![Eric Dumazet](/assets/img/avatar_default.png)
Some oprofile results obtained while using tbench on a 2x2 cpu machine were very surprising. For example, loopback_xmit() function was using high number of cpu cycles to perform the statistic updates, supposed to be real cheap since they use percpu data pcpu_lstats = netdev_priv(dev); lb_stats = per_cpu_ptr(pcpu_lstats, smp_processor_id()); lb_stats->packets++; /* HERE : serious contention */ lb_stats->bytes += skb->len; struct pcpu_lstats is a small structure containing two longs. It appears that on my 32bits platform, alloc_percpu(8) allocates a single cache line, instead of giving to each cpu a separate cache line. Using the following patch gave me impressive boost in various benchmarks ( 6 % in tbench) (all percpu_counters hit this bug too) Long term fix (ie >= 2.6.26) would be to let each CPU allocate their own block of memory, so that we dont need to roudup sizes to L1_CACHE_BYTES, or merging the SGI stuff of course... Note : SLUB vs SLAB is important here to *show* the improvement, since they dont have the same minimum allocation sizes (8 bytes vs 32 bytes). This could very well explain regressions some guys reported when they switched to SLUB. Signed-off-by: Eric Dumazet <dada1@cosmosbay.com> 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>
141 lines
3.9 KiB
C
141 lines
3.9 KiB
C
/*
|
|
* linux/mm/allocpercpu.c
|
|
*
|
|
* Separated from slab.c August 11, 2006 Christoph Lameter <clameter@sgi.com>
|
|
*/
|
|
#include <linux/mm.h>
|
|
#include <linux/module.h>
|
|
|
|
#ifndef cache_line_size
|
|
#define cache_line_size() L1_CACHE_BYTES
|
|
#endif
|
|
|
|
/**
|
|
* percpu_depopulate - depopulate per-cpu data for given cpu
|
|
* @__pdata: per-cpu data to depopulate
|
|
* @cpu: depopulate per-cpu data for this cpu
|
|
*
|
|
* Depopulating per-cpu data for a cpu going offline would be a typical
|
|
* use case. You need to register a cpu hotplug handler for that purpose.
|
|
*/
|
|
void percpu_depopulate(void *__pdata, int cpu)
|
|
{
|
|
struct percpu_data *pdata = __percpu_disguise(__pdata);
|
|
|
|
kfree(pdata->ptrs[cpu]);
|
|
pdata->ptrs[cpu] = NULL;
|
|
}
|
|
EXPORT_SYMBOL_GPL(percpu_depopulate);
|
|
|
|
/**
|
|
* percpu_depopulate_mask - depopulate per-cpu data for some cpu's
|
|
* @__pdata: per-cpu data to depopulate
|
|
* @mask: depopulate per-cpu data for cpu's selected through mask bits
|
|
*/
|
|
void __percpu_depopulate_mask(void *__pdata, cpumask_t *mask)
|
|
{
|
|
int cpu;
|
|
for_each_cpu_mask(cpu, *mask)
|
|
percpu_depopulate(__pdata, cpu);
|
|
}
|
|
EXPORT_SYMBOL_GPL(__percpu_depopulate_mask);
|
|
|
|
/**
|
|
* percpu_populate - populate per-cpu data for given cpu
|
|
* @__pdata: per-cpu data to populate further
|
|
* @size: size of per-cpu object
|
|
* @gfp: may sleep or not etc.
|
|
* @cpu: populate per-data for this cpu
|
|
*
|
|
* Populating per-cpu data for a cpu coming online would be a typical
|
|
* use case. You need to register a cpu hotplug handler for that purpose.
|
|
* Per-cpu object is populated with zeroed buffer.
|
|
*/
|
|
void *percpu_populate(void *__pdata, size_t size, gfp_t gfp, int cpu)
|
|
{
|
|
struct percpu_data *pdata = __percpu_disguise(__pdata);
|
|
int node = cpu_to_node(cpu);
|
|
|
|
/*
|
|
* We should make sure each CPU gets private memory.
|
|
*/
|
|
size = roundup(size, cache_line_size());
|
|
|
|
BUG_ON(pdata->ptrs[cpu]);
|
|
if (node_online(node))
|
|
pdata->ptrs[cpu] = kmalloc_node(size, gfp|__GFP_ZERO, node);
|
|
else
|
|
pdata->ptrs[cpu] = kzalloc(size, gfp);
|
|
return pdata->ptrs[cpu];
|
|
}
|
|
EXPORT_SYMBOL_GPL(percpu_populate);
|
|
|
|
/**
|
|
* percpu_populate_mask - populate per-cpu data for more cpu's
|
|
* @__pdata: per-cpu data to populate further
|
|
* @size: size of per-cpu object
|
|
* @gfp: may sleep or not etc.
|
|
* @mask: populate per-cpu data for cpu's selected through mask bits
|
|
*
|
|
* Per-cpu objects are populated with zeroed buffers.
|
|
*/
|
|
int __percpu_populate_mask(void *__pdata, size_t size, gfp_t gfp,
|
|
cpumask_t *mask)
|
|
{
|
|
cpumask_t populated = CPU_MASK_NONE;
|
|
int cpu;
|
|
|
|
for_each_cpu_mask(cpu, *mask)
|
|
if (unlikely(!percpu_populate(__pdata, size, gfp, cpu))) {
|
|
__percpu_depopulate_mask(__pdata, &populated);
|
|
return -ENOMEM;
|
|
} else
|
|
cpu_set(cpu, populated);
|
|
return 0;
|
|
}
|
|
EXPORT_SYMBOL_GPL(__percpu_populate_mask);
|
|
|
|
/**
|
|
* percpu_alloc_mask - initial setup of per-cpu data
|
|
* @size: size of per-cpu object
|
|
* @gfp: may sleep or not etc.
|
|
* @mask: populate per-data for cpu's selected through mask bits
|
|
*
|
|
* Populating per-cpu data for all online cpu's would be a typical use case,
|
|
* which is simplified by the percpu_alloc() wrapper.
|
|
* Per-cpu objects are populated with zeroed buffers.
|
|
*/
|
|
void *__percpu_alloc_mask(size_t size, gfp_t gfp, cpumask_t *mask)
|
|
{
|
|
/*
|
|
* We allocate whole cache lines to avoid false sharing
|
|
*/
|
|
size_t sz = roundup(nr_cpu_ids * sizeof(void *), cache_line_size());
|
|
void *pdata = kzalloc(sz, gfp);
|
|
void *__pdata = __percpu_disguise(pdata);
|
|
|
|
if (unlikely(!pdata))
|
|
return NULL;
|
|
if (likely(!__percpu_populate_mask(__pdata, size, gfp, mask)))
|
|
return __pdata;
|
|
kfree(pdata);
|
|
return NULL;
|
|
}
|
|
EXPORT_SYMBOL_GPL(__percpu_alloc_mask);
|
|
|
|
/**
|
|
* percpu_free - final cleanup of per-cpu data
|
|
* @__pdata: object to clean up
|
|
*
|
|
* We simply clean up any per-cpu object left. No need for the client to
|
|
* track and specify through a bis mask which per-cpu objects are to free.
|
|
*/
|
|
void percpu_free(void *__pdata)
|
|
{
|
|
if (unlikely(!__pdata))
|
|
return;
|
|
__percpu_depopulate_mask(__pdata, &cpu_possible_map);
|
|
kfree(__percpu_disguise(__pdata));
|
|
}
|
|
EXPORT_SYMBOL_GPL(percpu_free);
|