_type, ev->event_type,
> + "event_type mismatch enq=%d deq =%d",
> + attr->event_type, ev->event_type);


Simple + effective technique - I'll remember that one :)

Re: [dpdk-dev] [PATCH 26/39] app/test: octeontx multi queue enqueue and dequeue test

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 26/39] app/test: octeontx multi queue enqueue and 
> dequeue test
> 
> Added unit test case to verify enqueue and dequeue operations
> with multiple queues and a single port.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 27/39] app/test: octeontx eventdev priority test

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 27/39] app/test: octeontx eventdev priority test
> 
> Added unit test case to verify the priority associated with
> each event queue available in the device.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 28/39] app/test: add infrastructure for multicore octeontx tests

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 28/39] app/test: add infrastructure for multicore 
> octeontx tests
> 
> Add helper functions to launch and wait for n cores to complete the
> operation with deadlock detection.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 29/39] app/test: octeontx multi queue and multi core/port tests

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 29/39] app/test: octeontx multi queue and multi 
> core/port tests
> 
> Add unit test case to verify multi queue enqueue and multi core/port
> dequeue operation.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 30/39] app/test: octeontx single link establishment test

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 30/39] app/test: octeontx single link 
> establishment test
> 
> Add test case to verify queue to port single link establishment operation.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 31/39] app/test: octeontx multi link establishment test

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 31/39] app/test: octeontx multi link establishment 
> test
> 
> Add unit test case to verify queue to port multi link
> establishment operation.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 32/39] app/test: octeontx flow based two stage sched type test

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 32/39] app/test: octeontx flow based two stage 
> sched type test
> 
> Add flow based two stage pipeline test with all combination
> of schedule types.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH] vfio: add hotplug support

[Burakov, Anatoly]

Hi Alejandro,

I think the documentation should also be updated to reflect the fact that this 
patch adds VFIO support for hotplug.

Thanks,
Anatoly

Re: [dpdk-dev] [PATCH 33/39] app/test: octeontx queue based two stage sched type test

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 33/39] app/test: octeontx queue based two stage 
> sched type test
> 
> Add queue based two stage pipeline test with all combination
> of schedule types.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 34/39] app/test: octeontx flow based maximum stage pipeline

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 34/39] app/test: octeontx flow based maximum stage 
> pipeline
> 
> Add flow based pipeline test with maximum number of stages available
> in the device.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 35/39] app/test: octeontx queue based maximum stage pipeline

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 35/39] app/test: octeontx queue based maximum 
> stage pipeline
> 
> Add queue based pipeline test with maximum number of stages available
> in the device.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren 

Re: [dpdk-dev] [PATCH 36/39] app/test: octeontx queue and flow based max stage pipeline

[Van Haaren, Harry]

> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce 
> ; Van Haaren,
> Harry ; hemant.agra...@nxp.com; Eads, Gage 
> ;
> nipun.gu...@nxp.com; santosh.shu...@caviumnetworks.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH 36/39] app/test: octeontx queue and flow based max 
> stage pipeline
> 
> Add queue and flow based pipeline test with maximum number of
> stages available in the device.
> 
> Signed-off-by: Jerin Jacob 

Acked-by: Harry van Haaren