linux/fs/f2fs/segment.h

903 lines
27 KiB
C
Raw Normal View History

/* SPDX-License-Identifier: GPL-2.0 */
/*
* fs/f2fs/segment.h
*
* Copyright (c) 2012 Samsung Electronics Co., Ltd.
* http://www.samsung.com/
*/
#include <linux/blkdev.h>
#include <linux/backing-dev.h>
/* constant macro */
#define NULL_SEGNO ((unsigned int)(~0))
#define NULL_SECNO ((unsigned int)(~0))
#define DEF_RECLAIM_PREFREE_SEGMENTS 5 /* 5% over total segments */
#define DEF_MAX_RECLAIM_PREFREE_SEGMENTS 4096 /* 8GB in maximum */
#define F2FS_MIN_SEGMENTS 9 /* SB + 2 (CP + SIT + NAT) + SSA + MAIN */
/* L: Logical segment # in volume, R: Relative segment # in main area */
#define GET_L2R_SEGNO(free_i, segno) ((segno) - (free_i)->start_segno)
#define GET_R2L_SEGNO(free_i, segno) ((segno) + (free_i)->start_segno)
#define IS_DATASEG(t) ((t) <= CURSEG_COLD_DATA)
#define IS_NODESEG(t) ((t) >= CURSEG_HOT_NODE && (t) <= CURSEG_COLD_NODE)
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
static inline void sanity_check_seg_type(struct f2fs_sb_info *sbi,
unsigned short seg_type)
{
f2fs_bug_on(sbi, seg_type >= NR_PERSISTENT_LOG);
}
#define IS_HOT(t) ((t) == CURSEG_HOT_NODE || (t) == CURSEG_HOT_DATA)
#define IS_WARM(t) ((t) == CURSEG_WARM_NODE || (t) == CURSEG_WARM_DATA)
#define IS_COLD(t) ((t) == CURSEG_COLD_NODE || (t) == CURSEG_COLD_DATA)
#define IS_CURSEG(sbi, seg) \
(((seg) == CURSEG_I(sbi, CURSEG_HOT_DATA)->segno) || \
((seg) == CURSEG_I(sbi, CURSEG_WARM_DATA)->segno) || \
((seg) == CURSEG_I(sbi, CURSEG_COLD_DATA)->segno) || \
((seg) == CURSEG_I(sbi, CURSEG_HOT_NODE)->segno) || \
((seg) == CURSEG_I(sbi, CURSEG_WARM_NODE)->segno) || \
((seg) == CURSEG_I(sbi, CURSEG_COLD_NODE)->segno) || \
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
((seg) == CURSEG_I(sbi, CURSEG_COLD_DATA_PINNED)->segno) || \
((seg) == CURSEG_I(sbi, CURSEG_ALL_DATA_ATGC)->segno))
#define IS_CURSEC(sbi, secno) \
(((secno) == CURSEG_I(sbi, CURSEG_HOT_DATA)->segno / \
(sbi)->segs_per_sec) || \
((secno) == CURSEG_I(sbi, CURSEG_WARM_DATA)->segno / \
(sbi)->segs_per_sec) || \
((secno) == CURSEG_I(sbi, CURSEG_COLD_DATA)->segno / \
(sbi)->segs_per_sec) || \
((secno) == CURSEG_I(sbi, CURSEG_HOT_NODE)->segno / \
(sbi)->segs_per_sec) || \
((secno) == CURSEG_I(sbi, CURSEG_WARM_NODE)->segno / \
(sbi)->segs_per_sec) || \
((secno) == CURSEG_I(sbi, CURSEG_COLD_NODE)->segno / \
(sbi)->segs_per_sec) || \
((secno) == CURSEG_I(sbi, CURSEG_COLD_DATA_PINNED)->segno / \
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
(sbi)->segs_per_sec) || \
((secno) == CURSEG_I(sbi, CURSEG_ALL_DATA_ATGC)->segno / \
(sbi)->segs_per_sec))
#define MAIN_BLKADDR(sbi) \
(SM_I(sbi) ? SM_I(sbi)->main_blkaddr : \
le32_to_cpu(F2FS_RAW_SUPER(sbi)->main_blkaddr))
#define SEG0_BLKADDR(sbi) \
(SM_I(sbi) ? SM_I(sbi)->seg0_blkaddr : \
le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment0_blkaddr))
#define MAIN_SEGS(sbi) (SM_I(sbi)->main_segments)
#define MAIN_SECS(sbi) ((sbi)->total_sections)
#define TOTAL_SEGS(sbi) \
(SM_I(sbi) ? SM_I(sbi)->segment_count : \
le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count))
#define TOTAL_BLKS(sbi) (TOTAL_SEGS(sbi) << (sbi)->log_blocks_per_seg)
#define MAX_BLKADDR(sbi) (SEG0_BLKADDR(sbi) + TOTAL_BLKS(sbi))
#define SEGMENT_SIZE(sbi) (1ULL << ((sbi)->log_blocksize + \
(sbi)->log_blocks_per_seg))
#define START_BLOCK(sbi, segno) (SEG0_BLKADDR(sbi) + \
(GET_R2L_SEGNO(FREE_I(sbi), segno) << (sbi)->log_blocks_per_seg))
#define NEXT_FREE_BLKADDR(sbi, curseg) \
(START_BLOCK(sbi, (curseg)->segno) + (curseg)->next_blkoff)
#define GET_SEGOFF_FROM_SEG0(sbi, blk_addr) ((blk_addr) - SEG0_BLKADDR(sbi))
#define GET_SEGNO_FROM_SEG0(sbi, blk_addr) \
(GET_SEGOFF_FROM_SEG0(sbi, blk_addr) >> (sbi)->log_blocks_per_seg)
#define GET_BLKOFF_FROM_SEG0(sbi, blk_addr) \
(GET_SEGOFF_FROM_SEG0(sbi, blk_addr) & ((sbi)->blocks_per_seg - 1))
#define GET_SEGNO(sbi, blk_addr) \
f2fs: introduce DATA_GENERIC_ENHANCE Previously, f2fs_is_valid_blkaddr(, blkaddr, DATA_GENERIC) will check whether @blkaddr locates in main area or not. That check is weak, since the block address in range of main area can point to the address which is not valid in segment info table, and we can not detect such condition, we may suffer worse corruption as system continues running. So this patch introduce DATA_GENERIC_ENHANCE to enhance the sanity check which trigger SIT bitmap check rather than only range check. This patch did below changes as wel: - set SBI_NEED_FSCK in f2fs_is_valid_blkaddr(). - get rid of is_valid_data_blkaddr() to avoid panic if blkaddr is invalid. - introduce verify_fio_blkaddr() to wrap fio {new,old}_blkaddr validation check. - spread blkaddr check in: * f2fs_get_node_info() * __read_out_blkaddrs() * f2fs_submit_page_read() * ra_data_block() * do_recover_data() This patch can fix bug reported from bugzilla below: https://bugzilla.kernel.org/show_bug.cgi?id=203215 https://bugzilla.kernel.org/show_bug.cgi?id=203223 https://bugzilla.kernel.org/show_bug.cgi?id=203231 https://bugzilla.kernel.org/show_bug.cgi?id=203235 https://bugzilla.kernel.org/show_bug.cgi?id=203241 = Update by Jaegeuk Kim = DATA_GENERIC_ENHANCE enhanced to validate block addresses on read/write paths. But, xfstest/generic/446 compalins some generated kernel messages saying invalid bitmap was detected when reading a block. The reaons is, when we get the block addresses from extent_cache, there is no lock to synchronize it from truncating the blocks in parallel. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2019-04-15 07:26:32 +00:00
((!__is_valid_data_blkaddr(blk_addr)) ? \
NULL_SEGNO : GET_L2R_SEGNO(FREE_I(sbi), \
GET_SEGNO_FROM_SEG0(sbi, blk_addr)))
#define BLKS_PER_SEC(sbi) \
((sbi)->segs_per_sec * (sbi)->blocks_per_seg)
#define GET_SEC_FROM_SEG(sbi, segno) \
((segno) / (sbi)->segs_per_sec)
#define GET_SEG_FROM_SEC(sbi, secno) \
((secno) * (sbi)->segs_per_sec)
#define GET_ZONE_FROM_SEC(sbi, secno) \
((secno) / (sbi)->secs_per_zone)
#define GET_ZONE_FROM_SEG(sbi, segno) \
GET_ZONE_FROM_SEC(sbi, GET_SEC_FROM_SEG(sbi, segno))
#define GET_SUM_BLOCK(sbi, segno) \
((sbi)->sm_info->ssa_blkaddr + (segno))
#define GET_SUM_TYPE(footer) ((footer)->entry_type)
#define SET_SUM_TYPE(footer, type) ((footer)->entry_type = (type))
#define SIT_ENTRY_OFFSET(sit_i, segno) \
((segno) % (sit_i)->sents_per_block)
#define SIT_BLOCK_OFFSET(segno) \
((segno) / SIT_ENTRY_PER_BLOCK)
#define START_SEGNO(segno) \
(SIT_BLOCK_OFFSET(segno) * SIT_ENTRY_PER_BLOCK)
#define SIT_BLK_CNT(sbi) \
DIV_ROUND_UP(MAIN_SEGS(sbi), SIT_ENTRY_PER_BLOCK)
#define f2fs_bitmap_size(nr) \
(BITS_TO_LONGS(nr) * sizeof(unsigned long))
#define SECTOR_FROM_BLOCK(blk_addr) \
(((sector_t)blk_addr) << F2FS_LOG_SECTORS_PER_BLOCK)
#define SECTOR_TO_BLOCK(sectors) \
((sectors) >> F2FS_LOG_SECTORS_PER_BLOCK)
/*
* indicate a block allocation direction: RIGHT and LEFT.
* RIGHT means allocating new sections towards the end of volume.
* LEFT means the opposite direction.
*/
enum {
ALLOC_RIGHT = 0,
ALLOC_LEFT
};
/*
* In the victim_sel_policy->alloc_mode, there are two block allocation modes.
* LFS writes data sequentially with cleaning operations.
* SSR (Slack Space Recycle) reuses obsolete space without cleaning operations.
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
* AT_SSR (Age Threshold based Slack Space Recycle) merges fragments into
* fragmented segment which has similar aging degree.
*/
enum {
LFS = 0,
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
SSR,
AT_SSR,
};
/*
* In the victim_sel_policy->gc_mode, there are two gc, aka cleaning, modes.
* GC_CB is based on cost-benefit algorithm.
* GC_GREEDY is based on greedy algorithm.
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
* GC_AT is based on age-threshold algorithm.
*/
enum {
GC_CB = 0,
GC_GREEDY,
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
GC_AT,
ALLOC_NEXT,
FLUSH_DEVICE,
MAX_GC_POLICY,
};
/*
* BG_GC means the background cleaning job.
* FG_GC means the on-demand cleaning job.
* FORCE_FG_GC means on-demand cleaning job in background.
*/
enum {
BG_GC = 0,
FG_GC,
FORCE_FG_GC,
};
/* for a function parameter to select a victim segment */
struct victim_sel_policy {
int alloc_mode; /* LFS or SSR */
int gc_mode; /* GC_CB or GC_GREEDY */
unsigned long *dirty_bitmap; /* dirty segment/section bitmap */
unsigned int max_search; /*
* maximum # of segments/sections
* to search
*/
unsigned int offset; /* last scanned bitmap offset */
unsigned int ofs_unit; /* bitmap search unit */
unsigned int min_cost; /* minimum cost */
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
unsigned long long oldest_age; /* oldest age of segments having the same min cost */
unsigned int min_segno; /* segment # having min. cost */
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
unsigned long long age; /* mtime of GCed section*/
unsigned long long age_threshold;/* age threshold */
};
struct seg_entry {
unsigned int type:6; /* segment type like CURSEG_XXX_TYPE */
unsigned int valid_blocks:10; /* # of valid blocks */
unsigned int ckpt_valid_blocks:10; /* # of valid blocks last cp */
unsigned int padding:6; /* padding */
unsigned char *cur_valid_map; /* validity bitmap of blocks */
#ifdef CONFIG_F2FS_CHECK_FS
unsigned char *cur_valid_map_mir; /* mirror of current valid bitmap */
#endif
/*
* # of valid blocks and the validity bitmap stored in the last
* checkpoint pack. This information is used by the SSR mode.
*/
unsigned char *ckpt_valid_map; /* validity bitmap of blocks last cp */
unsigned char *discard_map;
unsigned long long mtime; /* modification time of the segment */
};
struct sec_entry {
unsigned int valid_blocks; /* # of valid blocks in a section */
};
struct segment_allocation {
void (*allocate_segment)(struct f2fs_sb_info *, int, bool);
};
#define MAX_SKIP_GC_COUNT 16
struct inmem_pages {
struct list_head list;
struct page *page;
block_t old_addr; /* for revoking when fail to commit */
};
struct sit_info {
const struct segment_allocation *s_ops;
block_t sit_base_addr; /* start block address of SIT area */
block_t sit_blocks; /* # of blocks used by SIT area */
block_t written_valid_blocks; /* # of valid blocks in main area */
char *bitmap; /* all bitmaps pointer */
char *sit_bitmap; /* SIT bitmap pointer */
#ifdef CONFIG_F2FS_CHECK_FS
char *sit_bitmap_mir; /* SIT bitmap mirror */
/* bitmap of segments to be ignored by GC in case of errors */
unsigned long *invalid_segmap;
#endif
unsigned int bitmap_size; /* SIT bitmap size */
unsigned long *tmp_map; /* bitmap for temporal use */
unsigned long *dirty_sentries_bitmap; /* bitmap for dirty sentries */
unsigned int dirty_sentries; /* # of dirty sentries */
unsigned int sents_per_block; /* # of SIT entries per block */
struct rw_semaphore sentry_lock; /* to protect SIT cache */
struct seg_entry *sentries; /* SIT segment-level cache */
struct sec_entry *sec_entries; /* SIT section-level cache */
/* for cost-benefit algorithm in cleaning procedure */
unsigned long long elapsed_time; /* elapsed time after mount */
unsigned long long mounted_time; /* mount time */
unsigned long long min_mtime; /* min. modification time */
unsigned long long max_mtime; /* max. modification time */
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
unsigned long long dirty_min_mtime; /* rerange candidates in GC_AT */
unsigned long long dirty_max_mtime; /* rerange candidates in GC_AT */
unsigned int last_victim[MAX_GC_POLICY]; /* last victim segment # */
};
struct free_segmap_info {
unsigned int start_segno; /* start segment number logically */
unsigned int free_segments; /* # of free segments */
unsigned int free_sections; /* # of free sections */
spinlock_t segmap_lock; /* free segmap lock */
unsigned long *free_segmap; /* free segment bitmap */
unsigned long *free_secmap; /* free section bitmap */
};
/* Notice: The order of dirty type is same with CURSEG_XXX in f2fs.h */
enum dirty_type {
DIRTY_HOT_DATA, /* dirty segments assigned as hot data logs */
DIRTY_WARM_DATA, /* dirty segments assigned as warm data logs */
DIRTY_COLD_DATA, /* dirty segments assigned as cold data logs */
DIRTY_HOT_NODE, /* dirty segments assigned as hot node logs */
DIRTY_WARM_NODE, /* dirty segments assigned as warm node logs */
DIRTY_COLD_NODE, /* dirty segments assigned as cold node logs */
DIRTY, /* to count # of dirty segments */
PRE, /* to count # of entirely obsolete segments */
NR_DIRTY_TYPE
};
struct dirty_seglist_info {
const struct victim_selection *v_ops; /* victim selction operation */
unsigned long *dirty_segmap[NR_DIRTY_TYPE];
unsigned long *dirty_secmap;
struct mutex seglist_lock; /* lock for segment bitmaps */
int nr_dirty[NR_DIRTY_TYPE]; /* # of dirty segments */
unsigned long *victim_secmap; /* background GC victims */
};
/* victim selection function for cleaning and SSR */
struct victim_selection {
int (*get_victim)(struct f2fs_sb_info *, unsigned int *,
f2fs: support age threshold based garbage collection There are several issues in current background GC algorithm: - valid blocks is one of key factors during cost overhead calculation, so if segment has less valid block, however even its age is young or it locates hot segment, CB algorithm will still choose the segment as victim, it's not appropriate. - GCed data/node will go to existing logs, no matter in-there datas' update frequency is the same or not, it may mix hot and cold data again. - GC alloctor mainly use LFS type segment, it will cost free segment more quickly. This patch introduces a new algorithm named age threshold based garbage collection to solve above issues, there are three steps mainly: 1. select a source victim: - set an age threshold, and select candidates beased threshold: e.g. 0 means youngest, 100 means oldest, if we set age threshold to 80 then select dirty segments which has age in range of [80, 100] as candiddates; - set candidate_ratio threshold, and select candidates based the ratio, so that we can shrink candidates to those oldest segments; - select target segment with fewest valid blocks in order to migrate blocks with minimum cost; 2. select a target victim: - select candidates beased age threshold; - set candidate_radius threshold, search candidates whose age is around source victims, searching radius should less than the radius threshold. - select target segment with most valid blocks in order to avoid migrating current target segment. 3. merge valid blocks from source victim into target victim with SSR alloctor. Test steps: - create 160 dirty segments: * half of them have 128 valid blocks per segment * left of them have 384 valid blocks per segment - run background GC Benefit: GC count and block movement count both decrease obviously: - Before: - Valid: 86 - Dirty: 1 - Prefree: 11 - Free: 6001 (6001) GC calls: 162 (BG: 220) - data segments : 160 (160) - node segments : 2 (2) Try to move 41454 blocks (BG: 41454) - data blocks : 40960 (40960) - node blocks : 494 (494) IPU: 0 blocks SSR: 0 blocks in 0 segments LFS: 41364 blocks in 81 segments - After: - Valid: 87 - Dirty: 0 - Prefree: 4 - Free: 6008 (6008) GC calls: 75 (BG: 76) - data segments : 74 (74) - node segments : 1 (1) Try to move 12813 blocks (BG: 12813) - data blocks : 12544 (12544) - node blocks : 269 (269) IPU: 0 blocks SSR: 12032 blocks in 77 segments LFS: 855 blocks in 2 segments Signed-off-by: Chao Yu <yuchao0@huawei.com> [Jaegeuk Kim: fix a bug along with pinfile in-mem segment & clean up] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-08-04 13:14:49 +00:00
int, int, char, unsigned long long);
};
/* for active log information */
struct curseg_info {
struct mutex curseg_mutex; /* lock for consistency */
struct f2fs_summary_block *sum_blk; /* cached summary block */
struct rw_semaphore journal_rwsem; /* protect journal area */
struct f2fs_journal *journal; /* cached journal info */
unsigned char alloc_type; /* current allocation type */
unsigned short seg_type; /* segment type like CURSEG_XXX_TYPE */
unsigned int segno; /* current segment number */
unsigned short next_blkoff; /* next block offset to write */
unsigned int zone; /* current zone number */
unsigned int next_segno; /* preallocated segment */
bool inited; /* indicate inmem log is inited */
};
f2fs: refactor flush_sit_entries codes for reducing SIT writes In commit aec71382c681 ("f2fs: refactor flush_nat_entries codes for reducing NAT writes"), we descripte the issue as below: "Although building NAT journal in cursum reduce the read/write work for NAT block, but previous design leave us lower performance when write checkpoint frequently for these cases: 1. if journal in cursum has already full, it's a bit of waste that we flush all nat entries to page for persistence, but not to cache any entries. 2. if journal in cursum is not full, we fill nat entries to journal util journal is full, then flush the left dirty entries to disk without merge journaled entries, so these journaled entries may be flushed to disk at next checkpoint but lost chance to flushed last time." Actually, we have the same problem in using SIT journal area. In this patch, firstly we will update sit journal with dirty entries as many as possible. Secondly if there is no space in sit journal, we will remove all entries in journal and walk through the whole dirty entry bitmap of sit, accounting dirty sit entries located in same SIT block to sit entry set. All entry sets are linked to list sit_entry_set in sm_info, sorted ascending order by count of entries in set. Later we flush entries in set which have fewest entries into journal as many as we can, and then flush dense set with merged entries to disk. In this way we can use sit journal area more effectively, also we will reduce SIT update, result in gaining in performance and saving lifetime of flash device. In my testing environment, it shows this patch can help to reduce SIT block update obviously. virtual machine + hard disk: fsstress -p 20 -n 400 -l 5 sit page num cp count sit pages/cp based 2006.50 1349.75 1.486 patched 1566.25 1463.25 1.070 Our latency of merging op is small when handling a great number of dirty SIT entries in flush_sit_entries: latency(ns) dirty sit count 36038 2151 49168 2123 37174 2232 Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2014-09-04 10:13:01 +00:00
struct sit_entry_set {
struct list_head set_list; /* link with all sit sets */
unsigned int start_segno; /* start segno of sits in set */
unsigned int entry_cnt; /* the # of sit entries in set */
};
/*
* inline functions
*/
static inline struct curseg_info *CURSEG_I(struct f2fs_sb_info *sbi, int type)
{
return (struct curseg_info *)(SM_I(sbi)->curseg_array + type);
}
static inline struct seg_entry *get_seg_entry(struct f2fs_sb_info *sbi,
unsigned int segno)
{
struct sit_info *sit_i = SIT_I(sbi);
return &sit_i->sentries[segno];
}
static inline struct sec_entry *get_sec_entry(struct f2fs_sb_info *sbi,
unsigned int segno)
{
struct sit_info *sit_i = SIT_I(sbi);
return &sit_i->sec_entries[GET_SEC_FROM_SEG(sbi, segno)];
}
static inline unsigned int get_valid_blocks(struct f2fs_sb_info *sbi,
unsigned int segno, bool use_section)
{
/*
* In order to get # of valid blocks in a section instantly from many
* segments, f2fs manages two counting structures separately.
*/
if (use_section && __is_large_section(sbi))
return get_sec_entry(sbi, segno)->valid_blocks;
else
return get_seg_entry(sbi, segno)->valid_blocks;
}
static inline unsigned int get_ckpt_valid_blocks(struct f2fs_sb_info *sbi,
unsigned int segno)
{
return get_seg_entry(sbi, segno)->ckpt_valid_blocks;
}
static inline void seg_info_from_raw_sit(struct seg_entry *se,
struct f2fs_sit_entry *rs)
{
se->valid_blocks = GET_SIT_VBLOCKS(rs);
se->ckpt_valid_blocks = GET_SIT_VBLOCKS(rs);
memcpy(se->cur_valid_map, rs->valid_map, SIT_VBLOCK_MAP_SIZE);
memcpy(se->ckpt_valid_map, rs->valid_map, SIT_VBLOCK_MAP_SIZE);
#ifdef CONFIG_F2FS_CHECK_FS
memcpy(se->cur_valid_map_mir, rs->valid_map, SIT_VBLOCK_MAP_SIZE);
#endif
se->type = GET_SIT_TYPE(rs);
se->mtime = le64_to_cpu(rs->mtime);
}
f2fs: rebuild sit page from sit info in mem This patch rebuild sit page from sit info in mem instead of issue a read io. I test this method and the result is as below: Pre: mmc_perf_test-12061 [001] ...1 976.819992: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [001] ...1 976.856446: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 998.976946: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 999.023269: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1022.060772: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1022.111034: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [002] ...1 1070.127643: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1070.187352: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1095.942124: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1095.995975: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1122.535091: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1122.586521: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [001] ...1 1147.897487: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [001] ...1 1147.959438: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1177.926951: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [002] ...1 1177.976823: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [002] ...1 1204.176087: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [002] ...1 1204.239046: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit Some sit flush consume more than 50ms. Now: mmc_perf_test-2187 [007] ...1 196.840684: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [007] ...1 196.841258: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [007] ...1 219.430582: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [007] ...1 219.431144: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [002] ...1 243.638678: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [000] ...1 243.638980: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [002] ...1 265.392180: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [002] ...1 265.392245: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [000] ...1 290.309051: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [000] ...1 290.309116: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [003] ...1 317.144209: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [003] ...1 317.145913: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [005] ...1 343.224954: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [005] ...1 343.225574: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [000] ...1 370.239846: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [000] ...1 370.241138: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [001] ...1 397.029043: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [001] ...1 397.030750: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [003] ...1 425.386377: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [003] ...1 425.387735: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit Most sit flush consume no more than 1ms. Signed-off-by: Yunlei He <heyunlei@huawei.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2018-01-25 09:27:11 +00:00
static inline void __seg_info_to_raw_sit(struct seg_entry *se,
struct f2fs_sit_entry *rs)
{
unsigned short raw_vblocks = (se->type << SIT_VBLOCKS_SHIFT) |
se->valid_blocks;
rs->vblocks = cpu_to_le16(raw_vblocks);
memcpy(rs->valid_map, se->cur_valid_map, SIT_VBLOCK_MAP_SIZE);
f2fs: rebuild sit page from sit info in mem This patch rebuild sit page from sit info in mem instead of issue a read io. I test this method and the result is as below: Pre: mmc_perf_test-12061 [001] ...1 976.819992: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [001] ...1 976.856446: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 998.976946: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 999.023269: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1022.060772: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1022.111034: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [002] ...1 1070.127643: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1070.187352: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1095.942124: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1095.995975: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1122.535091: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1122.586521: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [001] ...1 1147.897487: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [001] ...1 1147.959438: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1177.926951: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [002] ...1 1177.976823: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [002] ...1 1204.176087: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [002] ...1 1204.239046: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit Some sit flush consume more than 50ms. Now: mmc_perf_test-2187 [007] ...1 196.840684: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [007] ...1 196.841258: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [007] ...1 219.430582: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [007] ...1 219.431144: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [002] ...1 243.638678: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [000] ...1 243.638980: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [002] ...1 265.392180: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [002] ...1 265.392245: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [000] ...1 290.309051: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [000] ...1 290.309116: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [003] ...1 317.144209: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [003] ...1 317.145913: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [005] ...1 343.224954: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [005] ...1 343.225574: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [000] ...1 370.239846: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [000] ...1 370.241138: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [001] ...1 397.029043: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [001] ...1 397.030750: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [003] ...1 425.386377: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [003] ...1 425.387735: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit Most sit flush consume no more than 1ms. Signed-off-by: Yunlei He <heyunlei@huawei.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2018-01-25 09:27:11 +00:00
rs->mtime = cpu_to_le64(se->mtime);
}
static inline void seg_info_to_sit_page(struct f2fs_sb_info *sbi,
struct page *page, unsigned int start)
{
struct f2fs_sit_block *raw_sit;
struct seg_entry *se;
struct f2fs_sit_entry *rs;
unsigned int end = min(start + SIT_ENTRY_PER_BLOCK,
(unsigned long)MAIN_SEGS(sbi));
int i;
raw_sit = (struct f2fs_sit_block *)page_address(page);
memset(raw_sit, 0, PAGE_SIZE);
f2fs: rebuild sit page from sit info in mem This patch rebuild sit page from sit info in mem instead of issue a read io. I test this method and the result is as below: Pre: mmc_perf_test-12061 [001] ...1 976.819992: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [001] ...1 976.856446: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 998.976946: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 999.023269: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1022.060772: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1022.111034: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [002] ...1 1070.127643: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1070.187352: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1095.942124: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1095.995975: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1122.535091: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [003] ...1 1122.586521: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [001] ...1 1147.897487: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [001] ...1 1147.959438: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [003] ...1 1177.926951: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [002] ...1 1177.976823: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-12061 [002] ...1 1204.176087: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-12061 [002] ...1 1204.239046: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit Some sit flush consume more than 50ms. Now: mmc_perf_test-2187 [007] ...1 196.840684: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [007] ...1 196.841258: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [007] ...1 219.430582: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [007] ...1 219.431144: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [002] ...1 243.638678: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [000] ...1 243.638980: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [002] ...1 265.392180: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [002] ...1 265.392245: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [000] ...1 290.309051: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [000] ...1 290.309116: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [003] ...1 317.144209: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [003] ...1 317.145913: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [005] ...1 343.224954: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [005] ...1 343.225574: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [000] ...1 370.239846: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [000] ...1 370.241138: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [001] ...1 397.029043: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [001] ...1 397.030750: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit mmc_perf_test-2187 [003] ...1 425.386377: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start flush sit mmc_perf_test-2187 [003] ...1 425.387735: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = end flush sit Most sit flush consume no more than 1ms. Signed-off-by: Yunlei He <heyunlei@huawei.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2018-01-25 09:27:11 +00:00
for (i = 0; i < end - start; i++) {
rs = &raw_sit->entries[i];
se = get_seg_entry(sbi, start + i);
__seg_info_to_raw_sit(se, rs);
}
}
static inline void seg_info_to_raw_sit(struct seg_entry *se,
struct f2fs_sit_entry *rs)
{
__seg_info_to_raw_sit(se, rs);
memcpy(se->ckpt_valid_map, rs->valid_map, SIT_VBLOCK_MAP_SIZE);
se->ckpt_valid_blocks = se->valid_blocks;
}
static inline unsigned int find_next_inuse(struct free_segmap_info *free_i,
unsigned int max, unsigned int segno)
{
unsigned int ret;
spin_lock(&free_i->segmap_lock);
ret = find_next_bit(free_i->free_segmap, max, segno);
spin_unlock(&free_i->segmap_lock);
return ret;
}
static inline void __set_free(struct f2fs_sb_info *sbi, unsigned int segno)
{
struct free_segmap_info *free_i = FREE_I(sbi);
unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
unsigned int start_segno = GET_SEG_FROM_SEC(sbi, secno);
unsigned int next;
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
unsigned int usable_segs = f2fs_usable_segs_in_sec(sbi, segno);
spin_lock(&free_i->segmap_lock);
clear_bit(segno, free_i->free_segmap);
free_i->free_segments++;
next = find_next_bit(free_i->free_segmap,
start_segno + sbi->segs_per_sec, start_segno);
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
if (next >= start_segno + usable_segs) {
clear_bit(secno, free_i->free_secmap);
free_i->free_sections++;
}
spin_unlock(&free_i->segmap_lock);
}
static inline void __set_inuse(struct f2fs_sb_info *sbi,
unsigned int segno)
{
struct free_segmap_info *free_i = FREE_I(sbi);
unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
set_bit(segno, free_i->free_segmap);
free_i->free_segments--;
if (!test_and_set_bit(secno, free_i->free_secmap))
free_i->free_sections--;
}
static inline void __set_test_and_free(struct f2fs_sb_info *sbi,
unsigned int segno, bool inmem)
{
struct free_segmap_info *free_i = FREE_I(sbi);
unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
unsigned int start_segno = GET_SEG_FROM_SEC(sbi, secno);
unsigned int next;
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
unsigned int usable_segs = f2fs_usable_segs_in_sec(sbi, segno);
spin_lock(&free_i->segmap_lock);
if (test_and_clear_bit(segno, free_i->free_segmap)) {
free_i->free_segments++;
if (!inmem && IS_CURSEC(sbi, secno))
goto skip_free;
next = find_next_bit(free_i->free_segmap,
start_segno + sbi->segs_per_sec, start_segno);
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
if (next >= start_segno + usable_segs) {
if (test_and_clear_bit(secno, free_i->free_secmap))
free_i->free_sections++;
}
}
skip_free:
spin_unlock(&free_i->segmap_lock);
}
static inline void __set_test_and_inuse(struct f2fs_sb_info *sbi,
unsigned int segno)
{
struct free_segmap_info *free_i = FREE_I(sbi);
unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
spin_lock(&free_i->segmap_lock);
if (!test_and_set_bit(segno, free_i->free_segmap)) {
free_i->free_segments--;
if (!test_and_set_bit(secno, free_i->free_secmap))
free_i->free_sections--;
}
spin_unlock(&free_i->segmap_lock);
}
static inline void get_sit_bitmap(struct f2fs_sb_info *sbi,
void *dst_addr)
{
struct sit_info *sit_i = SIT_I(sbi);
#ifdef CONFIG_F2FS_CHECK_FS
if (memcmp(sit_i->sit_bitmap, sit_i->sit_bitmap_mir,
sit_i->bitmap_size))
f2fs_bug_on(sbi, 1);
#endif
memcpy(dst_addr, sit_i->sit_bitmap, sit_i->bitmap_size);
}
static inline block_t written_block_count(struct f2fs_sb_info *sbi)
{
return SIT_I(sbi)->written_valid_blocks;
}
static inline unsigned int free_segments(struct f2fs_sb_info *sbi)
{
return FREE_I(sbi)->free_segments;
}
static inline int reserved_segments(struct f2fs_sb_info *sbi)
{
return SM_I(sbi)->reserved_segments;
}
static inline unsigned int free_sections(struct f2fs_sb_info *sbi)
{
return FREE_I(sbi)->free_sections;
}
static inline unsigned int prefree_segments(struct f2fs_sb_info *sbi)
{
return DIRTY_I(sbi)->nr_dirty[PRE];
}
static inline unsigned int dirty_segments(struct f2fs_sb_info *sbi)
{
return DIRTY_I(sbi)->nr_dirty[DIRTY_HOT_DATA] +
DIRTY_I(sbi)->nr_dirty[DIRTY_WARM_DATA] +
DIRTY_I(sbi)->nr_dirty[DIRTY_COLD_DATA] +
DIRTY_I(sbi)->nr_dirty[DIRTY_HOT_NODE] +
DIRTY_I(sbi)->nr_dirty[DIRTY_WARM_NODE] +
DIRTY_I(sbi)->nr_dirty[DIRTY_COLD_NODE];
}
static inline int overprovision_segments(struct f2fs_sb_info *sbi)
{
return SM_I(sbi)->ovp_segments;
}
static inline int reserved_sections(struct f2fs_sb_info *sbi)
{
return GET_SEC_FROM_SEG(sbi, (unsigned int)reserved_segments(sbi));
}
static inline bool has_curseg_enough_space(struct f2fs_sb_info *sbi)
{
unsigned int node_blocks = get_pages(sbi, F2FS_DIRTY_NODES) +
get_pages(sbi, F2FS_DIRTY_DENTS);
unsigned int dent_blocks = get_pages(sbi, F2FS_DIRTY_DENTS);
unsigned int segno, left_blocks;
int i;
/* check current node segment */
for (i = CURSEG_HOT_NODE; i <= CURSEG_COLD_NODE; i++) {
segno = CURSEG_I(sbi, i)->segno;
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
left_blocks = f2fs_usable_blks_in_seg(sbi, segno) -
get_seg_entry(sbi, segno)->ckpt_valid_blocks;
if (node_blocks > left_blocks)
return false;
}
/* check current data segment */
segno = CURSEG_I(sbi, CURSEG_HOT_DATA)->segno;
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
left_blocks = f2fs_usable_blks_in_seg(sbi, segno) -
get_seg_entry(sbi, segno)->ckpt_valid_blocks;
if (dent_blocks > left_blocks)
return false;
return true;
}
static inline bool has_not_enough_free_secs(struct f2fs_sb_info *sbi,
int freed, int needed)
{
int node_secs = get_blocktype_secs(sbi, F2FS_DIRTY_NODES);
int dent_secs = get_blocktype_secs(sbi, F2FS_DIRTY_DENTS);
int imeta_secs = get_blocktype_secs(sbi, F2FS_DIRTY_IMETA);
if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
return false;
if (free_sections(sbi) + freed == reserved_sections(sbi) + needed &&
has_curseg_enough_space(sbi))
return false;
return (free_sections(sbi) + freed) <=
(node_secs + 2 * dent_secs + imeta_secs +
reserved_sections(sbi) + needed);
}
static inline bool f2fs_is_checkpoint_ready(struct f2fs_sb_info *sbi)
{
if (likely(!is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
return true;
if (likely(!has_not_enough_free_secs(sbi, 0, 0)))
return true;
return false;
}
static inline bool excess_prefree_segs(struct f2fs_sb_info *sbi)
{
return prefree_segments(sbi) > SM_I(sbi)->rec_prefree_segments;
}
static inline int utilization(struct f2fs_sb_info *sbi)
{
return div_u64((u64)valid_user_blocks(sbi) * 100,
sbi->user_block_count);
}
/*
* Sometimes f2fs may be better to drop out-of-place update policy.
* And, users can control the policy through sysfs entries.
* There are five policies with triggering conditions as follows.
* F2FS_IPU_FORCE - all the time,
* F2FS_IPU_SSR - if SSR mode is activated,
* F2FS_IPU_UTIL - if FS utilization is over threashold,
* F2FS_IPU_SSR_UTIL - if SSR mode is activated and FS utilization is over
* threashold,
* F2FS_IPU_FSYNC - activated in fsync path only for high performance flash
* storages. IPU will be triggered only if the # of dirty
* pages over min_fsync_blocks. (=default option)
* F2FS_IPU_ASYNC - do IPU given by asynchronous write requests.
* F2FS_IPU_NOCACHE - disable IPU bio cache.
* F2FS_IPUT_DISABLE - disable IPU. (=default option in LFS mode)
*/
#define DEF_MIN_IPU_UTIL 70
#define DEF_MIN_FSYNC_BLOCKS 8
#define DEF_MIN_HOT_BLOCKS 16
#define SMALL_VOLUME_SEGMENTS (16 * 512) /* 16GB */
enum {
F2FS_IPU_FORCE,
F2FS_IPU_SSR,
F2FS_IPU_UTIL,
F2FS_IPU_SSR_UTIL,
F2FS_IPU_FSYNC,
F2FS_IPU_ASYNC,
F2FS_IPU_NOCACHE,
};
static inline unsigned int curseg_segno(struct f2fs_sb_info *sbi,
int type)
{
struct curseg_info *curseg = CURSEG_I(sbi, type);
return curseg->segno;
}
static inline unsigned char curseg_alloc_type(struct f2fs_sb_info *sbi,
int type)
{
struct curseg_info *curseg = CURSEG_I(sbi, type);
return curseg->alloc_type;
}
static inline unsigned short curseg_blkoff(struct f2fs_sb_info *sbi, int type)
{
struct curseg_info *curseg = CURSEG_I(sbi, type);
return curseg->next_blkoff;
}
static inline void check_seg_range(struct f2fs_sb_info *sbi, unsigned int segno)
{
f2fs_bug_on(sbi, segno > TOTAL_SEGS(sbi) - 1);
}
f2fs: introduce DATA_GENERIC_ENHANCE Previously, f2fs_is_valid_blkaddr(, blkaddr, DATA_GENERIC) will check whether @blkaddr locates in main area or not. That check is weak, since the block address in range of main area can point to the address which is not valid in segment info table, and we can not detect such condition, we may suffer worse corruption as system continues running. So this patch introduce DATA_GENERIC_ENHANCE to enhance the sanity check which trigger SIT bitmap check rather than only range check. This patch did below changes as wel: - set SBI_NEED_FSCK in f2fs_is_valid_blkaddr(). - get rid of is_valid_data_blkaddr() to avoid panic if blkaddr is invalid. - introduce verify_fio_blkaddr() to wrap fio {new,old}_blkaddr validation check. - spread blkaddr check in: * f2fs_get_node_info() * __read_out_blkaddrs() * f2fs_submit_page_read() * ra_data_block() * do_recover_data() This patch can fix bug reported from bugzilla below: https://bugzilla.kernel.org/show_bug.cgi?id=203215 https://bugzilla.kernel.org/show_bug.cgi?id=203223 https://bugzilla.kernel.org/show_bug.cgi?id=203231 https://bugzilla.kernel.org/show_bug.cgi?id=203235 https://bugzilla.kernel.org/show_bug.cgi?id=203241 = Update by Jaegeuk Kim = DATA_GENERIC_ENHANCE enhanced to validate block addresses on read/write paths. But, xfstest/generic/446 compalins some generated kernel messages saying invalid bitmap was detected when reading a block. The reaons is, when we get the block addresses from extent_cache, there is no lock to synchronize it from truncating the blocks in parallel. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2019-04-15 07:26:32 +00:00
static inline void verify_fio_blkaddr(struct f2fs_io_info *fio)
{
struct f2fs_sb_info *sbi = fio->sbi;
f2fs: introduce DATA_GENERIC_ENHANCE Previously, f2fs_is_valid_blkaddr(, blkaddr, DATA_GENERIC) will check whether @blkaddr locates in main area or not. That check is weak, since the block address in range of main area can point to the address which is not valid in segment info table, and we can not detect such condition, we may suffer worse corruption as system continues running. So this patch introduce DATA_GENERIC_ENHANCE to enhance the sanity check which trigger SIT bitmap check rather than only range check. This patch did below changes as wel: - set SBI_NEED_FSCK in f2fs_is_valid_blkaddr(). - get rid of is_valid_data_blkaddr() to avoid panic if blkaddr is invalid. - introduce verify_fio_blkaddr() to wrap fio {new,old}_blkaddr validation check. - spread blkaddr check in: * f2fs_get_node_info() * __read_out_blkaddrs() * f2fs_submit_page_read() * ra_data_block() * do_recover_data() This patch can fix bug reported from bugzilla below: https://bugzilla.kernel.org/show_bug.cgi?id=203215 https://bugzilla.kernel.org/show_bug.cgi?id=203223 https://bugzilla.kernel.org/show_bug.cgi?id=203231 https://bugzilla.kernel.org/show_bug.cgi?id=203235 https://bugzilla.kernel.org/show_bug.cgi?id=203241 = Update by Jaegeuk Kim = DATA_GENERIC_ENHANCE enhanced to validate block addresses on read/write paths. But, xfstest/generic/446 compalins some generated kernel messages saying invalid bitmap was detected when reading a block. The reaons is, when we get the block addresses from extent_cache, there is no lock to synchronize it from truncating the blocks in parallel. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2019-04-15 07:26:32 +00:00
if (__is_valid_data_blkaddr(fio->old_blkaddr))
verify_blkaddr(sbi, fio->old_blkaddr, __is_meta_io(fio) ?
META_GENERIC : DATA_GENERIC);
verify_blkaddr(sbi, fio->new_blkaddr, __is_meta_io(fio) ?
META_GENERIC : DATA_GENERIC_ENHANCE);
}
/*
* Summary block is always treated as an invalid block
*/
static inline int check_block_count(struct f2fs_sb_info *sbi,
int segno, struct f2fs_sit_entry *raw_sit)
{
bool is_valid = test_bit_le(0, raw_sit->valid_map) ? true : false;
int valid_blocks = 0;
int cur_pos = 0, next_pos;
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
unsigned int usable_blks_per_seg = f2fs_usable_blks_in_seg(sbi, segno);
/* check bitmap with valid block count */
do {
if (is_valid) {
next_pos = find_next_zero_bit_le(&raw_sit->valid_map,
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
usable_blks_per_seg,
cur_pos);
valid_blocks += next_pos - cur_pos;
} else
next_pos = find_next_bit_le(&raw_sit->valid_map,
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
usable_blks_per_seg,
cur_pos);
cur_pos = next_pos;
is_valid = !is_valid;
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
} while (cur_pos < usable_blks_per_seg);
if (unlikely(GET_SIT_VBLOCKS(raw_sit) != valid_blocks)) {
f2fs_err(sbi, "Mismatch valid blocks %d vs. %d",
GET_SIT_VBLOCKS(raw_sit), valid_blocks);
set_sbi_flag(sbi, SBI_NEED_FSCK);
return -EFSCORRUPTED;
}
f2fs: fix to do sanity check on valid block count of segment As Jungyeon reported in bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=203233 - Overview When mounting the attached crafted image and running program, following errors are reported. Additionally, it hangs on sync after running program. The image is intentionally fuzzed from a normal f2fs image for testing. Compile options for F2FS are as follows. CONFIG_F2FS_FS=y CONFIG_F2FS_STAT_FS=y CONFIG_F2FS_FS_XATTR=y CONFIG_F2FS_FS_POSIX_ACL=y CONFIG_F2FS_CHECK_FS=y - Reproduces cc poc_13.c mkdir test mount -t f2fs tmp.img test cp a.out test cd test sudo ./a.out sync - Kernel messages F2FS-fs (sdb): Bitmap was wrongly set, blk:4608 kernel BUG at fs/f2fs/segment.c:2102! RIP: 0010:update_sit_entry+0x394/0x410 Call Trace: f2fs_allocate_data_block+0x16f/0x660 do_write_page+0x62/0x170 f2fs_do_write_node_page+0x33/0xa0 __write_node_page+0x270/0x4e0 f2fs_sync_node_pages+0x5df/0x670 f2fs_write_checkpoint+0x372/0x1400 f2fs_sync_fs+0xa3/0x130 f2fs_do_sync_file+0x1a6/0x810 do_fsync+0x33/0x60 __x64_sys_fsync+0xb/0x10 do_syscall_64+0x43/0xf0 entry_SYSCALL_64_after_hwframe+0x44/0xa9 sit.vblocks and sum valid block count in sit.valid_map may be inconsistent, segment w/ zero vblocks will be treated as free segment, while allocating in free segment, we may allocate a free block, if its bitmap is valid previously, it can cause kernel crash due to bitmap verification failure. Anyway, to avoid further serious metadata inconsistence and corruption, it is necessary and worth to detect SIT inconsistence. So let's enable check_block_count() to verify vblocks and valid_map all the time rather than do it only CONFIG_F2FS_CHECK_FS is enabled. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2019-04-15 07:30:51 +00:00
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
if (usable_blks_per_seg < sbi->blocks_per_seg)
f2fs_bug_on(sbi, find_next_bit_le(&raw_sit->valid_map,
sbi->blocks_per_seg,
usable_blks_per_seg) != sbi->blocks_per_seg);
/* check segment usage, and check boundary of a given segment number */
f2fs: support zone capacity less than zone size NVMe Zoned Namespace devices can have zone-capacity less than zone-size. Zone-capacity indicates the maximum number of sectors that are usable in a zone beginning from the first sector of the zone. This makes the sectors sectors after the zone-capacity till zone-size to be unusable. This patch set tracks zone-size and zone-capacity in zoned devices and calculate the usable blocks per segment and usable segments per section. If zone-capacity is less than zone-size mark only those segments which start before zone-capacity as free segments. All segments at and beyond zone-capacity are treated as permanently used segments. In cases where zone-capacity does not align with segment size the last segment will start before zone-capacity and end beyond the zone-capacity of the zone. For such spanning segments only sectors within the zone-capacity are used. During writes and GC manage the usable segments in a section and usable blocks per segment. Segments which are beyond zone-capacity are never allocated, and do not need to be garbage collected, only the segments which are before zone-capacity needs to garbage collected. For spanning segments based on the number of usable blocks in that segment, write to blocks only up to zone-capacity. Zone-capacity is device specific and cannot be configured by the user. Since NVMe ZNS device zones are sequentially write only, a block device with conventional zones or any normal block device is needed along with the ZNS device for the metadata operations of F2fs. A typical nvme-cli output of a zoned device shows zone start and capacity and write pointer as below: SLBA: 0x0 WP: 0x0 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x20000 WP: 0x20000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ SLBA: 0x40000 WP: 0x40000 Cap: 0x18800 State: EMPTY Type: SEQWRITE_REQ Here zone size is 64MB, capacity is 49MB, WP is at zone start as the zones are in EMPTY state. For each zone, only zone start + 49MB is usable area, any lba/sector after 49MB cannot be read or written to, the drive will fail any attempts to read/write. So, the second zone starts at 64MB and is usable till 113MB (64 + 49) and the range between 113 and 128MB is again unusable. The next zone starts at 128MB, and so on. Signed-off-by: Aravind Ramesh <aravind.ramesh@wdc.com> Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2020-07-16 12:56:56 +00:00
if (unlikely(GET_SIT_VBLOCKS(raw_sit) > usable_blks_per_seg
|| segno > TOTAL_SEGS(sbi) - 1)) {
f2fs_err(sbi, "Wrong valid blocks %d or segno %u",
GET_SIT_VBLOCKS(raw_sit), segno);
set_sbi_flag(sbi, SBI_NEED_FSCK);
return -EFSCORRUPTED;
}
return 0;
}
static inline pgoff_t current_sit_addr(struct f2fs_sb_info *sbi,
unsigned int start)
{
struct sit_info *sit_i = SIT_I(sbi);
unsigned int offset = SIT_BLOCK_OFFSET(start);
block_t blk_addr = sit_i->sit_base_addr + offset;
check_seg_range(sbi, start);
#ifdef CONFIG_F2FS_CHECK_FS
if (f2fs_test_bit(offset, sit_i->sit_bitmap) !=
f2fs_test_bit(offset, sit_i->sit_bitmap_mir))
f2fs_bug_on(sbi, 1);
#endif
/* calculate sit block address */
if (f2fs_test_bit(offset, sit_i->sit_bitmap))
blk_addr += sit_i->sit_blocks;
return blk_addr;
}
static inline pgoff_t next_sit_addr(struct f2fs_sb_info *sbi,
pgoff_t block_addr)
{
struct sit_info *sit_i = SIT_I(sbi);
block_addr -= sit_i->sit_base_addr;
if (block_addr < sit_i->sit_blocks)
block_addr += sit_i->sit_blocks;
else
block_addr -= sit_i->sit_blocks;
return block_addr + sit_i->sit_base_addr;
}
static inline void set_to_next_sit(struct sit_info *sit_i, unsigned int start)
{
unsigned int block_off = SIT_BLOCK_OFFSET(start);
f2fs_change_bit(block_off, sit_i->sit_bitmap);
#ifdef CONFIG_F2FS_CHECK_FS
f2fs_change_bit(block_off, sit_i->sit_bitmap_mir);
#endif
}
static inline unsigned long long get_mtime(struct f2fs_sb_info *sbi,
bool base_time)
{
struct sit_info *sit_i = SIT_I(sbi);
time64_t diff, now = ktime_get_boottime_seconds();
if (now >= sit_i->mounted_time)
return sit_i->elapsed_time + now - sit_i->mounted_time;
/* system time is set to the past */
if (!base_time) {
diff = sit_i->mounted_time - now;
if (sit_i->elapsed_time >= diff)
return sit_i->elapsed_time - diff;
return 0;
}
return sit_i->elapsed_time;
}
static inline void set_summary(struct f2fs_summary *sum, nid_t nid,
unsigned int ofs_in_node, unsigned char version)
{
sum->nid = cpu_to_le32(nid);
sum->ofs_in_node = cpu_to_le16(ofs_in_node);
sum->version = version;
}
static inline block_t start_sum_block(struct f2fs_sb_info *sbi)
{
return __start_cp_addr(sbi) +
le32_to_cpu(F2FS_CKPT(sbi)->cp_pack_start_sum);
}
static inline block_t sum_blk_addr(struct f2fs_sb_info *sbi, int base, int type)
{
return __start_cp_addr(sbi) +
le32_to_cpu(F2FS_CKPT(sbi)->cp_pack_total_block_count)
- (base + 1) + type;
}
static inline bool sec_usage_check(struct f2fs_sb_info *sbi, unsigned int secno)
{
if (IS_CURSEC(sbi, secno) || (sbi->cur_victim_sec == secno))
return true;
return false;
}
/*
* It is very important to gather dirty pages and write at once, so that we can
* submit a big bio without interfering other data writes.
* By default, 512 pages for directory data,
* 512 pages (2MB) * 8 for nodes, and
* 256 pages * 8 for meta are set.
*/
static inline int nr_pages_to_skip(struct f2fs_sb_info *sbi, int type)
{
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 21:13:28 +00:00
if (sbi->sb->s_bdi->wb.dirty_exceeded)
return 0;
if (type == DATA)
return sbi->blocks_per_seg;
else if (type == NODE)
return 8 * sbi->blocks_per_seg;
else if (type == META)
return 8 * BIO_MAX_PAGES;
else
return 0;
}
/*
* When writing pages, it'd better align nr_to_write for segment size.
*/
static inline long nr_pages_to_write(struct f2fs_sb_info *sbi, int type,
struct writeback_control *wbc)
{
long nr_to_write, desired;
if (wbc->sync_mode != WB_SYNC_NONE)
return 0;
nr_to_write = wbc->nr_to_write;
desired = BIO_MAX_PAGES;
if (type == NODE)
desired <<= 1;
wbc->nr_to_write = desired;
return desired - nr_to_write;
}
static inline void wake_up_discard_thread(struct f2fs_sb_info *sbi, bool force)
{
struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
bool wakeup = false;
int i;
if (force)
goto wake_up;
mutex_lock(&dcc->cmd_lock);
for (i = MAX_PLIST_NUM - 1; i >= 0; i--) {
if (i + 1 < dcc->discard_granularity)
break;
if (!list_empty(&dcc->pend_list[i])) {
wakeup = true;
break;
}
}
mutex_unlock(&dcc->cmd_lock);
if (!wakeup || !is_idle(sbi, DISCARD_TIME))
return;
wake_up:
dcc->discard_wake = 1;
wake_up_interruptible_all(&dcc->discard_wait_queue);
}