linux/fs/f2fs/gc.c
Jin Xu a26b7c8a01 f2fs: optimize gc for better performance
This patch improves the gc efficiency by optimizing the victim
selection policy. With this optimization, the random re-write
performance could increase up to 20%.

For f2fs, when disk is in shortage of free spaces, gc will selects
dirty segments and moves valid blocks around for making more space
available. The gc cost of a segment is determined by the valid blocks
in the segment. The less the valid blocks, the higher the efficiency.
The ideal victim segment is the one that has the most garbage blocks.

Currently, it searches up to 20 dirty segments for a victim segment.
The selected victim is not likely the best victim for gc when there
are much more dirty segments. Why not searching more dirty segments
for a better victim? The cost of searching dirty segments is
negligible in comparison to moving blocks.

In this patch, it enlarges the MAX_VICTIM_SEARCH to 4096 to make
the search more aggressively for a possible better victim. Since
it also applies to victim selection for SSR, it will likely improve
the SSR efficiency as well.

The test case is simple. It creates as many files until the disk full.
The size for each file is 32KB. Then it writes as many as 100000
records of 4KB size to random offsets of random files in sync mode.
The testing was done on a 2GB partition of a SDHC card. Let's see the
test result of f2fs without and with the patch.

---------------------------------------
2GB partition, SDHC
create 52023 files of size 32768 bytes
random re-write 100000 records of 4KB
---------------------------------------
| file creation (s) | rewrite time (s) | gc count | gc garbage blocks |
[no patch]  341         4227             1174          174840
[patched]   324         2958             645           106682

It's obvious that, with the patch, f2fs finishes the test in 20+% less
time than without the patch. And internally it does much less gc with
higher efficiency than before.

Since the performance improvement is related to gc, it might not be so
obvious for other tests that do not trigger gc as often as this one (
This is because f2fs selects dirty segments for SSR use most of the
time when free space is in shortage). The well-known iozone test tool
was not used for benchmarking the patch becuase it seems do not have
a test case that performs random re-write on a full disk.

This patch is the revised version based on the suggestion from
Jaegeuk Kim.

Signed-off-by: Jin Xu <jinuxstyle@gmail.com>
[Jaegeuk Kim: suggested simpler solution]
Reviewed-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2013-09-05 13:50:32 +09:00

750 lines
18 KiB
C

/*
* fs/f2fs/gc.c
*
* Copyright (c) 2012 Samsung Electronics Co., Ltd.
* http://www.samsung.com/
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*/
#include <linux/fs.h>
#include <linux/module.h>
#include <linux/backing-dev.h>
#include <linux/init.h>
#include <linux/f2fs_fs.h>
#include <linux/kthread.h>
#include <linux/delay.h>
#include <linux/freezer.h>
#include <linux/blkdev.h>
#include "f2fs.h"
#include "node.h"
#include "segment.h"
#include "gc.h"
#include <trace/events/f2fs.h>
static struct kmem_cache *winode_slab;
static int gc_thread_func(void *data)
{
struct f2fs_sb_info *sbi = data;
struct f2fs_gc_kthread *gc_th = sbi->gc_thread;
wait_queue_head_t *wq = &sbi->gc_thread->gc_wait_queue_head;
long wait_ms;
wait_ms = gc_th->min_sleep_time;
do {
if (try_to_freeze())
continue;
else
wait_event_interruptible_timeout(*wq,
kthread_should_stop(),
msecs_to_jiffies(wait_ms));
if (kthread_should_stop())
break;
if (sbi->sb->s_writers.frozen >= SB_FREEZE_WRITE) {
wait_ms = increase_sleep_time(gc_th, wait_ms);
continue;
}
/*
* [GC triggering condition]
* 0. GC is not conducted currently.
* 1. There are enough dirty segments.
* 2. IO subsystem is idle by checking the # of writeback pages.
* 3. IO subsystem is idle by checking the # of requests in
* bdev's request list.
*
* Note) We have to avoid triggering GCs too much frequently.
* Because it is possible that some segments can be
* invalidated soon after by user update or deletion.
* So, I'd like to wait some time to collect dirty segments.
*/
if (!mutex_trylock(&sbi->gc_mutex))
continue;
if (!is_idle(sbi)) {
wait_ms = increase_sleep_time(gc_th, wait_ms);
mutex_unlock(&sbi->gc_mutex);
continue;
}
if (has_enough_invalid_blocks(sbi))
wait_ms = decrease_sleep_time(gc_th, wait_ms);
else
wait_ms = increase_sleep_time(gc_th, wait_ms);
#ifdef CONFIG_F2FS_STAT_FS
sbi->bg_gc++;
#endif
/* if return value is not zero, no victim was selected */
if (f2fs_gc(sbi))
wait_ms = gc_th->no_gc_sleep_time;
} while (!kthread_should_stop());
return 0;
}
int start_gc_thread(struct f2fs_sb_info *sbi)
{
struct f2fs_gc_kthread *gc_th;
dev_t dev = sbi->sb->s_bdev->bd_dev;
int err = 0;
if (!test_opt(sbi, BG_GC))
goto out;
gc_th = kmalloc(sizeof(struct f2fs_gc_kthread), GFP_KERNEL);
if (!gc_th) {
err = -ENOMEM;
goto out;
}
gc_th->min_sleep_time = DEF_GC_THREAD_MIN_SLEEP_TIME;
gc_th->max_sleep_time = DEF_GC_THREAD_MAX_SLEEP_TIME;
gc_th->no_gc_sleep_time = DEF_GC_THREAD_NOGC_SLEEP_TIME;
gc_th->gc_idle = 0;
sbi->gc_thread = gc_th;
init_waitqueue_head(&sbi->gc_thread->gc_wait_queue_head);
sbi->gc_thread->f2fs_gc_task = kthread_run(gc_thread_func, sbi,
"f2fs_gc-%u:%u", MAJOR(dev), MINOR(dev));
if (IS_ERR(gc_th->f2fs_gc_task)) {
err = PTR_ERR(gc_th->f2fs_gc_task);
kfree(gc_th);
sbi->gc_thread = NULL;
}
out:
return err;
}
void stop_gc_thread(struct f2fs_sb_info *sbi)
{
struct f2fs_gc_kthread *gc_th = sbi->gc_thread;
if (!gc_th)
return;
kthread_stop(gc_th->f2fs_gc_task);
kfree(gc_th);
sbi->gc_thread = NULL;
}
static int select_gc_type(struct f2fs_gc_kthread *gc_th, int gc_type)
{
int gc_mode = (gc_type == BG_GC) ? GC_CB : GC_GREEDY;
if (gc_th && gc_th->gc_idle) {
if (gc_th->gc_idle == 1)
gc_mode = GC_CB;
else if (gc_th->gc_idle == 2)
gc_mode = GC_GREEDY;
}
return gc_mode;
}
static void select_policy(struct f2fs_sb_info *sbi, int gc_type,
int type, struct victim_sel_policy *p)
{
struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
if (p->alloc_mode == SSR) {
p->gc_mode = GC_GREEDY;
p->dirty_segmap = dirty_i->dirty_segmap[type];
p->max_search = dirty_i->nr_dirty[type];
p->ofs_unit = 1;
} else {
p->gc_mode = select_gc_type(sbi->gc_thread, gc_type);
p->dirty_segmap = dirty_i->dirty_segmap[DIRTY];
p->max_search = dirty_i->nr_dirty[DIRTY];
p->ofs_unit = sbi->segs_per_sec;
}
if (p->max_search > MAX_VICTIM_SEARCH)
p->max_search = MAX_VICTIM_SEARCH;
p->offset = sbi->last_victim[p->gc_mode];
}
static unsigned int get_max_cost(struct f2fs_sb_info *sbi,
struct victim_sel_policy *p)
{
/* SSR allocates in a segment unit */
if (p->alloc_mode == SSR)
return 1 << sbi->log_blocks_per_seg;
if (p->gc_mode == GC_GREEDY)
return (1 << sbi->log_blocks_per_seg) * p->ofs_unit;
else if (p->gc_mode == GC_CB)
return UINT_MAX;
else /* No other gc_mode */
return 0;
}
static unsigned int check_bg_victims(struct f2fs_sb_info *sbi)
{
struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
unsigned int hint = 0;
unsigned int secno;
/*
* If the gc_type is FG_GC, we can select victim segments
* selected by background GC before.
* Those segments guarantee they have small valid blocks.
*/
next:
secno = find_next_bit(dirty_i->victim_secmap, TOTAL_SECS(sbi), hint++);
if (secno < TOTAL_SECS(sbi)) {
if (sec_usage_check(sbi, secno))
goto next;
clear_bit(secno, dirty_i->victim_secmap);
return secno * sbi->segs_per_sec;
}
return NULL_SEGNO;
}
static unsigned int get_cb_cost(struct f2fs_sb_info *sbi, unsigned int segno)
{
struct sit_info *sit_i = SIT_I(sbi);
unsigned int secno = GET_SECNO(sbi, segno);
unsigned int start = secno * sbi->segs_per_sec;
unsigned long long mtime = 0;
unsigned int vblocks;
unsigned char age = 0;
unsigned char u;
unsigned int i;
for (i = 0; i < sbi->segs_per_sec; i++)
mtime += get_seg_entry(sbi, start + i)->mtime;
vblocks = get_valid_blocks(sbi, segno, sbi->segs_per_sec);
mtime = div_u64(mtime, sbi->segs_per_sec);
vblocks = div_u64(vblocks, sbi->segs_per_sec);
u = (vblocks * 100) >> sbi->log_blocks_per_seg;
/* Handle if the system time is changed by user */
if (mtime < sit_i->min_mtime)
sit_i->min_mtime = mtime;
if (mtime > sit_i->max_mtime)
sit_i->max_mtime = mtime;
if (sit_i->max_mtime != sit_i->min_mtime)
age = 100 - div64_u64(100 * (mtime - sit_i->min_mtime),
sit_i->max_mtime - sit_i->min_mtime);
return UINT_MAX - ((100 * (100 - u) * age) / (100 + u));
}
static unsigned int get_gc_cost(struct f2fs_sb_info *sbi, unsigned int segno,
struct victim_sel_policy *p)
{
if (p->alloc_mode == SSR)
return get_seg_entry(sbi, segno)->ckpt_valid_blocks;
/* alloc_mode == LFS */
if (p->gc_mode == GC_GREEDY)
return get_valid_blocks(sbi, segno, sbi->segs_per_sec);
else
return get_cb_cost(sbi, segno);
}
/*
* This function is called from two paths.
* One is garbage collection and the other is SSR segment selection.
* When it is called during GC, it just gets a victim segment
* and it does not remove it from dirty seglist.
* When it is called from SSR segment selection, it finds a segment
* which has minimum valid blocks and removes it from dirty seglist.
*/
static int get_victim_by_default(struct f2fs_sb_info *sbi,
unsigned int *result, int gc_type, int type, char alloc_mode)
{
struct dirty_seglist_info *dirty_i = DIRTY_I(sbi);
struct victim_sel_policy p;
unsigned int secno, max_cost;
int nsearched = 0;
p.alloc_mode = alloc_mode;
select_policy(sbi, gc_type, type, &p);
p.min_segno = NULL_SEGNO;
p.min_cost = max_cost = get_max_cost(sbi, &p);
mutex_lock(&dirty_i->seglist_lock);
if (p.alloc_mode == LFS && gc_type == FG_GC) {
p.min_segno = check_bg_victims(sbi);
if (p.min_segno != NULL_SEGNO)
goto got_it;
}
while (1) {
unsigned long cost;
unsigned int segno;
segno = find_next_bit(p.dirty_segmap,
TOTAL_SEGS(sbi), p.offset);
if (segno >= TOTAL_SEGS(sbi)) {
if (sbi->last_victim[p.gc_mode]) {
sbi->last_victim[p.gc_mode] = 0;
p.offset = 0;
continue;
}
break;
}
p.offset = ((segno / p.ofs_unit) * p.ofs_unit) + p.ofs_unit;
secno = GET_SECNO(sbi, segno);
if (sec_usage_check(sbi, secno))
continue;
if (gc_type == BG_GC && test_bit(secno, dirty_i->victim_secmap))
continue;
cost = get_gc_cost(sbi, segno, &p);
if (p.min_cost > cost) {
p.min_segno = segno;
p.min_cost = cost;
}
if (cost == max_cost)
continue;
if (nsearched++ >= p.max_search) {
sbi->last_victim[p.gc_mode] = segno;
break;
}
}
if (p.min_segno != NULL_SEGNO) {
got_it:
if (p.alloc_mode == LFS) {
secno = GET_SECNO(sbi, p.min_segno);
if (gc_type == FG_GC)
sbi->cur_victim_sec = secno;
else
set_bit(secno, dirty_i->victim_secmap);
}
*result = (p.min_segno / p.ofs_unit) * p.ofs_unit;
trace_f2fs_get_victim(sbi->sb, type, gc_type, &p,
sbi->cur_victim_sec,
prefree_segments(sbi), free_segments(sbi));
}
mutex_unlock(&dirty_i->seglist_lock);
return (p.min_segno == NULL_SEGNO) ? 0 : 1;
}
static const struct victim_selection default_v_ops = {
.get_victim = get_victim_by_default,
};
static struct inode *find_gc_inode(nid_t ino, struct list_head *ilist)
{
struct inode_entry *ie;
list_for_each_entry(ie, ilist, list)
if (ie->inode->i_ino == ino)
return ie->inode;
return NULL;
}
static void add_gc_inode(struct inode *inode, struct list_head *ilist)
{
struct inode_entry *new_ie;
if (inode == find_gc_inode(inode->i_ino, ilist)) {
iput(inode);
return;
}
repeat:
new_ie = kmem_cache_alloc(winode_slab, GFP_NOFS);
if (!new_ie) {
cond_resched();
goto repeat;
}
new_ie->inode = inode;
list_add_tail(&new_ie->list, ilist);
}
static void put_gc_inode(struct list_head *ilist)
{
struct inode_entry *ie, *next_ie;
list_for_each_entry_safe(ie, next_ie, ilist, list) {
iput(ie->inode);
list_del(&ie->list);
kmem_cache_free(winode_slab, ie);
}
}
static int check_valid_map(struct f2fs_sb_info *sbi,
unsigned int segno, int offset)
{
struct sit_info *sit_i = SIT_I(sbi);
struct seg_entry *sentry;
int ret;
mutex_lock(&sit_i->sentry_lock);
sentry = get_seg_entry(sbi, segno);
ret = f2fs_test_bit(offset, sentry->cur_valid_map);
mutex_unlock(&sit_i->sentry_lock);
return ret;
}
/*
* This function compares node address got in summary with that in NAT.
* On validity, copy that node with cold status, otherwise (invalid node)
* ignore that.
*/
static void gc_node_segment(struct f2fs_sb_info *sbi,
struct f2fs_summary *sum, unsigned int segno, int gc_type)
{
bool initial = true;
struct f2fs_summary *entry;
int off;
next_step:
entry = sum;
for (off = 0; off < sbi->blocks_per_seg; off++, entry++) {
nid_t nid = le32_to_cpu(entry->nid);
struct page *node_page;
/* stop BG_GC if there is not enough free sections. */
if (gc_type == BG_GC && has_not_enough_free_secs(sbi, 0))
return;
if (check_valid_map(sbi, segno, off) == 0)
continue;
if (initial) {
ra_node_page(sbi, nid);
continue;
}
node_page = get_node_page(sbi, nid);
if (IS_ERR(node_page))
continue;
/* set page dirty and write it */
if (gc_type == FG_GC) {
f2fs_wait_on_page_writeback(node_page, NODE, true);
set_page_dirty(node_page);
} else {
if (!PageWriteback(node_page))
set_page_dirty(node_page);
}
f2fs_put_page(node_page, 1);
stat_inc_node_blk_count(sbi, 1);
}
if (initial) {
initial = false;
goto next_step;
}
if (gc_type == FG_GC) {
struct writeback_control wbc = {
.sync_mode = WB_SYNC_ALL,
.nr_to_write = LONG_MAX,
.for_reclaim = 0,
};
sync_node_pages(sbi, 0, &wbc);
/*
* In the case of FG_GC, it'd be better to reclaim this victim
* completely.
*/
if (get_valid_blocks(sbi, segno, 1) != 0)
goto next_step;
}
}
/*
* Calculate start block index indicating the given node offset.
* Be careful, caller should give this node offset only indicating direct node
* blocks. If any node offsets, which point the other types of node blocks such
* as indirect or double indirect node blocks, are given, it must be a caller's
* bug.
*/
block_t start_bidx_of_node(unsigned int node_ofs, struct f2fs_inode_info *fi)
{
unsigned int indirect_blks = 2 * NIDS_PER_BLOCK + 4;
unsigned int bidx;
if (node_ofs == 0)
return 0;
if (node_ofs <= 2) {
bidx = node_ofs - 1;
} else if (node_ofs <= indirect_blks) {
int dec = (node_ofs - 4) / (NIDS_PER_BLOCK + 1);
bidx = node_ofs - 2 - dec;
} else {
int dec = (node_ofs - indirect_blks - 3) / (NIDS_PER_BLOCK + 1);
bidx = node_ofs - 5 - dec;
}
return bidx * ADDRS_PER_BLOCK + ADDRS_PER_INODE(fi);
}
static int check_dnode(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
struct node_info *dni, block_t blkaddr, unsigned int *nofs)
{
struct page *node_page;
nid_t nid;
unsigned int ofs_in_node;
block_t source_blkaddr;
nid = le32_to_cpu(sum->nid);
ofs_in_node = le16_to_cpu(sum->ofs_in_node);
node_page = get_node_page(sbi, nid);
if (IS_ERR(node_page))
return 0;
get_node_info(sbi, nid, dni);
if (sum->version != dni->version) {
f2fs_put_page(node_page, 1);
return 0;
}
*nofs = ofs_of_node(node_page);
source_blkaddr = datablock_addr(node_page, ofs_in_node);
f2fs_put_page(node_page, 1);
if (source_blkaddr != blkaddr)
return 0;
return 1;
}
static void move_data_page(struct inode *inode, struct page *page, int gc_type)
{
if (gc_type == BG_GC) {
if (PageWriteback(page))
goto out;
set_page_dirty(page);
set_cold_data(page);
} else {
struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
f2fs_wait_on_page_writeback(page, DATA, true);
if (clear_page_dirty_for_io(page) &&
S_ISDIR(inode->i_mode)) {
dec_page_count(sbi, F2FS_DIRTY_DENTS);
inode_dec_dirty_dents(inode);
}
set_cold_data(page);
do_write_data_page(page);
clear_cold_data(page);
}
out:
f2fs_put_page(page, 1);
}
/*
* This function tries to get parent node of victim data block, and identifies
* data block validity. If the block is valid, copy that with cold status and
* modify parent node.
* If the parent node is not valid or the data block address is different,
* the victim data block is ignored.
*/
static void gc_data_segment(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
struct list_head *ilist, unsigned int segno, int gc_type)
{
struct super_block *sb = sbi->sb;
struct f2fs_summary *entry;
block_t start_addr;
int off;
int phase = 0;
start_addr = START_BLOCK(sbi, segno);
next_step:
entry = sum;
for (off = 0; off < sbi->blocks_per_seg; off++, entry++) {
struct page *data_page;
struct inode *inode;
struct node_info dni; /* dnode info for the data */
unsigned int ofs_in_node, nofs;
block_t start_bidx;
/* stop BG_GC if there is not enough free sections. */
if (gc_type == BG_GC && has_not_enough_free_secs(sbi, 0))
return;
if (check_valid_map(sbi, segno, off) == 0)
continue;
if (phase == 0) {
ra_node_page(sbi, le32_to_cpu(entry->nid));
continue;
}
/* Get an inode by ino with checking validity */
if (check_dnode(sbi, entry, &dni, start_addr + off, &nofs) == 0)
continue;
if (phase == 1) {
ra_node_page(sbi, dni.ino);
continue;
}
ofs_in_node = le16_to_cpu(entry->ofs_in_node);
if (phase == 2) {
inode = f2fs_iget(sb, dni.ino);
if (IS_ERR(inode))
continue;
start_bidx = start_bidx_of_node(nofs, F2FS_I(inode));
data_page = find_data_page(inode,
start_bidx + ofs_in_node, false);
if (IS_ERR(data_page))
goto next_iput;
f2fs_put_page(data_page, 0);
add_gc_inode(inode, ilist);
} else {
inode = find_gc_inode(dni.ino, ilist);
if (inode) {
start_bidx = start_bidx_of_node(nofs,
F2FS_I(inode));
data_page = get_lock_data_page(inode,
start_bidx + ofs_in_node);
if (IS_ERR(data_page))
continue;
move_data_page(inode, data_page, gc_type);
stat_inc_data_blk_count(sbi, 1);
}
}
continue;
next_iput:
iput(inode);
}
if (++phase < 4)
goto next_step;
if (gc_type == FG_GC) {
f2fs_submit_bio(sbi, DATA, true);
/*
* In the case of FG_GC, it'd be better to reclaim this victim
* completely.
*/
if (get_valid_blocks(sbi, segno, 1) != 0) {
phase = 2;
goto next_step;
}
}
}
static int __get_victim(struct f2fs_sb_info *sbi, unsigned int *victim,
int gc_type, int type)
{
struct sit_info *sit_i = SIT_I(sbi);
int ret;
mutex_lock(&sit_i->sentry_lock);
ret = DIRTY_I(sbi)->v_ops->get_victim(sbi, victim, gc_type, type, LFS);
mutex_unlock(&sit_i->sentry_lock);
return ret;
}
static void do_garbage_collect(struct f2fs_sb_info *sbi, unsigned int segno,
struct list_head *ilist, int gc_type)
{
struct page *sum_page;
struct f2fs_summary_block *sum;
struct blk_plug plug;
/* read segment summary of victim */
sum_page = get_sum_page(sbi, segno);
if (IS_ERR(sum_page))
return;
blk_start_plug(&plug);
sum = page_address(sum_page);
switch (GET_SUM_TYPE((&sum->footer))) {
case SUM_TYPE_NODE:
gc_node_segment(sbi, sum->entries, segno, gc_type);
break;
case SUM_TYPE_DATA:
gc_data_segment(sbi, sum->entries, ilist, segno, gc_type);
break;
}
blk_finish_plug(&plug);
stat_inc_seg_count(sbi, GET_SUM_TYPE((&sum->footer)));
stat_inc_call_count(sbi->stat_info);
f2fs_put_page(sum_page, 1);
}
int f2fs_gc(struct f2fs_sb_info *sbi)
{
struct list_head ilist;
unsigned int segno, i;
int gc_type = BG_GC;
int nfree = 0;
int ret = -1;
INIT_LIST_HEAD(&ilist);
gc_more:
if (!(sbi->sb->s_flags & MS_ACTIVE))
goto stop;
if (gc_type == BG_GC && has_not_enough_free_secs(sbi, nfree)) {
gc_type = FG_GC;
write_checkpoint(sbi, false);
}
if (!__get_victim(sbi, &segno, gc_type, NO_CHECK_TYPE))
goto stop;
ret = 0;
for (i = 0; i < sbi->segs_per_sec; i++)
do_garbage_collect(sbi, segno + i, &ilist, gc_type);
if (gc_type == FG_GC) {
sbi->cur_victim_sec = NULL_SEGNO;
nfree++;
WARN_ON(get_valid_blocks(sbi, segno, sbi->segs_per_sec));
}
if (has_not_enough_free_secs(sbi, nfree))
goto gc_more;
if (gc_type == FG_GC)
write_checkpoint(sbi, false);
stop:
mutex_unlock(&sbi->gc_mutex);
put_gc_inode(&ilist);
return ret;
}
void build_gc_manager(struct f2fs_sb_info *sbi)
{
DIRTY_I(sbi)->v_ops = &default_v_ops;
}
int __init create_gc_caches(void)
{
winode_slab = f2fs_kmem_cache_create("f2fs_gc_inodes",
sizeof(struct inode_entry), NULL);
if (!winode_slab)
return -ENOMEM;
return 0;
}
void destroy_gc_caches(void)
{
kmem_cache_destroy(winode_slab);
}