From mboxrd@z Thu Jan 1 00:00:00 1970 From: Al Viro Subject: [PATCH] fib_trie: no need to delay vfree() Date: Mon, 6 May 2013 03:03:46 +0100 Message-ID: <20130506020346.GK25399@ZenIV.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: netdev@vger.kernel.org To: David Miller Return-path: Received: from zeniv.linux.org.uk ([195.92.253.2]:49384 "EHLO ZenIV.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751982Ab3EFCDt (ORCPT ); Sun, 5 May 2013 22:03:49 -0400 Content-Disposition: inline Sender: netdev-owner@vger.kernel.org List-ID: Now that vfree() can be called from interrupt contexts, there's no need to play games with schedule_work() to escape calling vfree() from RCU callbacks. Signed-off-by: Al Viro --- Not sure if it's 3.10 or 3.11 fodder - the patch is trivial and vfree() from RCU callbacks will be exercised in 3.10 anyway (both for large descriptor tables and for large sysv semaphore arrays). OTOH, it doesn't fix any bugs, so... Up to you. diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c index ff06b75..49616fe 100644 --- a/net/ipv4/fib_trie.c +++ b/net/ipv4/fib_trie.c @@ -125,7 +125,6 @@ struct tnode { unsigned int empty_children; /* KEYLENGTH bits needed */ union { struct rcu_head rcu; - struct work_struct work; struct tnode *tnode_free; }; struct rt_trie_node __rcu *child[0]; @@ -383,12 +382,6 @@ static struct tnode *tnode_alloc(size_t size) return vzalloc(size); } -static void __tnode_vfree(struct work_struct *arg) -{ - struct tnode *tn = container_of(arg, struct tnode, work); - vfree(tn); -} - static void __tnode_free_rcu(struct rcu_head *head) { struct tnode *tn = container_of(head, struct tnode, rcu); @@ -397,10 +390,8 @@ static void __tnode_free_rcu(struct rcu_head *head) if (size <= PAGE_SIZE) kfree(tn); - else { - INIT_WORK(&tn->work, __tnode_vfree); - schedule_work(&tn->work); - } + else + vfree(tn); } static inline void tnode_free(struct tnode *tn)