microsoft
/
WSL2-Linux-Kernel
зеркало из https://github.com/microsoft/WSL2-Linux-Kernel.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

Merge branch 'for-5.14' of git://git.kernel.org/pub/scm/linux/kernel/git/dennis/percpu 

Pull percpu updates from Dennis Zhou:

 - percpu chunk depopulation - depopulate backing pages for chunks with
   empty pages when we exceed a global threshold without those pages.
   This lets us reclaim a portion of memory that would previously be
   lost until the full chunk would be freed (possibly never).

 - memcg accounting cleanup - previously separate chunks were managed
   for normal allocations and __GFP_ACCOUNT allocations. These are now
   consolidated which cleans up the code quite a bit.

 - a few misc clean ups for clang warnings

* 'for-5.14' of git://git.kernel.org/pub/scm/linux/kernel/git/dennis/percpu:
  percpu: optimize locking in pcpu_balance_workfn()
  percpu: initialize best_upa variable
  percpu: rework memcg accounting
  mm, memcg: introduce mem_cgroup_kmem_disabled()
  mm, memcg: mark cgroup_memory_nosocket, nokmem and noswap as __ro_after_init
  percpu: make symbol 'pcpu_free_slot' static
  percpu: implement partial chunk depopulation
  percpu: use pcpu_free_slot instead of pcpu_nr_slots - 1
  percpu: factor out pcpu_check_block_hint()
  percpu: split __pcpu_balance_workfn()
  percpu: fix a comment about the chunks ordering
This commit is contained in:
Linus Torvalds 2021-07-01 17:17:24 -07:00
Родитель 19b4385922 e4d777003a
Коммит e267992f9e
7 изменённых файлов: 343 добавлений и 186 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

5
include/linux/memcontrol.h
Просмотреть файл

@ -1619,6 +1619,7 @@ static inline void set_shrinker_bit(struct mem_cgroup *memcg,
#endif #endif
#ifdef CONFIG_MEMCG_KMEM #ifdef CONFIG_MEMCG_KMEM
bool mem_cgroup_kmem_disabled(void);
int __memcg_kmem_charge_page(struct page *page, gfp_t gfp, int order); int __memcg_kmem_charge_page(struct page *page, gfp_t gfp, int order);
void __memcg_kmem_uncharge_page(struct page *page, int order); void __memcg_kmem_uncharge_page(struct page *page, int order);
@ -1672,6 +1673,10 @@ static inline int memcg_cache_id(struct mem_cgroup *memcg)
struct mem_cgroup *mem_cgroup_from_obj(void *p); struct mem_cgroup *mem_cgroup_from_obj(void *p);
#else #else
static inline bool mem_cgroup_kmem_disabled(void)
{
return true;
}
static inline int memcg_kmem_charge_page(struct page *page, gfp_t gfp, static inline int memcg_kmem_charge_page(struct page *page, gfp_t gfp,
int order) int order)

11
mm/memcontrol.c
Просмотреть файл

@ -81,14 +81,14 @@ DEFINE_PER_CPU(struct mem_cgroup *, int_active_memcg);
EXPORT_PER_CPU_SYMBOL_GPL(int_active_memcg); EXPORT_PER_CPU_SYMBOL_GPL(int_active_memcg);
/* Socket memory accounting disabled? */ /* Socket memory accounting disabled? */
static bool cgroup_memory_nosocket; static bool cgroup_memory_nosocket __ro_after_init;
/* Kernel memory accounting disabled? */ /* Kernel memory accounting disabled? */
bool cgroup_memory_nokmem; bool cgroup_memory_nokmem __ro_after_init;
/* Whether the swap controller is active */ /* Whether the swap controller is active */
#ifdef CONFIG_MEMCG_SWAP #ifdef CONFIG_MEMCG_SWAP
bool cgroup_memory_noswap __read_mostly; bool cgroup_memory_noswap __ro_after_init;
#else #else
#define cgroup_memory_noswap 1 #define cgroup_memory_noswap 1
#endif #endif
@ -256,6 +256,11 @@ struct cgroup_subsys_state *vmpressure_to_css(struct vmpressure *vmpr)
#ifdef CONFIG_MEMCG_KMEM #ifdef CONFIG_MEMCG_KMEM
extern spinlock_t css_set_lock; extern spinlock_t css_set_lock;
bool mem_cgroup_kmem_disabled(void)
{
return cgroup_memory_nokmem;
}
static void obj_cgroup_uncharge_pages(struct obj_cgroup *objcg, static void obj_cgroup_uncharge_pages(struct obj_cgroup *objcg,
unsigned int nr_pages); unsigned int nr_pages);

56
mm/percpu-internal.h
Просмотреть файл

@ -5,25 +5,6 @@
#include <linux/types.h> #include <linux/types.h>
#include <linux/percpu.h> #include <linux/percpu.h>
/*
* There are two chunk types: root and memcg-aware.
* Chunks of each type have separate slots list.
*
* Memcg-aware chunks have an attached vector of obj_cgroup pointers, which is
* used to store memcg membership data of a percpu object. Obj_cgroups are
* ref-counted pointers to a memory cgroup with an ability to switch dynamically
* to the parent memory cgroup. This allows to reclaim a deleted memory cgroup
* without reclaiming of all outstanding objects, which hold a reference at it.
*/
enum pcpu_chunk_type {
PCPU_CHUNK_ROOT,
#ifdef CONFIG_MEMCG_KMEM
PCPU_CHUNK_MEMCG,
#endif
PCPU_NR_CHUNK_TYPES,
PCPU_FAIL_ALLOC = PCPU_NR_CHUNK_TYPES
};
/* /*
* pcpu_block_md is the metadata block struct. * pcpu_block_md is the metadata block struct.
* Each chunk's bitmap is split into a number of full blocks. * Each chunk's bitmap is split into a number of full blocks.
@ -67,6 +48,8 @@ struct pcpu_chunk {
void *data; /* chunk data */ void *data; /* chunk data */
bool immutable; /* no [de]population allowed */ bool immutable; /* no [de]population allowed */
bool isolated; /* isolated from active chunk
slots */
int start_offset; /* the overlap with the previous int start_offset; /* the overlap with the previous
region to have a page aligned region to have a page aligned
base_addr */ base_addr */
@ -87,7 +70,9 @@ extern spinlock_t pcpu_lock;
extern struct list_head *pcpu_chunk_lists; extern struct list_head *pcpu_chunk_lists;
extern int pcpu_nr_slots; extern int pcpu_nr_slots;
extern int pcpu_nr_empty_pop_pages[]; extern int pcpu_sidelined_slot;
extern int pcpu_to_depopulate_slot;
extern int pcpu_nr_empty_pop_pages;
extern struct pcpu_chunk *pcpu_first_chunk; extern struct pcpu_chunk *pcpu_first_chunk;
extern struct pcpu_chunk *pcpu_reserved_chunk; extern struct pcpu_chunk *pcpu_reserved_chunk;
@ -128,37 +113,6 @@ static inline int pcpu_chunk_map_bits(struct pcpu_chunk *chunk)
return pcpu_nr_pages_to_map_bits(chunk->nr_pages); return pcpu_nr_pages_to_map_bits(chunk->nr_pages);
} }
#ifdef CONFIG_MEMCG_KMEM
static inline enum pcpu_chunk_type pcpu_chunk_type(struct pcpu_chunk *chunk)
{
if (chunk->obj_cgroups)
return PCPU_CHUNK_MEMCG;
return PCPU_CHUNK_ROOT;
}
static inline bool pcpu_is_memcg_chunk(enum pcpu_chunk_type chunk_type)
{
return chunk_type == PCPU_CHUNK_MEMCG;
}
#else
static inline enum pcpu_chunk_type pcpu_chunk_type(struct pcpu_chunk *chunk)
{
return PCPU_CHUNK_ROOT;
}
static inline bool pcpu_is_memcg_chunk(enum pcpu_chunk_type chunk_type)
{
return false;
}
#endif
static inline struct list_head *pcpu_chunk_list(enum pcpu_chunk_type chunk_type)
{
return &pcpu_chunk_lists[pcpu_nr_slots *
pcpu_is_memcg_chunk(chunk_type)];
}
#ifdef CONFIG_PERCPU_STATS #ifdef CONFIG_PERCPU_STATS
#include <linux/spinlock.h> #include <linux/spinlock.h>

10
mm/percpu-km.c
Просмотреть файл

@ -44,8 +44,7 @@ static void pcpu_depopulate_chunk(struct pcpu_chunk *chunk,
/* nada */ /* nada */
} }
static struct pcpu_chunk *pcpu_create_chunk(enum pcpu_chunk_type type, static struct pcpu_chunk *pcpu_create_chunk(gfp_t gfp)
gfp_t gfp)
{ {
const int nr_pages = pcpu_group_sizes[0] >> PAGE_SHIFT; const int nr_pages = pcpu_group_sizes[0] >> PAGE_SHIFT;
struct pcpu_chunk *chunk; struct pcpu_chunk *chunk;
@ -53,7 +52,7 @@ static struct pcpu_chunk *pcpu_create_chunk(enum pcpu_chunk_type type,
unsigned long flags; unsigned long flags;
int i; int i;
chunk = pcpu_alloc_chunk(type, gfp); chunk = pcpu_alloc_chunk(gfp);
if (!chunk) if (!chunk)
return NULL; return NULL;
@ -118,3 +117,8 @@ static int __init pcpu_verify_alloc_info(const struct pcpu_alloc_info *ai)
return 0; return 0;
} }
static bool pcpu_should_reclaim_chunk(struct pcpu_chunk *chunk)
{
return false;
}

44
mm/percpu-stats.c
Просмотреть файл

@ -34,15 +34,11 @@ static int find_max_nr_alloc(void)
{ {
struct pcpu_chunk *chunk; struct pcpu_chunk *chunk;
int slot, max_nr_alloc; int slot, max_nr_alloc;
enum pcpu_chunk_type type;
max_nr_alloc = 0; max_nr_alloc = 0;
for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++) for (slot = 0; slot < pcpu_nr_slots; slot++)
for (slot = 0; slot < pcpu_nr_slots; slot++) list_for_each_entry(chunk, &pcpu_chunk_lists[slot], list)
list_for_each_entry(chunk, &pcpu_chunk_list(type)[slot], max_nr_alloc = max(max_nr_alloc, chunk->nr_alloc);
list)
max_nr_alloc = max(max_nr_alloc,
chunk->nr_alloc);
return max_nr_alloc; return max_nr_alloc;
} }
@ -133,9 +129,6 @@ static void chunk_map_stats(struct seq_file *m, struct pcpu_chunk *chunk,
P("cur_min_alloc", cur_min_alloc); P("cur_min_alloc", cur_min_alloc);
P("cur_med_alloc", cur_med_alloc); P("cur_med_alloc", cur_med_alloc);
P("cur_max_alloc", cur_max_alloc); P("cur_max_alloc", cur_max_alloc);
#ifdef CONFIG_MEMCG_KMEM
P("memcg_aware", pcpu_is_memcg_chunk(pcpu_chunk_type(chunk)));
#endif
seq_putc(m, '\n'); seq_putc(m, '\n');
} }
@ -144,8 +137,6 @@ static int percpu_stats_show(struct seq_file *m, void *v)
struct pcpu_chunk *chunk; struct pcpu_chunk *chunk;
int slot, max_nr_alloc; int slot, max_nr_alloc;
int *buffer; int *buffer;
enum pcpu_chunk_type type;
int nr_empty_pop_pages;
alloc_buffer: alloc_buffer:
spin_lock_irq(&pcpu_lock); spin_lock_irq(&pcpu_lock);
@ -166,10 +157,6 @@ alloc_buffer:
goto alloc_buffer; goto alloc_buffer;
} }
nr_empty_pop_pages = 0;
for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++)
nr_empty_pop_pages += pcpu_nr_empty_pop_pages[type];
#define PL(X) \ #define PL(X) \
seq_printf(m, " %-20s: %12lld\n", #X, (long long int)pcpu_stats_ai.X) seq_printf(m, " %-20s: %12lld\n", #X, (long long int)pcpu_stats_ai.X)
@ -201,7 +188,7 @@ alloc_buffer:
PU(nr_max_chunks); PU(nr_max_chunks);
PU(min_alloc_size); PU(min_alloc_size);
PU(max_alloc_size); PU(max_alloc_size);
P("empty_pop_pages", nr_empty_pop_pages); P("empty_pop_pages", pcpu_nr_empty_pop_pages);
seq_putc(m, '\n'); seq_putc(m, '\n');
#undef PU #undef PU
@ -215,18 +202,17 @@ alloc_buffer:
chunk_map_stats(m, pcpu_reserved_chunk, buffer); chunk_map_stats(m, pcpu_reserved_chunk, buffer);
} }
for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++) { for (slot = 0; slot < pcpu_nr_slots; slot++) {
for (slot = 0; slot < pcpu_nr_slots; slot++) { list_for_each_entry(chunk, &pcpu_chunk_lists[slot], list) {
list_for_each_entry(chunk, &pcpu_chunk_list(type)[slot], if (chunk == pcpu_first_chunk)
list) { seq_puts(m, "Chunk: <- First Chunk\n");
if (chunk == pcpu_first_chunk) { else if (slot == pcpu_to_depopulate_slot)
seq_puts(m, "Chunk: <- First Chunk\n"); seq_puts(m, "Chunk (to_depopulate)\n");
chunk_map_stats(m, chunk, buffer); else if (slot == pcpu_sidelined_slot)
} else { seq_puts(m, "Chunk (sidelined):\n");
seq_puts(m, "Chunk:\n"); else
chunk_map_stats(m, chunk, buffer); seq_puts(m, "Chunk:\n");
} chunk_map_stats(m, chunk, buffer);
}
} }
} }

35
mm/percpu-vm.c
Просмотреть файл

@ -329,13 +329,12 @@ static void pcpu_depopulate_chunk(struct pcpu_chunk *chunk,
pcpu_free_pages(chunk, pages, page_start, page_end); pcpu_free_pages(chunk, pages, page_start, page_end);
} }
static struct pcpu_chunk *pcpu_create_chunk(enum pcpu_chunk_type type, static struct pcpu_chunk *pcpu_create_chunk(gfp_t gfp)
gfp_t gfp)
{ {
struct pcpu_chunk *chunk; struct pcpu_chunk *chunk;
struct vm_struct **vms; struct vm_struct **vms;
chunk = pcpu_alloc_chunk(type, gfp); chunk = pcpu_alloc_chunk(gfp);
if (!chunk) if (!chunk)
return NULL; return NULL;
@ -378,3 +377,33 @@ static int __init pcpu_verify_alloc_info(const struct pcpu_alloc_info *ai)
/* no extra restriction */ /* no extra restriction */
return 0; return 0;
} }
/**
* pcpu_should_reclaim_chunk - determine if a chunk should go into reclaim
* @chunk: chunk of interest
*
* This is the entry point for percpu reclaim. If a chunk qualifies, it is then
* isolated and managed in separate lists at the back of pcpu_slot: sidelined
* and to_depopulate respectively. The to_depopulate list holds chunks slated
* for depopulation. They no longer contribute to pcpu_nr_empty_pop_pages once
* they are on this list. Once depopulated, they are moved onto the sidelined
* list which enables them to be pulled back in for allocation if no other chunk
* can suffice the allocation.
*/
static bool pcpu_should_reclaim_chunk(struct pcpu_chunk *chunk)
{
/* do not reclaim either the first chunk or reserved chunk */
if (chunk == pcpu_first_chunk || chunk == pcpu_reserved_chunk)
return false;
/*
* If it is isolated, it may be on the sidelined list so move it back to
* the to_depopulate list. If we hit at least 1/4 pages empty pages AND
* there is no system-wide shortage of empty pages aside from this
* chunk, move it to the to_depopulate list.
*/
return ((chunk->isolated && chunk->nr_empty_pop_pages) ||
(pcpu_nr_empty_pop_pages >
(PCPU_EMPTY_POP_PAGES_HIGH + chunk->nr_empty_pop_pages) &&
chunk->nr_empty_pop_pages >= chunk->nr_pages / 4));
}

368
mm/percpu.c
Просмотреть файл

@ -99,7 +99,10 @@
#include "percpu-internal.h" #include "percpu-internal.h"
/* the slots are sorted by free bytes left, 1-31 bytes share the same slot */ /*
* The slots are sorted by the size of the biggest continuous free area.
* 1-31 bytes share the same slot.
*/
#define PCPU_SLOT_BASE_SHIFT 5 #define PCPU_SLOT_BASE_SHIFT 5
/* chunks in slots below this are subject to being sidelined on failed alloc */ /* chunks in slots below this are subject to being sidelined on failed alloc */
#define PCPU_SLOT_FAIL_THRESHOLD 3 #define PCPU_SLOT_FAIL_THRESHOLD 3
@ -132,6 +135,9 @@ static int pcpu_unit_size __ro_after_init;
static int pcpu_nr_units __ro_after_init; static int pcpu_nr_units __ro_after_init;
static int pcpu_atom_size __ro_after_init; static int pcpu_atom_size __ro_after_init;
int pcpu_nr_slots __ro_after_init; int pcpu_nr_slots __ro_after_init;
static int pcpu_free_slot __ro_after_init;
int pcpu_sidelined_slot __ro_after_init;
int pcpu_to_depopulate_slot __ro_after_init;
static size_t pcpu_chunk_struct_size __ro_after_init; static size_t pcpu_chunk_struct_size __ro_after_init;
/* cpus with the lowest and highest unit addresses */ /* cpus with the lowest and highest unit addresses */
@ -173,10 +179,10 @@ struct list_head *pcpu_chunk_lists __ro_after_init; /* chunk list slots */
static LIST_HEAD(pcpu_map_extend_chunks); static LIST_HEAD(pcpu_map_extend_chunks);
/* /*
* The number of empty populated pages by chunk type, protected by pcpu_lock. * The number of empty populated pages, protected by pcpu_lock.
* The reserved chunk doesn't contribute to the count. * The reserved chunk doesn't contribute to the count.
*/ */
int pcpu_nr_empty_pop_pages[PCPU_NR_CHUNK_TYPES]; int pcpu_nr_empty_pop_pages;
/* /*
* The number of populated pages in use by the allocator, protected by * The number of populated pages in use by the allocator, protected by
@ -234,7 +240,7 @@ static int __pcpu_size_to_slot(int size)
static int pcpu_size_to_slot(int size) static int pcpu_size_to_slot(int size)
{ {
if (size == pcpu_unit_size) if (size == pcpu_unit_size)
return pcpu_nr_slots - 1; return pcpu_free_slot;
return __pcpu_size_to_slot(size); return __pcpu_size_to_slot(size);
} }
@ -303,6 +309,25 @@ static unsigned long pcpu_block_off_to_off(int index, int off)
return index * PCPU_BITMAP_BLOCK_BITS + off; return index * PCPU_BITMAP_BLOCK_BITS + off;
} }
/**
* pcpu_check_block_hint - check against the contig hint
* @block: block of interest
* @bits: size of allocation
* @align: alignment of area (max PAGE_SIZE)
*
* Check to see if the allocation can fit in the block's contig hint.
* Note, a chunk uses the same hints as a block so this can also check against
* the chunk's contig hint.
*/
static bool pcpu_check_block_hint(struct pcpu_block_md *block, int bits,
size_t align)
{
int bit_off = ALIGN(block->contig_hint_start, align) -
block->contig_hint_start;
return bit_off + bits <= block->contig_hint;
}
/* /*
* pcpu_next_hint - determine which hint to use * pcpu_next_hint - determine which hint to use
* @block: block of interest * @block: block of interest
@ -507,13 +532,10 @@ static void __pcpu_chunk_move(struct pcpu_chunk *chunk, int slot,
bool move_front) bool move_front)
{ {
if (chunk != pcpu_reserved_chunk) { if (chunk != pcpu_reserved_chunk) {
struct list_head *pcpu_slot;
pcpu_slot = pcpu_chunk_list(pcpu_chunk_type(chunk));
if (move_front) if (move_front)
list_move(&chunk->list, &pcpu_slot[slot]); list_move(&chunk->list, &pcpu_chunk_lists[slot]);
else else
list_move_tail(&chunk->list, &pcpu_slot[slot]); list_move_tail(&chunk->list, &pcpu_chunk_lists[slot]);
} }
} }
@ -539,10 +561,36 @@ static void pcpu_chunk_relocate(struct pcpu_chunk *chunk, int oslot)
{ {
int nslot = pcpu_chunk_slot(chunk); int nslot = pcpu_chunk_slot(chunk);
/* leave isolated chunks in-place */
if (chunk->isolated)
return;
if (oslot != nslot) if (oslot != nslot)
__pcpu_chunk_move(chunk, nslot, oslot < nslot); __pcpu_chunk_move(chunk, nslot, oslot < nslot);
} }
static void pcpu_isolate_chunk(struct pcpu_chunk *chunk)
{
lockdep_assert_held(&pcpu_lock);
if (!chunk->isolated) {
chunk->isolated = true;
pcpu_nr_empty_pop_pages -= chunk->nr_empty_pop_pages;
}
list_move(&chunk->list, &pcpu_chunk_lists[pcpu_to_depopulate_slot]);
}
static void pcpu_reintegrate_chunk(struct pcpu_chunk *chunk)
{
lockdep_assert_held(&pcpu_lock);
if (chunk->isolated) {
chunk->isolated = false;
pcpu_nr_empty_pop_pages += chunk->nr_empty_pop_pages;
pcpu_chunk_relocate(chunk, -1);
}
}
/* /*
* pcpu_update_empty_pages - update empty page counters * pcpu_update_empty_pages - update empty page counters
* @chunk: chunk of interest * @chunk: chunk of interest
@ -555,8 +603,8 @@ static void pcpu_chunk_relocate(struct pcpu_chunk *chunk, int oslot)
static inline void pcpu_update_empty_pages(struct pcpu_chunk *chunk, int nr) static inline void pcpu_update_empty_pages(struct pcpu_chunk *chunk, int nr)
{ {
chunk->nr_empty_pop_pages += nr; chunk->nr_empty_pop_pages += nr;
if (chunk != pcpu_reserved_chunk) if (chunk != pcpu_reserved_chunk && !chunk->isolated)
pcpu_nr_empty_pop_pages[pcpu_chunk_type(chunk)] += nr; pcpu_nr_empty_pop_pages += nr;
} }
/* /*
@ -1063,14 +1111,11 @@ static int pcpu_find_block_fit(struct pcpu_chunk *chunk, int alloc_bits,
int bit_off, bits, next_off; int bit_off, bits, next_off;
/* /*
* Check to see if the allocation can fit in the chunk's contig hint. * This is an optimization to prevent scanning by assuming if the
* This is an optimization to prevent scanning by assuming if it * allocation cannot fit in the global hint, there is memory pressure
* cannot fit in the global hint, there is memory pressure and creating * and creating a new chunk would happen soon.
* a new chunk would happen soon.
*/ */
bit_off = ALIGN(chunk_md->contig_hint_start, align) - if (!pcpu_check_block_hint(chunk_md, alloc_bits, align))
chunk_md->contig_hint_start;
if (bit_off + alloc_bits > chunk_md->contig_hint)
return -1; return -1;
bit_off = pcpu_next_hint(chunk_md, alloc_bits); bit_off = pcpu_next_hint(chunk_md, alloc_bits);
@ -1352,7 +1397,7 @@ static struct pcpu_chunk * __init pcpu_alloc_first_chunk(unsigned long tmp_addr,
alloc_size); alloc_size);
#ifdef CONFIG_MEMCG_KMEM #ifdef CONFIG_MEMCG_KMEM
/* first chunk isn't memcg-aware */ /* first chunk is free to use */
chunk->obj_cgroups = NULL; chunk->obj_cgroups = NULL;
#endif #endif
pcpu_init_md_blocks(chunk); pcpu_init_md_blocks(chunk);
@ -1394,7 +1439,7 @@ static struct pcpu_chunk * __init pcpu_alloc_first_chunk(unsigned long tmp_addr,
return chunk; return chunk;
} }
static struct pcpu_chunk *pcpu_alloc_chunk(enum pcpu_chunk_type type, gfp_t gfp) static struct pcpu_chunk *pcpu_alloc_chunk(gfp_t gfp)
{ {
struct pcpu_chunk *chunk; struct pcpu_chunk *chunk;
int region_bits; int region_bits;
@ -1423,7 +1468,7 @@ static struct pcpu_chunk *pcpu_alloc_chunk(enum pcpu_chunk_type type, gfp_t gfp)
goto md_blocks_fail; goto md_blocks_fail;
#ifdef CONFIG_MEMCG_KMEM #ifdef CONFIG_MEMCG_KMEM
if (pcpu_is_memcg_chunk(type)) { if (!mem_cgroup_kmem_disabled()) {
chunk->obj_cgroups = chunk->obj_cgroups =
pcpu_mem_zalloc(pcpu_chunk_map_bits(chunk) * pcpu_mem_zalloc(pcpu_chunk_map_bits(chunk) *
sizeof(struct obj_cgroup *), gfp); sizeof(struct obj_cgroup *), gfp);
@ -1536,8 +1581,7 @@ static int pcpu_populate_chunk(struct pcpu_chunk *chunk,
int page_start, int page_end, gfp_t gfp); int page_start, int page_end, gfp_t gfp);
static void pcpu_depopulate_chunk(struct pcpu_chunk *chunk, static void pcpu_depopulate_chunk(struct pcpu_chunk *chunk,
int page_start, int page_end); int page_start, int page_end);
static struct pcpu_chunk *pcpu_create_chunk(enum pcpu_chunk_type type, static struct pcpu_chunk *pcpu_create_chunk(gfp_t gfp);
gfp_t gfp);
static void pcpu_destroy_chunk(struct pcpu_chunk *chunk); static void pcpu_destroy_chunk(struct pcpu_chunk *chunk);
static struct page *pcpu_addr_to_page(void *addr); static struct page *pcpu_addr_to_page(void *addr);
static int __init pcpu_verify_alloc_info(const struct pcpu_alloc_info *ai); static int __init pcpu_verify_alloc_info(const struct pcpu_alloc_info *ai);
@ -1580,25 +1624,25 @@ static struct pcpu_chunk *pcpu_chunk_addr_search(void *addr)
} }
#ifdef CONFIG_MEMCG_KMEM #ifdef CONFIG_MEMCG_KMEM
static enum pcpu_chunk_type pcpu_memcg_pre_alloc_hook(size_t size, gfp_t gfp, static bool pcpu_memcg_pre_alloc_hook(size_t size, gfp_t gfp,
struct obj_cgroup **objcgp) struct obj_cgroup **objcgp)
{ {
struct obj_cgroup *objcg; struct obj_cgroup *objcg;
if (!memcg_kmem_enabled() || !(gfp & __GFP_ACCOUNT)) if (!memcg_kmem_enabled() || !(gfp & __GFP_ACCOUNT))
return PCPU_CHUNK_ROOT; return true;
objcg = get_obj_cgroup_from_current(); objcg = get_obj_cgroup_from_current();
if (!objcg) if (!objcg)
return PCPU_CHUNK_ROOT; return true;
if (obj_cgroup_charge(objcg, gfp, size * num_possible_cpus())) { if (obj_cgroup_charge(objcg, gfp, size * num_possible_cpus())) {
obj_cgroup_put(objcg); obj_cgroup_put(objcg);
return PCPU_FAIL_ALLOC; return false;
} }
*objcgp = objcg; *objcgp = objcg;
return PCPU_CHUNK_MEMCG; return true;
} }
static void pcpu_memcg_post_alloc_hook(struct obj_cgroup *objcg, static void pcpu_memcg_post_alloc_hook(struct obj_cgroup *objcg,
@ -1608,7 +1652,7 @@ static void pcpu_memcg_post_alloc_hook(struct obj_cgroup *objcg,
if (!objcg) if (!objcg)
return; return;
if (chunk) { if (likely(chunk && chunk->obj_cgroups)) {
chunk->obj_cgroups[off >> PCPU_MIN_ALLOC_SHIFT] = objcg; chunk->obj_cgroups[off >> PCPU_MIN_ALLOC_SHIFT] = objcg;
rcu_read_lock(); rcu_read_lock();
@ -1625,10 +1669,12 @@ static void pcpu_memcg_free_hook(struct pcpu_chunk *chunk, int off, size_t size)
{ {
struct obj_cgroup *objcg; struct obj_cgroup *objcg;
if (!pcpu_is_memcg_chunk(pcpu_chunk_type(chunk))) if (unlikely(!chunk->obj_cgroups))
return; return;
objcg = chunk->obj_cgroups[off >> PCPU_MIN_ALLOC_SHIFT]; objcg = chunk->obj_cgroups[off >> PCPU_MIN_ALLOC_SHIFT];
if (!objcg)
return;
chunk->obj_cgroups[off >> PCPU_MIN_ALLOC_SHIFT] = NULL; chunk->obj_cgroups[off >> PCPU_MIN_ALLOC_SHIFT] = NULL;
obj_cgroup_uncharge(objcg, size * num_possible_cpus()); obj_cgroup_uncharge(objcg, size * num_possible_cpus());
@ -1642,10 +1688,10 @@ static void pcpu_memcg_free_hook(struct pcpu_chunk *chunk, int off, size_t size)
} }
#else /* CONFIG_MEMCG_KMEM */ #else /* CONFIG_MEMCG_KMEM */
static enum pcpu_chunk_type static bool
pcpu_memcg_pre_alloc_hook(size_t size, gfp_t gfp, struct obj_cgroup **objcgp) pcpu_memcg_pre_alloc_hook(size_t size, gfp_t gfp, struct obj_cgroup **objcgp)
{ {
return PCPU_CHUNK_ROOT; return true;
} }
static void pcpu_memcg_post_alloc_hook(struct obj_cgroup *objcg, static void pcpu_memcg_post_alloc_hook(struct obj_cgroup *objcg,
@ -1680,8 +1726,6 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
gfp_t pcpu_gfp; gfp_t pcpu_gfp;
bool is_atomic; bool is_atomic;
bool do_warn; bool do_warn;
enum pcpu_chunk_type type;
struct list_head *pcpu_slot;
struct obj_cgroup *objcg = NULL; struct obj_cgroup *objcg = NULL;
static int warn_limit = 10; static int warn_limit = 10;
struct pcpu_chunk *chunk, *next; struct pcpu_chunk *chunk, *next;
@ -1717,10 +1761,8 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
return NULL; return NULL;
} }
type = pcpu_memcg_pre_alloc_hook(size, gfp, &objcg); if (unlikely(!pcpu_memcg_pre_alloc_hook(size, gfp, &objcg)))
if (unlikely(type == PCPU_FAIL_ALLOC))
return NULL; return NULL;
pcpu_slot = pcpu_chunk_list(type);
if (!is_atomic) { if (!is_atomic) {
/* /*
@ -1758,8 +1800,9 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
restart: restart:
/* search through normal chunks */ /* search through normal chunks */
for (slot = pcpu_size_to_slot(size); slot < pcpu_nr_slots; slot++) { for (slot = pcpu_size_to_slot(size); slot <= pcpu_free_slot; slot++) {
list_for_each_entry_safe(chunk, next, &pcpu_slot[slot], list) { list_for_each_entry_safe(chunk, next, &pcpu_chunk_lists[slot],
list) {
off = pcpu_find_block_fit(chunk, bits, bit_align, off = pcpu_find_block_fit(chunk, bits, bit_align,
is_atomic); is_atomic);
if (off < 0) { if (off < 0) {
@ -1769,9 +1812,10 @@ restart:
} }
off = pcpu_alloc_area(chunk, bits, bit_align, off); off = pcpu_alloc_area(chunk, bits, bit_align, off);
if (off >= 0) if (off >= 0) {
pcpu_reintegrate_chunk(chunk);
goto area_found; goto area_found;
}
} }
} }
@ -1787,8 +1831,8 @@ restart:
goto fail; goto fail;
} }
if (list_empty(&pcpu_slot[pcpu_nr_slots - 1])) { if (list_empty(&pcpu_chunk_lists[pcpu_free_slot])) {
chunk = pcpu_create_chunk(type, pcpu_gfp); chunk = pcpu_create_chunk(pcpu_gfp);
if (!chunk) { if (!chunk) {
err = "failed to allocate new chunk"; err = "failed to allocate new chunk";
goto fail; goto fail;
@ -1832,7 +1876,7 @@ area_found:
mutex_unlock(&pcpu_alloc_mutex); mutex_unlock(&pcpu_alloc_mutex);
} }
if (pcpu_nr_empty_pop_pages[type] < PCPU_EMPTY_POP_PAGES_LOW) if (pcpu_nr_empty_pop_pages < PCPU_EMPTY_POP_PAGES_LOW)
pcpu_schedule_balance_work(); pcpu_schedule_balance_work();
/* clear the areas and return address relative to base address */ /* clear the areas and return address relative to base address */
@ -1930,33 +1974,28 @@ void __percpu *__alloc_reserved_percpu(size_t size, size_t align)
} }
/** /**
* __pcpu_balance_workfn - manage the amount of free chunks and populated pages * pcpu_balance_free - manage the amount of free chunks
* @type: chunk type * @empty_only: free chunks only if there are no populated pages
* *
* Reclaim all fully free chunks except for the first one. This is also * If empty_only is %false, reclaim all fully free chunks regardless of the
* responsible for maintaining the pool of empty populated pages. However, * number of populated pages. Otherwise, only reclaim chunks that have no
* it is possible that this is called when physical memory is scarce causing * populated pages.
* OOM killer to be triggered. We should avoid doing so until an actual *
* allocation causes the failure as it is possible that requests can be * CONTEXT:
* serviced from already backed regions. * pcpu_lock (can be dropped temporarily)
*/ */
static void __pcpu_balance_workfn(enum pcpu_chunk_type type) static void pcpu_balance_free(bool empty_only)
{ {
/* gfp flags passed to underlying allocators */
const gfp_t gfp = GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN;
LIST_HEAD(to_free); LIST_HEAD(to_free);
struct list_head *pcpu_slot = pcpu_chunk_list(type); struct list_head *free_head = &pcpu_chunk_lists[pcpu_free_slot];
struct list_head *free_head = &pcpu_slot[pcpu_nr_slots - 1];
struct pcpu_chunk *chunk, *next; struct pcpu_chunk *chunk, *next;
int slot, nr_to_pop, ret;
lockdep_assert_held(&pcpu_lock);
/* /*
* There's no reason to keep around multiple unused chunks and VM * There's no reason to keep around multiple unused chunks and VM
* areas can be scarce. Destroy all free chunks except for one. * areas can be scarce. Destroy all free chunks except for one.
*/ */
mutex_lock(&pcpu_alloc_mutex);
spin_lock_irq(&pcpu_lock);
list_for_each_entry_safe(chunk, next, free_head, list) { list_for_each_entry_safe(chunk, next, free_head, list) {
WARN_ON(chunk->immutable); WARN_ON(chunk->immutable);
@ -1964,11 +2003,14 @@ static void __pcpu_balance_workfn(enum pcpu_chunk_type type)
if (chunk == list_first_entry(free_head, struct pcpu_chunk, list)) if (chunk == list_first_entry(free_head, struct pcpu_chunk, list))
continue; continue;
list_move(&chunk->list, &to_free); if (!empty_only || chunk->nr_empty_pop_pages == 0)
list_move(&chunk->list, &to_free);
} }
spin_unlock_irq(&pcpu_lock); if (list_empty(&to_free))
return;
spin_unlock_irq(&pcpu_lock);
list_for_each_entry_safe(chunk, next, &to_free, list) { list_for_each_entry_safe(chunk, next, &to_free, list) {
unsigned int rs, re; unsigned int rs, re;
@ -1982,6 +2024,29 @@ static void __pcpu_balance_workfn(enum pcpu_chunk_type type)
pcpu_destroy_chunk(chunk); pcpu_destroy_chunk(chunk);
cond_resched(); cond_resched();
} }
spin_lock_irq(&pcpu_lock);
}
/**
* pcpu_balance_populated - manage the amount of populated pages
*
* Maintain a certain amount of populated pages to satisfy atomic allocations.
* It is possible that this is called when physical memory is scarce causing
* OOM killer to be triggered. We should avoid doing so until an actual
* allocation causes the failure as it is possible that requests can be
* serviced from already backed regions.
*
* CONTEXT:
* pcpu_lock (can be dropped temporarily)
*/
static void pcpu_balance_populated(void)
{
/* gfp flags passed to underlying allocators */
const gfp_t gfp = GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN;
struct pcpu_chunk *chunk;
int slot, nr_to_pop, ret;
lockdep_assert_held(&pcpu_lock);
/* /*
* Ensure there are certain number of free populated pages for * Ensure there are certain number of free populated pages for
@ -2000,23 +2065,21 @@ retry_pop:
pcpu_atomic_alloc_failed = false; pcpu_atomic_alloc_failed = false;
} else { } else {
nr_to_pop = clamp(PCPU_EMPTY_POP_PAGES_HIGH - nr_to_pop = clamp(PCPU_EMPTY_POP_PAGES_HIGH -
pcpu_nr_empty_pop_pages[type], pcpu_nr_empty_pop_pages,
0, PCPU_EMPTY_POP_PAGES_HIGH); 0, PCPU_EMPTY_POP_PAGES_HIGH);
} }
for (slot = pcpu_size_to_slot(PAGE_SIZE); slot < pcpu_nr_slots; slot++) { for (slot = pcpu_size_to_slot(PAGE_SIZE); slot <= pcpu_free_slot; slot++) {
unsigned int nr_unpop = 0, rs, re; unsigned int nr_unpop = 0, rs, re;
if (!nr_to_pop) if (!nr_to_pop)
break; break;
spin_lock_irq(&pcpu_lock); list_for_each_entry(chunk, &pcpu_chunk_lists[slot], list) {
list_for_each_entry(chunk, &pcpu_slot[slot], list) {
nr_unpop = chunk->nr_pages - chunk->nr_populated; nr_unpop = chunk->nr_pages - chunk->nr_populated;
if (nr_unpop) if (nr_unpop)
break; break;
} }
spin_unlock_irq(&pcpu_lock);
if (!nr_unpop) if (!nr_unpop)
continue; continue;
@ -2026,12 +2089,13 @@ retry_pop:
chunk->nr_pages) { chunk->nr_pages) {
int nr = min_t(int, re - rs, nr_to_pop); int nr = min_t(int, re - rs, nr_to_pop);
spin_unlock_irq(&pcpu_lock);
ret = pcpu_populate_chunk(chunk, rs, rs + nr, gfp); ret = pcpu_populate_chunk(chunk, rs, rs + nr, gfp);
cond_resched();
spin_lock_irq(&pcpu_lock);
if (!ret) { if (!ret) {
nr_to_pop -= nr; nr_to_pop -= nr;
spin_lock_irq(&pcpu_lock);
pcpu_chunk_populated(chunk, rs, rs + nr); pcpu_chunk_populated(chunk, rs, rs + nr);
spin_unlock_irq(&pcpu_lock);
} else { } else {
nr_to_pop = 0; nr_to_pop = 0;
} }
@ -2043,30 +2107,133 @@ retry_pop:
if (nr_to_pop) { if (nr_to_pop) {
/* ran out of chunks to populate, create a new one and retry */ /* ran out of chunks to populate, create a new one and retry */
chunk = pcpu_create_chunk(type, gfp); spin_unlock_irq(&pcpu_lock);
chunk = pcpu_create_chunk(gfp);
cond_resched();
spin_lock_irq(&pcpu_lock);
if (chunk) { if (chunk) {
spin_lock_irq(&pcpu_lock);
pcpu_chunk_relocate(chunk, -1); pcpu_chunk_relocate(chunk, -1);
spin_unlock_irq(&pcpu_lock);
goto retry_pop; goto retry_pop;
} }
} }
}
mutex_unlock(&pcpu_alloc_mutex); /**
* pcpu_reclaim_populated - scan over to_depopulate chunks and free empty pages
*
* Scan over chunks in the depopulate list and try to release unused populated
* pages back to the system. Depopulated chunks are sidelined to prevent
* repopulating these pages unless required. Fully free chunks are reintegrated
* and freed accordingly (1 is kept around). If we drop below the empty
* populated pages threshold, reintegrate the chunk if it has empty free pages.
* Each chunk is scanned in the reverse order to keep populated pages close to
* the beginning of the chunk.
*
* CONTEXT:
* pcpu_lock (can be dropped temporarily)
*
*/
static void pcpu_reclaim_populated(void)
{
struct pcpu_chunk *chunk;
struct pcpu_block_md *block;
int i, end;
lockdep_assert_held(&pcpu_lock);
restart:
/*
* Once a chunk is isolated to the to_depopulate list, the chunk is no
* longer discoverable to allocations whom may populate pages. The only
* other accessor is the free path which only returns area back to the
* allocator not touching the populated bitmap.
*/
while (!list_empty(&pcpu_chunk_lists[pcpu_to_depopulate_slot])) {
chunk = list_first_entry(&pcpu_chunk_lists[pcpu_to_depopulate_slot],
struct pcpu_chunk, list);
WARN_ON(chunk->immutable);
/*
* Scan chunk's pages in the reverse order to keep populated
* pages close to the beginning of the chunk.
*/
for (i = chunk->nr_pages - 1, end = -1; i >= 0; i--) {
/* no more work to do */
if (chunk->nr_empty_pop_pages == 0)
break;
/* reintegrate chunk to prevent atomic alloc failures */
if (pcpu_nr_empty_pop_pages < PCPU_EMPTY_POP_PAGES_HIGH) {
pcpu_reintegrate_chunk(chunk);
goto restart;
}
/*
* If the page is empty and populated, start or
* extend the (i, end) range. If i == 0, decrease
* i and perform the depopulation to cover the last
* (first) page in the chunk.
*/
block = chunk->md_blocks + i;
if (block->contig_hint == PCPU_BITMAP_BLOCK_BITS &&
test_bit(i, chunk->populated)) {
if (end == -1)
end = i;
if (i > 0)
continue;
i--;
}
/* depopulate if there is an active range */
if (end == -1)
continue;
spin_unlock_irq(&pcpu_lock);
pcpu_depopulate_chunk(chunk, i + 1, end + 1);
cond_resched();
spin_lock_irq(&pcpu_lock);
pcpu_chunk_depopulated(chunk, i + 1, end + 1);
/* reset the range and continue */
end = -1;
}
if (chunk->free_bytes == pcpu_unit_size)
pcpu_reintegrate_chunk(chunk);
else
list_move(&chunk->list,
&pcpu_chunk_lists[pcpu_sidelined_slot]);
}
} }
/** /**
* pcpu_balance_workfn - manage the amount of free chunks and populated pages * pcpu_balance_workfn - manage the amount of free chunks and populated pages
* @work: unused * @work: unused
* *
* Call __pcpu_balance_workfn() for each chunk type. * For each chunk type, manage the number of fully free chunks and the number of
* populated pages. An important thing to consider is when pages are freed and
* how they contribute to the global counts.
*/ */
static void pcpu_balance_workfn(struct work_struct *work) static void pcpu_balance_workfn(struct work_struct *work)
{ {
enum pcpu_chunk_type type; /*
* pcpu_balance_free() is called twice because the first time we may
* trim pages in the active pcpu_nr_empty_pop_pages which may cause us
* to grow other chunks. This then gives pcpu_reclaim_populated() time
* to move fully free chunks to the active list to be freed if
* appropriate.
*/
mutex_lock(&pcpu_alloc_mutex);
spin_lock_irq(&pcpu_lock);
for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++) pcpu_balance_free(false);
__pcpu_balance_workfn(type); pcpu_reclaim_populated();
pcpu_balance_populated();
pcpu_balance_free(true);
spin_unlock_irq(&pcpu_lock);
mutex_unlock(&pcpu_alloc_mutex);
} }
/** /**
@ -2085,7 +2252,6 @@ void free_percpu(void __percpu *ptr)
unsigned long flags; unsigned long flags;
int size, off; int size, off;
bool need_balance = false; bool need_balance = false;
struct list_head *pcpu_slot;
if (!ptr) if (!ptr)
return; return;
@ -2101,19 +2267,24 @@ void free_percpu(void __percpu *ptr)
size = pcpu_free_area(chunk, off); size = pcpu_free_area(chunk, off);
pcpu_slot = pcpu_chunk_list(pcpu_chunk_type(chunk));
pcpu_memcg_free_hook(chunk, off, size); pcpu_memcg_free_hook(chunk, off, size);
/* if there are more than one fully free chunks, wake up grim reaper */ /*
if (chunk->free_bytes == pcpu_unit_size) { * If there are more than one fully free chunks, wake up grim reaper.
* If the chunk is isolated, it may be in the process of being
* reclaimed. Let reclaim manage cleaning up of that chunk.
*/
if (!chunk->isolated && chunk->free_bytes == pcpu_unit_size) {
struct pcpu_chunk *pos; struct pcpu_chunk *pos;
list_for_each_entry(pos, &pcpu_slot[pcpu_nr_slots - 1], list) list_for_each_entry(pos, &pcpu_chunk_lists[pcpu_free_slot], list)
if (pos != chunk) { if (pos != chunk) {
need_balance = true; need_balance = true;
break; break;
} }
} else if (pcpu_should_reclaim_chunk(chunk)) {
pcpu_isolate_chunk(chunk);
need_balance = true;
} }
trace_percpu_free_percpu(chunk->base_addr, off, ptr); trace_percpu_free_percpu(chunk->base_addr, off, ptr);
@ -2414,7 +2585,6 @@ void __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
int map_size; int map_size;
unsigned long tmp_addr; unsigned long tmp_addr;
size_t alloc_size; size_t alloc_size;
enum pcpu_chunk_type type;
#define PCPU_SETUP_BUG_ON(cond) do { \ #define PCPU_SETUP_BUG_ON(cond) do { \
if (unlikely(cond)) { \ if (unlikely(cond)) { \
@ -2528,22 +2698,24 @@ void __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
pcpu_stats_save_ai(ai); pcpu_stats_save_ai(ai);
/* /*
* Allocate chunk slots. The additional last slot is for * Allocate chunk slots. The slots after the active slots are:
* empty chunks. * sidelined_slot - isolated, depopulated chunks
* free_slot - fully free chunks
* to_depopulate_slot - isolated, chunks to depopulate
*/ */
pcpu_nr_slots = __pcpu_size_to_slot(pcpu_unit_size) + 2; pcpu_sidelined_slot = __pcpu_size_to_slot(pcpu_unit_size) + 1;
pcpu_free_slot = pcpu_sidelined_slot + 1;
pcpu_to_depopulate_slot = pcpu_free_slot + 1;
pcpu_nr_slots = pcpu_to_depopulate_slot + 1;
pcpu_chunk_lists = memblock_alloc(pcpu_nr_slots * pcpu_chunk_lists = memblock_alloc(pcpu_nr_slots *
sizeof(pcpu_chunk_lists[0]) * sizeof(pcpu_chunk_lists[0]),
PCPU_NR_CHUNK_TYPES,
SMP_CACHE_BYTES); SMP_CACHE_BYTES);
if (!pcpu_chunk_lists) if (!pcpu_chunk_lists)
panic("%s: Failed to allocate %zu bytes\n", __func__, panic("%s: Failed to allocate %zu bytes\n", __func__,
pcpu_nr_slots * sizeof(pcpu_chunk_lists[0]) * pcpu_nr_slots * sizeof(pcpu_chunk_lists[0]));
PCPU_NR_CHUNK_TYPES);
for (type = 0; type < PCPU_NR_CHUNK_TYPES; type++) for (i = 0; i < pcpu_nr_slots; i++)
for (i = 0; i < pcpu_nr_slots; i++) INIT_LIST_HEAD(&pcpu_chunk_lists[i]);
INIT_LIST_HEAD(&pcpu_chunk_list(type)[i]);
/* /*
* The end of the static region needs to be aligned with the * The end of the static region needs to be aligned with the
@ -2580,7 +2752,7 @@ void __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
/* link the first chunk in */ /* link the first chunk in */
pcpu_first_chunk = chunk; pcpu_first_chunk = chunk;
pcpu_nr_empty_pop_pages[PCPU_CHUNK_ROOT] = pcpu_first_chunk->nr_empty_pop_pages; pcpu_nr_empty_pop_pages = pcpu_first_chunk->nr_empty_pop_pages;
pcpu_chunk_relocate(pcpu_first_chunk, -1); pcpu_chunk_relocate(pcpu_first_chunk, -1);
/* include all regions of the first chunk */ /* include all regions of the first chunk */
@ -2733,6 +2905,7 @@ static struct pcpu_alloc_info * __init __flatten pcpu_build_alloc_info(
* Related to atom_size, which could be much larger than the unit_size. * Related to atom_size, which could be much larger than the unit_size.
*/ */
last_allocs = INT_MAX; last_allocs = INT_MAX;
best_upa = 0;
for (upa = max_upa; upa; upa--) { for (upa = max_upa; upa; upa--) {
int allocs = 0, wasted = 0; int allocs = 0, wasted = 0;
@ -2759,6 +2932,7 @@ static struct pcpu_alloc_info * __init __flatten pcpu_build_alloc_info(
last_allocs = allocs; last_allocs = allocs;
best_upa = upa; best_upa = upa;
} }
BUG_ON(!best_upa);
upa = best_upa; upa = best_upa;
/* allocate and fill alloc_info */ /* allocate and fill alloc_info */