[dpdk-dev] [PATCH 0/8] remove mbuf seqn

[dpdk-dev] [PATCH 0/8] remove mbuf seqn

[David Marchand]

Followup on the work started by Thomas to free some space in the mbuf
structure.
This series applies on top of Thomas v3 series [1] and drops the seqn
field that had been deprecated.

1: https://patchwork.dpdk.org/project/dpdk/list/?series=13393

-- 
David Marchand

David Marchand (8):
  event/dpaa2: remove dead code
  crypto/scheduler: remove unused internal seqn
  net/ark: remove use of seqn for debug
  reorder: switch sequence number to dynamic mbuf field
  dpaa: switch sequence number to dynamic field
  fslmc: switch sequence number to dynamic field
  event: switch sequence number to dynamic field
  mbuf: remove seqn field

 app/test-eventdev/evt_main.c                  |  3 ++
 app/test-eventdev/test_order_common.c         |  2 +-
 app/test-eventdev/test_order_common.h         |  5 ++-
 app/test/test_reorder.c                       |  8 ++--
 doc/guides/rel_notes/deprecation.rst          |  1 -
 doc/guides/rel_notes/release_20_11.rst        |  3 ++
 drivers/bus/dpaa/dpaa_bus.c                   | 16 ++++++++
 drivers/bus/dpaa/rte_dpaa_bus.h               | 28 ++++++++++++++
 drivers/bus/dpaa/version.map                  |  1 +
 drivers/bus/fslmc/fslmc_bus.c                 | 17 +++++++++
 drivers/bus/fslmc/rte_fslmc.h                 | 23 +++++++++++
 drivers/bus/fslmc/version.map                 |  1 +
 drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c   | 18 +++++----
 drivers/crypto/dpaa_sec/dpaa_sec.c            |  6 +--
 .../crypto/scheduler/scheduler_pmd_private.h  |  1 -
 drivers/event/dpaa/dpaa_eventdev.c            |  6 +--
 drivers/event/dpaa2/dpaa2_eventdev.c          |  9 +++--
 drivers/event/dpaa2/dpaa2_eventdev_selftest.c | 17 ++-------
 drivers/event/octeontx/ssovf_evdev_selftest.c | 32 ++++++++--------
 drivers/event/octeontx2/otx2_evdev_selftest.c | 31 +++++++--------
 drivers/event/opdl/opdl_test.c                |  8 ++--
 drivers/event/sw/sw_evdev_selftest.c          | 34 +++++++++--------
 drivers/mempool/dpaa2/dpaa2_hw_mempool.h      |  2 -
 drivers/net/ark/ark_ethdev_rx.c               |  8 +---
 drivers/net/dpaa/dpaa_ethdev.h                |  7 ----
 drivers/net/dpaa/dpaa_rxtx.c                  |  6 +--
 drivers/net/dpaa2/dpaa2_rxtx.c                | 28 +++++++-------
 examples/packet_ordering/main.c               |  2 +-
 lib/librte_eventdev/rte_eventdev.c            | 21 +++++++++-
 lib/librte_eventdev/rte_eventdev.h            | 38 ++++++++++++++++---
 lib/librte_eventdev/version.map               |  2 +
 lib/librte_mbuf/rte_mbuf_core.h               | 15 +++-----
 lib/librte_reorder/rte_reorder.c              | 23 +++++++++--
 lib/librte_reorder/rte_reorder.h              | 21 ++++++++++
 lib/librte_reorder/version.map                |  6 +++
 35 files changed, 307 insertions(+), 142 deletions(-)

-- 
2.23.0

[dpdk-dev] [PATCH 1/8] event/dpaa2: remove dead code

[David Marchand]

This code has never been used since introduction.

Fixes: 653242c3375a ("event/dpaa2: add self test")

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/event/dpaa2/dpaa2_eventdev_selftest.c | 13 +------------
 1 file changed, 1 insertion(+), 12 deletions(-)

diff --git a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
index b1f3891484..5447db8a8a 100644
--- a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
+++ b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
@@ -47,17 +47,6 @@ struct event_attr {
 	uint8_t seq;
 };
 
-static uint32_t seqn_list_index;
-static int seqn_list[NUM_PACKETS];
-
-static void
-seqn_list_init(void)
-{
-	RTE_BUILD_BUG_ON(NUM_PACKETS < MAX_EVENTS);
-	memset(seqn_list, 0, sizeof(seqn_list));
-	seqn_list_index = 0;
-}
-
 struct test_core_param {
 	rte_atomic32_t *total_events;
 	uint64_t dequeue_tmo_ticks;
@@ -516,7 +505,7 @@ launch_workers_and_wait(int (*main_worker)(void *),
 		return 0;
 
 	rte_atomic32_set(&atomic_total_events, total_events);
-	seqn_list_init();
+	RTE_BUILD_BUG_ON(NUM_PACKETS < MAX_EVENTS);
 
 	param = malloc(sizeof(struct test_core_param) * nb_workers);
 	if (!param)
-- 
2.23.0

[dpdk-dev] [PATCH 2/8] crypto/scheduler: remove unused internal seqn

[David Marchand]

This field has been left behind after dropping its use.

Fixes: 8a48e039432b ("crypto/scheduler: optimize crypto op ordering")

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/crypto/scheduler/scheduler_pmd_private.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/crypto/scheduler/scheduler_pmd_private.h b/drivers/crypto/scheduler/scheduler_pmd_private.h
index adb4eb0632..4d33b9ab44 100644
--- a/drivers/crypto/scheduler/scheduler_pmd_private.h
+++ b/drivers/crypto/scheduler/scheduler_pmd_private.h
@@ -59,7 +59,6 @@ struct scheduler_qp_ctx {
 	uint32_t max_nb_objs;
 
 	struct rte_ring *order_ring;
-	uint32_t seqn;
 } __rte_cache_aligned;
 
 
-- 
2.23.0

[dpdk-dev] [PATCH 3/8] net/ark: remove use of seqn for debug

[David Marchand]

The seqn mbuf field is deprecated.
It is currently hacked for debug purpose, it could be changed to a
dynamic field but I see little value in the debug info it offers.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/net/ark/ark_ethdev_rx.c | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ark/ark_ethdev_rx.c b/drivers/net/ark/ark_ethdev_rx.c
index c5788498b3..cc0881b77c 100644
--- a/drivers/net/ark/ark_ethdev_rx.c
+++ b/drivers/net/ark/ark_ethdev_rx.c
@@ -302,8 +302,6 @@ eth_ark_recv_pkts(void *rx_queue,
 				mbuf->pkt_len = 63;
 				meta->pkt_len = 63;
 			}
-			/* seqn is only set under debug */
-			mbuf->seqn = cons_index;
 		}
 
 		if (unlikely(meta->pkt_len > ARK_RX_MAX_NOCHAIN))
@@ -360,8 +358,6 @@ eth_ark_rx_jumbo(struct ark_rx_queue *queue,
 		mbuf_prev = mbuf;
 		mbuf->data_len = data_len;
 		mbuf->data_off = 0;
-		if (ARK_DEBUG_CORE)
-			mbuf->seqn = cons_index;	/* for debug only */
 
 		cons_index += 1;
 	}
@@ -667,8 +663,8 @@ dump_mbuf_data(struct rte_mbuf *mbuf, uint16_t lo, uint16_t hi)
 {
 	uint16_t i, j;
 
-	ARK_PMD_LOG(DEBUG, " MBUF: %p len %d, off: %d, seq: %" PRIU32 "\n",
-		    mbuf, mbuf->pkt_len, mbuf->data_off, mbuf->seqn);
+	ARK_PMD_LOG(DEBUG, " MBUF: %p len %d, off: %d\n",
+		    mbuf, mbuf->pkt_len, mbuf->data_off);
 	for (i = lo; i < hi; i += 16) {
 		uint8_t *dp = RTE_PTR_ADD(mbuf->buf_addr, i);
 
-- 
2.23.0

Re: [dpdk-dev] [PATCH 3/8] net/ark: remove use of seqn for debug

[Ed Czeck]

Thank you

Acked-by: Ed Czeck <ed.czeck@atomicrules.com>

On Tue, Oct 27, 2020 at 6:13 PM David Marchand
<david.marchand@redhat.com> wrote:
>
> The seqn mbuf field is deprecated.
> It is currently hacked for debug purpose, it could be changed to a
> dynamic field but I see little value in the debug info it offers.
>
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> ---
>  drivers/net/ark/ark_ethdev_rx.c | 8 ++------
>  1 file changed, 2 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/net/ark/ark_ethdev_rx.c b/drivers/net/ark/ark_ethdev_rx.c
> index c5788498b3..cc0881b77c 100644
> --- a/drivers/net/ark/ark_ethdev_rx.c
> +++ b/drivers/net/ark/ark_ethdev_rx.c
> @@ -302,8 +302,6 @@ eth_ark_recv_pkts(void *rx_queue,
>                                 mbuf->pkt_len = 63;
>                                 meta->pkt_len = 63;
>                         }
> -                       /* seqn is only set under debug */
> -                       mbuf->seqn = cons_index;
>                 }
>
>                 if (unlikely(meta->pkt_len > ARK_RX_MAX_NOCHAIN))
> @@ -360,8 +358,6 @@ eth_ark_rx_jumbo(struct ark_rx_queue *queue,
>                 mbuf_prev = mbuf;
>                 mbuf->data_len = data_len;
>                 mbuf->data_off = 0;
> -               if (ARK_DEBUG_CORE)
> -                       mbuf->seqn = cons_index;        /* for debug only */
>
>                 cons_index += 1;
>         }
> @@ -667,8 +663,8 @@ dump_mbuf_data(struct rte_mbuf *mbuf, uint16_t lo, uint16_t hi)
>  {
>         uint16_t i, j;
>
> -       ARK_PMD_LOG(DEBUG, " MBUF: %p len %d, off: %d, seq: %" PRIU32 "\n",
> -                   mbuf, mbuf->pkt_len, mbuf->data_off, mbuf->seqn);
> +       ARK_PMD_LOG(DEBUG, " MBUF: %p len %d, off: %d\n",
> +                   mbuf, mbuf->pkt_len, mbuf->data_off);
>         for (i = lo; i < hi; i += 16) {
>                 uint8_t *dp = RTE_PTR_ADD(mbuf->buf_addr, i);
>
> --
> 2.23.0
>

[dpdk-dev] [PATCH 4/8] reorder: switch sequence number to dynamic mbuf field

[David Marchand]

The reorder library used sequence numbers stored in the deprecated field
seqn.
It is moved to a dynamic mbuf field in order to allow removal of seqn.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 app/test/test_reorder.c          |  8 ++++----
 examples/packet_ordering/main.c  |  2 +-
 lib/librte_reorder/rte_reorder.c | 23 ++++++++++++++++++++---
 lib/librte_reorder/rte_reorder.h | 21 +++++++++++++++++++++
 lib/librte_reorder/version.map   |  6 ++++++
 5 files changed, 52 insertions(+), 8 deletions(-)

diff --git a/app/test/test_reorder.c b/app/test/test_reorder.c
index 58fa9c71b5..1c4226da65 100644
--- a/app/test/test_reorder.c
+++ b/app/test/test_reorder.c
@@ -149,7 +149,7 @@ test_reorder_insert(void)
 	for (i = 0; i < num_bufs; i++) {
 		bufs[i] = rte_pktmbuf_alloc(p);
 		TEST_ASSERT_NOT_NULL(bufs[i], "Packet allocation failed\n");
-		bufs[i]->seqn = i;
+		*rte_reorder_seqn(bufs[i]) = i;
 	}
 
 	/* This should fill up order buffer:
@@ -183,7 +183,7 @@ test_reorder_insert(void)
 	bufs[4] = NULL;
 
 	/* early packet from current sequence window - full ready buffer */
-	bufs[5]->seqn = 2 * size;
+	*rte_reorder_seqn(bufs[5]) = 2 * size;
 	ret = rte_reorder_insert(b, bufs[5]);
 	if (!((ret == -1) && (rte_errno == ENOSPC))) {
 		printf("%s:%d: No error inserting early packet with full ready buffer\n",
@@ -194,7 +194,7 @@ test_reorder_insert(void)
 	bufs[5] = NULL;
 
 	/* late packet */
-	bufs[6]->seqn = 3 * size;
+	*rte_reorder_seqn(bufs[6]) = 3 * size;
 	ret = rte_reorder_insert(b, bufs[6]);
 	if (!((ret == -1) && (rte_errno == ERANGE))) {
 		printf("%s:%d: No error inserting late packet with seqn:"
@@ -250,7 +250,7 @@ test_reorder_drain(void)
 	for (i = 0; i < num_bufs; i++) {
 		bufs[i] = rte_pktmbuf_alloc(p);
 		TEST_ASSERT_NOT_NULL(bufs[i], "Packet allocation failed\n");
-		bufs[i]->seqn = i;
+		*rte_reorder_seqn(bufs[i]) = i;
 	}
 
 	/* Insert packet with seqn 1:
diff --git a/examples/packet_ordering/main.c b/examples/packet_ordering/main.c
index a79d77a321..4bea1982d5 100644
--- a/examples/packet_ordering/main.c
+++ b/examples/packet_ordering/main.c
@@ -451,7 +451,7 @@ rx_thread(struct rte_ring *ring_out)
 
 				/* mark sequence number */
 				for (i = 0; i < nb_rx_pkts; )
-					pkts[i++]->seqn = seqn++;
+					*rte_reorder_seqn(pkts[i++]) = seqn++;
 
 				/* enqueue to rx_to_workers ring */
 				ret = rte_ring_enqueue_burst(ring_out,
diff --git a/lib/librte_reorder/rte_reorder.c b/lib/librte_reorder/rte_reorder.c
index 3c9f0e2d08..9445853b79 100644
--- a/lib/librte_reorder/rte_reorder.c
+++ b/lib/librte_reorder/rte_reorder.c
@@ -8,6 +8,7 @@
 #include <rte_string_fns.h>
 #include <rte_log.h>
 #include <rte_mbuf.h>
+#include <rte_mbuf_dyn.h>
 #include <rte_eal_memconfig.h>
 #include <rte_errno.h>
 #include <rte_malloc.h>
@@ -29,6 +30,9 @@ EAL_REGISTER_TAILQ(rte_reorder_tailq)
 /* Macros for printing using RTE_LOG */
 #define RTE_LOGTYPE_REORDER	RTE_LOGTYPE_USER1
 
+#define RTE_REORDER_SEQN_DYNFIELD_NAME "rte_reorder_seqn_dynfield"
+int rte_reorder_seqn_dynfield_offset = -1;
+
 /* A generic circular buffer */
 struct cir_buffer {
 	unsigned int size;   /**< Number of entries that can be stored */
@@ -103,6 +107,11 @@ rte_reorder_create(const char *name, unsigned socket_id, unsigned int size)
 	struct rte_reorder_list *reorder_list;
 	const unsigned int bufsize = sizeof(struct rte_reorder_buffer) +
 					(2 * size * sizeof(struct rte_mbuf *));
+	static const struct rte_mbuf_dynfield reorder_seqn_dynfield_desc = {
+		.name = RTE_REORDER_SEQN_DYNFIELD_NAME,
+		.size = sizeof(rte_reorder_seqn_t),
+		.align = __alignof__(rte_reorder_seqn_t),
+	};
 
 	reorder_list = RTE_TAILQ_CAST(rte_reorder_tailq.head, rte_reorder_list);
 
@@ -120,6 +129,14 @@ rte_reorder_create(const char *name, unsigned socket_id, unsigned int size)
 		return NULL;
 	}
 
+	rte_reorder_seqn_dynfield_offset =
+		rte_mbuf_dynfield_register(&reorder_seqn_dynfield_desc);
+	if (rte_reorder_seqn_dynfield_offset < 0) {
+		RTE_LOG(ERR, REORDER, "Failed to register mbuf field for reorder sequence number\n");
+		rte_errno = ENOMEM;
+		return NULL;
+	}
+
 	rte_mcfg_tailq_write_lock();
 
 	/* guarantee there's no existing */
@@ -310,7 +327,7 @@ rte_reorder_insert(struct rte_reorder_buffer *b, struct rte_mbuf *mbuf)
 
 	order_buf = &b->order_buf;
 	if (!b->is_initialized) {
-		b->min_seqn = mbuf->seqn;
+		b->min_seqn = *rte_reorder_seqn(mbuf);
 		b->is_initialized = 1;
 	}
 
@@ -322,7 +339,7 @@ rte_reorder_insert(struct rte_reorder_buffer *b, struct rte_mbuf *mbuf)
 	 *	mbuf_seqn = 0x0010
 	 *	offset    = 0x0010 - 0xFFFD = 0x13
 	 */
-	offset = mbuf->seqn - b->min_seqn;
+	offset = *rte_reorder_seqn(mbuf) - b->min_seqn;
 
 	/*
 	 * action to take depends on offset.
@@ -352,7 +369,7 @@ rte_reorder_insert(struct rte_reorder_buffer *b, struct rte_mbuf *mbuf)
 			rte_errno = ENOSPC;
 			return -1;
 		}
-		offset = mbuf->seqn - b->min_seqn;
+		offset = *rte_reorder_seqn(mbuf) - b->min_seqn;
 		position = (order_buf->head + offset) & order_buf->mask;
 		order_buf->entries[position] = mbuf;
 	} else {
diff --git a/lib/librte_reorder/rte_reorder.h b/lib/librte_reorder/rte_reorder.h
index 6d39710088..ce6cd598e9 100644
--- a/lib/librte_reorder/rte_reorder.h
+++ b/lib/librte_reorder/rte_reorder.h
@@ -16,6 +16,7 @@
  */
 
 #include <rte_mbuf.h>
+#include <rte_mbuf_dyn.h>
 
 #ifdef __cplusplus
 extern "C" {
@@ -23,6 +24,26 @@ extern "C" {
 
 struct rte_reorder_buffer;
 
+typedef uint32_t rte_reorder_seqn_t;
+extern int rte_reorder_seqn_dynfield_offset;
+
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ *
+ * Read reorder sequence number from mbuf.
+ *
+ * @param mbuf Structure to read from.
+ * @return pointer to reorder sequence number.
+ */
+__rte_experimental
+static inline rte_reorder_seqn_t *
+rte_reorder_seqn(const struct rte_mbuf *mbuf)
+{
+	return RTE_MBUF_DYNFIELD(mbuf, rte_reorder_seqn_dynfield_offset,
+		rte_reorder_seqn_t *);
+}
+
 /**
  * Create a new reorder buffer instance
  *
diff --git a/lib/librte_reorder/version.map b/lib/librte_reorder/version.map
index 8c0220d324..d902a7fa12 100644
--- a/lib/librte_reorder/version.map
+++ b/lib/librte_reorder/version.map
@@ -11,3 +11,9 @@ DPDK_21 {
 
 	local: *;
 };
+
+EXPERIMENTAL {
+	global:
+
+	rte_reorder_seqn_dynfield_offset;
+};
-- 
2.23.0

[dpdk-dev] [PATCH 5/8] dpaa: switch sequence number to dynamic field

[David Marchand]

The dpaa drivers have been hacking the deprecated field seqn for
internal features.
It is moved to a dynamic mbuf field in order to allow removal of seqn.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/bus/dpaa/dpaa_bus.c        | 16 ++++++++++++++++
 drivers/bus/dpaa/rte_dpaa_bus.h    | 28 ++++++++++++++++++++++++++++
 drivers/bus/dpaa/version.map       |  1 +
 drivers/crypto/dpaa_sec/dpaa_sec.c |  6 +++---
 drivers/event/dpaa/dpaa_eventdev.c |  6 +++---
 drivers/net/dpaa/dpaa_ethdev.h     |  7 -------
 drivers/net/dpaa/dpaa_rxtx.c       |  6 +++---
 7 files changed, 54 insertions(+), 16 deletions(-)

diff --git a/drivers/bus/dpaa/dpaa_bus.c b/drivers/bus/dpaa/dpaa_bus.c
index c94c72106f..ece6a4c424 100644
--- a/drivers/bus/dpaa/dpaa_bus.c
+++ b/drivers/bus/dpaa/dpaa_bus.c
@@ -32,6 +32,7 @@
 #include <rte_ring.h>
 #include <rte_bus.h>
 #include <rte_mbuf_pool_ops.h>
+#include <rte_mbuf_dyn.h>
 
 #include <dpaa_of.h>
 #include <rte_dpaa_bus.h>
@@ -55,6 +56,9 @@ unsigned int dpaa_svr_family;
 
 RTE_DEFINE_PER_LCORE(struct dpaa_portal *, dpaa_io);
 
+#define DPAA_SEQN_DYNFIELD_NAME "dpaa_seqn_dynfield"
+int dpaa_seqn_dynfield_offset = -1;
+
 struct fm_eth_port_cfg *
 dpaa_get_eth_port_cfg(int dev_id)
 {
@@ -251,6 +255,11 @@ dpaa_clean_device_list(void)
 
 int rte_dpaa_portal_init(void *arg)
 {
+	static const struct rte_mbuf_dynfield dpaa_seqn_dynfield_desc = {
+		.name = DPAA_SEQN_DYNFIELD_NAME,
+		.size = sizeof(dpaa_seqn_t),
+		.align = __alignof__(dpaa_seqn_t),
+	};
 	unsigned int cpu, lcore = rte_lcore_id();
 	int ret;
 
@@ -264,6 +273,13 @@ int rte_dpaa_portal_init(void *arg)
 
 	cpu = rte_lcore_to_cpu_id(lcore);
 
+	dpaa_seqn_dynfield_offset =
+		rte_mbuf_dynfield_register(&dpaa_seqn_dynfield_desc);
+	if (dpaa_seqn_dynfield_offset < 0) {
+		DPAA_BUS_LOG(ERR, "Failed to register mbuf field for dpaa sequence number\n");
+		return -rte_errno;
+	}
+
 	/* Initialise bman thread portals */
 	ret = bman_thread_init();
 	if (ret) {
diff --git a/drivers/bus/dpaa/rte_dpaa_bus.h b/drivers/bus/dpaa/rte_dpaa_bus.h
index fdaa63a09b..5b8c44df72 100644
--- a/drivers/bus/dpaa/rte_dpaa_bus.h
+++ b/drivers/bus/dpaa/rte_dpaa_bus.h
@@ -7,6 +7,7 @@
 #define __RTE_DPAA_BUS_H__
 
 #include <rte_bus.h>
+#include <rte_mbuf_dyn.h>
 #include <rte_mempool.h>
 #include <dpaax_iova_table.h>
 
@@ -16,6 +17,33 @@
 #include <fsl_bman.h>
 #include <netcfg.h>
 
+/* This sequence number field is used to store event entry index for
+ * driver specific usage. For parallel mode queues, invalid
+ * index will be set and for atomic mode queues, valid value
+ * ranging from 1 to 16.
+ */
+#define DPAA_INVALID_MBUF_SEQN  0
+
+typedef uint32_t dpaa_seqn_t;
+extern int dpaa_seqn_dynfield_offset;
+
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ *
+ * Read dpaa sequence number from mbuf.
+ *
+ * @param mbuf Structure to read from.
+ * @return pointer to dpaa sequence number.
+ */
+__rte_experimental
+static inline dpaa_seqn_t *
+dpaa_seqn(const struct rte_mbuf *mbuf)
+{
+	return RTE_MBUF_DYNFIELD(mbuf, dpaa_seqn_dynfield_offset,
+		dpaa_seqn_t *);
+}
+
 #define DPAA_MEMPOOL_OPS_NAME	"dpaa"
 
 #define DEV_TO_DPAA_DEVICE(ptr)	\
diff --git a/drivers/bus/dpaa/version.map b/drivers/bus/dpaa/version.map
index 9bd2601213..fe4f9ac5aa 100644
--- a/drivers/bus/dpaa/version.map
+++ b/drivers/bus/dpaa/version.map
@@ -14,6 +14,7 @@ INTERNAL {
 	dpaa_get_qm_channel_pool;
 	dpaa_get_link_status;
 	dpaa_restart_link_autoneg;
+	dpaa_seqn_dynfield_offset;
 	dpaa_update_link_speed;
 	dpaa_intr_disable;
 	dpaa_intr_enable;
diff --git a/drivers/crypto/dpaa_sec/dpaa_sec.c b/drivers/crypto/dpaa_sec/dpaa_sec.c
index 55f457ac9a..44c742738f 100644
--- a/drivers/crypto/dpaa_sec/dpaa_sec.c
+++ b/drivers/crypto/dpaa_sec/dpaa_sec.c
@@ -1721,8 +1721,8 @@ dpaa_sec_enqueue_burst(void *qp, struct rte_crypto_op **ops,
 				DPAA_SEC_BURST : nb_ops;
 		for (loop = 0; loop < frames_to_send; loop++) {
 			op = *(ops++);
-			if (op->sym->m_src->seqn != 0) {
-				index = op->sym->m_src->seqn - 1;
+			if (*dpaa_seqn(op->sym->m_src) != 0) {
+				index = *dpaa_seqn(op->sym->m_src) - 1;
 				if (DPAA_PER_LCORE_DQRR_HELD & (1 << index)) {
 					/* QM_EQCR_DCA_IDXMASK = 0x0f */
 					flags[loop] = ((index & 0x0f) << 8);
@@ -3212,7 +3212,7 @@ dpaa_sec_process_atomic_event(void *event,
 	DPAA_PER_LCORE_DQRR_HELD |= 1 << index;
 	DPAA_PER_LCORE_DQRR_MBUF(index) = ctx->op->sym->m_src;
 	ev->impl_opaque = index + 1;
-	ctx->op->sym->m_src->seqn = (uint32_t)index + 1;
+	*dpaa_seqn(ctx->op->sym->m_src) = (uint32_t)index + 1;
 	*bufs = (void *)ctx->op;
 
 	rte_mempool_put(ctx->ctx_pool, (void *)ctx);
diff --git a/drivers/event/dpaa/dpaa_eventdev.c b/drivers/event/dpaa/dpaa_eventdev.c
index 07cd079768..01ddd0eb63 100644
--- a/drivers/event/dpaa/dpaa_eventdev.c
+++ b/drivers/event/dpaa/dpaa_eventdev.c
@@ -99,7 +99,7 @@ dpaa_event_enqueue_burst(void *port, const struct rte_event ev[],
 		case RTE_EVENT_OP_RELEASE:
 			qman_dca_index(ev[i].impl_opaque, 0);
 			mbuf = DPAA_PER_LCORE_DQRR_MBUF(i);
-			mbuf->seqn = DPAA_INVALID_MBUF_SEQN;
+			*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
 			DPAA_PER_LCORE_DQRR_HELD &= ~(1 << i);
 			DPAA_PER_LCORE_DQRR_SIZE--;
 			break;
@@ -206,7 +206,7 @@ dpaa_event_dequeue_burst(void *port, struct rte_event ev[],
 		if (DPAA_PER_LCORE_DQRR_HELD & (1 << i)) {
 			qman_dca_index(i, 0);
 			mbuf = DPAA_PER_LCORE_DQRR_MBUF(i);
-			mbuf->seqn = DPAA_INVALID_MBUF_SEQN;
+			*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
 			DPAA_PER_LCORE_DQRR_HELD &= ~(1 << i);
 			DPAA_PER_LCORE_DQRR_SIZE--;
 		}
@@ -276,7 +276,7 @@ dpaa_event_dequeue_burst_intr(void *port, struct rte_event ev[],
 		if (DPAA_PER_LCORE_DQRR_HELD & (1 << i)) {
 			qman_dca_index(i, 0);
 			mbuf = DPAA_PER_LCORE_DQRR_MBUF(i);
-			mbuf->seqn = DPAA_INVALID_MBUF_SEQN;
+			*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
 			DPAA_PER_LCORE_DQRR_HELD &= ~(1 << i);
 			DPAA_PER_LCORE_DQRR_SIZE--;
 		}
diff --git a/drivers/net/dpaa/dpaa_ethdev.h b/drivers/net/dpaa/dpaa_ethdev.h
index 1b8e120e8f..659bceb467 100644
--- a/drivers/net/dpaa/dpaa_ethdev.h
+++ b/drivers/net/dpaa/dpaa_ethdev.h
@@ -22,13 +22,6 @@
 #define DPAA_MBUF_HW_ANNOTATION		64
 #define DPAA_FD_PTA_SIZE		64
 
-/* mbuf->seqn will be used to store event entry index for
- * driver specific usage. For parallel mode queues, invalid
- * index will be set and for atomic mode queues, valid value
- * ranging from 1 to 16.
- */
-#define DPAA_INVALID_MBUF_SEQN  0
-
 /* we will re-use the HEADROOM for annotation in RX */
 #define DPAA_HW_BUF_RESERVE	0
 #define DPAA_PACKET_LAYOUT_ALIGN	64
diff --git a/drivers/net/dpaa/dpaa_rxtx.c b/drivers/net/dpaa/dpaa_rxtx.c
index e4f012c233..e2459d9b99 100644
--- a/drivers/net/dpaa/dpaa_rxtx.c
+++ b/drivers/net/dpaa/dpaa_rxtx.c
@@ -649,7 +649,7 @@ dpaa_rx_cb_parallel(void *event,
 	ev->queue_id = fq->ev.queue_id;
 	ev->priority = fq->ev.priority;
 	ev->impl_opaque = (uint8_t)DPAA_INVALID_MBUF_SEQN;
-	mbuf->seqn = DPAA_INVALID_MBUF_SEQN;
+	*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
 	*bufs = mbuf;
 
 	return qman_cb_dqrr_consume;
@@ -683,7 +683,7 @@ dpaa_rx_cb_atomic(void *event,
 	DPAA_PER_LCORE_DQRR_HELD |= 1 << index;
 	DPAA_PER_LCORE_DQRR_MBUF(index) = mbuf;
 	ev->impl_opaque = index + 1;
-	mbuf->seqn = (uint32_t)index + 1;
+	*dpaa_seqn(mbuf) = (uint32_t)index + 1;
 	*bufs = mbuf;
 
 	return qman_cb_dqrr_defer;
@@ -1078,7 +1078,7 @@ dpaa_eth_queue_tx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
 			if (dpaa_svr_family == SVR_LS1043A_FAMILY &&
 					(mbuf->data_off & 0x7F) != 0x0)
 				realloc_mbuf = 1;
-			seqn = mbuf->seqn;
+			seqn = *dpaa_seqn(mbuf);
 			if (seqn != DPAA_INVALID_MBUF_SEQN) {
 				index = seqn - 1;
 				if (DPAA_PER_LCORE_DQRR_HELD & (1 << index)) {
-- 
2.23.0

[dpdk-dev] [PATCH 6/8] fslmc: switch sequence number to dynamic field

[David Marchand]

The dpaa2 drivers have been hacking the deprecated field seqn for
internal features.
It is moved to a dynamic mbuf field in order to allow removal of seqn.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/bus/fslmc/fslmc_bus.c                 | 17 +++++++++++
 drivers/bus/fslmc/rte_fslmc.h                 | 23 +++++++++++++++
 drivers/bus/fslmc/version.map                 |  1 +
 drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c   | 18 ++++++------
 drivers/event/dpaa2/dpaa2_eventdev.c          |  9 +++---
 drivers/event/dpaa2/dpaa2_eventdev_selftest.c |  4 ++-
 drivers/mempool/dpaa2/dpaa2_hw_mempool.h      |  2 --
 drivers/net/dpaa2/dpaa2_rxtx.c                | 28 +++++++++----------
 8 files changed, 73 insertions(+), 29 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index beb3dd008f..db93669628 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -14,6 +14,7 @@
 #include <rte_devargs.h>
 #include <rte_memcpy.h>
 #include <rte_ethdev_driver.h>
+#include <rte_mbuf_dyn.h>
 
 #include <rte_fslmc.h>
 #include <fslmc_vfio.h>
@@ -27,6 +28,9 @@
 struct rte_fslmc_bus rte_fslmc_bus;
 uint8_t dpaa2_virt_mode;
 
+#define DPAA2_SEQN_DYNFIELD_NAME "dpaa2_seqn_dynfield"
+int dpaa2_seqn_dynfield_offset = -1;
+
 uint32_t
 rte_fslmc_get_device_count(enum rte_dpaa2_dev_type device_type)
 {
@@ -374,9 +378,22 @@ rte_fslmc_probe(void)
 	struct rte_dpaa2_device *dev;
 	struct rte_dpaa2_driver *drv;
 
+	static const struct rte_mbuf_dynfield dpaa2_seqn_dynfield_desc = {
+		.name = DPAA2_SEQN_DYNFIELD_NAME,
+		.size = sizeof(dpaa2_seqn_t),
+		.align = __alignof__(dpaa2_seqn_t),
+	};
+
 	if (TAILQ_EMPTY(&rte_fslmc_bus.device_list))
 		return 0;
 
+	dpaa2_seqn_dynfield_offset =
+		rte_mbuf_dynfield_register(&dpaa2_seqn_dynfield_desc);
+	if (dpaa2_seqn_dynfield_offset < 0) {
+		DPAA2_BUS_ERR("Failed to register mbuf field for dpaa sequence number");
+		return 0;
+	}
+
 	ret = fslmc_vfio_setup_group();
 	if (ret) {
 		DPAA2_BUS_ERR("Unable to setup VFIO %d", ret);
diff --git a/drivers/bus/fslmc/rte_fslmc.h b/drivers/bus/fslmc/rte_fslmc.h
index 80873fffc9..37ca6886f2 100644
--- a/drivers/bus/fslmc/rte_fslmc.h
+++ b/drivers/bus/fslmc/rte_fslmc.h
@@ -32,11 +32,34 @@ extern "C" {
 #include <rte_bus.h>
 #include <rte_tailq.h>
 #include <rte_devargs.h>
+#include <rte_mbuf.h>
+#include <rte_mbuf_dyn.h>
 
 #include <fslmc_vfio.h>
 
 #define FSLMC_OBJECT_MAX_LEN 32   /**< Length of each device on bus */
 
+#define DPAA2_INVALID_MBUF_SEQN        0
+
+typedef uint32_t dpaa2_seqn_t;
+extern int dpaa2_seqn_dynfield_offset;
+
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ *
+ * Read dpaa2 sequence number from mbuf.
+ *
+ * @param mbuf Structure to read from.
+ * @return pointer to dpaa2 sequence number.
+ */
+__rte_experimental
+static inline dpaa2_seqn_t *
+dpaa2_seqn(const struct rte_mbuf *mbuf)
+{
+	return RTE_MBUF_DYNFIELD(mbuf, dpaa2_seqn_dynfield_offset,
+		dpaa2_seqn_t *);
+}
 
 /** Device driver supports link state interrupt */
 #define RTE_DPAA2_DRV_INTR_LSC	0x0008
diff --git a/drivers/bus/fslmc/version.map b/drivers/bus/fslmc/version.map
index b169f5228a..f44c1a7988 100644
--- a/drivers/bus/fslmc/version.map
+++ b/drivers/bus/fslmc/version.map
@@ -19,6 +19,7 @@ INTERNAL {
 	dpaa2_free_eq_descriptors;
 	dpaa2_get_mcp_ptr;
 	dpaa2_io_portal;
+	dpaa2_seqn_dynfield_offset;
 	dpaa2_svr_family;
 	dpaa2_virt_mode;
 	dpbp_disable;
diff --git a/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c b/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
index afcd6bd063..ce1d50ce77 100644
--- a/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
+++ b/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
@@ -1472,13 +1472,15 @@ dpaa2_sec_enqueue_burst(void *qp, struct rte_crypto_op **ops,
 			dpaa2_eqcr_size : nb_ops;
 
 		for (loop = 0; loop < frames_to_send; loop++) {
-			if ((*ops)->sym->m_src->seqn) {
-			 uint8_t dqrr_index = (*ops)->sym->m_src->seqn - 1;
-
-			 flags[loop] = QBMAN_ENQUEUE_FLAG_DCA | dqrr_index;
-			 DPAA2_PER_LCORE_DQRR_SIZE--;
-			 DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dqrr_index);
-			 (*ops)->sym->m_src->seqn = DPAA2_INVALID_MBUF_SEQN;
+			if (*dpaa2_seqn((*ops)->sym->m_src)) {
+				uint8_t dqrr_index =
+					*dpaa2_seqn((*ops)->sym->m_src) - 1;
+
+				flags[loop] = QBMAN_ENQUEUE_FLAG_DCA | dqrr_index;
+				DPAA2_PER_LCORE_DQRR_SIZE--;
+				DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dqrr_index);
+				*dpaa2_seqn((*ops)->sym->m_src) =
+					DPAA2_INVALID_MBUF_SEQN;
 			}
 
 			/*Clear the unused FD fields before sending*/
@@ -3714,7 +3716,7 @@ dpaa2_sec_process_atomic_event(struct qbman_swp *swp __rte_unused,
 
 	ev->event_ptr = sec_fd_to_mbuf(fd);
 	dqrr_index = qbman_get_dqrr_idx(dq);
-	crypto_op->sym->m_src->seqn = dqrr_index + 1;
+	*dpaa2_seqn(crypto_op->sym->m_src) = dqrr_index + 1;
 	DPAA2_PER_LCORE_DQRR_SIZE++;
 	DPAA2_PER_LCORE_DQRR_HELD |= 1 << dqrr_index;
 	DPAA2_PER_LCORE_DQRR_MBUF(dqrr_index) = crypto_op->sym->m_src;
diff --git a/drivers/event/dpaa2/dpaa2_eventdev.c b/drivers/event/dpaa2/dpaa2_eventdev.c
index 95f03c8b9e..eeb2494bd0 100644
--- a/drivers/event/dpaa2/dpaa2_eventdev.c
+++ b/drivers/event/dpaa2/dpaa2_eventdev.c
@@ -131,8 +131,9 @@ dpaa2_eventdev_enqueue_burst(void *port, const struct rte_event ev[],
 			qbman_eq_desc_set_response(&eqdesc[loop], 0, 0);
 
 			if (event->sched_type == RTE_SCHED_TYPE_ATOMIC
-				&& event->mbuf->seqn) {
-				uint8_t dqrr_index = event->mbuf->seqn - 1;
+				&& *dpaa2_seqn(event->mbuf)) {
+				uint8_t dqrr_index =
+					*dpaa2_seqn(event->mbuf) - 1;
 
 				qbman_eq_desc_set_dca(&eqdesc[loop], 1,
 						      dqrr_index, 0);
@@ -249,7 +250,7 @@ static void dpaa2_eventdev_process_atomic(struct qbman_swp *swp,
 
 	rte_memcpy(ev, ev_temp, sizeof(struct rte_event));
 	rte_free(ev_temp);
-	ev->mbuf->seqn = dqrr_index + 1;
+	*dpaa2_seqn(ev->mbuf) = dqrr_index + 1;
 	DPAA2_PER_LCORE_DQRR_SIZE++;
 	DPAA2_PER_LCORE_DQRR_HELD |= 1 << dqrr_index;
 	DPAA2_PER_LCORE_DQRR_MBUF(dqrr_index) = ev->mbuf;
@@ -314,7 +315,7 @@ dpaa2_eventdev_dequeue_burst(void *port, struct rte_event ev[],
 		if (DPAA2_PER_LCORE_DQRR_HELD & (1 << i)) {
 			qbman_swp_dqrr_idx_consume(swp, i);
 			DPAA2_PER_LCORE_DQRR_SIZE--;
-			DPAA2_PER_LCORE_DQRR_MBUF(i)->seqn =
+			*dpaa2_seqn(DPAA2_PER_LCORE_DQRR_MBUF(i)) =
 				DPAA2_INVALID_MBUF_SEQN;
 		}
 		i++;
diff --git a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
index 5447db8a8a..cd7311a94d 100644
--- a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
+++ b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
@@ -19,6 +19,7 @@
 #include <rte_random.h>
 #include <rte_bus_vdev.h>
 #include <rte_test.h>
+#include <rte_fslmc.h>
 
 #include "dpaa2_eventdev.h"
 #include "dpaa2_eventdev_logs.h"
@@ -274,7 +275,8 @@ check_excess_events(uint8_t port)
 		valid_event = rte_event_dequeue_burst(evdev, port, &ev, 1, 0);
 
 		RTE_TEST_ASSERT_SUCCESS(valid_event,
-				"Unexpected valid event=%d", ev.mbuf->seqn);
+				"Unexpected valid event=%d",
+				*dpaa2_seqn(ev.mbuf));
 	}
 	return 0;
 }
diff --git a/drivers/mempool/dpaa2/dpaa2_hw_mempool.h b/drivers/mempool/dpaa2/dpaa2_hw_mempool.h
index 53fa1552d1..7c493b28e7 100644
--- a/drivers/mempool/dpaa2/dpaa2_hw_mempool.h
+++ b/drivers/mempool/dpaa2/dpaa2_hw_mempool.h
@@ -10,8 +10,6 @@
 
 #define DPAA2_MAX_BUF_POOLS	8
 
-#define DPAA2_INVALID_MBUF_SEQN	0
-
 struct buf_pool_cfg {
 	void *addr;
 	/**< The address from where DPAA2 will carve out the buffers */
diff --git a/drivers/net/dpaa2/dpaa2_rxtx.c b/drivers/net/dpaa2/dpaa2_rxtx.c
index 4dd1d5f578..6201de4606 100644
--- a/drivers/net/dpaa2/dpaa2_rxtx.c
+++ b/drivers/net/dpaa2/dpaa2_rxtx.c
@@ -710,7 +710,7 @@ dpaa2_dev_process_atomic_event(struct qbman_swp *swp __rte_unused,
 	ev->mbuf = eth_fd_to_mbuf(fd, rxq->eth_data->port_id);
 
 	dqrr_index = qbman_get_dqrr_idx(dq);
-	ev->mbuf->seqn = dqrr_index + 1;
+	*dpaa2_seqn(ev->mbuf) = dqrr_index + 1;
 	DPAA2_PER_LCORE_DQRR_SIZE++;
 	DPAA2_PER_LCORE_DQRR_HELD |= 1 << dqrr_index;
 	DPAA2_PER_LCORE_DQRR_MBUF(dqrr_index) = ev->mbuf;
@@ -736,9 +736,9 @@ dpaa2_dev_process_ordered_event(struct qbman_swp *swp,
 
 	ev->mbuf = eth_fd_to_mbuf(fd, rxq->eth_data->port_id);
 
-	ev->mbuf->seqn = DPAA2_ENQUEUE_FLAG_ORP;
-	ev->mbuf->seqn |= qbman_result_DQ_odpid(dq) << DPAA2_EQCR_OPRID_SHIFT;
-	ev->mbuf->seqn |= qbman_result_DQ_seqnum(dq) << DPAA2_EQCR_SEQNUM_SHIFT;
+	*dpaa2_seqn(ev->mbuf) = DPAA2_ENQUEUE_FLAG_ORP;
+	*dpaa2_seqn(ev->mbuf) |= qbman_result_DQ_odpid(dq) << DPAA2_EQCR_OPRID_SHIFT;
+	*dpaa2_seqn(ev->mbuf) |= qbman_result_DQ_seqnum(dq) << DPAA2_EQCR_SEQNUM_SHIFT;
 
 	qbman_swp_dqrr_consume(swp, dq);
 }
@@ -1063,14 +1063,14 @@ dpaa2_dev_tx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
 			dpaa2_eqcr_size : nb_pkts;
 
 		for (loop = 0; loop < frames_to_send; loop++) {
-			if ((*bufs)->seqn) {
-				uint8_t dqrr_index = (*bufs)->seqn - 1;
+			if (*dpaa2_seqn(*bufs)) {
+				uint8_t dqrr_index = *dpaa2_seqn(*bufs) - 1;
 
 				flags[loop] = QBMAN_ENQUEUE_FLAG_DCA |
 						dqrr_index;
 				DPAA2_PER_LCORE_DQRR_SIZE--;
 				DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dqrr_index);
-				(*bufs)->seqn = DPAA2_INVALID_MBUF_SEQN;
+				*dpaa2_seqn(*bufs) = DPAA2_INVALID_MBUF_SEQN;
 			}
 
 			if (likely(RTE_MBUF_DIRECT(*bufs))) {
@@ -1230,10 +1230,10 @@ dpaa2_set_enqueue_descriptor(struct dpaa2_queue *dpaa2_q,
 
 	qbman_eq_desc_set_fq(eqdesc, dpaa2_q->fqid);
 
-	if (m->seqn & DPAA2_ENQUEUE_FLAG_ORP) {
-		orpid = (m->seqn & DPAA2_EQCR_OPRID_MASK) >>
+	if (*dpaa2_seqn(m) & DPAA2_ENQUEUE_FLAG_ORP) {
+		orpid = (*dpaa2_seqn(m) & DPAA2_EQCR_OPRID_MASK) >>
 			DPAA2_EQCR_OPRID_SHIFT;
-		seqnum = (m->seqn & DPAA2_EQCR_SEQNUM_MASK) >>
+		seqnum = (*dpaa2_seqn(m) & DPAA2_EQCR_SEQNUM_MASK) >>
 			DPAA2_EQCR_SEQNUM_SHIFT;
 
 		if (!priv->en_loose_ordered) {
@@ -1255,12 +1255,12 @@ dpaa2_set_enqueue_descriptor(struct dpaa2_queue *dpaa2_q,
 			qbman_eq_desc_set_orp(eqdesc, 0, orpid, seqnum, 0);
 		}
 	} else {
-		dq_idx = m->seqn - 1;
+		dq_idx = *dpaa2_seqn(m) - 1;
 		qbman_eq_desc_set_dca(eqdesc, 1, dq_idx, 0);
 		DPAA2_PER_LCORE_DQRR_SIZE--;
 		DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dq_idx);
 	}
-	m->seqn = DPAA2_INVALID_MBUF_SEQN;
+	*dpaa2_seqn(m) = DPAA2_INVALID_MBUF_SEQN;
 }
 
 /* Callback to handle sending ordered packets through WRIOP based interface */
@@ -1314,7 +1314,7 @@ dpaa2_dev_tx_ordered(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
 			dpaa2_eqcr_size : nb_pkts;
 
 		if (!priv->en_loose_ordered) {
-			if ((*bufs)->seqn & DPAA2_ENQUEUE_FLAG_ORP) {
+			if (*dpaa2_seqn(*bufs) & DPAA2_ENQUEUE_FLAG_ORP) {
 				num_free_eq_desc = dpaa2_free_eq_descriptors();
 				if (num_free_eq_desc < frames_to_send)
 					frames_to_send = num_free_eq_desc;
@@ -1325,7 +1325,7 @@ dpaa2_dev_tx_ordered(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
 			/*Prepare enqueue descriptor*/
 			qbman_eq_desc_clear(&eqdesc[loop]);
 
-			if ((*bufs)->seqn) {
+			if (*dpaa2_seqn(*bufs)) {
 				/* Use only queue 0 for Tx in case of atomic/
 				 * ordered packets as packets can get unordered
 				 * when being tranmitted out from the interface
-- 
2.23.0

[dpdk-dev] [PATCH 7/8] event: switch sequence number to dynamic field

[David Marchand]

The eventdev drivers have been hacking the deprecated field seqn for
internal test usage.
It is moved to a dynamic mbuf field in order to allow removal of seqn.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 app/test-eventdev/evt_main.c                  |  3 ++
 app/test-eventdev/test_order_common.c         |  2 +-
 app/test-eventdev/test_order_common.h         |  5 ++-
 drivers/event/octeontx/ssovf_evdev_selftest.c | 32 ++++++++--------
 drivers/event/octeontx2/otx2_evdev_selftest.c | 31 +++++++--------
 drivers/event/opdl/opdl_test.c                |  8 ++--
 drivers/event/sw/sw_evdev_selftest.c          | 34 +++++++++--------
 lib/librte_eventdev/rte_eventdev.c            | 21 +++++++++-
 lib/librte_eventdev/rte_eventdev.h            | 38 ++++++++++++++++---
 lib/librte_eventdev/version.map               |  2 +
 10 files changed, 116 insertions(+), 60 deletions(-)

diff --git a/app/test-eventdev/evt_main.c b/app/test-eventdev/evt_main.c
index a8d304bab3..832bb21d7c 100644
--- a/app/test-eventdev/evt_main.c
+++ b/app/test-eventdev/evt_main.c
@@ -89,6 +89,9 @@ main(int argc, char **argv)
 	if (!evdevs)
 		rte_panic("no eventdev devices found\n");
 
+	if (rte_event_test_seqn_dynfield_register() < 0)
+		rte_panic("failed to register event dev sequence number\n");
+
 	/* Populate the default values of the options */
 	evt_options_default(&opt);
 
diff --git a/app/test-eventdev/test_order_common.c b/app/test-eventdev/test_order_common.c
index c5f7317440..d15ff80273 100644
--- a/app/test-eventdev/test_order_common.c
+++ b/app/test-eventdev/test_order_common.c
@@ -50,7 +50,7 @@ order_producer(void *arg)
 
 		const flow_id_t flow = (uintptr_t)m % nb_flows;
 		/* Maintain seq number per flow */
-		m->seqn = producer_flow_seq[flow]++;
+		*rte_event_test_seqn(m) = producer_flow_seq[flow]++;
 		flow_id_save(flow, m, &ev);
 
 		while (rte_event_enqueue_burst(dev_id, port, &ev, 1) != 1) {
diff --git a/app/test-eventdev/test_order_common.h b/app/test-eventdev/test_order_common.h
index 9e3415e421..d4ad31da46 100644
--- a/app/test-eventdev/test_order_common.h
+++ b/app/test-eventdev/test_order_common.h
@@ -89,9 +89,10 @@ order_process_stage_1(struct test_order *const t,
 {
 	const uint32_t flow = (uintptr_t)ev->mbuf % nb_flows;
 	/* compare the seqn against expected value */
-	if (ev->mbuf->seqn != expected_flow_seq[flow]) {
+	if (*rte_event_test_seqn(ev->mbuf) != expected_flow_seq[flow]) {
 		evt_err("flow=%x seqn mismatch got=%x expected=%x",
-			flow, ev->mbuf->seqn, expected_flow_seq[flow]);
+			flow, *rte_event_test_seqn(ev->mbuf),
+			expected_flow_seq[flow]);
 		t->err = true;
 		rte_smp_wmb();
 	}
diff --git a/drivers/event/octeontx/ssovf_evdev_selftest.c b/drivers/event/octeontx/ssovf_evdev_selftest.c
index 7a2b7ded25..b99889e2cc 100644
--- a/drivers/event/octeontx/ssovf_evdev_selftest.c
+++ b/drivers/event/octeontx/ssovf_evdev_selftest.c
@@ -300,7 +300,7 @@ inject_events(uint32_t flow_id, uint8_t event_type, uint8_t sub_event_type,
 		m = rte_pktmbuf_alloc(eventdev_test_mempool);
 		RTE_TEST_ASSERT_NOT_NULL(m, "mempool alloc failed");
 
-		m->seqn = i;
+		*rte_event_test_seqn(m) = i;
 		update_event_and_validation_attr(m, &ev, flow_id, event_type,
 			sub_event_type, sched_type, queue, port);
 		rte_event_enqueue_burst(evdev, port, &ev, 1);
@@ -320,7 +320,8 @@ check_excess_events(uint8_t port)
 		valid_event = rte_event_dequeue_burst(evdev, port, &ev, 1, 0);
 
 		RTE_TEST_ASSERT_SUCCESS(valid_event,
-				"Unexpected valid event=%d", ev.mbuf->seqn);
+			"Unexpected valid event=%d",
+			*rte_event_test_seqn(ev.mbuf));
 	}
 	return 0;
 }
@@ -425,8 +426,9 @@ static int
 validate_simple_enqdeq(uint32_t index, uint8_t port, struct rte_event *ev)
 {
 	RTE_SET_USED(port);
-	RTE_TEST_ASSERT_EQUAL(index, ev->mbuf->seqn, "index=%d != seqn=%d",
-			index, ev->mbuf->seqn);
+	RTE_TEST_ASSERT_EQUAL(index, *rte_event_test_seqn(ev->mbuf),
+		"index=%d != seqn=%d", index,
+		*rte_event_test_seqn(ev->mbuf));
 	return 0;
 }
 
@@ -509,10 +511,10 @@ validate_queue_priority(uint32_t index, uint8_t port, struct rte_event *ev)
 
 	expected_val += ev->queue_id;
 	RTE_SET_USED(port);
-	RTE_TEST_ASSERT_EQUAL(ev->mbuf->seqn, expected_val,
-	"seqn=%d index=%d expected=%d range=%d nb_queues=%d max_event=%d",
-			ev->mbuf->seqn, index, expected_val, range,
-			queue_count, MAX_EVENTS);
+	RTE_TEST_ASSERT_EQUAL(*rte_event_test_seqn(ev->mbuf), expected_val,
+		"seqn=%d index=%d expected=%d range=%d nb_queues=%d max_event=%d",
+		*rte_event_test_seqn(ev->mbuf), index, expected_val, range,
+		queue_count, MAX_EVENTS);
 	return 0;
 }
 
@@ -537,7 +539,7 @@ test_multi_queue_priority(void)
 		m = rte_pktmbuf_alloc(eventdev_test_mempool);
 		RTE_TEST_ASSERT_NOT_NULL(m, "mempool alloc failed");
 
-		m->seqn = i;
+		*rte_event_test_seqn(m) = i;
 		queue = i % queue_count;
 		update_event_and_validation_attr(m, &ev, 0, RTE_EVENT_TYPE_CPU,
 			0, RTE_SCHED_TYPE_PARALLEL, queue, 0);
@@ -904,7 +906,7 @@ worker_flow_based_pipeline(void *arg)
 			ev.op = RTE_EVENT_OP_FORWARD;
 			rte_event_enqueue_burst(evdev, port, &ev, 1);
 		} else if (ev.sub_event_type == 1) { /* Events from stage 1*/
-			if (seqn_list_update(ev.mbuf->seqn) == 0) {
+			if (seqn_list_update(*rte_event_test_seqn(ev.mbuf)) == 0) {
 				rte_pktmbuf_free(ev.mbuf);
 				rte_atomic32_sub(total_events, 1);
 			} else {
@@ -939,7 +941,7 @@ test_multiport_flow_sched_type_test(uint8_t in_sched_type,
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(
 		0x1 /*flow_id */,
 		RTE_EVENT_TYPE_CPU /* event_type */,
@@ -1059,7 +1061,7 @@ worker_group_based_pipeline(void *arg)
 			ev.op = RTE_EVENT_OP_FORWARD;
 			rte_event_enqueue_burst(evdev, port, &ev, 1);
 		} else if (ev.queue_id == 1) { /* Events from stage 1(group 1)*/
-			if (seqn_list_update(ev.mbuf->seqn) == 0) {
+			if (seqn_list_update(*rte_event_test_seqn(ev.mbuf)) == 0) {
 				rte_pktmbuf_free(ev.mbuf);
 				rte_atomic32_sub(total_events, 1);
 			} else {
@@ -1101,7 +1103,7 @@ test_multiport_queue_sched_type_test(uint8_t in_sched_type,
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(
 		0x1 /*flow_id */,
 		RTE_EVENT_TYPE_CPU /* event_type */,
@@ -1238,7 +1240,7 @@ launch_multi_port_max_stages_random_sched_type(int (*fn)(void *))
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(
 		0x1 /*flow_id */,
 		RTE_EVENT_TYPE_CPU /* event_type */,
@@ -1360,7 +1362,7 @@ worker_ordered_flow_producer(void *arg)
 		if (m == NULL)
 			continue;
 
-		m->seqn = counter++;
+		*rte_event_test_seqn(m) = counter++;
 
 		struct rte_event ev = {.event = 0, .u64 = 0};
 
diff --git a/drivers/event/octeontx2/otx2_evdev_selftest.c b/drivers/event/octeontx2/otx2_evdev_selftest.c
index 334a9ccb7c..c6381ac785 100644
--- a/drivers/event/octeontx2/otx2_evdev_selftest.c
+++ b/drivers/event/octeontx2/otx2_evdev_selftest.c
@@ -279,7 +279,7 @@ inject_events(uint32_t flow_id, uint8_t event_type, uint8_t sub_event_type,
 		m = rte_pktmbuf_alloc(eventdev_test_mempool);
 		RTE_TEST_ASSERT_NOT_NULL(m, "mempool alloc failed");
 
-		m->seqn = i;
+		*rte_event_test_seqn(m) = i;
 		update_event_and_validation_attr(m, &ev, flow_id, event_type,
 						 sub_event_type, sched_type,
 						 queue, port);
@@ -301,7 +301,7 @@ check_excess_events(uint8_t port)
 
 		RTE_TEST_ASSERT_SUCCESS(valid_event,
 					"Unexpected valid event=%d",
-					ev.mbuf->seqn);
+					*rte_event_test_seqn(ev.mbuf));
 	}
 	return 0;
 }
@@ -406,8 +406,9 @@ static int
 validate_simple_enqdeq(uint32_t index, uint8_t port, struct rte_event *ev)
 {
 	RTE_SET_USED(port);
-	RTE_TEST_ASSERT_EQUAL(index, ev->mbuf->seqn, "index=%d != seqn=%d",
-			      index, ev->mbuf->seqn);
+	RTE_TEST_ASSERT_EQUAL(index, *rte_event_test_seqn(ev->mbuf),
+		"index=%d != seqn=%d",
+		index, *rte_event_test_seqn(ev->mbuf));
 	return 0;
 }
 
@@ -493,10 +494,10 @@ validate_queue_priority(uint32_t index, uint8_t port, struct rte_event *ev)
 
 	expected_val += ev->queue_id;
 	RTE_SET_USED(port);
-	RTE_TEST_ASSERT_EQUAL(ev->mbuf->seqn, expected_val,
-	"seqn=%d index=%d expected=%d range=%d nb_queues=%d max_event=%d",
-			      ev->mbuf->seqn, index, expected_val, range,
-			      queue_count, MAX_EVENTS);
+	RTE_TEST_ASSERT_EQUAL(*rte_event_test_seqn(ev->mbuf), expected_val,
+		"seqn=%d index=%d expected=%d range=%d nb_queues=%d max_event=%d",
+		*rte_event_test_seqn(ev->mbuf), index, expected_val, range,
+		queue_count, MAX_EVENTS);
 	return 0;
 }
 
@@ -523,7 +524,7 @@ test_multi_queue_priority(void)
 		m = rte_pktmbuf_alloc(eventdev_test_mempool);
 		RTE_TEST_ASSERT_NOT_NULL(m, "mempool alloc failed");
 
-		m->seqn = i;
+		*rte_event_test_seqn(m) = i;
 		queue = i % queue_count;
 		update_event_and_validation_attr(m, &ev, 0, RTE_EVENT_TYPE_CPU,
 						 0, RTE_SCHED_TYPE_PARALLEL,
@@ -888,7 +889,7 @@ worker_flow_based_pipeline(void *arg)
 			ev.op = RTE_EVENT_OP_FORWARD;
 			rte_event_enqueue_burst(evdev, port, &ev, 1);
 		} else if (ev.sub_event_type == 1) { /* Events from stage 1*/
-			if (seqn_list_update(ev.mbuf->seqn) == 0) {
+			if (seqn_list_update(*rte_event_test_seqn(ev.mbuf)) == 0) {
 				rte_pktmbuf_free(ev.mbuf);
 				rte_atomic32_sub(total_events, 1);
 			} else {
@@ -923,7 +924,7 @@ test_multiport_flow_sched_type_test(uint8_t in_sched_type,
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(0x1 /*flow_id */,
 			    RTE_EVENT_TYPE_CPU /* event_type */,
 			    0 /* sub_event_type (stage 0) */,
@@ -1043,7 +1044,7 @@ worker_group_based_pipeline(void *arg)
 			ev.op = RTE_EVENT_OP_FORWARD;
 			rte_event_enqueue_burst(evdev, port, &ev, 1);
 		} else if (ev.queue_id == 1) { /* Events from stage 1(group 1)*/
-			if (seqn_list_update(ev.mbuf->seqn) == 0) {
+			if (seqn_list_update(*rte_event_test_seqn(ev.mbuf)) == 0) {
 				rte_pktmbuf_free(ev.mbuf);
 				rte_atomic32_sub(total_events, 1);
 			} else {
@@ -1084,7 +1085,7 @@ test_multiport_queue_sched_type_test(uint8_t in_sched_type,
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(0x1 /*flow_id */,
 			    RTE_EVENT_TYPE_CPU /* event_type */,
 			    0 /* sub_event_type (stage 0) */,
@@ -1222,7 +1223,7 @@ launch_multi_port_max_stages_random_sched_type(int (*fn)(void *))
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(0x1 /*flow_id */,
 			    RTE_EVENT_TYPE_CPU /* event_type */,
 			    0 /* sub_event_type (stage 0) */,
@@ -1348,7 +1349,7 @@ worker_ordered_flow_producer(void *arg)
 		if (m == NULL)
 			continue;
 
-		m->seqn = counter++;
+		*rte_event_test_seqn(m) = counter++;
 
 		struct rte_event ev = {.event = 0, .u64 = 0};
 
diff --git a/drivers/event/opdl/opdl_test.c b/drivers/event/opdl/opdl_test.c
index e7a32fbd31..cbf33d38f7 100644
--- a/drivers/event/opdl/opdl_test.c
+++ b/drivers/event/opdl/opdl_test.c
@@ -256,7 +256,7 @@ ordered_basic(struct test *t)
 		ev.queue_id = t->qid[0];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = mbufs[i];
-		mbufs[i]->seqn = MAGIC_SEQN + i;
+		*rte_event_test_seqn(mbufs[i]) = MAGIC_SEQN + i;
 
 		/* generate pkt and enqueue */
 		err = rte_event_enqueue_burst(evdev, t->port[rx_port], &ev, 1);
@@ -281,7 +281,7 @@ ordered_basic(struct test *t)
 			rte_event_dev_dump(evdev, stdout);
 			return -1;
 		}
-		seq = deq_ev[i].mbuf->seqn  - MAGIC_SEQN;
+		seq = *rte_event_test_seqn(deq_ev[i].mbuf)  - MAGIC_SEQN;
 
 		if (seq != (i-1)) {
 			PMD_DRV_LOG(ERR, " seq test failed ! eq is %d , "
@@ -396,7 +396,7 @@ atomic_basic(struct test *t)
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.flow_id = 1;
 		ev.mbuf = mbufs[i];
-		mbufs[i]->seqn = MAGIC_SEQN + i;
+		*rte_event_test_seqn(mbufs[i]) = MAGIC_SEQN + i;
 
 		/* generate pkt and enqueue */
 		err = rte_event_enqueue_burst(evdev, t->port[rx_port], &ev, 1);
@@ -625,7 +625,7 @@ single_link_w_stats(struct test *t)
 		ev.queue_id = t->qid[0];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = mbufs[i];
-		mbufs[i]->seqn = 1234 + i;
+		*rte_event_test_seqn(mbufs[i]) = 1234 + i;
 
 		/* generate pkt and enqueue */
 		err = rte_event_enqueue_burst(evdev, t->port[rx_port], &ev, 1);
diff --git a/drivers/event/sw/sw_evdev_selftest.c b/drivers/event/sw/sw_evdev_selftest.c
index ad4fc0eed7..47f5b55651 100644
--- a/drivers/event/sw/sw_evdev_selftest.c
+++ b/drivers/event/sw/sw_evdev_selftest.c
@@ -380,7 +380,7 @@ run_prio_packet_test(struct test *t)
 			printf("%d: gen of pkt failed\n", __LINE__);
 			return -1;
 		}
-		arp->seqn = MAGIC_SEQN[i];
+		*rte_event_test_seqn(arp) = MAGIC_SEQN[i];
 
 		ev = (struct rte_event){
 			.priority = PRIORITY[i],
@@ -419,7 +419,7 @@ run_prio_packet_test(struct test *t)
 		rte_event_dev_dump(evdev, stdout);
 		return -1;
 	}
-	if (ev.mbuf->seqn != MAGIC_SEQN[1]) {
+	if (*rte_event_test_seqn(ev.mbuf) != MAGIC_SEQN[1]) {
 		printf("%d: first packet out not highest priority\n",
 				__LINE__);
 		rte_event_dev_dump(evdev, stdout);
@@ -433,7 +433,7 @@ run_prio_packet_test(struct test *t)
 		rte_event_dev_dump(evdev, stdout);
 		return -1;
 	}
-	if (ev2.mbuf->seqn != MAGIC_SEQN[0]) {
+	if (*rte_event_test_seqn(ev2.mbuf) != MAGIC_SEQN[0]) {
 		printf("%d: second packet out not lower priority\n",
 				__LINE__);
 		rte_event_dev_dump(evdev, stdout);
@@ -477,7 +477,7 @@ test_single_directed_packet(struct test *t)
 	}
 
 	const uint32_t MAGIC_SEQN = 4711;
-	arp->seqn = MAGIC_SEQN;
+	*rte_event_test_seqn(arp) = MAGIC_SEQN;
 
 	/* generate pkt and enqueue */
 	err = rte_event_enqueue_burst(evdev, rx_enq, &ev, 1);
@@ -516,7 +516,7 @@ test_single_directed_packet(struct test *t)
 		return -1;
 	}
 
-	if (ev.mbuf->seqn != MAGIC_SEQN) {
+	if (*rte_event_test_seqn(ev.mbuf) != MAGIC_SEQN) {
 		printf("%d: error magic sequence number not dequeued\n",
 				__LINE__);
 		return -1;
@@ -934,7 +934,7 @@ xstats_tests(struct test *t)
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = arp;
 		ev.flow_id = 7;
-		arp->seqn = i;
+		*rte_event_test_seqn(arp) = i;
 
 		int err = rte_event_enqueue_burst(evdev, t->port[0], &ev, 1);
 		if (err != 1) {
@@ -1485,7 +1485,7 @@ xstats_id_reset_tests(struct test *t)
 		ev.queue_id = t->qid[i];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = arp;
-		arp->seqn = i;
+		*rte_event_test_seqn(arp) = i;
 
 		int err = rte_event_enqueue_burst(evdev, t->port[0], &ev, 1);
 		if (err != 1) {
@@ -1873,7 +1873,7 @@ qid_priorities(struct test *t)
 		ev.queue_id = t->qid[i];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = arp;
-		arp->seqn = i;
+		*rte_event_test_seqn(arp) = i;
 
 		int err = rte_event_enqueue_burst(evdev, t->port[0], &ev, 1);
 		if (err != 1) {
@@ -1894,7 +1894,7 @@ qid_priorities(struct test *t)
 		return -1;
 	}
 	for (i = 0; i < 3; i++) {
-		if (ev[i].mbuf->seqn != 2-i) {
+		if (*rte_event_test_seqn(ev[i].mbuf) != 2-i) {
 			printf(
 				"%d: qid priority test: seqn %d incorrectly prioritized\n",
 					__LINE__, i);
@@ -2371,7 +2371,7 @@ single_packet(struct test *t)
 	ev.mbuf = arp;
 	ev.queue_id = 0;
 	ev.flow_id = 3;
-	arp->seqn = MAGIC_SEQN;
+	*rte_event_test_seqn(arp) = MAGIC_SEQN;
 
 	err = rte_event_enqueue_burst(evdev, t->port[rx_enq], &ev, 1);
 	if (err != 1) {
@@ -2411,7 +2411,7 @@ single_packet(struct test *t)
 	}
 
 	err = test_event_dev_stats_get(evdev, &stats);
-	if (ev.mbuf->seqn != MAGIC_SEQN) {
+	if (*rte_event_test_seqn(ev.mbuf) != MAGIC_SEQN) {
 		printf("%d: magic sequence number not dequeued\n", __LINE__);
 		return -1;
 	}
@@ -2684,7 +2684,7 @@ parallel_basic(struct test *t, int check_order)
 		ev.queue_id = t->qid[0];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = mbufs[i];
-		mbufs[i]->seqn = MAGIC_SEQN + i;
+		*rte_event_test_seqn(mbufs[i]) = MAGIC_SEQN + i;
 
 		/* generate pkt and enqueue */
 		err = rte_event_enqueue_burst(evdev, t->port[rx_port], &ev, 1);
@@ -2739,10 +2739,12 @@ parallel_basic(struct test *t, int check_order)
 	/* Check to see if the sequence numbers are in expected order */
 	if (check_order) {
 		for (j = 0 ; j < deq_pkts ; j++) {
-			if (deq_ev[j].mbuf->seqn != MAGIC_SEQN + j) {
-				printf(
-					"%d: Incorrect sequence number(%d) from port %d\n",
-					__LINE__, mbufs_out[j]->seqn, tx_port);
+			if (*rte_event_test_seqn(deq_ev[j].mbuf) !=
+					MAGIC_SEQN + j) {
+				printf("%d: Incorrect sequence number(%d) from port %d\n",
+					__LINE__,
+					*rte_event_test_seqn(mbufs_out[j]),
+					tx_port);
 				return -1;
 			}
 		}
diff --git a/lib/librte_eventdev/rte_eventdev.c b/lib/librte_eventdev/rte_eventdev.c
index 322453c532..61ff6d3404 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -109,6 +109,22 @@ rte_event_dev_info_get(uint8_t dev_id, struct rte_event_dev_info *dev_info)
 	return 0;
 }
 
+#define RTE_EVENT_TEST_SEQN_DYNFIELD_NAME "rte_event_test_seqn_dynfield"
+int rte_event_test_seqn_dynfield_offset = -1;
+
+int
+rte_event_test_seqn_dynfield_register(void)
+{
+	static const struct rte_mbuf_dynfield event_test_seqn_dynfield_desc = {
+		.name = RTE_EVENT_TEST_SEQN_DYNFIELD_NAME,
+		.size = sizeof(rte_event_test_seqn_t),
+		.align = __alignof__(rte_event_test_seqn_t),
+	};
+	rte_event_test_seqn_dynfield_offset =
+		rte_mbuf_dynfield_register(&event_test_seqn_dynfield_desc);
+	return rte_event_test_seqn_dynfield_offset;
+}
+
 int
 rte_event_eth_rx_adapter_caps_get(uint8_t dev_id, uint16_t eth_port_id,
 				uint32_t *caps)
@@ -1247,8 +1263,11 @@ int rte_event_dev_selftest(uint8_t dev_id)
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
 	struct rte_eventdev *dev = &rte_eventdevs[dev_id];
 
-	if (dev->dev_ops->dev_selftest != NULL)
+	if (dev->dev_ops->dev_selftest != NULL) {
+		if (rte_event_test_seqn_dynfield_register() < 0)
+			return -ENOMEM;
 		return (*dev->dev_ops->dev_selftest)();
+	}
 	return -ENOTSUP;
 }
 
diff --git a/lib/librte_eventdev/rte_eventdev.h b/lib/librte_eventdev/rte_eventdev.h
index ce1fc2ce0f..1656ff8dce 100644
--- a/lib/librte_eventdev/rte_eventdev.h
+++ b/lib/librte_eventdev/rte_eventdev.h
@@ -211,13 +211,15 @@ extern "C" {
 #endif
 
 #include <rte_common.h>
+#include <rte_compat.h>
 #include <rte_config.h>
+#include <rte_mbuf.h>
+#include <rte_mbuf_dyn.h>
 #include <rte_memory.h>
 #include <rte_errno.h>
 
 #include "rte_eventdev_trace_fp.h"
 
-struct rte_mbuf; /* we just use mbuf pointers; no need to include rte_mbuf.h */
 struct rte_event;
 
 /* Event device capability bitmap flags */
@@ -570,9 +572,9 @@ struct rte_event_queue_conf {
 	 */
 	uint32_t nb_atomic_order_sequences;
 	/**< The maximum number of outstanding events waiting to be
-	 * reordered by this queue. In other words, the number of entries in
-	 * this queue’s reorder buffer.When the number of events in the
-	 * reorder buffer reaches to *nb_atomic_order_sequences* then the
+	 * event_tested by this queue. In other words, the number of entries in
+	 * this queue’s event_test buffer.When the number of events in the
+	 * event_test buffer reaches to *nb_atomic_order_sequences* then the
 	 * scheduler cannot schedule the events from this queue and invalid
 	 * event will be returned from dequeue until one or more entries are
 	 * freed up/released.
@@ -935,7 +937,7 @@ rte_event_dev_close(uint8_t dev_id);
  * Event ordering is based on the received event(s), but also other
  * (newly allocated or stored) events are ordered when enqueued within the same
  * ordered context. Events not enqueued (e.g. released or stored) within the
- * context are  considered missing from reordering and are skipped at this time
+ * context are  considered missing from event_testing and are skipped at this time
  * (but can be ordered again within another context).
  *
  * @see rte_event_queue_setup(), rte_event_dequeue_burst(), RTE_EVENT_OP_RELEASE
@@ -1021,7 +1023,7 @@ rte_event_dev_close(uint8_t dev_id);
  * then this function hints the scheduler that the user has done all that need
  * to maintain event order in the current ordered context.
  * The scheduler is allowed to release the ordered context of this port and
- * avoid reordering any following enqueues.
+ * avoid event_testing any following enqueues.
  *
  * Early ordered context release may increase parallelism and thus system
  * performance.
@@ -1111,6 +1113,30 @@ struct rte_event {
 	};
 };
 
+typedef uint32_t rte_event_test_seqn_t;
+extern int rte_event_test_seqn_dynfield_offset;
+
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ *
+ * Read test sequence number from mbuf.
+ *
+ * @param mbuf Structure to read from.
+ * @return pointer to test sequence number.
+ */
+__rte_experimental
+static inline rte_event_test_seqn_t *
+rte_event_test_seqn(const struct rte_mbuf *mbuf)
+{
+	return RTE_MBUF_DYNFIELD(mbuf, rte_event_test_seqn_dynfield_offset,
+		rte_event_test_seqn_t *);
+}
+
+__rte_experimental
+int
+rte_event_test_seqn_dynfield_register(void);
+
 /* Ethdev Rx adapter capability bitmap flags */
 #define RTE_EVENT_ETH_RX_ADAPTER_CAP_INTERNAL_PORT	0x1
 /**< This flag is sent when the packet transfer mechanism is in HW.
diff --git a/lib/librte_eventdev/version.map b/lib/librte_eventdev/version.map
index 8ae8420f9b..e49382ba99 100644
--- a/lib/librte_eventdev/version.map
+++ b/lib/librte_eventdev/version.map
@@ -138,4 +138,6 @@ EXPERIMENTAL {
 	__rte_eventdev_trace_port_setup;
 	# added in 20.11
 	rte_event_pmd_pci_probe_named;
+	rte_event_test_seqn_dynfield_offset;
+	rte_event_test_seqn_dynfield_register;
 };
-- 
2.23.0

Re: [dpdk-dev] [PATCH 7/8] event: switch sequence number to dynamic field

[David Marchand]

On Tue, Oct 27, 2020 at 11:16 PM David Marchand
<david.marchand@redhat.com> wrote:
> diff --git a/lib/librte_eventdev/rte_eventdev.h b/lib/librte_eventdev/rte_eventdev.h
> index ce1fc2ce0f..1656ff8dce 100644
> --- a/lib/librte_eventdev/rte_eventdev.h
> +++ b/lib/librte_eventdev/rte_eventdev.h
> @@ -211,13 +211,15 @@ extern "C" {
>  #endif
>
>  #include <rte_common.h>
> +#include <rte_compat.h>
>  #include <rte_config.h>
> +#include <rte_mbuf.h>
> +#include <rte_mbuf_dyn.h>
>  #include <rte_memory.h>
>  #include <rte_errno.h>
>
>  #include "rte_eventdev_trace_fp.h"
>
> -struct rte_mbuf; /* we just use mbuf pointers; no need to include rte_mbuf.h */
>  struct rte_event;
>
>  /* Event device capability bitmap flags */
> @@ -570,9 +572,9 @@ struct rte_event_queue_conf {
>          */
>         uint32_t nb_atomic_order_sequences;
>         /**< The maximum number of outstanding events waiting to be
> -        * reordered by this queue. In other words, the number of entries in
> -        * this queue’s reorder buffer.When the number of events in the
> -        * reorder buffer reaches to *nb_atomic_order_sequences* then the
> +        * event_tested by this queue. In other words, the number of entries in
> +        * this queue’s event_test buffer.When the number of events in the
> +        * event_test buffer reaches to *nb_atomic_order_sequences* then the

v2 tomorrow, without this %s/reorder/test_event/g...


>          * scheduler cannot schedule the events from this queue and invalid
>          * event will be returned from dequeue until one or more entries are
>          * freed up/released.
> @@ -935,7 +937,7 @@ rte_event_dev_close(uint8_t dev_id);
>   * Event ordering is based on the received event(s), but also other
>   * (newly allocated or stored) events are ordered when enqueued within the same
>   * ordered context. Events not enqueued (e.g. released or stored) within the
> - * context are  considered missing from reordering and are skipped at this time
> + * context are  considered missing from event_testing and are skipped at this time
>   * (but can be ordered again within another context).
>   *
>   * @see rte_event_queue_setup(), rte_event_dequeue_burst(), RTE_EVENT_OP_RELEASE
> @@ -1021,7 +1023,7 @@ rte_event_dev_close(uint8_t dev_id);
>   * then this function hints the scheduler that the user has done all that need
>   * to maintain event order in the current ordered context.
>   * The scheduler is allowed to release the ordered context of this port and
> - * avoid reordering any following enqueues.
> + * avoid event_testing any following enqueues.
>   *
>   * Early ordered context release may increase parallelism and thus system
>   * performance.


-- 
David Marchand

Re: [dpdk-dev] [PATCH 7/8] event: switch sequence number to dynamic field

[Jerin Jacob]

On Wed, Oct 28, 2020 at 3:46 AM David Marchand
<david.marchand@redhat.com> wrote:
>
> The eventdev drivers have been hacking the deprecated field seqn for
> internal test usage.
> It is moved to a dynamic mbuf field in order to allow removal of seqn.
>
> Signed-off-by: David Marchand <david.marchand@redhat.com>

Overall looks good some minor comments:
# use eventdev: instead of event: in git commit


> ---
>  app/test-eventdev/evt_main.c                  |  3 ++
>  app/test-eventdev/test_order_common.c         |  2 +-
>  app/test-eventdev/test_order_common.h         |  5 ++-
>  drivers/event/octeontx/ssovf_evdev_selftest.c | 32 ++++++++--------
>  drivers/event/octeontx2/otx2_evdev_selftest.c | 31 +++++++--------
>  drivers/event/opdl/opdl_test.c                |  8 ++--
>  drivers/event/sw/sw_evdev_selftest.c          | 34 +++++++++--------
>  lib/librte_eventdev/rte_eventdev.c            | 21 +++++++++-
>  lib/librte_eventdev/rte_eventdev.h            | 38 ++++++++++++++++---
>  lib/librte_eventdev/version.map               |  2 +
>  10 files changed, 116 insertions(+), 60 deletions(-)
>
> diff --git a/app/test-eventdev/evt_main.c b/app/test-eventdev/evt_main.c
> index a8d304bab3..832bb21d7c 100644
> --- a/app/test-eventdev/evt_main.c
> +++ b/app/test-eventdev/evt_main.c
> @@ -89,6 +89,9 @@ main(int argc, char **argv)
>         if (!evdevs)
>                 rte_panic("no eventdev devices found\n");
>
> +       if (rte_event_test_seqn_dynfield_register() < 0)
> +               rte_panic("failed to register event dev sequence number\n");
> +
>         /* Populate the default values of the options */
>         evt_options_default(&opt);
>
> diff --git a/app/test-eventdev/test_order_common.c b/app/test-eventdev/test_order_common.c
> index c5f7317440..d15ff80273 100644
> --- a/app/test-eventdev/test_order_common.c
> +++ b/app/test-eventdev/test_order_common.c
> @@ -50,7 +50,7 @@ order_producer(void *arg)
>
>                 const flow_id_t flow = (uintptr_t)m % nb_flows;
>                 /* Maintain seq number per flow */
> -               m->seqn = producer_flow_seq[flow]++;
> +               *rte_event_test_seqn(m) = producer_flow_seq[flow]++;
>                 flow_id_save(flow, m, &ev);
>
>                 while (rte_event_enqueue_burst(dev_id, port, &ev, 1) != 1) {
> diff --git a/app/test-eventdev/test_order_common.h b/app/test-eventdev/test_order_common.h
> index 9e3415e421..d4ad31da46 100644
> --- a/app/test-eventdev/test_order_common.h
> +++ b/app/test-eventdev/test_order_common.h
> @@ -89,9 +89,10 @@ order_process_stage_1(struct test_order *const t,
>  {
>         const uint32_t flow = (uintptr_t)ev->mbuf % nb_flows;
>         /* compare the seqn against expected value */
> -       if (ev->mbuf->seqn != expected_flow_seq[flow]) {
> +       if (*rte_event_test_seqn(ev->mbuf) != expected_flow_seq[flow]) {
>                 evt_err("flow=%x seqn mismatch got=%x expected=%x",
> -                       flow, ev->mbuf->seqn, expected_flow_seq[flow]);
> +                       flow, *rte_event_test_seqn(ev->mbuf),
> +                       expected_flow_seq[flow]);
>                 t->err = true;
>                 rte_smp_wmb();
>         }

# Since rte_event_test_seqn_dynfield_register() is the public API, I
would like to limit the scope
only to self test so that  rte_event_test_seqn_dynfield_register()
need not be exposed.
Could you have a separate application-specific dynamic field here?
# Also this patch used in fastpath, better to have offset stored in
fastpath cache line.
see http://mails.dpdk.org/archives/dev/2020-October/189588.html


> +#define RTE_EVENT_TEST_SEQN_DYNFIELD_NAME "rte_event_test_seqn_dynfield"
> +int rte_event_test_seqn_dynfield_offset = -1;
> +
> +int
> +rte_event_test_seqn_dynfield_register(void)
> +{
> +       static const struct rte_mbuf_dynfield event_test_seqn_dynfield_desc = {
> +               .name = RTE_EVENT_TEST_SEQN_DYNFIELD_NAME,
> +               .size = sizeof(rte_event_test_seqn_t),
> +               .align = __alignof__(rte_event_test_seqn_t),
> +       };
> +       rte_event_test_seqn_dynfield_offset =
> +               rte_mbuf_dynfield_register(&event_test_seqn_dynfield_desc);
> +       return rte_event_test_seqn_dynfield_offset;
> +}
> +
>  int
>  rte_event_eth_rx_adapter_caps_get(uint8_t dev_id, uint16_t eth_port_id,
>                                 uint32_t *caps)
> @@ -1247,8 +1263,11 @@ int rte_event_dev_selftest(uint8_t dev_id)
>         RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
>         struct rte_eventdev *dev = &rte_eventdevs[dev_id];
>
> -       if (dev->dev_ops->dev_selftest != NULL)
> +       if (dev->dev_ops->dev_selftest != NULL) {
> +               if (rte_event_test_seqn_dynfield_register() < 0)
> +                       return -ENOMEM;
>                 return (*dev->dev_ops->dev_selftest)();
> +       }
>         return -ENOTSUP;
>  }
>
> diff --git a/lib/librte_eventdev/rte_eventdev.h b/lib/librte_eventdev/rte_eventdev.h
> index ce1fc2ce0f..1656ff8dce 100644
> --- a/lib/librte_eventdev/rte_eventdev.h
> +++ b/lib/librte_eventdev/rte_eventdev.h
> @@ -211,13 +211,15 @@ extern "C" {
>  #endif
>
>  #include <rte_common.h>
> +#include <rte_compat.h>
>  #include <rte_config.h>
> +#include <rte_mbuf.h>
> +#include <rte_mbuf_dyn.h>
>  #include <rte_memory.h>
>  #include <rte_errno.h>
>
>  #include "rte_eventdev_trace_fp.h"
>
> -struct rte_mbuf; /* we just use mbuf pointers; no need to include rte_mbuf.h */
>  struct rte_event;
>
>  /* Event device capability bitmap flags */
> @@ -570,9 +572,9 @@ struct rte_event_queue_conf {
>          */
>         uint32_t nb_atomic_order_sequences;
>         /**< The maximum number of outstanding events waiting to be
> -        * reordered by this queue. In other words, the number of entries in
> -        * this queue’s reorder buffer.When the number of events in the
> -        * reorder buffer reaches to *nb_atomic_order_sequences* then the
> +        * event_tested by this queue. In other words, the number of entries in
> +        * this queue’s event_test buffer.When the number of events in the
> +        * event_test buffer reaches to *nb_atomic_order_sequences* then the
>          * scheduler cannot schedule the events from this queue and invalid
>          * event will be returned from dequeue until one or more entries are
>          * freed up/released.
> @@ -935,7 +937,7 @@ rte_event_dev_close(uint8_t dev_id);
>   * Event ordering is based on the received event(s), but also other
>   * (newly allocated or stored) events are ordered when enqueued within the same
>   * ordered context. Events not enqueued (e.g. released or stored) within the
> - * context are  considered missing from reordering and are skipped at this time
> + * context are  considered missing from event_testing and are skipped at this time
>   * (but can be ordered again within another context).
>   *
>   * @see rte_event_queue_setup(), rte_event_dequeue_burst(), RTE_EVENT_OP_RELEASE
> @@ -1021,7 +1023,7 @@ rte_event_dev_close(uint8_t dev_id);
>   * then this function hints the scheduler that the user has done all that need
>   * to maintain event order in the current ordered context.
>   * The scheduler is allowed to release the ordered context of this port and
> - * avoid reordering any following enqueues.
> + * avoid event_testing any following enqueues.
>   *
>   * Early ordered context release may increase parallelism and thus system
>   * performance.
> @@ -1111,6 +1113,30 @@ struct rte_event {
>         };
>  };
>
> +typedef uint32_t rte_event_test_seqn_t;
> +extern int rte_event_test_seqn_dynfield_offset;
> +
> +/**
> + * @warning
> + * @b EXPERIMENTAL: this API may change without prior notice
> + *
> + * Read test sequence number from mbuf.
> + *
> + * @param mbuf Structure to read from.
> + * @return pointer to test sequence number.
> + */
> +__rte_experimental
> +static inline rte_event_test_seqn_t *
> +rte_event_test_seqn(const struct rte_mbuf *mbuf)
> +{
> +       return RTE_MBUF_DYNFIELD(mbuf, rte_event_test_seqn_dynfield_offset,
> +               rte_event_test_seqn_t *);
> +}
> +
> +__rte_experimental
> +int
> +rte_event_test_seqn_dynfield_register(void);
> +
>  /* Ethdev Rx adapter capability bitmap flags */
>  #define RTE_EVENT_ETH_RX_ADAPTER_CAP_INTERNAL_PORT     0x1
>  /**< This flag is sent when the packet transfer mechanism is in HW.
> diff --git a/lib/librte_eventdev/version.map b/lib/librte_eventdev/version.map
> index 8ae8420f9b..e49382ba99 100644
> --- a/lib/librte_eventdev/version.map
> +++ b/lib/librte_eventdev/version.map
> @@ -138,4 +138,6 @@ EXPERIMENTAL {
>         __rte_eventdev_trace_port_setup;
>         # added in 20.11
>         rte_event_pmd_pci_probe_named;
> +       rte_event_test_seqn_dynfield_offset;
> +       rte_event_test_seqn_dynfield_register;

Could you change symbol name to rte_event_dev_selftest_seqn_dynfield_offset()
to limit the scope only to self test. also, it is not required to expose
rte_event_test_seqn_dynfield_register() in that case.

>  };
> --
> 2.23.0
>

Re: [dpdk-dev] [PATCH 7/8] event: switch sequence number to dynamic field

[David Marchand]

On Wed, Oct 28, 2020 at 8:27 AM Jerin Jacob <jerinjacobk@gmail.com> wrote:
> > diff --git a/app/test-eventdev/test_order_common.h b/app/test-eventdev/test_order_common.h
> > index 9e3415e421..d4ad31da46 100644
> > --- a/app/test-eventdev/test_order_common.h
> > +++ b/app/test-eventdev/test_order_common.h
> > @@ -89,9 +89,10 @@ order_process_stage_1(struct test_order *const t,
> >  {
> >         const uint32_t flow = (uintptr_t)ev->mbuf % nb_flows;
> >         /* compare the seqn against expected value */
> > -       if (ev->mbuf->seqn != expected_flow_seq[flow]) {
> > +       if (*rte_event_test_seqn(ev->mbuf) != expected_flow_seq[flow]) {
> >                 evt_err("flow=%x seqn mismatch got=%x expected=%x",
> > -                       flow, ev->mbuf->seqn, expected_flow_seq[flow]);
> > +                       flow, *rte_event_test_seqn(ev->mbuf),
> > +                       expected_flow_seq[flow]);
> >                 t->err = true;
> >                 rte_smp_wmb();
> >         }
>
> # Since rte_event_test_seqn_dynfield_register() is the public API, I
> would like to limit the scope
> only to self test so that  rte_event_test_seqn_dynfield_register()
> need not be exposed.
> Could you have a separate application-specific dynamic field here?
> # Also this patch used in fastpath, better to have offset stored in
> fastpath cache line.
> see http://mails.dpdk.org/archives/dev/2020-October/189588.html

Ack for a test app dynamic field.
On the second comment, I'll wait for Thomas respin.


> > diff --git a/lib/librte_eventdev/version.map b/lib/librte_eventdev/version.map
> > index 8ae8420f9b..e49382ba99 100644
> > --- a/lib/librte_eventdev/version.map
> > +++ b/lib/librte_eventdev/version.map
> > @@ -138,4 +138,6 @@ EXPERIMENTAL {
> >         __rte_eventdev_trace_port_setup;
> >         # added in 20.11
> >         rte_event_pmd_pci_probe_named;
> > +       rte_event_test_seqn_dynfield_offset;
> > +       rte_event_test_seqn_dynfield_register;
>
> Could you change symbol name to rte_event_dev_selftest_seqn_dynfield_offset()
> to limit the scope only to self test. also, it is not required to expose
> rte_event_test_seqn_dynfield_register() in that case.

How about moving this to rte_eventdev_pmd.h and make it a pmd only API?
=> rte_event_pmd_ prefix

I would mark it rte_internal too.


-- 
David Marchand

Re: [dpdk-dev] [PATCH 7/8] event: switch sequence number to dynamic field

[Jerin Jacob]

On Wed, Oct 28, 2020 at 2:25 PM David Marchand
<david.marchand@redhat.com> wrote:
>
> On Wed, Oct 28, 2020 at 8:27 AM Jerin Jacob <jerinjacobk@gmail.com> wrote:
> > > diff --git a/app/test-eventdev/test_order_common.h b/app/test-eventdev/test_order_common.h
> > > index 9e3415e421..d4ad31da46 100644
> > > --- a/app/test-eventdev/test_order_common.h
> > > +++ b/app/test-eventdev/test_order_common.h

> >
> > Could you change symbol name to rte_event_dev_selftest_seqn_dynfield_offset()
> > to limit the scope only to self test. also, it is not required to expose
> > rte_event_test_seqn_dynfield_register() in that case.
>
> How about moving this to rte_eventdev_pmd.h and make it a pmd only API?
> => rte_event_pmd_ prefix

+1

>
> I would mark it rte_internal too.

Ack.

>
>
> --
> David Marchand
>

[dpdk-dev] [PATCH 8/8] mbuf: remove seqn field

[David Marchand]

As announced in the deprecation note, the field seqn is removed to give
more space to the dynamic fields.

This is how the mbuf layout looks like (pahole-style):

word  type                              name                byte  size
 0    void *                            buf_addr;         /*   0 +  8 */
 1    rte_iova_t                        buf_iova          /*   8 +  8 */
      /* --- RTE_MARKER64               rearm_data;                   */
 2    uint16_t                          data_off;         /*  16 +  2 */
      uint16_t                          refcnt;           /*  18 +  2 */
      uint16_t                          nb_segs;          /*  20 +  2 */
      uint16_t                          port;             /*  22 +  2 */
 3    uint64_t                          ol_flags;         /*  24 +  8 */
      /* --- RTE_MARKER                 rx_descriptor_fields1;        */
 4    uint32_t             union        packet_type;      /*  32 +  4 */
      uint32_t                          pkt_len;          /*  36 +  4 */
 5    uint16_t                          data_len;         /*  40 +  2 */
      uint16_t                          vlan_tci;         /*  42 +  2 */
 5.5  uint64_t             union        hash;             /*  44 +  8 */
 6.5  uint16_t                          vlan_tci_outer;   /*  52 +  2 */
      uint16_t                          buf_len;          /*  54 +  2 */
 7    uint64_t                          timestamp;        /*  56 +  8 */
      /* --- RTE_MARKER                 cacheline1;                   */
 8    struct rte_mempool *              pool;             /*  64 +  8 */
 9    struct rte_mbuf *                 next;             /*  72 +  8 */
10    uint64_t             union        tx_offload;       /*  80 +  8 */
11    struct rte_mbuf_ext_shared_info * shinfo;           /*  88 +  8 */
12    uint16_t                          priv_size;        /*  96 +  2 */
      uint16_t                          timesync;         /*  98 +  2 */
12.5  uint32_t                          dynfield1[7];     /* 100 + 28 */
16    /* --- END                                             128      */

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 doc/guides/rel_notes/deprecation.rst   |  1 -
 doc/guides/rel_notes/release_20_11.rst |  3 +++
 lib/librte_mbuf/rte_mbuf_core.h        | 15 ++++++---------
 3 files changed, 9 insertions(+), 10 deletions(-)

diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
index 0f6f1df12a..fe3fd3956c 100644
--- a/doc/guides/rel_notes/deprecation.rst
+++ b/doc/guides/rel_notes/deprecation.rst
@@ -87,7 +87,6 @@ Deprecation Notices
   The following static fields will be moved as dynamic:
 
   - ``timestamp``
-  - ``seqn``
 
   As a consequence, the layout of the ``struct rte_mbuf`` will be re-arranged,
   avoiding impact on vectorized implementation of the driver datapaths,
diff --git a/doc/guides/rel_notes/release_20_11.rst b/doc/guides/rel_notes/release_20_11.rst
index 3d7edbfdbb..c0a9fc96aa 100644
--- a/doc/guides/rel_notes/release_20_11.rst
+++ b/doc/guides/rel_notes/release_20_11.rst
@@ -429,6 +429,9 @@ API Changes
 * mbuf: Removed the unioned fields ``userdata`` and ``udata64``
   from the structure ``rte_mbuf``. It is replaced with dynamic fields.
 
+* mbuf: Removed the field ``seqn`` from the structure ``rte_mbuf``.
+  It is replaced with dynamic fields.
+
 * pci: Removed the ``rte_kernel_driver`` enum defined in rte_dev.h and
   replaced with a private enum in the PCI subsystem.
 
diff --git a/lib/librte_mbuf/rte_mbuf_core.h b/lib/librte_mbuf/rte_mbuf_core.h
index a65eaaf692..3fb5abda3c 100644
--- a/lib/librte_mbuf/rte_mbuf_core.h
+++ b/lib/librte_mbuf/rte_mbuf_core.h
@@ -640,6 +640,11 @@ struct rte_mbuf {
 		};
 	};
 
+	/** Shared data for external buffer attached to mbuf. See
+	 * rte_pktmbuf_attach_extbuf().
+	 */
+	struct rte_mbuf_ext_shared_info *shinfo;
+
 	/** Size of the application private data. In case of an indirect
 	 * mbuf, it stores the direct mbuf private data size.
 	 */
@@ -648,15 +653,7 @@ struct rte_mbuf {
 	/** Timesync flags for use with IEEE1588. */
 	uint16_t timesync;
 
-	/** Sequence number. See also rte_reorder_insert(). */
-	uint32_t seqn;
-
-	/** Shared data for external buffer attached to mbuf. See
-	 * rte_pktmbuf_attach_extbuf().
-	 */
-	struct rte_mbuf_ext_shared_info *shinfo;
-
-	uint64_t dynfield1[3]; /**< Reserved for dynamic fields. */
+	uint32_t dynfield1[7]; /**< Reserved for dynamic fields. */
 } __rte_cache_aligned;
 
 /**
-- 
2.23.0

Re: [dpdk-dev] [PATCH 8/8] mbuf: remove seqn field

[Andrew Rybchenko]

On 10/28/20 1:13 AM, David Marchand wrote:
> As announced in the deprecation note, the field seqn is removed to give
> more space to the dynamic fields.
> 
> This is how the mbuf layout looks like (pahole-style):
> 
> word  type                              name                byte  size
>  0    void *                            buf_addr;         /*   0 +  8 */
>  1    rte_iova_t                        buf_iova          /*   8 +  8 */
>       /* --- RTE_MARKER64               rearm_data;                   */
>  2    uint16_t                          data_off;         /*  16 +  2 */
>       uint16_t                          refcnt;           /*  18 +  2 */
>       uint16_t                          nb_segs;          /*  20 +  2 */
>       uint16_t                          port;             /*  22 +  2 */
>  3    uint64_t                          ol_flags;         /*  24 +  8 */
>       /* --- RTE_MARKER                 rx_descriptor_fields1;        */
>  4    uint32_t             union        packet_type;      /*  32 +  4 */
>       uint32_t                          pkt_len;          /*  36 +  4 */
>  5    uint16_t                          data_len;         /*  40 +  2 */
>       uint16_t                          vlan_tci;         /*  42 +  2 */
>  5.5  uint64_t             union        hash;             /*  44 +  8 */
>  6.5  uint16_t                          vlan_tci_outer;   /*  52 +  2 */
>       uint16_t                          buf_len;          /*  54 +  2 */
>  7    uint64_t                          timestamp;        /*  56 +  8 */
>       /* --- RTE_MARKER                 cacheline1;                   */
>  8    struct rte_mempool *              pool;             /*  64 +  8 */
>  9    struct rte_mbuf *                 next;             /*  72 +  8 */
> 10    uint64_t             union        tx_offload;       /*  80 +  8 */
> 11    struct rte_mbuf_ext_shared_info * shinfo;           /*  88 +  8 */
> 12    uint16_t                          priv_size;        /*  96 +  2 */
>       uint16_t                          timesync;         /*  98 +  2 */
> 12.5  uint32_t                          dynfield1[7];     /* 100 + 28 */
> 16    /* --- END                                             128      */
> 
> Signed-off-by: David Marchand <david.marchand@redhat.com>

Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>

[dpdk-dev] [PATCH v2 0/9] remove mbuf seqn

[David Marchand]

Followup on the work started by Thomas to free some space in the mbuf
structure.
This series applies on top of Thomas v4 series [1] and drops the seqn
field that had been deprecated.

1: https://patchwork.dpdk.org/project/dpdk/list/?series=13416

-- 
David Marchand

Changelog since v1:
- rebased on latest Thomas series,
- introduced a specific seqn for the eventdev test app,
- moved eventdev seqn as a pmd only thing,

David Marchand (9):
  event/dpaa2: remove dead code
  crypto/scheduler: remove unused internal seqn
  net/ark: remove use of seqn for debug
  reorder: switch sequence number to dynamic mbuf field
  dpaa: switch sequence number to dynamic mbuf field
  fslmc: switch sequence number to dynamic mbuf field
  eventdev: switch sequence number to dynamic mbuf field
  app/eventdev: switch sequence number to dynamic mbuf field
  mbuf: remove seqn field

 app/test-eventdev/test_order_common.c         | 14 +++++++-
 app/test-eventdev/test_order_common.h         | 13 +++++--
 app/test/test_reorder.c                       |  8 ++---
 doc/guides/rel_notes/deprecation.rst          |  1 -
 doc/guides/rel_notes/release_20_11.rst        |  3 ++
 drivers/bus/dpaa/dpaa_bus.c                   | 16 +++++++++
 drivers/bus/dpaa/rte_dpaa_bus.h               | 28 +++++++++++++++
 drivers/bus/dpaa/version.map                  |  1 +
 drivers/bus/fslmc/fslmc_bus.c                 | 17 +++++++++
 drivers/bus/fslmc/rte_fslmc.h                 | 23 ++++++++++++
 drivers/bus/fslmc/version.map                 |  1 +
 drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c   | 18 +++++-----
 drivers/crypto/dpaa_sec/dpaa_sec.c            |  6 ++--
 .../crypto/scheduler/scheduler_pmd_private.h  |  1 -
 drivers/event/dpaa/dpaa_eventdev.c            |  6 ++--
 drivers/event/dpaa2/dpaa2_eventdev.c          |  9 ++---
 drivers/event/dpaa2/dpaa2_eventdev_selftest.c | 17 +++------
 drivers/event/octeontx/ssovf_evdev_selftest.c | 32 +++++++++--------
 drivers/event/octeontx2/otx2_evdev_selftest.c | 36 +++++++++++--------
 drivers/event/opdl/opdl_test.c                |  8 ++---
 drivers/event/sw/sw_evdev_selftest.c          | 34 +++++++++---------
 drivers/mempool/dpaa2/dpaa2_hw_mempool.h      |  2 --
 drivers/net/ark/ark_ethdev_rx.c               |  8 ++---
 drivers/net/dpaa/dpaa_ethdev.h                |  7 ----
 drivers/net/dpaa/dpaa_rxtx.c                  |  6 ++--
 drivers/net/dpaa2/dpaa2_rxtx.c                | 28 +++++++--------
 examples/packet_ordering/main.c               |  2 +-
 lib/librte_eventdev/rte_eventdev.c            | 14 +++++++-
 lib/librte_eventdev/rte_eventdev_pmd.h        | 20 +++++++++++
 lib/librte_eventdev/version.map               |  6 ++++
 lib/librte_mbuf/rte_mbuf_core.h               | 15 ++++----
 lib/librte_reorder/rte_reorder.c              | 23 ++++++++++--
 lib/librte_reorder/rte_reorder.h              | 21 +++++++++++
 lib/librte_reorder/version.map                |  6 ++++
 34 files changed, 314 insertions(+), 136 deletions(-)

-- 
2.23.0

[dpdk-dev] [PATCH v2 1/9] event/dpaa2: remove dead code

[David Marchand]

This code has never been used since introduction.

Fixes: 653242c3375a ("event/dpaa2: add self test")

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/event/dpaa2/dpaa2_eventdev_selftest.c | 13 +------------
 1 file changed, 1 insertion(+), 12 deletions(-)

diff --git a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
index b1f3891484..5447db8a8a 100644
--- a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
+++ b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
@@ -47,17 +47,6 @@ struct event_attr {
 	uint8_t seq;
 };
 
-static uint32_t seqn_list_index;
-static int seqn_list[NUM_PACKETS];
-
-static void
-seqn_list_init(void)
-{
-	RTE_BUILD_BUG_ON(NUM_PACKETS < MAX_EVENTS);
-	memset(seqn_list, 0, sizeof(seqn_list));
-	seqn_list_index = 0;
-}
-
 struct test_core_param {
 	rte_atomic32_t *total_events;
 	uint64_t dequeue_tmo_ticks;
@@ -516,7 +505,7 @@ launch_workers_and_wait(int (*main_worker)(void *),
 		return 0;
 
 	rte_atomic32_set(&atomic_total_events, total_events);
-	seqn_list_init();
+	RTE_BUILD_BUG_ON(NUM_PACKETS < MAX_EVENTS);
 
 	param = malloc(sizeof(struct test_core_param) * nb_workers);
 	if (!param)
-- 
2.23.0

Re: [dpdk-dev] [PATCH v2 1/9] event/dpaa2: remove dead code

[Nipun Gupta]

Acked-by: Nipun Gupta <nipun.gupta@nxp.com>

> -----Original Message-----
> From: David Marchand <david.marchand@redhat.com>
> Sent: Wednesday, October 28, 2020 5:50 PM
> To: dev@dpdk.org
> Cc: Hemant Agrawal <hemant.agrawal@nxp.com>; Nipun Gupta
> <nipun.gupta@nxp.com>
> Subject: [PATCH v2 1/9] event/dpaa2: remove dead code
> 
> This code has never been used since introduction.
> 
> Fixes: 653242c3375a ("event/dpaa2: add self test")
> 
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> ---
>  drivers/event/dpaa2/dpaa2_eventdev_selftest.c | 13 +------------
>  1 file changed, 1 insertion(+), 12 deletions(-)
> 
> diff --git a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
> b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
> index b1f3891484..5447db8a8a 100644
> --- a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
> +++ b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
> @@ -47,17 +47,6 @@ struct event_attr {
>  	uint8_t seq;
>  };
> 
> -static uint32_t seqn_list_index;
> -static int seqn_list[NUM_PACKETS];
> -
> -static void
> -seqn_list_init(void)
> -{
> -	RTE_BUILD_BUG_ON(NUM_PACKETS < MAX_EVENTS);
> -	memset(seqn_list, 0, sizeof(seqn_list));
> -	seqn_list_index = 0;
> -}
> -
>  struct test_core_param {
>  	rte_atomic32_t *total_events;
>  	uint64_t dequeue_tmo_ticks;
> @@ -516,7 +505,7 @@ launch_workers_and_wait(int (*main_worker)(void *),
>  		return 0;
> 
>  	rte_atomic32_set(&atomic_total_events, total_events);
> -	seqn_list_init();
> +	RTE_BUILD_BUG_ON(NUM_PACKETS < MAX_EVENTS);
> 
>  	param = malloc(sizeof(struct test_core_param) * nb_workers);
>  	if (!param)
> --
> 2.23.0

[dpdk-dev] [PATCH v2 2/9] crypto/scheduler: remove unused internal seqn

[David Marchand]

This field has been left behind after dropping its use.

Fixes: 8a48e039432b ("crypto/scheduler: optimize crypto op ordering")

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/crypto/scheduler/scheduler_pmd_private.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/crypto/scheduler/scheduler_pmd_private.h b/drivers/crypto/scheduler/scheduler_pmd_private.h
index adb4eb0632..4d33b9ab44 100644
--- a/drivers/crypto/scheduler/scheduler_pmd_private.h
+++ b/drivers/crypto/scheduler/scheduler_pmd_private.h
@@ -59,7 +59,6 @@ struct scheduler_qp_ctx {
 	uint32_t max_nb_objs;
 
 	struct rte_ring *order_ring;
-	uint32_t seqn;
 } __rte_cache_aligned;
 
 
-- 
2.23.0

[dpdk-dev] [PATCH v2 3/9] net/ark: remove use of seqn for debug

[David Marchand]

The seqn mbuf field is deprecated.
It is currently hacked for debug purpose, it could be changed to a
dynamic field but I see little value in the debug info it offers.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/net/ark/ark_ethdev_rx.c | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ark/ark_ethdev_rx.c b/drivers/net/ark/ark_ethdev_rx.c
index 825b4791be..c24cc00e2f 100644
--- a/drivers/net/ark/ark_ethdev_rx.c
+++ b/drivers/net/ark/ark_ethdev_rx.c
@@ -302,8 +302,6 @@ eth_ark_recv_pkts(void *rx_queue,
 				mbuf->pkt_len = 63;
 				meta->pkt_len = 63;
 			}
-			/* seqn is only set under debug */
-			mbuf->seqn = cons_index;
 		}
 
 		if (unlikely(meta->pkt_len > ARK_RX_MAX_NOCHAIN))
@@ -360,8 +358,6 @@ eth_ark_rx_jumbo(struct ark_rx_queue *queue,
 		mbuf_prev = mbuf;
 		mbuf->data_len = data_len;
 		mbuf->data_off = 0;
-		if (ARK_DEBUG_CORE)
-			mbuf->seqn = cons_index;	/* for debug only */
 
 		cons_index += 1;
 	}
@@ -667,8 +663,8 @@ dump_mbuf_data(struct rte_mbuf *mbuf, uint16_t lo, uint16_t hi)
 {
 	uint16_t i, j;
 
-	ARK_PMD_LOG(DEBUG, " MBUF: %p len %d, off: %d, seq: %" PRIU32 "\n",
-		    mbuf, mbuf->pkt_len, mbuf->data_off, mbuf->seqn);
+	ARK_PMD_LOG(DEBUG, " MBUF: %p len %d, off: %d\n",
+		    mbuf, mbuf->pkt_len, mbuf->data_off);
 	for (i = lo; i < hi; i += 16) {
 		uint8_t *dp = RTE_PTR_ADD(mbuf->buf_addr, i);
 
-- 
2.23.0

[dpdk-dev] [PATCH v2 4/9] reorder: switch sequence number to dynamic mbuf field

[David Marchand]

The reorder library used sequence numbers stored in the deprecated field
seqn.
It is moved to a dynamic mbuf field in order to allow removal of seqn.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 app/test/test_reorder.c          |  8 ++++----
 examples/packet_ordering/main.c  |  2 +-
 lib/librte_reorder/rte_reorder.c | 23 ++++++++++++++++++++---
 lib/librte_reorder/rte_reorder.h | 21 +++++++++++++++++++++
 lib/librte_reorder/version.map   |  6 ++++++
 5 files changed, 52 insertions(+), 8 deletions(-)

diff --git a/app/test/test_reorder.c b/app/test/test_reorder.c
index 58fa9c71b5..1c4226da65 100644
--- a/app/test/test_reorder.c
+++ b/app/test/test_reorder.c
@@ -149,7 +149,7 @@ test_reorder_insert(void)
 	for (i = 0; i < num_bufs; i++) {
 		bufs[i] = rte_pktmbuf_alloc(p);
 		TEST_ASSERT_NOT_NULL(bufs[i], "Packet allocation failed\n");
-		bufs[i]->seqn = i;
+		*rte_reorder_seqn(bufs[i]) = i;
 	}
 
 	/* This should fill up order buffer:
@@ -183,7 +183,7 @@ test_reorder_insert(void)
 	bufs[4] = NULL;
 
 	/* early packet from current sequence window - full ready buffer */
-	bufs[5]->seqn = 2 * size;
+	*rte_reorder_seqn(bufs[5]) = 2 * size;
 	ret = rte_reorder_insert(b, bufs[5]);
 	if (!((ret == -1) && (rte_errno == ENOSPC))) {
 		printf("%s:%d: No error inserting early packet with full ready buffer\n",
@@ -194,7 +194,7 @@ test_reorder_insert(void)
 	bufs[5] = NULL;
 
 	/* late packet */
-	bufs[6]->seqn = 3 * size;
+	*rte_reorder_seqn(bufs[6]) = 3 * size;
 	ret = rte_reorder_insert(b, bufs[6]);
 	if (!((ret == -1) && (rte_errno == ERANGE))) {
 		printf("%s:%d: No error inserting late packet with seqn:"
@@ -250,7 +250,7 @@ test_reorder_drain(void)
 	for (i = 0; i < num_bufs; i++) {
 		bufs[i] = rte_pktmbuf_alloc(p);
 		TEST_ASSERT_NOT_NULL(bufs[i], "Packet allocation failed\n");
-		bufs[i]->seqn = i;
+		*rte_reorder_seqn(bufs[i]) = i;
 	}
 
 	/* Insert packet with seqn 1:
diff --git a/examples/packet_ordering/main.c b/examples/packet_ordering/main.c
index a79d77a321..4bea1982d5 100644
--- a/examples/packet_ordering/main.c
+++ b/examples/packet_ordering/main.c
@@ -451,7 +451,7 @@ rx_thread(struct rte_ring *ring_out)
 
 				/* mark sequence number */
 				for (i = 0; i < nb_rx_pkts; )
-					pkts[i++]->seqn = seqn++;
+					*rte_reorder_seqn(pkts[i++]) = seqn++;
 
 				/* enqueue to rx_to_workers ring */
 				ret = rte_ring_enqueue_burst(ring_out,
diff --git a/lib/librte_reorder/rte_reorder.c b/lib/librte_reorder/rte_reorder.c
index 3c9f0e2d08..9445853b79 100644
--- a/lib/librte_reorder/rte_reorder.c
+++ b/lib/librte_reorder/rte_reorder.c
@@ -8,6 +8,7 @@
 #include <rte_string_fns.h>
 #include <rte_log.h>
 #include <rte_mbuf.h>
+#include <rte_mbuf_dyn.h>
 #include <rte_eal_memconfig.h>
 #include <rte_errno.h>
 #include <rte_malloc.h>
@@ -29,6 +30,9 @@ EAL_REGISTER_TAILQ(rte_reorder_tailq)
 /* Macros for printing using RTE_LOG */
 #define RTE_LOGTYPE_REORDER	RTE_LOGTYPE_USER1
 
+#define RTE_REORDER_SEQN_DYNFIELD_NAME "rte_reorder_seqn_dynfield"
+int rte_reorder_seqn_dynfield_offset = -1;
+
 /* A generic circular buffer */
 struct cir_buffer {
 	unsigned int size;   /**< Number of entries that can be stored */
@@ -103,6 +107,11 @@ rte_reorder_create(const char *name, unsigned socket_id, unsigned int size)
 	struct rte_reorder_list *reorder_list;
 	const unsigned int bufsize = sizeof(struct rte_reorder_buffer) +
 					(2 * size * sizeof(struct rte_mbuf *));
+	static const struct rte_mbuf_dynfield reorder_seqn_dynfield_desc = {
+		.name = RTE_REORDER_SEQN_DYNFIELD_NAME,
+		.size = sizeof(rte_reorder_seqn_t),
+		.align = __alignof__(rte_reorder_seqn_t),
+	};
 
 	reorder_list = RTE_TAILQ_CAST(rte_reorder_tailq.head, rte_reorder_list);
 
@@ -120,6 +129,14 @@ rte_reorder_create(const char *name, unsigned socket_id, unsigned int size)
 		return NULL;
 	}
 
+	rte_reorder_seqn_dynfield_offset =
+		rte_mbuf_dynfield_register(&reorder_seqn_dynfield_desc);
+	if (rte_reorder_seqn_dynfield_offset < 0) {
+		RTE_LOG(ERR, REORDER, "Failed to register mbuf field for reorder sequence number\n");
+		rte_errno = ENOMEM;
+		return NULL;
+	}
+
 	rte_mcfg_tailq_write_lock();
 
 	/* guarantee there's no existing */
@@ -310,7 +327,7 @@ rte_reorder_insert(struct rte_reorder_buffer *b, struct rte_mbuf *mbuf)
 
 	order_buf = &b->order_buf;
 	if (!b->is_initialized) {
-		b->min_seqn = mbuf->seqn;
+		b->min_seqn = *rte_reorder_seqn(mbuf);
 		b->is_initialized = 1;
 	}
 
@@ -322,7 +339,7 @@ rte_reorder_insert(struct rte_reorder_buffer *b, struct rte_mbuf *mbuf)
 	 *	mbuf_seqn = 0x0010
 	 *	offset    = 0x0010 - 0xFFFD = 0x13
 	 */
-	offset = mbuf->seqn - b->min_seqn;
+	offset = *rte_reorder_seqn(mbuf) - b->min_seqn;
 
 	/*
 	 * action to take depends on offset.
@@ -352,7 +369,7 @@ rte_reorder_insert(struct rte_reorder_buffer *b, struct rte_mbuf *mbuf)
 			rte_errno = ENOSPC;
 			return -1;
 		}
-		offset = mbuf->seqn - b->min_seqn;
+		offset = *rte_reorder_seqn(mbuf) - b->min_seqn;
 		position = (order_buf->head + offset) & order_buf->mask;
 		order_buf->entries[position] = mbuf;
 	} else {
diff --git a/lib/librte_reorder/rte_reorder.h b/lib/librte_reorder/rte_reorder.h
index 6d39710088..9de0240374 100644
--- a/lib/librte_reorder/rte_reorder.h
+++ b/lib/librte_reorder/rte_reorder.h
@@ -16,6 +16,7 @@
  */
 
 #include <rte_mbuf.h>
+#include <rte_mbuf_dyn.h>
 
 #ifdef __cplusplus
 extern "C" {
@@ -23,6 +24,26 @@ extern "C" {
 
 struct rte_reorder_buffer;
 
+typedef uint32_t rte_reorder_seqn_t;
+extern int rte_reorder_seqn_dynfield_offset;
+
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ *
+ * Read reorder sequence number from mbuf.
+ *
+ * @param mbuf Structure to read from.
+ * @return pointer to reorder sequence number.
+ */
+__rte_experimental
+static inline rte_reorder_seqn_t *
+rte_reorder_seqn(struct rte_mbuf *mbuf)
+{
+	return RTE_MBUF_DYNFIELD(mbuf, rte_reorder_seqn_dynfield_offset,
+		rte_reorder_seqn_t *);
+}
+
 /**
  * Create a new reorder buffer instance
  *
diff --git a/lib/librte_reorder/version.map b/lib/librte_reorder/version.map
index 8c0220d324..d902a7fa12 100644
--- a/lib/librte_reorder/version.map
+++ b/lib/librte_reorder/version.map
@@ -11,3 +11,9 @@ DPDK_21 {
 
 	local: *;
 };
+
+EXPERIMENTAL {
+	global:
+
+	rte_reorder_seqn_dynfield_offset;
+};
-- 
2.23.0

Re: [dpdk-dev] [PATCH v2 4/9] reorder: switch sequence number to dynamic mbuf field

[Andrew Rybchenko]

On 10/28/20 3:20 PM, David Marchand wrote:
> The reorder library used sequence numbers stored in the deprecated field
> seqn.
> It is moved to a dynamic mbuf field in order to allow removal of seqn.
> 
> Signed-off-by: David Marchand <david.marchand@redhat.com>

Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>

[dpdk-dev] [PATCH v2 5/9] dpaa: switch sequence number to dynamic mbuf field

[David Marchand]

The dpaa drivers have been hacking the deprecated field seqn for
internal features.
It is moved to a dynamic mbuf field in order to allow removal of seqn.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/bus/dpaa/dpaa_bus.c        | 16 ++++++++++++++++
 drivers/bus/dpaa/rte_dpaa_bus.h    | 28 ++++++++++++++++++++++++++++
 drivers/bus/dpaa/version.map       |  1 +
 drivers/crypto/dpaa_sec/dpaa_sec.c |  6 +++---
 drivers/event/dpaa/dpaa_eventdev.c |  6 +++---
 drivers/net/dpaa/dpaa_ethdev.h     |  7 -------
 drivers/net/dpaa/dpaa_rxtx.c       |  6 +++---
 7 files changed, 54 insertions(+), 16 deletions(-)

diff --git a/drivers/bus/dpaa/dpaa_bus.c b/drivers/bus/dpaa/dpaa_bus.c
index c94c72106f..ece6a4c424 100644
--- a/drivers/bus/dpaa/dpaa_bus.c
+++ b/drivers/bus/dpaa/dpaa_bus.c
@@ -32,6 +32,7 @@
 #include <rte_ring.h>
 #include <rte_bus.h>
 #include <rte_mbuf_pool_ops.h>
+#include <rte_mbuf_dyn.h>
 
 #include <dpaa_of.h>
 #include <rte_dpaa_bus.h>
@@ -55,6 +56,9 @@ unsigned int dpaa_svr_family;
 
 RTE_DEFINE_PER_LCORE(struct dpaa_portal *, dpaa_io);
 
+#define DPAA_SEQN_DYNFIELD_NAME "dpaa_seqn_dynfield"
+int dpaa_seqn_dynfield_offset = -1;
+
 struct fm_eth_port_cfg *
 dpaa_get_eth_port_cfg(int dev_id)
 {
@@ -251,6 +255,11 @@ dpaa_clean_device_list(void)
 
 int rte_dpaa_portal_init(void *arg)
 {
+	static const struct rte_mbuf_dynfield dpaa_seqn_dynfield_desc = {
+		.name = DPAA_SEQN_DYNFIELD_NAME,
+		.size = sizeof(dpaa_seqn_t),
+		.align = __alignof__(dpaa_seqn_t),
+	};
 	unsigned int cpu, lcore = rte_lcore_id();
 	int ret;
 
@@ -264,6 +273,13 @@ int rte_dpaa_portal_init(void *arg)
 
 	cpu = rte_lcore_to_cpu_id(lcore);
 
+	dpaa_seqn_dynfield_offset =
+		rte_mbuf_dynfield_register(&dpaa_seqn_dynfield_desc);
+	if (dpaa_seqn_dynfield_offset < 0) {
+		DPAA_BUS_LOG(ERR, "Failed to register mbuf field for dpaa sequence number\n");
+		return -rte_errno;
+	}
+
 	/* Initialise bman thread portals */
 	ret = bman_thread_init();
 	if (ret) {
diff --git a/drivers/bus/dpaa/rte_dpaa_bus.h b/drivers/bus/dpaa/rte_dpaa_bus.h
index fdaa63a09b..48d5cf4625 100644
--- a/drivers/bus/dpaa/rte_dpaa_bus.h
+++ b/drivers/bus/dpaa/rte_dpaa_bus.h
@@ -7,6 +7,7 @@
 #define __RTE_DPAA_BUS_H__
 
 #include <rte_bus.h>
+#include <rte_mbuf_dyn.h>
 #include <rte_mempool.h>
 #include <dpaax_iova_table.h>
 
@@ -16,6 +17,33 @@
 #include <fsl_bman.h>
 #include <netcfg.h>
 
+/* This sequence number field is used to store event entry index for
+ * driver specific usage. For parallel mode queues, invalid
+ * index will be set and for atomic mode queues, valid value
+ * ranging from 1 to 16.
+ */
+#define DPAA_INVALID_MBUF_SEQN  0
+
+typedef uint32_t dpaa_seqn_t;
+extern int dpaa_seqn_dynfield_offset;
+
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ *
+ * Read dpaa sequence number from mbuf.
+ *
+ * @param mbuf Structure to read from.
+ * @return pointer to dpaa sequence number.
+ */
+__rte_experimental
+static inline dpaa_seqn_t *
+dpaa_seqn(struct rte_mbuf *mbuf)
+{
+	return RTE_MBUF_DYNFIELD(mbuf, dpaa_seqn_dynfield_offset,
+		dpaa_seqn_t *);
+}
+
 #define DPAA_MEMPOOL_OPS_NAME	"dpaa"
 
 #define DEV_TO_DPAA_DEVICE(ptr)	\
diff --git a/drivers/bus/dpaa/version.map b/drivers/bus/dpaa/version.map
index 9bd2601213..fe4f9ac5aa 100644
--- a/drivers/bus/dpaa/version.map
+++ b/drivers/bus/dpaa/version.map
@@ -14,6 +14,7 @@ INTERNAL {
 	dpaa_get_qm_channel_pool;
 	dpaa_get_link_status;
 	dpaa_restart_link_autoneg;
+	dpaa_seqn_dynfield_offset;
 	dpaa_update_link_speed;
 	dpaa_intr_disable;
 	dpaa_intr_enable;
diff --git a/drivers/crypto/dpaa_sec/dpaa_sec.c b/drivers/crypto/dpaa_sec/dpaa_sec.c
index 55f457ac9a..44c742738f 100644
--- a/drivers/crypto/dpaa_sec/dpaa_sec.c
+++ b/drivers/crypto/dpaa_sec/dpaa_sec.c
@@ -1721,8 +1721,8 @@ dpaa_sec_enqueue_burst(void *qp, struct rte_crypto_op **ops,
 				DPAA_SEC_BURST : nb_ops;
 		for (loop = 0; loop < frames_to_send; loop++) {
 			op = *(ops++);
-			if (op->sym->m_src->seqn != 0) {
-				index = op->sym->m_src->seqn - 1;
+			if (*dpaa_seqn(op->sym->m_src) != 0) {
+				index = *dpaa_seqn(op->sym->m_src) - 1;
 				if (DPAA_PER_LCORE_DQRR_HELD & (1 << index)) {
 					/* QM_EQCR_DCA_IDXMASK = 0x0f */
 					flags[loop] = ((index & 0x0f) << 8);
@@ -3212,7 +3212,7 @@ dpaa_sec_process_atomic_event(void *event,
 	DPAA_PER_LCORE_DQRR_HELD |= 1 << index;
 	DPAA_PER_LCORE_DQRR_MBUF(index) = ctx->op->sym->m_src;
 	ev->impl_opaque = index + 1;
-	ctx->op->sym->m_src->seqn = (uint32_t)index + 1;
+	*dpaa_seqn(ctx->op->sym->m_src) = (uint32_t)index + 1;
 	*bufs = (void *)ctx->op;
 
 	rte_mempool_put(ctx->ctx_pool, (void *)ctx);
diff --git a/drivers/event/dpaa/dpaa_eventdev.c b/drivers/event/dpaa/dpaa_eventdev.c
index 07cd079768..01ddd0eb63 100644
--- a/drivers/event/dpaa/dpaa_eventdev.c
+++ b/drivers/event/dpaa/dpaa_eventdev.c
@@ -99,7 +99,7 @@ dpaa_event_enqueue_burst(void *port, const struct rte_event ev[],
 		case RTE_EVENT_OP_RELEASE:
 			qman_dca_index(ev[i].impl_opaque, 0);
 			mbuf = DPAA_PER_LCORE_DQRR_MBUF(i);
-			mbuf->seqn = DPAA_INVALID_MBUF_SEQN;
+			*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
 			DPAA_PER_LCORE_DQRR_HELD &= ~(1 << i);
 			DPAA_PER_LCORE_DQRR_SIZE--;
 			break;
@@ -206,7 +206,7 @@ dpaa_event_dequeue_burst(void *port, struct rte_event ev[],
 		if (DPAA_PER_LCORE_DQRR_HELD & (1 << i)) {
 			qman_dca_index(i, 0);
 			mbuf = DPAA_PER_LCORE_DQRR_MBUF(i);
-			mbuf->seqn = DPAA_INVALID_MBUF_SEQN;
+			*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
 			DPAA_PER_LCORE_DQRR_HELD &= ~(1 << i);
 			DPAA_PER_LCORE_DQRR_SIZE--;
 		}
@@ -276,7 +276,7 @@ dpaa_event_dequeue_burst_intr(void *port, struct rte_event ev[],
 		if (DPAA_PER_LCORE_DQRR_HELD & (1 << i)) {
 			qman_dca_index(i, 0);
 			mbuf = DPAA_PER_LCORE_DQRR_MBUF(i);
-			mbuf->seqn = DPAA_INVALID_MBUF_SEQN;
+			*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
 			DPAA_PER_LCORE_DQRR_HELD &= ~(1 << i);
 			DPAA_PER_LCORE_DQRR_SIZE--;
 		}
diff --git a/drivers/net/dpaa/dpaa_ethdev.h b/drivers/net/dpaa/dpaa_ethdev.h
index 1b8e120e8f..659bceb467 100644
--- a/drivers/net/dpaa/dpaa_ethdev.h
+++ b/drivers/net/dpaa/dpaa_ethdev.h
@@ -22,13 +22,6 @@
 #define DPAA_MBUF_HW_ANNOTATION		64
 #define DPAA_FD_PTA_SIZE		64
 
-/* mbuf->seqn will be used to store event entry index for
- * driver specific usage. For parallel mode queues, invalid
- * index will be set and for atomic mode queues, valid value
- * ranging from 1 to 16.
- */
-#define DPAA_INVALID_MBUF_SEQN  0
-
 /* we will re-use the HEADROOM for annotation in RX */
 #define DPAA_HW_BUF_RESERVE	0
 #define DPAA_PACKET_LAYOUT_ALIGN	64
diff --git a/drivers/net/dpaa/dpaa_rxtx.c b/drivers/net/dpaa/dpaa_rxtx.c
index e4f012c233..e2459d9b99 100644
--- a/drivers/net/dpaa/dpaa_rxtx.c
+++ b/drivers/net/dpaa/dpaa_rxtx.c
@@ -649,7 +649,7 @@ dpaa_rx_cb_parallel(void *event,
 	ev->queue_id = fq->ev.queue_id;
 	ev->priority = fq->ev.priority;
 	ev->impl_opaque = (uint8_t)DPAA_INVALID_MBUF_SEQN;
-	mbuf->seqn = DPAA_INVALID_MBUF_SEQN;
+	*dpaa_seqn(mbuf) = DPAA_INVALID_MBUF_SEQN;
 	*bufs = mbuf;
 
 	return qman_cb_dqrr_consume;
@@ -683,7 +683,7 @@ dpaa_rx_cb_atomic(void *event,
 	DPAA_PER_LCORE_DQRR_HELD |= 1 << index;
 	DPAA_PER_LCORE_DQRR_MBUF(index) = mbuf;
 	ev->impl_opaque = index + 1;
-	mbuf->seqn = (uint32_t)index + 1;
+	*dpaa_seqn(mbuf) = (uint32_t)index + 1;
 	*bufs = mbuf;
 
 	return qman_cb_dqrr_defer;
@@ -1078,7 +1078,7 @@ dpaa_eth_queue_tx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
 			if (dpaa_svr_family == SVR_LS1043A_FAMILY &&
 					(mbuf->data_off & 0x7F) != 0x0)
 				realloc_mbuf = 1;
-			seqn = mbuf->seqn;
+			seqn = *dpaa_seqn(mbuf);
 			if (seqn != DPAA_INVALID_MBUF_SEQN) {
 				index = seqn - 1;
 				if (DPAA_PER_LCORE_DQRR_HELD & (1 << index)) {
-- 
2.23.0

[dpdk-dev] [PATCH v2 6/9] fslmc: switch sequence number to dynamic mbuf field

[David Marchand]

The dpaa2 drivers have been hacking the deprecated field seqn for
internal features.
It is moved to a dynamic mbuf field in order to allow removal of seqn.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/bus/fslmc/fslmc_bus.c                 | 17 +++++++++++
 drivers/bus/fslmc/rte_fslmc.h                 | 23 +++++++++++++++
 drivers/bus/fslmc/version.map                 |  1 +
 drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c   | 18 ++++++------
 drivers/event/dpaa2/dpaa2_eventdev.c          |  9 +++---
 drivers/event/dpaa2/dpaa2_eventdev_selftest.c |  4 ++-
 drivers/mempool/dpaa2/dpaa2_hw_mempool.h      |  2 --
 drivers/net/dpaa2/dpaa2_rxtx.c                | 28 +++++++++----------
 8 files changed, 73 insertions(+), 29 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index beb3dd008f..db93669628 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -14,6 +14,7 @@
 #include <rte_devargs.h>
 #include <rte_memcpy.h>
 #include <rte_ethdev_driver.h>
+#include <rte_mbuf_dyn.h>
 
 #include <rte_fslmc.h>
 #include <fslmc_vfio.h>
@@ -27,6 +28,9 @@
 struct rte_fslmc_bus rte_fslmc_bus;
 uint8_t dpaa2_virt_mode;
 
+#define DPAA2_SEQN_DYNFIELD_NAME "dpaa2_seqn_dynfield"
+int dpaa2_seqn_dynfield_offset = -1;
+
 uint32_t
 rte_fslmc_get_device_count(enum rte_dpaa2_dev_type device_type)
 {
@@ -374,9 +378,22 @@ rte_fslmc_probe(void)
 	struct rte_dpaa2_device *dev;
 	struct rte_dpaa2_driver *drv;
 
+	static const struct rte_mbuf_dynfield dpaa2_seqn_dynfield_desc = {
+		.name = DPAA2_SEQN_DYNFIELD_NAME,
+		.size = sizeof(dpaa2_seqn_t),
+		.align = __alignof__(dpaa2_seqn_t),
+	};
+
 	if (TAILQ_EMPTY(&rte_fslmc_bus.device_list))
 		return 0;
 
+	dpaa2_seqn_dynfield_offset =
+		rte_mbuf_dynfield_register(&dpaa2_seqn_dynfield_desc);
+	if (dpaa2_seqn_dynfield_offset < 0) {
+		DPAA2_BUS_ERR("Failed to register mbuf field for dpaa sequence number");
+		return 0;
+	}
+
 	ret = fslmc_vfio_setup_group();
 	if (ret) {
 		DPAA2_BUS_ERR("Unable to setup VFIO %d", ret);
diff --git a/drivers/bus/fslmc/rte_fslmc.h b/drivers/bus/fslmc/rte_fslmc.h
index 80873fffc9..37d45dffe5 100644
--- a/drivers/bus/fslmc/rte_fslmc.h
+++ b/drivers/bus/fslmc/rte_fslmc.h
@@ -32,11 +32,34 @@ extern "C" {
 #include <rte_bus.h>
 #include <rte_tailq.h>
 #include <rte_devargs.h>
+#include <rte_mbuf.h>
+#include <rte_mbuf_dyn.h>
 
 #include <fslmc_vfio.h>
 
 #define FSLMC_OBJECT_MAX_LEN 32   /**< Length of each device on bus */
 
+#define DPAA2_INVALID_MBUF_SEQN        0
+
+typedef uint32_t dpaa2_seqn_t;
+extern int dpaa2_seqn_dynfield_offset;
+
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice
+ *
+ * Read dpaa2 sequence number from mbuf.
+ *
+ * @param mbuf Structure to read from.
+ * @return pointer to dpaa2 sequence number.
+ */
+__rte_experimental
+static inline dpaa2_seqn_t *
+dpaa2_seqn(struct rte_mbuf *mbuf)
+{
+	return RTE_MBUF_DYNFIELD(mbuf, dpaa2_seqn_dynfield_offset,
+		dpaa2_seqn_t *);
+}
 
 /** Device driver supports link state interrupt */
 #define RTE_DPAA2_DRV_INTR_LSC	0x0008
diff --git a/drivers/bus/fslmc/version.map b/drivers/bus/fslmc/version.map
index b169f5228a..f44c1a7988 100644
--- a/drivers/bus/fslmc/version.map
+++ b/drivers/bus/fslmc/version.map
@@ -19,6 +19,7 @@ INTERNAL {
 	dpaa2_free_eq_descriptors;
 	dpaa2_get_mcp_ptr;
 	dpaa2_io_portal;
+	dpaa2_seqn_dynfield_offset;
 	dpaa2_svr_family;
 	dpaa2_virt_mode;
 	dpbp_disable;
diff --git a/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c b/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
index afcd6bd063..ce1d50ce77 100644
--- a/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
+++ b/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
@@ -1472,13 +1472,15 @@ dpaa2_sec_enqueue_burst(void *qp, struct rte_crypto_op **ops,
 			dpaa2_eqcr_size : nb_ops;
 
 		for (loop = 0; loop < frames_to_send; loop++) {
-			if ((*ops)->sym->m_src->seqn) {
-			 uint8_t dqrr_index = (*ops)->sym->m_src->seqn - 1;
-
-			 flags[loop] = QBMAN_ENQUEUE_FLAG_DCA | dqrr_index;
-			 DPAA2_PER_LCORE_DQRR_SIZE--;
-			 DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dqrr_index);
-			 (*ops)->sym->m_src->seqn = DPAA2_INVALID_MBUF_SEQN;
+			if (*dpaa2_seqn((*ops)->sym->m_src)) {
+				uint8_t dqrr_index =
+					*dpaa2_seqn((*ops)->sym->m_src) - 1;
+
+				flags[loop] = QBMAN_ENQUEUE_FLAG_DCA | dqrr_index;
+				DPAA2_PER_LCORE_DQRR_SIZE--;
+				DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dqrr_index);
+				*dpaa2_seqn((*ops)->sym->m_src) =
+					DPAA2_INVALID_MBUF_SEQN;
 			}
 
 			/*Clear the unused FD fields before sending*/
@@ -3714,7 +3716,7 @@ dpaa2_sec_process_atomic_event(struct qbman_swp *swp __rte_unused,
 
 	ev->event_ptr = sec_fd_to_mbuf(fd);
 	dqrr_index = qbman_get_dqrr_idx(dq);
-	crypto_op->sym->m_src->seqn = dqrr_index + 1;
+	*dpaa2_seqn(crypto_op->sym->m_src) = dqrr_index + 1;
 	DPAA2_PER_LCORE_DQRR_SIZE++;
 	DPAA2_PER_LCORE_DQRR_HELD |= 1 << dqrr_index;
 	DPAA2_PER_LCORE_DQRR_MBUF(dqrr_index) = crypto_op->sym->m_src;
diff --git a/drivers/event/dpaa2/dpaa2_eventdev.c b/drivers/event/dpaa2/dpaa2_eventdev.c
index 95f03c8b9e..eeb2494bd0 100644
--- a/drivers/event/dpaa2/dpaa2_eventdev.c
+++ b/drivers/event/dpaa2/dpaa2_eventdev.c
@@ -131,8 +131,9 @@ dpaa2_eventdev_enqueue_burst(void *port, const struct rte_event ev[],
 			qbman_eq_desc_set_response(&eqdesc[loop], 0, 0);
 
 			if (event->sched_type == RTE_SCHED_TYPE_ATOMIC
-				&& event->mbuf->seqn) {
-				uint8_t dqrr_index = event->mbuf->seqn - 1;
+				&& *dpaa2_seqn(event->mbuf)) {
+				uint8_t dqrr_index =
+					*dpaa2_seqn(event->mbuf) - 1;
 
 				qbman_eq_desc_set_dca(&eqdesc[loop], 1,
 						      dqrr_index, 0);
@@ -249,7 +250,7 @@ static void dpaa2_eventdev_process_atomic(struct qbman_swp *swp,
 
 	rte_memcpy(ev, ev_temp, sizeof(struct rte_event));
 	rte_free(ev_temp);
-	ev->mbuf->seqn = dqrr_index + 1;
+	*dpaa2_seqn(ev->mbuf) = dqrr_index + 1;
 	DPAA2_PER_LCORE_DQRR_SIZE++;
 	DPAA2_PER_LCORE_DQRR_HELD |= 1 << dqrr_index;
 	DPAA2_PER_LCORE_DQRR_MBUF(dqrr_index) = ev->mbuf;
@@ -314,7 +315,7 @@ dpaa2_eventdev_dequeue_burst(void *port, struct rte_event ev[],
 		if (DPAA2_PER_LCORE_DQRR_HELD & (1 << i)) {
 			qbman_swp_dqrr_idx_consume(swp, i);
 			DPAA2_PER_LCORE_DQRR_SIZE--;
-			DPAA2_PER_LCORE_DQRR_MBUF(i)->seqn =
+			*dpaa2_seqn(DPAA2_PER_LCORE_DQRR_MBUF(i)) =
 				DPAA2_INVALID_MBUF_SEQN;
 		}
 		i++;
diff --git a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
index 5447db8a8a..cd7311a94d 100644
--- a/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
+++ b/drivers/event/dpaa2/dpaa2_eventdev_selftest.c
@@ -19,6 +19,7 @@
 #include <rte_random.h>
 #include <rte_bus_vdev.h>
 #include <rte_test.h>
+#include <rte_fslmc.h>
 
 #include "dpaa2_eventdev.h"
 #include "dpaa2_eventdev_logs.h"
@@ -274,7 +275,8 @@ check_excess_events(uint8_t port)
 		valid_event = rte_event_dequeue_burst(evdev, port, &ev, 1, 0);
 
 		RTE_TEST_ASSERT_SUCCESS(valid_event,
-				"Unexpected valid event=%d", ev.mbuf->seqn);
+				"Unexpected valid event=%d",
+				*dpaa2_seqn(ev.mbuf));
 	}
 	return 0;
 }
diff --git a/drivers/mempool/dpaa2/dpaa2_hw_mempool.h b/drivers/mempool/dpaa2/dpaa2_hw_mempool.h
index 53fa1552d1..7c493b28e7 100644
--- a/drivers/mempool/dpaa2/dpaa2_hw_mempool.h
+++ b/drivers/mempool/dpaa2/dpaa2_hw_mempool.h
@@ -10,8 +10,6 @@
 
 #define DPAA2_MAX_BUF_POOLS	8
 
-#define DPAA2_INVALID_MBUF_SEQN	0
-
 struct buf_pool_cfg {
 	void *addr;
 	/**< The address from where DPAA2 will carve out the buffers */
diff --git a/drivers/net/dpaa2/dpaa2_rxtx.c b/drivers/net/dpaa2/dpaa2_rxtx.c
index 4dd1d5f578..6201de4606 100644
--- a/drivers/net/dpaa2/dpaa2_rxtx.c
+++ b/drivers/net/dpaa2/dpaa2_rxtx.c
@@ -710,7 +710,7 @@ dpaa2_dev_process_atomic_event(struct qbman_swp *swp __rte_unused,
 	ev->mbuf = eth_fd_to_mbuf(fd, rxq->eth_data->port_id);
 
 	dqrr_index = qbman_get_dqrr_idx(dq);
-	ev->mbuf->seqn = dqrr_index + 1;
+	*dpaa2_seqn(ev->mbuf) = dqrr_index + 1;
 	DPAA2_PER_LCORE_DQRR_SIZE++;
 	DPAA2_PER_LCORE_DQRR_HELD |= 1 << dqrr_index;
 	DPAA2_PER_LCORE_DQRR_MBUF(dqrr_index) = ev->mbuf;
@@ -736,9 +736,9 @@ dpaa2_dev_process_ordered_event(struct qbman_swp *swp,
 
 	ev->mbuf = eth_fd_to_mbuf(fd, rxq->eth_data->port_id);
 
-	ev->mbuf->seqn = DPAA2_ENQUEUE_FLAG_ORP;
-	ev->mbuf->seqn |= qbman_result_DQ_odpid(dq) << DPAA2_EQCR_OPRID_SHIFT;
-	ev->mbuf->seqn |= qbman_result_DQ_seqnum(dq) << DPAA2_EQCR_SEQNUM_SHIFT;
+	*dpaa2_seqn(ev->mbuf) = DPAA2_ENQUEUE_FLAG_ORP;
+	*dpaa2_seqn(ev->mbuf) |= qbman_result_DQ_odpid(dq) << DPAA2_EQCR_OPRID_SHIFT;
+	*dpaa2_seqn(ev->mbuf) |= qbman_result_DQ_seqnum(dq) << DPAA2_EQCR_SEQNUM_SHIFT;
 
 	qbman_swp_dqrr_consume(swp, dq);
 }
@@ -1063,14 +1063,14 @@ dpaa2_dev_tx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
 			dpaa2_eqcr_size : nb_pkts;
 
 		for (loop = 0; loop < frames_to_send; loop++) {
-			if ((*bufs)->seqn) {
-				uint8_t dqrr_index = (*bufs)->seqn - 1;
+			if (*dpaa2_seqn(*bufs)) {
+				uint8_t dqrr_index = *dpaa2_seqn(*bufs) - 1;
 
 				flags[loop] = QBMAN_ENQUEUE_FLAG_DCA |
 						dqrr_index;
 				DPAA2_PER_LCORE_DQRR_SIZE--;
 				DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dqrr_index);
-				(*bufs)->seqn = DPAA2_INVALID_MBUF_SEQN;
+				*dpaa2_seqn(*bufs) = DPAA2_INVALID_MBUF_SEQN;
 			}
 
 			if (likely(RTE_MBUF_DIRECT(*bufs))) {
@@ -1230,10 +1230,10 @@ dpaa2_set_enqueue_descriptor(struct dpaa2_queue *dpaa2_q,
 
 	qbman_eq_desc_set_fq(eqdesc, dpaa2_q->fqid);
 
-	if (m->seqn & DPAA2_ENQUEUE_FLAG_ORP) {
-		orpid = (m->seqn & DPAA2_EQCR_OPRID_MASK) >>
+	if (*dpaa2_seqn(m) & DPAA2_ENQUEUE_FLAG_ORP) {
+		orpid = (*dpaa2_seqn(m) & DPAA2_EQCR_OPRID_MASK) >>
 			DPAA2_EQCR_OPRID_SHIFT;
-		seqnum = (m->seqn & DPAA2_EQCR_SEQNUM_MASK) >>
+		seqnum = (*dpaa2_seqn(m) & DPAA2_EQCR_SEQNUM_MASK) >>
 			DPAA2_EQCR_SEQNUM_SHIFT;
 
 		if (!priv->en_loose_ordered) {
@@ -1255,12 +1255,12 @@ dpaa2_set_enqueue_descriptor(struct dpaa2_queue *dpaa2_q,
 			qbman_eq_desc_set_orp(eqdesc, 0, orpid, seqnum, 0);
 		}
 	} else {
-		dq_idx = m->seqn - 1;
+		dq_idx = *dpaa2_seqn(m) - 1;
 		qbman_eq_desc_set_dca(eqdesc, 1, dq_idx, 0);
 		DPAA2_PER_LCORE_DQRR_SIZE--;
 		DPAA2_PER_LCORE_DQRR_HELD &= ~(1 << dq_idx);
 	}
-	m->seqn = DPAA2_INVALID_MBUF_SEQN;
+	*dpaa2_seqn(m) = DPAA2_INVALID_MBUF_SEQN;
 }
 
 /* Callback to handle sending ordered packets through WRIOP based interface */
@@ -1314,7 +1314,7 @@ dpaa2_dev_tx_ordered(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
 			dpaa2_eqcr_size : nb_pkts;
 
 		if (!priv->en_loose_ordered) {
-			if ((*bufs)->seqn & DPAA2_ENQUEUE_FLAG_ORP) {
+			if (*dpaa2_seqn(*bufs) & DPAA2_ENQUEUE_FLAG_ORP) {
 				num_free_eq_desc = dpaa2_free_eq_descriptors();
 				if (num_free_eq_desc < frames_to_send)
 					frames_to_send = num_free_eq_desc;
@@ -1325,7 +1325,7 @@ dpaa2_dev_tx_ordered(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
 			/*Prepare enqueue descriptor*/
 			qbman_eq_desc_clear(&eqdesc[loop]);
 
-			if ((*bufs)->seqn) {
+			if (*dpaa2_seqn(*bufs)) {
 				/* Use only queue 0 for Tx in case of atomic/
 				 * ordered packets as packets can get unordered
 				 * when being tranmitted out from the interface
-- 
2.23.0

[dpdk-dev] [PATCH v2 7/9] eventdev: switch sequence number to dynamic mbuf field

[David Marchand]

The eventdev drivers have been hacking the deprecated field seqn for
internal test usage.
It is moved to a dynamic mbuf field in order to allow removal of seqn.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
Changelog since v1:
- split eventdev test app changes from this patch (moved to next patch),
- moved seqn as a pmd only API, fixed namespace and marked as internal,

---
 drivers/event/octeontx/ssovf_evdev_selftest.c | 32 +++++++++--------
 drivers/event/octeontx2/otx2_evdev_selftest.c | 36 +++++++++++--------
 drivers/event/opdl/opdl_test.c                |  8 ++---
 drivers/event/sw/sw_evdev_selftest.c          | 34 +++++++++---------
 lib/librte_eventdev/rte_eventdev.c            | 14 +++++++-
 lib/librte_eventdev/rte_eventdev_pmd.h        | 20 +++++++++++
 lib/librte_eventdev/version.map               |  6 ++++
 7 files changed, 99 insertions(+), 51 deletions(-)

diff --git a/drivers/event/octeontx/ssovf_evdev_selftest.c b/drivers/event/octeontx/ssovf_evdev_selftest.c
index 7a2b7ded25..528f99dd84 100644
--- a/drivers/event/octeontx/ssovf_evdev_selftest.c
+++ b/drivers/event/octeontx/ssovf_evdev_selftest.c
@@ -300,7 +300,7 @@ inject_events(uint32_t flow_id, uint8_t event_type, uint8_t sub_event_type,
 		m = rte_pktmbuf_alloc(eventdev_test_mempool);
 		RTE_TEST_ASSERT_NOT_NULL(m, "mempool alloc failed");
 
-		m->seqn = i;
+		*rte_event_pmd_selftest_seqn(m) = i;
 		update_event_and_validation_attr(m, &ev, flow_id, event_type,
 			sub_event_type, sched_type, queue, port);
 		rte_event_enqueue_burst(evdev, port, &ev, 1);
@@ -320,7 +320,8 @@ check_excess_events(uint8_t port)
 		valid_event = rte_event_dequeue_burst(evdev, port, &ev, 1, 0);
 
 		RTE_TEST_ASSERT_SUCCESS(valid_event,
-				"Unexpected valid event=%d", ev.mbuf->seqn);
+			"Unexpected valid event=%d",
+			*rte_event_pmd_selftest_seqn(ev.mbuf));
 	}
 	return 0;
 }
@@ -425,8 +426,9 @@ static int
 validate_simple_enqdeq(uint32_t index, uint8_t port, struct rte_event *ev)
 {
 	RTE_SET_USED(port);
-	RTE_TEST_ASSERT_EQUAL(index, ev->mbuf->seqn, "index=%d != seqn=%d",
-			index, ev->mbuf->seqn);
+	RTE_TEST_ASSERT_EQUAL(index, *rte_event_pmd_selftest_seqn(ev->mbuf),
+		"index=%d != seqn=%d", index,
+		*rte_event_pmd_selftest_seqn(ev->mbuf));
 	return 0;
 }
 
@@ -509,10 +511,10 @@ validate_queue_priority(uint32_t index, uint8_t port, struct rte_event *ev)
 
 	expected_val += ev->queue_id;
 	RTE_SET_USED(port);
-	RTE_TEST_ASSERT_EQUAL(ev->mbuf->seqn, expected_val,
-	"seqn=%d index=%d expected=%d range=%d nb_queues=%d max_event=%d",
-			ev->mbuf->seqn, index, expected_val, range,
-			queue_count, MAX_EVENTS);
+	RTE_TEST_ASSERT_EQUAL(*rte_event_pmd_selftest_seqn(ev->mbuf), expected_val,
+		"seqn=%d index=%d expected=%d range=%d nb_queues=%d max_event=%d",
+		*rte_event_pmd_selftest_seqn(ev->mbuf), index, expected_val, range,
+		queue_count, MAX_EVENTS);
 	return 0;
 }
 
@@ -537,7 +539,7 @@ test_multi_queue_priority(void)
 		m = rte_pktmbuf_alloc(eventdev_test_mempool);
 		RTE_TEST_ASSERT_NOT_NULL(m, "mempool alloc failed");
 
-		m->seqn = i;
+		*rte_event_pmd_selftest_seqn(m) = i;
 		queue = i % queue_count;
 		update_event_and_validation_attr(m, &ev, 0, RTE_EVENT_TYPE_CPU,
 			0, RTE_SCHED_TYPE_PARALLEL, queue, 0);
@@ -904,7 +906,7 @@ worker_flow_based_pipeline(void *arg)
 			ev.op = RTE_EVENT_OP_FORWARD;
 			rte_event_enqueue_burst(evdev, port, &ev, 1);
 		} else if (ev.sub_event_type == 1) { /* Events from stage 1*/
-			if (seqn_list_update(ev.mbuf->seqn) == 0) {
+			if (seqn_list_update(*rte_event_pmd_selftest_seqn(ev.mbuf)) == 0) {
 				rte_pktmbuf_free(ev.mbuf);
 				rte_atomic32_sub(total_events, 1);
 			} else {
@@ -939,7 +941,7 @@ test_multiport_flow_sched_type_test(uint8_t in_sched_type,
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(
 		0x1 /*flow_id */,
 		RTE_EVENT_TYPE_CPU /* event_type */,
@@ -1059,7 +1061,7 @@ worker_group_based_pipeline(void *arg)
 			ev.op = RTE_EVENT_OP_FORWARD;
 			rte_event_enqueue_burst(evdev, port, &ev, 1);
 		} else if (ev.queue_id == 1) { /* Events from stage 1(group 1)*/
-			if (seqn_list_update(ev.mbuf->seqn) == 0) {
+			if (seqn_list_update(*rte_event_pmd_selftest_seqn(ev.mbuf)) == 0) {
 				rte_pktmbuf_free(ev.mbuf);
 				rte_atomic32_sub(total_events, 1);
 			} else {
@@ -1101,7 +1103,7 @@ test_multiport_queue_sched_type_test(uint8_t in_sched_type,
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(
 		0x1 /*flow_id */,
 		RTE_EVENT_TYPE_CPU /* event_type */,
@@ -1238,7 +1240,7 @@ launch_multi_port_max_stages_random_sched_type(int (*fn)(void *))
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(
 		0x1 /*flow_id */,
 		RTE_EVENT_TYPE_CPU /* event_type */,
@@ -1360,7 +1362,7 @@ worker_ordered_flow_producer(void *arg)
 		if (m == NULL)
 			continue;
 
-		m->seqn = counter++;
+		*rte_event_pmd_selftest_seqn(m) = counter++;
 
 		struct rte_event ev = {.event = 0, .u64 = 0};
 
diff --git a/drivers/event/octeontx2/otx2_evdev_selftest.c b/drivers/event/octeontx2/otx2_evdev_selftest.c
index 334a9ccb7c..48bfaf893d 100644
--- a/drivers/event/octeontx2/otx2_evdev_selftest.c
+++ b/drivers/event/octeontx2/otx2_evdev_selftest.c
@@ -279,7 +279,7 @@ inject_events(uint32_t flow_id, uint8_t event_type, uint8_t sub_event_type,
 		m = rte_pktmbuf_alloc(eventdev_test_mempool);
 		RTE_TEST_ASSERT_NOT_NULL(m, "mempool alloc failed");
 
-		m->seqn = i;
+		*rte_event_pmd_selftest_seqn(m) = i;
 		update_event_and_validation_attr(m, &ev, flow_id, event_type,
 						 sub_event_type, sched_type,
 						 queue, port);
@@ -301,7 +301,7 @@ check_excess_events(uint8_t port)
 
 		RTE_TEST_ASSERT_SUCCESS(valid_event,
 					"Unexpected valid event=%d",
-					ev.mbuf->seqn);
+					*rte_event_pmd_selftest_seqn(ev.mbuf));
 	}
 	return 0;
 }
@@ -406,8 +406,9 @@ static int
 validate_simple_enqdeq(uint32_t index, uint8_t port, struct rte_event *ev)
 {
 	RTE_SET_USED(port);
-	RTE_TEST_ASSERT_EQUAL(index, ev->mbuf->seqn, "index=%d != seqn=%d",
-			      index, ev->mbuf->seqn);
+	RTE_TEST_ASSERT_EQUAL(index, *rte_event_pmd_selftest_seqn(ev->mbuf),
+		"index=%d != seqn=%d",
+		index, *rte_event_pmd_selftest_seqn(ev->mbuf));
 	return 0;
 }
 
@@ -493,10 +494,11 @@ validate_queue_priority(uint32_t index, uint8_t port, struct rte_event *ev)
 
 	expected_val += ev->queue_id;
 	RTE_SET_USED(port);
-	RTE_TEST_ASSERT_EQUAL(ev->mbuf->seqn, expected_val,
-	"seqn=%d index=%d expected=%d range=%d nb_queues=%d max_event=%d",
-			      ev->mbuf->seqn, index, expected_val, range,
-			      queue_count, MAX_EVENTS);
+	RTE_TEST_ASSERT_EQUAL(
+		*rte_event_pmd_selftest_seqn(ev->mbuf), expected_val,
+		"seqn=%d index=%d expected=%d range=%d nb_queues=%d max_event=%d",
+		*rte_event_pmd_selftest_seqn(ev->mbuf), index, expected_val,
+		range, queue_count, MAX_EVENTS);
 	return 0;
 }
 
@@ -523,7 +525,7 @@ test_multi_queue_priority(void)
 		m = rte_pktmbuf_alloc(eventdev_test_mempool);
 		RTE_TEST_ASSERT_NOT_NULL(m, "mempool alloc failed");
 
-		m->seqn = i;
+		*rte_event_pmd_selftest_seqn(m) = i;
 		queue = i % queue_count;
 		update_event_and_validation_attr(m, &ev, 0, RTE_EVENT_TYPE_CPU,
 						 0, RTE_SCHED_TYPE_PARALLEL,
@@ -888,7 +890,9 @@ worker_flow_based_pipeline(void *arg)
 			ev.op = RTE_EVENT_OP_FORWARD;
 			rte_event_enqueue_burst(evdev, port, &ev, 1);
 		} else if (ev.sub_event_type == 1) { /* Events from stage 1*/
-			if (seqn_list_update(ev.mbuf->seqn) == 0) {
+			uint32_t seqn = *rte_event_pmd_selftest_seqn(ev.mbuf);
+
+			if (seqn_list_update(seqn) == 0) {
 				rte_pktmbuf_free(ev.mbuf);
 				rte_atomic32_sub(total_events, 1);
 			} else {
@@ -923,7 +927,7 @@ test_multiport_flow_sched_type_test(uint8_t in_sched_type,
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(0x1 /*flow_id */,
 			    RTE_EVENT_TYPE_CPU /* event_type */,
 			    0 /* sub_event_type (stage 0) */,
@@ -1043,7 +1047,9 @@ worker_group_based_pipeline(void *arg)
 			ev.op = RTE_EVENT_OP_FORWARD;
 			rte_event_enqueue_burst(evdev, port, &ev, 1);
 		} else if (ev.queue_id == 1) { /* Events from stage 1(group 1)*/
-			if (seqn_list_update(ev.mbuf->seqn) == 0) {
+			uint32_t seqn = *rte_event_pmd_selftest_seqn(ev.mbuf);
+
+			if (seqn_list_update(seqn) == 0) {
 				rte_pktmbuf_free(ev.mbuf);
 				rte_atomic32_sub(total_events, 1);
 			} else {
@@ -1084,7 +1090,7 @@ test_multiport_queue_sched_type_test(uint8_t in_sched_type,
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(0x1 /*flow_id */,
 			    RTE_EVENT_TYPE_CPU /* event_type */,
 			    0 /* sub_event_type (stage 0) */,
@@ -1222,7 +1228,7 @@ launch_multi_port_max_stages_random_sched_type(int (*fn)(void *))
 		return 0;
 	}
 
-	/* Injects events with m->seqn=0 to total_events */
+	/* Injects events with a 0 sequence number to total_events */
 	ret = inject_events(0x1 /*flow_id */,
 			    RTE_EVENT_TYPE_CPU /* event_type */,
 			    0 /* sub_event_type (stage 0) */,
@@ -1348,7 +1354,7 @@ worker_ordered_flow_producer(void *arg)
 		if (m == NULL)
 			continue;
 
-		m->seqn = counter++;
+		*rte_event_pmd_selftest_seqn(m) = counter++;
 
 		struct rte_event ev = {.event = 0, .u64 = 0};
 
diff --git a/drivers/event/opdl/opdl_test.c b/drivers/event/opdl/opdl_test.c
index e7a32fbd31..e4fc70a440 100644
--- a/drivers/event/opdl/opdl_test.c
+++ b/drivers/event/opdl/opdl_test.c
@@ -256,7 +256,7 @@ ordered_basic(struct test *t)
 		ev.queue_id = t->qid[0];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = mbufs[i];
-		mbufs[i]->seqn = MAGIC_SEQN + i;
+		*rte_event_pmd_selftest_seqn(mbufs[i]) = MAGIC_SEQN + i;
 
 		/* generate pkt and enqueue */
 		err = rte_event_enqueue_burst(evdev, t->port[rx_port], &ev, 1);
@@ -281,7 +281,7 @@ ordered_basic(struct test *t)
 			rte_event_dev_dump(evdev, stdout);
 			return -1;
 		}
-		seq = deq_ev[i].mbuf->seqn  - MAGIC_SEQN;
+		seq = *rte_event_pmd_selftest_seqn(deq_ev[i].mbuf)  - MAGIC_SEQN;
 
 		if (seq != (i-1)) {
 			PMD_DRV_LOG(ERR, " seq test failed ! eq is %d , "
@@ -396,7 +396,7 @@ atomic_basic(struct test *t)
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.flow_id = 1;
 		ev.mbuf = mbufs[i];
-		mbufs[i]->seqn = MAGIC_SEQN + i;
+		*rte_event_pmd_selftest_seqn(mbufs[i]) = MAGIC_SEQN + i;
 
 		/* generate pkt and enqueue */
 		err = rte_event_enqueue_burst(evdev, t->port[rx_port], &ev, 1);
@@ -625,7 +625,7 @@ single_link_w_stats(struct test *t)
 		ev.queue_id = t->qid[0];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = mbufs[i];
-		mbufs[i]->seqn = 1234 + i;
+		*rte_event_pmd_selftest_seqn(mbufs[i]) = 1234 + i;
 
 		/* generate pkt and enqueue */
 		err = rte_event_enqueue_burst(evdev, t->port[rx_port], &ev, 1);
diff --git a/drivers/event/sw/sw_evdev_selftest.c b/drivers/event/sw/sw_evdev_selftest.c
index ad4fc0eed7..4555f799df 100644
--- a/drivers/event/sw/sw_evdev_selftest.c
+++ b/drivers/event/sw/sw_evdev_selftest.c
@@ -380,7 +380,7 @@ run_prio_packet_test(struct test *t)
 			printf("%d: gen of pkt failed\n", __LINE__);
 			return -1;
 		}
-		arp->seqn = MAGIC_SEQN[i];
+		*rte_event_pmd_selftest_seqn(arp) = MAGIC_SEQN[i];
 
 		ev = (struct rte_event){
 			.priority = PRIORITY[i],
@@ -419,7 +419,7 @@ run_prio_packet_test(struct test *t)
 		rte_event_dev_dump(evdev, stdout);
 		return -1;
 	}
-	if (ev.mbuf->seqn != MAGIC_SEQN[1]) {
+	if (*rte_event_pmd_selftest_seqn(ev.mbuf) != MAGIC_SEQN[1]) {
 		printf("%d: first packet out not highest priority\n",
 				__LINE__);
 		rte_event_dev_dump(evdev, stdout);
@@ -433,7 +433,7 @@ run_prio_packet_test(struct test *t)
 		rte_event_dev_dump(evdev, stdout);
 		return -1;
 	}
-	if (ev2.mbuf->seqn != MAGIC_SEQN[0]) {
+	if (*rte_event_pmd_selftest_seqn(ev2.mbuf) != MAGIC_SEQN[0]) {
 		printf("%d: second packet out not lower priority\n",
 				__LINE__);
 		rte_event_dev_dump(evdev, stdout);
@@ -477,7 +477,7 @@ test_single_directed_packet(struct test *t)
 	}
 
 	const uint32_t MAGIC_SEQN = 4711;
-	arp->seqn = MAGIC_SEQN;
+	*rte_event_pmd_selftest_seqn(arp) = MAGIC_SEQN;
 
 	/* generate pkt and enqueue */
 	err = rte_event_enqueue_burst(evdev, rx_enq, &ev, 1);
@@ -516,7 +516,7 @@ test_single_directed_packet(struct test *t)
 		return -1;
 	}
 
-	if (ev.mbuf->seqn != MAGIC_SEQN) {
+	if (*rte_event_pmd_selftest_seqn(ev.mbuf) != MAGIC_SEQN) {
 		printf("%d: error magic sequence number not dequeued\n",
 				__LINE__);
 		return -1;
@@ -934,7 +934,7 @@ xstats_tests(struct test *t)
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = arp;
 		ev.flow_id = 7;
-		arp->seqn = i;
+		*rte_event_pmd_selftest_seqn(arp) = i;
 
 		int err = rte_event_enqueue_burst(evdev, t->port[0], &ev, 1);
 		if (err != 1) {
@@ -1485,7 +1485,7 @@ xstats_id_reset_tests(struct test *t)
 		ev.queue_id = t->qid[i];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = arp;
-		arp->seqn = i;
+		*rte_event_pmd_selftest_seqn(arp) = i;
 
 		int err = rte_event_enqueue_burst(evdev, t->port[0], &ev, 1);
 		if (err != 1) {
@@ -1873,7 +1873,7 @@ qid_priorities(struct test *t)
 		ev.queue_id = t->qid[i];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = arp;
-		arp->seqn = i;
+		*rte_event_pmd_selftest_seqn(arp) = i;
 
 		int err = rte_event_enqueue_burst(evdev, t->port[0], &ev, 1);
 		if (err != 1) {
@@ -1894,7 +1894,7 @@ qid_priorities(struct test *t)
 		return -1;
 	}
 	for (i = 0; i < 3; i++) {
-		if (ev[i].mbuf->seqn != 2-i) {
+		if (*rte_event_pmd_selftest_seqn(ev[i].mbuf) != 2-i) {
 			printf(
 				"%d: qid priority test: seqn %d incorrectly prioritized\n",
 					__LINE__, i);
@@ -2371,7 +2371,7 @@ single_packet(struct test *t)
 	ev.mbuf = arp;
 	ev.queue_id = 0;
 	ev.flow_id = 3;
-	arp->seqn = MAGIC_SEQN;
+	*rte_event_pmd_selftest_seqn(arp) = MAGIC_SEQN;
 
 	err = rte_event_enqueue_burst(evdev, t->port[rx_enq], &ev, 1);
 	if (err != 1) {
@@ -2411,7 +2411,7 @@ single_packet(struct test *t)
 	}
 
 	err = test_event_dev_stats_get(evdev, &stats);
-	if (ev.mbuf->seqn != MAGIC_SEQN) {
+	if (*rte_event_pmd_selftest_seqn(ev.mbuf) != MAGIC_SEQN) {
 		printf("%d: magic sequence number not dequeued\n", __LINE__);
 		return -1;
 	}
@@ -2684,7 +2684,7 @@ parallel_basic(struct test *t, int check_order)
 		ev.queue_id = t->qid[0];
 		ev.op = RTE_EVENT_OP_NEW;
 		ev.mbuf = mbufs[i];
-		mbufs[i]->seqn = MAGIC_SEQN + i;
+		*rte_event_pmd_selftest_seqn(mbufs[i]) = MAGIC_SEQN + i;
 
 		/* generate pkt and enqueue */
 		err = rte_event_enqueue_burst(evdev, t->port[rx_port], &ev, 1);
@@ -2739,10 +2739,12 @@ parallel_basic(struct test *t, int check_order)
 	/* Check to see if the sequence numbers are in expected order */
 	if (check_order) {
 		for (j = 0 ; j < deq_pkts ; j++) {
-			if (deq_ev[j].mbuf->seqn != MAGIC_SEQN + j) {
-				printf(
-					"%d: Incorrect sequence number(%d) from port %d\n",
-					__LINE__, mbufs_out[j]->seqn, tx_port);
+			if (*rte_event_pmd_selftest_seqn(deq_ev[j].mbuf) !=
+					MAGIC_SEQN + j) {
+				printf("%d: Incorrect sequence number(%d) from port %d\n",
+					__LINE__,
+					*rte_event_pmd_selftest_seqn(mbufs_out[j]),
+					tx_port);
 				return -1;
 			}
 		}
diff --git a/lib/librte_eventdev/rte_eventdev.c b/lib/librte_eventdev/rte_eventdev.c
index 322453c532..994bd1eaa9 100644
--- a/lib/librte_eventdev/rte_eventdev.c
+++ b/lib/librte_eventdev/rte_eventdev.c
@@ -1242,13 +1242,25 @@ int rte_event_dev_xstats_reset(uint8_t dev_id,
 	return -ENOTSUP;
 }
 
+int rte_event_pmd_selftest_seqn_dynfield_offset = -1;
+
 int rte_event_dev_selftest(uint8_t dev_id)
 {
 	RTE_EVENTDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+	static const struct rte_mbuf_dynfield test_seqn_dynfield_desc = {
+		.name = "rte_event_pmd_selftest_seqn_dynfield",
+		.size = sizeof(rte_event_pmd_selftest_seqn_t),
+		.align = __alignof__(rte_event_pmd_selftest_seqn_t),
+	};
 	struct rte_eventdev *dev = &rte_eventdevs[dev_id];
 
-	if (dev->dev_ops->dev_selftest != NULL)
+	if (dev->dev_ops->dev_selftest != NULL) {
+		rte_event_pmd_selftest_seqn_dynfield_offset =
+			rte_mbuf_dynfield_register(&test_seqn_dynfield_desc);
+		if (rte_event_pmd_selftest_seqn_dynfield_offset < 0)
+			return -ENOMEM;
 		return (*dev->dev_ops->dev_selftest)();
+	}
 	return -ENOTSUP;
 }
 
diff --git a/lib/librte_eventdev/rte_eventdev_pmd.h b/lib/librte_eventdev/rte_eventdev_pmd.h
index d118b9e5ba..27be376ed1 100644
--- a/lib/librte_eventdev/rte_eventdev_pmd.h
+++ b/lib/librte_eventdev/rte_eventdev_pmd.h
@@ -20,10 +20,13 @@ extern "C" {
 #include <string.h>
 
 #include <rte_common.h>
+#include <rte_compat.h>
 #include <rte_config.h>
 #include <rte_dev.h>
 #include <rte_log.h>
 #include <rte_malloc.h>
+#include <rte_mbuf.h>
+#include <rte_mbuf_dyn.h>
 
 #include "rte_eventdev.h"
 #include "rte_event_timer_adapter_pmd.h"
@@ -635,6 +638,23 @@ typedef int (*eventdev_eth_rx_adapter_stats_reset)
  */
 typedef int (*eventdev_selftest)(void);
 
+typedef uint32_t rte_event_pmd_selftest_seqn_t;
+extern int rte_event_pmd_selftest_seqn_dynfield_offset;
+
+/**
+ * Read test sequence number from mbuf.
+ *
+ * @param mbuf Structure to read from.
+ * @return pointer to test sequence number.
+ */
+__rte_internal
+static inline rte_event_pmd_selftest_seqn_t *
+rte_event_pmd_selftest_seqn(struct rte_mbuf *mbuf)
+{
+	return RTE_MBUF_DYNFIELD(mbuf,
+		rte_event_pmd_selftest_seqn_dynfield_offset,
+		rte_event_pmd_selftest_seqn_t *);
+}
 
 struct rte_cryptodev;
 
diff --git a/lib/librte_eventdev/version.map b/lib/librte_eventdev/version.map
index 8ae8420f9b..3e5c09cfdb 100644
--- a/lib/librte_eventdev/version.map
+++ b/lib/librte_eventdev/version.map
@@ -139,3 +139,9 @@ EXPERIMENTAL {
 	# added in 20.11
 	rte_event_pmd_pci_probe_named;
 };
+
+INTERNAL {
+	global:
+
+	rte_event_pmd_selftest_seqn_dynfield_offset;
+};
-- 
2.23.0

[dpdk-dev] [PATCH v2 8/9] app/eventdev: switch sequence number to dynamic mbuf field

[David Marchand]

The order test stored a sequence number in the deprecated mbuf field
seqn.
It is moved to a dynamic field in order to allow removal of seqn.

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 app/test-eventdev/test_order_common.c | 14 +++++++++++++-
 app/test-eventdev/test_order_common.h | 13 +++++++++++--
 2 files changed, 24 insertions(+), 3 deletions(-)

diff --git a/app/test-eventdev/test_order_common.c b/app/test-eventdev/test_order_common.c
index 01a44bcd75..04456d56db 100644
--- a/app/test-eventdev/test_order_common.c
+++ b/app/test-eventdev/test_order_common.c
@@ -48,7 +48,7 @@ order_producer(void *arg)
 
 		const flow_id_t flow = (uintptr_t)m % nb_flows;
 		/* Maintain seq number per flow */
-		m->seqn = producer_flow_seq[flow]++;
+		*order_mbuf_seqn(t, m) = producer_flow_seq[flow]++;
 		order_flow_id_save(t, flow, m, &ev);
 
 		while (rte_event_enqueue_burst(dev_id, port, &ev, 1) != 1) {
@@ -141,6 +141,11 @@ order_test_setup(struct evt_test *test, struct evt_options *opt)
 		.size = sizeof(flow_id_t),
 		.align = __alignof__(flow_id_t),
 	};
+	static const struct rte_mbuf_dynfield seqn_dynfield_desc = {
+		.name = "test_event_dynfield_seqn",
+		.size = sizeof(seqn_t),
+		.align = __alignof__(seqn_t),
+	};
 
 	test_order = rte_zmalloc_socket(test->name, sizeof(struct test_order),
 				RTE_CACHE_LINE_SIZE, opt->socket_id);
@@ -158,6 +163,13 @@ order_test_setup(struct evt_test *test, struct evt_options *opt)
 		return -rte_errno;
 	}
 
+	t->seqn_dynfield_offset =
+		rte_mbuf_dynfield_register(&seqn_dynfield_desc);
+	if (t->seqn_dynfield_offset < 0) {
+		evt_err("failed to register mbuf field");
+		return -rte_errno;
+	}
+
 	t->producer_flow_seq = rte_zmalloc_socket("test_producer_flow_seq",
 				 sizeof(*t->producer_flow_seq) * opt->nb_flows,
 				RTE_CACHE_LINE_SIZE, opt->socket_id);
diff --git a/app/test-eventdev/test_order_common.h b/app/test-eventdev/test_order_common.h
index 90eac96fc8..5ef8404938 100644
--- a/app/test-eventdev/test_order_common.h
+++ b/app/test-eventdev/test_order_common.h
@@ -22,6 +22,7 @@
 #define BURST_SIZE 16
 
 typedef uint32_t flow_id_t;
+typedef uint32_t seqn_t;
 
 struct test_order;
 
@@ -53,6 +54,7 @@ struct test_order {
 	uint64_t nb_pkts;
 	struct rte_mempool *pool;
 	int flow_id_dynfield_offset;
+	int seqn_dynfield_offset;
 	struct prod_data prod;
 	struct worker_data worker[EVT_MAX_PORTS];
 	uint32_t *producer_flow_seq;
@@ -77,6 +79,12 @@ order_flow_id_save(struct test_order *t, flow_id_t flow_id,
 	event->mbuf = mbuf;
 }
 
+static inline seqn_t *
+order_mbuf_seqn(struct test_order *t, struct rte_mbuf *mbuf)
+{
+	return RTE_MBUF_DYNFIELD(mbuf, t->seqn_dynfield_offset, seqn_t *);
+}
+
 static inline int
 order_nb_event_ports(struct evt_options *opt)
 {
@@ -91,9 +99,10 @@ order_process_stage_1(struct test_order *const t,
 {
 	const uint32_t flow = (uintptr_t)ev->mbuf % nb_flows;
 	/* compare the seqn against expected value */
-	if (ev->mbuf->seqn != expected_flow_seq[flow]) {
+	if (*order_mbuf_seqn(t, ev->mbuf) != expected_flow_seq[flow]) {
 		evt_err("flow=%x seqn mismatch got=%x expected=%x",
-			flow, ev->mbuf->seqn, expected_flow_seq[flow]);
+			flow, *order_mbuf_seqn(t, ev->mbuf),
+			expected_flow_seq[flow]);
 		t->err = true;
 		rte_smp_wmb();
 	}
-- 
2.23.0

[dpdk-dev] [PATCH v2 9/9] mbuf: remove seqn field

[David Marchand]

As announced in the deprecation note, the field seqn is removed to give
more space to the dynamic fields.

This is how the mbuf layout looks like (pahole-style):

word  type                              name                byte  size
 0    void *                            buf_addr;         /*   0 +  8 */
 1    rte_iova_t                        buf_iova          /*   8 +  8 */
      /* --- RTE_MARKER64               rearm_data;                   */
 2    uint16_t                          data_off;         /*  16 +  2 */
      uint16_t                          refcnt;           /*  18 +  2 */
      uint16_t                          nb_segs;          /*  20 +  2 */
      uint16_t                          port;             /*  22 +  2 */
 3    uint64_t                          ol_flags;         /*  24 +  8 */
      /* --- RTE_MARKER                 rx_descriptor_fields1;        */
 4    uint32_t             union        packet_type;      /*  32 +  4 */
      uint32_t                          pkt_len;          /*  36 +  4 */
 5    uint16_t                          data_len;         /*  40 +  2 */
      uint16_t                          vlan_tci;         /*  42 +  2 */
 5.5  uint64_t             union        hash;             /*  44 +  8 */
 6.5  uint16_t                          vlan_tci_outer;   /*  52 +  2 */
      uint16_t                          buf_len;          /*  54 +  2 */
 7    uint64_t                          timestamp;        /*  56 +  8 */
      /* --- RTE_MARKER                 cacheline1;                   */
 8    struct rte_mempool *              pool;             /*  64 +  8 */
 9    struct rte_mbuf *                 next;             /*  72 +  8 */
10    uint64_t             union        tx_offload;       /*  80 +  8 */
11    struct rte_mbuf_ext_shared_info * shinfo;           /*  88 +  8 */
12    uint16_t                          priv_size;        /*  96 +  2 */
      uint16_t                          timesync;         /*  98 +  2 */
12.5  uint32_t                          dynfield1[7];     /* 100 + 28 */
16    /* --- END                                             128      */

Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
---
 doc/guides/rel_notes/deprecation.rst   |  1 -
 doc/guides/rel_notes/release_20_11.rst |  3 +++
 lib/librte_mbuf/rte_mbuf_core.h        | 15 ++++++---------
 3 files changed, 9 insertions(+), 10 deletions(-)

diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
index 0f6f1df12a..fe3fd3956c 100644
--- a/doc/guides/rel_notes/deprecation.rst
+++ b/doc/guides/rel_notes/deprecation.rst
@@ -87,7 +87,6 @@ Deprecation Notices
   The following static fields will be moved as dynamic:
 
   - ``timestamp``
-  - ``seqn``
 
   As a consequence, the layout of the ``struct rte_mbuf`` will be re-arranged,
   avoiding impact on vectorized implementation of the driver datapaths,
diff --git a/doc/guides/rel_notes/release_20_11.rst b/doc/guides/rel_notes/release_20_11.rst
index 3d7edbfdbb..c0a9fc96aa 100644
--- a/doc/guides/rel_notes/release_20_11.rst
+++ b/doc/guides/rel_notes/release_20_11.rst
@@ -429,6 +429,9 @@ API Changes
 * mbuf: Removed the unioned fields ``userdata`` and ``udata64``
   from the structure ``rte_mbuf``. It is replaced with dynamic fields.
 
+* mbuf: Removed the field ``seqn`` from the structure ``rte_mbuf``.
+  It is replaced with dynamic fields.
+
 * pci: Removed the ``rte_kernel_driver`` enum defined in rte_dev.h and
   replaced with a private enum in the PCI subsystem.
 
diff --git a/lib/librte_mbuf/rte_mbuf_core.h b/lib/librte_mbuf/rte_mbuf_core.h
index a65eaaf692..3fb5abda3c 100644
--- a/lib/librte_mbuf/rte_mbuf_core.h
+++ b/lib/librte_mbuf/rte_mbuf_core.h
@@ -640,6 +640,11 @@ struct rte_mbuf {
 		};
 	};
 
+	/** Shared data for external buffer attached to mbuf. See
+	 * rte_pktmbuf_attach_extbuf().
+	 */
+	struct rte_mbuf_ext_shared_info *shinfo;
+
 	/** Size of the application private data. In case of an indirect
 	 * mbuf, it stores the direct mbuf private data size.
 	 */
@@ -648,15 +653,7 @@ struct rte_mbuf {
 	/** Timesync flags for use with IEEE1588. */
 	uint16_t timesync;
 
-	/** Sequence number. See also rte_reorder_insert(). */
-	uint32_t seqn;
-
-	/** Shared data for external buffer attached to mbuf. See
-	 * rte_pktmbuf_attach_extbuf().
-	 */
-	struct rte_mbuf_ext_shared_info *shinfo;
-
-	uint64_t dynfield1[3]; /**< Reserved for dynamic fields. */
+	uint32_t dynfield1[7]; /**< Reserved for dynamic fields. */
 } __rte_cache_aligned;
 
 /**
-- 
2.23.0

Re: [dpdk-dev] [PATCH v2 9/9] mbuf: remove seqn field

[Thomas Monjalon]

28/10/2020 13:20, David Marchand:
> As announced in the deprecation note, the field seqn is removed to give
> more space to the dynamic fields.
> 
> This is how the mbuf layout looks like (pahole-style):
> 
> word  type                              name                byte  size
>  0    void *                            buf_addr;         /*   0 +  8 */
>  1    rte_iova_t                        buf_iova          /*   8 +  8 */
>       /* --- RTE_MARKER64               rearm_data;                   */
>  2    uint16_t                          data_off;         /*  16 +  2 */
>       uint16_t                          refcnt;           /*  18 +  2 */
>       uint16_t                          nb_segs;          /*  20 +  2 */
>       uint16_t                          port;             /*  22 +  2 */
>  3    uint64_t                          ol_flags;         /*  24 +  8 */
>       /* --- RTE_MARKER                 rx_descriptor_fields1;        */
>  4    uint32_t             union        packet_type;      /*  32 +  4 */
>       uint32_t                          pkt_len;          /*  36 +  4 */
>  5    uint16_t                          data_len;         /*  40 +  2 */
>       uint16_t                          vlan_tci;         /*  42 +  2 */
>  5.5  uint64_t             union        hash;             /*  44 +  8 */
>  6.5  uint16_t                          vlan_tci_outer;   /*  52 +  2 */
>       uint16_t                          buf_len;          /*  54 +  2 */
>  7    uint64_t                          timestamp;        /*  56 +  8 */
>       /* --- RTE_MARKER                 cacheline1;                   */
>  8    struct rte_mempool *              pool;             /*  64 +  8 */
>  9    struct rte_mbuf *                 next;             /*  72 +  8 */
> 10    uint64_t             union        tx_offload;       /*  80 +  8 */
> 11    struct rte_mbuf_ext_shared_info * shinfo;           /*  88 +  8 */
> 12    uint16_t                          priv_size;        /*  96 +  2 */
>       uint16_t                          timesync;         /*  98 +  2 */
> 12.5  uint32_t                          dynfield1[7];     /* 100 + 28 */
> 16    /* --- END                                             128      */
> 
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>

Acked-by: Thomas Monjalon <thomas@monjalon.net>

Re: [dpdk-dev] [PATCH v2 0/9] remove mbuf seqn

[Thomas Monjalon]

28/10/2020 13:20, David Marchand:
> Followup on the work started by Thomas to free some space in the mbuf
> structure.
> This series applies on top of Thomas v4 series [1] and drops the seqn
> field that had been deprecated.

Thanks for the follow-up.

Applied, thanks.

Next step: removing timestamp and deciding what to get instead.