[dpdk-dev] [PATCH] test: support to check AVX512F

[Zhiyong Yang]

The CPUs which support AVX512 have been released. Add the support
of checking AVX512F instruction set.

Signed-off-by: Zhiyong Yang 
---
 test/test/test_cpuflags.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/test/test/test_cpuflags.c b/test/test/test_cpuflags.c
index 0e5ebe788..08c166139 100644
--- a/test/test/test_cpuflags.c
+++ b/test/test/test_cpuflags.c
@@ -171,6 +171,9 @@ test_cpuflags(void)
printf("Check for AVX2:\t\t");
CHECK_FOR_FLAG(RTE_CPUFLAG_AVX2);
 
+   printf("Check for AVX512F:\t");
+   CHECK_FOR_FLAG(RTE_CPUFLAG_AVX512F);
+
printf("Check for TRBOBST:\t");
CHECK_FOR_FLAG(RTE_CPUFLAG_TRBOBST);
 
-- 
2.13.3

Re: [dpdk-dev] [PATCH] vhost: adaptively batch small guest memory copies

[Yuanhan Liu]

On Thu, Aug 24, 2017 at 10:19:39AM +0800, Tiwei Bie wrote:
> This patch adaptively batches the small guest memory copies.
> By batching the small copies, the efficiency of executing the
> memory LOAD instructions can be improved greatly, because the
> memory LOAD latency can be effectively hidden by the pipeline.
> We saw great performance boosts for small packets PVP test.
> 
> This patch improves the performance for small packets, and has
> distinguished the packets by size. So although the performance
> for big packets doesn't change, it makes it relatively easy to
> do some special optimizations for the big packets too.

The number showed in other replies looks really impressive. Great work!
This patch also looks good to me. I have one minor comment though.

[...]
> +/*
> + * Structure contains the info for each batched memory copy.
> + */
> +struct burst_copy_elem {
> + void *dst;
> + void *src;
> + uint32_t len;
> + uint64_t log_addr;
> +};

Like the title says, it's more about batch (but not burst). Also, it's
not a good idea to mix burst and batch. I'd suggest you to use the term
"batch" consistently.

--yliu

Re: [dpdk-dev] [PATCH 07/21] vhost: add iotlb helper functions

[Yuanhan Liu]

On Thu, Aug 31, 2017 at 11:50:09AM +0200, Maxime Coquelin wrote:
> diff --git a/lib/librte_vhost/iotlb.c b/lib/librte_vhost/iotlb.c
> new file mode 100644
> index 0..1b739dae5
> --- /dev/null
> +++ b/lib/librte_vhost/iotlb.c
> @@ -0,0 +1,231 @@
> +/*-
> + *   BSD LICENSE
> + *
> + *   Copyright (c) 2017 Red Hat, Inc.
> + *   Copyright (c) 2017 Maxime Coquelin 

I'm not a lawer, but I have been told many years before, that you don't
have the copyright for the code you write for open source project, the
company you work for does.

Thus, it's more common to see something like following:
Copyright , ... the commany ...
Author:  Some One <...@...>

However, as you may have noticed, it's not common to put the authorship
in the source files. Though I don't object it.

[...]
> +#define IOTLB_CACHE_SIZE 1024
> +
> +static void vhost_user_iotlb_cache_remove_all(struct vhost_virtqueue *vq)
   
Note that it's not the DPDK coding style to define a function.

> +{
> + struct vhost_iotlb_entry *node, *temp_node;
> +
> + rte_rwlock_write_lock(&vq->iotlb_lock);
> +
> + TAILQ_FOREACH_SAFE(node, &vq->iotlb_list, next, temp_node) {
> + TAILQ_REMOVE(&vq->iotlb_list, node, next);
> + rte_mempool_put(vq->iotlb_pool, node);
> + }
> +
> + rte_rwlock_write_unlock(&vq->iotlb_lock);
> +}
> +
> +void vhost_user_iotlb_cache_insert(struct vhost_virtqueue *vq, uint64_t iova,
> + uint64_t uaddr, uint64_t size, uint8_t perm)
> +{
> + struct vhost_iotlb_entry *node, *new_node;
> + int ret;
> +
> + ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
> + if (ret) {
> + RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool empty, invalidate 
> cache\n");

It's a cache, why not considering remove one to get space for new one?

> + vhost_user_iotlb_cache_remove_all(vq);
> + ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
> + if (ret) {
> + RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool still empty, 
> failure\n");
> + return;
> + }
> + }
> +
> + new_node->iova = iova;
> + new_node->uaddr = uaddr;
> + new_node->size = size;
> + new_node->perm = perm;
> +
> + rte_rwlock_write_lock(&vq->iotlb_lock);
> +
> + TAILQ_FOREACH(node, &vq->iotlb_list, next) {
> + /*
> +  * Entries must be invalidated before being updated.
> +  * So if iova already in list, assume identical.
> +  */
> + if (node->iova == new_node->iova) {
> + rte_mempool_put(vq->iotlb_pool, new_node);
> + goto unlock;
> + } else if (node->iova > new_node->iova) {
> + TAILQ_INSERT_BEFORE(node, new_node, next);
> + goto unlock;
> + }
> + }
> +
> + TAILQ_INSERT_TAIL(&vq->iotlb_list, new_node, next);
> +
> +unlock:
> + rte_rwlock_write_unlock(&vq->iotlb_lock);
> +}
> +
> +void vhost_user_iotlb_cache_remove(struct vhost_virtqueue *vq,
> + uint64_t iova, uint64_t size)
> +{
> + struct vhost_iotlb_entry *node, *temp_node;
> +
> + if (unlikely(!size))
> + return;
> +
> + rte_rwlock_write_lock(&vq->iotlb_lock);
> +
> + TAILQ_FOREACH_SAFE(node, &vq->iotlb_list, next, temp_node) {
> + /* Sorted list */

I'd like to put such comments at the struct declartion, so that you don't
have to mention it many times that it's a sorted list.

> + if (unlikely(node->iova >= iova + size)) {
> + break;
> + } else if ((node->iova < iova + size) &&
> + (iova < node->iova + node->size)) {
> + TAILQ_REMOVE(&vq->iotlb_list, node, next);
> + rte_mempool_put(vq->iotlb_pool, node);
> + continue;
> + }
> + }
> +
> + rte_rwlock_write_unlock(&vq->iotlb_lock);
> +}
> +
[...]
> +int vhost_user_iotlb_init(struct virtio_net *dev, int vq_index)
> +{
> + char pool_name[RTE_MEMPOOL_NAMESIZE];
> + struct vhost_virtqueue *vq = dev->virtqueue[vq_index];
> + int ret = -1, socket;
> +
> + if (vq->iotlb_pool) {
> + /*
> +  * The cache has already been initialized,
> +  * just drop all entries
> +  */
> + vhost_user_iotlb_cache_remove_all(vq);
> + return 0;
> + }
> +
> +#ifdef RTE_LIBRTE_VHOST_NUMA
> + ret = get_mempolicy(&socket, NULL, 0, vq, MPOL_F_NODE | MPOL_F_ADDR);
> +#endif
> + if (ret)
> + socket = 0;
> +
> + rte_rwlock_init(&vq->iotlb_lock);
> +
> + TAILQ_INIT(&vq->iotlb_list);
> +
> + snprintf(pool_name, sizeof(pool_name), "iotlb_cache_%d_%d",
> + dev->vid, vq_index);

iotlb_cache is too generic. Adding a "vhost" prefix?

--yliu

Re: [dpdk-dev] [PATCH 07/21] vhost: add iotlb helper functions

[Maxime Coquelin]

On 09/08/2017 10:08 AM, Yuanhan Liu wrote:

On Thu, Aug 31, 2017 at 11:50:09AM +0200, Maxime Coquelin wrote:

diff --git a/lib/librte_vhost/iotlb.c b/lib/librte_vhost/iotlb.c
new file mode 100644
index 0..1b739dae5
--- /dev/null
+++ b/lib/librte_vhost/iotlb.c
@@ -0,0 +1,231 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright (c) 2017 Red Hat, Inc.
+ *   Copyright (c) 2017 Maxime Coquelin 


I'm not a lawer, but I have been told many years before, that you don't
have the copyright for the code you write for open source project, the
company you work for does.

Thus, it's more common to see something like following:
Copyright , ... the commany ...
Author:  Some One <...@...>

However, as you may have noticed, it's not common to put the authorship
in the source files. Though I don't object it.


I'm not a lawyer too. At least in other projects, it seems common the
author puts his name as copyright owner.

I have no issue to change it to only keep Red Hat's one though.


[...]

+#define IOTLB_CACHE_SIZE 1024
+
+static void vhost_user_iotlb_cache_remove_all(struct vhost_virtqueue *vq)


Note that it's not the DPDK coding style to define a function.


Ok, I guess you mean:
static void
vhost_user_iotlb_cache_remove_all(struct vhost_virtqueue *vq) ?


+{
+   struct vhost_iotlb_entry *node, *temp_node;
+
+   rte_rwlock_write_lock(&vq->iotlb_lock);
+
+   TAILQ_FOREACH_SAFE(node, &vq->iotlb_list, next, temp_node) {
+   TAILQ_REMOVE(&vq->iotlb_list, node, next);
+   rte_mempool_put(vq->iotlb_pool, node);
+   }
+
+   rte_rwlock_write_unlock(&vq->iotlb_lock);
+}
+
+void vhost_user_iotlb_cache_insert(struct vhost_virtqueue *vq, uint64_t iova,
+   uint64_t uaddr, uint64_t size, uint8_t perm)
+{
+   struct vhost_iotlb_entry *node, *new_node;
+   int ret;
+
+   ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
+   if (ret) {
+   RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool empty, invalidate 
cache\n");


It's a cache, why not considering remove one to get space for new one?


It would mean having to track every lookups not to remove hot entries,
which would have an impact on performance.

Moreover, the idea is to have the cache large enough, else you could
face packet drops due to random cache misses.

We might consider to improve it, but I consider it an optimization that
could be implemented later if needed.


+   vhost_user_iotlb_cache_remove_all(vq);
+   ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
+   if (ret) {
+   RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool still empty, 
failure\n");
+   return;
+   }
+   }
+
+   new_node->iova = iova;
+   new_node->uaddr = uaddr;
+   new_node->size = size;
+   new_node->perm = perm;
+
+   rte_rwlock_write_lock(&vq->iotlb_lock);
+
+   TAILQ_FOREACH(node, &vq->iotlb_list, next) {
+   /*
+* Entries must be invalidated before being updated.
+* So if iova already in list, assume identical.
+*/
+   if (node->iova == new_node->iova) {
+   rte_mempool_put(vq->iotlb_pool, new_node);
+   goto unlock;
+   } else if (node->iova > new_node->iova) {
+   TAILQ_INSERT_BEFORE(node, new_node, next);
+   goto unlock;
+   }
+   }
+
+   TAILQ_INSERT_TAIL(&vq->iotlb_list, new_node, next);
+
+unlock:
+   rte_rwlock_write_unlock(&vq->iotlb_lock);
+}
+
+void vhost_user_iotlb_cache_remove(struct vhost_virtqueue *vq,
+   uint64_t iova, uint64_t size)
+{
+   struct vhost_iotlb_entry *node, *temp_node;
+
+   if (unlikely(!size))
+   return;
+
+   rte_rwlock_write_lock(&vq->iotlb_lock);
+
+   TAILQ_FOREACH_SAFE(node, &vq->iotlb_list, next, temp_node) {
+   /* Sorted list */


I'd like to put such comments at the struct declartion, so that you don't
have to mention it many times that it's a sorted list.


Ok, I'll comment directly in struct declaration.


+   if (unlikely(node->iova >= iova + size)) {
+   break;
+   } else if ((node->iova < iova + size) &&
+   (iova < node->iova + node->size)) {
+   TAILQ_REMOVE(&vq->iotlb_list, node, next);
+   rte_mempool_put(vq->iotlb_pool, node);
+   continue;
+   }
+   }
+
+   rte_rwlock_write_unlock(&vq->iotlb_lock);
+}
+

[...]

+int vhost_user_iotlb_init(struct virtio_net *dev, int vq_index)
+{
+   char pool_name[RTE_MEMPOOL_NAMESIZE];
+   struct vhost_virtqueue *vq = dev->virtqueue[vq_index];
+   int ret = -1, socket;
+
+   if (vq->iotlb_pool) {
+   /*
+   

Re: [dpdk-dev] [PATCH 07/21] vhost: add iotlb helper functions

[Yuanhan Liu]

On Fri, Sep 08, 2017 at 10:24:58AM +0200, Maxime Coquelin wrote:
> 
> 
> On 09/08/2017 10:08 AM, Yuanhan Liu wrote:
> >On Thu, Aug 31, 2017 at 11:50:09AM +0200, Maxime Coquelin wrote:
> >>diff --git a/lib/librte_vhost/iotlb.c b/lib/librte_vhost/iotlb.c
> >>new file mode 100644
> >>index 0..1b739dae5
> >>--- /dev/null
> >>+++ b/lib/librte_vhost/iotlb.c
> >>@@ -0,0 +1,231 @@
> >>+/*-
> >>+ *   BSD LICENSE
> >>+ *
> >>+ *   Copyright (c) 2017 Red Hat, Inc.
> >>+ *   Copyright (c) 2017 Maxime Coquelin 
> >
> >I'm not a lawer, but I have been told many years before, that you don't
> >have the copyright for the code you write for open source project, the
> >company you work for does.
> >
> >Thus, it's more common to see something like following:
> > Copyright , ... the commany ...
> > Author:  Some One <...@...>
> >
> >However, as you may have noticed, it's not common to put the authorship
> >in the source files. Though I don't object it.
> 
> I'm not a lawyer too. At least in other projects, it seems common the
> author puts his name as copyright owner.
> 
> I have no issue to change it to only keep Red Hat's one though.

That's up to you. What I said before was JFYI :)

> >[...]
> >>+#define IOTLB_CACHE_SIZE 1024
> >>+
> >>+static void vhost_user_iotlb_cache_remove_all(struct vhost_virtqueue *vq)
> >
> >Note that it's not the DPDK coding style to define a function.
> 
> Ok, I guess you mean:
> static void
> vhost_user_iotlb_cache_remove_all(struct vhost_virtqueue *vq) ?

Yep.

> >>+{
> >>+   struct vhost_iotlb_entry *node, *temp_node;
> >>+
> >>+   rte_rwlock_write_lock(&vq->iotlb_lock);
> >>+
> >>+   TAILQ_FOREACH_SAFE(node, &vq->iotlb_list, next, temp_node) {
> >>+   TAILQ_REMOVE(&vq->iotlb_list, node, next);
> >>+   rte_mempool_put(vq->iotlb_pool, node);
> >>+   }
> >>+
> >>+   rte_rwlock_write_unlock(&vq->iotlb_lock);
> >>+}
> >>+
> >>+void vhost_user_iotlb_cache_insert(struct vhost_virtqueue *vq, uint64_t 
> >>iova,
> >>+   uint64_t uaddr, uint64_t size, uint8_t perm)
> >>+{
> >>+   struct vhost_iotlb_entry *node, *new_node;
> >>+   int ret;
> >>+
> >>+   ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
> >>+   if (ret) {
> >>+   RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool empty, invalidate 
> >>cache\n");
> >
> >It's a cache, why not considering remove one to get space for new one?
> 
> It would mean having to track every lookups not to remove hot entries,
> which would have an impact on performance.

You were removing all caches, how can we do worse than that? Even a
random evict would be better. Or, more simply, just to remove the
head or the tail?

--yliu

> Moreover, the idea is to have the cache large enough, else you could
> face packet drops due to random cache misses.
> 
> We might consider to improve it, but I consider it an optimization that
> could be implemented later if needed.
> 
> >>+   vhost_user_iotlb_cache_remove_all(vq);
> >>+   ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
> >>+   if (ret) {
> >>+   RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool still empty, 
> >>failure\n");
> >>+   return;
> >>+   }
> >>+   }

[dpdk-dev] [PATCH v2 01/30] bus/fslmc: qbman replace word copy with memcpy

[Hemant Agrawal]

From: Haiying Wang 

The word_copy is not as efficient as expected, so remove it from
this driver.

Signed-off-by: Haiying Wang 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/qbman/include/compat.h | 42 
 drivers/bus/fslmc/qbman/qbman_portal.c   | 12 -
 drivers/bus/fslmc/qbman/qbman_sys.h  |  9 ---
 3 files changed, 5 insertions(+), 58 deletions(-)

diff --git a/drivers/bus/fslmc/qbman/include/compat.h 
b/drivers/bus/fslmc/qbman/include/compat.h
index 529f1ea..7b69fd1 100644
--- a/drivers/bus/fslmc/qbman/include/compat.h
+++ b/drivers/bus/fslmc/qbman/include/compat.h
@@ -229,48 +229,6 @@ typedef uint32_t   phandle;
 #define __raw_readl(p) (*(const volatile unsigned int *)(p))
 #define __raw_writel(v, p) {*(volatile unsigned int *)(p) = (v); }
 
-/* memcpy() stuff - when you know alignments in advance */
-#ifdef CONFIG_TRY_BETTER_MEMCPY
-static inline void copy_words(void *dest, const void *src, size_t sz)
-{
-   u32 *__dest = dest;
-   const u32 *__src = src;
-   size_t __sz = sz >> 2;
-
-   QBMAN_BUG_ON((unsigned long)dest & 0x3);
-   QBMAN_BUG_ON((unsigned long)src & 0x3);
-   QBMAN_BUG_ON(sz & 0x3);
-   while (__sz--)
-   *(__dest++) = *(__src++);
-}
-
-static inline void copy_shorts(void *dest, const void *src, size_t sz)
-{
-   u16 *__dest = dest;
-   const u16 *__src = src;
-   size_t __sz = sz >> 1;
-
-   QBMAN_BUG_ON((unsigned long)dest & 0x1);
-   QBMAN_BUG_ON((unsigned long)src & 0x1);
-   QBMAN_BUG_ON(sz & 0x1);
-   while (__sz--)
-   *(__dest++) = *(__src++);
-}
-
-static inline void copy_bytes(void *dest, const void *src, size_t sz)
-{
-   u8 *__dest = dest;
-   const u8 *__src = src;
-
-   while (sz--)
-   *(__dest++) = *(__src++);
-}
-#else
-#define copy_words memcpy
-#define copy_shorts memcpy
-#define copy_bytes memcpy
-#endif
-
 /* Completion stuff */
 #define DECLARE_COMPLETION(n) int n = 0
 #define complete(n) { *n = 1; }
diff --git a/drivers/bus/fslmc/qbman/qbman_portal.c 
b/drivers/bus/fslmc/qbman/qbman_portal.c
index dd62e9a..f360760 100644
--- a/drivers/bus/fslmc/qbman/qbman_portal.c
+++ b/drivers/bus/fslmc/qbman/qbman_portal.c
@@ -480,8 +480,8 @@ static int qbman_swp_enqueue_array_mode(struct qbman_swp *s,
return -EBUSY;
p = qbman_cena_write_start_wo_shadow(&s->sys,
QBMAN_CENA_SWP_EQCR(EQAR_IDX(eqar)));
-   word_copy(&p[1], &cl[1], 7);
-   word_copy(&p[8], fd, sizeof(*fd) >> 2);
+   memcpy(&p[1], &cl[1], 28);
+   memcpy(&p[8], fd, sizeof(*fd));
/* Set the verb byte, have to substitute in the valid-bit */
lwsync();
p[0] = cl[0] | EQAR_VB(eqar);
@@ -512,8 +512,8 @@ static int qbman_swp_enqueue_ring_mode(struct qbman_swp *s,
 
p = qbman_cena_write_start_wo_shadow(&s->sys,
QBMAN_CENA_SWP_EQCR(s->eqcr.pi & 7));
-   word_copy(&p[1], &cl[1], 7);
-   word_copy(&p[8], fd, sizeof(*fd) >> 2);
+   memcpy(&p[1], &cl[1], 28);
+   memcpy(&p[8], fd, sizeof(*fd));
lwsync();
/* Set the verb byte, have to substitute in the valid-bit */
p[0] = cl[0] | s->eqcr.pi_vb;
@@ -549,9 +549,7 @@ int qbman_swp_fill_ring(struct qbman_swp *s,
}
p = qbman_cena_write_start_wo_shadow(&s->sys,
QBMAN_CENA_SWP_EQCR((s->eqcr.pi/* +burst_index */) & 7));
-   /* word_copy(&p[1], &cl[1], 7); */
memcpy(&p[1], &cl[1], 7 * 4);
-   /* word_copy(&p[8], fd, sizeof(*fd) >> 2); */
memcpy(&p[8], fd, sizeof(struct qbman_fd));
 
/* lwsync(); */
@@ -799,7 +797,7 @@ int qbman_swp_pull(struct qbman_swp *s, struct 
qbman_pull_desc *d)
 */
qb_attr_code_encode(&code_pull_token, cl, s->desc.idx + 1);
p = qbman_cena_write_start_wo_shadow(&s->sys, QBMAN_CENA_SWP_VDQCR);
-   word_copy(&p[1], &cl[1], 3);
+   memcpy(&p[1], &cl[1], 12);
/* Set the verb byte, have to substitute in the valid-bit */
lwsync();
p[0] = cl[0] | s->vdq.valid_bit;
diff --git a/drivers/bus/fslmc/qbman/qbman_sys.h 
b/drivers/bus/fslmc/qbman/qbman_sys.h
index 5dbcaa5..9ea55de 100644
--- a/drivers/bus/fslmc/qbman/qbman_sys.h
+++ b/drivers/bus/fslmc/qbman/qbman_sys.h
@@ -47,15 +47,6 @@
 #undef QBMAN_CINH_TRACE
 #undef QBMAN_CENA_TRACE
 
-static inline void word_copy(void *d, const void *s, unsigned int cnt)
-{
-   uint32_t *dd = d;
-   const uint32_t *ss = s;
-
-   while (cnt--)
-   *(dd++) = *(ss++);
-}
-
 /* Currently, the CENA support code expects each 32-bit word to be written in
  * host order, and these are converted to hardware (little-endian) order on
  * command submission. However, 64-bit quantities are must be written (and 
read)
-- 
2.7.4

[dpdk-dev] [PATCH v2 00/30] NXP DPAA2 PMD updates

[Hemant Agrawal]

This patchset includes the hw driver upgrades and additional
nic feature implementations.

patches 1..8 - upgrades the qbman hw driver
patches 9..11 - upgrades the MC version to 10.3.1
patches 12..13 - adds the support for LX2160 platform
patches 14..30 - adds various features and cleanups in
NXP DPAA2 pmd mainly - links status, RSS update,
extra stats etc.

This patch is based on Shreyansh patchsetfor bus refactoring [1].

1. NXP DPAA2: Refactor bus scan/probe code
http://dpdk.org/ml/archives/dev/2017-August/073545.html

v2: 
1. improve the checksum offload suport as per fuser config
2. remvoe checksum errors.
3. some cleanups on log messages.

Ashish Jain (2):
  config/dpaa2: change max lores to 16
  bus/fslmc: add support for LX2160 platform

Haiying Wang (4):
  bus/fslmc: qbman replace word copy with memcpy
  bus/fslmc: add qbman API to do enqueue with multiple frames
  bus/fslmc: support up to 32 frames in one volatile dequeue
  bus/fslmc: enhance the QBMAN CENA mode

Hemant Agrawal (21):
  bus/fslmc: remove the export for qbman version function
  bus/fslmc: qbman remove unused funcs and align names
  bus/fslmc: cleanup compat file
  bus/fslmc: clean the qbman support code
  net/dpaa2: add support for congestion overhead
  bus/fslmc: add support to check dpbp presence
  bus/fslmc: cleanup the dpaa2 interrupt support
  net/dpaa2: add support for link status event
  bus/fslmc: enable link status interrupt
  net/dpaa2: check physical link state on up cmd
  net/dpaa2: improve error and logs for flow distribution
  net/dpaa2: increase the dist param to 64 bit
  net/dpaa2: remove RSS restriction with num of queues
  net/dpaa2: add support for RSS hash update and get
  bus/dpaa2: add support for hw extra stats API
  net/dpaa2: add support for extra stats
  net/dpaa2: fix the Tx handling of non HW pool bufs
  net/dpaa2: log that VLAN extend offload not supported
  net/dpaa2: checksum support as per user config
  net/dpaa2: improve debug messaging
  bus/dpaa2: improve debug log messages

Shreyansh Jain (3):
  bus/fslmc: update MC to 10.3.x
  net/dpaa2: update MC to 10.3.x
  crypto/dpaa2_sec: update MC to 10.3.x

 config/defconfig_arm64-dpaa2-linuxapp-gcc  |2 +-
 doc/guides/nics/features/dpaa2.ini |2 +
 drivers/bus/dpaa/base/fman/fman_hw.c   |   30 +
 drivers/bus/dpaa/include/fsl_fman.h|3 +
 drivers/bus/dpaa/rte_bus_dpaa_version.map  |1 +
 drivers/bus/fslmc/Makefile |1 +
 drivers/bus/fslmc/fslmc_vfio.c |  131 +-
 drivers/bus/fslmc/fslmc_vfio.h |8 +-
 drivers/bus/fslmc/mc/dpbp.c|  182 ++-
 drivers/bus/fslmc/mc/dpci.c|  202 ++-
 drivers/bus/fslmc/mc/dpcon.c   |  163 ++-
 drivers/bus/fslmc/mc/dpio.c|  230 +++-
 drivers/bus/fslmc/mc/dpmng.c   |   33 +-
 drivers/bus/fslmc/mc/fsl_dpbp.h|  191 +--
 drivers/bus/fslmc/mc/fsl_dpbp_cmd.h|  125 +-
 drivers/bus/fslmc/mc/fsl_dpci.h|  257 +---
 drivers/bus/fslmc/mc/fsl_dpci_cmd.h|  222 ++--
 drivers/bus/fslmc/mc/fsl_dpcon.h   |  186 +--
 drivers/bus/fslmc/mc/fsl_dpcon_cmd.h   |  193 +--
 drivers/bus/fslmc/mc/fsl_dpio.h|  299 ++---
 drivers/bus/fslmc/mc/fsl_dpio_cmd.h|  178 +--
 drivers/bus/fslmc/mc/fsl_dpmng.h   |   41 +-
 drivers/bus/fslmc/mc/fsl_dpmng_cmd.h   |   41 +-
 drivers/bus/fslmc/mc/fsl_mc_cmd.h  |  217 ++--
 drivers/bus/fslmc/mc/fsl_mc_sys.h  |   36 +-
 drivers/bus/fslmc/mc/mc_sys.c  |5 +-
 drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c   |9 +-
 drivers/bus/fslmc/portal/dpaa2_hw_dpci.c   |2 +-
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c   |   46 +-
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h|2 +
 drivers/bus/fslmc/qbman/include/compat.h   |  311 +
 drivers/bus/fslmc/qbman/include/fsl_qbman_base.h   |4 -
 drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h |  183 ++-
 drivers/bus/fslmc/qbman/qbman_portal.c |  973 ++
 drivers/bus/fslmc/qbman/qbman_portal.h |  140 +--
 drivers/bus/fslmc/qbman/qbman_private.h|  174 ---
 drivers/bus/fslmc/qbman/qbman_sys.h|  144 ++-
 drivers/bus/fslmc/qbman/qbman_sys_decl.h   |   25 +-
 drivers/bus/fslmc/rte_bus_fslmc_version.map|9 +-
 drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c|6 +-
 drivers/crypto/dpaa2_sec/mc/dpseci.c   |  676 ++
 drivers/crypto/dpaa2_sec/mc/fsl_dpseci.h   |  782 
 drivers/crypto/dpaa2_sec/mc/fsl_dpseci_cmd.h   |  387 +++---
 drivers/event/dpaa2/dpaa2_eventdev.c   |2 +-
 drivers/net/dpaa/dpaa_ethdev.c  

[dpdk-dev] [PATCH v2 02/30] bus/fslmc: remove the export for qbman version function

[Hemant Agrawal]

This func is not required to be used outside of the qbman driver.

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/qbman/include/fsl_qbman_base.h   | 4 
 drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h | 2 --
 drivers/bus/fslmc/qbman/qbman_portal.c | 7 ---
 drivers/bus/fslmc/qbman/qbman_portal.h | 1 +
 drivers/bus/fslmc/qbman/qbman_private.h| 4 
 5 files changed, 5 insertions(+), 13 deletions(-)

diff --git a/drivers/bus/fslmc/qbman/include/fsl_qbman_base.h 
b/drivers/bus/fslmc/qbman/include/fsl_qbman_base.h
index ee4b772..1415960 100644
--- a/drivers/bus/fslmc/qbman/include/fsl_qbman_base.h
+++ b/drivers/bus/fslmc/qbman/include/fsl_qbman_base.h
@@ -38,10 +38,6 @@ typedef uint64_t  dma_addr_t;
  *
  */
 
-#define QMAN_REV_4000   0x0400
-#define QMAN_REV_4100   0x0401
-#define QMAN_REV_4101   0x04010001
-
 /**
  * struct qbman_block_desc - qbman block descriptor structure
  * @ccsr_reg_bar: CCSR register map.
diff --git a/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h 
b/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
index 9e9047e..23c3d13 100644
--- a/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
+++ b/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
@@ -1132,6 +1132,4 @@ int qbman_swp_send_multiple(struct qbman_swp *s,
 
 int qbman_check_command_complete(struct qbman_swp *s,
 const struct qbman_result *dq);
-
-int qbman_get_version(void);
 #endif /* !_FSL_QBMAN_PORTAL_H */
diff --git a/drivers/bus/fslmc/qbman/qbman_portal.c 
b/drivers/bus/fslmc/qbman/qbman_portal.c
index f360760..97df703 100644
--- a/drivers/bus/fslmc/qbman/qbman_portal.c
+++ b/drivers/bus/fslmc/qbman/qbman_portal.c
@@ -105,8 +105,6 @@ struct qb_attr_code code_sdqcr_dqsrc = QB_CODE(0, 0, 16);
 #define MAX_QBMAN_PORTALS  35
 static struct qbman_swp *portal_idx_map[MAX_QBMAN_PORTALS];
 
-uint32_t qman_version;
-
 /*/
 /* Portal constructor/destructor */
 /*/
@@ -1579,8 +1577,3 @@ int qbman_swp_send_multiple(struct qbman_swp *s,
 
return sent;
 }
-
-int qbman_get_version(void)
-{
-   return qman_version;
-}
diff --git a/drivers/bus/fslmc/qbman/qbman_portal.h 
b/drivers/bus/fslmc/qbman/qbman_portal.h
index 7aa1d4f..8018048 100644
--- a/drivers/bus/fslmc/qbman/qbman_portal.h
+++ b/drivers/bus/fslmc/qbman/qbman_portal.h
@@ -29,6 +29,7 @@
 #include "qbman_private.h"
 #include 
 
+uint32_t qman_version;
 /* All QBMan command and result structures use this "valid bit" encoding */
 #define QB_VALID_BIT ((uint32_t)0x80)
 
diff --git a/drivers/bus/fslmc/qbman/qbman_private.h 
b/drivers/bus/fslmc/qbman/qbman_private.h
index 292ec6a..b98c330 100644
--- a/drivers/bus/fslmc/qbman/qbman_private.h
+++ b/drivers/bus/fslmc/qbman/qbman_private.h
@@ -171,4 +171,8 @@ static inline void hexdump(const void *ptr, size_t sz)
__hexdump(start, end, p, sz, c);
 }
 
+#define QMAN_REV_4000   0x0400
+#define QMAN_REV_4100   0x0401
+#define QMAN_REV_4101   0x04010001
+
 #include "qbman_sys.h"
-- 
2.7.4

[dpdk-dev] [PATCH v2 03/30] bus/fslmc: add qbman API to do enqueue with multiple frames

[Hemant Agrawal]

From: Haiying Wang 

Clean it up and update the prototype.

Signed-off-by: Haiying Wang 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h |  32 ++--
 drivers/bus/fslmc/qbman/qbman_portal.c | 200 +++--
 drivers/bus/fslmc/rte_bus_fslmc_version.map|   3 +-
 drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c|   2 +-
 drivers/event/dpaa2/dpaa2_eventdev.c   |   2 +-
 drivers/net/dpaa2/dpaa2_rxtx.c |   2 +-
 6 files changed, 83 insertions(+), 158 deletions(-)

diff --git a/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h 
b/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
index 23c3d13..fe1cc94 100644
--- a/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
+++ b/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
@@ -914,19 +914,33 @@ void qbman_eq_desc_set_dca(struct qbman_eq_desc *d, int 
enable,
 int qbman_swp_enqueue(struct qbman_swp *s, const struct qbman_eq_desc *d,
  const struct qbman_fd *fd);
 /**
- * qbman_swp_enqueue_multiple_eqdesc() - Enqueue multiple frames with separte
- * enqueue descriptors.
+ * qbman_swp_enqueue_multiple() - Enqueue multiple frames with same
+ eq descriptor
  * @s: the software portal used for enqueue.
- * @d: the enqueue descriptors
+ * @d: the enqueue descriptor.
  * @fd: the frame descriptor to be enqueued.
  * @num_frames: the number of the frames to be enqueued.
  *
  * Return the number of enqueued frames, -EBUSY if the EQCR is not ready.
  */
-int qbman_swp_enqueue_multiple_eqdesc(struct qbman_swp *s,
+int qbman_swp_enqueue_multiple(struct qbman_swp *s,
   const struct qbman_eq_desc *d,
   const struct qbman_fd *fd,
   int num_frames);
+/**
+ * qbman_swp_enqueue_multiple_desc() - Enqueue multiple frames with
+ *individual eq descriptor.
+ * @s: the software portal used for enqueue.
+ * @d: the enqueue descriptor.
+ * @fd: the frame descriptor to be enqueued.
+ * @num_frames: the number of the frames to be enqueued.
+ *
+ * Return the number of enqueued frames, -EBUSY if the EQCR is not ready.
+ */
+int qbman_swp_enqueue_multiple_desc(struct qbman_swp *s,
+   const struct qbman_eq_desc *d,
+   const struct qbman_fd *fd,
+   int num_frames);
 
 /* TODO:
  * qbman_swp_enqueue_thresh() - Set threshold for EQRI interrupt.
@@ -1119,16 +1133,6 @@ int qbman_swp_CDAN_disable(struct qbman_swp *s, uint16_t 
channelid);
  */
 int qbman_swp_CDAN_set_context_enable(struct qbman_swp *s, uint16_t channelid,
  uint64_t ctx);
-int qbman_swp_fill_ring(struct qbman_swp *s,
-   const struct qbman_eq_desc *d,
-  const struct qbman_fd *fd,
-  uint8_t burst_index);
-int qbman_swp_flush_ring(struct qbman_swp *s);
-void qbman_sync(void);
-int qbman_swp_send_multiple(struct qbman_swp *s,
-   const struct qbman_eq_desc *d,
-   const struct qbman_fd *fd,
-   int frames_to_send);
 
 int qbman_check_command_complete(struct qbman_swp *s,
 const struct qbman_result *dq);
diff --git a/drivers/bus/fslmc/qbman/qbman_portal.c 
b/drivers/bus/fslmc/qbman/qbman_portal.c
index 97df703..f212829 100644
--- a/drivers/bus/fslmc/qbman/qbman_portal.c
+++ b/drivers/bus/fslmc/qbman/qbman_portal.c
@@ -525,15 +525,26 @@ static int qbman_swp_enqueue_ring_mode(struct qbman_swp 
*s,
return 0;
 }
 
-int qbman_swp_fill_ring(struct qbman_swp *s,
-   const struct qbman_eq_desc *d,
-   const struct qbman_fd *fd,
-   __attribute__((unused)) uint8_t burst_index)
+int qbman_swp_enqueue(struct qbman_swp *s, const struct qbman_eq_desc *d,
+ const struct qbman_fd *fd)
+{
+   if (s->sys.eqcr_mode == qman_eqcr_vb_array)
+   return qbman_swp_enqueue_array_mode(s, d, fd);
+   else/* Use ring mode by default */
+   return qbman_swp_enqueue_ring_mode(s, d, fd);
+}
+
+int qbman_swp_enqueue_multiple(struct qbman_swp *s,
+  const struct qbman_eq_desc *d,
+  const struct qbman_fd *fd,
+  int num_frames)
 {
uint32_t *p;
const uint32_t *cl = qb_cl(d);
-   uint32_t eqcr_ci;
+   uint32_t eqcr_ci, eqcr_pi;
uint8_t diff;
+   int i, num_enqueued = 0;
+   uint64_t addr_cena;
 
if (!s->eqcr.available) {
eqcr_ci = s->eqcr.ci;
@@ -543,62 +554,58 @@ int qbman_swp_fill_ring(struct qbman_swp *s,
   eqcr_ci, s->eqcr.ci);
s->eqcr.available += diff;
if (!diff)
-   

[dpdk-dev] [PATCH v2 06/30] bus/fslmc: qbman remove unused funcs and align names

[Hemant Agrawal]

name alignment for check command and result functions
putting them as separate functions instead of changing the original
functions.

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h | 146 ++--
 drivers/bus/fslmc/qbman/qbman_portal.c | 731 ++---
 drivers/bus/fslmc/qbman/qbman_portal.h |   6 +-
 drivers/bus/fslmc/qbman/qbman_private.h|   2 +-
 drivers/bus/fslmc/qbman/qbman_sys.h|  12 -
 drivers/bus/fslmc/rte_bus_fslmc_version.map|   3 +-
 drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c|   4 +-
 drivers/net/dpaa2/dpaa2_rxtx.c |  10 +-
 8 files changed, 449 insertions(+), 465 deletions(-)

diff --git a/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h 
b/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
index fe1cc94..24a6d4b 100644
--- a/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
+++ b/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
@@ -194,11 +194,38 @@ void qbman_swp_interrupt_set_inhibit(struct qbman_swp *p, 
int inhibit);
 /**
  * struct qbman_result - structure for qbman dequeue response and/or
  * notification.
- * @dont_manipulate_directly: the 16 32bit data to represent the whole
+ * @donot_manipulate_directly: the 16 32bit data to represent the whole
  * possible qbman dequeue result.
  */
 struct qbman_result {
-   uint32_t dont_manipulate_directly[16];
+   union {
+   struct common {
+   uint8_t verb;
+   uint8_t reserved[63];
+   } common;
+   struct dq {
+   uint8_t verb;
+   uint8_t stat;
+   __le16 seqnum;
+   __le16 oprid;
+   uint8_t reserved;
+   uint8_t tok;
+   __le32 fqid;
+   uint32_t reserved2;
+   __le32 fq_byte_cnt;
+   __le32 fq_frm_cnt;
+   __le64 fqd_ctx;
+   uint8_t fd[32];
+   } dq;
+   struct scn {
+   uint8_t verb;
+   uint8_t stat;
+   uint8_t state;
+   uint8_t reserved;
+   __le32 rid_tok;
+   __le64 ctx;
+   } scn;
+   };
 };
 
 /* TODO:
@@ -254,11 +281,21 @@ void qbman_swp_push_set(struct qbman_swp *s, uint8_t 
channel_idx, int enable);
 
 /**
  * struct qbman_pull_desc - the structure for pull dequeue descriptor
- * @dont_manipulate_directly: the 6 32bit data to represent the whole
- * possible settings for pull dequeue descriptor.
  */
 struct qbman_pull_desc {
-   uint32_t dont_manipulate_directly[6];
+   union {
+   uint32_t donot_manipulate_directly[16];
+   struct pull {
+   uint8_t verb;
+   uint8_t numf;
+   uint8_t tok;
+   uint8_t reserved;
+   uint32_t dq_src;
+   uint64_t rsp_addr;
+   uint64_t rsp_addr_virt;
+   uint8_t padding[40];
+   } pull;
+   };
 };
 
 enum qbman_pull_type_e {
@@ -415,7 +452,20 @@ struct qbman_result *qbman_get_dqrr_from_idx(struct 
qbman_swp *s, uint8_t idx);
  * dequeue result.
  */
 int qbman_result_has_new_result(struct qbman_swp *s,
-   const struct qbman_result *dq);
+   struct qbman_result *dq);
+
+/**
+ * qbman_check_command_complete() - Check if the previous issued dq commnd
+ * is completed and results are available in memory.
+ * @s: the software portal object.
+ * @dq: the dequeue result read from the memory.
+ *
+ * Return 1 for getting a valid dequeue result, or 0 for not getting a valid
+ * dequeue result.
+ */
+int qbman_check_command_complete(struct qbman_result *dq);
+
+int qbman_check_new_result(struct qbman_result *dq);
 
 /*  */
 /* Parsing dequeue entries (DQRR and user-provided storage) */
@@ -537,7 +587,7 @@ int qbman_result_is_FQPN(const struct qbman_result *dq);
  *
  * Return the state field.
  */
-uint32_t qbman_result_DQ_flags(const struct qbman_result *dq);
+uint8_t qbman_result_DQ_flags(const struct qbman_result *dq);
 
 /**
  * qbman_result_DQ_is_pull() - Check whether the dq response is from a pull
@@ -648,24 +698,6 @@ uint32_t qbman_result_SCN_rid(const struct qbman_result 
*scn);
  */
 uint64_t qbman_result_SCN_ctx(const struct qbman_result *scn);
 
-/**
- * qbman_result_SCN_state_in_mem() - Get the state in notification written
- * in memory
- * @scn: the state change notification.
- *
- * Return the state.
- */
-uint8_t qbman_result_SCN_state_in_mem(const struct qbman_result *scn);
-
-/**
- * qbman_result_SCN_rid_in_mem() - Get the resource id in notification writ

[dpdk-dev] [PATCH v2 05/30] bus/fslmc: enhance the QBMAN CENA mode

[Hemant Agrawal]

From: Haiying Wang 

Signed-off-by: Haiying Wang 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/qbman/qbman_sys.h | 7 +--
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/drivers/bus/fslmc/qbman/qbman_sys.h 
b/drivers/bus/fslmc/qbman/qbman_sys.h
index 9ea55de..47da595 100644
--- a/drivers/bus/fslmc/qbman/qbman_sys.h
+++ b/drivers/bus/fslmc/qbman/qbman_sys.h
@@ -217,7 +217,6 @@ static inline void 
qbman_cena_write_complete_wo_shadow(struct qbman_swp_sys *s,
 #ifdef QBMAN_CENA_TRACE
pr_info("qbman_cena_write_complete(%p:%d:0x%03x)\n",
s->addr_cena, s->idx, offset);
-   hexdump(cmd, 64);
 #endif
dcbf(s->addr_cena + offset);
 }
@@ -251,11 +250,7 @@ static inline void *qbman_cena_read_wo_shadow(struct 
qbman_swp_sys *s,
 {
 #ifdef QBMAN_CENA_TRACE
pr_info("qbman_cena_read(%p:%d:0x%03x) %p\n",
-   s->addr_cena, s->idx, offset, shadow);
-#endif
-
-#ifdef QBMAN_CENA_TRACE
-   hexdump(shadow, 64);
+   s->addr_cena, s->idx, offset);
 #endif
return s->addr_cena + offset;
 }
-- 
2.7.4

[dpdk-dev] [PATCH v2 07/30] bus/fslmc: cleanup compat file

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/qbman/include/compat.h | 188 +--
 drivers/bus/fslmc/qbman/qbman_private.h  |   2 +-
 2 files changed, 4 insertions(+), 186 deletions(-)

diff --git a/drivers/bus/fslmc/qbman/include/compat.h 
b/drivers/bus/fslmc/qbman/include/compat.h
index 7b69fd1..8885d60 100644
--- a/drivers/bus/fslmc/qbman/include/compat.h
+++ b/drivers/bus/fslmc/qbman/include/compat.h
@@ -65,16 +65,9 @@
  */
 
 /* Required compiler attributes */
-#define __user
 #define likely(x)  __builtin_expect(!!(x), 1)
 #define unlikely(x)__builtin_expect(!!(x), 0)
 #define cacheline_aligned __attribute__((aligned(L1_CACHE_BYTES)))
-#undef container_of
-#define container_of(ptr, type, member) ({ \
-   typeof(((type *)0)->member)(*__mptr) = (ptr); \
-   (type *)((char *)__mptr - offsetof(type, member)); })
-#define __stringify_1(x) #x
-#define __stringify(x) __stringify_1(x)
 
 #ifdef ARRAY_SIZE
 #undef ARRAY_SIZE
@@ -88,17 +81,8 @@ typedef uint32_t u32;
 typedef uint64_t   u64;
 typedef uint64_t   dma_addr_t;
 typedef cpu_set_t  cpumask_t;
-typedefu32 compat_uptr_t;
-
-static inline void __user *compat_ptr(compat_uptr_t uptr)
-{
-   return (void __user *)(unsigned long)uptr;
-}
-
-static inline compat_uptr_t ptr_to_compat(void __user *uptr)
-{
-   return (u32)(unsigned long)uptr;
-}
+typedef unsigned int   gfp_t;
+typedef uint32_t   phandle;
 
 /* I/O operations */
 static inline u32 in_be32(volatile void *__p)
@@ -124,12 +108,11 @@ static inline void out_be32(volatile void *__p, u32 val)
 #define pr_warn(fmt, args...)   prflush("WARN:" fmt, ##args)
 #define pr_info(fmt, args...)   prflush(fmt, ##args)
 
+#ifdef RTE_LIBRTE_DPAA2_DEBUG_BUS
 #ifdef pr_debug
 #undef pr_debug
 #endif
 #define pr_debug(fmt, args...) {}
-#define might_sleep_if(c) {}
-#define msleep(x) {}
 #define WARN_ON(c, str) \
 do { \
static int warned_##__LINE__; \
@@ -147,103 +130,20 @@ do { \
 
 #define ALIGN(x, a) (((x) + ((typeof(x))(a) - 1)) & ~((typeof(x))(a) - 1))
 
-//
-/* Linked-lists */
-//
-
-struct list_head {
-   struct list_head *prev;
-   struct list_head *next;
-};
-
-#define LIST_HEAD(n) \
-struct list_head n = { \
-   .prev = &n, \
-   .next = &n \
-}
-
-#define INIT_LIST_HEAD(p) \
-do { \
-   struct list_head *__p298 = (p); \
-   __p298->next = __p298; \
-   __p298->prev = __p298->next; \
-} while (0)
-#define list_entry(node, type, member) \
-   (type *)((void *)node - offsetof(type, member))
-#define list_empty(p) \
-({ \
-   const struct list_head *__p298 = (p); \
-   ((__p298->next == __p298) && (__p298->prev == __p298)); \
-})
-#define list_add(p, l) \
-do { \
-   struct list_head *__p298 = (p); \
-   struct list_head *__l298 = (l); \
-   __p298->next = __l298->next; \
-   __p298->prev = __l298; \
-   __l298->next->prev = __p298; \
-   __l298->next = __p298; \
-} while (0)
-#define list_add_tail(p, l) \
-do { \
-   struct list_head *__p298 = (p); \
-   struct list_head *__l298 = (l); \
-   __p298->prev = __l298->prev; \
-   __p298->next = __l298; \
-   __l298->prev->next = __p298; \
-   __l298->prev = __p298; \
-} while (0)
-#define list_for_each(i, l)\
-   for (i = (l)->next; i != (l); i = i->next)
-#define list_for_each_safe(i, j, l)\
-   for (i = (l)->next, j = i->next; i != (l);  \
-i = j, j = i->next)
-#define list_for_each_entry(i, l, name) \
-   for (i = list_entry((l)->next, typeof(*i), name); &i->name != (l); \
-   i = list_entry(i->name.next, typeof(*i), name))
-#define list_for_each_entry_safe(i, j, l, name) \
-   for (i = list_entry((l)->next, typeof(*i), name), \
-   j = list_entry(i->name.next, typeof(*j), name); \
-   &i->name != (l); \
-   i = j, j = list_entry(j->name.next, typeof(*j), name))
-#define list_del(i) \
-do { \
-   (i)->next->prev = (i)->prev; \
-   (i)->prev->next = (i)->next; \
-} while (0)
-
 /* Other miscellaneous interfaces our APIs depend on; */
-
 #define lower_32_bits(x) ((u32)(x))
 #define upper_32_bits(x) ((u32)(((x) >> 16) >> 16))
 
 /* Compiler/type stuff */
-typedef unsigned int   gfp_t;
-typedef uint32_t   phandle;
 
 #define __iomem
-#define EINTR  4
-#define ENODEV 19
 #define GFP_KERNEL 0
 #define __raw_readb(p) (*(const volatile unsigned char *)(p))
 #define __raw_readl(p) (*(const volatile unsigned int *)(p))
 #define __raw_writel(v, p) {*(volatile unsigned int *)(p) = (v); }
 
-/* Completion stuff */
-#define DECLARE_COMPLETION(n) int n = 0
-#define complete(n) { *n = 1; }
-#define wait_for_completion(n) \
-do { \
-   while (!*n) { \
-   bman_poll(); \
-   qman_poll(); \
-   } \
-   *n = 0; \
-} while (0)
-
 /* Allocator stuff *

[dpdk-dev] [PATCH v2 08/30] bus/fslmc: clean the qbman support code

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/qbman/include/compat.h   |  87 +-
 drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h |   5 +-
 drivers/bus/fslmc/qbman/qbman_portal.c |  26 +--
 drivers/bus/fslmc/qbman/qbman_portal.h | 133 ++-
 drivers/bus/fslmc/qbman/qbman_private.h| 178 -
 drivers/bus/fslmc/qbman/qbman_sys.h| 124 --
 drivers/bus/fslmc/qbman/qbman_sys_decl.h   |  25 +--
 7 files changed, 102 insertions(+), 476 deletions(-)
 delete mode 100644 drivers/bus/fslmc/qbman/qbman_private.h

diff --git a/drivers/bus/fslmc/qbman/include/compat.h 
b/drivers/bus/fslmc/qbman/include/compat.h
index 8885d60..97904b8 100644
--- a/drivers/bus/fslmc/qbman/include/compat.h
+++ b/drivers/bus/fslmc/qbman/include/compat.h
@@ -30,32 +30,17 @@
 #ifndef HEADER_COMPAT_H
 #define HEADER_COMPAT_H
 
-#include 
-
 #ifndef _GNU_SOURCE
 #define _GNU_SOURCE
 #endif
 #include 
 #include 
-#include 
 #include 
 #include 
-#include 
-#include 
-#include 
-#include 
-#include 
 #include 
-#include 
-#include 
-#include 
 #include 
-#include 
-#include 
-#include 
-#include 
-#include 
 #include 
+#include 
 #include 
 
 /* The following definitions are primarily to allow the single-source driver
@@ -67,35 +52,9 @@
 /* Required compiler attributes */
 #define likely(x)  __builtin_expect(!!(x), 1)
 #define unlikely(x)__builtin_expect(!!(x), 0)
-#define cacheline_aligned __attribute__((aligned(L1_CACHE_BYTES)))
-
-#ifdef ARRAY_SIZE
-#undef ARRAY_SIZE
-#endif
-#define ARRAY_SIZE(a) (sizeof(a) / sizeof((a)[0]))
 
 /* Required types */
-typedef uint8_tu8;
-typedef uint16_t   u16;
-typedef uint32_t   u32;
-typedef uint64_t   u64;
 typedef uint64_t   dma_addr_t;
-typedef cpu_set_t  cpumask_t;
-typedef unsigned int   gfp_t;
-typedef uint32_t   phandle;
-
-/* I/O operations */
-static inline u32 in_be32(volatile void *__p)
-{
-   volatile u32 *p = __p;
-   return *p;
-}
-
-static inline void out_be32(volatile void *__p, u32 val)
-{
-   volatile u32 *p = __p;
-   *p = val;
-}
 
 /* Debugging */
 #define prflush(fmt, args...) \
@@ -112,8 +71,8 @@ static inline void out_be32(volatile void *__p, u32 val)
 #ifdef pr_debug
 #undef pr_debug
 #endif
-#define pr_debug(fmt, args...) {}
-#define WARN_ON(c, str) \
+#define pr_debug(fmt, args...) printf(fmt, ##args)
+#define QBMAN_BUG_ON(c) \
 do { \
static int warned_##__LINE__; \
if ((c) && !warned_##__LINE__) { \
@@ -122,53 +81,23 @@ do { \
warned_##__LINE__ = 1; \
} \
 } while (0)
-#ifdef CONFIG_BUGON
-#define QBMAN_BUG_ON(c) WARN_ON(c, "BUG")
 #else
 #define QBMAN_BUG_ON(c) {}
+#define pr_debug(fmt, args...) {}
 #endif
 
-#define ALIGN(x, a) (((x) + ((typeof(x))(a) - 1)) & ~((typeof(x))(a) - 1))
-
 /* Other miscellaneous interfaces our APIs depend on; */
-#define lower_32_bits(x) ((u32)(x))
-#define upper_32_bits(x) ((u32)(((x) >> 16) >> 16))
 
-/* Compiler/type stuff */
+#define lower_32_bits(x) ((uint32_t)(x))
+#define upper_32_bits(x) ((uint32_t)(((x) >> 16) >> 16))
+
 
 #define __iomem
-#define GFP_KERNEL 0
+
 #define __raw_readb(p) (*(const volatile unsigned char *)(p))
 #define __raw_readl(p) (*(const volatile unsigned int *)(p))
 #define __raw_writel(v, p) {*(volatile unsigned int *)(p) = (v); }
 
-/* Allocator stuff */
-#define kmalloc(sz, t) malloc(sz)
-#define kfree(p)   { if (p) free(p); }
-static inline void *kzalloc(size_t sz, gfp_t __foo __rte_unused)
-{
-   void *ptr = malloc(sz);
-
-   if (ptr)
-   memset(ptr, 0, sz);
-   return ptr;
-}
-
-static inline unsigned long get_zeroed_page(gfp_t __foo __rte_unused)
-{
-   void *p;
-
-   if (posix_memalign(&p, 4096, 4096))
-   return 0;
-   memset(p, 0, 4096);
-   return (unsigned long)p;
-}
-
-static inline void free_page(unsigned long p)
-{
-   free((void *)p);
-}
-
 #define atomic_trte_atomic32_t
 #define atomic_read(v)  rte_atomic32_read(v)
 #define atomic_set(v, i)rte_atomic32_set(v, i)
diff --git a/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h 
b/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
index 24a6d4b..1e65660 100644
--- a/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
+++ b/drivers/bus/fslmc/qbman/include/fsl_qbman_portal.h
@@ -329,7 +329,7 @@ void qbman_pull_desc_clear(struct qbman_pull_desc *d);
  */
 void qbman_pull_desc_set_storage(struct qbman_pull_desc *d,
 struct qbman_result *storage,
-dma_addr_t storage_phys,
+uint64_t storage_phys,
 int stash);
 /**
  * qbman_pull_desc_set_numframes() - Set the number of frames to be dequeued.
@@ -796,7 +796,6 @@ struct qbman_eq_desc {
uint8_t wae;
uint8_t rspid;

[dpdk-dev] [PATCH v2 04/30] bus/fslmc: support up to 32 frames in one volatile dequeue

[Hemant Agrawal]

From: Haiying Wang 

QMan5.0 supports up to 32 frames in one volatile dequeue
command. For the older Qman versions which only support
up to 16 frames, the highest bit in NUMF will be ignored.

Signed-off-by: Haiying Wang 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/qbman/qbman_portal.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/bus/fslmc/qbman/qbman_portal.c 
b/drivers/bus/fslmc/qbman/qbman_portal.c
index f212829..7fc78cd 100644
--- a/drivers/bus/fslmc/qbman/qbman_portal.c
+++ b/drivers/bus/fslmc/qbman/qbman_portal.c
@@ -704,7 +704,7 @@ static struct qb_attr_code code_pull_dct = QB_CODE(0, 0, 2);
 static struct qb_attr_code code_pull_dt = QB_CODE(0, 2, 2);
 static struct qb_attr_code code_pull_rls = QB_CODE(0, 4, 1);
 static struct qb_attr_code code_pull_stash = QB_CODE(0, 5, 1);
-static struct qb_attr_code code_pull_numframes = QB_CODE(0, 8, 4);
+static struct qb_attr_code code_pull_numframes = QB_CODE(0, 8, 5);
 static struct qb_attr_code code_pull_token = QB_CODE(0, 16, 8);
 static struct qb_attr_code code_pull_dqsource = QB_CODE(1, 0, 24);
 static struct qb_attr_code code_pull_rsp_lo = QB_CODE(2, 0, 32);
@@ -743,7 +743,6 @@ void qbman_pull_desc_set_numframes(struct qbman_pull_desc 
*d, uint8_t numframes)
 {
uint32_t *cl = qb_cl(d);
 
-   QBMAN_BUG_ON(!numframes || (numframes > 16));
qb_attr_code_encode(&code_pull_numframes, cl,
(uint32_t)(numframes - 1));
 }
-- 
2.7.4

[dpdk-dev] [PATCH v2 11/30] crypto/dpaa2_sec: update MC to 10.3.x

[Hemant Agrawal]

From: Shreyansh Jain 

Signed-off-by: Shreyansh Jain 
---
 drivers/crypto/dpaa2_sec/mc/dpseci.c | 676 +--
 drivers/crypto/dpaa2_sec/mc/fsl_dpseci.h | 782 ---
 drivers/crypto/dpaa2_sec/mc/fsl_dpseci_cmd.h | 387 ++---
 3 files changed, 808 insertions(+), 1037 deletions(-)

diff --git a/drivers/crypto/dpaa2_sec/mc/dpseci.c 
b/drivers/crypto/dpaa2_sec/mc/dpseci.c
index 4a10962..2a216af 100644
--- a/drivers/crypto/dpaa2_sec/mc/dpseci.c
+++ b/drivers/crypto/dpaa2_sec/mc/dpseci.c
@@ -37,18 +37,34 @@
  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
  * POSSIBILITY OF SUCH DAMAGE.
  */
-
 #include 
 #include 
 #include 
 #include 
 
-int
-dpseci_open(struct fsl_mc_io *mc_io,
-   uint32_t cmd_flags,
-   int dpseci_id,
-   uint16_t *token)
+/**
+ * dpseci_open() - Open a control session for the specified object
+ * @mc_io: Pointer to MC portal's I/O object
+ * @cmd_flags: Command flags; one or more of 'MC_CMD_FLAG_'
+ * @dpseci_id: DPSECI unique ID
+ * @token: Returned token; use in subsequent API calls
+ *
+ * This function can be used to open a control session for an
+ * already created object; an object may have been declared in
+ * the DPL or by calling the dpseci_create() function.
+ * This function returns a unique authentication token,
+ * associated with the specific object ID and the specific MC
+ * portal; this token must be used in all subsequent commands for
+ * this specific object.
+ *
+ * Return: '0' on Success; Error code otherwise.
+ */
+int dpseci_open(struct fsl_mc_io *mc_io,
+   uint32_t cmd_flags,
+   int dpseci_id,
+   uint16_t *token)
 {
+   struct dpseci_cmd_open *cmd_params;
struct mc_command cmd = { 0 };
int err;
 
@@ -56,23 +72,34 @@ dpseci_open(struct fsl_mc_io *mc_io,
cmd.header = mc_encode_cmd_header(DPSECI_CMDID_OPEN,
  cmd_flags,
  0);
-   DPSECI_CMD_OPEN(cmd, dpseci_id);
+   cmd_params = (struct dpseci_cmd_open *)cmd.params;
+   cmd_params->dpseci_id = cpu_to_le32(dpseci_id);
 
-   /* send command to mc */
+   /* send command to mc*/
err = mc_send_command(mc_io, &cmd);
if (err)
return err;
 
/* retrieve response parameters */
-   *token = MC_CMD_HDR_READ_TOKEN(cmd.header);
+   *token = mc_cmd_hdr_read_token(&cmd);
 
return 0;
 }
 
-int
-dpseci_close(struct fsl_mc_io *mc_io,
-uint32_t cmd_flags,
-uint16_t token)
+/**
+ * dpseci_close() - Close the control session of the object
+ * @mc_io: Pointer to MC portal's I/O object
+ * @cmd_flags: Command flags; one or more of 'MC_CMD_FLAG_'
+ * @token: Token of DPSECI object
+ *
+ * After this function is called, no further operations are
+ * allowed on the object without opening a new control session.
+ *
+ * Return: '0' on Success; Error code otherwise.
+ */
+int dpseci_close(struct fsl_mc_io *mc_io,
+uint32_t cmd_flags,
+uint16_t token)
 {
struct mc_command cmd = { 0 };
 
@@ -81,478 +108,569 @@ dpseci_close(struct fsl_mc_io *mc_io,
  cmd_flags,
  token);
 
-   /* send command to mc */
+   /* send command to mc*/
return mc_send_command(mc_io, &cmd);
 }
 
-int
-dpseci_create(struct fsl_mc_io *mc_io,
- uint16_t dprc_token,
- uint32_t cmd_flags,
- const struct dpseci_cfg *cfg,
- uint32_t *obj_id)
+/**
+ * dpseci_create() - Create the DPSECI object
+ * @mc_io: Pointer to MC portal's I/O object
+ * @dprc_token:Parent container token; '0' for default container
+ * @cmd_flags: Command flags; one or more of 'MC_CMD_FLAG_'
+ * @cfg:   Configuration structure
+ * @obj_id:Returned object id
+ *
+ * Create the DPSECI object, allocate required resources and
+ * perform required initialization.
+ *
+ * The object can be created either by declaring it in the
+ * DPL file, or by calling this function.
+ *
+ * The function accepts an authentication token of a parent
+ * container that this object should be assigned to. The token
+ * can be '0' so the object will be assigned to the default container.
+ * The newly created object can be opened with the returned
+ * object id and using the container's associated tokens and MC portals.
+ *
+ * Return: '0' on Success; Error code otherwise.
+ */
+int dpseci_create(struct fsl_mc_io *mc_io,
+ uint16_t dprc_token,
+ uint32_t cmd_flags,
+ const struct dpseci_cfg *cfg,
+ uint32_t *obj_id)
 {
+   struct dpseci_cmd_create *cmd_params;
struct mc_command cmd = { 0 };
-   int err;
+   int err, i;
 
/* prepare command */
cmd.header = mc_encode_c

[dpdk-dev] [PATCH v2 09/30] bus/fslmc: update MC to 10.3.x

[Hemant Agrawal]

From: Shreyansh Jain 

Signed-off-by: Shreyansh Jain 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/fslmc_vfio.c   |  19 +++
 drivers/bus/fslmc/mc/dpbp.c  | 182 +
 drivers/bus/fslmc/mc/dpci.c  | 202 ---
 drivers/bus/fslmc/mc/dpcon.c | 163 ---
 drivers/bus/fslmc/mc/dpio.c  | 230 ---
 drivers/bus/fslmc/mc/dpmng.c |  33 +++-
 drivers/bus/fslmc/mc/fsl_dpbp.h  | 191 +-
 drivers/bus/fslmc/mc/fsl_dpbp_cmd.h  | 125 ++-
 drivers/bus/fslmc/mc/fsl_dpci.h  | 257 +++---
 drivers/bus/fslmc/mc/fsl_dpci_cmd.h  | 222 ++
 drivers/bus/fslmc/mc/fsl_dpcon.h | 186 --
 drivers/bus/fslmc/mc/fsl_dpcon_cmd.h | 193 --
 drivers/bus/fslmc/mc/fsl_dpio.h  | 299 +--
 drivers/bus/fslmc/mc/fsl_dpio_cmd.h  | 178 -
 drivers/bus/fslmc/mc/fsl_dpmng.h |  41 ++---
 drivers/bus/fslmc/mc/fsl_dpmng_cmd.h |  41 +++--
 drivers/bus/fslmc/mc/fsl_mc_cmd.h| 217 +++--
 drivers/bus/fslmc/mc/fsl_mc_sys.h|  36 ++---
 drivers/bus/fslmc/mc/mc_sys.c|   5 +-
 19 files changed, 1452 insertions(+), 1368 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
index 42a99e8..cdc982b 100644
--- a/drivers/bus/fslmc/fslmc_vfio.c
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -58,6 +58,7 @@
 
 #include "rte_fslmc.h"
 #include "fslmc_vfio.h"
+#include 
 
 #include "portal/dpaa2_hw_pvt.h"
 #include "portal/dpaa2_hw_dpio.h"
@@ -417,6 +418,8 @@ fslmc_process_mcp(struct rte_dpaa2_device *dev)
 {
int64_t v_addr;
char *dev_name;
+   struct fsl_mc_io dpmng  = {0};
+   struct mc_version mc_ver_info = {0};
 
rte_mcp_ptr_list = malloc(sizeof(void *) * 1);
if (!rte_mcp_ptr_list) {
@@ -441,6 +444,22 @@ fslmc_process_mcp(struct rte_dpaa2_device *dev)
return -1;
}
 
+   /* check the MC version compatibility */
+   dpmng.regs = (void *)v_addr;
+   if (mc_get_version(&dpmng, CMD_PRI_LOW, &mc_ver_info))
+   RTE_LOG(WARNING, PMD, "\tmc_get_version failed\n");
+
+   if ((mc_ver_info.major != MC_VER_MAJOR) ||
+   (mc_ver_info.minor < MC_VER_MINOR)) {
+   RTE_LOG(ERR, PMD, "DPAA2 MC version not compatible!"
+   " Expected %d.%d.x, Detected %d.%d.%d\n",
+   MC_VER_MAJOR, MC_VER_MINOR,
+   mc_ver_info.major, mc_ver_info.minor,
+   mc_ver_info.revision);
+   free(rte_mcp_ptr_list);
+   rte_mcp_ptr_list = NULL;
+   return -1;
+   }
rte_mcp_ptr_list[0] = (void *)v_addr;
 
return 0;
diff --git a/drivers/bus/fslmc/mc/dpbp.c b/drivers/bus/fslmc/mc/dpbp.c
index fd9a52d..a846245 100644
--- a/drivers/bus/fslmc/mc/dpbp.c
+++ b/drivers/bus/fslmc/mc/dpbp.c
@@ -5,7 +5,7 @@
  *   BSD LICENSE
  *
  * Copyright 2013-2016 Freescale Semiconductor Inc.
- * Copyright 2016 NXP.
+ * Copyright 2016-2017 NXP.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
@@ -42,19 +42,37 @@
 #include 
 #include 
 
+/**
+ * dpbp_open() - Open a control session for the specified object.
+ * @mc_io: Pointer to MC portal's I/O object
+ * @cmd_flags: Command flags; one or more of 'MC_CMD_FLAG_'
+ * @dpbp_id:   DPBP unique ID
+ * @token: Returned token; use in subsequent API calls
+ *
+ * This function can be used to open a control session for an
+ * already created object; an object may have been declared in
+ * the DPL or by calling the dpbp_create function.
+ * This function returns a unique authentication token,
+ * associated with the specific object ID and the specific MC
+ * portal; this token must be used in all subsequent commands for
+ * this specific object
+ *
+ * Return: '0' on Success; Error code otherwise.
+ */
 int dpbp_open(struct fsl_mc_io *mc_io,
  uint32_t cmd_flags,
  int dpbp_id,
  uint16_t *token)
 {
+   struct dpbp_cmd_open *cmd_params;
struct mc_command cmd = { 0 };
int err;
 
/* prepare command */
cmd.header = mc_encode_cmd_header(DPBP_CMDID_OPEN,
- cmd_flags,
- 0);
-   DPBP_CMD_OPEN(cmd, dpbp_id);
+ cmd_flags, 0);
+   cmd_params = (struct dpbp_cmd_open *)cmd.params;
+   cmd_params->dpbp_id = cpu_to_le32(dpbp_id);
 
/* send command to mc*/
err = mc_send_command(mc_io, &cmd);
@@ -62,11 +80,22 @@ int dpbp_open(struct fsl_mc_io *mc_io,
return err;
 
/* retrieve response parameters */
-   *token = MC_CMD_HDR_READ_TOKEN(cmd.header);

[dpdk-dev] [PATCH v2 12/30] config/dpaa2: change max lores to 16

[Hemant Agrawal]

From: Ashish Jain 

To support new LX2 series

Signed-off-by: Ashish Jain 
---
 config/defconfig_arm64-dpaa2-linuxapp-gcc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/config/defconfig_arm64-dpaa2-linuxapp-gcc 
b/config/defconfig_arm64-dpaa2-linuxapp-gcc
index 8a42944..91f4993 100644
--- a/config/defconfig_arm64-dpaa2-linuxapp-gcc
+++ b/config/defconfig_arm64-dpaa2-linuxapp-gcc
@@ -39,7 +39,7 @@ CONFIG_RTE_ARCH_ARM_TUNE="cortex-a72"
 #
 # Compile Environment Abstraction Layer
 #
-CONFIG_RTE_MAX_LCORE=8
+CONFIG_RTE_MAX_LCORE=16
 CONFIG_RTE_MAX_NUMA_NODES=1
 CONFIG_RTE_CACHE_LINE_SIZE=64
 
-- 
2.7.4

[dpdk-dev] [PATCH v2 15/30] bus/fslmc: add support to check dpbp presence

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c| 7 +++
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h | 1 +
 drivers/bus/fslmc/rte_bus_fslmc_version.map | 1 +
 3 files changed, 9 insertions(+)

diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
index 1ddd280..7e96115 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
@@ -129,6 +129,13 @@ void dpaa2_free_dpbp_dev(struct dpaa2_dpbp_dev *dpbp)
}
 }
 
+int dpaa2_dpbp_supported(void)
+{
+   if (TAILQ_EMPTY(&dpbp_dev_list))
+   return -1;
+   return 0;
+}
+
 static struct rte_dpaa2_object rte_dpaa2_dpbp_obj = {
.dev_type = DPAA2_BPOOL,
.create = dpaa2_create_dpbp_device,
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
index bb52a15..8f39cfb 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
@@ -366,6 +366,7 @@ void set_swp_active_dqs(uint16_t dpio_index, struct 
qbman_result *dqs)
 }
 struct dpaa2_dpbp_dev *dpaa2_alloc_dpbp_dev(void);
 void dpaa2_free_dpbp_dev(struct dpaa2_dpbp_dev *dpbp);
+int dpaa2_dpbp_supported(void);
 
 struct dpaa2_dpci_dev *rte_dpaa2_alloc_dpci_dev(void);
 void rte_dpaa2_free_dpci_dev(struct dpaa2_dpci_dev *dpci);
diff --git a/drivers/bus/fslmc/rte_bus_fslmc_version.map 
b/drivers/bus/fslmc/rte_bus_fslmc_version.map
index 3cc7dad..7b25248 100644
--- a/drivers/bus/fslmc/rte_bus_fslmc_version.map
+++ b/drivers/bus/fslmc/rte_bus_fslmc_version.map
@@ -83,6 +83,7 @@ DPDK_17.08 {
 DPDK_17.11 {
global:
 
+   dpaa2_dpbp_supported;
rte_dpaa2_dev_type;
 
 } DPDK_17.08;
-- 
2.7.4

[dpdk-dev] [PATCH v2 13/30] bus/fslmc: add support for LX2160 platform

[Hemant Agrawal]

From: Ashish Jain 

Signed-off-by: Ashish Jain 
---
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 22 ++
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h  |  1 +
 2 files changed, 15 insertions(+), 8 deletions(-)

diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index 731..8db1f6c 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -90,20 +90,22 @@ static int dpaa2_cluster_sz = 2;
  * Cluster 1 (ID = x02) : CPU0, CPU1, CPU2, CPU3;
  * Cluster 2 (ID = x03) : CPU4, CPU5, CPU6, CPU7;
  */
-
-/* Set the STASH Destination depending on Current CPU ID.
- * e.g. Valid values of SDEST are 4,5,6,7. Where,
- * CPU 0-1 will have SDEST 4
- * CPU 2-3 will have SDEST 5.and so on.
+/* For LX2160 platform There are four clusters with following mapping:
+ * Cluster 1 (ID = x00) : CPU0, CPU1;
+ * Cluster 2 (ID = x01) : CPU2, CPU3;
+ * Cluster 3 (ID = x02) : CPU4, CPU5;
+ * Cluster 4 (ID = x03) : CPU6, CPU7;
+ * Cluster 1 (ID = x04) : CPU8, CPU9;
+ * Cluster 2 (ID = x05) : CPU10, CP11;
+ * Cluster 3 (ID = x06) : CPU12, CPU13;
+ * Cluster 4 (ID = x07) : CPU14, CPU15;
  */
+
 static int
 dpaa2_core_cluster_sdest(int cpu_id)
 {
int x = cpu_id / dpaa2_cluster_sz;
 
-   if (x > 3)
-   x = 3;
-
return dpaa2_core_cluster_base + x;
 }
 
@@ -278,6 +280,10 @@ dpaa2_configure_stashing(struct dpaa2_dpio_dev *dpio_dev, 
int cpu_id)
dpaa2_core_cluster_base = 0x02;
dpaa2_cluster_sz = 4;
PMD_INIT_LOG(DEBUG, "\tLS108x (A53) Platform Detected");
+   } else if ((mc_plat_info.svr & 0x) == SVR_LX2160A) {
+   dpaa2_core_cluster_base = 0x00;
+   dpaa2_cluster_sz = 2;
+   PMD_INIT_LOG(DEBUG, "\tLX2160 Platform Detected");
}
first_time = 1;
}
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
index 5d7a828..bb52a15 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
@@ -51,6 +51,7 @@
 #define SVR_LS1080A 0x8703
 #define SVR_LS2080A 0x8701
 #define SVR_LS2088A 0x8709
+#define SVR_LX2160A 0x8736
 
 #ifndef ETH_VLAN_HLEN
 #define ETH_VLAN_HLEN   4 /** < Vlan Header Length */
-- 
2.7.4

[dpdk-dev] [PATCH v2 16/30] bus/fslmc: cleanup the dpaa2 interrupt support

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/Makefile  |   1 +
 drivers/bus/fslmc/fslmc_vfio.c  | 108 +++-
 drivers/bus/fslmc/fslmc_vfio.h  |   8 ++-
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c|  18 +++--
 drivers/bus/fslmc/rte_bus_fslmc_version.map |   2 +
 5 files changed, 110 insertions(+), 27 deletions(-)

diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index d1b790b..37da1b0 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -51,6 +51,7 @@ CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc/mc
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc/qbman/include
 CFLAGS += -I$(RTE_SDK)/lib/librte_eal/linuxapp/eal
+CFLAGS += -I$(RTE_SDK)/lib/librte_eal/common
 
 # versioning export map
 EXPORT_MAP := rte_bus_fslmc_version.map
diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
index cdc982b..ab1df36 100644
--- a/drivers/bus/fslmc/fslmc_vfio.c
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -46,6 +46,7 @@
 #include 
 #include 
 
+#include 
 #include 
 #include 
 #include 
@@ -335,36 +336,107 @@ static int64_t vfio_map_mcp_obj(struct fslmc_vfio_group 
*group, char *mcp_obj)
 
 #define IRQ_SET_BUF_LEN  (sizeof(struct vfio_irq_set) + sizeof(int))
 
-int rte_dpaa2_intr_enable(struct rte_intr_handle *intr_handle,
- uint32_t index)
+int rte_dpaa2_intr_enable(struct rte_intr_handle *intr_handle, int index)
 {
-   struct vfio_irq_set *irq_set;
+   int len, ret;
char irq_set_buf[IRQ_SET_BUF_LEN];
-   int *fd_ptr, fd, ret;
+   struct vfio_irq_set *irq_set;
+   int *fd_ptr;
+
+   len = sizeof(irq_set_buf);
 
-   /* Prepare vfio_irq_set structure and SET the IRQ in VFIO */
-   /* Give the eventfd to VFIO */
-   fd = eventfd(0, 0);
irq_set = (struct vfio_irq_set *)irq_set_buf;
-   irq_set->argsz = sizeof(irq_set_buf);
+   irq_set->argsz = len;
irq_set->count = 1;
-   irq_set->flags = VFIO_IRQ_SET_DATA_EVENTFD |
-VFIO_IRQ_SET_ACTION_TRIGGER;
+   irq_set->flags =
+   VFIO_IRQ_SET_DATA_EVENTFD | VFIO_IRQ_SET_ACTION_TRIGGER;
irq_set->index = index;
irq_set->start = 0;
fd_ptr = (int *)&irq_set->data;
-   *fd_ptr = fd;
+   *fd_ptr = intr_handle->fd;
 
ret = ioctl(intr_handle->vfio_dev_fd, VFIO_DEVICE_SET_IRQS, irq_set);
-   if (ret < 0) {
-   FSLMC_VFIO_LOG(ERR, "Unable to set IRQ in VFIO, ret: %d\n",
-  ret);
-   return -1;
+   if (ret) {
+   RTE_LOG(ERR, EAL, "Error:dpaa2 SET IRQs fd=%d, err = %d(%s)\n",
+   intr_handle->fd, errno, strerror(errno));
+   return ret;
}
 
-   /* Set the FD and update the flags */
-   intr_handle->fd = fd;
-   return 0;
+   return ret;
+}
+
+int rte_dpaa2_intr_disable(struct rte_intr_handle *intr_handle, int index)
+{
+   struct vfio_irq_set *irq_set;
+   char irq_set_buf[IRQ_SET_BUF_LEN];
+   int len, ret;
+
+   len = sizeof(struct vfio_irq_set);
+
+   irq_set = (struct vfio_irq_set *)irq_set_buf;
+   irq_set->argsz = len;
+   irq_set->flags = VFIO_IRQ_SET_DATA_NONE | VFIO_IRQ_SET_ACTION_TRIGGER;
+   irq_set->index = index;
+   irq_set->start = 0;
+   irq_set->count = 0;
+
+   ret = ioctl(intr_handle->vfio_dev_fd, VFIO_DEVICE_SET_IRQS, irq_set);
+   if (ret)
+   RTE_LOG(ERR, EAL,
+   "Error disabling dpaa2 interrupts for fd %d\n",
+   intr_handle->fd);
+
+   return ret;
+}
+
+/* set up interrupt support (but not enable interrupts) */
+int
+rte_dpaa2_vfio_setup_intr(struct rte_intr_handle *intr_handle,
+ int vfio_dev_fd,
+ int num_irqs)
+{
+   int i, ret;
+
+   /* start from MSI-X interrupt type */
+   for (i = 0; i < num_irqs; i++) {
+   struct vfio_irq_info irq_info = { .argsz = sizeof(irq_info) };
+   int fd = -1;
+
+   irq_info.index = i;
+
+   ret = ioctl(vfio_dev_fd, VFIO_DEVICE_GET_IRQ_INFO, &irq_info);
+   if (ret < 0) {
+   FSLMC_VFIO_LOG(ERR,
+  "cannot get IRQ(%d) info, error %i (%s)",
+  i, errno, strerror(errno));
+   return -1;
+   }
+
+   /* if this vector cannot be used with eventfd,
+* fail if we explicitly
+* specified interrupt type, otherwise continue
+*/
+   if ((irq_info.flags & VFIO_IRQ_INFO_EVENTFD) == 0)
+   continue;
+
+   /* set up an eventfd for interrupts */
+   fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
+   if (fd < 0) {
+

[dpdk-dev] [PATCH v2 14/30] net/dpaa2: add support for congestion overhead

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_ethdev.c | 1 +
 drivers/net/dpaa2/dpaa2_ethdev.h | 1 +
 2 files changed, 2 insertions(+)

diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 945dcc7..d7950a5 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -419,6 +419,7 @@ dpaa2_dev_rx_queue_setup(struct rte_eth_dev *dev,
/*enabling per rx queue congestion control */
taildrop.threshold = CONG_THRESHOLD_RX_Q;
taildrop.units = DPNI_CONGESTION_UNIT_BYTES;
+   taildrop.oal = CONG_RX_OAL;
PMD_INIT_LOG(DEBUG, "Enabling Early Drop on queue = %d",
 rx_queue_id);
ret = dpni_set_taildrop(dpni, CMD_PRI_LOW, priv->token,
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.h b/drivers/net/dpaa2/dpaa2_ethdev.h
index 07e1da6..6ee1a9f 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.h
+++ b/drivers/net/dpaa2/dpaa2_ethdev.h
@@ -61,6 +61,7 @@
  * currently considering 32 KB packets
  */
 #define CONG_THRESHOLD_RX_Q  (64 * 1024)
+#define CONG_RX_OAL128
 
 /* Size of the input SMMU mapped memory required by MC */
 #define DIST_PARAM_IOVA_SIZE 256
-- 
2.7.4

[dpdk-dev] [PATCH v2 17/30] net/dpaa2: add support for link status event

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini  |   1 +
 drivers/net/dpaa2/dpaa2_ethdev.c| 123 +++
 drivers/net/dpaa2/mc/dpni.c | 233 
 drivers/net/dpaa2/mc/fsl_dpni.h |  49 
 drivers/net/dpaa2/mc/fsl_dpni_cmd.h |  50 
 5 files changed, 456 insertions(+)

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index 146e087..ba4321c 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -6,6 +6,7 @@
 [Features]
 Speed capabilities   = P
 Link status  = Y
+Link status event= Y
 Queue start/stop = Y
 Jumbo frame  = Y
 MTU update   = Y
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index d7950a5..49dc42b 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -54,6 +54,8 @@
 
 static struct rte_dpaa2_driver rte_dpaa2_pmd;
 static int dpaa2_dev_uninit(struct rte_eth_dev *eth_dev);
+static int dpaa2_dev_link_update(struct rte_eth_dev *dev,
+int wait_to_complete);
 static int dpaa2_dev_set_link_up(struct rte_eth_dev *dev);
 static int dpaa2_dev_set_link_down(struct rte_eth_dev *dev);
 static int dpaa2_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu);
@@ -344,6 +346,10 @@ dpaa2_eth_dev_configure(struct rte_eth_dev *dev)
return ret;
}
}
+
+   /* update the current status */
+   dpaa2_dev_link_update(dev, 0);
+
return 0;
 }
 
@@ -556,9 +562,87 @@ dpaa2_supported_ptypes_get(struct rte_eth_dev *dev)
return NULL;
 }
 
+/**
+ * Dpaa2 link Interrupt handler
+ *
+ * @param param
+ *  The address of parameter (struct rte_eth_dev *) regsitered before.
+ *
+ * @return
+ *  void
+ */
+static void
+dpaa2_interrupt_handler(void *param)
+{
+   struct rte_eth_dev *dev = param;
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+   struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
+   int ret;
+   int irq_index = DPNI_IRQ_INDEX;
+   unsigned int status, clear = 0;
+
+   PMD_INIT_FUNC_TRACE();
+
+   if (dpni == NULL) {
+   RTE_LOG(ERR, PMD, "dpni is NULL");
+   return;
+   }
+
+   ret = dpni_get_irq_status(dpni, CMD_PRI_LOW, priv->token,
+ irq_index, &status);
+   if (unlikely(ret)) {
+   RTE_LOG(ERR, PMD, "Can't get irq status (err %d)", ret);
+   clear = 0x;
+   goto out;
+   }
+
+   if (status & DPNI_IRQ_EVENT_LINK_CHANGED) {
+   clear = DPNI_IRQ_EVENT_LINK_CHANGED;
+   dpaa2_dev_link_update(dev, 0);
+   /* calling all the apps registered for link status event */
+   _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC,
+ NULL, NULL);
+   }
+out:
+   ret = dpni_clear_irq_status(dpni, CMD_PRI_LOW, priv->token,
+   irq_index, clear);
+   if (unlikely(ret))
+   RTE_LOG(ERR, PMD, "Can't clear irq status (err %d)", ret);
+}
+
+static int
+dpaa2_eth_setup_irqs(struct rte_eth_dev *dev, int enable)
+{
+   int err = 0;
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+   struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
+   int irq_index = DPNI_IRQ_INDEX;
+   unsigned int mask = DPNI_IRQ_EVENT_LINK_CHANGED;
+
+   PMD_INIT_FUNC_TRACE();
+
+   err = dpni_set_irq_mask(dpni, CMD_PRI_LOW, priv->token,
+   irq_index, mask);
+   if (err < 0) {
+   PMD_INIT_LOG(ERR, "Error: dpni_set_irq_mask():%d (%s)", err,
+strerror(-err));
+   return err;
+   }
+
+   err = dpni_set_irq_enable(dpni, CMD_PRI_LOW, priv->token,
+ irq_index, enable);
+   if (err < 0)
+   PMD_INIT_LOG(ERR, "Error: dpni_set_irq_enable():%d (%s)", err,
+strerror(-err));
+
+   return err;
+}
+
 static int
 dpaa2_dev_start(struct rte_eth_dev *dev)
 {
+   struct rte_device *rdev = dev->device;
+   struct rte_dpaa2_device *dpaa2_dev;
struct rte_eth_dev_data *data = dev->data;
struct dpaa2_dev_priv *priv = data->dev_private;
struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
@@ -568,6 +652,10 @@ dpaa2_dev_start(struct rte_eth_dev *dev)
struct dpni_queue_id qid;
struct dpaa2_queue *dpaa2_q;
int ret, i;
+   struct rte_intr_handle *intr_handle;
+
+   dpaa2_dev = container_of(rdev, struct rte_dpaa2_device, device);
+   intr_handle = &dpaa2_dev->intr_handle;
 
PMD_INIT_FUNC_TRACE();
 
@@ -647,6 +735,24 @@ dpaa2_dev_start(struct rte_eth_dev *dev)
if (priv->max_vlan_filters)
dpaa2_vlan_offload_set(de

[dpdk-dev] [PATCH v2 18/30] bus/fslmc: enable link status interrupt

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/fslmc_vfio.c | 4 
 1 file changed, 4 insertions(+)

diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
index ab1df36..81f0779 100644
--- a/drivers/bus/fslmc/fslmc_vfio.c
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -464,6 +464,10 @@ fslmc_process_iodevices(struct rte_dpaa2_device *dev)
}
 
switch (dev->dev_type) {
+   case DPAA2_ETH:
+   rte_dpaa2_vfio_setup_intr(&dev->intr_handle, dev_fd,
+ device_info.num_irqs);
+   break;
case DPAA2_CON:
case DPAA2_IO:
case DPAA2_CI:
-- 
2.7.4

[dpdk-dev] [PATCH v2 21/30] net/dpaa2: increase the dist param to 64 bit

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/base/dpaa2_hw_dpni.c | 6 +++---
 drivers/net/dpaa2/dpaa2_ethdev.h   | 2 +-
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c 
b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
index 7e5ce64..e3ab90a 100644
--- a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
+++ b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
@@ -51,12 +51,12 @@
 
 static int
 dpaa2_distset_to_dpkg_profile_cfg(
-   uint32_t req_dist_set,
+   uint64_t req_dist_set,
struct dpkg_profile_cfg *kg_cfg);
 
 int
 dpaa2_setup_flow_dist(struct rte_eth_dev *eth_dev,
- uint32_t req_dist_set)
+ uint64_t req_dist_set)
 {
struct dpaa2_dev_priv *priv = eth_dev->data->dev_private;
struct fsl_mc_io *dpni = priv->hw;
@@ -148,7 +148,7 @@ int dpaa2_remove_flow_dist(
 
 static int
 dpaa2_distset_to_dpkg_profile_cfg(
-   uint32_t req_dist_set,
+   uint64_t req_dist_set,
struct dpkg_profile_cfg *kg_cfg)
 {
uint32_t loop = 0, i = 0, dist_field = 0;
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.h b/drivers/net/dpaa2/dpaa2_ethdev.h
index 6ee1a9f..7b14ae0 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.h
+++ b/drivers/net/dpaa2/dpaa2_ethdev.h
@@ -93,7 +93,7 @@ struct dpaa2_dev_priv {
 };
 
 int dpaa2_setup_flow_dist(struct rte_eth_dev *eth_dev,
- uint32_t req_dist_set);
+ uint64_t req_dist_set);
 
 int dpaa2_remove_flow_dist(struct rte_eth_dev *eth_dev,
   uint8_t tc_index);
-- 
2.7.4

[dpdk-dev] [PATCH v2 19/30] net/dpaa2: check physical link state on up cmd

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_ethdev.c | 23 +++
 1 file changed, 15 insertions(+), 8 deletions(-)

diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 49dc42b..74a61b5 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -666,7 +666,7 @@ dpaa2_dev_start(struct rte_eth_dev *dev)
return ret;
}
 
-   /* Power up the phy. Needed to make the link go Up */
+   /* Power up the phy. Needed to make the link go UP */
dpaa2_dev_set_link_up(dev);
 
ret = dpni_get_qdid(dpni, CMD_PRI_LOW, priv->token,
@@ -1137,8 +1137,6 @@ dpaa2_dev_link_update(struct rte_eth_dev *dev,
struct rte_eth_link link, old;
struct dpni_link_state state = {0};
 
-   PMD_INIT_FUNC_TRACE();
-
if (dpni == NULL) {
RTE_LOG(ERR, PMD, "dpni is NULL\n");
return 0;
@@ -1171,7 +1169,7 @@ dpaa2_dev_link_update(struct rte_eth_dev *dev,
if (link.link_status)
PMD_DRV_LOG(INFO, "Port %d Link is Up\n", dev->data->port_id);
else
-   PMD_DRV_LOG(INFO, "Port %d Link is Down\n", dev->data->port_id);
+   PMD_DRV_LOG(INFO, "Port %d Link is Down", dev->data->port_id);
return 0;
 }
 
@@ -1186,8 +1184,7 @@ dpaa2_dev_set_link_up(struct rte_eth_dev *dev)
struct dpaa2_dev_priv *priv;
struct fsl_mc_io *dpni;
int en = 0;
-
-   PMD_INIT_FUNC_TRACE();
+   struct dpni_link_state state = {0};
 
priv = dev->data->dev_private;
dpni = (struct fsl_mc_io *)priv->hw;
@@ -1213,11 +1210,21 @@ dpaa2_dev_set_link_up(struct rte_eth_dev *dev)
return -EINVAL;
}
}
+   ret = dpni_get_link_state(dpni, CMD_PRI_LOW, priv->token, &state);
+   if (ret < 0) {
+   RTE_LOG(ERR, PMD, "error: dpni_get_link_state %d\n", ret);
+   return -1;
+   }
+
/* changing tx burst function to start enqueues */
dev->tx_pkt_burst = dpaa2_dev_tx;
-   dev->data->dev_link.link_status = 1;
+   dev->data->dev_link.link_status = state.up;
 
-   PMD_DRV_LOG(INFO, "Port %d Link UP successful", dev->data->port_id);
+   if (state.up)
+   PMD_DRV_LOG(INFO, "Port %d Link is set as UP",
+   dev->data->port_id);
+   else
+   PMD_DRV_LOG(INFO, "Port %d Link is DOWN", dev->data->port_id);
return ret;
 }
 
-- 
2.7.4

[dpdk-dev] [PATCH v2 20/30] net/dpaa2: improve error and logs for flow distribution

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/base/dpaa2_hw_dpni.c | 39 +-
 1 file changed, 24 insertions(+), 15 deletions(-)

diff --git a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c 
b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
index 1269dd2..7e5ce64 100644
--- a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
+++ b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
@@ -49,7 +49,7 @@
 
 #include "../dpaa2_ethdev.h"
 
-static void
+static int
 dpaa2_distset_to_dpkg_profile_cfg(
uint32_t req_dist_set,
struct dpkg_profile_cfg *kg_cfg);
@@ -68,20 +68,26 @@ dpaa2_setup_flow_dist(struct rte_eth_dev *eth_dev,
p_params = rte_malloc(
NULL, DIST_PARAM_IOVA_SIZE, RTE_CACHE_LINE_SIZE);
if (!p_params) {
-   RTE_LOG(ERR, PMD, "Memory unavaialble\n");
+   PMD_INIT_LOG(ERR, "Memory unavailable");
return -ENOMEM;
}
memset(p_params, 0, DIST_PARAM_IOVA_SIZE);
memset(&tc_cfg, 0, sizeof(struct dpni_rx_tc_dist_cfg));
 
-   dpaa2_distset_to_dpkg_profile_cfg(req_dist_set, &kg_cfg);
+   ret = dpaa2_distset_to_dpkg_profile_cfg(req_dist_set, &kg_cfg);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "given rss_hf (%lx) not supported",
+req_dist_set);
+   rte_free(p_params);
+   return ret;
+   }
tc_cfg.key_cfg_iova = (uint64_t)(DPAA2_VADDR_TO_IOVA(p_params));
tc_cfg.dist_size = eth_dev->data->nb_rx_queues;
tc_cfg.dist_mode = DPNI_DIST_MODE_HASH;
 
ret = dpkg_prepare_key_cfg(&kg_cfg, p_params);
if (ret) {
-   RTE_LOG(ERR, PMD, "Unable to prepare extract parameters\n");
+   PMD_INIT_LOG(ERR, "Unable to prepare extract parameters");
rte_free(p_params);
return ret;
}
@@ -90,9 +96,9 @@ dpaa2_setup_flow_dist(struct rte_eth_dev *eth_dev,
  &tc_cfg);
rte_free(p_params);
if (ret) {
-   RTE_LOG(ERR, PMD,
-   "Setting distribution for Rx failed with err: %d\n",
-   ret);
+   PMD_INIT_LOG(ERR,
+"Setting distribution for Rx failed with err: %d",
+ret);
return ret;
}
 
@@ -113,7 +119,7 @@ int dpaa2_remove_flow_dist(
p_params = rte_malloc(
NULL, DIST_PARAM_IOVA_SIZE, RTE_CACHE_LINE_SIZE);
if (!p_params) {
-   RTE_LOG(ERR, PMD, "Memory unavaialble\n");
+   PMD_INIT_LOG(ERR, "Memory unavailable");
return -ENOMEM;
}
memset(p_params, 0, DIST_PARAM_IOVA_SIZE);
@@ -125,7 +131,7 @@ int dpaa2_remove_flow_dist(
 
ret = dpkg_prepare_key_cfg(&kg_cfg, p_params);
if (ret) {
-   RTE_LOG(ERR, PMD, "Unable to prepare extract parameters\n");
+   PMD_INIT_LOG(ERR, "Unable to prepare extract parameters");
rte_free(p_params);
return ret;
}
@@ -134,13 +140,13 @@ int dpaa2_remove_flow_dist(
  &tc_cfg);
rte_free(p_params);
if (ret)
-   RTE_LOG(ERR, PMD,
-   "Setting distribution for Rx failed with err: %d\n",
-   ret);
+   PMD_INIT_LOG(ERR,
+"Setting distribution for Rx failed with err:%d",
+ret);
return ret;
 }
 
-static void
+static int
 dpaa2_distset_to_dpkg_profile_cfg(
uint32_t req_dist_set,
struct dpkg_profile_cfg *kg_cfg)
@@ -276,14 +282,17 @@ dpaa2_distset_to_dpkg_profile_cfg(
break;
 
default:
-   PMD_DRV_LOG(WARNING, "Bad flow distribution"
-   " option %x\n", dist_field);
+   PMD_INIT_LOG(WARNING,
+"Unsupported flow dist option %x",
+dist_field);
+   return -EINVAL;
}
}
req_dist_set = req_dist_set >> 1;
loop++;
}
kg_cfg->num_extracts = i;
+   return 0;
 }
 
 int
-- 
2.7.4

[dpdk-dev] [PATCH v2 22/30] net/dpaa2: remove RSS restriction with num of queues

[Hemant Agrawal]

DPAA2 HW does not have such restrictions.

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_ethdev.c | 11 ---
 1 file changed, 11 deletions(-)

diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 74a61b5..0fe7a13 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -326,18 +326,7 @@ dpaa2_eth_dev_configure(struct rte_eth_dev *dev)
}
}
 
-   /* Check for correct configuration */
-   if (eth_conf->rxmode.mq_mode != ETH_MQ_RX_RSS &&
-   data->nb_rx_queues > 1) {
-   PMD_INIT_LOG(ERR, "Distribution is not enabled, "
-   "but Rx queues more than 1\n");
-   return -1;
-   }
-
if (eth_conf->rxmode.mq_mode == ETH_MQ_RX_RSS) {
-   /* Return in case number of Rx queues is 1 */
-   if (data->nb_rx_queues == 1)
-   return 0;
ret = dpaa2_setup_flow_dist(dev,
eth_conf->rx_adv_conf.rss_conf.rss_hf);
if (ret) {
-- 
2.7.4

[dpdk-dev] [PATCH v2 23/30] net/dpaa2: add support for RSS hash update and get

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_ethdev.c | 42 
 1 file changed, 42 insertions(+)

diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 0fe7a13..aebcc71 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -1418,6 +1418,46 @@ dpaa2_flow_ctrl_set(struct rte_eth_dev *dev, struct 
rte_eth_fc_conf *fc_conf)
return ret;
 }
 
+static int
+dpaa2_dev_rss_hash_update(struct rte_eth_dev *dev,
+ struct rte_eth_rss_conf *rss_conf)
+{
+   struct rte_eth_dev_data *data = dev->data;
+   struct rte_eth_conf *eth_conf = &data->dev_conf;
+   int ret;
+
+   PMD_INIT_FUNC_TRACE();
+
+   if (rss_conf->rss_hf) {
+   ret = dpaa2_setup_flow_dist(dev, rss_conf->rss_hf);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "unable to set flow dist");
+   return ret;
+   }
+   } else {
+   ret = dpaa2_remove_flow_dist(dev, 0);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "unable to remove flow dist");
+   return ret;
+   }
+   }
+   eth_conf->rx_adv_conf.rss_conf.rss_hf = rss_conf->rss_hf;
+   return 0;
+}
+
+static int
+dpaa2_dev_rss_hash_conf_get(struct rte_eth_dev *dev,
+   struct rte_eth_rss_conf *rss_conf)
+{
+   struct rte_eth_dev_data *data = dev->data;
+   struct rte_eth_conf *eth_conf = &data->dev_conf;
+
+   /* dpaa2 does not support rss_key, so length should be 0*/
+   rss_conf->rss_key_len = 0;
+   rss_conf->rss_hf = eth_conf->rx_adv_conf.rss_conf.rss_hf;
+   return 0;
+}
+
 static struct eth_dev_ops dpaa2_ethdev_ops = {
.dev_configure= dpaa2_eth_dev_configure,
.dev_start= dpaa2_dev_start,
@@ -1447,6 +1487,8 @@ static struct eth_dev_ops dpaa2_ethdev_ops = {
.mac_addr_add = dpaa2_dev_add_mac_addr,
.mac_addr_remove  = dpaa2_dev_remove_mac_addr,
.mac_addr_set = dpaa2_dev_set_mac_addr,
+   .rss_hash_update  = dpaa2_dev_rss_hash_update,
+   .rss_hash_conf_get= dpaa2_dev_rss_hash_conf_get,
 };
 
 static int
-- 
2.7.4

[dpdk-dev] [PATCH v2 25/30] net/dpaa2: add support for extra stats

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini |   1 +
 drivers/net/dpaa/dpaa_ethdev.c | 143 ++
 drivers/net/dpaa/dpaa_ethdev.h |  40 +
 drivers/net/dpaa2/dpaa2_ethdev.c   | 174 -
 4 files changed, 356 insertions(+), 2 deletions(-)

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index ba4321c..6ebbab4 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -21,6 +21,7 @@ L3 checksum offload  = Y
 L4 checksum offload  = Y
 Packet type parsing  = Y
 Basic stats  = Y
+Extended stats   = Y
 FW version   = Y
 Linux VFIO   = Y
 ARMv8= Y
diff --git a/drivers/net/dpaa/dpaa_ethdev.c b/drivers/net/dpaa/dpaa_ethdev.c
index e94cf7c..82a87f6 100644
--- a/drivers/net/dpaa/dpaa_ethdev.c
+++ b/drivers/net/dpaa/dpaa_ethdev.c
@@ -78,6 +78,40 @@ static int is_global_init;
 
 static struct rte_dpaa_driver rte_dpaa_pmd;
 
+struct rte_dpaa_xstats_name_off {
+   char name[RTE_ETH_XSTATS_NAME_SIZE];
+   uint32_t offset;
+};
+
+static const struct rte_dpaa_xstats_name_off dpaa_xstats_strings[] = {
+   {"rx_align_err",
+   offsetof(struct dpaa_if_stats, raln)},
+   {"rx_valid_pause",
+   offsetof(struct dpaa_if_stats, rxpf)},
+   {"rx_fcs_err",
+   offsetof(struct dpaa_if_stats, rfcs)},
+   {"rx_vlan_frame",
+   offsetof(struct dpaa_if_stats, rvlan)},
+   {"rx_frame_err",
+   offsetof(struct dpaa_if_stats, rerr)},
+   {"rx_drop_err",
+   offsetof(struct dpaa_if_stats, rdrp)},
+   {"rx_undersized",
+   offsetof(struct dpaa_if_stats, rund)},
+   {"rx_oversize_err",
+   offsetof(struct dpaa_if_stats, rovr)},
+   {"rx_fragment_pkt",
+   offsetof(struct dpaa_if_stats, rfrg)},
+   {"tx_valid_pause",
+   offsetof(struct dpaa_if_stats, txpf)},
+   {"tx_fcs_err",
+   offsetof(struct dpaa_if_stats, terr)},
+   {"tx_vlan_frame",
+   offsetof(struct dpaa_if_stats, tvlan)},
+   {"rx_undersized",
+   offsetof(struct dpaa_if_stats, tund)},
+};
+
 static int
 dpaa_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
 {
@@ -271,6 +305,110 @@ static void dpaa_eth_stats_reset(struct rte_eth_dev *dev)
fman_if_stats_reset(dpaa_intf->fif);
 }
 
+static int
+dpaa_dev_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *xstats,
+   unsigned int n)
+{
+   struct dpaa_if *dpaa_intf = dev->data->dev_private;
+   unsigned int i = 0, num = RTE_DIM(dpaa_xstats_strings);
+   uint64_t values[sizeof(struct dpaa_if_stats) / 8];
+
+   if (xstats == NULL)
+   return 0;
+
+   if (n < num)
+   return num;
+
+   fman_if_stats_get_all(dpaa_intf->fif, values,
+ sizeof(struct dpaa_if_stats) / 8);
+
+   for (i = 0; i < num; i++) {
+   xstats[i].id = i;
+   xstats[i].value = values[dpaa_xstats_strings[i].offset / 8];
+   }
+   return i;
+}
+
+static int
+dpaa_xstats_get_names(__rte_unused struct rte_eth_dev *dev,
+ struct rte_eth_xstat_name *xstats_names,
+ __rte_unused unsigned int limit)
+{
+   unsigned int i, stat_cnt = RTE_DIM(dpaa_xstats_strings);
+
+   if (xstats_names != NULL)
+   for (i = 0; i < stat_cnt; i++)
+   snprintf(xstats_names[i].name,
+sizeof(xstats_names[i].name),
+"%s",
+dpaa_xstats_strings[i].name);
+
+   return stat_cnt;
+}
+
+static int
+dpaa_xstats_get_by_id(struct rte_eth_dev *dev, const uint64_t *ids,
+ uint64_t *values, unsigned int n)
+{
+   unsigned int i, stat_cnt = RTE_DIM(dpaa_xstats_strings);
+   uint64_t values_copy[sizeof(struct dpaa_if_stats) / 8];
+
+   if (!ids) {
+   struct dpaa_if *dpaa_intf = dev->data->dev_private;
+
+   if (n < stat_cnt)
+   return stat_cnt;
+
+   if (!values)
+   return 0;
+
+   fman_if_stats_get_all(dpaa_intf->fif, values_copy,
+ sizeof(struct dpaa_if_stats));
+
+   for (i = 0; i < stat_cnt; i++)
+   values[i] =
+   values_copy[dpaa_xstats_strings[i].offset / 8];
+
+   return stat_cnt;
+   }
+
+   dpaa_xstats_get_by_id(dev, NULL, values_copy, stat_cnt);
+
+   for (i = 0; i < n; i++) {
+   if (ids[i] >= stat_cnt) {
+   DPAA_PMD_ERR("id value isn't valid");
+   return -1;
+   }
+   values[i] = values_copy[ids[i]];
+   }
+   return n;
+}
+
+static int
+dpaa_xs

[dpdk-dev] [PATCH v2 24/30] bus/dpaa2: add support for hw extra stats API

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/dpaa/base/fman/fman_hw.c  | 30 ++
 drivers/bus/dpaa/include/fsl_fman.h   |  3 +++
 drivers/bus/dpaa/rte_bus_dpaa_version.map |  1 +
 3 files changed, 34 insertions(+)

diff --git a/drivers/bus/dpaa/base/fman/fman_hw.c 
b/drivers/bus/dpaa/base/fman/fman_hw.c
index be59858..90ded59 100644
--- a/drivers/bus/dpaa/base/fman/fman_hw.c
+++ b/drivers/bus/dpaa/base/fman/fman_hw.c
@@ -234,6 +234,36 @@ fman_if_stats_get(struct fman_if *p, struct rte_eth_stats 
*stats)
 }
 
 void
+fman_if_stats_get_all(struct fman_if *p, uint64_t *value, int n)
+{
+   struct __fman_if *m = container_of(p, struct __fman_if, __if);
+   struct memac_regs *regs = m->ccsr_map;
+   int i;
+   uint64_t base_offset = offsetof(struct memac_regs, reoct_l);
+
+   for (i = 0; i < n; i++)
+   value[i] = ((u64)in_be32((char *)regs
+   + base_offset + 8 * i + 4)) << 32 |
+   ((u64)in_be32((char *)regs
+   + base_offset + 8 * i));
+}
+
+void
+fman_if_stats_get_all(struct fman_if *p, uint64_t *value, int n)
+{
+   struct __fman_if *m = container_of(p, struct __fman_if, __if);
+   struct memac_regs *regs = m->ccsr_map;
+   int i;
+   uint64_t base_offset = offsetof(struct memac_regs, reoct_l);
+
+   for (i = 0; i < n; i++)
+   value[i] = ((u64)in_be32((char *)regs
+   + base_offset + 8 * i + 4)) << 32 |
+   ((u64)in_be32((char *)regs
+   + base_offset + 8 * i));
+}
+
+void
 fman_if_stats_reset(struct fman_if *p)
 {
struct __fman_if *m = container_of(p, struct __fman_if, __if);
diff --git a/drivers/bus/dpaa/include/fsl_fman.h 
b/drivers/bus/dpaa/include/fsl_fman.h
index 421a581..9eeb3cd 100644
--- a/drivers/bus/dpaa/include/fsl_fman.h
+++ b/drivers/bus/dpaa/include/fsl_fman.h
@@ -87,6 +87,9 @@ void fman_if_stats_get(struct fman_if *p, struct 
rte_eth_stats *stats);
 /* Reset the FMAN statistics */
 void fman_if_stats_reset(struct fman_if *p);
 
+/* Get all of the FMAN statistics */
+void fman_if_stats_get_all(struct fman_if *p, uint64_t *value, int n);
+
 /* Set ignore pause option for a specific interface */
 void fman_if_set_rx_ignore_pause_frames(struct fman_if *p, bool enable);
 
diff --git a/drivers/bus/dpaa/rte_bus_dpaa_version.map 
b/drivers/bus/dpaa/rte_bus_dpaa_version.map
index f231231..a8ac56d 100644
--- a/drivers/bus/dpaa/rte_bus_dpaa_version.map
+++ b/drivers/bus/dpaa/rte_bus_dpaa_version.map
@@ -30,6 +30,7 @@ DPDK_17.11 {
fman_if_set_maxfrm;
fman_if_set_mcast_filter_table;
fman_if_stats_get;
+   fman_if_stats_get_all;
fman_if_stats_reset;
 netcfg_acquire;
netcfg_release;
-- 
2.7.4

[dpdk-dev] [PATCH v2 27/30] net/dpaa2: log that VLAN extend offload not supported

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_ethdev.c | 6 ++
 1 file changed, 6 insertions(+)

diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index a274e12..589ae6b 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -182,6 +182,12 @@ dpaa2_vlan_offload_set(struct rte_eth_dev *dev, int mask)
RTE_LOG(ERR, PMD, "Unable to set vlan filter = %d\n",
ret);
}
+
+   if (mask & ETH_VLAN_EXTEND_MASK) {
+   if (dev->data->dev_conf.rxmode.hw_vlan_extend)
+   RTE_LOG(INFO, PMD,
+   "VLAN extend offload not supported\n");
+   }
 }
 
 static int
-- 
2.7.4

[dpdk-dev] [PATCH v2 26/30] net/dpaa2: fix the Tx handling of non HW pool bufs

[Hemant Agrawal]

The current code is sending 8 packet in each internal loop.
In some of the conditions, mbuf is being allocated or freed.
In case of error, the code is returning without taking care of
such buffer. It is better to send already prepared buffer and err
for the current failure only.

Fixes: 9e5f3e6d3658 ("net/dpaa2: handle non-hardware backed buffer pool")
Cc: sta...@dpdk.org

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_rxtx.c | 30 +-
 1 file changed, 21 insertions(+), 9 deletions(-)

diff --git a/drivers/net/dpaa2/dpaa2_rxtx.c b/drivers/net/dpaa2/dpaa2_rxtx.c
index 75a06f5..8611d2e 100644
--- a/drivers/net/dpaa2/dpaa2_rxtx.c
+++ b/drivers/net/dpaa2/dpaa2_rxtx.c
@@ -350,7 +350,6 @@ eth_copy_mbuf_to_fd(struct rte_mbuf *mbuf,
if (rte_dpaa2_mbuf_alloc_bulk(
rte_dpaa2_bpid_info[bpid].bp_list->mp, &mb, 1)) {
PMD_TX_LOG(WARNING, "Unable to allocated DPAA2 buffer");
-   rte_pktmbuf_free(mbuf);
return -1;
}
m = (struct rte_mbuf *)mb;
@@ -382,8 +381,6 @@ eth_copy_mbuf_to_fd(struct rte_mbuf *mbuf,
rte_dpaa2_bpid_info[DPAA2_GET_FD_BPID(fd)].meta_data_size,
DPAA2_GET_FD_OFFSET(fd),
DPAA2_GET_FD_LEN(fd));
-   /*free the original packet */
-   rte_pktmbuf_free(mbuf);
 
return 0;
 }
@@ -582,7 +579,7 @@ dpaa2_dev_tx(void *queue, struct rte_mbuf **bufs, uint16_t 
nb_pkts)
/* Not a hw_pkt pool allocated frame */
if (!mp) {
PMD_TX_LOG(ERR, "err: no bpool attached");
-   goto skip_tx;
+   goto send_n_return;
}
if (mp->ops_index != priv->bp_list->dpaa2_ops_index) {
PMD_TX_LOG(ERR, "non hw offload bufffer ");
@@ -595,24 +592,25 @@ dpaa2_dev_tx(void *queue, struct rte_mbuf **bufs, 
uint16_t nb_pkts)
PMD_TX_LOG(ERR,
   "err: no bpool attached");
num_tx = 0;
-   goto skip_tx;
+   goto send_n_return;
}
if (unlikely((*bufs)->nb_segs > 1)) {
PMD_TX_LOG(ERR, "S/G support not added"
" for non hw offload buffer");
-   goto skip_tx;
+   goto send_n_return;
}
if (eth_copy_mbuf_to_fd(*bufs,
&fd_arr[loop], bpid)) {
-   bufs++;
-   continue;
+   goto send_n_return;
}
+   /* free the original packet */
+   rte_pktmbuf_free(*bufs);
} else {
bpid = mempool_to_bpid(mp);
if (unlikely((*bufs)->nb_segs > 1)) {
if (eth_mbuf_to_sg_fd(*bufs,
&fd_arr[loop], bpid))
-   goto skip_tx;
+   goto send_n_return;
} else {
eth_mbuf_to_fd(*bufs,
   &fd_arr[loop], bpid);
@@ -630,6 +628,20 @@ dpaa2_dev_tx(void *queue, struct rte_mbuf **bufs, uint16_t 
nb_pkts)
dpaa2_q->tx_pkts += frames_to_send;
nb_pkts -= frames_to_send;
}
+   return num_tx;
+
+send_n_return:
+   /* send any already prepared fd */
+   if (loop) {
+   unsigned int i = 0;
+
+   while (i < loop) {
+   i += qbman_swp_enqueue_multiple(swp, &eqdesc,
+   &fd_arr[i], loop - i);
+   }
+   num_tx += loop;
+   dpaa2_q->tx_pkts += loop;
+   }
 skip_tx:
return num_tx;
 }
-- 
2.7.4

[dpdk-dev] [PATCH v2 30/30] bus/dpaa2: improve debug log messages

[Hemant Agrawal]

enable the printing of objects during debuging.
use RTE_LOG to avoid function name printing for object name.

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c | 2 +-
 drivers/bus/fslmc/portal/dpaa2_hw_dpci.c | 2 +-
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c | 6 ++
 3 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
index 7e96115..334e1f5 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
@@ -98,7 +98,7 @@ dpaa2_create_dpbp_device(int vdev_fd __rte_unused,
 
TAILQ_INSERT_TAIL(&dpbp_dev_list, dpbp_node, next);
 
-   PMD_INIT_LOG(DEBUG, "DPAA2: Added [dpbp.%d]", dpbp_id);
+   RTE_LOG(DEBUG, PMD, "DPAA2: Added [dpbp.%d]\n", dpbp_id);
 
return 0;
 }
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpci.c 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpci.c
index b60a745..ae189c7 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpci.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpci.c
@@ -140,7 +140,7 @@ rte_dpaa2_create_dpci_device(int vdev_fd __rte_unused,
 
TAILQ_INSERT_TAIL(&dpci_dev_list, dpci_node, next);
 
-   PMD_INIT_LOG(DEBUG, "DPAA2: Added [dpci.%d]", dpci_id);
+   RTE_LOG(DEBUG, PMD, "DPAA2: Added [dpci.%d]\n", dpci_id);
 
return 0;
 }
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index ff41ce4..f00070f 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -210,7 +210,7 @@ configure_dpio_qbman_swp(struct dpaa2_dpio_dev *dpio_dev)
return -1;
}
 
-   PMD_DRV_LOG(DEBUG, "\t Allocated  DPIO Portal[%p]", dpio_dev->dpio);
+   PMD_DRV_LOG(DEBUG, "Allocated  DPIO Portal[%p]", dpio_dev->dpio);
dpio_dev->dpio->regs = dpio_dev->mc_portal;
if (dpio_open(dpio_dev->dpio, CMD_PRI_LOW, dpio_dev->hw_id,
  &dpio_dev->token)) {
@@ -242,8 +242,6 @@ configure_dpio_qbman_swp(struct dpaa2_dpio_dev *dpio_dev)
return -1;
}
 
-   PMD_INIT_LOG(DEBUG, "Qbman Portal ID %d", attr.qbman_portal_id);
-
/* Configure & setup SW portal */
p_des.block = NULL;
p_des.idx = attr.qbman_portal_id;
@@ -502,7 +500,7 @@ dpaa2_create_dpio_device(int vdev_fd,
}
 
TAILQ_INSERT_TAIL(&dpio_dev_list, dpio_dev, next);
-   PMD_INIT_LOG(DEBUG, "DPAA2: Added [dpio.%d]", object_id);
+   RTE_LOG(DEBUG, PMD, "DPAA2: Added [dpio.%d]\n", object_id);
 
return 0;
 }
-- 
2.7.4

[dpdk-dev] [PATCH v2 29/30] net/dpaa2: improve debug messaging

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_ethdev.c | 14 --
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 4e833db..333e4e6 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -426,8 +426,8 @@ dpaa2_dev_rx_queue_setup(struct rte_eth_dev *dev,
 
PMD_INIT_FUNC_TRACE();
 
-   PMD_INIT_LOG(DEBUG, "dev =%p, queue =%d, pool = %p, conf =%p",
-dev, rx_queue_id, mb_pool, rx_conf);
+   PMD_DRV_LOG(DEBUG, "dev =%p, queue =%d, pool = %p, conf =%p",
+   dev, rx_queue_id, mb_pool, rx_conf);
 
if (!priv->bp_list || priv->bp_list->mp != mb_pool) {
bpid = mempool_to_bpid(mb_pool);
@@ -476,8 +476,8 @@ dpaa2_dev_rx_queue_setup(struct rte_eth_dev *dev,
taildrop.threshold = CONG_THRESHOLD_RX_Q;
taildrop.units = DPNI_CONGESTION_UNIT_BYTES;
taildrop.oal = CONG_RX_OAL;
-   PMD_INIT_LOG(DEBUG, "Enabling Early Drop on queue = %d",
-rx_queue_id);
+   PMD_DRV_LOG(DEBUG, "Enabling Early Drop on queue = %d",
+   rx_queue_id);
ret = dpni_set_taildrop(dpni, CMD_PRI_LOW, priv->token,
DPNI_CP_QUEUE, DPNI_QUEUE_RX,
dpaa2_q->tc_index, flow_id, &taildrop);
@@ -983,7 +983,7 @@ dpaa2_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
PMD_DRV_LOG(ERR, "setting the max frame length failed");
return -1;
}
-   PMD_DRV_LOG(INFO, "MTU is configured %d for the device\n", mtu);
+   PMD_DRV_LOG(INFO, "MTU is configured %d for the device", mtu);
return 0;
 }
 
@@ -1738,7 +1738,7 @@ dpaa2_dev_init(struct rte_eth_dev *eth_dev)
priv->nb_tx_queues = attr.num_tx_tcs;
 
PMD_DRV_LOG(DEBUG, "RX-TC= %d, nb_rx_queues= %d, nb_tx_queues=%d",
-   priv->num_tc, priv->nb_rx_queues, priv->nb_tx_queues);
+   priv->num_rx_tc, priv->nb_rx_queues, priv->nb_tx_queues);
 
priv->hw = dpni_dev;
priv->hw_id = hw_id;
@@ -1805,6 +1805,7 @@ dpaa2_dev_init(struct rte_eth_dev *eth_dev)
eth_dev->tx_pkt_burst = dpaa2_dev_tx;
rte_fslmc_vfio_dmamap();
 
+   RTE_LOG(INFO, PMD, "%s: netdev created\n", eth_dev->data->name);
return 0;
 init_err:
dpaa2_dev_uninit(eth_dev);
@@ -1865,6 +1866,7 @@ dpaa2_dev_uninit(struct rte_eth_dev *eth_dev)
eth_dev->rx_pkt_burst = NULL;
eth_dev->tx_pkt_burst = NULL;
 
+   RTE_LOG(INFO, PMD, "%s: netdev created\n", eth_dev->data->name);
return 0;
 }
 
-- 
2.7.4

[dpdk-dev] [PATCH v2 28/30] net/dpaa2: checksum support as per user config

[Hemant Agrawal]

Instead of enabling the RX checksum by default, make it
enable only with user ethernet configuration

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_ethdev.c | 65 +---
 1 file changed, 35 insertions(+), 30 deletions(-)

diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 589ae6b..4e833db 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -334,8 +334,10 @@ dpaa2_alloc_rx_tx_queues(struct rte_eth_dev *dev)
 static int
 dpaa2_eth_dev_configure(struct rte_eth_dev *dev)
 {
-   struct rte_eth_dev_data *data = dev->data;
-   struct rte_eth_conf *eth_conf = &data->dev_conf;
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+   struct fsl_mc_io *dpni = priv->hw;
+   struct rte_eth_conf *eth_conf = &dev->data->dev_conf;
+   int rx_ip_csum_offload = false;
int ret;
 
PMD_INIT_FUNC_TRACE();
@@ -364,6 +366,37 @@ dpaa2_eth_dev_configure(struct rte_eth_dev *dev)
}
}
 
+   if (eth_conf->rxmode.hw_ip_checksum)
+   rx_ip_csum_offload = true;
+
+   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
+  DPNI_OFF_RX_L3_CSUM, rx_ip_csum_offload);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "Error to set RX l3 csum:Error = %d\n", ret);
+   return ret;
+   }
+
+   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
+  DPNI_OFF_RX_L4_CSUM, rx_ip_csum_offload);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "Error to get RX l4 csum:Error = %d\n", ret);
+   return ret;
+   }
+
+   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
+  DPNI_OFF_TX_L3_CSUM, true);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "Error to set TX l3 csum:Error = %d\n", ret);
+   return ret;
+   }
+
+   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
+  DPNI_OFF_TX_L4_CSUM, true);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "Error to get TX l4 csum:Error = %d\n", ret);
+   return ret;
+   }
+
/* update the current status */
dpaa2_dev_link_update(dev, 0);
 
@@ -707,34 +740,6 @@ dpaa2_dev_start(struct rte_eth_dev *dev)
dpaa2_q->fqid = qid.fqid;
}
 
-   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
-  DPNI_OFF_RX_L3_CSUM, true);
-   if (ret) {
-   PMD_INIT_LOG(ERR, "Error to set RX l3 csum:Error = %d\n", ret);
-   return ret;
-   }
-
-   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
-  DPNI_OFF_RX_L4_CSUM, true);
-   if (ret) {
-   PMD_INIT_LOG(ERR, "Error to get RX l4 csum:Error = %d\n", ret);
-   return ret;
-   }
-
-   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
-  DPNI_OFF_TX_L3_CSUM, true);
-   if (ret) {
-   PMD_INIT_LOG(ERR, "Error to set TX l3 csum:Error = %d\n", ret);
-   return ret;
-   }
-
-   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
-  DPNI_OFF_TX_L4_CSUM, true);
-   if (ret) {
-   PMD_INIT_LOG(ERR, "Error to get TX l4 csum:Error = %d\n", ret);
-   return ret;
-   }
-
/*checksum errors, send them to normal path and set it in annotation */
err_cfg.errors = DPNI_ERROR_L3CE | DPNI_ERROR_L4CE;
 
-- 
2.7.4

Re: [dpdk-dev] [dpdk-dev, 01/17] build: add initial infrastructure for meson & ninja builds

[Bruce Richardson]

On Thu, Sep 07, 2017 at 12:21:57PM -0400, Neil Horman wrote:
> On Fri, Sep 01, 2017 at 11:04:00AM +0100, Bruce Richardson wrote:
> > To build with meson and ninja, we need some initial infrastructure in
> > place. The build files for meson always need to be called "meson.build",
> > and options get placed in meson_options.txt
> > 
> > This commit adds a top-level meson.build file, which sets up the global
> > variables for tracking drivers, libraries, etc., and then includes other
> > build files, before finishing by writing the global build configuration
> > header file and a DPDK pkgconfig file at the end, using some of those same
> > globals.
> > 
> > >From the top level build file, the only include file thus far is for the
> > config folder, which does some other setup of global configuration
> > parameters, including pulling in architecture specific parameters from an
> > architectural subdirectory. A number of configuration build options are
> > provided for the project to tune a number of global variables which will be
> > used later e.g. max numa nodes, max cores, etc. These settings all make
> > their way to the global build config header "rte_build_config.h". There is
> > also a file "rte_config.h", which includes "rte_build_config.h", and this
> > file is meant to hold other build-time values which are present in our
> > current static build configuration but are not normally meant for
> > user-configuration. Ideally, over time, the values placed here should be
> > moved to the individual libraries or drivers which want those values.
> > 
> > Signed-off-by: Bruce Richardson 
> > Reviewed-by: Harry van Haaren 
> 
> I feel like I need to underscore my previous concern here.  While I'm not
> opposed per-se to a new build system, I am very concerned about the burden 
> that
> switching places on downstream consumers, in particular distributions (since I
> represent one of them).  Moving to a new build system with new tools means 
> those
> tools need to be packaged, tested and shipped, which is a significant work
> effort.  While it might be a net gain long term, its something you need to 
> keep
> in mind when making these changes.
> 
Understood. If there is anything we/I can do to make this transition
easier, please flag it for consideration.

> I know you've said that we will be keepting the existing build system,
> I just need to be sure everyone understands just how important that
> is.
> 
What is your feeling here, in terms of timescale. After any new system
reaches feature parity, how long would you estimate that we would need
to support the existing makefile system before it would be safe to
deprecate it? Should we start a deprecation plan, or is it best just to
commit to support both until we get all - or almost all - downstream
consumers switched over? While I wouldn't push for deprecating the old
system any time soon, and I wouldn't consider maintaining the two
unduly burdensome, it's not something we want to do in the long term.

> Though perhaps the time frame for keeping the current build system as priarmy 
> is
> less concerning, as feature parity is even more critical.  That is to say, the
> new build system must be able to produce the same configurations that the
> current build system does.  Without it I don't think anyone will be able to 
> use
> it consistently, and that will leave a great number of users in a very poor
> position.  I think getting a little closer to parity with the current system 
> is
> warranted.  I'd suggest as a gating factor:
> 
> 1) Building on all supported arches
> 2) Cross building on all supported arches
> 3) Proper identification of targeted machine (i.e. equivalent of the machine
> component of the current build system)
> 
The question there is gating factor for what? Presumably not for merging
into the staging tree. But for merging into the main tree for releases?
I'd push back a little on that, as the new system does not interfere in
any way with the old, and by keeping it in a staging tree until it
reaches full feature parity will make the job considerably harder. For
example, it means that anyone submitting a new driver or library has to
submit the code and makefiles in one set and the meson patches in a
separate one for a separate build tree. It also makes it less likely
that people will try out the new system and find the issues with it, and
help fill in the gaps. While I can understand us not recommending the
new build system until it reaches feature parity, I think there are a
lot of benefits to be got by making it widely available, even if it's
incomplete. 

On a semi-related note, some folks here are chomping at the bit for it
to get mainlined, as they want the improved build time for recompiles
for speed up their development process. They can work off the posted
patches, but it's more painful than having it locked in.

> Specific notes inline
> 
> > ---
> >  config/meson.build | 69 +
> >  config/

Re: [dpdk-dev] [dpdk-dev, 02/17] eal: add eal library to meson build

[Bruce Richardson]

On Thu, Sep 07, 2017 at 12:25:29PM -0400, Neil Horman wrote:
> On Fri, Sep 01, 2017 at 11:04:01AM +0100, Bruce Richardson wrote:
> > Support building the EAL with meson and ninja. This involves a number of
> > different meson.build files for iterating through all the different
> > subdirectories in the EAL. The library itself will be compiled on build but
> > the header files are only copied from their initial location once "ninja
> > install" is run. Instead, we use meson dependency tracking to ensure that
> > other libraries which use the EAL headers can find them in their original
> > locations.
> > 
> > Note: this does not include building kernel modules on either BSD or Linux
> > 
> > Signed-off-by: Bruce Richardson 
> > Reviewed-by: Harry van Haaren 
> > ---
> >  config/rte_config.h| 11 
> >  lib/librte_eal/bsdapp/eal/meson.build  | 67 
> > 
> >  lib/librte_eal/bsdapp/meson.build  | 32 ++
> >  lib/librte_eal/common/arch/meson.build | 33 ++
> >  lib/librte_eal/common/arch/x86/meson.build | 32 ++
> >  lib/librte_eal/common/include/arch/meson.build | 33 ++
> >  lib/librte_eal/common/include/arch/x86/meson.build | 48 +++
> >  lib/librte_eal/common/include/meson.build  | 71 
> > +
> >  lib/librte_eal/common/meson.build  | 71 
> > +
> >  lib/librte_eal/linuxapp/eal/meson.build| 72 
> > ++
> >  lib/librte_eal/linuxapp/meson.build| 32 ++
> >  lib/librte_eal/meson.build | 44 +
> >  lib/meson.build| 32 ++
> >  meson.build|  3 +-
> >  14 files changed, 580 insertions(+), 1 deletion(-)
> >  create mode 100644 lib/librte_eal/bsdapp/eal/meson.build
> >  create mode 100644 lib/librte_eal/bsdapp/meson.build
> >  create mode 100644 lib/librte_eal/common/arch/meson.build
> >  create mode 100644 lib/librte_eal/common/arch/x86/meson.build
> >  create mode 100644 lib/librte_eal/common/include/arch/meson.build
> >  create mode 100644 lib/librte_eal/common/include/arch/x86/meson.build
> >  create mode 100644 lib/librte_eal/common/include/meson.build
> >  create mode 100644 lib/librte_eal/common/meson.build
> >  create mode 100644 lib/librte_eal/linuxapp/eal/meson.build
> >  create mode 100644 lib/librte_eal/linuxapp/meson.build
> >  create mode 100644 lib/librte_eal/meson.build
> >  create mode 100644 lib/meson.build
> > 
> > diff --git a/config/rte_config.h b/config/rte_config.h
> > index 79b0db90f..252b087ad 100644
> > --- a/config/rte_config.h
> > +++ b/config/rte_config.h
> > @@ -47,4 +47,15 @@
> >  
> >  #include 
> >  
> > +/** library defines /
> > +
> > +/* EAL defines */
> > +#define RTE_MAX_MEMSEG 512
> > +#define RTE_MAX_MEMZONE 2560
> > +#define RTE_MAX_TAILQ 32
> > +#define RTE_LOG_LEVEL RTE_LOG_INFO
> > +#define RTE_LOG_DP_LEVEL RTE_LOG_INFO
> > +#define RTE_BACKTRACE 1
> > +#define RTE_EAL_VFIO 1
> > +
> >  #endif /* _RTE_CONFIG_H_ */
> > diff --git a/lib/librte_eal/bsdapp/eal/meson.build 
> > b/lib/librte_eal/bsdapp/eal/meson.build
> > new file mode 100644
> > index 0..bc9c3f0cb
> > --- /dev/null
> > +++ b/lib/librte_eal/bsdapp/eal/meson.build
> > @@ -0,0 +1,67 @@
> > +#   BSD LICENSE
> > +#
> > +#   Copyright(c) 2017 Intel Corporation. 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
> >

Re: [dpdk-dev] [PATCH 07/21] vhost: add iotlb helper functions

[Maxime Coquelin]

On 09/08/2017 10:36 AM, Yuanhan Liu wrote:

On Fri, Sep 08, 2017 at 10:24:58AM +0200, Maxime Coquelin wrote:



On 09/08/2017 10:08 AM, Yuanhan Liu wrote:

On Thu, Aug 31, 2017 at 11:50:09AM +0200, Maxime Coquelin wrote:

diff --git a/lib/librte_vhost/iotlb.c b/lib/librte_vhost/iotlb.c
new file mode 100644
index 0..1b739dae5
--- /dev/null
+++ b/lib/librte_vhost/iotlb.c
@@ -0,0 +1,231 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright (c) 2017 Red Hat, Inc.
+ *   Copyright (c) 2017 Maxime Coquelin 


I'm not a lawer, but I have been told many years before, that you don't
have the copyright for the code you write for open source project, the
company you work for does.

Thus, it's more common to see something like following:
Copyright , ... the commany ...
Author:  Some One <...@...>

However, as you may have noticed, it's not common to put the authorship
in the source files. Though I don't object it.


I'm not a lawyer too. At least in other projects, it seems common the
author puts his name as copyright owner.

I have no issue to change it to only keep Red Hat's one though.


That's up to you. What I said before was JFYI :)


[...]

+#define IOTLB_CACHE_SIZE 1024
+
+static void vhost_user_iotlb_cache_remove_all(struct vhost_virtqueue *vq)


Note that it's not the DPDK coding style to define a function.


Ok, I guess you mean:
static void
vhost_user_iotlb_cache_remove_all(struct vhost_virtqueue *vq) ?


Yep.


+{
+   struct vhost_iotlb_entry *node, *temp_node;
+
+   rte_rwlock_write_lock(&vq->iotlb_lock);
+
+   TAILQ_FOREACH_SAFE(node, &vq->iotlb_list, next, temp_node) {
+   TAILQ_REMOVE(&vq->iotlb_list, node, next);
+   rte_mempool_put(vq->iotlb_pool, node);
+   }
+
+   rte_rwlock_write_unlock(&vq->iotlb_lock);
+}
+
+void vhost_user_iotlb_cache_insert(struct vhost_virtqueue *vq, uint64_t iova,
+   uint64_t uaddr, uint64_t size, uint8_t perm)
+{
+   struct vhost_iotlb_entry *node, *new_node;
+   int ret;
+
+   ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
+   if (ret) {
+   RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool empty, invalidate 
cache\n");


It's a cache, why not considering remove one to get space for new one?


It would mean having to track every lookups not to remove hot entries,
which would have an impact on performance.


You were removing all caches, how can we do worse than that? Even a
random evict would be better. Or, more simply, just to remove the
head or the tail?


I think removing head or tail could cause deadlocks.
For example it needs to translate from 0x0 to 0x2000, with page size
being 0x1000.

If cache is full, when inserting 0x1000-0x1fff, it will remove
0x0-0xfff, so a miss will be sent for 0x0-0x and 0x1000-0x1fff will
be remove at insert time, etc...


Note that we really need to size the cache large enough for performance
purpose, so having the cache full could be cause by either buggy or
malicious guest.



--yliu


Moreover, the idea is to have the cache large enough, else you could
face packet drops due to random cache misses.

We might consider to improve it, but I consider it an optimization that
could be implemented later if needed.


+   vhost_user_iotlb_cache_remove_all(vq);
+   ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
+   if (ret) {
+   RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool still empty, 
failure\n");
+   return;
+   }
+   }

Re: [dpdk-dev] [dpdk-dev, 16/17] build: add option to version libs using DPDK version

[Bruce Richardson]

On Thu, Sep 07, 2017 at 01:07:19PM -0400, Neil Horman wrote:
> On Fri, Sep 01, 2017 at 11:04:15AM +0100, Bruce Richardson wrote:
> > Normally, each library has it's own version number based on the ABI.
> > Add an option to have all libs just use the DPDK version number as
> > the .so version.
> > 
> > Signed-off-by: Bruce Richardson  ---
> > drivers/meson.build | 8 +++- lib/meson.build | 8 +++-
> > meson_options.txt   | 1 + 3 files changed, 15 insertions(+), 2
> > deletions(-)
> > 
> > diff --git a/drivers/meson.build b/drivers/meson.build index
> > d7a614f83..76d610a5b 100644 --- a/drivers/meson.build +++
> > b/drivers/meson.build @@ -92,6 +92,12 @@ foreach
> > class:driver_classes depends: [pmdinfogen, tmp_lib]) endforeach
> >  
> > +   if get_option('per_library_versions') +
> > so_version = '@0@.1'.format(version) +  else +
> > so_version = meson.project_version() +  endif +
> 
> I'm not sure this is a good idea.  If we default to using the defined
> project version number, we commit to, by default incrementing the
> library version number on each release, even if the ABI hasn't
> changed, which means that the purpose of versioning (creating
> compatibility between library releases), is rendered useless.
>
Yes, I agree on the compatibility, so the default for this option is to
have individual .so versions, as right now. This option itself was added
at the request of some distros [1][2], who wanted to just use a single
version number across all DPDK.

/Bruce

[1] http://dpdk.org/ml/archives/dev/2017-August/072969.html
[2] http://dpdk.org/ml/archives/dev/2017-August/072976.html

Re: [dpdk-dev] [PATCH] cmdline: fix warning for unused return value

[Bruce Richardson]

On Thu, Sep 07, 2017 at 09:46:18PM +0200, Olivier MATZ wrote:
> On Thu, Sep 07, 2017 at 10:50:20AM -0700, Stephen Hemminger wrote:
> > On Thu,  7 Sep 2017 14:09:23 +0100
> > Bruce Richardson  wrote:
> > 
> > > When DPDK is compiled on Ubuntu with extra warnings turned on, there is a
> > > warning about the return value from write() being unchecked. Rather than
> > > having builds disable the warning, which may mask other cases we do care
> > > about, we can add a dummy use of the return value in the code to silence 
> > > it
> > > in this instance.
> > > 
> > > Signed-off-by: Bruce Richardson 
> > > ---
> > >  lib/librte_cmdline/cmdline.c | 3 ++-
> > >  1 file changed, 2 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/lib/librte_cmdline/cmdline.c b/lib/librte_cmdline/cmdline.c
> > > index a9c47be..d749165 100644
> > > --- a/lib/librte_cmdline/cmdline.c
> > > +++ b/lib/librte_cmdline/cmdline.c
> > > @@ -205,7 +205,8 @@ cmdline_printf(const struct cmdline *cl, const char 
> > > *fmt, ...)
> > >   }
> > >   if (ret >= BUFSIZ)
> > >   ret = BUFSIZ - 1;
> > > - write(cl->s_out, buf, ret);
> > > + ret = write(cl->s_out, buf, ret);
> > > + (void)ret;
> > 
> > That is an ugly way to fix the warning.
> > If write fails, the user has probably logged out.
> 
> Here, we are writing on stdout.
> I don't think it's worst than calling printf() without checking the
> return value.
> 
> By the way, it looks strange to this code is compiled, because
> -D _GNU_SOURCE is passed to the compiler, and the code is:
> 
>   #ifdef _GNU_SOURCE
>   ...
>   #else
>   ...
>   write()
>   #endif
> 
> It does not mean we shouldn't fix it, but maybe it hides something
> else. Bruce, I bet you are testing your new build framework?
> 
> Olivier

Yep, and meson has the -Wunused-result flag on by default, and on Ubuntu
write is marked as needing the result check. I don't like disabling
warnings, so I'd rather fix it this way, for the sake of on extra line
of ugliness. Open to better options though (other than not using Ubuntu,
which is fine with me, but some others seem to object to that :-) )

Re: [dpdk-dev] [PATCH] eal: cleanup strerror function

[Bruce Richardson]

On Thu, Sep 07, 2017 at 10:51:52AM -0700, Stephen Hemminger wrote:
> On Thu,  7 Sep 2017 14:09:56 +0100
> Bruce Richardson  wrote:
> 
> > +   /* BSD puts a colon in the "unknown error" messages, Linux doesn't */
> > +#ifdef RTE_EXEC_ENV_BSDAPP
> > +   static const char *sep = ":";
> > +#else
> > +   static const char *sep = "";
> > +#endif
> 
> This is seems unnecessary to me just have the message be different.

I'd rather keep it consistent for the sake of the couple of lines of
code. What would be better is if there was a C variable we could use for
this to avoid macros, but I not aware of one right now that could work.
Again, open to suggestions on that. Even with keeping the macros, this
is still cleaner than the previous version.

/Bruce

Re: [dpdk-dev] [PATCH] mk: add silvermont to replace atom as a target

[Bruce Richardson]

On Fri, Sep 08, 2017 at 11:28:52AM +0800, Xiaoyun Li wrote:
> The -march=atom flag is for older atom CPUs and don't support SSE4 which
> is the minimum reqiurement for DPDK. And in fact, the current atom CPUs
> support SSE4. So this patch removes atom as a target for DPDK builds and
> adds a silvermont replacement instead.
> 
> Signed-off-by: Xiaoyun Li 
> ---
Ideally this should be posted as a v2 replacing the previous patch
solution you proposed so that it's threaded with that email discussion.
However, that doesn't affect the patch itself which looks reasonable.

Acked-by: Bruce Richardson 

Re: [dpdk-dev] [PATCH v1 0/3] net/mlx4: additional interrupt handling fixes

[Ferruh Yigit]

On 9/5/2017 1:56 PM, Adrien Mazarguil wrote:
> While the previous interrupt rework improved the situation, it failed to
> address one last issue pointed out by Matan: RMV/LSC events may be missed
> and not reported to the application.
> 
> Adrien Mazarguil (3):
>   net/mlx4: fix unhandled event debug message
>   net/mlx4: fix rescheduled link status check
>   net/mlx4: merge interrupt collector function

Series applied to dpdk-next-net/master, thanks.

[dpdk-dev] [RFC] ethdev: add administrative information in per-port info

[Ivan Boule]

To help administrative tasks on DPDK ports, add in the data structure
rte_eth_dev_info the following per-port information to be supplied
by the dev_infos_get() function exported by a Poll Mode Driver:

- the set of supported link modes,
- the set of advertised link modes,
- the type of port connector,
- autonegotiation enabled or not.

Set new administrative fields to a default value in rte_eth_dev_info_get()
before invoking the dev_infos_get() function exported by a PMD, if any.

Once this API change is accepted, the dev_infos_get() function of PMDs
will be updated accordingly to set these new fields, along with the
port_infos_display() function of the testpmd to display them.

Signed-off-by: Ivan Boule 
---
 lib/librte_ether/rte_ethdev.c |   1 +
 lib/librte_ether/rte_ethdev.h | 112 ++
 2 files changed, 113 insertions(+)

diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
index 0597641..4ca51e1 100644
--- a/lib/librte_ether/rte_ethdev.c
+++ b/lib/librte_ether/rte_ethdev.c
@@ -1897,6 +1897,7 @@ rte_eth_dev_info_get(uint8_t port_id, struct 
rte_eth_dev_info *dev_info)
memset(dev_info, 0, sizeof(struct rte_eth_dev_info));
dev_info->rx_desc_lim = lim;
dev_info->tx_desc_lim = lim;
+   dev_info->connector = RTE_ETH_CONNECTOR_OTHER;
 
RTE_FUNC_PTR_OR_RET(*dev->dev_ops->dev_infos_get);
(*dev->dev_ops->dev_infos_get)(dev, dev_info);
diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
index 0adf327..ac49380 100644
--- a/lib/librte_ether/rte_ethdev.h
+++ b/lib/librte_ether/rte_ethdev.h
@@ -935,6 +935,113 @@ struct rte_pci_device;
 /**
  * Ethernet device information
  */
+
+/* Types of port connector. */
+#define RTE_ETH_CONNECTOR_TP0x00 /**< Twisted Pair */
+#define RTE_ETH_CONNECTOR_AUI   0x01 /**< Attachment Unit Interface */
+#define RTE_ETH_CONNECTOR_MII   0x02 /**< Media-Independent Interface */
+#define RTE_ETH_CONNECTOR_FIBRE 0x03 /**< Fiber */
+#define RTE_ETH_CONNECTOR_DA0x05 /**< Direct Attach */
+#define RTE_ETH_CONNECTOR_NONE  0xef
+#define RTE_ETH_CONNECTOR_OTHER 0xff
+
+/* Link modes. */
+#define RTE_ETH_LINK_MODE_10baseT_Half_BIT0
+#define RTE_ETH_LINK_MODE_10baseT_Full_BIT1
+#define RTE_ETH_LINK_MODE_100baseT_Half_BIT   2
+#define RTE_ETH_LINK_MODE_100baseT_Full_BIT   3
+#define RTE_ETH_LINK_MODE_1000baseT_Half_BIT  4
+#define RTE_ETH_LINK_MODE_1000baseT_Full_BIT  5
+#define RTE_ETH_LINK_MODE_Autoneg_BIT 6
+#define RTE_ETH_LINK_MODE_TP_BIT  7
+#define RTE_ETH_LINK_MODE_AUI_BIT 8
+#define RTE_ETH_LINK_MODE_MII_BIT 9
+#define RTE_ETH_LINK_MODE_FIBRE_BIT  10
+#define RTE_ETH_LINK_MODE_BNC_BIT11
+#define RTE_ETH_LINK_MODE_1baseT_Full_BIT12
+#define RTE_ETH_LINK_MODE_Pause_BIT  13
+#define RTE_ETH_LINK_MODE_Asym_Pause_BIT 14
+#define RTE_ETH_LINK_MODE_2500baseX_Full_BIT 15
+#define RTE_ETH_LINK_MODE_Backplane_BIT  16
+#define RTE_ETH_LINK_MODE_1000baseKX_Full_BIT17
+#define RTE_ETH_LINK_MODE_1baseKX4_Full_BIT  18
+#define RTE_ETH_LINK_MODE_1baseKR_Full_BIT   19
+#define RTE_ETH_LINK_MODE_1baseR_FEC_BIT 20
+#define RTE_ETH_LINK_MODE_2baseMLD2_Full_BIT 21
+#define RTE_ETH_LINK_MODE_2baseKR2_Full_BIT  22
+#define RTE_ETH_LINK_MODE_4baseKR4_Full_BIT  23
+#define RTE_ETH_LINK_MODE_4baseCR4_Full_BIT  24
+#define RTE_ETH_LINK_MODE_4baseSR4_Full_BIT  25
+#define RTE_ETH_LINK_MODE_4baseLR4_Full_BIT  26
+#define RTE_ETH_LINK_MODE_56000baseKR4_Full_BIT  27
+#define RTE_ETH_LINK_MODE_56000baseCR4_Full_BIT  28
+#define RTE_ETH_LINK_MODE_56000baseSR4_Full_BIT  29
+#define RTE_ETH_LINK_MODE_56000baseLR4_Full_BIT  30
+#define RTE_ETH_LINK_MODE_25000baseCR_Full_BIT   31
+#define RTE_ETH_LINK_MODE_25000baseKR_Full_BIT   32
+#define RTE_ETH_LINK_MODE_25000baseSR_Full_BIT   33
+#define RTE_ETH_LINK_MODE_5baseCR2_Full_BIT  34
+#define RTE_ETH_LINK_MODE_5baseKR2_Full_BIT  35
+#define RTE_ETH_LINK_MODE_10baseKR4_Full_BIT 36
+#define RTE_ETH_LINK_MODE_10baseSR4_Full_BIT 37
+#define RTE_ETH_LINK_MODE_10baseCR4_Full_BIT 38
+#define RTE_ETH_LINK_MODE_10baseLR4_ER4_Full_BIT 39
+#define RTE_ETH_LINK_MODE_5baseSR2_Full_BIT  40
+#define RTE_ETH_LINK_MODE_1000baseX_Full_BIT 41
+#define RTE_ETH_LINK_MODE_1baseCR_Full_BIT   42
+#define RTE_ETH_LINK_MODE_1baseSR_Full_BIT   43
+#define RTE_ETH_LINK_MODE_1baseLR_Full_BIT   44
+#define RTE_ETH_LINK_MODE_1baseLRM_Full_BIT  45
+#define RTE_ETH_LINK_MODE_1baseER_Full_BIT   46
+#define RTE_ETH_LINK_MODE_2500baseT_Full_BIT 47
+#define RTE_ETH_LINK_MODE_5000baseT_Full_BIT 48
+
+#define 

Re: [dpdk-dev] [PATCH 07/21] vhost: add iotlb helper functions

[Yuanhan Liu]

On Fri, Sep 08, 2017 at 10:50:49AM +0200, Maxime Coquelin wrote:
> +{
> + struct vhost_iotlb_entry *node, *temp_node;
> +
> + rte_rwlock_write_lock(&vq->iotlb_lock);
> +
> + TAILQ_FOREACH_SAFE(node, &vq->iotlb_list, next, temp_node) {
> + TAILQ_REMOVE(&vq->iotlb_list, node, next);
> + rte_mempool_put(vq->iotlb_pool, node);
> + }
> +
> + rte_rwlock_write_unlock(&vq->iotlb_lock);
> +}
> +
> +void vhost_user_iotlb_cache_insert(struct vhost_virtqueue *vq, uint64_t 
> iova,
> + uint64_t uaddr, uint64_t size, uint8_t perm)
> +{
> + struct vhost_iotlb_entry *node, *new_node;
> + int ret;
> +
> + ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
> + if (ret) {
> + RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool empty, invalidate 
> cache\n");
> >>>
> >>>It's a cache, why not considering remove one to get space for new one?
> >>
> >>It would mean having to track every lookups not to remove hot entries,
> >>which would have an impact on performance.
> >
> >You were removing all caches, how can we do worse than that? Even a
> >random evict would be better. Or, more simply, just to remove the
> >head or the tail?
> 
> I think removing head or tail could cause deadlocks.
> For example it needs to translate from 0x0 to 0x2000, with page size
> being 0x1000.
> 
> If cache is full, when inserting 0x1000-0x1fff, it will remove
> 0x0-0xfff, so a miss will be sent for 0x0-0x and 0x1000-0x1fff will
> be remove at insert time, etc...

Okay, that means we can't simply remove the head or tail.

> Note that we really need to size the cache large enough for performance
> purpose, so having the cache full could be cause by either buggy or
> malicious guest.

I agree. But for the malicious guest, it could lead to a DDOS attack:
assume it keeps vhost running out of cache and then vhost keeps printing
above message.

What I suggested was to evict one (by some polices) to get a space for
the new one we want to insert. Note that it won't be a performance issue,
IMO, due to we only do that when we run out of caches, which, according
to your sayings, should not happen in normal cases.

--yliu

> >>Moreover, the idea is to have the cache large enough, else you could
> >>face packet drops due to random cache misses.
> >>
> >>We might consider to improve it, but I consider it an optimization that
> >>could be implemented later if needed.
> >>
> + vhost_user_iotlb_cache_remove_all(vq);
> + ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
> + if (ret) {
> + RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool still empty, 
> failure\n");
> + return;
> + }
> + }

Re: [dpdk-dev] [PATCH 07/21] vhost: add iotlb helper functions

[Maxime Coquelin]

On 09/08/2017 11:21 AM, Yuanhan Liu wrote:

On Fri, Sep 08, 2017 at 10:50:49AM +0200, Maxime Coquelin wrote:

+{
+   struct vhost_iotlb_entry *node, *temp_node;
+
+   rte_rwlock_write_lock(&vq->iotlb_lock);
+
+   TAILQ_FOREACH_SAFE(node, &vq->iotlb_list, next, temp_node) {
+   TAILQ_REMOVE(&vq->iotlb_list, node, next);
+   rte_mempool_put(vq->iotlb_pool, node);
+   }
+
+   rte_rwlock_write_unlock(&vq->iotlb_lock);
+}
+
+void vhost_user_iotlb_cache_insert(struct vhost_virtqueue *vq, uint64_t iova,
+   uint64_t uaddr, uint64_t size, uint8_t perm)
+{
+   struct vhost_iotlb_entry *node, *new_node;
+   int ret;
+
+   ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
+   if (ret) {
+   RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool empty, invalidate 
cache\n");


It's a cache, why not considering remove one to get space for new one?


It would mean having to track every lookups not to remove hot entries,
which would have an impact on performance.


You were removing all caches, how can we do worse than that? Even a
random evict would be better. Or, more simply, just to remove the
head or the tail?


I think removing head or tail could cause deadlocks.
For example it needs to translate from 0x0 to 0x2000, with page size
being 0x1000.

If cache is full, when inserting 0x1000-0x1fff, it will remove
0x0-0xfff, so a miss will be sent for 0x0-0x and 0x1000-0x1fff will
be remove at insert time, etc...


Okay, that means we can't simply remove the head or tail.


Note that we really need to size the cache large enough for performance
purpose, so having the cache full could be cause by either buggy or
malicious guest.


I agree. But for the malicious guest, it could lead to a DDOS attack:
assume it keeps vhost running out of cache and then vhost keeps printing
above message.

What I suggested was to evict one (by some polices) to get a space for
the new one we want to insert. Note that it won't be a performance issue,
IMO, due to we only do that when we run out of caches, which, according
to your sayings, should not happen in normal cases.


Ok, so let's randomly remove one entry. What about using something like:
rte_rand() % IOTLB_CACHE_SIZE



--yliu


Moreover, the idea is to have the cache large enough, else you could
face packet drops due to random cache misses.

We might consider to improve it, but I consider it an optimization that
could be implemented later if needed.


+   vhost_user_iotlb_cache_remove_all(vq);
+   ret = rte_mempool_get(vq->iotlb_pool, (void **)&new_node);
+   if (ret) {
+   RTE_LOG(ERR, VHOST_CONFIG, "IOTLB pool still empty, 
failure\n");
+   return;
+   }
+   }

Re: [dpdk-dev] [PATCH v3 1/4] net/softnic: add softnic PMD

[Singh, Jasvinder]

Hi Ferruh,

Thank you for the review and feedback. Please see inline response;

> -Original Message-
> From: Yigit, Ferruh
> Sent: Tuesday, September 5, 2017 3:53 PM
> To: Singh, Jasvinder ; dev@dpdk.org
> Cc: Dumitrescu, Cristian ;
> tho...@monjalon.net
> Subject: Re: [PATCH v3 1/4] net/softnic: add softnic PMD
> 
> On 8/11/2017 1:49 PM, Jasvinder Singh wrote:
> > Add SoftNIC PMD to provide SW fall-back for ethdev APIs.
> >
> > Signed-off-by: Cristian Dumitrescu 
> > Signed-off-by: Jasvinder Singh 
> > ---
> > v3 changes:
> > - rebase to dpdk17.08 release
> >
> > v2 changes:
> > - fix build errors
> > - rebased to TM APIs v6 plus dpdk master
> >
> >  MAINTAINERS |   5 +
> >  config/common_base  |   5 +
> >  drivers/net/Makefile|   5 +
> >  drivers/net/softnic/Makefile|  56 +++
> >  drivers/net/softnic/rte_eth_softnic.c   | 609
> 
> >  drivers/net/softnic/rte_eth_softnic.h   |  54 +++
> >  drivers/net/softnic/rte_eth_softnic_internals.h | 114 +
> >  drivers/net/softnic/rte_eth_softnic_version.map |   7 +
> >  mk/rte.app.mk   |   5 +-
> 
> Also documentation updates are required:
> - .ini file
> - PMD documentation .rst file
> - I believe it is good to update release note about new PMD
> - release notes library version info, since this has public API

Will send documentation patch.

> <...>
> 
> > +EXPORT_MAP := rte_eth_softnic_version.map
> 
> rte_pmd_... to be consistent.
> 
> <...>

Will do.

> > +#
> > +# Export include files
> > +#
> > +SYMLINK-y-include +=rte_eth_softnic.h
> 
> space after +=
> 

Will add space.
> 
> > diff --git a/drivers/net/softnic/rte_eth_softnic.c
> > b/drivers/net/softnic/rte_eth_softnic.c
> <...>
> > +
> > +static struct rte_vdev_driver pmd_drv;
> 
> Why this is required, already defined below.
> And for naming, pmd=poll mode driver, drv=driver, makes "poll mode driver
> driver"
> 

Ok. will correct this.

> <...>
> 
> > +static int
> > +pmd_rx_queue_setup(struct rte_eth_dev *dev,
> > +   uint16_t rx_queue_id,
> > +   uint16_t nb_rx_desc __rte_unused,
> > +   unsigned int socket_id,
> > +   const struct rte_eth_rxconf *rx_conf __rte_unused,
> > +   struct rte_mempool *mb_pool __rte_unused) {
> > +   struct pmd_internals *p = dev->data->dev_private;
> > +
> > +   if (p->params.soft.intrusive == 0) {
> > +   struct pmd_rx_queue *rxq;
> > +
> > +   rxq = rte_zmalloc_socket(p->params.soft.name,
> > +   sizeof(struct pmd_rx_queue), 0, socket_id);
> > +   if (rxq == NULL)
> > +   return -1;
> 
> return -ENOMEM ?

Ok.
 
> > +
> > +   rxq->hard.port_id = p->hard.port_id;
> > +   rxq->hard.rx_queue_id = rx_queue_id;
> > +   dev->data->rx_queues[rx_queue_id] = rxq;
> > +   } else {
> > +   struct rte_eth_dev *hard_dev =
> > +   &rte_eth_devices[p->hard.port_id];> +
>   void *rxq = hard_dev->data->rx_queues[rx_queue_id];
> > +
> > +   if (rxq == NULL)
> > +   return -1;
> > +
> > +   dev->data->rx_queues[rx_queue_id] = rxq;
> 
> This assigns underlying hw queue as this soft PMD queue, what happens if
> two different cores, one polls the actual hw device and other polls the this
> virtual device, since both are indeed same queues?

Once soft device is created and attached to hard device, application has to 
reads packets from/writes packets to the "soft" port instead of the "hard" port 
as soft device is feature rich
version of the hard device (See Cover letter notes). The RX and TX queues of 
the "soft" port are thread safe, as any ethdev. 
 
> > +   }
> > +   return 0;
> > +}
> > +
> 
> <...>
> 
> > +static __rte_always_inline int
> > +rte_pmd_softnic_run_default(struct rte_eth_dev *dev) {
> > +   struct pmd_internals *p = dev->data->dev_private;
> > +
> > +   /* Persistent context: Read Only (update not required) */
> > +   struct rte_mbuf **pkts = p->soft.def.pkts;
> > +   uint16_t nb_tx_queues = dev->data->nb_tx_queues;
> > +
> > +   /* Persistent context: Read - Write (update required) */
> > +   uint32_t txq_pos = p->soft.def.txq_pos;
> > +   uint32_t pkts_len = p->soft.def.pkts_len;
> > +   uint32_t flush_count = p->soft.def.flush_count;
> > +
> > +   /* Not part of the persistent context */
> > +   uint32_t pos;
> > +   uint16_t i;
> > +
> > +   /* Soft device TXQ read, Hard device TXQ write */
> > +   for (i = 0; i < nb_tx_queues; i++) {
> > +   struct rte_ring *txq = dev->data->tx_queues[txq_pos];
> > +
> > +   /* Read soft device TXQ burst to packet enqueue buffer */
> > +   pkts_len += rte_ring_sc_dequeue_burst(txq,
> > +   (void **) &pkts[pkts_len],
> > +   DEFAULT_BURST_SIZE,
> > +   NULL);
> > +
> > +   /* Increment soft device TXQ */
> > +   txq_pos++;
> > 

Re: [dpdk-dev] [dpdk-stable] [PATCH] net/ixgbe: fix adding multiple mirror type in a rule

[Ferruh Yigit]

On 9/8/2017 1:23 AM, Wu, Jingjing wrote:
> 
> 
>> -Original Message-
>> From: Dai, Wei
>> Sent: Tuesday, September 5, 2017 5:17 PM
>> To: Lu, Wenzhuo ; Ananyev, Konstantin
>> ; Wu, Jingjing 
>> Cc: dev@dpdk.org; Dai, Wei ; sta...@dpdk.org
>> Subject: [PATCH] net/ixgbe: fix adding multiple mirror type in a rule
>>
>> mirror rule_type can be a bit OR result of multiple mirror type of a rule. 
>> Before
>> the commit which introduced this issue, the code supported adding multiple
>> mirror type in a rule.
>>
>> Fixes: 7ba29a76b196 ("ethdev: rename and extend the mirror type")
>> Cc: sta...@dpdk.org
>>
>> Signed-off-by: Wei Dai 
> Acked-by: Jingjing Wu 

Applied to dpdk-next-net/master, thanks.

Re: [dpdk-dev] [PATCH v3 1/4] net/softnic: add softnic PMD

[Ferruh Yigit]

On 9/8/2017 10:30 AM, Singh, Jasvinder wrote:
> Hi Ferruh,
> 
> Thank you for the review and feedback. Please see inline response;
> 
>> -Original Message-
>> From: Yigit, Ferruh
>> Sent: Tuesday, September 5, 2017 3:53 PM
>> To: Singh, Jasvinder ; dev@dpdk.org
>> Cc: Dumitrescu, Cristian ;
>> tho...@monjalon.net
>> Subject: Re: [PATCH v3 1/4] net/softnic: add softnic PMD
>>
>> On 8/11/2017 1:49 PM, Jasvinder Singh wrote:
>>> Add SoftNIC PMD to provide SW fall-back for ethdev APIs.
>>>
>>> Signed-off-by: Cristian Dumitrescu 
>>> Signed-off-by: Jasvinder Singh 

<...>

>>> +
>>> +   /* Default */
>>> +   status = default_init(p, params, numa_node);
>>> +   if (status) {
>>> +   rte_free(p);
>>> +   return NULL;
>>> +   }
>>> +
>>> +   return p;
>>> +}
>>> +
>>> +static void
>>> +pmd_free(struct pmd_internals *p)
>>> +{
>>> +   default_free(p);
>>
>> p->hard.name also needs to be freed here.
> 
> No, we don't allocate any memory to this varibale as it points to the value 
> retrieved from the rte_eth_dev_get_port_by_name();

I guess it is otherway around, the rte_eth_dev_get_port_by_name() uses
hard.name to get and store the port_id of the underlying hw.

how hard.name set, if I don't miss anything, it is strdup from devargs:

--
ret = rte_kvargs_process(kvlist, PMD_PARAM_HARD_NAME, &get_string,
&p->hard.name);
--
get_string()
*(char **)extra_args = strdup(value);
--

Re: [dpdk-dev] [PATCH v4 2/2] net/mlx5: support device removal event

[Ferruh Yigit]

On 9/6/2017 4:03 PM, Matan Azrad wrote:
> Extend the LSC event handling to support the device removal as well.
> 
> The mlx5 event handling has been made capable of receiving and
> signaling several event types at once.
> 
> This support includes next:
> 1. Removal event detection according to the user configuration.
> 2. Calling to all registered mlx5 removal callbacks.
> 3. Capabilities extension to include removal interrupt handling.
> 
> Signed-off-by: Matan Azrad 

Hi Matan,

Can you please update features file [1] to document the support [2] ?

[1]
doc/guides/nics/features/mlx5.ini

[2]
"Removal event="

Re: [dpdk-dev] [PATCH] vhost: adaptively batch small guest memory copies

[Tiwei Bie]

On Fri, Sep 08, 2017 at 03:41:08PM +0800, Yuanhan Liu wrote:
> On Thu, Aug 24, 2017 at 10:19:39AM +0800, Tiwei Bie wrote:
> > This patch adaptively batches the small guest memory copies.
> > By batching the small copies, the efficiency of executing the
> > memory LOAD instructions can be improved greatly, because the
> > memory LOAD latency can be effectively hidden by the pipeline.
> > We saw great performance boosts for small packets PVP test.
> > 
> > This patch improves the performance for small packets, and has
> > distinguished the packets by size. So although the performance
> > for big packets doesn't change, it makes it relatively easy to
> > do some special optimizations for the big packets too.
> 
> The number showed in other replies looks really impressive. Great work!
> This patch also looks good to me. I have one minor comment though.
> 
> [...]
> > +/*
> > + * Structure contains the info for each batched memory copy.
> > + */
> > +struct burst_copy_elem {
> > +   void *dst;
> > +   void *src;
> > +   uint32_t len;
> > +   uint64_t log_addr;
> > +};
> 
> Like the title says, it's more about batch (but not burst). Also, it's
> not a good idea to mix burst and batch. I'd suggest you to use the term
> "batch" consistently.
> 

Okay, I'll change it. Thank you! :-)

Best regards,
Tiwei Bie

Re: [dpdk-dev] [PATCH v3 1/4] net/softnic: add softnic PMD

[Singh, Jasvinder]

> -Original Message-
> From: Yigit, Ferruh
> Sent: Friday, September 8, 2017 10:49 AM
> To: Singh, Jasvinder ; dev@dpdk.org
> Cc: Dumitrescu, Cristian ;
> tho...@monjalon.net
> Subject: Re: [PATCH v3 1/4] net/softnic: add softnic PMD
> 
> On 9/8/2017 10:30 AM, Singh, Jasvinder wrote:
> > Hi Ferruh,
> >
> > Thank you for the review and feedback. Please see inline response;
> >
> >> -Original Message-
> >> From: Yigit, Ferruh
> >> Sent: Tuesday, September 5, 2017 3:53 PM
> >> To: Singh, Jasvinder ; dev@dpdk.org
> >> Cc: Dumitrescu, Cristian ;
> >> tho...@monjalon.net
> >> Subject: Re: [PATCH v3 1/4] net/softnic: add softnic PMD
> >>
> >> On 8/11/2017 1:49 PM, Jasvinder Singh wrote:
> >>> Add SoftNIC PMD to provide SW fall-back for ethdev APIs.
> >>>
> >>> Signed-off-by: Cristian Dumitrescu 
> >>> Signed-off-by: Jasvinder Singh 
> 
> <...>
> 
> >>> +
> >>> + /* Default */
> >>> + status = default_init(p, params, numa_node);
> >>> + if (status) {
> >>> + rte_free(p);
> >>> + return NULL;
> >>> + }
> >>> +
> >>> + return p;
> >>> +}
> >>> +
> >>> +static void
> >>> +pmd_free(struct pmd_internals *p)
> >>> +{
> >>> + default_free(p);
> >>
> >> p->hard.name also needs to be freed here.
> >
> > No, we don't allocate any memory to this varibale as it points to the
> > value retrieved from the rte_eth_dev_get_port_by_name();
> 
> I guess it is otherway around, the rte_eth_dev_get_port_by_name() uses
> hard.name to get and store the port_id of the underlying hw.
> 
> how hard.name set, if I don't miss anything, it is strdup from devargs:
> 
> --
> ret = rte_kvargs_process(kvlist, PMD_PARAM_HARD_NAME, &get_string,
> &p->hard.name);
> --
> get_string()
>   *(char **)extra_args = strdup(value);
> --

Yes, it is set using above, will  correct that. Thanks.

[dpdk-dev] [PATCH v5] net/mlx5: support device removal event

[Matan Azrad]

Extend the LSC event handling to support the device removal as well.

The mlx5 event handling has been made capable of receiving and
signaling several event types at once.

This support includes next:
1. Removal event detection according to the user configuration.
2. Calling to all registered mlx5 removal callbacks.
3. Capabilities extension to include removal interrupt handling.

Signed-off-by: Matan Azrad 
---
 doc/guides/nics/features/mlx5.ini |  1 +
 drivers/net/mlx5/mlx5.c   |  2 +-
 drivers/net/mlx5/mlx5_ethdev.c| 13 +++--
 3 files changed, 13 insertions(+), 3 deletions(-)

V5 for adding support in mlx5 feature doc.

diff --git a/doc/guides/nics/features/mlx5.ini 
b/doc/guides/nics/features/mlx5.ini
index 2913591..20a1461 100644
--- a/doc/guides/nics/features/mlx5.ini
+++ b/doc/guides/nics/features/mlx5.ini
@@ -7,6 +7,7 @@
 Speed capabilities   = Y
 Link status  = Y
 Link status event= Y
+Removal event= Y
 Rx interrupt = Y
 Queue start/stop = Y
 MTU update   = Y
diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 32081a4..99a2fb3 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -870,7 +870,7 @@ static struct rte_pci_driver mlx5_driver = {
},
.id_table = mlx5_pci_id_map,
.probe = mlx5_pci_probe,
-   .drv_flags = RTE_PCI_DRV_INTR_LSC,
+   .drv_flags = RTE_PCI_DRV_INTR_LSC | RTE_PCI_DRV_INTR_RMV,
 };
 
 /**
diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index 64a8db8..b57de4b 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -1172,6 +1172,9 @@ priv_dev_status_handler(struct priv *priv)
event.event_type == IBV_EVENT_PORT_ERR) &&
(priv->dev->data->dev_conf.intr_conf.lsc == 1))
ret |= (1 << RTE_ETH_EVENT_INTR_LSC);
+   else if (event.event_type == IBV_EVENT_DEVICE_FATAL &&
+   priv->dev->data->dev_conf.intr_conf.rmv == 1)
+   ret |= (1 << RTE_ETH_EVENT_INTR_RMV);
else
DEBUG("event type %d on port %d not handled",
  event.event_type, event.element.port_num);
@@ -1227,6 +1230,9 @@ mlx5_dev_interrupt_handler(void *cb_arg)
if (events & (1 << RTE_ETH_EVENT_INTR_LSC))
_rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL,
  NULL);
+   if (events & (1 << RTE_ETH_EVENT_INTR_RMV))
+   _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_RMV, NULL,
+ NULL);
 }
 
 /**
@@ -1240,7 +1246,8 @@ mlx5_dev_interrupt_handler(void *cb_arg)
 void
 priv_dev_interrupt_handler_uninstall(struct priv *priv, struct rte_eth_dev 
*dev)
 {
-   if (!dev->data->dev_conf.intr_conf.lsc)
+   if (!dev->data->dev_conf.intr_conf.lsc &&
+   !dev->data->dev_conf.intr_conf.rmv)
return;
rte_intr_callback_unregister(&priv->intr_handle,
 mlx5_dev_interrupt_handler,
@@ -1265,7 +1272,8 @@ priv_dev_interrupt_handler_install(struct priv *priv, 
struct rte_eth_dev *dev)
 {
int rc, flags;
 
-   if (!dev->data->dev_conf.intr_conf.lsc)
+   if (!dev->data->dev_conf.intr_conf.lsc &&
+   !dev->data->dev_conf.intr_conf.rmv)
return;
assert(priv->ctx->async_fd > 0);
flags = fcntl(priv->ctx->async_fd, F_GETFL);
@@ -1273,6 +1281,7 @@ priv_dev_interrupt_handler_install(struct priv *priv, 
struct rte_eth_dev *dev)
if (rc < 0) {
INFO("failed to change file descriptor async event queue");
dev->data->dev_conf.intr_conf.lsc = 0;
+   dev->data->dev_conf.intr_conf.rmv = 0;
} else {
priv->intr_handle.fd = priv->ctx->async_fd;
priv->intr_handle.type = RTE_INTR_HANDLE_EXT;
-- 
2.7.4

Re: [dpdk-dev] [RFC PATCH 0/1] IPSec Inline and look aside crypto offload

[Akhil Goyal]

Hi  Jerin,

On 9/6/2017 9:23 PM, Jerin Jacob wrote:

-Original Message-

Date: Thu, 31 Aug 2017 15:09:45 +0100
From: Radu Nicolau 
To: Thomas Monjalon , Akhil Goyal 
CC: dev@dpdk.org, bor...@mellanox.com, declan.dohe...@intel.com,
  avia...@mellanox.com, sandeep.ma...@nxp.com, hemant.agra...@nxp.com,
  pablo.de.lara.gua...@intel.com
Subject: Re: [dpdk-dev] [RFC PATCH 0/1] IPSec Inline and look aside crypto
  offload
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101
  Thunderbird/52.1.0


On 8/31/2017 2:14 PM, Thomas Monjalon wrote:

31/08/2017 12:52, Akhil Goyal:

On 8/31/2017 3:36 PM, Thomas Monjalon wrote:

31/08/2017 11:37, Akhil Goyal:

On 8/29/2017 8:19 PM, Thomas Monjalon wrote:

25/07/2017 13:21, Akhil Goyal:

2. Ipsec inline(RTE_SECURITY_SESS_ETH_INLINE_CRYPTO) - This is when the
crypto operations are performed by ethernet device instead of crypto
device. This is also without protocol knowledge inside the ethernet device

If the ethernet device can act as a crypto device, this function
should be offered via the cryptodev interface.

yes this could be thought of but the intent was to keep cryptodev and
ethdev separate, as this would create confusion and will become
difficult to manage.

I think the reverse: it is confusing to do crypto operations through
ethdev interface.
If a device can do "standalone crypto" and networking, it should appear as
2 different ports in my opinion.


How is it different from mode RTE_SECURITY_SESS_NONE?

In RTE_SECURITY_SESS_NONE - crypto device is used for crypto operations.
In RTE_SECURITY_SESS_ETH_INLINE_CRYPTO - ethernet device is used for
crypto operations.
For details of the data path of this mode, refer to the covernote of RFC
patch from Boris.
http://dpdk.org/ml/archives/dev/2017-July/070793.html

For implementation of this mode, see patches from Radu,
http://dpdk.org/ml/archives/dev/2017-August/073587.html

Boris RFC uses rte_flow.
Radu implementation does not use rte_flow.
So I still don't understand the big picture.
Boris asked the question and had no answer.

I'll answer here: it was an omission from my side; v2 of the will include
rte_flow usage, derived from Boris RFC.



Cavium would like to contribute to the definition of this specification
as our HW supports the IPSec offload.

I was trying to review the latest patch. But it looks like there are
multiple versions of the header file floating around. like,

http://dpdk.org/ml/archives/dev/2017-August/073587.html
http://dpdk.org/ml/archives/dev/2017-August/073738.html

Can some one tell which one is latest one to review?

Previously for rte_flow, rte_eventdev specification, etc we had some
header file sign off before jumping to the RFC implementation. IMO, That
model was useful where all the vendors could make inline comments on the
proposal instead of maintaining in the draft repo.  So it possible for
sending the latest revision of the header file patch on the mailing list
for the inline comments.



The RFC remained for some time, there were not many comments. so we all 
agreed moved to implementation. That is the point we requested for the 
repo.


The Cavium comments came bit late.

Currently I have just consolidated the patches in the draft repo and I 
am going rebase it and post to mailing list as well in next 1-2 days.


Since, the implementation is started, we will request any subsequent 
comments as an incremental patches.



Akhil,

Based on your v2 version, we could map a lot with our HW. However, there
are three top level quires for the further review.

1) Some HW cannot offload all types of packets(like IP fragmented
packets) and/or there may have back pressure momentarily from IPSec offload
engine (like Queue is full) etc. So in that case what is the expected behavior
a) Is it an offload driver responsibility to take care of that or
b) Is it passed to application as encrypted packets(in case of inbound)
and the application has to take or of them.



It will depend on the HW capability. If the HW is not supporting the 
fragmented etc packets, they will come as an encrypted packed to the 
application and application need to take care of them.



2) In case of inbound traffic, What is the packet format from offload
driver. i.e
a) Will ESP header will be removed from the packet area after the
decryption.

It depend on the session action type. e.g. for inline crypto, the header 
will be intact. for inline proto, the headers will be removed.

In any case, we need to improve the documentation.


3) We have a few feature like, anti-replay check, SA expiry((byte/time)
notification, etc from HW/FW. So it is not clear from the specification
on the contract between between offload driver vs application
responsibility? Can you give some insight on that? Especially
the error notification scheme if it is an offload driver responsibility.



Anti-replay, SA expiry management is still in my todo  list.
The responsibilities will depend on the amount of offloading the HW/FW 
is o

Re: [dpdk-dev] [dpdk-dev, 01/17] build: add initial infrastructure for meson & ninja builds

[Neil Horman]

On Fri, Sep 08, 2017 at 09:50:26AM +0100, Bruce Richardson wrote:
> On Thu, Sep 07, 2017 at 12:21:57PM -0400, Neil Horman wrote:
> > On Fri, Sep 01, 2017 at 11:04:00AM +0100, Bruce Richardson wrote:
> > > To build with meson and ninja, we need some initial infrastructure in
> > > place. The build files for meson always need to be called "meson.build",
> > > and options get placed in meson_options.txt
> > > 
> > > This commit adds a top-level meson.build file, which sets up the global
> > > variables for tracking drivers, libraries, etc., and then includes other
> > > build files, before finishing by writing the global build configuration
> > > header file and a DPDK pkgconfig file at the end, using some of those same
> > > globals.
> > > 
> > > >From the top level build file, the only include file thus far is for the
> > > config folder, which does some other setup of global configuration
> > > parameters, including pulling in architecture specific parameters from an
> > > architectural subdirectory. A number of configuration build options are
> > > provided for the project to tune a number of global variables which will 
> > > be
> > > used later e.g. max numa nodes, max cores, etc. These settings all make
> > > their way to the global build config header "rte_build_config.h". There is
> > > also a file "rte_config.h", which includes "rte_build_config.h", and this
> > > file is meant to hold other build-time values which are present in our
> > > current static build configuration but are not normally meant for
> > > user-configuration. Ideally, over time, the values placed here should be
> > > moved to the individual libraries or drivers which want those values.
> > > 
> > > Signed-off-by: Bruce Richardson 
> > > Reviewed-by: Harry van Haaren 
> > 
> > I feel like I need to underscore my previous concern here.  While I'm not
> > opposed per-se to a new build system, I am very concerned about the burden 
> > that
> > switching places on downstream consumers, in particular distributions 
> > (since I
> > represent one of them).  Moving to a new build system with new tools means 
> > those
> > tools need to be packaged, tested and shipped, which is a significant work
> > effort.  While it might be a net gain long term, its something you need to 
> > keep
> > in mind when making these changes.
> > 
> Understood. If there is anything we/I can do to make this transition
> easier, please flag it for consideration.
> 
Thank you, I appreciate that.

> > I know you've said that we will be keepting the existing build system,
> > I just need to be sure everyone understands just how important that
> > is.
> > 
> What is your feeling here, in terms of timescale. After any new system
> reaches feature parity, how long would you estimate that we would need
> to support the existing makefile system before it would be safe to
> deprecate it? Should we start a deprecation plan, or is it best just to
> commit to support both until we get all - or almost all - downstream
> consumers switched over? While I wouldn't push for deprecating the old
> system any time soon, and I wouldn't consider maintaining the two
> unduly burdensome, it's not something we want to do in the long term.
> 
I was hoping to avoid putting a specific time frame on it, but its a fair
question to ask.  I feel like any particular timetable is somewhat arbitrary.
Keith suggested a year, which is likely as good as any in my mind.  To put a bit
more detail behind it, a RHEL release cycle is anywhere from 6 to 18 months, so
a year fits well.  If we assume starting a few weeks back when you first
proposed this change, that its going to be merged, that gives us time to package
the build components, build the new package using them, get it through a qa
cycle and fix anything that pops up as a result.  That way, when the switch is
made, it can be done with an immediate deprecation of the old build system with
a level of confidence that some of the more esoteric build targets/configs will
likely work.

> > Though perhaps the time frame for keeping the current build system as 
> > priarmy is
> > less concerning, as feature parity is even more critical.  That is to say, 
> > the
> > new build system must be able to produce the same configurations that the
> > current build system does.  Without it I don't think anyone will be able to 
> > use
> > it consistently, and that will leave a great number of users in a very poor
> > position.  I think getting a little closer to parity with the current 
> > system is
> > warranted.  I'd suggest as a gating factor:
> > 
> > 1) Building on all supported arches
> > 2) Cross building on all supported arches
> > 3) Proper identification of targeted machine (i.e. equivalent of the machine
> > component of the current build system)
> > 
> The question there is gating factor for what? Presumably not for merging
> into the staging tree. But for merging into the main tree for releases?
> I'd push back a little on that, as the new syste

Re: [dpdk-dev] [PATCH] ixgbe: eliminate duplicate filterlist symbols

[David Harton (dharton)]

Hi Konstantin,

A gentle reminder to weigh in on this patch and the 
discussion between Ferruh and myself below.

Thanks,
Dave

> -Original Message-
> From: David Harton (dharton)
> 
> > -Original Message-
> > From: Ferruh Yigit [mailto:ferruh.yi...@intel.com]
> >
> > On 8/25/2017 8:21 PM, David Harton wrote:
> > > Some compilers generate warnings for duplicate symbols for the set
> > > of filter lists current defined in ixgbe_ethdev.h.
> > > This commits moves the defintion and declaration to the source file
> > > that actually uses them and provides a function to initialize the
> > > values akin to its flush function.
> > >
> > > Signed-off-by: David Harton 
> > > ---
> > >  drivers/net/ixgbe/ixgbe_ethdev.c |  8 ++--
> > > drivers/net/ixgbe/ixgbe_ethdev.h |  7 +--
> > >  drivers/net/ixgbe/ixgbe_flow.c   | 18 ++
> > >  3 files changed, 21 insertions(+), 12 deletions(-)
> > >
> > > diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c
> > > b/drivers/net/ixgbe/ixgbe_ethdev.c
> > > index 1ec5aaf..ed21af5 100644
> > > --- a/drivers/net/ixgbe/ixgbe_ethdev.c
> > > +++ b/drivers/net/ixgbe/ixgbe_ethdev.c
> > > @@ -1332,12 +1332,8 @@ struct rte_ixgbe_xstats_name_off {
> > >   /* initialize l2 tunnel filter list & hash */
> > >   ixgbe_l2_tn_filter_init(eth_dev);
> > >
> > > - TAILQ_INIT(&filter_ntuple_list);
> > > - TAILQ_INIT(&filter_ethertype_list);
> > > - TAILQ_INIT(&filter_syn_list);
> > > - TAILQ_INIT(&filter_fdir_list);
> > > - TAILQ_INIT(&filter_l2_tunnel_list);
> > > - TAILQ_INIT(&ixgbe_flow_list);
> > > + /* initialize flow filter lists */
> > > + ixgbe_filterlist_init();
> >
> > Will it work if list initialization converted to the static
> > initialization? That would remove requirement of this function call.
> 
> I was trying to keep the behavior the same:
>   .probe -> eth_ixgbe_dev_init -> ixgbe_filterlist_init
>   .remove -> eth_ixgbe_dev_uninit -> ixgbe_filterlist_flush
> 
> While I think the below would work I always like mirror/parallel
> relationships like ctor/dtor for objects.
> 
> Is that reasonable?
> 
> Thanks,
> Dave
> 
> >
> > something like:
> >
> > diff --git a/drivers/net/ixgbe/ixgbe_flow.c
> > b/drivers/net/ixgbe/ixgbe_flow.c index c8645f056..6184eb4af 100644
> > --- a/drivers/net/ixgbe/ixgbe_flow.c
> > +++ b/drivers/net/ixgbe/ixgbe_flow.c
> > @@ -79,6 +79,24 @@
> >  #define IXGBE_MAX_N_TUPLE_PRIO 7
> >  #define IXGBE_MAX_FLX_SOURCE_OFF 62
> >
> > +static TAILQ_HEAD(ixgbe_ntuple_filter_list, ixgbe_ntuple_filter_ele)
> > +   filter_ntuple_list =
> > +TAILQ_HEAD_INITIALIZER(filter_ntuple_list);
> > +
> > +static TAILQ_HEAD(ixgbe_ethertype_filter_list,
> > ixgbe_ethertype_filter_ele)
> > +   filter_ethertype_list =
> > TAILQ_HEAD_INITIALIZER(filter_ethertype_list);
> > +
> > +static TAILQ_HEAD(ixgbe_syn_filter_list, ixgbe_eth_syn_filter_ele)
> > +   filter_syn_list = TAILQ_HEAD_INITIALIZER(filter_syn_list);
> > +
> > +static TAILQ_HEAD(ixgbe_fdir_rule_filter_list, ixgbe_fdir_rule_ele)
> > +   filter_fdir_list = TAILQ_HEAD_INITIALIZER(filter_fdir_list);
> > +
> > +static TAILQ_HEAD(ixgbe_l2_tunnel_filter_list,
> > ixgbe_eth_l2_tunnel_conf_ele)
> > +   filter_l2_tunnel_list =
> > TAILQ_HEAD_INITIALIZER(filter_l2_tunnel_list);
> > +
> > +static TAILQ_HEAD(ixgbe_flow_mem_list, ixgbe_flow_mem)
> > +   ixgbe_flow_list = TAILQ_HEAD_INITIALIZER(ixgbe_flow_list);
> > +

[dpdk-dev] [PATCH v2] vhost: adaptively batch small guest memory copies

[Tiwei Bie]

This patch adaptively batches the small guest memory copies.
By batching the small copies, the efficiency of executing the
memory LOAD instructions can be improved greatly, because the
memory LOAD latency can be effectively hidden by the pipeline.
We saw great performance boosts for small packets PVP test.

This patch improves the performance for small packets, and has
distinguished the packets by size. So although the performance
for big packets doesn't change, it makes it relatively easy to
do some special optimizations for the big packets too.

Signed-off-by: Tiwei Bie 
Signed-off-by: Zhihong Wang 
Signed-off-by: Zhiyong Yang 
Reviewed-by: Maxime Coquelin 
---
This optimization depends on the CPU internal pipeline design.
So further tests (e.g. ARM) from the community is appreciated.

v2: s/burst_copy/batch_copy/g

 lib/librte_vhost/vhost.c  |   2 +-
 lib/librte_vhost/vhost.h  |  13 +++
 lib/librte_vhost/vhost_user.c |  12 +++
 lib/librte_vhost/virtio_net.c | 240 --
 4 files changed, 209 insertions(+), 58 deletions(-)

diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c
index 0b6aa1c..474b6e4 100644
--- a/lib/librte_vhost/vhost.c
+++ b/lib/librte_vhost/vhost.c
@@ -102,7 +102,7 @@ free_device(struct virtio_net *dev)
vq = dev->virtqueue[i];
 
rte_free(vq->shadow_used_ring);
-
+   rte_free(vq->batch_copy_elems);
rte_free(vq);
}
 
diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
index 0f294f3..78fd895 100644
--- a/lib/librte_vhost/vhost.h
+++ b/lib/librte_vhost/vhost.h
@@ -81,6 +81,16 @@ struct zcopy_mbuf {
 };
 TAILQ_HEAD(zcopy_mbuf_list, zcopy_mbuf);
 
+/*
+ * Structure contains the info for each batched memory copy.
+ */
+struct batch_copy_elem {
+   void *dst;
+   void *src;
+   uint32_t len;
+   uint64_t log_addr;
+};
+
 /**
  * Structure contains variables relevant to RX/TX virtqueues.
  */
@@ -114,6 +124,9 @@ struct vhost_virtqueue {
 
struct vring_used_elem  *shadow_used_ring;
uint16_tshadow_used_idx;
+
+   struct batch_copy_elem  *batch_copy_elems;
+   uint16_tbatch_copy_nb_elems;
 } __rte_cache_aligned;
 
 /* Old kernels have no such macros defined */
diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
index ad2e8d3..b62e382 100644
--- a/lib/librte_vhost/vhost_user.c
+++ b/lib/librte_vhost/vhost_user.c
@@ -230,6 +230,15 @@ vhost_user_set_vring_num(struct virtio_net *dev,
return -1;
}
 
+   vq->batch_copy_elems = rte_malloc(NULL,
+   vq->size * sizeof(struct batch_copy_elem),
+   RTE_CACHE_LINE_SIZE);
+   if (!vq->batch_copy_elems) {
+   RTE_LOG(ERR, VHOST_CONFIG,
+   "failed to allocate memory for batching copy.\n");
+   return -1;
+   }
+
return 0;
 }
 
@@ -741,6 +750,9 @@ vhost_user_get_vring_base(struct virtio_net *dev,
rte_free(vq->shadow_used_ring);
vq->shadow_used_ring = NULL;
 
+   rte_free(vq->batch_copy_elems);
+   vq->batch_copy_elems = NULL;
+
return 0;
 }
 
diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c
index a5f0eeb..f8732df 100644
--- a/lib/librte_vhost/virtio_net.c
+++ b/lib/librte_vhost/virtio_net.c
@@ -49,6 +49,8 @@
 
 #define MAX_PKT_BURST 32
 
+#define MAX_BATCH_LEN 256
+
 static bool
 is_valid_virt_queue_idx(uint32_t idx, int is_tx, uint32_t nr_vring)
 {
@@ -105,6 +107,31 @@ update_shadow_used_ring(struct vhost_virtqueue *vq,
vq->shadow_used_ring[i].len = len;
 }
 
+static inline void
+do_data_copy_enqueue(struct virtio_net *dev, struct vhost_virtqueue *vq)
+{
+   struct batch_copy_elem *elem = vq->batch_copy_elems;
+   uint16_t count = vq->batch_copy_nb_elems;
+   int i;
+
+   for (i = 0; i < count; i++) {
+   rte_memcpy(elem[i].dst, elem[i].src, elem[i].len);
+   vhost_log_write(dev, elem[i].log_addr, elem[i].len);
+   PRINT_PACKET(dev, (uintptr_t)elem[i].dst, elem[i].len, 0);
+   }
+}
+
+static inline void
+do_data_copy_dequeue(struct vhost_virtqueue *vq)
+{
+   struct batch_copy_elem *elem = vq->batch_copy_elems;
+   uint16_t count = vq->batch_copy_nb_elems;
+   int i;
+
+   for (i = 0; i < count; i++)
+   rte_memcpy(elem[i].dst, elem[i].src, elem[i].len);
+}
+
 /* avoid write operation when necessary, to lessen cache issues */
 #define ASSIGN_UNLESS_EQUAL(var, val) do { \
if ((var) != (val)) \
@@ -168,8 +195,9 @@ virtio_enqueue_offload(struct rte_mbuf *m_buf, struct 
virtio_net_hdr *net_hdr)
 }
 
 static __rte_always_inline int
-copy_mbuf_to_desc(struct virtio_net *dev, struct vring_desc *descs,
- struct rte_mbuf *m, uint16_t desc_idx, uint32_t size)
+copy_mbuf_to_desc(struct virtio_net *dev, struct 

Re: [dpdk-dev] [PATCH] i40e: fix i40e_validate_mac_addr to permit multicast addresses

[David Harton (dharton)]

Hi Jingjing/Beilei,

A kind reminder to review the patch and the discussion between 
Ferruh and myself.

Thanks,
Dave

> -Original Message-
> From: David Harton (dharton)
> > -Original Message-
> > From: Ferruh Yigit [mailto:ferruh.yi...@intel.com]
> >
> > On 8/22/2017 11:21 PM, David Harton wrote:
> > > The i40e maintains a single MAC filter table for both unicast and
> > > multicast addresses.  The i40e_validate_mac_addr function was
> > > preventing multicast addresses from being added to the table via
> > > i40evf_add_mac_addr.  Fixed the issue by removing the multicast
> > > address check in i40e_validate_mac_addr.
> > >
> > > Signed-off-by: David Harton 
> > > ---
> > >  drivers/net/i40e/base/i40e_common.c | 12 +---
> > >  drivers/net/i40e/i40e_ethdev.c  |  3 ++-
> > >  2 files changed, 7 insertions(+), 8 deletions(-)
> > >
> > > diff --git a/drivers/net/i40e/base/i40e_common.c
> > > b/drivers/net/i40e/base/i40e_common.c
> > > index 900d379..9779854 100644
> > > --- a/drivers/net/i40e/base/i40e_common.c
> > > +++ b/drivers/net/i40e/base/i40e_common.c
> > > @@ -969,10 +969,10 @@ struct i40e_rx_ptype_decoded
> > > i40e_ptype_lookup[] = {
> > >
> > >
> > >  /**
> > > - * i40e_validate_mac_addr - Validate unicast MAC address
> > > + * i40e_validate_mac_addr - Validate MAC address
> > >   * @mac_addr: pointer to MAC address
> > >   *
> > > - * Tests a MAC address to ensure it is a valid Individual Address
> > > + * Tests a MAC address to ensure it is a valid Address
> > >   **/
> > >  enum i40e_status_code i40e_validate_mac_addr(u8 *mac_addr)  { @@
> > > -980,13 +980,11 @@ enum i40e_status_code i40e_validate_mac_addr(u8
> > > *mac_addr)
> > >
> > >   DEBUGFUNC("i40e_validate_mac_addr");
> > >
> > > - /* Broadcast addresses ARE multicast addresses
> > > -  * Make sure it is not a multicast address
> > > + /*
> > >* Reject the zero address
> > >*/
> > > - if (I40E_IS_MULTICAST(mac_addr) ||
> > > - (mac_addr[0] == 0 && mac_addr[1] == 0 && mac_addr[2] == 0 &&
> > > -   mac_addr[3] == 0 && mac_addr[4] == 0 && mac_addr[5] == 0))
> > > + if (mac_addr[0] == 0 && mac_addr[1] == 0 && mac_addr[2] == 0 &&
> > > + mac_addr[3] == 0 && mac_addr[4] == 0 && mac_addr[5] == 0)
> > >   status = I40E_ERR_INVALID_MAC_ADDR;
> > >
> > >   return status;
> > > diff --git a/drivers/net/i40e/i40e_ethdev.c
> > > b/drivers/net/i40e/i40e_ethdev.c index 5f26e24..00b6082 100644
> > > --- a/drivers/net/i40e/i40e_ethdev.c
> > > +++ b/drivers/net/i40e/i40e_ethdev.c
> > > @@ -1199,7 +1199,8 @@ eth_i40e_dev_init(struct rte_eth_dev *dev)
> > >
> > >   /* Get and check the mac address */
> > >   i40e_get_mac_addr(hw, hw->mac.addr);
> > > - if (i40e_validate_mac_addr(hw->mac.addr) != I40E_SUCCESS) {
> > > + if (i40e_validate_mac_addr(hw->mac.addr) != I40E_SUCCESS ||
> > > + I40E_IS_MULTICAST(hw->mac.addr)) {
> >
> > As far as I can see this is to set PF permanent mac address during
> > init(),
> > i40e_macaddr_add() can be used to set multicast address filters.
> > Why do you want to set multicast address as mac permanent address?
> >
> > Any chance that you want to update i40evf_add_mac_addr() to let
> > multicast addresses?
> 
> No.  I'm preserving the existing behavior here.  The previous call used to
> ensure that the permanent mac address here was both non-null and not a
> multi-cast address.
> 
> Since I removed the multi cast check from i40e_validate_mac_addr() I added
> the multicast check here.  If "not valid" or a multicast address fail.
> 
> Regards,
> Dave
> 
> >
> > >   PMD_INIT_LOG(ERR, "mac address is not valid");
> > >   ret = -EIO;
> > >   goto err_get_mac_addr;
> > >

Re: [dpdk-dev] [PATCH] i40e: fix i40e_validate_mac_addr to permit multicast addresses

[Ferruh Yigit]

On 9/8/2017 1:51 PM, David Harton (dharton) wrote:
> Hi Jingjing/Beilei,
> 
> A kind reminder to review the patch and the discussion between 
> Ferruh and myself.

Hi David,

To clarify, your reply answered my concern, so I am OK, still I believe
this should be reviewed by driver maintainer.

Thanks,
ferruh

> 
> Thanks,
> Dave
> 
>> -Original Message-
>> From: David Harton (dharton)
>>> -Original Message-
>>> From: Ferruh Yigit [mailto:ferruh.yi...@intel.com]
>>>
>>> On 8/22/2017 11:21 PM, David Harton wrote:
 The i40e maintains a single MAC filter table for both unicast and
 multicast addresses.  The i40e_validate_mac_addr function was
 preventing multicast addresses from being added to the table via
 i40evf_add_mac_addr.  Fixed the issue by removing the multicast
 address check in i40e_validate_mac_addr.

 Signed-off-by: David Harton 
 ---
  drivers/net/i40e/base/i40e_common.c | 12 +---
  drivers/net/i40e/i40e_ethdev.c  |  3 ++-
  2 files changed, 7 insertions(+), 8 deletions(-)

 diff --git a/drivers/net/i40e/base/i40e_common.c
 b/drivers/net/i40e/base/i40e_common.c
 index 900d379..9779854 100644
 --- a/drivers/net/i40e/base/i40e_common.c
 +++ b/drivers/net/i40e/base/i40e_common.c
 @@ -969,10 +969,10 @@ struct i40e_rx_ptype_decoded
 i40e_ptype_lookup[] = {


  /**
 - * i40e_validate_mac_addr - Validate unicast MAC address
 + * i40e_validate_mac_addr - Validate MAC address
   * @mac_addr: pointer to MAC address
   *
 - * Tests a MAC address to ensure it is a valid Individual Address
 + * Tests a MAC address to ensure it is a valid Address
   **/
  enum i40e_status_code i40e_validate_mac_addr(u8 *mac_addr)  { @@
 -980,13 +980,11 @@ enum i40e_status_code i40e_validate_mac_addr(u8
 *mac_addr)

DEBUGFUNC("i40e_validate_mac_addr");

 -  /* Broadcast addresses ARE multicast addresses
 -   * Make sure it is not a multicast address
 +  /*
 * Reject the zero address
 */
 -  if (I40E_IS_MULTICAST(mac_addr) ||
 -  (mac_addr[0] == 0 && mac_addr[1] == 0 && mac_addr[2] == 0 &&
 -mac_addr[3] == 0 && mac_addr[4] == 0 && mac_addr[5] == 0))
 +  if (mac_addr[0] == 0 && mac_addr[1] == 0 && mac_addr[2] == 0 &&
 +  mac_addr[3] == 0 && mac_addr[4] == 0 && mac_addr[5] == 0)
status = I40E_ERR_INVALID_MAC_ADDR;

return status;
 diff --git a/drivers/net/i40e/i40e_ethdev.c
 b/drivers/net/i40e/i40e_ethdev.c index 5f26e24..00b6082 100644
 --- a/drivers/net/i40e/i40e_ethdev.c
 +++ b/drivers/net/i40e/i40e_ethdev.c
 @@ -1199,7 +1199,8 @@ eth_i40e_dev_init(struct rte_eth_dev *dev)

/* Get and check the mac address */
i40e_get_mac_addr(hw, hw->mac.addr);
 -  if (i40e_validate_mac_addr(hw->mac.addr) != I40E_SUCCESS) {
 +  if (i40e_validate_mac_addr(hw->mac.addr) != I40E_SUCCESS ||
 +  I40E_IS_MULTICAST(hw->mac.addr)) {
>>>
>>> As far as I can see this is to set PF permanent mac address during
>>> init(),
>>> i40e_macaddr_add() can be used to set multicast address filters.
>>> Why do you want to set multicast address as mac permanent address?
>>>
>>> Any chance that you want to update i40evf_add_mac_addr() to let
>>> multicast addresses?
>>
>> No.  I'm preserving the existing behavior here.  The previous call used to
>> ensure that the permanent mac address here was both non-null and not a
>> multi-cast address.
>>
>> Since I removed the multi cast check from i40e_validate_mac_addr() I added
>> the multicast check here.  If "not valid" or a multicast address fail.
>>
>> Regards,
>> Dave
>>
>>>
PMD_INIT_LOG(ERR, "mac address is not valid");
ret = -EIO;
goto err_get_mac_addr;

> 

[dpdk-dev] [PATCH v7] ether: add support for vtune task tracing

[ilia . kurakin]

From: Ilia Kurakin 

The patch simplifies DPDK applications analysis for developers which use
Intel® VTune Amplifier.

The empty cycles are such iterations that yielded no RX packets. As far as
DPDK is running in poll mode, wasting cycles is equal to wasting CPU time.
Tracing such iterations can identify that device is underutilized. Tracing
empty cycles becomes even more critical if a system uses a lot of Ethernet
ports.

The patch gives possibility to analyze empty cycles without changing
application code. All needs to be done is just to reconfigure and rebuild
the DPDK itself with CONFIG_RTE_ETHDEV_PROFILE_ITT_WASTED_RX_ITERATIONS
enbled. The important thing here is that this does not affect DPDK code.
The profiling code is not being compiled if user does not specify config flag.

The patch provides common way to inject RX queues profiling and VTune specific
implementation.


Signed-off-by: Ilia Kurakin 

---

-V2 change:
ITT tasks collection is moved to rx callback

-V3 change:
rte_ethdev_profile.c created, all profile specific code moved there.

Added generic profile function

-V4 change:
checkpatch issues fixed

Added documentation topic

-V5 change:
Documentation fixes

-V6 change:
Documentation changes are moved to different patch

-V7 change:
Documentation and code changes are merged to a single patch.
Added detailed description.


 config/common_base|   1 +
 doc/guides/prog_guide/profile_app.rst |  37 +++-
 lib/librte_ether/Makefile |   1 +
 lib/librte_ether/rte_ethdev.c |   4 +
 lib/librte_ether/rte_ethdev_profile.c | 156 ++
 lib/librte_ether/rte_ethdev_profile.h |  52 
 6 files changed, 250 insertions(+), 1 deletion(-)
 create mode 100644 lib/librte_ether/rte_ethdev_profile.c
 create mode 100644 lib/librte_ether/rte_ethdev_profile.h

diff --git a/config/common_base b/config/common_base
index 5e97a08..12f6be9 100644
--- a/config/common_base
+++ b/config/common_base
@@ -136,6 +136,7 @@ CONFIG_RTE_MAX_QUEUES_PER_PORT=1024
 CONFIG_RTE_LIBRTE_IEEE1588=n
 CONFIG_RTE_ETHDEV_QUEUE_STAT_CNTRS=16
 CONFIG_RTE_ETHDEV_RXTX_CALLBACKS=y
+CONFIG_RTE_ETHDEV_PROFILE_ITT_WASTED_RX_ITERATIONS=n
 
 #
 # Turn off Tx preparation stage
diff --git a/doc/guides/prog_guide/profile_app.rst 
b/doc/guides/prog_guide/profile_app.rst
index 54b546a..ca1c91f 100644
--- a/doc/guides/prog_guide/profile_app.rst
+++ b/doc/guides/prog_guide/profile_app.rst
@@ -39,7 +39,8 @@ Profiling on x86
 
 
 Intel processors provide performance counters to monitor events.
-Some tools provided by Intel, such as VTune, can be used to profile and 
benchmark an application.
+Some tools provided by Intel, such as Intel® VTune™ Amplifier, can be used
+to profile and benchmark an application.
 See the *VTune Performance Analyzer Essentials* publication from Intel Press 
for more information.
 
 For a DPDK application, this can be done in a Linux* application environment 
only.
@@ -59,6 +60,40 @@ Refer to the
 for details about application profiling.
 
 
+Empty cycles tracing
+
+
+Iterations that yielded no RX packets (empty cycles, wasted iterations) can
+be analyzed using VTune Amplifier. This profiling employs the
+`Instrumentation and Tracing Technology (ITT) API
+`_
+feature of VTune Amplifier and requires only reconfiguring the DPDK library,
+no changes in a DPDK application are needed.
+
+To trace wasted iterations on RX queues, first reconfigure DPDK with
+``CONFIG_RTE_ETHDEV_RXTX_CALLBACKS`` and
+``CONFIG_RTE_ETHDEV_PROFILE_ITT_WASTED_RX_ITERATIONS`` enabled.
+
+Then rebuild DPDK, specifying paths to the ITT header and library, which can
+be found in any VTune Amplifier distribution in the *include* and *lib*
+directories respectively:
+
+.. code-block:: console
+
+make EXTRA_CFLAGS=-I \
+ EXTRA_LDLIBS="-L -littnotify"
+
+Finally, to see wasted iterations in your performance analysis results,
+select the *"Analyze user tasks, events, and counters"* checkbox in the
+*"Analysis Type"* tab when configuring analysis via VTune Amplifier GUI.
+Alternatively, when running VTune Amplifier via command line, specify
+``-knob enable-user-tasks=true`` option.
+
+Collected regions of wasted iterations will be marked on VTune Amplifier's
+timeline as ITT tasks. These ITT tasks have predefined names, containing
+Ethernet device and RX queue identifiers.
+
+
 Profiling on ARM64
 --
 
diff --git a/lib/librte_ether/Makefile b/lib/librte_ether/Makefile
index 27d9766..ac1622b 100644
--- a/lib/librte_ether/Makefile
+++ b/lib/librte_ether/Makefile
@@ -46,6 +46,7 @@ LIBABIVER := 6
 SRCS-y += rte_ethdev.c
 SRCS-y += rte_flow.c
 SRCS-y += rte_tm.c
+SRCS-y += rte_ethdev_profile.c
 
 #
 # Export include files
diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
index 0597641..0f020b3 100644
--- a/lib/librte_ether/rte_et

Re: [dpdk-dev] [PATCH v2] devtools: rework abi checker script

[Neil Horman]

On Wed, Sep 06, 2017 at 04:51:01PM +0200, Olivier Matz wrote:
> The intiatial version of the script had some limitations:
> - cannot work on a non-clean workspace
> - environment variables are not documented
> - no compilation log in case of failure
> - return success even it abi is incompatible
> 
> This patch addresses these issues and rework the code.
> 
> Signed-off-by: Olivier Matz 
> ---
> 
> v1->v2:
> - use /usr/bin/env to find bash (which is required)
> - fix displayed path to html reports
> - reword help for -f option
> 
This still doesn't seem to work.  Running the following:
./validate-abi.sh v17.05 v17.08

produces no report, and the end of the abi-check.log file contains:

[nhorman@hmswarspite abi-check]$ tail -n 10 ./abi-check.log 
CMD: abi-dumper librte_vhost.so -o 
/home/nhorman/git/dpdk/devtools/abi-check/222555480/librte_vhost.so.dump -lver 
222555480
Reading debug-info
WARNING: incompatible build option detected: -O0 (required -Og for better 
analysis)
Creating ABI dump

The object ABI has been dumped to:
  /home/nhorman/git/dpdk/devtools/abi-check/222555480/librte_vhost.so.dump
CMD: cd ../..
CMD: git clone ./.. /home/nhorman/git/dpdk/devtools/abi-check/02657b4ad
fatal: repository './..' does not exist

The warning I assume should be fixed to, but the second clone failing seems to
be the most salient bit here

Neil

>  devtools/validate-abi.sh | 389 
> ---
>  1 file changed, 197 insertions(+), 192 deletions(-)
> 
> diff --git a/devtools/validate-abi.sh b/devtools/validate-abi.sh
> index 0accc99b1..d747db189 100755
> --- a/devtools/validate-abi.sh
> +++ b/devtools/validate-abi.sh
> @@ -1,7 +1,8 @@
> -#!/bin/sh
> +#!/usr/bin/env bash
>  #   BSD LICENSE
>  #
>  #   Copyright(c) 2015 Neil Horman. All rights reserved.
> +#   Copyright(c) 2017 6WIND S.A.
>  #   All rights reserved.
>  #
>  #   Redistribution and use in source and binary forms, with or without
> @@ -27,236 +28,240 @@
>  #   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
>  #   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
>  
> -TAG1=$1
> -TAG2=$2
> -TARGET=$3
> -ABI_DIR=`mktemp -d -p /tmp ABI.XX`
> +set -e
>  
> -usage() {
> - echo "$0   "
> -}
> +abicheck=abi-compliance-checker
> +abidump=abi-dumper
> +default_dst=abi-check
> +default_target=x86_64-native-linuxapp-gcc
>  
> -log() {
> - local level=$1
> - shift
> - echo "$*"
> +# trap on error
> +err_report() {
> +echo "$0: error at line $1"
>  }
> +trap 'err_report $LINENO' ERR
>  
> -validate_tags() {
> +print_usage () {
> + cat <<- END_OF_HELP
> + $(basename $0) [options]  
>  
> - if [ -z "$HASH1" ]
> - then
> - echo "invalid revision: $TAG1"
> - return
> - fi
> - if [ -z "$HASH2" ]
> - then
> - echo "invalid revision: $TAG2"
> - return
> - fi
> + This script compares the ABI of 2 git revisions of the current
> + workspace. The output is a html report and a compilation log.
> +
> + The objective is to make sure that applications built against
> + DSOs from the first revision can still run when executed using
> + the DSOs built from the second revision.
> +
> +  and  are git commit id or tags.
> +
> + Options:
> +   -hshow this help
> +   -j   enable parallel compilation with  threads
> +   -vshow compilation logs on the console
> +   -d   change working directory (default is ${default_dst})
> +   -tthe dpdk target to use (default is ${default_target})
> +   -foverwrite existing files in destination directory
> +
> + The script returns 0 on success, or the value of last failing
> + call of ${abicheck} (incompatible abi or the tool has run with errors).
> + The errors returned by ${abidump} are ignored.
> +
> + END_OF_HELP
>  }
>  
> -validate_args() {
> - if [ -z "$TAG1" ]
> - then
> - echo "Must Specify REV1"
> - return
> - fi
> - if [ -z "$TAG2" ]
> - then
> - echo "Must Specify REV2"
> - return
> - fi
> - if [ -z "$TARGET" ]
> - then
> - echo "Must Specify a build target"
> +# log in the file, and on stdout if verbose
> +# $1: level string
> +# $2: string to be logged
> +log() {
> + echo "$1: $2"
> + if [ "${verbose}" != "true" ]; then
> + echo "$1: $2" >&3
>   fi
>  }
>  
> +# launch a command and log it, taking care of surrounding spaces with quotes
> +cmd() {
> + local i s whitespace
> + s=""
> + whitespace="[[:space:]]"
> + for i in "$@"; do
> + if [[ $i =~ $whitespace ]]; then
> + i=\"$i\"
> + fi
> + if [ -z "$s" ]; then
> + s="$i"
> + else
> + s="$s $i"
> + fi
> + done
> +
> + log "CMD" "$s"
> + "$@"

Re: [dpdk-dev] [dpdk-dev, 01/17] build: add initial infrastructure for meson & ninja builds

[Bruce Richardson]

On Fri, Sep 08, 2017 at 07:57:06AM -0400, Neil Horman wrote:
> On Fri, Sep 08, 2017 at 09:50:26AM +0100, Bruce Richardson wrote:
> > On Thu, Sep 07, 2017 at 12:21:57PM -0400, Neil Horman wrote:
> > > On Fri, Sep 01, 2017 at 11:04:00AM +0100, Bruce Richardson wrote:
> > > > To build with meson and ninja, we need some initial infrastructure in
> > > > place. The build files for meson always need to be called "meson.build",
> > > > and options get placed in meson_options.txt
> > > > 
> > > > This commit adds a top-level meson.build file, which sets up the global
> > > > variables for tracking drivers, libraries, etc., and then includes other
> > > > build files, before finishing by writing the global build configuration
> > > > header file and a DPDK pkgconfig file at the end, using some of those 
> > > > same
> > > > globals.
> > > > 
> > > > >From the top level build file, the only include file thus far is for 
> > > > >the
> > > > config folder, which does some other setup of global configuration
> > > > parameters, including pulling in architecture specific parameters from 
> > > > an
> > > > architectural subdirectory. A number of configuration build options are
> > > > provided for the project to tune a number of global variables which 
> > > > will be
> > > > used later e.g. max numa nodes, max cores, etc. These settings all make
> > > > their way to the global build config header "rte_build_config.h". There 
> > > > is
> > > > also a file "rte_config.h", which includes "rte_build_config.h", and 
> > > > this
> > > > file is meant to hold other build-time values which are present in our
> > > > current static build configuration but are not normally meant for
> > > > user-configuration. Ideally, over time, the values placed here should be
> > > > moved to the individual libraries or drivers which want those values.
> > > > 
> > > > Signed-off-by: Bruce Richardson 
> > > > Reviewed-by: Harry van Haaren 
> > > 
> > > I feel like I need to underscore my previous concern here.  While I'm not
> > > opposed per-se to a new build system, I am very concerned about the 
> > > burden that
> > > switching places on downstream consumers, in particular distributions 
> > > (since I
> > > represent one of them).  Moving to a new build system with new tools 
> > > means those
> > > tools need to be packaged, tested and shipped, which is a significant work
> > > effort.  While it might be a net gain long term, its something you need 
> > > to keep
> > > in mind when making these changes.
> > > 
> > Understood. If there is anything we/I can do to make this transition
> > easier, please flag it for consideration.
> > 
> Thank you, I appreciate that.
> 
> > > I know you've said that we will be keepting the existing build system,
> > > I just need to be sure everyone understands just how important that
> > > is.
> > > 
> > What is your feeling here, in terms of timescale. After any new system
> > reaches feature parity, how long would you estimate that we would need
> > to support the existing makefile system before it would be safe to
> > deprecate it? Should we start a deprecation plan, or is it best just to
> > commit to support both until we get all - or almost all - downstream
> > consumers switched over? While I wouldn't push for deprecating the old
> > system any time soon, and I wouldn't consider maintaining the two
> > unduly burdensome, it's not something we want to do in the long term.
> > 
> I was hoping to avoid putting a specific time frame on it, but its a fair
> question to ask.  I feel like any particular timetable is somewhat arbitrary.
> Keith suggested a year, which is likely as good as any in my mind.  To put a 
> bit
> more detail behind it, a RHEL release cycle is anywhere from 6 to 18 months, 
> so
> a year fits well.  If we assume starting a few weeks back when you first
> proposed this change, that its going to be merged, that gives us time to 
> package
> the build components, build the new package using them, get it through a qa
> cycle and fix anything that pops up as a result.  That way, when the switch is
> made, it can be done with an immediate deprecation of the old build system 
> with
> a level of confidence that some of the more esoteric build targets/configs 
> will
> likely work.
> 
> > > Though perhaps the time frame for keeping the current build system as 
> > > priarmy is
> > > less concerning, as feature parity is even more critical.  That is to 
> > > say, the
> > > new build system must be able to produce the same configurations that the
> > > current build system does.  Without it I don't think anyone will be able 
> > > to use
> > > it consistently, and that will leave a great number of users in a very 
> > > poor
> > > position.  I think getting a little closer to parity with the current 
> > > system is
> > > warranted.  I'd suggest as a gating factor:
> > > 
> > > 1) Building on all supported arches
> > > 2) Cross building on all supported arches
> > > 3) Proper identificat

Re: [dpdk-dev] [RFC 0/5] Port Representor for control and monitoring of VF devices

[Alejandro Lucero]

On Thu, Sep 7, 2017 at 2:13 PM, Declan Doherty 
wrote:

> On 07/09/2017 11:01 AM, Alejandro Lucero wrote:
>
>> I understand this is the representor idea suiting Intel cards but it does
>> not cover other possibilities.
>>
>> At least, Netronome and Mellanox require a representor not just for
>> controlling a VF, but also for sending and receiving packets through the
>> representor PMD.
>>
>>
> Hey Alejandro,
>
> What we've proposed here doesn't preclude the support of a data path on
> the representor pmd as the functionality which the PMD supports is
> dependent on how the parent device running the representor broker
> configures each representor instance. As you note in the case of our
> current generation of IO devices supporting a data path doesn't make sense,
> but I think this framework could support that functionality if required.
>
>
Hi Declan,

Tell me if I'm wrong, but this looks to me like kernel netdev ndo_set_vf_*
functions, maybe with the idea of having more possibilities than those
currently available with the kernel. The representor idea I refer to is
completely different, and it is related to using SRIOV with OVS offload. It
has other possibilities but this is the current main reason. Some
references:


https://www.slideshare.net/Netronome/using-agilio-smartnics-for-openstack-networking-acceleration


https://netdevconf.org/1.2/slides/oct6/04_gerlitz_efraim_introduction_to_switchdev_sriov_offloads.pdf

I sent an abstract for a presentation in next Dublin Users meeting, and
>> discussing the representor idea was in the agenda.
>>
>>
> I've also submitted an presentation which I got notification was accepted
> this morning which is based on the RFC, so maybe we should collaborate on a
> presentation if there is a large overlap?


Unfortunately, I did not get my abstract approved. Maybe you could add some
slide about this other view for the shake of discussion.


>
>
> By the way, I remember there was reticence about adding control plane to
>> DPDK. What is the current official DPDK position in this regard?
>>
>>
> I also remember there was concerns about adding control plane APIs to
> DPDK, hence the RFC, but I think the advantage of the representor port
> approach is that there are no new public API being introduced, with only
> back-end support in the PMDs which wish to support this functionality.
>
>
I'd like to know what is the opinion of the DPDK community regarding the
control path.


>
>
>>
>> On Thu, Sep 7, 2017 at 9:35 AM, Mohammad Abdul Awal <
>> mohammad.abdul.a...@intel.com> wrote:
>>
>> Signed-off-by: Mohammad Abdul Awal 
>>> Signed-off-by: Remy Horton 
>>> Signed-off-by: Declan Doherty 
>>> ---
>>> This RFC introduces a port representor based model for the control,
>>> management
>>> and monitoring of SR-IOV virtual function (VF) devices for control plane
>>> applications which have bound the devices physical function.
>>>
>>> Port Representors are virtual poll mode drivers (PMD) which provide a
>>> logical
>>> representation in DPDK for VF ports for control and monitoring. Each port
>>> representor PMD represents a single VF and is associated with it's parent
>>> physical function (PF) PMD which provides the back-end hooks for the
>>> representor
>>> device ops and defines the control domain to which that port belongs.This
>>> allows to use existing DPDK APIs to monitor and control the port without
>>> the
>>> need to create and maintain VF specific APIs.
>>>
>>>
>>> +-+   +---+  +---+
>>> |Control Plane|   |   Data Plane  |  |   Data Plane  |
>>> | Application |   |   Application |  |   Application |
>>> +-+   +---+  +---+
>>> | eth dev api |   |  eth dev api  |  |  eth dev api  |
>>> +-+   +---+  +---+
>>> +---+  +---+  +---+   +---+  +---+
>>> |  PF0  |  | Port  |  | Port  |   |VF0 PMD|  |VF0 PMD|
>>> |  PMD  <--+ Rep 0 |  | Rep 1 |   +---+  +--++
>>> |   |  | PMD   |  | PMD   | |
>>> +---+--^+  +---+  +-+-+ |
>>> |  ||  ||
>>> |  ++  ||
>>> |  ||
>>> |  ||
>>> ++  |
>>> |   |  HW (logical view)   | |  |
>>> | --+--+ +---+ +---+---+ |  |
>>> | |   PF   | |  VF0  | |  VF1  | |  |
>>> | || |   | |   ++
>>> | ++ +---+ +---+ |
>>> | ++ |
>>> | |VEB | |
>>> | +---

Re: [dpdk-dev] [dpdk-dev, 01/17] build: add initial infrastructure for meson & ninja builds

[Bruce Richardson]

+Thomas and Qian from our validation team, in case they have any insight
on build CI.

On Fri, Sep 08, 2017 at 02:55:08PM +0100, Bruce Richardson wrote:
> On Fri, Sep 08, 2017 at 07:57:06AM -0400, Neil Horman wrote:
> > On Fri, Sep 08, 2017 at 09:50:26AM +0100, Bruce Richardson wrote:
> > > On Thu, Sep 07, 2017 at 12:21:57PM -0400, Neil Horman wrote:
> > > > On Fri, Sep 01, 2017 at 11:04:00AM +0100, Bruce Richardson wrote:
> > > > > To build with meson and ninja, we need some initial infrastructure in
> > > > > place. The build files for meson always need to be called 
> > > > > "meson.build",
> > > > > and options get placed in meson_options.txt
> > > > > 
> > > > > This commit adds a top-level meson.build file, which sets up the 
> > > > > global
> > > > > variables for tracking drivers, libraries, etc., and then includes 
> > > > > other
> > > > > build files, before finishing by writing the global build 
> > > > > configuration
> > > > > header file and a DPDK pkgconfig file at the end, using some of those 
> > > > > same
> > > > > globals.
> > > > > 
> > > > > >From the top level build file, the only include file thus far is for 
> > > > > >the
> > > > > config folder, which does some other setup of global configuration
> > > > > parameters, including pulling in architecture specific parameters 
> > > > > from an
> > > > > architectural subdirectory. A number of configuration build options 
> > > > > are
> > > > > provided for the project to tune a number of global variables which 
> > > > > will be
> > > > > used later e.g. max numa nodes, max cores, etc. These settings all 
> > > > > make
> > > > > their way to the global build config header "rte_build_config.h". 
> > > > > There is
> > > > > also a file "rte_config.h", which includes "rte_build_config.h", and 
> > > > > this
> > > > > file is meant to hold other build-time values which are present in our
> > > > > current static build configuration but are not normally meant for
> > > > > user-configuration. Ideally, over time, the values placed here should 
> > > > > be
> > > > > moved to the individual libraries or drivers which want those values.
> > > > > 
> > > > > Signed-off-by: Bruce Richardson 
> > > > > Reviewed-by: Harry van Haaren 
> > > > 
> > > > I feel like I need to underscore my previous concern here.  While I'm 
> > > > not
> > > > opposed per-se to a new build system, I am very concerned about the 
> > > > burden that
> > > > switching places on downstream consumers, in particular distributions 
> > > > (since I
> > > > represent one of them).  Moving to a new build system with new tools 
> > > > means those
> > > > tools need to be packaged, tested and shipped, which is a significant 
> > > > work
> > > > effort.  While it might be a net gain long term, its something you need 
> > > > to keep
> > > > in mind when making these changes.
> > > > 
> > > Understood. If there is anything we/I can do to make this transition
> > > easier, please flag it for consideration.
> > > 
> > Thank you, I appreciate that.
> > 
> > > > I know you've said that we will be keepting the existing build system,
> > > > I just need to be sure everyone understands just how important that
> > > > is.
> > > > 
> > > What is your feeling here, in terms of timescale. After any new system
> > > reaches feature parity, how long would you estimate that we would need
> > > to support the existing makefile system before it would be safe to
> > > deprecate it? Should we start a deprecation plan, or is it best just to
> > > commit to support both until we get all - or almost all - downstream
> > > consumers switched over? While I wouldn't push for deprecating the old
> > > system any time soon, and I wouldn't consider maintaining the two
> > > unduly burdensome, it's not something we want to do in the long term.
> > > 
> > I was hoping to avoid putting a specific time frame on it, but its a fair
> > question to ask.  I feel like any particular timetable is somewhat 
> > arbitrary.
> > Keith suggested a year, which is likely as good as any in my mind.  To put 
> > a bit
> > more detail behind it, a RHEL release cycle is anywhere from 6 to 18 
> > months, so
> > a year fits well.  If we assume starting a few weeks back when you first
> > proposed this change, that its going to be merged, that gives us time to 
> > package
> > the build components, build the new package using them, get it through a qa
> > cycle and fix anything that pops up as a result.  That way, when the switch 
> > is
> > made, it can be done with an immediate deprecation of the old build system 
> > with
> > a level of confidence that some of the more esoteric build targets/configs 
> > will
> > likely work.
> > 
> > > > Though perhaps the time frame for keeping the current build system as 
> > > > priarmy is
> > > > less concerning, as feature parity is even more critical.  That is to 
> > > > say, the
> > > > new build system must be able to produce the same configurations that 
> > > > the
> > > 

Re: [dpdk-dev] [PATCH v2 1/2] crypto/openssl: replace evp APIs with HMAC APIs

[De Lara Guarch, Pablo]

Hi Akhil,

> -Original Message-
> From: Akhil Goyal [mailto:akhil.go...@nxp.com]
> Sent: Tuesday, August 29, 2017 7:59 AM
> To: dev@dpdk.org; De Lara Guarch, Pablo
> 
> Cc: hemant.agra...@nxp.com; Doherty, Declan
> ; Akhil Goyal 
> Subject: [PATCH v2 1/2] crypto/openssl: replace evp APIs with HMAC APIs
> 
> in case of HMAC the openssl APIs HMAC_XXX give better performance for
> all HMAC cases as compared with EVP_XXX
> 
> Signed-off-by: Akhil Goyal 
> ---
> changes in v2:
> patch split in two patches as per Pablo's recommendations
> 
>  drivers/crypto/openssl/rte_openssl_pmd.c | 37 +---

I just come across an issue with this patch on openssl 1.1.0 (below).
Unfortunately, I have already applied the patch in the subtree, but if you 
could send a patch to fix this,
I can integrate as part of that patch.

Thanks,
Pablo

drivers/crypto/openssl/rte_openssl_pmd_private.h:168:14: error: field 'ctx' has 
incomplete type
 HMAC_CTX ctx;
  ^~~
In file included from drivers/crypto/openssl/rte_openssl_pmd_ops.c:39:0:
drivers/crypto/openssl/rte_openssl_pmd_private.h:168:14: error: field 'ctx' has 
incomplete type
 HMAC_CTX ctx;
  ^~~
drivers/crypto/openssl/rte_openssl_pmd.c: In function 
'openssl_set_session_auth_parameters':
drivers/crypto/openssl/rte_openssl_pmd.c:440:3: error: implicit declaration of 
function 'HMAC_CTX_init'; did you mean 'HMAC_CTX_new'? 
[-Werror=implicit-function-declaration]
   HMAC_CTX_init(&sess->auth.hmac.ctx);
   ^
   HMAC_CTX_new

drivers/crypto/openssl/rte_openssl_pmd.c:440:3: error: nested extern 
declaration of 'HMAC_CTX_init' [-Werror=nested-externs]
make[4]: *** [mk/internal/rte.compile-pre.mk:140: rte_openssl_pmd_ops.o] Error 1
make[4]: *** Waiting for unfinished jobs
drivers/crypto/openssl/rte_openssl_pmd.c: In function 'openssl_reset_session':
drivers/crypto/openssl/rte_openssl_pmd.c:588:3: error: implicit declaration of 
function 'HMAC_CTX_cleanup'; did you mean 'HMAC_CTX_get_md'? 
[-Werror=implicit-function-declaration]
   HMAC_CTX_cleanup(&sess->auth.hmac.ctx);
   ^~~~
   HMAC_CTX_get_md
drivers/crypto/openssl/rte_openssl_pmd.c:588:3: error: nested extern 
declaration of 'HMAC_CTX_cleanup' [-Werror=nested-externs]
cc1: all warnings being treated as errors
make[4]: *** [mk/internal/rte.compile-pre.mk:140: rte_openssl_pmd.o] Error 1

[dpdk-dev] [PATCH 2/2] net/sfc: free mbufs in bulks on simple EF10 Tx datapath reap

[Andrew Rybchenko]

From: Ivan Malov 

Signed-off-by: Ivan Malov 
Signed-off-by: Andrew Rybchenko 
---
 doc/guides/nics/sfc_efx.rst   |  4 +++-
 drivers/net/sfc/sfc_dp_tx.h   |  2 ++
 drivers/net/sfc/sfc_ef10_tx.c | 15 ++-
 drivers/net/sfc/sfc_ethdev.c  |  6 ++
 drivers/net/sfc/sfc_tx.c  | 17 +
 5 files changed, 42 insertions(+), 2 deletions(-)

diff --git a/doc/guides/nics/sfc_efx.rst b/doc/guides/nics/sfc_efx.rst
index 973a4a0..028b980 100644
--- a/doc/guides/nics/sfc_efx.rst
+++ b/doc/guides/nics/sfc_efx.rst
@@ -245,12 +245,14 @@ boolean parameters value.
   features available and required by the datapath implementation.
   **efx** chooses libefx-based datapath which supports VLAN insertion
   (full-feature firmware variant only), TSO and multi-segment mbufs.
+  Mbuf segments may come from different mempools, and mbuf reference
+  counters are treated responsibly.
   **ef10** chooses EF10 (SFN7xxx, SFN8xxx) native datapath which is
   more efficient than libefx-based but has no VLAN insertion and TSO
   support yet.
   **ef10_simple** chooses EF10 (SFN7xxx, SFN8xxx) native datapath which
   is even more faster then **ef10** but does not support multi-segment
-  mbufs.
+  mbufs, disallows multiple mempools and neglects mbuf reference counters.
 
 - ``perf_profile`` [auto|throughput|low-latency] (default **throughput**)
 
diff --git a/drivers/net/sfc/sfc_dp_tx.h b/drivers/net/sfc/sfc_dp_tx.h
index db2a70b..94d1b10 100644
--- a/drivers/net/sfc/sfc_dp_tx.h
+++ b/drivers/net/sfc/sfc_dp_tx.h
@@ -142,6 +142,8 @@ struct sfc_dp_tx {
 #define SFC_DP_TX_FEAT_TSO 0x2
 #define SFC_DP_TX_FEAT_MULTI_SEG   0x4
 #define SFC_DP_TX_FEAT_MULTI_PROCESS   0x8
+#define SFC_DP_TX_FEAT_MULTI_POOL  0x10
+#define SFC_DP_TX_FEAT_REFCNT  0x20
sfc_dp_tx_qcreate_t *qcreate;
sfc_dp_tx_qdestroy_t*qdestroy;
sfc_dp_tx_qstart_t  *qstart;
diff --git a/drivers/net/sfc/sfc_ef10_tx.c b/drivers/net/sfc/sfc_ef10_tx.c
index 5127a7a..9047b3e 100644
--- a/drivers/net/sfc/sfc_ef10_tx.c
+++ b/drivers/net/sfc/sfc_ef10_tx.c
@@ -401,14 +401,25 @@ struct sfc_ef10_txq {
pending += sfc_ef10_tx_process_events(txq);
 
if (pending != completed) {
+   struct rte_mbuf *bulk[SFC_TX_REAP_BULK_SIZE];
+   unsigned int nb = 0;
+
do {
struct sfc_ef10_tx_sw_desc *txd;
 
txd = &txq->sw_ring[completed & ptr_mask];
 
-   rte_pktmbuf_free_seg(txd->mbuf);
+   if (nb == RTE_DIM(bulk)) {
+   rte_mempool_put_bulk(bulk[0]->pool,
+(void *)bulk, nb);
+   nb = 0;
+   }
+
+   bulk[nb++] = txd->mbuf;
} while (++completed != pending);
 
+   rte_mempool_put_bulk(bulk[0]->pool, (void *)bulk, nb);
+
txq->completed = completed;
}
 
@@ -614,6 +625,8 @@ struct sfc_dp_tx sfc_ef10_tx = {
.hw_fw_caps = SFC_DP_HW_FW_CAP_EF10,
},
.features   = SFC_DP_TX_FEAT_MULTI_SEG |
+ SFC_DP_TX_FEAT_MULTI_POOL |
+ SFC_DP_TX_FEAT_REFCNT |
  SFC_DP_TX_FEAT_MULTI_PROCESS,
.qcreate= sfc_ef10_tx_qcreate,
.qdestroy   = sfc_ef10_tx_qdestroy,
diff --git a/drivers/net/sfc/sfc_ethdev.c b/drivers/net/sfc/sfc_ethdev.c
index 2b037d8..26e8c93 100644
--- a/drivers/net/sfc/sfc_ethdev.c
+++ b/drivers/net/sfc/sfc_ethdev.c
@@ -145,6 +145,12 @@
if (~sa->dp_tx->features & SFC_DP_TX_FEAT_MULTI_SEG)
dev_info->default_txconf.txq_flags |= ETH_TXQ_FLAGS_NOMULTSEGS;
 
+   if (~sa->dp_tx->features & SFC_DP_TX_FEAT_MULTI_POOL)
+   dev_info->default_txconf.txq_flags |= ETH_TXQ_FLAGS_NOMULTMEMP;
+
+   if (~sa->dp_tx->features & SFC_DP_TX_FEAT_REFCNT)
+   dev_info->default_txconf.txq_flags |= ETH_TXQ_FLAGS_NOREFCOUNT;
+
 #if EFSYS_OPT_RX_SCALE
if (sa->rss_support != EFX_RX_SCALE_UNAVAILABLE) {
dev_info->reta_size = EFX_RSS_TBL_SIZE;
diff --git a/drivers/net/sfc/sfc_tx.c b/drivers/net/sfc/sfc_tx.c
index bf59601..4ea7bd7 100644
--- a/drivers/net/sfc/sfc_tx.c
+++ b/drivers/net/sfc/sfc_tx.c
@@ -91,6 +91,21 @@
rc = EINVAL;
}
 
+   if (((flags & ETH_TXQ_FLAGS_NOMULTMEMP) == 0) &&
+   (~sa->dp_tx->features & SFC_DP_TX_FEAT_MULTI_POOL)) {
+   sfc_err(sa, "multi-mempool is not supported by %s datapath",
+   sa->dp_tx->dp.name);
+   rc = EINVAL;
+   }
+
+   if (((flags & ETH_TXQ_FLAGS_NOREFCOUNT) == 0) &&
+   (~sa->dp_tx->features & SFC_DP_TX_FEAT_REFCNT)) {
+   sfc_err(sa,
+   "mbuf reference counters are neglected by

[dpdk-dev] [PATCH 1/2] net/sfc: free mbufs in bulks on EF10 native Tx datapath reap

[Andrew Rybchenko]

From: Ivan Malov 

Signed-off-by: Ivan Malov 
Signed-off-by: Andrew Rybchenko 
---
 drivers/net/sfc/sfc_ef10_tx.c | 48 ---
 drivers/net/sfc/sfc_tweak.h   |  3 +++
 2 files changed, 44 insertions(+), 7 deletions(-)

diff --git a/drivers/net/sfc/sfc_ef10_tx.c b/drivers/net/sfc/sfc_ef10_tx.c
index 182fc23..5127a7a 100644
--- a/drivers/net/sfc/sfc_ef10_tx.c
+++ b/drivers/net/sfc/sfc_ef10_tx.c
@@ -158,17 +158,35 @@ struct sfc_ef10_txq {
pending += sfc_ef10_tx_process_events(txq);
 
if (pending != completed) {
+   struct rte_mbuf *bulk[SFC_TX_REAP_BULK_SIZE];
+   unsigned int nb = 0;
+
do {
struct sfc_ef10_tx_sw_desc *txd;
+   struct rte_mbuf *m;
 
txd = &txq->sw_ring[completed & ptr_mask];
+   if (txd->mbuf == NULL)
+   continue;
 
-   if (txd->mbuf != NULL) {
-   rte_pktmbuf_free(txd->mbuf);
-   txd->mbuf = NULL;
+   m = rte_pktmbuf_prefree_seg(txd->mbuf);
+   txd->mbuf = NULL;
+   if (m == NULL)
+   continue;
+
+   if ((nb == RTE_DIM(bulk)) ||
+   ((nb != 0) && (m->pool != bulk[0]->pool))) {
+   rte_mempool_put_bulk(bulk[0]->pool,
+(void *)bulk, nb);
+   nb = 0;
}
+
+   bulk[nb++] = m;
} while (++completed != pending);
 
+   if (nb != 0)
+   rte_mempool_put_bulk(bulk[0]->pool, (void *)bulk, nb);
+
txq->completed = completed;
}
 
@@ -325,6 +343,7 @@ struct sfc_ef10_txq {
do {
phys_addr_t seg_addr = rte_mbuf_data_dma_addr(m_seg);
unsigned int seg_len = rte_pktmbuf_data_len(m_seg);
+   unsigned int id = added & ptr_mask;
 
SFC_ASSERT(seg_len <= SFC_EF10_TX_DMA_DESC_LEN_MAX);
 
@@ -332,15 +351,30 @@ struct sfc_ef10_txq {
 
sfc_ef10_tx_qdesc_dma_create(seg_addr,
seg_len, (pkt_len == 0),
-   &txq->txq_hw_ring[added & ptr_mask]);
+   &txq->txq_hw_ring[id]);
+
+   /*
+* rte_pktmbuf_free() is commonly used in DPDK for
+* recycling packets - the function checks every
+* segment's reference counter and returns the
+* buffer to its pool whenever possible;
+* nevertheless, freeing mbuf segments one by one
+* may entail some performance decline;
+* from this point, sfc_efx_tx_reap() does the same job
+* on its own and frees buffers in bulks (all mbufs
+* within a bulk belong to the same pool);
+* from this perspective, individual segment pointers
+* must be associated with the corresponding SW
+* descriptors independently so that only one loop
+* is sufficient on reap to inspect all the buffers
+*/
+   txq->sw_ring[id].mbuf = m_seg;
+
++added;
 
} while ((m_seg = m_seg->next) != 0);
 
dma_desc_space -= (added - pkt_start);
-
-   /* Assign mbuf to the last used desc */
-   txq->sw_ring[(added - 1) & ptr_mask].mbuf = *pktp;
}
 
if (likely(added != txq->added)) {
diff --git a/drivers/net/sfc/sfc_tweak.h b/drivers/net/sfc/sfc_tweak.h
index 4ef7fc8..fd2f75c 100644
--- a/drivers/net/sfc/sfc_tweak.h
+++ b/drivers/net/sfc/sfc_tweak.h
@@ -53,4 +53,7 @@
 /** Default free threshold follows recommendations from DPDK documentation */
 #define SFC_TX_DEFAULT_FREE_THRESH 32
 
+/** Number of mbufs to be freed in bulk in a single call */
+#define SFC_TX_REAP_BULK_SIZE  32
+
 #endif /* _SFC_TWEAK_H_ */
-- 
1.8.2.3

[dpdk-dev] [PATCH v2 1/2] eal/malloc: fix RTE malloc debug macro

[Xueming Li]

This patch replaces broken macro RTE_LIBRTE_MALLOC_DEBUG with
RTE_MALLOC_DEBUG.

Fixes: af75078fece3 ("first public release")

Cc: Sergio Gonzalez Monroy 
Signed-off-by: Xueming Li 
---
 lib/librte_eal/common/malloc_elem.h |  4 ++--
 test/test/test_malloc.c | 10 +-
 2 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/lib/librte_eal/common/malloc_elem.h 
b/lib/librte_eal/common/malloc_elem.h
index f04b2d1e4..ce39129d9 100644
--- a/lib/librte_eal/common/malloc_elem.h
+++ b/lib/librte_eal/common/malloc_elem.h
@@ -53,13 +53,13 @@ struct malloc_elem {
volatile enum elem_state state;
uint32_t pad;
size_t size;
-#ifdef RTE_LIBRTE_MALLOC_DEBUG
+#ifdef RTE_MALLOC_DEBUG
uint64_t header_cookie; /* Cookie marking start of data */
/* trailer cookie at start + size */
 #endif
 } __rte_cache_aligned;
 
-#ifndef RTE_LIBRTE_MALLOC_DEBUG
+#ifndef RTE_MALLOC_DEBUG
 static const unsigned MALLOC_ELEM_TRAILER_LEN = 0;
 
 /* dummy function - just check if pointer is non-null */
diff --git a/test/test/test_malloc.c b/test/test/test_malloc.c
index 013fd4407..5558acda4 100644
--- a/test/test/test_malloc.c
+++ b/test/test/test_malloc.c
@@ -108,7 +108,7 @@ test_align_overlap_per_lcore(__attribute__((unused)) void 
*arg)
}
for(j = 0; j < 1000 ; j++) {
if( *(char *)p1 != 0) {
-   printf("rte_zmalloc didn't zero"
+   printf("rte_zmalloc didn't zero "
   "the allocated memory\n");
ret = -1;
}
@@ -180,7 +180,7 @@ test_reordered_free_per_lcore(__attribute__((unused)) void 
*arg)
}
for(j = 0; j < 1000 ; j++) {
if( *(char *)p1 != 0) {
-   printf("rte_zmalloc didn't zero"
+   printf("rte_zmalloc didn't zero "
   "the allocated memory\n");
ret = -1;
}
@@ -293,7 +293,7 @@ test_multi_alloc_statistics(void)
struct rte_malloc_socket_stats pre_stats, post_stats ,first_stats, 
second_stats;
size_t size = 2048;
int align = 1024;
-#ifndef RTE_LIBRTE_MALLOC_DEBUG
+#ifndef RTE_MALLOC_DEBUG
int trailer_size = 0;
 #else
int trailer_size = RTE_CACHE_LINE_SIZE;
@@ -623,7 +623,7 @@ test_rte_malloc_validate(void)
const size_t request_size = 1024;
size_t allocated_size;
char *data_ptr = rte_malloc(NULL, request_size, RTE_CACHE_LINE_SIZE);
-#ifdef RTE_LIBRTE_MALLOC_DEBUG
+#ifdef RTE_MALLOC_DEBUG
int retval;
char *over_write_vals = NULL;
 #endif
@@ -645,7 +645,7 @@ test_rte_malloc_validate(void)
if (allocated_size < request_size)
err_return();
 
-#ifdef RTE_LIBRTE_MALLOC_DEBUG
+#ifdef RTE_MALLOC_DEBUG
 
/** change the header to be bad */
char save_buf[64];
-- 
2.13.3

[dpdk-dev] [PATCH v2 2/2] eal/malloc: fix RTE malloc element free

[Xueming Li]

malloc_elem_free() is clearing(setting to 0) the trailer cookie when
RTE_MALLOC_DEBUG is enabled. In case of joining free neighbor element,
part of joined memory is not getting cleared due to missing the length
of trailer cookie in the middle.

This patch fixes calculation of free memory length to be cleared in
malloc_elem_free() by including trailer cookie.

Fixes: af75078fece3 ("first public release")

Cc: Sergio Gonzalez Monroy 
Signed-off-by: Xueming Li 
---
 lib/librte_eal/common/malloc_elem.c | 8 
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/lib/librte_eal/common/malloc_elem.c 
b/lib/librte_eal/common/malloc_elem.c
index 150769057..889dffd21 100644
--- a/lib/librte_eal/common/malloc_elem.c
+++ b/lib/librte_eal/common/malloc_elem.c
@@ -275,14 +275,14 @@ malloc_elem_free(struct malloc_elem *elem)
return -1;
 
rte_spinlock_lock(&(elem->heap->lock));
-   size_t sz = elem->size - sizeof(*elem);
+   size_t sz = elem->size - sizeof(*elem) - MALLOC_ELEM_TRAILER_LEN;
uint8_t *ptr = (uint8_t *)&elem[1];
struct malloc_elem *next = RTE_PTR_ADD(elem, elem->size);
if (next->state == ELEM_FREE){
/* remove from free list, join to this one */
elem_free_list_remove(next);
join_elem(elem, next);
-   sz += sizeof(*elem);
+   sz += (sizeof(*elem) + MALLOC_ELEM_TRAILER_LEN);
}
 
/* check if previous element is free, if so join with it and return,
@@ -291,8 +291,8 @@ malloc_elem_free(struct malloc_elem *elem)
if (elem->prev != NULL && elem->prev->state == ELEM_FREE) {
elem_free_list_remove(elem->prev);
join_elem(elem->prev, elem);
-   sz += sizeof(*elem);
-   ptr -= sizeof(*elem);
+   sz += (sizeof(*elem) + MALLOC_ELEM_TRAILER_LEN);
+   ptr -= (sizeof(*elem) + MALLOC_ELEM_TRAILER_LEN);
elem = elem->prev;
}
malloc_elem_free_list_insert(elem);
-- 
2.13.3

[dpdk-dev] [PATCH v2 0/3] eventdev: add attribute based get APIs

[Harry van Haaren]

This patchset refactors the eventdev API to be more flexible
and capable. In particular, the API is capable of returning an
error value if an invalid device, port or attribute ID is passed
in, which was not possible with the previous APIs.

The implementation of this patchset is based on a v1 patch[1],
and after some discussion this API was seen as the best solution.

In terms of flexibility, the attribute id allows addition of new
common eventdev layer attributes without breaking ABI or adding
new functions. Note that these attributes are not data-path, and
that PMDs should continue to use the xstats API for reporting any
unique PMD statistics that are available.

Regarding API/ABI compatibility, I have removed the functions from
the .map files - please review the .map file changes for ABI issues
carefully.

The last patch of this series adds a started attribute to the device,
allowing the application to query if a device is currently running.

-Harry

[1] http://dpdk.org/dev/patchwork/patch/27152/


Harry van Haaren (3):
  eventdev: add port attribute function
  eventdev: add dev attribute get function
  eventdev: add queue attribute function

 lib/librte_eventdev/rte_eventdev.c   |  88 +++--
 lib/librte_eventdev/rte_eventdev.h   | 115 ++---
 lib/librte_eventdev/rte_eventdev_version.map |  12 ++-
 test/test/test_eventdev.c| 132 +++--
 test/test/test_eventdev_octeontx.c   | 143 ---
 5 files changed, 334 insertions(+), 156 deletions(-)

-- 
2.7.4

[dpdk-dev] [PATCH v2 1/4] eventdev: add port attribute function

[Harry van Haaren]

This commit reworks the port functions to retrieve information
about the port, like the enq or deq depths. Note that "port count"
is a device attribute, and is added in a later patch for dev attributes.

Signed-off-by: Harry van Haaren 
---
 lib/librte_eventdev/rte_eventdev.c   | 33 +++
 lib/librte_eventdev/rte_eventdev.h   | 49 
 lib/librte_eventdev/rte_eventdev_version.map |  7 
 test/test/test_eventdev.c| 16 ++---
 4 files changed, 59 insertions(+), 46 deletions(-)

diff --git a/lib/librte_eventdev/rte_eventdev.c 
b/lib/librte_eventdev/rte_eventdev.c
index bbb3805..6b867b3 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -744,30 +744,35 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
 }
 
 uint8_t
-rte_event_port_dequeue_depth(uint8_t dev_id, uint8_t port_id)
+rte_event_port_count(uint8_t dev_id)
 {
struct rte_eventdev *dev;
 
dev = &rte_eventdevs[dev_id];
-   return dev->data->ports_dequeue_depth[port_id];
+   return dev->data->nb_ports;
 }
 
-uint8_t
-rte_event_port_enqueue_depth(uint8_t dev_id, uint8_t port_id)
+int
+rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, uint32_t attr_id,
+   uint32_t *attr_value /*out */)
 {
struct rte_eventdev *dev;
-
+   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
dev = &rte_eventdevs[dev_id];
-   return dev->data->ports_enqueue_depth[port_id];
-}
-
-uint8_t
-rte_event_port_count(uint8_t dev_id)
-{
-   struct rte_eventdev *dev;
+   if (!is_valid_port(dev, port_id)) {
+   RTE_EDEV_LOG_ERR("Invalid port_id=%" PRIu8, port_id);
+   return -EINVAL;
+   }
 
-   dev = &rte_eventdevs[dev_id];
-   return dev->data->nb_ports;
+   switch(attr_id) {
+   case RTE_EVENT_PORT_ATTR_ENQ_DEPTH:
+   *attr_value = dev->data->ports_enqueue_depth[port_id];
+   break;
+   case RTE_EVENT_PORT_ATTR_DEQ_DEPTH:
+   *attr_value = dev->data->ports_dequeue_depth[port_id];
+   break;
+   };
+   return 0;
 }
 
 int
diff --git a/lib/librte_eventdev/rte_eventdev.h 
b/lib/librte_eventdev/rte_eventdev.h
index 128bc52..fadc209 100644
--- a/lib/librte_eventdev/rte_eventdev.h
+++ b/lib/librte_eventdev/rte_eventdev.h
@@ -715,47 +715,40 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
 const struct rte_event_port_conf *port_conf);
 
 /**
- * Get the number of dequeue queue depth configured for event port designated
- * by its *port_id* on a specific event device
+ * Get the number of ports on a specific event device
  *
  * @param dev_id
  *   Event device identifier.
- * @param port_id
- *   Event port identifier.
  * @return
- *   - The number of configured dequeue queue depth
- *
- * @see rte_event_dequeue_burst()
+ *   - The number of configured ports
  */
 uint8_t
-rte_event_port_dequeue_depth(uint8_t dev_id, uint8_t port_id);
+rte_event_port_count(uint8_t dev_id);
 
 /**
- * Get the number of enqueue queue depth configured for event port designated
- * by its *port_id* on a specific event device
- *
- * @param dev_id
- *   Event device identifier.
- * @param port_id
- *   Event port identifier.
- * @return
- *   - The number of configured enqueue queue depth
- *
- * @see rte_event_enqueue_burst()
+ * The queue depth of the port on the enqueue side
  */
-uint8_t
-rte_event_port_enqueue_depth(uint8_t dev_id, uint8_t port_id);
+#define RTE_EVENT_PORT_ATTR_ENQ_DEPTH 0
+/**
+ * The queue depth of the port on the dequeue side
+ */
+#define RTE_EVENT_PORT_ATTR_DEQ_DEPTH 1
 
 /**
- * Get the number of ports on a specific event device
+ * Get an attribute from a port.
  *
- * @param dev_id
- *   Event device identifier.
- * @return
- *   - The number of configured ports
+ * @param dev_id Eventdev id
+ * @param port_id Eventdev port id
+ * @param attr_id The attribute ID to retrieve
+ * @param[out] attr_value A pointer that will be filled in with the attribute
+ * value if successful
+ *
+ * @retval 0 Successfully returned value
+ * -EINVAL Invalid device, port or attr_id, or attr_value was NULL
  */
-uint8_t
-rte_event_port_count(uint8_t dev_id);
+int
+rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, uint32_t attr_id,
+   uint32_t *attr_value /*out */);
 
 /**
  * Start an event device.
diff --git a/lib/librte_eventdev/rte_eventdev_version.map 
b/lib/librte_eventdev/rte_eventdev_version.map
index 4c48e5f..a0adde3 100644
--- a/lib/librte_eventdev/rte_eventdev_version.map
+++ b/lib/librte_eventdev/rte_eventdev_version.map
@@ -51,3 +51,10 @@ DPDK_17.08 {
rte_event_ring_init;
rte_event_ring_lookup;
 } DPDK_17.05;
+
+EXPERIMENTAL {
+   global:
+
+   rte_event_port_attr_get;
+
+} DPDK_17.08;
diff --git a/test/test/test_eventdev.c b/test/test/test_eventdev.c
index f766191..f3ec470 100644
--

[dpdk-dev] [PATCH v2 2/4] eventdev: add dev attribute get function

[Harry van Haaren]

This commit adds a device attribute function, allowing flexible
fetching of device attributes, like port count or queue count.
The unit tests and .map file are updated to the new function.

Signed-off-by: Harry van Haaren 
---
 lib/librte_eventdev/rte_eventdev.c   | 24 --
 lib/librte_eventdev/rte_eventdev.h   | 28 ---
 lib/librte_eventdev/rte_eventdev_version.map |  2 +-
 test/test/test_eventdev.c| 36 +++---
 test/test/test_eventdev_octeontx.c   | 72 
 5 files changed, 124 insertions(+), 38 deletions(-)

diff --git a/lib/librte_eventdev/rte_eventdev.c 
b/lib/librte_eventdev/rte_eventdev.c
index 6b867b3..177dc6b 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -743,13 +743,27 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
return 0;
 }
 
-uint8_t
-rte_event_port_count(uint8_t dev_id)
+int
+rte_event_dev_attr_get(uint8_t dev_id, uint32_t attr_id,
+  uint32_t *attr_value /*out */)
 {
struct rte_eventdev *dev;
 
+   if(!attr_value)
+   return -EINVAL;
+   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
dev = &rte_eventdevs[dev_id];
-   return dev->data->nb_ports;
+
+   switch(attr_id) {
+   case RTE_EVENT_DEV_ATTR_PORT_COUNT:
+   *attr_value = dev->data->nb_ports;
+   break;
+   case RTE_EVENT_DEV_ATTR_QUEUE_COUNT:
+   *attr_value = dev->data->nb_queues;
+   break;
+   }
+
+   return 0;
 }
 
 int
@@ -757,6 +771,10 @@ rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, 
uint32_t attr_id,
uint32_t *attr_value /*out */)
 {
struct rte_eventdev *dev;
+
+   if(!attr_value)
+   return -EINVAL;
+
RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
dev = &rte_eventdevs[dev_id];
if (!is_valid_port(dev, port_id)) {
diff --git a/lib/librte_eventdev/rte_eventdev.h 
b/lib/librte_eventdev/rte_eventdev.h
index fadc209..e931eb2 100644
--- a/lib/librte_eventdev/rte_eventdev.h
+++ b/lib/librte_eventdev/rte_eventdev.h
@@ -715,15 +715,29 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
 const struct rte_event_port_conf *port_conf);
 
 /**
- * Get the number of ports on a specific event device
+ * The count of ports.
+ */
+#define RTE_EVENT_DEV_ATTR_PORT_COUNT 0
+/**
+ * The count of queues.
+ */
+#define RTE_EVENT_DEV_ATTR_QUEUE_COUNT 1
+
+/**
+ * Get an attribute from a device.
  *
- * @param dev_id
- *   Event device identifier.
- * @return
- *   - The number of configured ports
+ * @param dev_id Eventdev id
+ * @param attr_id The attribute ID to retrieve
+ * @param[out] attr_value A pointer that will be filled in with the attribute
+ * value if successful.
+ *
+ * @retval 0 Successfully retrieved attribute value
+ * -EINVAL Invalid device or  *attr_id* provided, or *attr_value*
+ * is NULL
  */
-uint8_t
-rte_event_port_count(uint8_t dev_id);
+int
+rte_event_dev_attr_get(uint8_t dev_id, uint32_t attr_id,
+  uint32_t *attr_value /*out */);
 
 /**
  * The queue depth of the port on the enqueue side
diff --git a/lib/librte_eventdev/rte_eventdev_version.map 
b/lib/librte_eventdev/rte_eventdev_version.map
index a0adde3..484a071 100644
--- a/lib/librte_eventdev/rte_eventdev_version.map
+++ b/lib/librte_eventdev/rte_eventdev_version.map
@@ -21,7 +21,6 @@ DPDK_17.05 {
rte_event_port_setup;
rte_event_port_dequeue_depth;
rte_event_port_enqueue_depth;
-   rte_event_port_count;
rte_event_port_link;
rte_event_port_unlink;
rte_event_port_links_get;
@@ -55,6 +54,7 @@ DPDK_17.08 {
 EXPERIMENTAL {
global:
 
+   rte_event_dev_attr_get;
rte_event_port_attr_get;
 
 } DPDK_17.08;
diff --git a/test/test/test_eventdev.c b/test/test/test_eventdev.c
index f3ec470..a87b113 100644
--- a/test/test/test_eventdev.c
+++ b/test/test/test_eventdev.c
@@ -386,11 +386,16 @@ test_eventdev_port_default_conf_get(void)
ret = rte_event_port_default_conf_get(TEST_DEV_ID, 0, NULL);
TEST_ASSERT(ret == -EINVAL, "Expected -EINVAL, %d", ret);
 
+   uint32_t port_count;
+   TEST_ASSERT_SUCCESS(rte_event_dev_attr_get(TEST_DEV_ID,
+   RTE_EVENT_DEV_ATTR_PORT_COUNT,
+   &port_count), "Port count get failed");
+
ret = rte_event_port_default_conf_get(TEST_DEV_ID,
-   rte_event_port_count(TEST_DEV_ID) + 1, NULL);
+   port_count + 1, NULL);
TEST_ASSERT(ret == -EINVAL, "Expected -EINVAL, %d", ret);
 
-   for (i = 0; i < rte_event_port_count(TEST_DEV_ID); i++) {
+   for (i = 0; i < (int)port_count; i++) {
ret = rte_event_port_default_conf_get(TEST_DEV_ID, i,
&pconf)

[dpdk-dev] [PATCH v2 3/4] eventdev: add queue attribute function

[Harry van Haaren]

This commit adds a generic queue attribute function. It also removes
the previous rte_event_queue_priority() and priority() functions, and
updates the map files and unit tests to use the new attr functions.

Signed-off-by: Harry van Haaren 
---
 lib/librte_eventdev/rte_eventdev.c   | 47 
 lib/librte_eventdev/rte_eventdev.h   | 49 -
 lib/librte_eventdev/rte_eventdev_version.map |  3 +-
 test/test/test_eventdev.c| 80 +---
 test/test/test_eventdev_octeontx.c   | 75 +++---
 5 files changed, 166 insertions(+), 88 deletions(-)

diff --git a/lib/librte_eventdev/rte_eventdev.c 
b/lib/librte_eventdev/rte_eventdev.c
index 177dc6b..85c0960 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -609,27 +609,6 @@ rte_event_queue_setup(uint8_t dev_id, uint8_t queue_id,
return (*dev->dev_ops->queue_setup)(dev, queue_id, queue_conf);
 }
 
-uint8_t
-rte_event_queue_count(uint8_t dev_id)
-{
-   struct rte_eventdev *dev;
-
-   dev = &rte_eventdevs[dev_id];
-   return dev->data->nb_queues;
-}
-
-uint8_t
-rte_event_queue_priority(uint8_t dev_id, uint8_t queue_id)
-{
-   struct rte_eventdev *dev;
-
-   dev = &rte_eventdevs[dev_id];
-   if (dev->data->event_dev_cap & RTE_EVENT_DEV_CAP_QUEUE_QOS)
-   return dev->data->queues_prio[queue_id];
-   else
-   return RTE_EVENT_DEV_PRIORITY_NORMAL;
-}
-
 static inline int
 is_valid_port(struct rte_eventdev *dev, uint8_t port_id)
 {
@@ -794,6 +773,32 @@ rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, 
uint32_t attr_id,
 }
 
 int
+rte_event_queue_attr_get(uint8_t dev_id, uint8_t queue_id, uint32_t attr_id,
+   uint32_t *attr_value /*out */)
+{
+   struct rte_eventdev *dev;
+
+   if(!attr_value)
+   return -EINVAL;
+
+   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+   dev = &rte_eventdevs[dev_id];
+   if (!is_valid_queue(dev, queue_id)) {
+   RTE_EDEV_LOG_ERR("Invalid queue_id=%" PRIu8, queue_id);
+   return -EINVAL;
+   }
+
+   switch(attr_id) {
+   case RTE_EVENT_QUEUE_ATTR_PRIORITY:
+   *attr_value = RTE_EVENT_DEV_PRIORITY_NORMAL;
+   if (dev->data->event_dev_cap & RTE_EVENT_DEV_CAP_QUEUE_QOS)
+   *attr_value = dev->data->queues_prio[queue_id];
+   break;
+   };
+   return 0;
+}
+
+int
 rte_event_port_link(uint8_t dev_id, uint8_t port_id,
const uint8_t queues[], const uint8_t priorities[],
uint16_t nb_links)
diff --git a/lib/librte_eventdev/rte_eventdev.h 
b/lib/librte_eventdev/rte_eventdev.h
index e931eb2..cd3026d 100644
--- a/lib/librte_eventdev/rte_eventdev.h
+++ b/lib/librte_eventdev/rte_eventdev.h
@@ -606,33 +606,6 @@ int
 rte_event_queue_setup(uint8_t dev_id, uint8_t queue_id,
  const struct rte_event_queue_conf *queue_conf);
 
-/**
- * Get the number of event queues on a specific event device
- *
- * @param dev_id
- *   Event device identifier.
- * @return
- *   - The number of configured event queues
- */
-uint8_t
-rte_event_queue_count(uint8_t dev_id);
-
-/**
- * Get the priority of the event queue on a specific event device
- *
- * @param dev_id
- *   Event device identifier.
- * @param queue_id
- *   Event queue identifier.
- * @return
- *   - If the device has RTE_EVENT_DEV_CAP_QUEUE_QOS capability then the
- *configured priority of the event queue in
- *[RTE_EVENT_DEV_PRIORITY_HIGHEST, RTE_EVENT_DEV_PRIORITY_LOWEST] range
- *else the value RTE_EVENT_DEV_PRIORITY_NORMAL
- */
-uint8_t
-rte_event_queue_priority(uint8_t dev_id, uint8_t queue_id);
-
 /* Event port specific APIs */
 
 /** Event port configuration structure */
@@ -765,6 +738,28 @@ rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, 
uint32_t attr_id,
uint32_t *attr_value /*out */);
 
 /**
+ * The priority of the queue.
+ */
+#define RTE_EVENT_QUEUE_ATTR_PRIORITY 0
+
+/**
+ * Get an attribute from a queue.
+ *
+ * @param dev_id Eventdev id
+ * @param queue_id Eventdev queue id
+ * @param attr_id The attribute ID to retrieve
+ * @param[out] attr_value A pointer that will be filled in with the attribute
+ * value if successful
+ *
+ * @retval 0 Successfully returned value
+ * -EINVAL invalid device, queue or attr_id provided, or attr_value
+ * was NULL
+ */
+int
+rte_event_queue_attr_get(uint8_t dev_id, uint8_t queue_id, uint32_t attr_id,
+   uint32_t *attr_value /*out */);
+
+/**
  * Start an event device.
  *
  * The device start step is the last one and consists of setting the event
diff --git a/lib/librte_eventdev/rte_eventdev_version.map 
b/lib/librte_eventdev/rte_eventdev_version.map
index 484a071..c3854f4 100644
--- a/lib/librte_eventdev/rte_eventdev_version.map
+++ b/lib/librte_event

[dpdk-dev] [PATCH v2 4/4] eventdev: add device started attribute

[Harry van Haaren]

This commit adds an attribute to the eventdev, allowing applications
to retrieve if the eventdev is running or stopped. Note that no API
or ABI changes were required in adding the statistic, and code changes
are minimal.

Signed-off-by: Harry van Haaren 
---
 lib/librte_eventdev/rte_eventdev.c | 3 +++
 lib/librte_eventdev/rte_eventdev.h | 4 
 2 files changed, 7 insertions(+)

diff --git a/lib/librte_eventdev/rte_eventdev.c 
b/lib/librte_eventdev/rte_eventdev.c
index 85c0960..9ec5d42 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -740,6 +740,9 @@ rte_event_dev_attr_get(uint8_t dev_id, uint32_t attr_id,
case RTE_EVENT_DEV_ATTR_QUEUE_COUNT:
*attr_value = dev->data->nb_queues;
break;
+   case RTE_EVENT_DEV_ATTR_STARTED:
+   *attr_value = dev->data->dev_started;
+   break;
}
 
return 0;
diff --git a/lib/librte_eventdev/rte_eventdev.h 
b/lib/librte_eventdev/rte_eventdev.h
index cd3026d..f76d9f9 100644
--- a/lib/librte_eventdev/rte_eventdev.h
+++ b/lib/librte_eventdev/rte_eventdev.h
@@ -695,6 +695,10 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
  * The count of queues.
  */
 #define RTE_EVENT_DEV_ATTR_QUEUE_COUNT 1
+/**
+ * The state of the device, returns zero if stopped, non-zero when running.
+ */
+#define RTE_EVENT_DEV_ATTR_STARTED 2
 
 /**
  * Get an attribute from a device.
-- 
2.7.4

[dpdk-dev] [PATCH v3 0/4] eventdev: add attribute based get APIs

[Harry van Haaren]

This patchset refactors the eventdev API to be more flexible
and capable. In particular, the API is capable of returning an
error value if an invalid device, port or attribute ID is passed
in, which was not possible with the previous APIs.

The implementation of this patchset is based on a v1 patch[1],
and after some discussion this API was seen as the best solution.

In terms of flexibility, the attribute id allows addition of new
common eventdev layer attributes without breaking ABI or adding
new functions. Note that these attributes are not data-path, and
that PMDs should continue to use the xstats API for reporting any
unique PMD statistics that are available.

Regarding API/ABI compatibility, I have removed the functions from
the .map files - please review the .map file changes for ABI issues
carefully.

The last patch of this series adds a started attribute to the device,
allowing the application to query if a device is currently running.

-Harry

[1] http://dpdk.org/dev/patchwork/patch/27152/

---

v3:
- Fix checkpatch issues... somehow I broke my checkpatch script :/

v2:
- New APIs design based on discussion of initial patch.



Harry van Haaren (4):
  eventdev: add port attribute function
  eventdev: add dev attribute get function
  eventdev: add queue attribute function
  eventdev: add device started attribute

 lib/librte_eventdev/rte_eventdev.c   |  91 +++--
 lib/librte_eventdev/rte_eventdev.h   | 118 +++---
 lib/librte_eventdev/rte_eventdev_version.map |  12 ++-
 test/test/test_eventdev.c| 132 +++--
 test/test/test_eventdev_octeontx.c   | 143 ---
 5 files changed, 340 insertions(+), 156 deletions(-)

-- 
2.7.4

[dpdk-dev] [PATCH v3 1/4] eventdev: add port attribute function

[Harry van Haaren]

This commit reworks the port functions to retrieve information
about the port, like the enq or deq depths. Note that "port count"
is a device attribute, and is added in a later patch for dev attributes.

Signed-off-by: Harry van Haaren 
---
 lib/librte_eventdev/rte_eventdev.c   | 33 +++
 lib/librte_eventdev/rte_eventdev.h   | 49 
 lib/librte_eventdev/rte_eventdev_version.map |  7 
 test/test/test_eventdev.c| 16 ++---
 4 files changed, 59 insertions(+), 46 deletions(-)

diff --git a/lib/librte_eventdev/rte_eventdev.c 
b/lib/librte_eventdev/rte_eventdev.c
index bbb3805..a02ff0a 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -744,30 +744,35 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
 }
 
 uint8_t
-rte_event_port_dequeue_depth(uint8_t dev_id, uint8_t port_id)
+rte_event_port_count(uint8_t dev_id)
 {
struct rte_eventdev *dev;
 
dev = &rte_eventdevs[dev_id];
-   return dev->data->ports_dequeue_depth[port_id];
+   return dev->data->nb_ports;
 }
 
-uint8_t
-rte_event_port_enqueue_depth(uint8_t dev_id, uint8_t port_id)
+int
+rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, uint32_t attr_id,
+   uint32_t *attr_value /*out */)
 {
struct rte_eventdev *dev;
-
+   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
dev = &rte_eventdevs[dev_id];
-   return dev->data->ports_enqueue_depth[port_id];
-}
-
-uint8_t
-rte_event_port_count(uint8_t dev_id)
-{
-   struct rte_eventdev *dev;
+   if (!is_valid_port(dev, port_id)) {
+   RTE_EDEV_LOG_ERR("Invalid port_id=%" PRIu8, port_id);
+   return -EINVAL;
+   }
 
-   dev = &rte_eventdevs[dev_id];
-   return dev->data->nb_ports;
+   switch (attr_id) {
+   case RTE_EVENT_PORT_ATTR_ENQ_DEPTH:
+   *attr_value = dev->data->ports_enqueue_depth[port_id];
+   break;
+   case RTE_EVENT_PORT_ATTR_DEQ_DEPTH:
+   *attr_value = dev->data->ports_dequeue_depth[port_id];
+   break;
+   };
+   return 0;
 }
 
 int
diff --git a/lib/librte_eventdev/rte_eventdev.h 
b/lib/librte_eventdev/rte_eventdev.h
index 128bc52..fadc209 100644
--- a/lib/librte_eventdev/rte_eventdev.h
+++ b/lib/librte_eventdev/rte_eventdev.h
@@ -715,47 +715,40 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
 const struct rte_event_port_conf *port_conf);
 
 /**
- * Get the number of dequeue queue depth configured for event port designated
- * by its *port_id* on a specific event device
+ * Get the number of ports on a specific event device
  *
  * @param dev_id
  *   Event device identifier.
- * @param port_id
- *   Event port identifier.
  * @return
- *   - The number of configured dequeue queue depth
- *
- * @see rte_event_dequeue_burst()
+ *   - The number of configured ports
  */
 uint8_t
-rte_event_port_dequeue_depth(uint8_t dev_id, uint8_t port_id);
+rte_event_port_count(uint8_t dev_id);
 
 /**
- * Get the number of enqueue queue depth configured for event port designated
- * by its *port_id* on a specific event device
- *
- * @param dev_id
- *   Event device identifier.
- * @param port_id
- *   Event port identifier.
- * @return
- *   - The number of configured enqueue queue depth
- *
- * @see rte_event_enqueue_burst()
+ * The queue depth of the port on the enqueue side
  */
-uint8_t
-rte_event_port_enqueue_depth(uint8_t dev_id, uint8_t port_id);
+#define RTE_EVENT_PORT_ATTR_ENQ_DEPTH 0
+/**
+ * The queue depth of the port on the dequeue side
+ */
+#define RTE_EVENT_PORT_ATTR_DEQ_DEPTH 1
 
 /**
- * Get the number of ports on a specific event device
+ * Get an attribute from a port.
  *
- * @param dev_id
- *   Event device identifier.
- * @return
- *   - The number of configured ports
+ * @param dev_id Eventdev id
+ * @param port_id Eventdev port id
+ * @param attr_id The attribute ID to retrieve
+ * @param[out] attr_value A pointer that will be filled in with the attribute
+ * value if successful
+ *
+ * @retval 0 Successfully returned value
+ * -EINVAL Invalid device, port or attr_id, or attr_value was NULL
  */
-uint8_t
-rte_event_port_count(uint8_t dev_id);
+int
+rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, uint32_t attr_id,
+   uint32_t *attr_value /*out */);
 
 /**
  * Start an event device.
diff --git a/lib/librte_eventdev/rte_eventdev_version.map 
b/lib/librte_eventdev/rte_eventdev_version.map
index 4c48e5f..a0adde3 100644
--- a/lib/librte_eventdev/rte_eventdev_version.map
+++ b/lib/librte_eventdev/rte_eventdev_version.map
@@ -51,3 +51,10 @@ DPDK_17.08 {
rte_event_ring_init;
rte_event_ring_lookup;
 } DPDK_17.05;
+
+EXPERIMENTAL {
+   global:
+
+   rte_event_port_attr_get;
+
+} DPDK_17.08;
diff --git a/test/test/test_eventdev.c b/test/test/test_eventdev.c
index f766191..f3ec470 100644
-

[dpdk-dev] [PATCH v3 4/4] eventdev: add device started attribute

[Harry van Haaren]

This commit adds an attribute to the eventdev, allowing applications
to retrieve if the eventdev is running or stopped. Note that no API
or ABI changes were required in adding the statistic, and code changes
are minimal.

Signed-off-by: Harry van Haaren 
---
 lib/librte_eventdev/rte_eventdev.c | 3 +++
 lib/librte_eventdev/rte_eventdev.h | 4 
 2 files changed, 7 insertions(+)

diff --git a/lib/librte_eventdev/rte_eventdev.c 
b/lib/librte_eventdev/rte_eventdev.c
index 3756ec7..87a1e19 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -740,6 +740,9 @@ rte_event_dev_attr_get(uint8_t dev_id, uint32_t attr_id,
case RTE_EVENT_DEV_ATTR_QUEUE_COUNT:
*attr_value = dev->data->nb_queues;
break;
+   case RTE_EVENT_DEV_ATTR_STARTED:
+   *attr_value = dev->data->dev_started;
+   break;
}
 
return 0;
diff --git a/lib/librte_eventdev/rte_eventdev.h 
b/lib/librte_eventdev/rte_eventdev.h
index cd3026d..f76d9f9 100644
--- a/lib/librte_eventdev/rte_eventdev.h
+++ b/lib/librte_eventdev/rte_eventdev.h
@@ -695,6 +695,10 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
  * The count of queues.
  */
 #define RTE_EVENT_DEV_ATTR_QUEUE_COUNT 1
+/**
+ * The state of the device, returns zero if stopped, non-zero when running.
+ */
+#define RTE_EVENT_DEV_ATTR_STARTED 2
 
 /**
  * Get an attribute from a device.
-- 
2.7.4

[dpdk-dev] [PATCH v3 3/4] eventdev: add queue attribute function

[Harry van Haaren]

This commit adds a generic queue attribute function. It also removes
the previous rte_event_queue_priority() and priority() functions, and
updates the map files and unit tests to use the new attr functions.

Signed-off-by: Harry van Haaren 
---
 lib/librte_eventdev/rte_eventdev.c   | 47 
 lib/librte_eventdev/rte_eventdev.h   | 49 -
 lib/librte_eventdev/rte_eventdev_version.map |  3 +-
 test/test/test_eventdev.c| 80 +---
 test/test/test_eventdev_octeontx.c   | 75 +++---
 5 files changed, 166 insertions(+), 88 deletions(-)

diff --git a/lib/librte_eventdev/rte_eventdev.c 
b/lib/librte_eventdev/rte_eventdev.c
index 4b1c0be..3756ec7 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -609,27 +609,6 @@ rte_event_queue_setup(uint8_t dev_id, uint8_t queue_id,
return (*dev->dev_ops->queue_setup)(dev, queue_id, queue_conf);
 }
 
-uint8_t
-rte_event_queue_count(uint8_t dev_id)
-{
-   struct rte_eventdev *dev;
-
-   dev = &rte_eventdevs[dev_id];
-   return dev->data->nb_queues;
-}
-
-uint8_t
-rte_event_queue_priority(uint8_t dev_id, uint8_t queue_id)
-{
-   struct rte_eventdev *dev;
-
-   dev = &rte_eventdevs[dev_id];
-   if (dev->data->event_dev_cap & RTE_EVENT_DEV_CAP_QUEUE_QOS)
-   return dev->data->queues_prio[queue_id];
-   else
-   return RTE_EVENT_DEV_PRIORITY_NORMAL;
-}
-
 static inline int
 is_valid_port(struct rte_eventdev *dev, uint8_t port_id)
 {
@@ -794,6 +773,32 @@ rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, 
uint32_t attr_id,
 }
 
 int
+rte_event_queue_attr_get(uint8_t dev_id, uint8_t queue_id, uint32_t attr_id,
+   uint32_t *attr_value /*out */)
+{
+   struct rte_eventdev *dev;
+
+   if (!attr_value)
+   return -EINVAL;
+
+   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+   dev = &rte_eventdevs[dev_id];
+   if (!is_valid_queue(dev, queue_id)) {
+   RTE_EDEV_LOG_ERR("Invalid queue_id=%" PRIu8, queue_id);
+   return -EINVAL;
+   }
+
+   switch (attr_id) {
+   case RTE_EVENT_QUEUE_ATTR_PRIORITY:
+   *attr_value = RTE_EVENT_DEV_PRIORITY_NORMAL;
+   if (dev->data->event_dev_cap & RTE_EVENT_DEV_CAP_QUEUE_QOS)
+   *attr_value = dev->data->queues_prio[queue_id];
+   break;
+   };
+   return 0;
+}
+
+int
 rte_event_port_link(uint8_t dev_id, uint8_t port_id,
const uint8_t queues[], const uint8_t priorities[],
uint16_t nb_links)
diff --git a/lib/librte_eventdev/rte_eventdev.h 
b/lib/librte_eventdev/rte_eventdev.h
index e931eb2..cd3026d 100644
--- a/lib/librte_eventdev/rte_eventdev.h
+++ b/lib/librte_eventdev/rte_eventdev.h
@@ -606,33 +606,6 @@ int
 rte_event_queue_setup(uint8_t dev_id, uint8_t queue_id,
  const struct rte_event_queue_conf *queue_conf);
 
-/**
- * Get the number of event queues on a specific event device
- *
- * @param dev_id
- *   Event device identifier.
- * @return
- *   - The number of configured event queues
- */
-uint8_t
-rte_event_queue_count(uint8_t dev_id);
-
-/**
- * Get the priority of the event queue on a specific event device
- *
- * @param dev_id
- *   Event device identifier.
- * @param queue_id
- *   Event queue identifier.
- * @return
- *   - If the device has RTE_EVENT_DEV_CAP_QUEUE_QOS capability then the
- *configured priority of the event queue in
- *[RTE_EVENT_DEV_PRIORITY_HIGHEST, RTE_EVENT_DEV_PRIORITY_LOWEST] range
- *else the value RTE_EVENT_DEV_PRIORITY_NORMAL
- */
-uint8_t
-rte_event_queue_priority(uint8_t dev_id, uint8_t queue_id);
-
 /* Event port specific APIs */
 
 /** Event port configuration structure */
@@ -765,6 +738,28 @@ rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, 
uint32_t attr_id,
uint32_t *attr_value /*out */);
 
 /**
+ * The priority of the queue.
+ */
+#define RTE_EVENT_QUEUE_ATTR_PRIORITY 0
+
+/**
+ * Get an attribute from a queue.
+ *
+ * @param dev_id Eventdev id
+ * @param queue_id Eventdev queue id
+ * @param attr_id The attribute ID to retrieve
+ * @param[out] attr_value A pointer that will be filled in with the attribute
+ * value if successful
+ *
+ * @retval 0 Successfully returned value
+ * -EINVAL invalid device, queue or attr_id provided, or attr_value
+ * was NULL
+ */
+int
+rte_event_queue_attr_get(uint8_t dev_id, uint8_t queue_id, uint32_t attr_id,
+   uint32_t *attr_value /*out */);
+
+/**
  * Start an event device.
  *
  * The device start step is the last one and consists of setting the event
diff --git a/lib/librte_eventdev/rte_eventdev_version.map 
b/lib/librte_eventdev/rte_eventdev_version.map
index 484a071..c3854f4 100644
--- a/lib/librte_eventdev/rte_eventdev_version.map
+++ b/lib/librte_eve

[dpdk-dev] [PATCH v3 2/4] eventdev: add dev attribute get function

[Harry van Haaren]

This commit adds a device attribute function, allowing flexible
fetching of device attributes, like port count or queue count.
The unit tests and .map file are updated to the new function.

Signed-off-by: Harry van Haaren 
---
 lib/librte_eventdev/rte_eventdev.c   | 24 --
 lib/librte_eventdev/rte_eventdev.h   | 28 ---
 lib/librte_eventdev/rte_eventdev_version.map |  2 +-
 test/test/test_eventdev.c| 36 +++---
 test/test/test_eventdev_octeontx.c   | 72 
 5 files changed, 124 insertions(+), 38 deletions(-)

diff --git a/lib/librte_eventdev/rte_eventdev.c 
b/lib/librte_eventdev/rte_eventdev.c
index a02ff0a..4b1c0be 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -743,13 +743,27 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
return 0;
 }
 
-uint8_t
-rte_event_port_count(uint8_t dev_id)
+int
+rte_event_dev_attr_get(uint8_t dev_id, uint32_t attr_id,
+  uint32_t *attr_value /*out */)
 {
struct rte_eventdev *dev;
 
+   if (!attr_value)
+   return -EINVAL;
+   RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
dev = &rte_eventdevs[dev_id];
-   return dev->data->nb_ports;
+
+   switch (attr_id) {
+   case RTE_EVENT_DEV_ATTR_PORT_COUNT:
+   *attr_value = dev->data->nb_ports;
+   break;
+   case RTE_EVENT_DEV_ATTR_QUEUE_COUNT:
+   *attr_value = dev->data->nb_queues;
+   break;
+   }
+
+   return 0;
 }
 
 int
@@ -757,6 +771,10 @@ rte_event_port_attr_get(uint8_t dev_id, uint8_t port_id, 
uint32_t attr_id,
uint32_t *attr_value /*out */)
 {
struct rte_eventdev *dev;
+
+   if (!attr_value)
+   return -EINVAL;
+
RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
dev = &rte_eventdevs[dev_id];
if (!is_valid_port(dev, port_id)) {
diff --git a/lib/librte_eventdev/rte_eventdev.h 
b/lib/librte_eventdev/rte_eventdev.h
index fadc209..e931eb2 100644
--- a/lib/librte_eventdev/rte_eventdev.h
+++ b/lib/librte_eventdev/rte_eventdev.h
@@ -715,15 +715,29 @@ rte_event_port_setup(uint8_t dev_id, uint8_t port_id,
 const struct rte_event_port_conf *port_conf);
 
 /**
- * Get the number of ports on a specific event device
+ * The count of ports.
+ */
+#define RTE_EVENT_DEV_ATTR_PORT_COUNT 0
+/**
+ * The count of queues.
+ */
+#define RTE_EVENT_DEV_ATTR_QUEUE_COUNT 1
+
+/**
+ * Get an attribute from a device.
  *
- * @param dev_id
- *   Event device identifier.
- * @return
- *   - The number of configured ports
+ * @param dev_id Eventdev id
+ * @param attr_id The attribute ID to retrieve
+ * @param[out] attr_value A pointer that will be filled in with the attribute
+ * value if successful.
+ *
+ * @retval 0 Successfully retrieved attribute value
+ * -EINVAL Invalid device or  *attr_id* provided, or *attr_value*
+ * is NULL
  */
-uint8_t
-rte_event_port_count(uint8_t dev_id);
+int
+rte_event_dev_attr_get(uint8_t dev_id, uint32_t attr_id,
+  uint32_t *attr_value /*out */);
 
 /**
  * The queue depth of the port on the enqueue side
diff --git a/lib/librte_eventdev/rte_eventdev_version.map 
b/lib/librte_eventdev/rte_eventdev_version.map
index a0adde3..484a071 100644
--- a/lib/librte_eventdev/rte_eventdev_version.map
+++ b/lib/librte_eventdev/rte_eventdev_version.map
@@ -21,7 +21,6 @@ DPDK_17.05 {
rte_event_port_setup;
rte_event_port_dequeue_depth;
rte_event_port_enqueue_depth;
-   rte_event_port_count;
rte_event_port_link;
rte_event_port_unlink;
rte_event_port_links_get;
@@ -55,6 +54,7 @@ DPDK_17.08 {
 EXPERIMENTAL {
global:
 
+   rte_event_dev_attr_get;
rte_event_port_attr_get;
 
 } DPDK_17.08;
diff --git a/test/test/test_eventdev.c b/test/test/test_eventdev.c
index f3ec470..a87b113 100644
--- a/test/test/test_eventdev.c
+++ b/test/test/test_eventdev.c
@@ -386,11 +386,16 @@ test_eventdev_port_default_conf_get(void)
ret = rte_event_port_default_conf_get(TEST_DEV_ID, 0, NULL);
TEST_ASSERT(ret == -EINVAL, "Expected -EINVAL, %d", ret);
 
+   uint32_t port_count;
+   TEST_ASSERT_SUCCESS(rte_event_dev_attr_get(TEST_DEV_ID,
+   RTE_EVENT_DEV_ATTR_PORT_COUNT,
+   &port_count), "Port count get failed");
+
ret = rte_event_port_default_conf_get(TEST_DEV_ID,
-   rte_event_port_count(TEST_DEV_ID) + 1, NULL);
+   port_count + 1, NULL);
TEST_ASSERT(ret == -EINVAL, "Expected -EINVAL, %d", ret);
 
-   for (i = 0; i < rte_event_port_count(TEST_DEV_ID); i++) {
+   for (i = 0; i < (int)port_count; i++) {
ret = rte_event_port_default_conf_get(TEST_DEV_ID, i,
&pco

Re: [dpdk-dev] [PATCH 01/17] build: add initial infrastructure for meson & ninja builds

[Van Haaren, Harry]

> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Bruce Richardson
> Sent: Friday, September 1, 2017 11:04 AM
> To: dev@dpdk.org
> Cc: Richardson, Bruce 
> Subject: [dpdk-dev] [PATCH 01/17] build: add initial infrastructure for meson 
> &
> ninja builds



> diff --git a/config/meson.build b/config/meson.build

> +# set the machine type and cflags for it
> +machine = get_option('machine')
> +dpdk_conf.set('RTE_MACHINE', machine)
> +add_project_arguments('-march=@0@'.format(machine), language: 'c')
> +# some libs depend on maths lib
> +add_project_link_arguments('-lm', language: 'c')
> +
> +# add -include rte_config to cflags
> +add_project_arguments('-include', 'rte_config.h', language: 'c')
> +
> +# disable any unwanted warnings
> +unwanted_warnings = [
> + '-Wno-address-of-packed-member',
> + '-Wno-format-truncation'
> +]


Feedback from usage while developing new features;
- Mis-matched integer sign comparison doesn't cause a warning
- And -Werror isn't set by default

Adding these as per below fixes things... but for cleanliness "unwanted 
warnings" should probably be renamed, we want more warnings! :D

# disable any unwanted warnings
unwanted_warnings = [
'-Werror',
'-Wno-address-of-packed-member',
'-Wno-format-truncation',
'-Wsign-compare',
]

The performance of Ninja is amazing for testing shared-object builds and 
version.map changes - feedback is instant. Makes light work of re-checking 
ABI/API stuff.

> +foreach arg: unwanted_warnings
> + if cc.has_argument(arg)
> + add_project_arguments(arg, language: 'c')
> + endif
> +endforeach
> +

Re: [dpdk-dev] [PATCH v2 1/2] eal/malloc: fix RTE malloc debug macro

[Stephen Hemminger]

On Fri,  8 Sep 2017 22:50:54 +0800
Xueming Li  wrote:

> - printf("rte_zmalloc didn't zero"
> + printf("rte_zmalloc didn't zero "
>  "the allocated memory\n");

Please don't break error messages onto two lines.
It makes it harder to use tools like grep and google searches to find where 
message is located.

Checkpatch and other tools make an exception for long lines in quoted strings.

Re: [dpdk-dev] [PATCH 01/17] build: add initial infrastructure for meson & ninja builds

[Richardson, Bruce]

> -Original Message-
> From: Van Haaren, Harry
> Sent: Friday, September 8, 2017 5:03 PM
> To: Richardson, Bruce ; dev@dpdk.org
> Cc: Richardson, Bruce 
> Subject: RE: [dpdk-dev] [PATCH 01/17] build: add initial infrastructure
> for meson & ninja builds
> 
> > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Bruce Richardson
> > Sent: Friday, September 1, 2017 11:04 AM
> > To: dev@dpdk.org
> > Cc: Richardson, Bruce 
> > Subject: [dpdk-dev] [PATCH 01/17] build: add initial infrastructure
> > for meson & ninja builds
> 
> 
> 
> > diff --git a/config/meson.build b/config/meson.build
> 
> > +# set the machine type and cflags for it machine =
> > +get_option('machine') dpdk_conf.set('RTE_MACHINE', machine)
> > +add_project_arguments('-march=@0@'.format(machine), language: 'c') #
> > +some libs depend on maths lib add_project_link_arguments('-lm',
> > +language: 'c')
> > +
> > +# add -include rte_config to cflags
> > +add_project_arguments('-include', 'rte_config.h', language: 'c')
> > +
> > +# disable any unwanted warnings
> > +unwanted_warnings = [
> > +   '-Wno-address-of-packed-member',
> > +   '-Wno-format-truncation'
> > +]
> 
> 
> Feedback from usage while developing new features;
> - Mis-matched integer sign comparison doesn't cause a warning
> - And -Werror isn't set by default
> 
Will fix the former in V2.
For the latter, I already have a patch for it that approximates the behavior of 
the existing system - i.e. a "developer build" setting, which is "off" by 
default, except when you are building from a git checkout in which case it 
defaults to "on". However, I don't think I'll include that in V2, as having 
Werror when we are not reasonably sure we have fixed all warnings in the code 
for all supported compilers and distros is a bad idea IMHO.

> Adding these as per below fixes things... but for cleanliness "unwanted
> warnings" should probably be renamed, we want more warnings! :D
> 
> # disable any unwanted warnings
> unwanted_warnings = [
>   '-Werror',
>   '-Wno-address-of-packed-member',
>   '-Wno-format-truncation',
>   '-Wsign-compare',
> ]
> 
Actually, I'd rather keep it as a list of warnings to disable for now. Enabling 
warnings should be handled differently, as we may only enable specific warnings 
depending on the build type being done.

/Bruce

Re: [dpdk-dev] [PATCH v2 1/4] net/i40e: implement dynamic mapping of sw flow types to hw pctypes

[Ferruh Yigit]

On 9/1/2017 4:02 PM, Kirill Rybalchenko wrote:
> Implement dynamic mapping of software flow types to hardware pctypes.
> This allows to add new flow types and pctypes for DDP without changing
> API of the driver. The mapping table is located in private
> data area for particular network adapter and can be individually
> modified with set of appropriate functions.
> 
> Signed-off-by: Kirill Rybalchenko 
> ---
> v2
> Re-arrange patchset to avoid compillation errors.
> Remove usage of statically defined flow types and pctypes.
<...>

> + for (i = 0; i < I40E_FLOW_TYPE_MAX; i++) {
> + if (flags & (1ULL << i))
> + hena |= adapter->pcypes_tbl[i];

s/pcypes_tbl/pctypes_tbl/ ?

<...>

> @@ -6668,16 +6601,9 @@ static void
>  i40e_pf_disable_rss(struct i40e_pf *pf)
>  {
>   struct i40e_hw *hw = I40E_PF_TO_HW(pf);
> - uint64_t hena;
>  
> - hena = (uint64_t)i40e_read_rx_ctl(hw, I40E_PFQF_HENA(0));
> - hena |= ((uint64_t)i40e_read_rx_ctl(hw, I40E_PFQF_HENA(1))) << 32;
> - if (hw->mac.type == I40E_MAC_X722)
> - hena &= ~I40E_RSS_HENA_ALL_X722;
> - else
> - hena &= ~I40E_RSS_HENA_ALL;
> - i40e_write_rx_ctl(hw, I40E_PFQF_HENA(0), (uint32_t)hena);
> - i40e_write_rx_ctl(hw, I40E_PFQF_HENA(1), (uint32_t)(hena >> 32));

Above logic keeps intact bits of I40E_PFQF_HENA register that are not
supported RSS has filter. But below new code just sets all to zero.

I don't know why above prefer this logic, but would you mind separating
this decision into own its patch? I mean in this patch keep this as it
is, and make new patch to switch new register values.

> + i40e_write_rx_ctl(hw, I40E_PFQF_HENA(0), 0);
> + i40e_write_rx_ctl(hw, I40E_PFQF_HENA(1), 0);
>   I40E_WRITE_FLUSH(hw);
>  }
>  
> @@ -6749,23 +6675,14 @@ static int
>  i40e_hw_rss_hash_set(struct i40e_pf *pf, struct rte_eth_rss_conf *rss_conf)
>  {
>   struct i40e_hw *hw = I40E_PF_TO_HW(pf);
> - uint64_t rss_hf;
>   uint64_t hena;
>   int ret;
>  
> - ret = i40e_set_rss_key(pf->main_vsi, rss_conf->rss_key,
> -rss_conf->rss_key_len);
> + ret = i40e_set_rss_key(pf->main_vsi, rss_conf->rss_key, 
> rss_conf->rss_key_len);

I think there is no change here, and original code is correct as syntax,
please keep it unchanged.

>   if (ret)
>   return ret;
>  
> - rss_hf = rss_conf->rss_hf;
> - hena = (uint64_t)i40e_read_rx_ctl(hw, I40E_PFQF_HENA(0));
> - hena |= ((uint64_t)i40e_read_rx_ctl(hw, I40E_PFQF_HENA(1))) << 32;
> - if (hw->mac.type == I40E_MAC_X722)
> - hena &= ~I40E_RSS_HENA_ALL_X722;
> - else
> - hena &= ~I40E_RSS_HENA_ALL;
> - hena |= i40e_config_hena(rss_hf, hw->mac.type);

Same register comment as above one.

> + hena = i40e_config_hena(rss_conf->rss_hf, pf->adapter);
>   i40e_write_rx_ctl(hw, I40E_PFQF_HENA(0), (uint32_t)hena);
>   i40e_write_rx_ctl(hw, I40E_PFQF_HENA(1), (uint32_t)(hena >> 32));
>   I40E_WRITE_FLUSH(hw);

<...>

> @@ -7820,29 +7736,31 @@ i40e_get_hash_filter_global_config(struct i40e_hw *hw,
>   PMD_DRV_LOG(DEBUG, "Hash function is %s",
>   (reg & I40E_GLQF_CTL_HTOEP_MASK) ? "Toeplitz" : "Simple XOR");
>  
> - for (i = 0; mask && i < RTE_ETH_FLOW_MAX; i++) {
> - if (!(mask & (1UL << i)))
> - continue;
> - mask &= ~(1UL << i);
> - /* Bit set indicats the coresponding flow type is supported */
> - g_cfg->valid_bit_mask[0] |= (1UL << i);
> - /* if flowtype is invalid, continue */
> - if (!I40E_VALID_FLOW(i))
> - continue;
> - pctype = i40e_flowtype_to_pctype(i);
> - reg = i40e_read_rx_ctl(hw, I40E_GLQF_HSYM(pctype));
> - if (reg & I40E_GLQF_HSYM_SYMH_ENA_MASK)
> - g_cfg->sym_hash_enable_mask[0] |= (1UL << i);
> + g_cfg->valid_bit_mask[0] = (uint32_t)adapter->flow_types_msk;

Loosing data here by casting to 32bit, intentional?

> +
> + for (i = 0; i < UINT32_BIT; i++) {

Why use UINT32_BIT? not I40E_FLOW_TYPE_MAX? Is it because
sym_hash_enable_mask is 32bits? Why not increase its stroge if we need
64 flow type / pctype ?

> + if (adapter->pcypes_tbl[i]) {
> + for (j = 0; j < I40E_PCTYPE_MAX; j++) {
> + if (adapter->pcypes_tbl[i] & (1ULL << j)) {
> + reg = i40e_read_rx_ctl(hw, 
> I40E_GLQF_HSYM(j));
> + if (reg & I40E_GLQF_HSYM_SYMH_ENA_MASK) 
> {
> + g_cfg->sym_hash_enable_mask[0] 
> |=
> + (1UL << 
> i);
> + }
> + }
> + }
> + }
>   }
>  
>   return 0;
>  }

<...>

> @@ -7885,64 +7803,26 @@

Re: [dpdk-dev] [PATCH v2 4/4] ethdev: remove unnecessary check for new flow type

[Ferruh Yigit]

On 9/1/2017 4:02 PM, Kirill Rybalchenko wrote:
> Remove unnecessary check for new flow type for rss hash
> filter update.
> 
> Signed-off-by: Kirill Rybalchenko 
> ---
>  lib/librte_ether/rte_ethdev.c | 8 
>  1 file changed, 8 deletions(-)
> 
> diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
> index 0597641..c470997 100644
> --- a/lib/librte_ether/rte_ethdev.c
> +++ b/lib/librte_ether/rte_ethdev.c
> @@ -2263,16 +2263,8 @@ int
>  rte_eth_dev_rss_hash_update(uint8_t port_id, struct rte_eth_rss_conf 
> *rss_conf)
>  {
>   struct rte_eth_dev *dev;
> - uint16_t rss_hash_protos;
>  
>   RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
> - rss_hash_protos = rss_conf->rss_hf;
> - if ((rss_hash_protos != 0) &&
> - ((rss_hash_protos & ETH_RSS_PROTO_MASK) == 0)) {

Can intention here be:
((rss_hash_protos & ~(ETH_RSS_PROTO_MASK)) != 0)

> - RTE_PMD_DEBUG_TRACE("Invalid rss_hash_protos=0x%x\n",
> - rss_hash_protos);
> - return -EINVAL;
> - }
>   dev = &rte_eth_devices[port_id];
>   RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rss_hash_update, -ENOTSUP);
>   return (*dev->dev_ops->rss_hash_update)(dev, rss_conf);
> 

Re: [dpdk-dev] [PATCH v2 2/4] net/i40e: add new functions to manipulate with pctype mapping table

[Ferruh Yigit]

On 9/1/2017 4:02 PM, Kirill Rybalchenko wrote:
> Add new functions which allow modify, return or reset to default
> the contents of flow type to pctype dynamic mapping table.
> 
> Signed-off-by: Kirill Rybalchenko 
> ---
>  drivers/net/i40e/rte_pmd_i40e.c | 98 
> +
>  drivers/net/i40e/rte_pmd_i40e.h | 61 +

.map file needs to be updated for shared build.

<...>

> +int rte_pmd_i40e_flow_type_mapping_get(
> + uint8_t port,
> + struct rte_pmd_i40e_flow_type_mapping *mapping_items,
> + uint16_t size,
> + uint16_t *count,
> + uint8_t valid_only)> +{
> + struct rte_eth_dev *dev;
> + struct i40e_adapter *ad;
> + int n = 0;
> + uint16_t i;
> +
> + RTE_ETH_VALID_PORTID_OR_ERR_RET(port, -ENODEV);
> +
> + dev = &rte_eth_devices[port];
> +
> + if (!is_i40e_supported(dev))
> + return -ENOTSUP;
> +
> + ad = I40E_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
> +
> + for (i = 0; i < I40E_FLOW_TYPE_MAX; i++) {
> + if (n >= size)
> + break;

This returns partial table and success, how caller can know if it get
full table or not?

Is returning partial table required, specially when custom offset is not
supported..

> + if (valid_only && ad->pcypes_tbl[i] == 0ULL)
> + continue;
> + mapping_items[n].flow_type = i;
> + mapping_items[n].pctype = ad->pcypes_tbl[i];
> + n++;
> + }
> +
> + *count = n;
> + return 0;
> +}
<...>

> @@ -657,4 +657,65 @@ int rte_pmd_i40e_ptype_mapping_replace(uint8_t port,
>  int rte_pmd_i40e_add_vf_mac_addr(uint8_t port, uint16_t vf_id,
>struct ether_addr *mac_addr);
>  
> +
> +struct rte_pmd_i40e_flow_type_mapping {
> + uint8_t flow_type; /**< software defined flow type*/

flow_type are RTE_ETH_FLOW_* values defined in rte_eth_ctrl.h which is
public header.

> + uint64_t pctype; /**< hardware defined pctype */

But how user can get pctype values, since this is part of public API,
and if application would like to use these values, is there any public
header list them?

> +};
> +
> +/**
> + * Update hardware defined pctype to software defined flow type
> + * mapping table.
> + *
> + * @param port
> + *pointer to port identifier of the device.
> + * @param mapping_items
> + *the base address of the mapping items array.
> + * @param count
> + *number of mapping items.
> + * @param exclusive
> + *the flag indicate different ptype mapping update method.
> + *-(0) only overwrite referred PTYPE mapping,
> + *   keep other PTYPEs mapping unchanged.
> + *-(!0) overwrite referred PTYPE mapping,
> + *   set other PTYPEs maps to PTYPE_UNKNOWN.
> + */
> +int rte_pmd_i40e_flow_type_mapping_update(
> + uint8_t port,
> + struct rte_pmd_i40e_flow_type_mapping *mapping_items,
> + uint16_t count,
> + uint8_t exclusive);

May prefer boolean here.

> +
> +/**
> + * Get software defined flow type to hardware defined pctype
> + * mapping items.
> + *
> + * @param port
> + *pointer to port identifier of the device.
> + * @param mapping_items
> + *the base address of the array to store returned items.
> + * @param size
> + *the size of the input array.
> + * @param count
> + *the place to store the number of returned items.
> + * @param valid_only
> + *-(0) return full mapping table.
> + *-(!0) only return mapping items which flow_type != 
> RTE_ETH_FLOW_UNKNOWN.
> + */
> +int rte_pmd_i40e_flow_type_mapping_get(
> + uint8_t port,
> + struct rte_pmd_i40e_flow_type_mapping *mapping_items,
> + uint16_t size,
> + uint16_t *count,
> + uint8_t valid_only);

May prefer boolean here.

<...>

Re: [dpdk-dev] [PATCH v2 3/4] app/testpmd: add new commands to manipulate with pctype mapping

[Ferruh Yigit]

On 9/1/2017 4:02 PM, Kirill Rybalchenko wrote:
> Add new commands to manipulate with dynamic flow type to
> pctype mapping table in i40e PMD.
> Commands allow to print table, modify it and reset to default value.
> 
> Signed-off-by: Kirill Rybalchenko 
> ---
>  app/test-pmd/cmdline.c  | 311 
> +++-
>  doc/guides/testpmd_app_ug/testpmd_funcs.rst |  21 ++
>  2 files changed, 322 insertions(+), 10 deletions(-)
> 
> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
> index 0144191..f7d0733 100644
> --- a/app/test-pmd/cmdline.c
> +++ b/app/test-pmd/cmdline.c
> @@ -637,6 +637,16 @@ static void cmd_help_long_parsed(void *parsed_result,
>   "ptype mapping update (port_id) (hw_ptype) (sw_ptype)\n"
>   "Update a ptype mapping item on a port\n\n"
>  
> + "pctype mapping reset (port_id)\n"
> + "Reset flow type to pctype mapping on a port\n\n"
> +
> + "pctype mapping get (port_id)\n"
> + "Get flow ptype to pctype mapping on a port\n\n"
> +
> + "pctype mapping update (port_id) 
> (pctype_id_0[,pctype_id_1]*)"
> + " (flow_type_id)\n"
> + "Update a flow type to pctype mapping item on a 
> port\n\n"

This is another root level command for PMD specific feature, can this be
under "port" or "set" command?

> +
>   , list_pkt_forwarding_modes()
>   );
>   }
> @@ -681,7 +691,8 @@ static void cmd_help_long_parsed(void *parsed_result,
>   "Set 
> crc-strip/scatter/rx-checksum/hardware-vlan/drop_en"
>   " for ports.\n\n"
>  
> - "port config all rss 
> (all|ip|tcp|udp|sctp|ether|port|vxlan|geneve|nvgre|none)\n"
> + "port config all rss 
> (all|ip|tcp|udp|sctp|ether|port|vxlan|"
> + "geneve|nvgre|none|)\n"

Why not use existing defined hash functions but use custom one, is there
a missing one in ethdev?

>   "Set the RSS mode.\n\n"
>  
>   "port config port-id rss reta 
> (hash,queue)[,(hash,queue)]\n"
> @@ -878,8 +889,8 @@ static void cmd_help_long_parsed(void *parsed_result,
>   "set_hash_input_set (port_id) (ipv4|ipv4-frag|"
>   "ipv4-tcp|ipv4-udp|ipv4-sctp|ipv4-other|ipv6|"
>   "ipv6-frag|ipv6-tcp|ipv6-udp|ipv6-sctp|ipv6-other|"
> - "l2_payload) (ovlan|ivlan|src-ipv4|dst-ipv4|src-ipv6|"
> - "dst-ipv6|ipv4-tos|ipv4-proto|ipv6-tc|"
> + "l2_payload|) 
> (ovlan|ivlan|src-ipv4|dst-ipv4|"
> + "src-ipv6|dst-ipv6|ipv4-tos|ipv4-proto|ipv6-tc|"
>   "ipv6-next-header|udp-src-port|udp-dst-port|"
>   "tcp-src-port|tcp-dst-port|sctp-src-port|"
>   "sctp-dst-port|sctp-veri-tag|udp-key|gre-key|fld-1st|"
> @@ -1716,6 +1727,8 @@ cmd_config_rss_parsed(void *parsed_result,
>   rss_conf.rss_hf = ETH_RSS_NVGRE;
>   else if (!strcmp(res->value, "none"))
>   rss_conf.rss_hf = 0;
> + else if (isdigit(res->value[0]) && atoi(res->value) > 0 && 
> atoi(res->value) < 64)
> + rss_conf.rss_hf = 1ULL << atoi(res->value);
>   else {
>   printf("Unknown parameter\n");
>   return;
> @@ -1739,14 +1752,13 @@ cmdline_parse_token_string_t cmd_config_rss_all =
>  cmdline_parse_token_string_t cmd_config_rss_name =
>   TOKEN_STRING_INITIALIZER(struct cmd_config_rss, name, "rss");
>  cmdline_parse_token_string_t cmd_config_rss_value =
> - TOKEN_STRING_INITIALIZER(struct cmd_config_rss, value,
> - "all#ip#tcp#udp#sctp#ether#port#vxlan#geneve#nvgre#none");
> + TOKEN_STRING_INITIALIZER(struct cmd_config_rss, value, NULL);

I guess this will prevent auto completion.

<...>

> +/* pctype mapping get */
> +
> +#define FLOW_TYPE_MAX 64
> +#define PCTYPE_MAX64

These should not be defined by application I believe.

<...>

> +static void
> +cmd_pctype_mapping_update_parsed(
> + void *parsed_result,
> + __attribute__((unused)) struct cmdline *cl,
> + __attribute__((unused)) void *data)
> +{
> + struct cmd_pctype_mapping_update_result *res = parsed_result;
> + int ret = -ENOTSUP;
> +#ifdef RTE_LIBRTE_I40E_PMD
> + struct rte_pmd_i40e_flow_type_mapping mapping;
> +#endif
> + unsigned int nb_item, i;
> + unsigned int pctype_list[PCTYPE_MAX];
> +
> + if (port_id_is_invalid(res->port_id, ENABLED_WARN))
> + return;
> +
> + nb_item = parse_item_list(res->pctype_list, "pctypes", PCTYPE_MAX,
> +   pctype_list, 1);

How user knows which values to fill the pctype_list?

More importantly, if this is an API call instead of user defined values,
how application should know which values to use? I 

Re: [dpdk-dev] [PATCH v3 0/4] net/softnic: sw fall-back pmd for traffic mgmt and others

[Dumitrescu, Cristian]

> -Original Message-
> From: Singh, Jasvinder
> Sent: Friday, August 11, 2017 1:49 PM
> To: dev@dpdk.org
> Cc: Dumitrescu, Cristian ; Yigit, Ferruh
> ; tho...@monjalon.net
> Subject: [PATCH v3 0/4] net/softnic: sw fall-back pmd for traffic mgmt and
> others
> 
> The SoftNIC PMD is intended to provide SW fall-back options for specific
> ethdev APIs in a generic way to the NICs not supporting those features.
> 
> Currently, the only implemented ethdev API is Traffic Management (TM),
> but other ethdev APIs such as rte_flow, traffic metering & policing, etc
> can be easily implemented.
> 
> Overview:
> * Generic: The SoftNIC PMD works with any "hard" PMD that implements
> the
>   ethdev API. It does not change the "hard" PMD in any way.
> * Creation: For any given "hard" ethdev port, the user can decide to
>   create an associated "soft" ethdev port to drive the "hard" port. The
>   "soft" port is a virtual device that can be created at app start-up
>   through EAL vdev arg or later through the virtual device API.
> * Configuration: The app explicitly decides which features are to be
>   enabled on the "soft" port and which features are still to be used from
>   the "hard" port. The app continues to explicitly configure both the
>   "hard" and the "soft" ports after the creation of the "soft" port.
> * RX/TX: The app reads packets from/writes packets to the "soft" port
>   instead of the "hard" port. The RX and TX queues of the "soft" port are
>   thread safe, as any ethdev.
> * Execution: The "soft" port is a feature-rich NIC implemented by the CPU,
>   so the run function of the "soft" port has to be executed by the CPU in
>   order to get packets moving between "hard" port and the app.
> * Meets the NFV vision: The app should be (almost) agnostic about the NIC
>   implementation (different vendors/models, HW-SW mix), the app should
> not
>   require changes to use different NICs, the app should use the same API
>   for all NICs. If a NIC does not implement a specific feature, the HW
>   should be augmented with SW to meet the functionality while still
>   preserving the same API.
> 
> Traffic Management SW fall-back overview:
> * Implements the ethdev traffic management API (rte_tm.h).
> * Based on the existing librte_sched DPDK library.
> 
> Example: Create "soft" port for "hard" port ":04:00.1", enable the TM
> feature with default settings:
>   --vdev 'net_softnic0,hard_name=:04:00.1,soft_tm=on'
> 
> Q1: Why generic name, if only TM is supported (for now)?
> A1: The intention is to have SoftNIC PMD implement many other (all?)
> ethdev APIs under a single "ideal" ethdev, hence the generic name.
> The initial motivation is TM API, but the mechanism is generic and can
> be used for many other ethdev APIs. Somebody looking to provide SW
> fall-back for other ethdev API is likely to end up inventing the same,
> hence it would be good to consolidate all under a single PMD and have
> the user explicitly enable/disable the features it needs for each
> "soft" device.
> 
> Q2: Are there any performance requirements for SoftNIC?
> A2: Yes, performance should be great/decent for every feature, otherwise
> the SW fall-back is unusable, thus useless.
> 
> Q3: Why not change the "hard" device (and keep a single device) instead of
> creating a new "soft" device (and thus having two devices)?
> A3: This is not possible with the current librte_ether ethdev
> implementation. The ethdev->dev_ops are defined as constant structure,
> so it cannot be changed per device (nor per PMD). The new ops also
> need memory space to store their context data structures, which
> requires updating the ethdev->data->dev_private of the existing
> device; at best, maybe a resize of ethdev->data->dev_private could be
> done, assuming that librte_ether will introduce a way to find out its
> size, but this cannot be done while device is running. Other side
> effects might exist, as the changes are very intrusive, plus it likely
> needs more changes in librte_ether.
> 
> Q4: Why not call the SW fall-back dev_ops directly in librte_ether for
> devices which do not support the specific feature? If the device
> supports the capability, let's call its dev_ops, otherwise call the
> SW fall-back dev_ops.
> A4: First, similar reasons to Q&A3. This fixes the need to change
> ethdev->dev_ops of the device, but it does not do anything to fix the
> other significant issue of where to store the context data structures
> needed by the SW fall-back functions (which, in this approach, are
> called implicitly by librte_ether).
> Second, the SW fall-back options should not be restricted arbitrarily
> by the librte_ether library, the decision should belong to the app.
> For example, the TM SW fall-back should not be limited to only
> librte_sched, which (like any SW fall-back) is limited to a specific
> hierarchy and fe

[dpdk-dev] [PATCH] event/sw: allow fwd and rel when out of credits

[Gage Eads]

When forwarding or releasing events, the operation would fail if the port
has 0 inflight credits and cannot acquire more, or the inflight count
exceeds the port's new event threshold.

This patch fixes that by counting the number of new events in the burst,
and applying the credit and new event threshold checks accordingly.

Signed-off-by: Gage Eads 
---
 drivers/event/sw/sw_evdev_worker.c | 14 +-
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/drivers/event/sw/sw_evdev_worker.c 
b/drivers/event/sw/sw_evdev_worker.c
index d76d3d5..b3b3b17 100644
--- a/drivers/event/sw/sw_evdev_worker.c
+++ b/drivers/event/sw/sw_evdev_worker.c
@@ -85,14 +85,18 @@ sw_event_enqueue_burst(void *port, const struct rte_event 
ev[], uint16_t num)
struct sw_port *p = port;
struct sw_evdev *sw = (void *)p->sw;
uint32_t sw_inflights = rte_atomic32_read(&sw->inflights);
-
-   if (unlikely(p->inflight_max < sw_inflights))
-   return 0;
+   int new = 0;
 
if (num > PORT_ENQUEUE_MAX_BURST_SIZE)
num = PORT_ENQUEUE_MAX_BURST_SIZE;
 
-   if (p->inflight_credits < num) {
+   for (i = 0; i < num; i++)
+   new += (ev[i].op == RTE_EVENT_OP_NEW);
+
+   if (unlikely(new > 0 && p->inflight_max < sw_inflights))
+   return 0;
+
+   if (p->inflight_credits < new) {
/* check if event enqueue brings port over max threshold */
uint32_t credit_update_quanta = sw->credit_update_quanta;
if (sw_inflights + credit_update_quanta > sw->nb_events_limit)
@@ -101,7 +105,7 @@ sw_event_enqueue_burst(void *port, const struct rte_event 
ev[], uint16_t num)
rte_atomic32_add(&sw->inflights, credit_update_quanta);
p->inflight_credits += (credit_update_quanta);
 
-   if (p->inflight_credits < num)
+   if (p->inflight_credits < new)
return 0;
}
 
-- 
2.7.4

Re: [dpdk-dev] [PATCH v2 1/2] net/i40e: fix clear xstats bug in vf port

[Wu, Jingjing]

> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Zhao1, Wei
> Sent: Friday, September 1, 2017 10:30 AM
> To: Yigit, Ferruh ; dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2 1/2] net/i40e: fix clear xstats bug in vf 
> port
> 
> Hi,  Ferruh
> 
> > -Original Message-
> > From: Yigit, Ferruh
> > Sent: Friday, September 1, 2017 12:54 AM
> > To: Zhao1, Wei ; dev@dpdk.org
> > Subject: Re: [dpdk-dev] [PATCH v2 1/2] net/i40e: fix clear xstats bug in vf
> > port
> >
> > On 8/29/2017 3:28 AM, Wei Zhao wrote:
> > > There is a bug in vf clear xstats command, it do not record the
> > > statics data in offset struct member.So, vf need to keep record of
> > > xstats data from pf and update the statics according to offset.
> > >
> > > Fixes: da61cd0849766 ("i40evf: add extended stats")
> > >
> > > Signed-off-by: Wei Zhao 
> > >
> > > ---
> > >
> > > Changes in v2:
> > >
> > >  fix patch log check warning.
> > > ---
> > >  app/test-pmd/config.c |  6 ++--
> > >  drivers/net/i40e/i40e_ethdev_vf.c | 64
> > > ++-
> > >  2 files changed, 67 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c index
> > > 3ae3e1c..14131d6 100644
> > > --- a/app/test-pmd/config.c
> > > +++ b/app/test-pmd/config.c
> > > @@ -203,8 +203,10 @@ nic_stats_display(portid_t port_id)
> > >   if (diff_cycles > 0)
> > >   diff_cycles = prev_cycles[port_id] - diff_cycles;
> > >
> > > - diff_pkts_rx = stats.ipackets - prev_pkts_rx[port_id];
> > > - diff_pkts_tx = stats.opackets - prev_pkts_tx[port_id];
> > > + diff_pkts_rx = (stats.ipackets > prev_pkts_rx[port_id]) ?
> > > + (stats.ipackets - prev_pkts_rx[port_id]) : 0;
> > > + diff_pkts_tx = (stats.opackets > prev_pkts_tx[port_id]) ?
> > > + (stats.opackets - prev_pkts_tx[port_id]) : 0;
> >
> > I guess this testpmd update is not directly related to this patch, but to 
> > protect
> > testpmd against value overflow? Can this be another patch?
> 
> Nonono, this code change is directly related to this patch, if we do not do 
> this code
> change, the
> diff_pkts_rx and diff_pkts_tx statistic data will be wrong  when the first 
> time after clear
> xstats command.
>
Yes, the fix will make the error happen, but this is the fix of the clear 
xstats issue.
You can create a separate patch for it just to make clearer.

<..>
> This bug only appear after use CLI "clear port xstats 0". So it is not easy 
> to detect this
> bug.
> After using this fix patch ,the big user who report this issue has feed back 
> it work well
> now.
> The root cause is not so complicated, when the pf which admin this vf is in 
> kernel state,
> DPDK can not
> Give pf the info to clear and update offset command, so vf can only keep 
> record the
> offset data in DPDK
> VF port locally.
> 
That was because i40evf PMD doesn't support the ops "stats_reset", but 
"xstats_reset"
Is implemented when xstats are introduced in i40vf PMD.
I think you also need to add the ops "stats_reset", testing can cover this 
basic case then.

Thanks
Jingjing

Re: [dpdk-dev] [PATCH] net/i40e: fix vsi vlan stripping

[Wu, Jingjing]

> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Zang MingJie
> Sent: Thursday, August 24, 2017 9:29 PM
> To: dev@dpdk.org
> Cc: Zang MingJie 
> Subject: [dpdk-dev] [PATCH] net/i40e: fix vsi vlan stripping
> 
> Function i40e_vsi_config_vlan_stripping doesn't strip vlan tag for vf.
> The patch should fix the problem.
> 
> Signed-off-by: Zang MingJie 
> ---
>  drivers/net/i40e/i40e_ethdev.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
> index 5f26e24a3..cd48ebbc1 100644
> --- a/drivers/net/i40e/i40e_ethdev.c
> +++ b/drivers/net/i40e/i40e_ethdev.c
> @@ -5189,7 +5189,7 @@ i40e_vsi_config_vlan_stripping(struct i40e_vsi *vsi, 
> bool on)
>   }
> 
>   if (on)
> - vlan_flags = I40E_AQ_VSI_PVLAN_EMOD_STR_BOTH;
> + vlan_flags = I40E_AQ_VSI_PVLAN_EMOD_STR;
>   else
>   vlan_flags = I40E_AQ_VSI_PVLAN_EMOD_NOTHING;
>   vsi->info.valid_sections =

How did you find this issue? In our cases, it works well.
And according to the datasheet, the EMOD_STR meaning Hide vlan but not strip 
them to descriptor.

Thanks
Jingjing

Re: [dpdk-dev] [PATCH 5/6] i40e: remove unnecessary cast of rte_memcpy

[Wu, Jingjing]

> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Stephen Hemminger
> Sent: Wednesday, August 23, 2017 11:45 PM
> To: dev@dpdk.org
> Cc: Stephen Hemminger 
> Subject: [dpdk-dev] [PATCH 5/6] i40e: remove unnecessary cast of rte_memcpy
> 
> Signed-off-by: Stephen Hemminger 
Acked-by: Jingjing Wu