aboutsummaryrefslogtreecommitdiff
path: root/fs/f2fs/gc.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/f2fs/gc.c')
-rw-r--r--fs/f2fs/gc.c186
1 files changed, 129 insertions, 57 deletions
diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
index 88bfc3dff496..fa3d2e2df8e7 100644
--- a/fs/f2fs/gc.c
+++ b/fs/f2fs/gc.c
@@ -32,13 +32,14 @@ static int gc_thread_func(void *data)
wait_ms = gc_th->min_sleep_time;
+ set_freezable();
do {
+ wait_event_interruptible_timeout(*wq,
+ kthread_should_stop() || freezing(current),
+ msecs_to_jiffies(wait_ms));
+
if (try_to_freeze())
continue;
- else
- wait_event_interruptible_timeout(*wq,
- kthread_should_stop(),
- msecs_to_jiffies(wait_ms));
if (kthread_should_stop())
break;
@@ -48,8 +49,10 @@ static int gc_thread_func(void *data)
}
#ifdef CONFIG_F2FS_FAULT_INJECTION
- if (time_to_inject(sbi, FAULT_CHECKPOINT))
+ if (time_to_inject(sbi, FAULT_CHECKPOINT)) {
+ f2fs_show_injection_info(FAULT_CHECKPOINT);
f2fs_stop_checkpoint(sbi, false);
+ }
#endif
/*
@@ -82,7 +85,7 @@ static int gc_thread_func(void *data)
stat_inc_bggc_count(sbi);
/* if return value is not zero, no victim was selected */
- if (f2fs_gc(sbi, test_opt(sbi, FORCE_FG_GC), true))
+ if (f2fs_gc(sbi, test_opt(sbi, FORCE_FG_GC), true, NULL_SEGNO))
wait_ms = gc_th->no_gc_sleep_time;
trace_f2fs_background_gc(sbi->sb, wait_ms,
@@ -166,10 +169,15 @@ static void select_policy(struct f2fs_sb_info *sbi, int gc_type,
p->ofs_unit = sbi->segs_per_sec;
}
- if (p->max_search > sbi->max_victim_search)
+ /* we need to check every dirty segments in the FG_GC case */
+ if (gc_type != FG_GC && p->max_search > sbi->max_victim_search)
p->max_search = sbi->max_victim_search;
- p->offset = sbi->last_victim[p->gc_mode];
+ /* let's select beginning hot/small space first */
+ if (type == CURSEG_HOT_DATA || IS_NODESEG(type))
+ p->offset = 0;
+ else
+ p->offset = SIT_I(sbi)->last_victim[p->gc_mode];
}
static unsigned int get_max_cost(struct f2fs_sb_info *sbi,
@@ -179,7 +187,7 @@ static unsigned int get_max_cost(struct f2fs_sb_info *sbi,
if (p->alloc_mode == SSR)
return sbi->blocks_per_seg;
if (p->gc_mode == GC_GREEDY)
- return sbi->blocks_per_seg * p->ofs_unit;
+ return 2 * sbi->blocks_per_seg * p->ofs_unit;
else if (p->gc_mode == GC_CB)
return UINT_MAX;
else /* No other gc_mode */
@@ -199,8 +207,12 @@ static unsigned int check_bg_victims(struct f2fs_sb_info *sbi)
for_each_set_bit(secno, dirty_i->victim_secmap, MAIN_SECS(sbi)) {
if (sec_usage_check(sbi, secno))
continue;
+
+ if (no_fggc_candidate(sbi, secno))
+ continue;
+
clear_bit(secno, dirty_i->victim_secmap);
- return secno * sbi->segs_per_sec;
+ return GET_SEG_FROM_SEC(sbi, secno);
}
return NULL_SEGNO;
}
@@ -208,8 +220,8 @@ static unsigned int check_bg_victims(struct f2fs_sb_info *sbi)
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 int secno = GET_SEC_FROM_SEG(sbi, segno);
+ unsigned int start = GET_SEG_FROM_SEC(sbi, secno);
unsigned long long mtime = 0;
unsigned int vblocks;
unsigned char age = 0;
@@ -218,7 +230,7 @@ static unsigned int get_cb_cost(struct f2fs_sb_info *sbi, unsigned int segno)
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);
+ vblocks = get_valid_blocks(sbi, segno, true);
mtime = div_u64(mtime, sbi->segs_per_sec);
vblocks = div_u64(vblocks, sbi->segs_per_sec);
@@ -237,15 +249,34 @@ static unsigned int get_cb_cost(struct f2fs_sb_info *sbi, unsigned int segno)
return UINT_MAX - ((100 * (100 - u) * age) / (100 + u));
}
+static unsigned int get_greedy_cost(struct f2fs_sb_info *sbi,
+ unsigned int segno)
+{
+ unsigned int valid_blocks =
+ get_valid_blocks(sbi, segno, true);
+
+ return IS_DATASEG(get_seg_entry(sbi, segno)->type) ?
+ valid_blocks * 2 : valid_blocks;
+}
+
+static unsigned int get_ssr_cost(struct f2fs_sb_info *sbi,
+ unsigned int segno)
+{
+ struct seg_entry *se = get_seg_entry(sbi, segno);
+
+ return se->ckpt_valid_blocks > se->valid_blocks ?
+ se->ckpt_valid_blocks : se->valid_blocks;
+}
+
static inline 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;
+ return get_ssr_cost(sbi, segno);
/* alloc_mode == LFS */
if (p->gc_mode == GC_GREEDY)
- return get_valid_blocks(sbi, segno, sbi->segs_per_sec);
+ return get_greedy_cost(sbi, segno);
else
return get_cb_cost(sbi, segno);
}
@@ -274,6 +305,7 @@ 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 sit_info *sm = SIT_I(sbi);
struct victim_sel_policy p;
unsigned int secno, last_victim;
unsigned int last_segment = MAIN_SEGS(sbi);
@@ -287,10 +319,18 @@ static int get_victim_by_default(struct f2fs_sb_info *sbi,
p.min_segno = NULL_SEGNO;
p.min_cost = get_max_cost(sbi, &p);
+ if (*result != NULL_SEGNO) {
+ if (IS_DATASEG(get_seg_entry(sbi, *result)->type) &&
+ get_valid_blocks(sbi, *result, false) &&
+ !sec_usage_check(sbi, GET_SEC_FROM_SEG(sbi, *result)))
+ p.min_segno = *result;
+ goto out;
+ }
+
if (p.max_search == 0)
goto out;
- last_victim = sbi->last_victim[p.gc_mode];
+ last_victim = sm->last_victim[p.gc_mode];
if (p.alloc_mode == LFS && gc_type == FG_GC) {
p.min_segno = check_bg_victims(sbi);
if (p.min_segno != NULL_SEGNO)
@@ -303,9 +343,10 @@ static int get_victim_by_default(struct f2fs_sb_info *sbi,
segno = find_next_bit(p.dirty_segmap, last_segment, p.offset);
if (segno >= last_segment) {
- if (sbi->last_victim[p.gc_mode]) {
- last_segment = sbi->last_victim[p.gc_mode];
- sbi->last_victim[p.gc_mode] = 0;
+ if (sm->last_victim[p.gc_mode]) {
+ last_segment =
+ sm->last_victim[p.gc_mode];
+ sm->last_victim[p.gc_mode] = 0;
p.offset = 0;
continue;
}
@@ -322,13 +363,15 @@ static int get_victim_by_default(struct f2fs_sb_info *sbi,
nsearched++;
}
-
- secno = GET_SECNO(sbi, segno);
+ secno = GET_SEC_FROM_SEG(sbi, segno);
if (sec_usage_check(sbi, secno))
goto next;
if (gc_type == BG_GC && test_bit(secno, dirty_i->victim_secmap))
goto next;
+ if (gc_type == FG_GC && p.alloc_mode == LFS &&
+ no_fggc_candidate(sbi, secno))
+ goto next;
cost = get_gc_cost(sbi, segno, &p);
@@ -338,17 +381,18 @@ static int get_victim_by_default(struct f2fs_sb_info *sbi,
}
next:
if (nsearched >= p.max_search) {
- if (!sbi->last_victim[p.gc_mode] && segno <= last_victim)
- sbi->last_victim[p.gc_mode] = last_victim + 1;
+ if (!sm->last_victim[p.gc_mode] && segno <= last_victim)
+ sm->last_victim[p.gc_mode] = last_victim + 1;
else
- sbi->last_victim[p.gc_mode] = segno + 1;
+ sm->last_victim[p.gc_mode] = segno + 1;
+ sm->last_victim[p.gc_mode] %= MAIN_SEGS(sbi);
break;
}
}
if (p.min_segno != NULL_SEGNO) {
got_it:
if (p.alloc_mode == LFS) {
- secno = GET_SECNO(sbi, p.min_segno);
+ secno = GET_SEC_FROM_SEG(sbi, p.min_segno);
if (gc_type == FG_GC)
sbi->cur_victim_sec = secno;
else
@@ -531,8 +575,10 @@ static bool is_alive(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
get_node_info(sbi, nid, dni);
if (sum->version != dni->version) {
- f2fs_put_page(node_page, 1);
- return false;
+ f2fs_msg(sbi->sb, KERN_WARNING,
+ "%s: valid data with mismatched node version.",
+ __func__);
+ set_sbi_flag(sbi, SBI_NEED_FSCK);
}
*nofs = ofs_of_node(node_page);
@@ -550,9 +596,11 @@ static void move_encrypted_block(struct inode *inode, block_t bidx,
struct f2fs_io_info fio = {
.sbi = F2FS_I_SB(inode),
.type = DATA,
+ .temp = COLD,
.op = REQ_OP_READ,
.op_flags = 0,
.encrypted_page = NULL,
+ .in_list = false,
};
struct dnode_of_data dn;
struct f2fs_summary sum;
@@ -569,6 +617,9 @@ static void move_encrypted_block(struct inode *inode, block_t bidx,
if (!check_valid_map(F2FS_I_SB(inode), segno, off))
goto out;
+ if (f2fs_is_atomic_file(inode))
+ goto out;
+
set_new_dnode(&dn, inode, NULL, NULL, 0);
err = get_dnode_of_data(&dn, bidx, LOOKUP_NODE);
if (err)
@@ -593,7 +644,7 @@ static void move_encrypted_block(struct inode *inode, block_t bidx,
fio.new_blkaddr = fio.old_blkaddr = dn.data_blkaddr;
allocate_data_block(fio.sbi, NULL, fio.old_blkaddr, &newaddr,
- &sum, CURSEG_COLD_DATA);
+ &sum, CURSEG_COLD_DATA, NULL, false);
fio.encrypted_page = pagecache_get_page(META_MAPPING(fio.sbi), newaddr,
FGP_LOCK | FGP_CREAT, GFP_NOFS);
@@ -631,7 +682,7 @@ static void move_encrypted_block(struct inode *inode, block_t bidx,
fio.op = REQ_OP_WRITE;
fio.op_flags = REQ_SYNC;
fio.new_blkaddr = newaddr;
- f2fs_submit_page_mbio(&fio);
+ f2fs_submit_page_write(&fio);
f2fs_update_data_blkaddr(&dn, newaddr);
set_inode_flag(inode, FI_APPEND_WRITE);
@@ -661,6 +712,9 @@ static void move_data_page(struct inode *inode, block_t bidx, int gc_type,
if (!check_valid_map(F2FS_I_SB(inode), segno, off))
goto out;
+ if (f2fs_is_atomic_file(inode))
+ goto out;
+
if (gc_type == BG_GC) {
if (PageWriteback(page))
goto out;
@@ -670,10 +724,13 @@ static void move_data_page(struct inode *inode, block_t bidx, int gc_type,
struct f2fs_io_info fio = {
.sbi = F2FS_I_SB(inode),
.type = DATA,
+ .temp = COLD,
.op = REQ_OP_WRITE,
.op_flags = REQ_SYNC,
+ .old_blkaddr = NULL_ADDR,
.page = page,
.encrypted_page = NULL,
+ .need_lock = LOCK_REQ,
};
bool is_dirty = PageDirty(page);
int err;
@@ -865,7 +922,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi,
GET_SUM_BLOCK(sbi, segno));
f2fs_put_page(sum_page, 0);
- if (get_valid_blocks(sbi, segno, 1) == 0 ||
+ if (get_valid_blocks(sbi, segno, false) == 0 ||
!PageUptodate(sum_page) ||
unlikely(f2fs_cp_error(sbi)))
goto next;
@@ -880,7 +937,6 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi,
* - mutex_lock(sentry_lock) - change_curseg()
* - lock_page(sum_page)
*/
-
if (type == SUM_TYPE_NODE)
gc_node_segment(sbi, sum->entries, segno, gc_type);
else
@@ -893,13 +949,13 @@ next:
}
if (gc_type == FG_GC)
- f2fs_submit_merged_bio(sbi,
- (type == SUM_TYPE_NODE) ? NODE : DATA, WRITE);
+ f2fs_submit_merged_write(sbi,
+ (type == SUM_TYPE_NODE) ? NODE : DATA);
blk_finish_plug(&plug);
if (gc_type == FG_GC &&
- get_valid_blocks(sbi, start_segno, sbi->segs_per_sec) == 0)
+ get_valid_blocks(sbi, start_segno, true) == 0)
sec_freed = 1;
stat_inc_call_count(sbi->stat_info);
@@ -907,13 +963,14 @@ next:
return sec_freed;
}
-int f2fs_gc(struct f2fs_sb_info *sbi, bool sync, bool background)
+int f2fs_gc(struct f2fs_sb_info *sbi, bool sync,
+ bool background, unsigned int segno)
{
- unsigned int segno;
int gc_type = sync ? FG_GC : BG_GC;
int sec_freed = 0;
- int ret = -EINVAL;
+ int ret;
struct cp_control cpc;
+ unsigned int init_segno = segno;
struct gc_inode_list gc_list = {
.ilist = LIST_HEAD_INIT(gc_list.ilist),
.iroot = RADIX_TREE_INIT(GFP_NOFS),
@@ -921,39 +978,35 @@ int f2fs_gc(struct f2fs_sb_info *sbi, bool sync, bool background)
cpc.reason = __get_cp_reason(sbi);
gc_more:
- segno = NULL_SEGNO;
-
- if (unlikely(!(sbi->sb->s_flags & MS_ACTIVE)))
+ if (unlikely(!(sbi->sb->s_flags & MS_ACTIVE))) {
+ ret = -EINVAL;
goto stop;
+ }
if (unlikely(f2fs_cp_error(sbi))) {
ret = -EIO;
goto stop;
}
- if (gc_type == BG_GC && has_not_enough_free_secs(sbi, sec_freed, 0)) {
- gc_type = FG_GC;
+ if (gc_type == BG_GC && has_not_enough_free_secs(sbi, 0, 0)) {
/*
- * If there is no victim and no prefree segment but still not
- * enough free sections, we should flush dent/node blocks and do
- * garbage collections.
+ * For example, if there are many prefree_segments below given
+ * threshold, we can make them free by checkpoint. Then, we
+ * secure free segments which doesn't need fggc any more.
*/
- if (__get_victim(sbi, &segno, gc_type) ||
- prefree_segments(sbi)) {
- ret = write_checkpoint(sbi, &cpc);
- if (ret)
- goto stop;
- segno = NULL_SEGNO;
- } else if (has_not_enough_free_secs(sbi, 0, 0)) {
+ if (prefree_segments(sbi)) {
ret = write_checkpoint(sbi, &cpc);
if (ret)
goto stop;
}
- } else if (gc_type == BG_GC && !background) {
- /* f2fs_balance_fs doesn't need to do BG_GC in critical path. */
- goto stop;
+ if (has_not_enough_free_secs(sbi, 0, 0))
+ gc_type = FG_GC;
}
- if (segno == NULL_SEGNO && !__get_victim(sbi, &segno, gc_type))
+ ret = -EINVAL;
+ /* f2fs_balance_fs doesn't need to do BG_GC in critical path. */
+ if (gc_type == BG_GC && !background)
+ goto stop;
+ if (!__get_victim(sbi, &segno, gc_type))
goto stop;
ret = 0;
@@ -965,13 +1018,17 @@ gc_more:
sbi->cur_victim_sec = NULL_SEGNO;
if (!sync) {
- if (has_not_enough_free_secs(sbi, sec_freed, 0))
+ if (has_not_enough_free_secs(sbi, sec_freed, 0)) {
+ segno = NULL_SEGNO;
goto gc_more;
+ }
if (gc_type == FG_GC)
ret = write_checkpoint(sbi, &cpc);
}
stop:
+ SIT_I(sbi)->last_victim[ALLOC_NEXT] = 0;
+ SIT_I(sbi)->last_victim[FLUSH_DEVICE] = init_segno;
mutex_unlock(&sbi->gc_mutex);
put_gc_inode(&gc_list);
@@ -983,5 +1040,20 @@ stop:
void build_gc_manager(struct f2fs_sb_info *sbi)
{
+ u64 main_count, resv_count, ovp_count;
+
DIRTY_I(sbi)->v_ops = &default_v_ops;
+
+ /* threshold of # of valid blocks in a section for victims of FG_GC */
+ main_count = SM_I(sbi)->main_segments << sbi->log_blocks_per_seg;
+ resv_count = SM_I(sbi)->reserved_segments << sbi->log_blocks_per_seg;
+ ovp_count = SM_I(sbi)->ovp_segments << sbi->log_blocks_per_seg;
+
+ sbi->fggc_threshold = div64_u64((main_count - ovp_count) *
+ BLKS_PER_SEC(sbi), (main_count - resv_count));
+
+ /* give warm/cold data area from slower device */
+ if (sbi->s_ndevs && sbi->segs_per_sec == 1)
+ SIT_I(sbi)->last_victim[ALLOC_NEXT] =
+ GET_SEGNO(sbi, FDEV(0).end_blk) + 1;
}