From: Sebastian Sanchez <sebastian.sanchez@intel.com>
Date: Fri, 26 May 2017 05:35:12 -0700
Subject: IB/hfi1: Don't remove RB entry when not needed.
Patch-mainline: v4.14-rc1
Git-commit: 7be85676f1d13c77a7e0c72e04903bfd39580d4f
References: bsc#1060463 FATE#323043

An RB tree is used for the SDMA pinning cache. Cache
entries are extracted and reinserted from the tree
in case the address range for it changes. However,
if the address range for the entry doesn't change,
deleting the entry from the RB tree is not necessary.

This affects performance since the tree needs to be
rebalanced for each insertion, and this happens in
the hot path. Optimize RB search by not removing
entries when it's not needed.

Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Reviewed-by: Mitko Haralanov <mitko.haralanov@intel.com>
Signed-off-by: Sebastian Sanchez <sebastian.sanchez@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Acked-by: Thomas Bogendoerfer <tbogendoerfer@suse.de>
---
 drivers/infiniband/hw/hfi1/mmu_rb.c    |   14 ++++++++++----
 drivers/infiniband/hw/hfi1/mmu_rb.h    |    5 +++--
 drivers/infiniband/hw/hfi1/user_sdma.c |   21 +++++++++++++++------
 3 files changed, 28 insertions(+), 12 deletions(-)

--- a/drivers/infiniband/hw/hfi1/mmu_rb.c
+++ b/drivers/infiniband/hw/hfi1/mmu_rb.c
@@ -214,21 +214,27 @@ static struct mmu_rb_node *__mmu_rb_sear
 	return node;
 }
 
-struct mmu_rb_node *hfi1_mmu_rb_extract(struct mmu_rb_handler *handler,
-					unsigned long addr, unsigned long len)
+bool hfi1_mmu_rb_remove_unless_exact(struct mmu_rb_handler *handler,
+				     unsigned long addr, unsigned long len,
+				     struct mmu_rb_node **rb_node)
 {
 	struct mmu_rb_node *node;
 	unsigned long flags;
+	bool ret = false;
 
 	spin_lock_irqsave(&handler->lock, flags);
 	node = __mmu_rb_search(handler, addr, len);
 	if (node) {
+		if (node->addr == addr && node->len == len)
+			goto unlock;
 		__mmu_int_rb_remove(node, &handler->root);
 		list_del(&node->list); /* remove from LRU list */
+		ret = true;
 	}
+unlock:
 	spin_unlock_irqrestore(&handler->lock, flags);
-
-	return node;
+	*rb_node = node;
+	return ret;
 }
 
 void hfi1_mmu_rb_evict(struct mmu_rb_handler *handler, void *evict_arg)
--- a/drivers/infiniband/hw/hfi1/mmu_rb.h
+++ b/drivers/infiniband/hw/hfi1/mmu_rb.h
@@ -81,7 +81,8 @@ int hfi1_mmu_rb_insert(struct mmu_rb_han
 void hfi1_mmu_rb_evict(struct mmu_rb_handler *handler, void *evict_arg);
 void hfi1_mmu_rb_remove(struct mmu_rb_handler *handler,
 			struct mmu_rb_node *mnode);
-struct mmu_rb_node *hfi1_mmu_rb_extract(struct mmu_rb_handler *handler,
-					unsigned long addr, unsigned long len);
+bool hfi1_mmu_rb_remove_unless_exact(struct mmu_rb_handler *handler,
+				     unsigned long addr, unsigned long len,
+				     struct mmu_rb_node **rb_node);
 
 #endif /* _HFI1_MMU_RB_H */
--- a/drivers/infiniband/hw/hfi1/user_sdma.c
+++ b/drivers/infiniband/hw/hfi1/user_sdma.c
@@ -1165,14 +1165,23 @@ static int pin_vector_pages(struct user_
 	struct hfi1_user_sdma_pkt_q *pq = req->pq;
 	struct sdma_mmu_node *node = NULL;
 	struct mmu_rb_node *rb_node;
+	bool extracted;
 
-	rb_node = hfi1_mmu_rb_extract(pq->handler,
-				      (unsigned long)iovec->iov.iov_base,
-				      iovec->iov.iov_len);
-	if (rb_node)
+	extracted =
+		hfi1_mmu_rb_remove_unless_exact(pq->handler,
+						(unsigned long)
+						iovec->iov.iov_base,
+						iovec->iov.iov_len, &rb_node);
+	if (rb_node) {
 		node = container_of(rb_node, struct sdma_mmu_node, rb);
-	else
-		rb_node = NULL;
+		if (!extracted) {
+			atomic_inc(&node->refcount);
+			iovec->pages = node->pages;
+			iovec->npages = node->npages;
+			iovec->node = node;
+			return 0;
+		}
+	}
 
 	if (!node) {
 		node = kzalloc(sizeof(*node), GFP_KERNEL);