40064aeca3
The percpu memory allocator is experiencing scalability issues when
allocating and freeing large numbers of counters as in BPF.
Additionally, there is a corner case where iteration is triggered over
all chunks if the contig_hint is the right size, but wrong alignment.
This patch replaces the area map allocator with a basic bitmap allocator
implementation. Each subsequent patch will introduce new features and
replace full scanning functions with faster non-scanning options when
possible.
Implementation:
This patchset removes the area map allocator in favor of a bitmap
allocator backed by metadata blocks. The primary goal is to provide
consistency in performance and memory footprint with a focus on small
allocations (< 64 bytes). The bitmap removes the heavy memmove from the
freeing critical path and provides a consistent memory footprint. The
metadata blocks provide a bound on the amount of scanning required by
maintaining a set of hints.
In an effort to make freeing fast, the metadata is updated on the free
path if the new free area makes a page free, a block free, or spans
across blocks. This causes the chunk's contig hint to potentially be
smaller than what it could allocate by up to the smaller of a page or a
block. If the chunk's contig hint is contained within a block, a check
occurs and the hint is kept accurate. Metadata is always kept accurate
on allocation, so there will not be a situation where a chunk has a
later contig hint than available.
Evaluation:
I have primarily done testing against a simple workload of allocation of
1 million objects (2^20) of varying size. Deallocation was done by in
order, alternating, and in reverse. These numbers were collected after
rebasing ontop of a80099a152. I present the worst-case numbers here:
Area Map Allocator:
Object Size | Alloc Time (ms) | Free Time (ms)
----------------------------------------------
4B | 310 | 4770
16B | 557 | 1325
64B | 436 | 273
256B | 776 | 131
1024B | 3280 | 122
Bitmap Allocator:
Object Size | Alloc Time (ms) | Free Time (ms)
----------------------------------------------
4B | 490 | 70
16B | 515 | 75
64B | 610 | 80
256B | 950 | 100
1024B | 3520 | 200
This data demonstrates the inability for the area map allocator to
handle less than ideal situations. In the best case of reverse
deallocation, the area map allocator was able to perform within range
of the bitmap allocator. In the worst case situation, freeing took
nearly 5 seconds for 1 million 4-byte objects. The bitmap allocator
dramatically improves the consistency of the free path. The small
allocations performed nearly identical regardless of the freeing
pattern.
While it does add to the allocation latency, the allocation scenario
here is optimal for the area map allocator. The area map allocator runs
into trouble when it is allocating in chunks where the latter half is
full. It is difficult to replicate this, so I present a variant where
the pages are second half filled. Freeing was done sequentially. Below
are the numbers for this scenario:
Area Map Allocator:
Object Size | Alloc Time (ms) | Free Time (ms)
----------------------------------------------
4B | 4118 | 4892
16B | 1651 | 1163
64B | 598 | 285
256B | 771 | 158
1024B | 3034 | 160
Bitmap Allocator:
Object Size | Alloc Time (ms) | Free Time (ms)
----------------------------------------------
4B | 481 | 67
16B | 506 | 69
64B | 636 | 75
256B | 892 | 90
1024B | 3262 | 147
The data shows a parabolic curve of performance for the area map
allocator. This is due to the memmove operation being the dominant cost
with the lower object sizes as more objects are packed in a chunk and at
higher object sizes, the traversal of the chunk slots is the dominating
cost. The bitmap allocator suffers this problem as well. The above data
shows the inability to scale for the allocation path with the area map
allocator and that the bitmap allocator demonstrates consistent
performance in general.
The second problem of additional scanning can result in the area map
allocator completing in 52 minutes when trying to allocate 1 million
4-byte objects with 8-byte alignment. The same workload takes
approximately 16 seconds to complete for the bitmap allocator.
V2:
Fixed a bug in pcpu_alloc_first_chunk end_offset was setting the bitmap
using bytes instead of bits.
Added a comment to pcpu_cnt_pop_pages to explain bitmap_weight.
Signed-off-by: Dennis Zhou <dennisszhou@gmail.com>
Reviewed-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
137 lines
4.3 KiB
C
137 lines
4.3 KiB
C
#ifndef __LINUX_PERCPU_H
|
|
#define __LINUX_PERCPU_H
|
|
|
|
#include <linux/mmdebug.h>
|
|
#include <linux/preempt.h>
|
|
#include <linux/smp.h>
|
|
#include <linux/cpumask.h>
|
|
#include <linux/printk.h>
|
|
#include <linux/pfn.h>
|
|
#include <linux/init.h>
|
|
|
|
#include <asm/percpu.h>
|
|
|
|
/* enough to cover all DEFINE_PER_CPUs in modules */
|
|
#ifdef CONFIG_MODULES
|
|
#define PERCPU_MODULE_RESERVE (8 << 10)
|
|
#else
|
|
#define PERCPU_MODULE_RESERVE 0
|
|
#endif
|
|
|
|
/* minimum unit size, also is the maximum supported allocation size */
|
|
#define PCPU_MIN_UNIT_SIZE PFN_ALIGN(32 << 10)
|
|
|
|
/* minimum allocation size and shift in bytes */
|
|
#define PCPU_MIN_ALLOC_SHIFT 2
|
|
#define PCPU_MIN_ALLOC_SIZE (1 << PCPU_MIN_ALLOC_SHIFT)
|
|
|
|
/*
|
|
* Percpu allocator can serve percpu allocations before slab is
|
|
* initialized which allows slab to depend on the percpu allocator.
|
|
* The following two parameters decide how much resource to
|
|
* preallocate for this. Keep PERCPU_DYNAMIC_RESERVE equal to or
|
|
* larger than PERCPU_DYNAMIC_EARLY_SIZE.
|
|
*/
|
|
#define PERCPU_DYNAMIC_EARLY_SLOTS 128
|
|
#define PERCPU_DYNAMIC_EARLY_SIZE (12 << 10)
|
|
|
|
/*
|
|
* PERCPU_DYNAMIC_RESERVE indicates the amount of free area to piggy
|
|
* back on the first chunk for dynamic percpu allocation if arch is
|
|
* manually allocating and mapping it for faster access (as a part of
|
|
* large page mapping for example).
|
|
*
|
|
* The following values give between one and two pages of free space
|
|
* after typical minimal boot (2-way SMP, single disk and NIC) with
|
|
* both defconfig and a distro config on x86_64 and 32. More
|
|
* intelligent way to determine this would be nice.
|
|
*/
|
|
#if BITS_PER_LONG > 32
|
|
#define PERCPU_DYNAMIC_RESERVE (28 << 10)
|
|
#else
|
|
#define PERCPU_DYNAMIC_RESERVE (20 << 10)
|
|
#endif
|
|
|
|
extern void *pcpu_base_addr;
|
|
extern const unsigned long *pcpu_unit_offsets;
|
|
|
|
struct pcpu_group_info {
|
|
int nr_units; /* aligned # of units */
|
|
unsigned long base_offset; /* base address offset */
|
|
unsigned int *cpu_map; /* unit->cpu map, empty
|
|
* entries contain NR_CPUS */
|
|
};
|
|
|
|
struct pcpu_alloc_info {
|
|
size_t static_size;
|
|
size_t reserved_size;
|
|
size_t dyn_size;
|
|
size_t unit_size;
|
|
size_t atom_size;
|
|
size_t alloc_size;
|
|
size_t __ai_size; /* internal, don't use */
|
|
int nr_groups; /* 0 if grouping unnecessary */
|
|
struct pcpu_group_info groups[];
|
|
};
|
|
|
|
enum pcpu_fc {
|
|
PCPU_FC_AUTO,
|
|
PCPU_FC_EMBED,
|
|
PCPU_FC_PAGE,
|
|
|
|
PCPU_FC_NR,
|
|
};
|
|
extern const char * const pcpu_fc_names[PCPU_FC_NR];
|
|
|
|
extern enum pcpu_fc pcpu_chosen_fc;
|
|
|
|
typedef void * (*pcpu_fc_alloc_fn_t)(unsigned int cpu, size_t size,
|
|
size_t align);
|
|
typedef void (*pcpu_fc_free_fn_t)(void *ptr, size_t size);
|
|
typedef void (*pcpu_fc_populate_pte_fn_t)(unsigned long addr);
|
|
typedef int (pcpu_fc_cpu_distance_fn_t)(unsigned int from, unsigned int to);
|
|
|
|
extern struct pcpu_alloc_info * __init pcpu_alloc_alloc_info(int nr_groups,
|
|
int nr_units);
|
|
extern void __init pcpu_free_alloc_info(struct pcpu_alloc_info *ai);
|
|
|
|
extern int __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
|
|
void *base_addr);
|
|
|
|
#ifdef CONFIG_NEED_PER_CPU_EMBED_FIRST_CHUNK
|
|
extern int __init pcpu_embed_first_chunk(size_t reserved_size, size_t dyn_size,
|
|
size_t atom_size,
|
|
pcpu_fc_cpu_distance_fn_t cpu_distance_fn,
|
|
pcpu_fc_alloc_fn_t alloc_fn,
|
|
pcpu_fc_free_fn_t free_fn);
|
|
#endif
|
|
|
|
#ifdef CONFIG_NEED_PER_CPU_PAGE_FIRST_CHUNK
|
|
extern int __init pcpu_page_first_chunk(size_t reserved_size,
|
|
pcpu_fc_alloc_fn_t alloc_fn,
|
|
pcpu_fc_free_fn_t free_fn,
|
|
pcpu_fc_populate_pte_fn_t populate_pte_fn);
|
|
#endif
|
|
|
|
extern void __percpu *__alloc_reserved_percpu(size_t size, size_t align);
|
|
extern bool __is_kernel_percpu_address(unsigned long addr, unsigned long *can_addr);
|
|
extern bool is_kernel_percpu_address(unsigned long addr);
|
|
|
|
#if !defined(CONFIG_SMP) || !defined(CONFIG_HAVE_SETUP_PER_CPU_AREA)
|
|
extern void __init setup_per_cpu_areas(void);
|
|
#endif
|
|
|
|
extern void __percpu *__alloc_percpu_gfp(size_t size, size_t align, gfp_t gfp);
|
|
extern void __percpu *__alloc_percpu(size_t size, size_t align);
|
|
extern void free_percpu(void __percpu *__pdata);
|
|
extern phys_addr_t per_cpu_ptr_to_phys(void *addr);
|
|
|
|
#define alloc_percpu_gfp(type, gfp) \
|
|
(typeof(type) __percpu *)__alloc_percpu_gfp(sizeof(type), \
|
|
__alignof__(type), gfp)
|
|
#define alloc_percpu(type) \
|
|
(typeof(type) __percpu *)__alloc_percpu(sizeof(type), \
|
|
__alignof__(type))
|
|
|
|
#endif /* __LINUX_PERCPU_H */
|