percpu_counter: FBC_BATCH should be a variable
For NR_CPUS >= 16 values, FBC_BATCH is 2*NR_CPUS
Considering more and more distros are using high NR_CPUS values, it makes
sense to use a more sensible value for FBC_BATCH, and get rid of NR_CPUS.
A sensible value is 2*num_online_cpus(), with a minimum value of 32 (This
minimum value helps branch prediction in __percpu_counter_add())
We already have a hotcpu notifier, so we can adjust FBC_BATCH dynamically.
We rename FBC_BATCH to percpu_counter_batch since its not a constant
anymore.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Acked-by: David S. Miller <davem@davemloft.net>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index b0537c8..6c46c64 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1225,11 +1225,11 @@
} while (0)
#ifdef CONFIG_SMP
-/* Each CPU can accumulate FBC_BATCH blocks in their local
+/* Each CPU can accumulate percpu_counter_batch blocks in their local
* counters. So we need to make sure we have free blocks more
- * than FBC_BATCH * nr_cpu_ids. Also add a window of 4 times.
+ * than percpu_counter_batch * nr_cpu_ids. Also add a window of 4 times.
*/
-#define EXT4_FREEBLOCKS_WATERMARK (4 * (FBC_BATCH * nr_cpu_ids))
+#define EXT4_FREEBLOCKS_WATERMARK (4 * (percpu_counter_batch * nr_cpu_ids))
#else
#define EXT4_FREEBLOCKS_WATERMARK 0
#endif
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 6702a49..98d3fe7 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -2498,7 +2498,7 @@
/*
* switch to non delalloc mode if we are running low
* on free block. The free block accounting via percpu
- * counters can get slightly wrong with FBC_BATCH getting
+ * counters can get slightly wrong with percpu_counter_batch getting
* accumulated on each CPU without updating global counters
* Delalloc need an accurate free block accounting. So switch
* to non delalloc when we are near to error range.
diff --git a/include/linux/percpu_counter.h b/include/linux/percpu_counter.h
index 9007ccd..99de7a3 100644
--- a/include/linux/percpu_counter.h
+++ b/include/linux/percpu_counter.h
@@ -24,11 +24,7 @@
s32 *counters;
};
-#if NR_CPUS >= 16
-#define FBC_BATCH (NR_CPUS*2)
-#else
-#define FBC_BATCH (NR_CPUS*4)
-#endif
+extern int percpu_counter_batch;
int percpu_counter_init(struct percpu_counter *fbc, s64 amount);
int percpu_counter_init_irq(struct percpu_counter *fbc, s64 amount);
@@ -39,7 +35,7 @@
static inline void percpu_counter_add(struct percpu_counter *fbc, s64 amount)
{
- __percpu_counter_add(fbc, amount, FBC_BATCH);
+ __percpu_counter_add(fbc, amount, percpu_counter_batch);
}
static inline s64 percpu_counter_sum_positive(struct percpu_counter *fbc)
diff --git a/lib/percpu_counter.c b/lib/percpu_counter.c
index b255b93..a60bd80 100644
--- a/lib/percpu_counter.c
+++ b/lib/percpu_counter.c
@@ -9,10 +9,8 @@
#include <linux/cpu.h>
#include <linux/module.h>
-#ifdef CONFIG_HOTPLUG_CPU
static LIST_HEAD(percpu_counters);
static DEFINE_MUTEX(percpu_counters_lock);
-#endif
void percpu_counter_set(struct percpu_counter *fbc, s64 amount)
{
@@ -111,13 +109,24 @@
}
EXPORT_SYMBOL(percpu_counter_destroy);
-#ifdef CONFIG_HOTPLUG_CPU
+int percpu_counter_batch __read_mostly = 32;
+EXPORT_SYMBOL(percpu_counter_batch);
+
+static void compute_batch_value(void)
+{
+ int nr = num_online_cpus();
+
+ percpu_counter_batch = max(32, nr*2);
+}
+
static int __cpuinit percpu_counter_hotcpu_callback(struct notifier_block *nb,
unsigned long action, void *hcpu)
{
+#ifdef CONFIG_HOTPLUG_CPU
unsigned int cpu;
struct percpu_counter *fbc;
+ compute_batch_value();
if (action != CPU_DEAD)
return NOTIFY_OK;
@@ -134,13 +143,14 @@
spin_unlock_irqrestore(&fbc->lock, flags);
}
mutex_unlock(&percpu_counters_lock);
+#endif
return NOTIFY_OK;
}
static int __init percpu_counter_startup(void)
{
+ compute_batch_value();
hotcpu_notifier(percpu_counter_hotcpu_callback, 0);
return 0;
}
module_init(percpu_counter_startup);
-#endif
