From patchwork Mon Jun 17 15:13:56 2013 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jeff Layton X-Patchwork-Id: 2733881 Return-Path: X-Original-To: patchwork-cifs-client@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork2.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.19.201]) by patchwork2.web.kernel.org (Postfix) with ESMTP id B468EC0AB1 for ; Mon, 17 Jun 2013 15:15:33 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 42B4E2036F for ; Mon, 17 Jun 2013 15:15:28 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 5DBE520318 for ; Mon, 17 Jun 2013 15:15:23 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752538Ab3FQPOr (ORCPT ); Mon, 17 Jun 2013 11:14:47 -0400 Received: from mx1.redhat.com ([209.132.183.28]:16719 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752533Ab3FQPOp (ORCPT ); Mon, 17 Jun 2013 11:14:45 -0400 Received: from int-mx10.intmail.prod.int.phx2.redhat.com (int-mx10.intmail.prod.int.phx2.redhat.com [10.5.11.23]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id r5HFEDi7009519 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 17 Jun 2013 11:14:13 -0400 Received: from sikun.lab.eng.rdu2.redhat.com (sikun.lab.eng.rdu2.redhat.com [10.8.0.43]) by int-mx10.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id r5HFDuWe009391; Mon, 17 Jun 2013 11:14:12 -0400 From: Jeff Layton To: viro@zeniv.linux.org.uk, matthew@wil.cx, bfields@fieldses.org Cc: dhowells@redhat.com, sage@inktank.com, smfrench@gmail.com, swhiteho@redhat.com, Trond.Myklebust@netapp.com, akpm@linux-foundation.org, linux-kernel@vger.kernel.org, linux-afs@lists.infradead.org, ceph-devel@vger.kernel.org, linux-cifs@vger.kernel.org, samba-technical@lists.samba.org, cluster-devel@redhat.com, linux-nfs@vger.kernel.org, linux-fsdevel@vger.kernel.org, piastryyy@gmail.com Subject: [PATCH v3 13/13] locks: move file_lock_list to a set of percpu hlist_heads and convert file_lock_lock to an lglock Date: Mon, 17 Jun 2013 11:13:56 -0400 Message-Id: <1371482036-15958-14-git-send-email-jlayton@redhat.com> In-Reply-To: <1371482036-15958-1-git-send-email-jlayton@redhat.com> References: <1371482036-15958-1-git-send-email-jlayton@redhat.com> X-Scanned-By: MIMEDefang 2.68 on 10.5.11.23 Sender: linux-cifs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-cifs@vger.kernel.org X-Spam-Status: No, score=-8.0 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP The file_lock_list is only used for /proc/locks. The vastly common case is for locks to be put onto the list and come off again, without ever being traversed. Help optimize for this use-case by moving to percpu hlist_head-s. At the same time, we can make the locking less contentious by moving to an lglock. When iterating over the lists for /proc/locks, we must take the global lock and then iterate over each CPU's list in turn. This change necessitates a new fl_link_cpu field to keep track of which CPU the entry is on. On x86_64 at least, this field is placed within an existing hole in the struct to avoid growing the size. Signed-off-by: Jeff Layton Acked-by: J. Bruce Fields --- fs/locks.c | 68 ++++++++++++++++++++++++++++++++++++--------------- include/linux/fs.h | 1 + 2 files changed, 49 insertions(+), 20 deletions(-) diff --git a/fs/locks.c b/fs/locks.c index 5db80c7..be9ca57 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -127,6 +127,8 @@ #include #include #include +#include +#include #include @@ -155,11 +157,12 @@ int lease_break_time = 45; for (lockp = &inode->i_flock; *lockp != NULL; lockp = &(*lockp)->fl_next) /* - * The global file_lock_list is only used for displaying /proc/locks. Protected - * by the file_lock_lock. + * The global file_lock_list is only used for displaying /proc/locks, so we + * keep a list on each CPU, with each list protected by its own spinlock via + * the file_lock_lglock. */ -static HLIST_HEAD(file_lock_list); -static DEFINE_SPINLOCK(file_lock_lock); +DEFINE_STATIC_LGLOCK(file_lock_lglock); +static DEFINE_PER_CPU(struct hlist_head, file_lock_list); /* * The blocked_hash is used to find POSIX lock loops for deadlock detection. @@ -490,20 +493,30 @@ static int posix_same_owner(struct file_lock *fl1, struct file_lock *fl2) return fl1->fl_owner == fl2->fl_owner; } +/* Must be called with the i_lock held! */ static inline void locks_insert_global_locks(struct file_lock *waiter) { - spin_lock(&file_lock_lock); - hlist_add_head(&waiter->fl_link, &file_lock_list); - spin_unlock(&file_lock_lock); + lg_local_lock(&file_lock_lglock); + waiter->fl_link_cpu = smp_processor_id(); + hlist_add_head(&waiter->fl_link, this_cpu_ptr(&file_lock_list)); + lg_local_unlock(&file_lock_lglock); } +/* Must be called with the i_lock held! */ static inline void locks_delete_global_locks(struct file_lock *waiter) { - spin_lock(&file_lock_lock); + /* + * Avoid taking lock if already unhashed. This is safe since this check + * is done while holding the i_lock, and new insertions into the list + * also require that it be held. + */ + if (hlist_unhashed(&waiter->fl_link)) + return; + lg_local_lock_cpu(&file_lock_lglock, waiter->fl_link_cpu); hlist_del_init(&waiter->fl_link); - spin_unlock(&file_lock_lock); + lg_local_unlock_cpu(&file_lock_lglock, waiter->fl_link_cpu); } static unsigned long @@ -2228,6 +2241,11 @@ EXPORT_SYMBOL_GPL(vfs_cancel_lock); #include #include +struct locks_iterator { + int li_cpu; + loff_t li_pos; +}; + static void lock_get_status(struct seq_file *f, struct file_lock *fl, loff_t id, char *pfx) { @@ -2301,39 +2319,41 @@ static void lock_get_status(struct seq_file *f, struct file_lock *fl, static int locks_show(struct seq_file *f, void *v) { + struct locks_iterator *iter = f->private; struct file_lock *fl, *bfl; fl = hlist_entry(v, struct file_lock, fl_link); - lock_get_status(f, fl, *((loff_t *)f->private), ""); + lock_get_status(f, fl, iter->li_pos, ""); list_for_each_entry(bfl, &fl->fl_block, fl_block) - lock_get_status(f, bfl, *((loff_t *)f->private), " ->"); + lock_get_status(f, bfl, iter->li_pos, " ->"); return 0; } static void *locks_start(struct seq_file *f, loff_t *pos) { - loff_t *p = f->private; + struct locks_iterator *iter = f->private; - spin_lock(&file_lock_lock); + iter->li_pos = *pos + 1; + lg_global_lock(&file_lock_lglock); spin_lock(&blocked_lock_lock); - *p = (*pos + 1); - return seq_hlist_start(&file_lock_list, *pos); + return seq_hlist_start_percpu(&file_lock_list, &iter->li_cpu, *pos); } static void *locks_next(struct seq_file *f, void *v, loff_t *pos) { - loff_t *p = f->private; - ++*p; - return seq_hlist_next(v, &file_lock_list, pos); + struct locks_iterator *iter = f->private; + + ++iter->li_pos; + return seq_hlist_next_percpu(v, &file_lock_list, &iter->li_cpu, pos); } static void locks_stop(struct seq_file *f, void *v) { spin_unlock(&blocked_lock_lock); - spin_unlock(&file_lock_lock); + lg_global_unlock(&file_lock_lglock); } static const struct seq_operations locks_seq_operations = { @@ -2345,7 +2365,8 @@ static const struct seq_operations locks_seq_operations = { static int locks_open(struct inode *inode, struct file *filp) { - return seq_open_private(filp, &locks_seq_operations, sizeof(loff_t)); + return seq_open_private(filp, &locks_seq_operations, + sizeof(struct locks_iterator)); } static const struct file_operations proc_locks_operations = { @@ -2445,9 +2466,16 @@ EXPORT_SYMBOL(lock_may_write); static int __init filelock_init(void) { + int i; + filelock_cache = kmem_cache_create("file_lock_cache", sizeof(struct file_lock), 0, SLAB_PANIC, NULL); + lg_lock_init(&file_lock_lglock, "file_lock_lglock"); + + for_each_possible_cpu(i) + INIT_HLIST_HEAD(per_cpu_ptr(&file_lock_list, i)); + return 0; } diff --git a/include/linux/fs.h b/include/linux/fs.h index 232a345..18e59b8 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -953,6 +953,7 @@ struct file_lock { unsigned int fl_flags; unsigned char fl_type; unsigned int fl_pid; + int fl_link_cpu; /* what cpu's list is this on? */ struct pid *fl_nspid; wait_queue_head_t fl_wait; struct file *fl_file;