Commit 93197a36 authored by Nathan Lynch's avatar Nathan Lynch Committed by Benjamin Herrenschmidt

powerpc: Rewrite sysfs processor cache info code

The current code for providing processor cache information in sysfs
has the following deficiencies:
- several complex functions that are hard to understand
- implicit recursion (cache_desc_release -> kobject_put -> cache_desc_release)
- explicit recursion (create_cache_index_info)
- use of two per-cpu arrays when one would suffice
- duplication of work on systems where CPUs share cache

Also, when I looked at implementing support for a shared_cpu_map
attribute, it was pretty much impossible to handle hotplug without
checking every single online CPU's cache_desc list and fixing things
up... not that this is a hot path, but it would have introduced
O(n^2)-ish behavior during boot.  Addressing this involved rethinking
the core data structures used, which didn't lend itself to an
incremental approach.

This implementation maintains a "forest" (potentially more than one
tree) of cache objects which reflects the system's cache topology.
Cache objects are instantiated as needed as CPUs come online.  A
per-cpu array is used mainly for sysfs-related bookkeeping; the
objects in the array just point to the appropriate points in the
forest.

This maintains compatibility with the existing code and includes some
enhancements:
- Implement the shared_cpu_map attribute, which is essential for
  enabling userspace to discover the system's overall cache topology.
- Use cache-block-size properties if cache-line-size is not available.

I chose to place this implementation in a new file since it would have
roughly doubled the size of sysfs.c, which is already kind of messy.
Signed-off-by: default avatarNathan Lynch <ntl@pobox.com>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
parent 5c9a2606
......@@ -29,7 +29,7 @@ endif
obj-y := cputable.o ptrace.o syscalls.o \
irq.o align.o signal_32.o pmc.o vdso.o \
init_task.o process.o systbl.o idle.o \
signal.o sysfs.o
signal.o sysfs.o cacheinfo.o
obj-y += vdso32/
obj-$(CONFIG_PPC64) += setup_64.o sys_ppc32.o \
signal_64.o ptrace32.o \
......
This diff is collapsed.
#ifndef _PPC_CACHEINFO_H
#define _PPC_CACHEINFO_H
/* These are just hooks for sysfs.c to use. */
extern void cacheinfo_cpu_online(unsigned int cpu_id);
extern void cacheinfo_cpu_offline(unsigned int cpu_id);
#endif /* _PPC_CACHEINFO_H */
......@@ -18,6 +18,8 @@
#include <asm/machdep.h>
#include <asm/smp.h>
#include "cacheinfo.h"
#ifdef CONFIG_PPC64
#include <asm/paca.h>
#include <asm/lppaca.h>
......@@ -25,8 +27,6 @@
static DEFINE_PER_CPU(struct cpu, cpu_devices);
static DEFINE_PER_CPU(struct kobject *, cache_toplevel);
/*
* SMT snooze delay stuff, 64-bit only for now
*/
......@@ -343,283 +343,6 @@ static struct sysdev_attribute pa6t_attrs[] = {
#endif /* HAS_PPC_PMC_PA6T */
#endif /* HAS_PPC_PMC_CLASSIC */
struct cache_desc {
struct kobject kobj;
struct cache_desc *next;
const char *type; /* Instruction, Data, or Unified */
u32 size; /* total cache size in KB */
u32 line_size; /* in bytes */
u32 nr_sets; /* number of sets */
u32 level; /* e.g. 1, 2, 3... */
u32 associativity; /* e.g. 8-way... 0 is fully associative */
};
DEFINE_PER_CPU(struct cache_desc *, cache_desc);
static struct cache_desc *kobj_to_cache_desc(struct kobject *k)
{
return container_of(k, struct cache_desc, kobj);
}
static void cache_desc_release(struct kobject *k)
{
struct cache_desc *desc = kobj_to_cache_desc(k);
pr_debug("%s: releasing %s\n", __func__, kobject_name(k));
if (desc->next)
kobject_put(&desc->next->kobj);
kfree(kobj_to_cache_desc(k));
}
static ssize_t cache_desc_show(struct kobject *k, struct attribute *attr, char *buf)
{
struct kobj_attribute *kobj_attr;
kobj_attr = container_of(attr, struct kobj_attribute, attr);
return kobj_attr->show(k, kobj_attr, buf);
}
static struct sysfs_ops cache_desc_sysfs_ops = {
.show = cache_desc_show,
};
static struct kobj_type cache_desc_type = {
.release = cache_desc_release,
.sysfs_ops = &cache_desc_sysfs_ops,
};
static ssize_t cache_size_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
{
struct cache_desc *cache = kobj_to_cache_desc(k);
return sprintf(buf, "%uK\n", cache->size);
}
static struct kobj_attribute cache_size_attr =
__ATTR(size, 0444, cache_size_show, NULL);
static ssize_t cache_line_size_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
{
struct cache_desc *cache = kobj_to_cache_desc(k);
return sprintf(buf, "%u\n", cache->line_size);
}
static struct kobj_attribute cache_line_size_attr =
__ATTR(coherency_line_size, 0444, cache_line_size_show, NULL);
static ssize_t cache_nr_sets_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
{
struct cache_desc *cache = kobj_to_cache_desc(k);
return sprintf(buf, "%u\n", cache->nr_sets);
}
static struct kobj_attribute cache_nr_sets_attr =
__ATTR(number_of_sets, 0444, cache_nr_sets_show, NULL);
static ssize_t cache_type_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
{
struct cache_desc *cache = kobj_to_cache_desc(k);
return sprintf(buf, "%s\n", cache->type);
}
static struct kobj_attribute cache_type_attr =
__ATTR(type, 0444, cache_type_show, NULL);
static ssize_t cache_level_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
{
struct cache_desc *cache = kobj_to_cache_desc(k);
return sprintf(buf, "%u\n", cache->level);
}
static struct kobj_attribute cache_level_attr =
__ATTR(level, 0444, cache_level_show, NULL);
static ssize_t cache_assoc_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
{
struct cache_desc *cache = kobj_to_cache_desc(k);
return sprintf(buf, "%u\n", cache->associativity);
}
static struct kobj_attribute cache_assoc_attr =
__ATTR(ways_of_associativity, 0444, cache_assoc_show, NULL);
struct cache_desc_info {
const char *type;
const char *size_prop;
const char *line_size_prop;
const char *nr_sets_prop;
};
/* PowerPC Processor binding says the [di]-cache-* must be equal on
* unified caches, so just use d-cache properties. */
static struct cache_desc_info ucache_info = {
.type = "Unified",
.size_prop = "d-cache-size",
.line_size_prop = "d-cache-line-size",
.nr_sets_prop = "d-cache-sets",
};
static struct cache_desc_info dcache_info = {
.type = "Data",
.size_prop = "d-cache-size",
.line_size_prop = "d-cache-line-size",
.nr_sets_prop = "d-cache-sets",
};
static struct cache_desc_info icache_info = {
.type = "Instruction",
.size_prop = "i-cache-size",
.line_size_prop = "i-cache-line-size",
.nr_sets_prop = "i-cache-sets",
};
static struct cache_desc * __cpuinit create_cache_desc(struct device_node *np, struct kobject *parent, int index, int level, struct cache_desc_info *info)
{
const u32 *cache_line_size;
struct cache_desc *new;
const u32 *cache_size;
const u32 *nr_sets;
int rc;
new = kzalloc(sizeof(*new), GFP_KERNEL);
if (!new)
return NULL;
rc = kobject_init_and_add(&new->kobj, &cache_desc_type, parent,
"index%d", index);
if (rc)
goto err;
/* type */
new->type = info->type;
rc = sysfs_create_file(&new->kobj, &cache_type_attr.attr);
WARN_ON(rc);
/* level */
new->level = level;
rc = sysfs_create_file(&new->kobj, &cache_level_attr.attr);
WARN_ON(rc);
/* size */
cache_size = of_get_property(np, info->size_prop, NULL);
if (cache_size) {
new->size = *cache_size / 1024;
rc = sysfs_create_file(&new->kobj,
&cache_size_attr.attr);
WARN_ON(rc);
}
/* coherency_line_size */
cache_line_size = of_get_property(np, info->line_size_prop, NULL);
if (cache_line_size) {
new->line_size = *cache_line_size;
rc = sysfs_create_file(&new->kobj,
&cache_line_size_attr.attr);
WARN_ON(rc);
}
/* number_of_sets */
nr_sets = of_get_property(np, info->nr_sets_prop, NULL);
if (nr_sets) {
new->nr_sets = *nr_sets;
rc = sysfs_create_file(&new->kobj,
&cache_nr_sets_attr.attr);
WARN_ON(rc);
}
/* ways_of_associativity */
if (new->nr_sets == 1) {
/* fully associative */
new->associativity = 0;
goto create_assoc;
}
if (new->nr_sets && new->size && new->line_size) {
/* If we have values for all of these we can derive
* the associativity. */
new->associativity =
((new->size * 1024) / new->nr_sets) / new->line_size;
create_assoc:
rc = sysfs_create_file(&new->kobj,
&cache_assoc_attr.attr);
WARN_ON(rc);
}
return new;
err:
kfree(new);
return NULL;
}
static bool cache_is_unified(struct device_node *np)
{
return of_get_property(np, "cache-unified", NULL);
}
static struct cache_desc * __cpuinit create_cache_index_info(struct device_node *np, struct kobject *parent, int index, int level)
{
struct device_node *next_cache;
struct cache_desc *new, **end;
pr_debug("%s(node = %s, index = %d)\n", __func__, np->full_name, index);
if (cache_is_unified(np)) {
new = create_cache_desc(np, parent, index, level,
&ucache_info);
} else {
new = create_cache_desc(np, parent, index, level,
&dcache_info);
if (new) {
index++;
new->next = create_cache_desc(np, parent, index, level,
&icache_info);
}
}
if (!new)
return NULL;
end = &new->next;
while (*end)
end = &(*end)->next;
next_cache = of_find_next_cache_node(np);
if (!next_cache)
goto out;
*end = create_cache_index_info(next_cache, parent, ++index, ++level);
of_node_put(next_cache);
out:
return new;
}
static void __cpuinit create_cache_info(struct sys_device *sysdev)
{
struct kobject *cache_toplevel;
struct device_node *np = NULL;
int cpu = sysdev->id;
cache_toplevel = kobject_create_and_add("cache", &sysdev->kobj);
if (!cache_toplevel)
return;
per_cpu(cache_toplevel, cpu) = cache_toplevel;
np = of_get_cpu_node(cpu, NULL);
if (np != NULL) {
per_cpu(cache_desc, cpu) =
create_cache_index_info(np, cache_toplevel, 0, 1);
of_node_put(np);
}
return;
}
static void __cpuinit register_cpu_online(unsigned int cpu)
{
struct cpu *c = &per_cpu(cpu_devices, cpu);
......@@ -684,25 +407,10 @@ static void __cpuinit register_cpu_online(unsigned int cpu)
sysdev_create_file(s, &attr_dscr);
#endif /* CONFIG_PPC64 */
create_cache_info(s);
cacheinfo_cpu_online(cpu);
}
#ifdef CONFIG_HOTPLUG_CPU
static void remove_cache_info(struct sys_device *sysdev)
{
struct kobject *cache_toplevel;
struct cache_desc *cache_desc;
int cpu = sysdev->id;
cache_desc = per_cpu(cache_desc, cpu);
if (cache_desc != NULL)
kobject_put(&cache_desc->kobj);
cache_toplevel = per_cpu(cache_toplevel, cpu);
if (cache_toplevel != NULL)
kobject_put(cache_toplevel);
}
static void unregister_cpu_online(unsigned int cpu)
{
struct cpu *c = &per_cpu(cpu_devices, cpu);
......@@ -769,7 +477,7 @@ static void unregister_cpu_online(unsigned int cpu)
sysdev_remove_file(s, &attr_dscr);
#endif /* CONFIG_PPC64 */
remove_cache_info(s);
cacheinfo_cpu_offline(cpu);
}
#endif /* CONFIG_HOTPLUG_CPU */
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment