Re: [PATCH v2 14/14] locks: move file_lock_list to a set of percpuhlist_heads and convert file_lock_lock to an lglock

[J. Bruce Fields]

On Tue, Jun 11, 2013 at 07:09:08AM -0400, Jeff Layton wrote:
> 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 its real size.

ACK.--b.

> 
> Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
> ---
>  fs/locks.c         |   57 +++++++++++++++++++++++++++++++++++----------------
>  include/linux/fs.h |    1 +
>  2 files changed, 40 insertions(+), 18 deletions(-)
> 
> diff --git a/fs/locks.c b/fs/locks.c
> index 8124fc1..094eb4d 100644
> --- a/fs/locks.c
> +++ b/fs/locks.c
> @@ -127,6 +127,8 @@
>  #include <linux/rcupdate.h>
>  #include <linux/pid_namespace.h>
>  #include <linux/hashtable.h>
> +#include <linux/percpu.h>
> +#include <linux/lglock.h>
>  
>  #include <asm/uaccess.h>
>  
> @@ -165,8 +167,8 @@ int lease_break_time = 45;
>  static DEFINE_SPINLOCK(blocked_hash_lock);
>  static DEFINE_HASHTABLE(blocked_hash, BLOCKED_HASH_BITS);
>  
> -static DEFINE_SPINLOCK(file_lock_lock);
> -static HLIST_HEAD(file_lock_list);
> +DEFINE_STATIC_LGLOCK(file_lock_lglock);
> +static DEFINE_PER_CPU(struct hlist_head, file_lock_list);
>  
>  static struct kmem_cache *filelock_cache __read_mostly;
>  
> @@ -512,17 +514,21 @@ locks_delete_global_blocked(struct file_lock *waiter)
>  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);
>  }
>  
>  static inline void
>  locks_delete_global_locks(struct file_lock *waiter)
>  {
> -	spin_lock(&file_lock_lock);
> +	/* avoid taking lock if already unhashed */
> +	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);
>  }
>  
>  /* Remove waiter from blocker's block list.
> @@ -2228,6 +2234,11 @@ EXPORT_SYMBOL_GPL(vfs_cancel_lock);
>  #include <linux/proc_fs.h>
>  #include <linux/seq_file.h>
>  
> +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)
>  {
> @@ -2302,16 +2313,17 @@ static void lock_get_status(struct seq_file *f, struct file_lock *fl,
>  static int locks_show(struct seq_file *f, void *v)
>  {
>  	int bkt;
> +	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, "");
>  
>  	spin_lock(&blocked_hash_lock);
>  	hash_for_each(blocked_hash, bkt, bfl, fl_link) {
>  		if (bfl->fl_next == fl)
> -			lock_get_status(f, bfl, *((loff_t *)f->private), " ->");
> +			lock_get_status(f, bfl, iter->li_pos, " ->");
>  	}
>  	spin_unlock(&blocked_hash_lock);
>  
> @@ -2320,23 +2332,24 @@ static int locks_show(struct seq_file *f, void *v)
>  
>  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);
> -	*p = (*pos + 1);
> -	return seq_hlist_start(&file_lock_list, *pos);
> +	iter->li_pos = *pos + 1;
> +	lg_global_lock(&file_lock_lglock);
> +	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(&file_lock_lock);
> +	lg_global_unlock(&file_lock_lglock);
>  }
>  
>  static const struct seq_operations locks_seq_operations = {
> @@ -2348,7 +2361,8 @@ 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 = {
> @@ -2448,9 +2462,16 @@ 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;
> -- 
> 1.7.1
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/