diff mbox series

rhashtable: Drop raw RCU deref in nested_table_free

Message ID 20200603081243.GA30134@gondor.apana.org.au
State Accepted
Delegated to: David Miller
Headers show
Series rhashtable: Drop raw RCU deref in nested_table_free | expand

Commit Message

Herbert Xu June 3, 2020, 8:12 a.m. UTC 
This patch replaces some unnecessary uses of rcu_dereference_raw
in the rhashtable code with rcu_dereference_protected.

The top-level nested table entry is only marked as RCU because it
shares the same type as the tree entries underneath it.  So it
doesn't need any RCU protection.

We also don't need RCU protection when we're freeing a nested RCU
table because by this stage we've long passed a memory barrier
when anyone could change the nested table.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Comments

David Miller June 6, 2020, 10:51 p.m. UTC | #1 
From: Herbert Xu <herbert@gondor.apana.org.au>
Date: Wed, 3 Jun 2020 18:12:43 +1000

> This patch replaces some unnecessary uses of rcu_dereference_raw
> in the rhashtable code with rcu_dereference_protected.
> 
> The top-level nested table entry is only marked as RCU because it
> shares the same type as the tree entries underneath it.  So it
> doesn't need any RCU protection.
> 
> We also don't need RCU protection when we're freeing a nested RCU
> table because by this stage we've long passed a memory barrier
> when anyone could change the nested table.
> 
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Applied, thanks Herbert.
diff mbox series

Patch

diff --git a/lib/rhashtable.c b/lib/rhashtable.c
index bdb7e4cadf05..9f6890aedd1a 100644
--- a/lib/rhashtable.c
+++ b/lib/rhashtable.c
@@ -63,13 +63,22 @@  EXPORT_SYMBOL_GPL(lockdep_rht_bucket_is_held);
 #define ASSERT_RHT_MUTEX(HT)
 #endif
 
+static inline union nested_table *nested_table_top(
+	const struct bucket_table *tbl)
+{
+	/* The top-level bucket entry does not need RCU protection
+	 * because it's set at the same time as tbl->nest.
+	 */
+	return (void *)rcu_dereference_protected(tbl->buckets[0], 1);
+}
+
 static void nested_table_free(union nested_table *ntbl, unsigned int size)
 {
 	const unsigned int shift = PAGE_SHIFT - ilog2(sizeof(void *));
 	const unsigned int len = 1 << shift;
 	unsigned int i;
 
-	ntbl = rcu_dereference_raw(ntbl->table);
+	ntbl = rcu_dereference_protected(ntbl->table, 1);
 	if (!ntbl)
 		return;
 
@@ -89,7 +98,7 @@  static void nested_bucket_table_free(const struct bucket_table *tbl)
 	union nested_table *ntbl;
 	unsigned int i;
 
-	ntbl = (union nested_table *)rcu_dereference_raw(tbl->buckets[0]);
+	ntbl = nested_table_top(tbl);
 
 	for (i = 0; i < len; i++)
 		nested_table_free(ntbl + i, size);
@@ -1173,7 +1182,7 @@  struct rhash_lock_head **__rht_bucket_nested(const struct bucket_table *tbl,
 	unsigned int subhash = hash;
 	union nested_table *ntbl;
 
-	ntbl = (union nested_table *)rcu_dereference_raw(tbl->buckets[0]);
+	ntbl = nested_table_top(tbl);
 	ntbl = rht_dereference_bucket_rcu(ntbl[index].table, tbl, hash);
 	subhash >>= tbl->nest;
 
@@ -1213,7 +1222,7 @@  struct rhash_lock_head **rht_bucket_nested_insert(struct rhashtable *ht,
 	unsigned int size = tbl->size >> tbl->nest;
 	union nested_table *ntbl;
 
-	ntbl = (union nested_table *)rcu_dereference_raw(tbl->buckets[0]);
+	ntbl = nested_table_top(tbl);
 	hash >>= tbl->nest;
 	ntbl = nested_table_alloc(ht, &ntbl[index].table,
 				  size <= (1 << shift));