ibmvnic: Increase max subcrq indirect entries with fallback

POWER8 support a maximum of 16 subcrq indirect descriptor entries per
 H_SEND_SUB_CRQ_INDIRECT call, while POWER9 and newer hypervisors
 support up to 128 entries. Increasing the max number of indirect
descriptor entries improves batching efficiency and reduces
hcall overhead, which enhances throughput under large workload on POWER9+.

Currently, ibmvnic driver always uses a fixed number of max indirect
descriptor entries (16). send_subcrq_indirect() treats all hypervisor
errors the same:
 - Cleanup and Drop the entire batch of descriptors.
 - Return an error to the caller.
 - Rely on TCP/IP retransmissions to recover.
 - If the hypervisor returns H_PARAMETER (e.g., because 128
   entries are not supported on POWER8), the driver will continue
   to drop batches, resulting in unnecessary packet loss.

In this patch:
Raise the default maximum indirect entries to 128 to improve ibmvnic
batching on morden platform. But also gracefully fall back to
16 entries for Power 8 systems.

Since there is no VIO interface to query the hypervisor’s supported
limit, vnic handles send_subcrq_indirect() H_PARAMETER errors:
 - On first H_PARAMETER failure, log the failure context
 - Reduce max_indirect_entries to 16 and allow the single batch to drop.
 - Subsequent calls automatically use the correct lower limit,
    avoiding repeated drops.

The goal is to  optimizes performance on modern systems while handles
falling back for older POWER8 hypervisors.

Performance shows 40% improvements with MTU (1500) on largework load.

Signed-off-by: Mingming Cao <mmc@linux.ibm.com>
Reviewed-by: Brian King <bjking1@linux.ibm.com>
Reviewed-by: Haren Myneni <haren@linux.ibm.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20250821130215.97960-1-mmc@linux.ibm.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

1 files changed, 52 insertions, 7 deletions

diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c
index eec971567aac..3808148c1fc7 100644
--- a/drivers/net/ethernet/ibm/ibmvnic.c
+++ b/drivers/net/ethernet/ibm/ibmvnic.c
@@ -756,6 +756,17 @@ static void deactivate_rx_pools(struct ibmvnic_adapter *adapter)
 		adapter->rx_pool[i].active = 0;
 }
 
+static void ibmvnic_set_safe_max_ind_descs(struct ibmvnic_adapter *adapter)
+{
+	if (adapter->cur_max_ind_descs > IBMVNIC_SAFE_IND_DESC) {
+		netdev_info(adapter->netdev,
+			    "set max ind descs from %u to safe limit %u\n",
+			    adapter->cur_max_ind_descs,
+			    IBMVNIC_SAFE_IND_DESC);
+		adapter->cur_max_ind_descs = IBMVNIC_SAFE_IND_DESC;
+	}
+}
+
 static void replenish_rx_pool(struct ibmvnic_adapter *adapter,
 			      struct ibmvnic_rx_pool *pool)
 {
@@ -843,7 +854,7 @@ static void replenish_rx_pool(struct ibmvnic_adapter *adapter,
 		sub_crq->rx_add.len = cpu_to_be32(pool->buff_size << shift);
 
 		/* if send_subcrq_indirect queue is full, flush to VIOS */
-		if (ind_bufp->index == IBMVNIC_MAX_IND_DESCS ||
+		if (ind_bufp->index == adapter->cur_max_ind_descs ||
 		    i == count - 1) {
 			lpar_rc =
 				send_subcrq_indirect(adapter, handle,
@@ -862,6 +873,14 @@ static void replenish_rx_pool(struct ibmvnic_adapter *adapter,
 failure:
 	if (lpar_rc != H_PARAMETER && lpar_rc != H_CLOSED)
 		dev_err_ratelimited(dev, "rx: replenish packet buffer failed\n");
+
+	/* Detect platform limit H_PARAMETER */
+	if (lpar_rc == H_PARAMETER)
+		ibmvnic_set_safe_max_ind_descs(adapter);
+
+	/* For all error case, temporarily drop only this batch
+	 * Rely on TCP/IP retransmissions to retry and recover
+	 */
 	for (i = ind_bufp->index - 1; i >= 0; --i) {
 		struct ibmvnic_rx_buff *rx_buff;
 
@@ -2381,16 +2400,28 @@ static int ibmvnic_tx_scrq_flush(struct ibmvnic_adapter *adapter,
 		rc = send_subcrq_direct(adapter, handle,
 					(u64 *)ind_bufp->indir_arr);
 
-	if (rc)
+	if (rc) {
+		dev_err_ratelimited(&adapter->vdev->dev,
+				    "tx_flush failed, rc=%u (%llu entries dma=%pad handle=%llx)\n",
+				    rc, entries, &dma_addr, handle);
+		/* Detect platform limit H_PARAMETER */
+		if (rc == H_PARAMETER)
+			ibmvnic_set_safe_max_ind_descs(adapter);
+
+		/* For all error case, temporarily drop only this batch
+		 * Rely on TCP/IP retransmissions to retry and recover
+		 */
 		ibmvnic_tx_scrq_clean_buffer(adapter, tx_scrq);
-	else
+	} else {
 		ind_bufp->index = 0;
+	}
 	return rc;
 }
 
 static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 {
 	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+	u32 cur_max_ind_descs = adapter->cur_max_ind_descs;
 	int queue_num = skb_get_queue_mapping(skb);
 	u8 *hdrs = (u8 *)&adapter->tx_rx_desc_req;
 	struct device *dev = &adapter->vdev->dev;
@@ -2590,7 +2621,7 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 	tx_crq.v1.n_crq_elem = num_entries;
 	tx_buff->num_entries = num_entries;
 	/* flush buffer if current entry can not fit */
-	if (num_entries + ind_bufp->index > IBMVNIC_MAX_IND_DESCS) {
+	if (num_entries + ind_bufp->index > cur_max_ind_descs) {
 		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq, true);
 		if (lpar_rc != H_SUCCESS)
 			goto tx_flush_err;
@@ -2603,7 +2634,7 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 	ind_bufp->index += num_entries;
 	if (__netdev_tx_sent_queue(txq, skb->len,
 				   netdev_xmit_more() &&
-				   ind_bufp->index < IBMVNIC_MAX_IND_DESCS)) {
+				   ind_bufp->index < cur_max_ind_descs)) {
 		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq, true);
 		if (lpar_rc != H_SUCCESS)
 			goto tx_err;
@@ -4006,7 +4037,7 @@ static void release_sub_crq_queue(struct ibmvnic_adapter *adapter,
 	}
 
 	dma_free_coherent(dev,
-			  IBMVNIC_IND_ARR_SZ,
+			  IBMVNIC_IND_MAX_ARR_SZ,
 			  scrq->ind_buf.indir_arr,
 			  scrq->ind_buf.indir_dma);
 
@@ -4063,7 +4094,7 @@ static struct ibmvnic_sub_crq_queue *init_sub_crq_queue(struct ibmvnic_adapter
 
 	scrq->ind_buf.indir_arr =
 		dma_alloc_coherent(dev,
-				   IBMVNIC_IND_ARR_SZ,
+				   IBMVNIC_IND_MAX_ARR_SZ,
 				   &scrq->ind_buf.indir_dma,
 				   GFP_KERNEL);
 
@@ -6369,6 +6400,19 @@ static int ibmvnic_reset_init(struct ibmvnic_adapter *adapter, bool reset)
 			rc = reset_sub_crq_queues(adapter);
 		}
 	} else {
+		if (adapter->reset_reason == VNIC_RESET_MOBILITY) {
+			/* After an LPM, reset the max number of indirect
+			 * subcrq descriptors per H_SEND_SUB_CRQ_INDIRECT
+			 * hcall to the default max (e.g POWER8 -> POWER10)
+			 *
+			 * If the new destination platform does not support
+			 * the higher limit max (e.g. POWER10-> POWER8 LPM)
+			 * H_PARAMETER will trigger automatic fallback to the
+			 * safe minimum limit.
+			 */
+			adapter->cur_max_ind_descs = IBMVNIC_MAX_IND_DESCS;
+		}
+
 		rc = init_sub_crqs(adapter);
 	}
 
@@ -6520,6 +6564,7 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id)
 
 	adapter->wait_for_reset = false;
 	adapter->last_reset_time = jiffies;
+	adapter->cur_max_ind_descs = IBMVNIC_MAX_IND_DESCS;
 
 	rc = register_netdev(netdev);
 	if (rc) {