Patchwork [dpdk-dev,19/30] net/sfc: remove RxQ control from shared RxQ info

login
register
mail settings
Submitter Andrew Rybchenko
Date Feb. 7, 2019, 12:17 p.m.
Message ID <1549541873-17403-20-git-send-email-arybchenko@solarflare.com>
Download mbox | patch
Permalink /patch/720439/
State New
Headers show

Comments

Andrew Rybchenko - Feb. 7, 2019, 12:17 p.m.
libefx is not multi-process aware and all related structures
should be moved to primary process only.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
---
 drivers/net/sfc/sfc.h        |  3 ++
 drivers/net/sfc/sfc_ethdev.c |  2 +-
 drivers/net/sfc/sfc_flow.c   |  8 ++--
 drivers/net/sfc/sfc_rx.c     | 77 ++++++++++++++++++++++--------------
 drivers/net/sfc/sfc_rx.h     |  4 +-
 5 files changed, 57 insertions(+), 37 deletions(-)

Patch

diff --git a/drivers/net/sfc/sfc.h b/drivers/net/sfc/sfc.h
index f04127bbc..f19a9dafb 100644
--- a/drivers/net/sfc/sfc.h
+++ b/drivers/net/sfc/sfc.h
@@ -110,6 +110,8 @@  struct sfc_intr {
 	boolean_t			lsc_intr;
 };
 
+struct sfc_rxq;
+
 struct sfc_rxq_info;
 struct sfc_txq_info;
 struct sfc_dp_rx;
@@ -256,6 +258,7 @@  struct sfc_adapter {
 
 	unsigned int			rxq_count;
 	struct sfc_rxq_info		*rxq_info;
+	struct sfc_rxq			*rxq_ctrl;
 
 	unsigned int			txq_count;
 	struct sfc_txq_info		*txq_info;
diff --git a/drivers/net/sfc/sfc_ethdev.c b/drivers/net/sfc/sfc_ethdev.c
index a00ea9ce4..3d38bf661 100644
--- a/drivers/net/sfc/sfc_ethdev.c
+++ b/drivers/net/sfc/sfc_ethdev.c
@@ -1196,7 +1196,7 @@  sfc_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id)
 	if (sa->state != SFC_ADAPTER_STARTED)
 		goto fail_not_started;
 
-	if (sa->rxq_info[rx_queue_id].rxq == NULL)
+	if (sa->rxq_info[rx_queue_id].state != SFC_RXQ_INITIALIZED)
 		goto fail_not_setup;
 
 	rc = sfc_rx_qstart(sa, rx_queue_id);
diff --git a/drivers/net/sfc/sfc_flow.c b/drivers/net/sfc/sfc_flow.c
index 41d388735..e20c2e612 100644
--- a/drivers/net/sfc/sfc_flow.c
+++ b/drivers/net/sfc/sfc_flow.c
@@ -1244,7 +1244,7 @@  sfc_flow_parse_queue(struct sfc_adapter *sa,
 	if (queue->index >= sa->rxq_count)
 		return -EINVAL;
 
-	rxq = sa->rxq_info[queue->index].rxq;
+	rxq = &sa->rxq_ctrl[queue->index];
 	flow->spec.template.efs_dmaq_id = (uint16_t)rxq->hw_index;
 
 	return 0;
@@ -1269,7 +1269,7 @@  sfc_flow_parse_rss(struct sfc_adapter *sa,
 		return -EINVAL;
 
 	rxq_sw_index = sa->rxq_count - 1;
-	rxq = sa->rxq_info[rxq_sw_index].rxq;
+	rxq = &sa->rxq_ctrl[rxq_sw_index];
 	rxq_hw_index_min = rxq->hw_index;
 	rxq_hw_index_max = 0;
 
@@ -1279,7 +1279,7 @@  sfc_flow_parse_rss(struct sfc_adapter *sa,
 		if (rxq_sw_index >= sa->rxq_count)
 			return -EINVAL;
 
-		rxq = sa->rxq_info[rxq_sw_index].rxq;
+		rxq = &sa->rxq_ctrl[rxq_sw_index];
 
 		if (rxq->hw_index < rxq_hw_index_min)
 			rxq_hw_index_min = rxq->hw_index;
@@ -1344,7 +1344,7 @@  sfc_flow_parse_rss(struct sfc_adapter *sa,
 	for (i = 0; i < RTE_DIM(sfc_rss_conf->rss_tbl); ++i) {
 		unsigned int nb_queues = action_rss->queue_num;
 		unsigned int rxq_sw_index = action_rss->queue[i % nb_queues];
-		struct sfc_rxq *rxq = sa->rxq_info[rxq_sw_index].rxq;
+		struct sfc_rxq *rxq = &sa->rxq_ctrl[rxq_sw_index];
 
 		sfc_rss_conf->rss_tbl[i] = rxq->hw_index - rxq_hw_index_min;
 	}
diff --git a/drivers/net/sfc/sfc_rx.c b/drivers/net/sfc/sfc_rx.c
index 3c1e77974..eb4875fec 100644
--- a/drivers/net/sfc/sfc_rx.c
+++ b/drivers/net/sfc/sfc_rx.c
@@ -395,12 +395,17 @@  sfc_rxq_info_by_dp_rxq(const struct sfc_dp_rxq *dp_rxq)
 struct sfc_rxq *
 sfc_rxq_by_dp_rxq(const struct sfc_dp_rxq *dp_rxq)
 {
-	struct sfc_rxq_info *rxq_info;
+	const struct sfc_dp_queue *dpq = &dp_rxq->dpq;
+	struct rte_eth_dev *eth_dev;
+	struct sfc_adapter *sa;
 
-	rxq_info = sfc_rxq_info_by_dp_rxq(dp_rxq);
+	SFC_ASSERT(rte_eth_dev_is_valid_port(dpq->port_id));
+	eth_dev = &rte_eth_devices[dpq->port_id];
+
+	sa = eth_dev->data->dev_private;
 
-	SFC_ASSERT(rxq_info->rxq != NULL);
-	return rxq_info->rxq;
+	SFC_ASSERT(dpq->queue_id < sa->rxq_count);
+	return &sa->rxq_ctrl[dpq->queue_id];
 }
 
 static sfc_dp_rx_qsize_up_rings_t sfc_efx_rx_qsize_up_rings;
@@ -563,9 +568,10 @@  sfc_rx_qflush(struct sfc_adapter *sa, unsigned int sw_index)
 	int rc;
 
 	rxq_info = &sa->rxq_info[sw_index];
-	rxq = rxq_info->rxq;
 	SFC_ASSERT(rxq_info->state & SFC_RXQ_STARTED);
 
+	rxq = &sa->rxq_ctrl[sw_index];
+
 	/*
 	 * Retry Rx queue flushing in the case of flush failed or
 	 * timeout. In the worst case it can delay for 6 seconds.
@@ -674,10 +680,9 @@  sfc_rx_qstart(struct sfc_adapter *sa, unsigned int sw_index)
 	SFC_ASSERT(sw_index < sa->rxq_count);
 
 	rxq_info = &sa->rxq_info[sw_index];
-	rxq = rxq_info->rxq;
-	SFC_ASSERT(rxq != NULL);
 	SFC_ASSERT(rxq_info->state == SFC_RXQ_INITIALIZED);
 
+	rxq = &sa->rxq_ctrl[sw_index];
 	evq = rxq->evq;
 
 	rc = sfc_ev_qstart(evq, sfc_evq_index_by_rxq_sw_index(sa, sw_index));
@@ -764,9 +769,8 @@  sfc_rx_qstop(struct sfc_adapter *sa, unsigned int sw_index)
 	SFC_ASSERT(sw_index < sa->rxq_count);
 
 	rxq_info = &sa->rxq_info[sw_index];
-	rxq = rxq_info->rxq;
 
-	if (rxq == NULL || rxq_info->state == SFC_RXQ_INITIALIZED)
+	if (rxq_info->state == SFC_RXQ_INITIALIZED)
 		return;
 	SFC_ASSERT(rxq_info->state & SFC_RXQ_STARTED);
 
@@ -774,6 +778,7 @@  sfc_rx_qstop(struct sfc_adapter *sa, unsigned int sw_index)
 	sa->eth_dev->data->rx_queue_state[sw_index] =
 		RTE_ETH_QUEUE_STATE_STOPPED;
 
+	rxq = &sa->rxq_ctrl[sw_index];
 	sa->priv.dp_rx->qstop(rxq_info->dp, &rxq->evq->read_ptr);
 
 	if (sw_index == 0)
@@ -1026,14 +1031,7 @@  sfc_rx_qinit(struct sfc_adapter *sa, unsigned int sw_index,
 	if (rc != 0)
 		goto fail_ev_qinit;
 
-	rc = ENOMEM;
-	rxq = rte_zmalloc_socket("sfc-rxq", sizeof(*rxq), RTE_CACHE_LINE_SIZE,
-				 socket_id);
-	if (rxq == NULL)
-		goto fail_rxq_alloc;
-
-	rxq_info->rxq = rxq;
-
+	rxq = &sa->rxq_ctrl[sw_index];
 	rxq->evq = evq;
 	rxq->hw_index = sw_index;
 	rxq_info->refill_threshold =
@@ -1083,10 +1081,6 @@  sfc_rx_qinit(struct sfc_adapter *sa, unsigned int sw_index,
 	sfc_dma_free(sa, &rxq->mem);
 
 fail_dma_alloc:
-	rxq_info->rxq = NULL;
-	rte_free(rxq);
-
-fail_rxq_alloc:
 	sfc_ev_qfini(evq);
 
 fail_ev_qinit:
@@ -1109,21 +1103,20 @@  sfc_rx_qfini(struct sfc_adapter *sa, unsigned int sw_index)
 
 	rxq_info = &sa->rxq_info[sw_index];
 
-	rxq = rxq_info->rxq;
 	SFC_ASSERT(rxq_info->state == SFC_RXQ_INITIALIZED);
 
 	sa->priv.dp_rx->qdestroy(rxq_info->dp);
 	rxq_info->dp = NULL;
 
-	rxq_info->rxq = NULL;
+	rxq_info->state &= ~SFC_RXQ_INITIALIZED;
 	rxq_info->entries = 0;
 
+	rxq = &sa->rxq_ctrl[sw_index];
+
 	sfc_dma_free(sa, &rxq->mem);
 
 	sfc_ev_qfini(rxq->evq);
 	rxq->evq = NULL;
-
-	rte_free(rxq);
 }
 
 /*
@@ -1366,7 +1359,7 @@  sfc_rx_start(struct sfc_adapter *sa)
 		goto fail_rss_config;
 
 	for (sw_index = 0; sw_index < sa->rxq_count; ++sw_index) {
-		if (sa->rxq_info[sw_index].rxq != NULL &&
+		if (sa->rxq_info[sw_index].state == SFC_RXQ_INITIALIZED &&
 		    (!sa->rxq_info[sw_index].deferred_start ||
 		     sa->rxq_info[sw_index].deferred_started)) {
 			rc = sfc_rx_qstart(sa, sw_index);
@@ -1398,7 +1391,7 @@  sfc_rx_stop(struct sfc_adapter *sa)
 
 	sw_index = sa->rxq_count;
 	while (sw_index-- > 0) {
-		if (sa->rxq_info[sw_index].rxq != NULL)
+		if (sa->rxq_info[sw_index].state & SFC_RXQ_STARTED)
 			sfc_rx_qstop(sa, sw_index);
 	}
 
@@ -1476,7 +1469,7 @@  sfc_rx_fini_queues(struct sfc_adapter *sa, unsigned int nb_rx_queues)
 
 	sw_index = sa->rxq_count;
 	while (--sw_index >= (int)nb_rx_queues) {
-		if (sa->rxq_info[sw_index].rxq != NULL)
+		if (sa->rxq_info[sw_index].state & SFC_RXQ_INITIALIZED)
 			sfc_rx_qfini(sa, sw_index);
 	}
 
@@ -1516,8 +1509,18 @@  sfc_rx_configure(struct sfc_adapter *sa)
 						 sa->socket_id);
 		if (sa->rxq_info == NULL)
 			goto fail_rxqs_alloc;
+
+		/*
+		 * Allocate primary process only RxQ control from heap
+		 * since it should not be shared.
+		 */
+		rc = ENOMEM;
+		sa->rxq_ctrl = calloc(nb_rx_queues, sizeof(sa->rxq_ctrl[0]));
+		if (sa->rxq_ctrl == NULL)
+			goto fail_rxqs_ctrl_alloc;
 	} else {
 		struct sfc_rxq_info *new_rxq_info;
+		struct sfc_rxq *new_rxq_ctrl;
 
 		if (nb_rx_queues < sa->rxq_count)
 			sfc_rx_fini_queues(sa, nb_rx_queues);
@@ -1529,11 +1532,22 @@  sfc_rx_configure(struct sfc_adapter *sa)
 		if (new_rxq_info == NULL && nb_rx_queues > 0)
 			goto fail_rxqs_realloc;
 
+		rc = ENOMEM;
+		new_rxq_ctrl = realloc(sa->rxq_ctrl,
+				       nb_rx_queues * sizeof(sa->rxq_ctrl[0]));
+		if (new_rxq_ctrl == NULL && nb_rx_queues > 0)
+			goto fail_rxqs_ctrl_realloc;
+
 		sa->rxq_info = new_rxq_info;
-		if (nb_rx_queues > sa->rxq_count)
+		sa->rxq_ctrl = new_rxq_ctrl;
+		if (nb_rx_queues > sa->rxq_count) {
 			memset(&sa->rxq_info[sa->rxq_count], 0,
 			       (nb_rx_queues - sa->rxq_count) *
 			       sizeof(sa->rxq_info[0]));
+			memset(&sa->rxq_ctrl[sa->rxq_count], 0,
+			       (nb_rx_queues - sa->rxq_count) *
+			       sizeof(sa->rxq_ctrl[0]));
+		}
 	}
 
 	while (sa->rxq_count < nb_rx_queues) {
@@ -1565,7 +1579,9 @@  sfc_rx_configure(struct sfc_adapter *sa)
 
 fail_rx_process_adv_conf_rss:
 fail_rx_qinit_info:
+fail_rxqs_ctrl_realloc:
 fail_rxqs_realloc:
+fail_rxqs_ctrl_alloc:
 fail_rxqs_alloc:
 	sfc_rx_close(sa);
 
@@ -1588,6 +1604,9 @@  sfc_rx_close(struct sfc_adapter *sa)
 
 	rss->channels = 0;
 
+	free(sa->rxq_ctrl);
+	sa->rxq_ctrl = NULL;
+
 	rte_free(sa->rxq_info);
 	sa->rxq_info = NULL;
 }
diff --git a/drivers/net/sfc/sfc_rx.h b/drivers/net/sfc/sfc_rx.h
index 963d487c0..73b70249c 100644
--- a/drivers/net/sfc/sfc_rx.h
+++ b/drivers/net/sfc/sfc_rx.h
@@ -50,8 +50,7 @@  enum sfc_rxq_state_bit {
 };
 
 /**
- * Receive queue control information.
- * Allocated on the socket specified on the queue setup.
+ * Receive queue control primary process-only information.
  */
 struct sfc_rxq {
 	struct sfc_evq		*evq;
@@ -110,7 +109,6 @@  struct sfc_rxq_info {
 	unsigned int		entries;
 	efx_rxq_type_t		type;
 	unsigned int		type_flags;
-	struct sfc_rxq		*rxq;
 	struct sfc_dp_rxq	*dp;
 	boolean_t		deferred_start;
 	boolean_t		deferred_started;