The nr_dentry stat is a globally touched cacheline and atomic operation
twice over the lifetime of a dentry. It is used for the benfit of userspace
only. Turn it into a per-cpu counter and always decrement it in d_free instead
of doing various batching operations to reduce lock hold times in the callers.
Based on an earlier patch from Nick Piggin <npiggin@xxxxxxx>.
Signed-off-by: Christoph Hellwig <hch@xxxxxx>
Index: linux-2.6/fs/dcache.c
===================================================================
--- linux-2.6.orig/fs/dcache.c 2010-10-10 09:36:28.918003777 +0200
+++ linux-2.6/fs/dcache.c 2010-10-10 09:36:31.325003987 +0200
@@ -67,6 +67,19 @@ struct dentry_stat_t dentry_stat = {
.age_limit = 45,
};
+static struct percpu_counter nr_dentry __cacheline_aligned_in_smp;
+static struct percpu_counter nr_dentry_unused __cacheline_aligned_in_smp;
+
+#if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS)
+int proc_nr_dentry(ctl_table *table, int write, void __user *buffer,
+ size_t *lenp, loff_t *ppos)
+{
+ dentry_stat.nr_dentry = percpu_counter_sum_positive(&nr_dentry);
+ dentry_stat.nr_unused = percpu_counter_sum_positive(&nr_dentry_unused);
+ return proc_dointvec(table, write, buffer, lenp, ppos);
+}
+#endif
+
static void __d_free(struct rcu_head *head)
{
struct dentry *dentry = container_of(head, struct dentry, d_u.d_rcu);
@@ -78,13 +91,14 @@ static void __d_free(struct rcu_head *he
}
/*
- * no dcache_lock, please. The caller must decrement dentry_stat.nr_dentry
- * inside dcache_lock.
+ * no dcache_lock, please.
*/
static void d_free(struct dentry *dentry)
{
+ percpu_counter_dec(&nr_dentry);
if (dentry->d_op && dentry->d_op->d_release)
dentry->d_op->d_release(dentry);
+
/* if dentry was never inserted into hash, immediate free is OK */
if (hlist_unhashed(&dentry->d_hash))
__d_free(&dentry->d_u.d_rcu);
@@ -125,14 +139,14 @@ static void dentry_lru_add(struct dentry
{
list_add(&dentry->d_lru, &dentry->d_sb->s_dentry_lru);
dentry->d_sb->s_nr_dentry_unused++;
- dentry_stat.nr_unused++;
+ percpu_counter_inc(&nr_dentry_unused);
}
static void dentry_lru_add_tail(struct dentry *dentry)
{
list_add_tail(&dentry->d_lru, &dentry->d_sb->s_dentry_lru);
dentry->d_sb->s_nr_dentry_unused++;
- dentry_stat.nr_unused++;
+ percpu_counter_inc(&nr_dentry_unused);
}
static void dentry_lru_del(struct dentry *dentry)
@@ -140,7 +154,7 @@ static void dentry_lru_del(struct dentry
if (!list_empty(&dentry->d_lru)) {
list_del(&dentry->d_lru);
dentry->d_sb->s_nr_dentry_unused--;
- dentry_stat.nr_unused--;
+ percpu_counter_dec(&nr_dentry_unused);
}
}
@@ -149,7 +163,7 @@ static void dentry_lru_del_init(struct d
if (likely(!list_empty(&dentry->d_lru))) {
list_del_init(&dentry->d_lru);
dentry->d_sb->s_nr_dentry_unused--;
- dentry_stat.nr_unused--;
+ percpu_counter_dec(&nr_dentry_unused);
}
}
@@ -168,7 +182,6 @@ static struct dentry *d_kill(struct dent
struct dentry *parent;
list_del(&dentry->d_u.d_child);
- dentry_stat.nr_dentry--; /* For d_free, below */
/*drops the locks, at that point nobody can reach this dentry */
dentry_iput(dentry);
if (IS_ROOT(dentry))
@@ -314,7 +327,6 @@ int d_invalidate(struct dentry * dentry)
EXPORT_SYMBOL(d_invalidate);
/* This should be called _only_ with dcache_lock held */
-
static inline struct dentry * __dget_locked(struct dentry *dentry)
{
atomic_inc(&dentry->d_count);
@@ -534,7 +546,7 @@ static void prune_dcache(int count)
{
struct super_block *sb, *p = NULL;
int w_count;
- int unused = dentry_stat.nr_unused;
+ int unused = percpu_counter_sum_positive(&nr_dentry_unused);
int prune_ratio;
int pruned;
@@ -699,20 +711,13 @@ static void shrink_dcache_for_umount_sub
* otherwise we ascend to the parent and move to the
* next sibling if there is one */
if (!parent)
- goto out;
-
+ return;
dentry = parent;
-
} while (list_empty(&dentry->d_subdirs));
dentry = list_entry(dentry->d_subdirs.next,
struct dentry, d_u.d_child);
}
-out:
- /* several dentries were freed, need to correct nr_dentry */
- spin_lock(&dcache_lock);
- dentry_stat.nr_dentry -= detached;
- spin_unlock(&dcache_lock);
}
/*
@@ -896,12 +901,16 @@ EXPORT_SYMBOL(shrink_dcache_parent);
*/
static int shrink_dcache_memory(struct shrinker *shrink, int nr, gfp_t gfp_mask)
{
+ int nr_unused;
+
if (nr) {
if (!(gfp_mask & __GFP_FS))
return -1;
prune_dcache(nr);
}
- return (dentry_stat.nr_unused / 100) * sysctl_vfs_cache_pressure;
+
+ nr_unused = percpu_counter_sum_positive(&nr_dentry_unused);
+ return (nr_unused / 100) * sysctl_vfs_cache_pressure;
}
static struct shrinker dcache_shrinker = {
@@ -968,9 +977,10 @@ struct dentry *d_alloc(struct dentry * p
spin_lock(&dcache_lock);
if (parent)
list_add(&dentry->d_u.d_child, &parent->d_subdirs);
- dentry_stat.nr_dentry++;
spin_unlock(&dcache_lock);
+ percpu_counter_inc(&nr_dentry);
+
return dentry;
}
EXPORT_SYMBOL(d_alloc);
@@ -2417,6 +2427,9 @@ static void __init dcache_init(void)
{
int loop;
+ percpu_counter_init(&nr_dentry, 0);
+ percpu_counter_init(&nr_dentry_unused, 0);
+
/*
* A constructor could be added for stable state like the lists,
* but it is probably not worth it because of the cache nature
Index: linux-2.6/include/linux/fs.h
===================================================================
--- linux-2.6.orig/include/linux/fs.h 2010-10-10 09:36:05.400003777 +0200
+++ linux-2.6/include/linux/fs.h 2010-10-10 09:36:31.326004057 +0200
@@ -2483,6 +2483,8 @@ ssize_t simple_attr_write(struct file *f
struct ctl_table;
int proc_nr_files(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp, loff_t *ppos);
+int proc_nr_dentry(struct ctl_table *table, int write,
+ void __user *buffer, size_t *lenp, loff_t *ppos);
int proc_nr_inodes(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp, loff_t *ppos);
int __init get_filesystem_list(char *buf);
Index: linux-2.6/kernel/sysctl.c
===================================================================
--- linux-2.6.orig/kernel/sysctl.c 2010-10-10 09:36:05.412003707 +0200
+++ linux-2.6/kernel/sysctl.c 2010-10-10 09:36:31.331004057 +0200
@@ -1377,7 +1377,7 @@ static struct ctl_table fs_table[] = {
.data = &dentry_stat,
.maxlen = 6*sizeof(int),
.mode = 0444,
- .proc_handler = proc_dointvec,
+ .proc_handler = proc_nr_dentry,
},
{
.procname = "overflowuid",
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
