erofs: introduce partial-referenced pclusters

Due to deduplication for compressed data, pclusters can be partially
referenced with their prefixes.

Together with the user-space implementation, it enables EROFS
variable-length global compressed data deduplication with rolling
hash.

Link: https://lore.kernel.org/r/20220923014915.4362-1-hsiangkao@linux.alibaba.com
Reviewed-by: Yue Hu <huyue2@coolpad.com>
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
This commit is contained in:
Gao Xiang 2022-09-23 09:49:15 +08:00
parent b15b2e307c
commit 5c2a64252c
7 changed files with 23 additions and 2 deletions

View File

@ -217,6 +217,9 @@ again:
strm->buf.out_size = min_t(u32, outlen,
PAGE_SIZE - pageofs);
outlen -= strm->buf.out_size;
if (!rq->out[no] && rq->fillgaps) /* deduped */
rq->out[no] = erofs_allocpage(pagepool,
GFP_KERNEL | __GFP_NOFAIL);
if (rq->out[no])
strm->buf.out = kmap(rq->out[no]) + pageofs;
pageofs = 0;

View File

@ -26,6 +26,7 @@
#define EROFS_FEATURE_INCOMPAT_COMPR_HEAD2 0x00000008
#define EROFS_FEATURE_INCOMPAT_ZTAILPACKING 0x00000010
#define EROFS_FEATURE_INCOMPAT_FRAGMENTS 0x00000020
#define EROFS_FEATURE_INCOMPAT_DEDUPE 0x00000020
#define EROFS_ALL_FEATURE_INCOMPAT \
(EROFS_FEATURE_INCOMPAT_ZERO_PADDING | \
EROFS_FEATURE_INCOMPAT_COMPR_CFGS | \
@ -34,7 +35,8 @@
EROFS_FEATURE_INCOMPAT_DEVICE_TABLE | \
EROFS_FEATURE_INCOMPAT_COMPR_HEAD2 | \
EROFS_FEATURE_INCOMPAT_ZTAILPACKING | \
EROFS_FEATURE_INCOMPAT_FRAGMENTS)
EROFS_FEATURE_INCOMPAT_FRAGMENTS | \
EROFS_FEATURE_INCOMPAT_DEDUPE)
#define EROFS_SB_EXTSLOT_SIZE 16
@ -371,6 +373,9 @@ enum {
#define Z_EROFS_VLE_DI_CLUSTER_TYPE_BITS 2
#define Z_EROFS_VLE_DI_CLUSTER_TYPE_BIT 0
/* (noncompact only, HEAD) This pcluster refers to partial decompressed data */
#define Z_EROFS_VLE_DI_PARTIAL_REF (1 << 15)
/*
* D0_CBLKCNT will be marked _only_ at the 1st non-head lcluster to store the
* compressed block count of a compressed extent (in logical clusters, aka.

View File

@ -291,6 +291,7 @@ EROFS_FEATURE_FUNCS(device_table, incompat, INCOMPAT_DEVICE_TABLE)
EROFS_FEATURE_FUNCS(compr_head2, incompat, INCOMPAT_COMPR_HEAD2)
EROFS_FEATURE_FUNCS(ztailpacking, incompat, INCOMPAT_ZTAILPACKING)
EROFS_FEATURE_FUNCS(fragments, incompat, INCOMPAT_FRAGMENTS)
EROFS_FEATURE_FUNCS(dedupe, incompat, INCOMPAT_DEDUPE)
EROFS_FEATURE_FUNCS(sb_chksum, compat, COMPAT_SB_CHKSUM)
/* atomic flag definitions */
@ -392,6 +393,7 @@ enum {
BH_Encoded = BH_PrivateStart,
BH_FullMapped,
BH_Fragment,
BH_Partialref,
};
/* Has a disk mapping */
@ -404,6 +406,8 @@ enum {
#define EROFS_MAP_FULL_MAPPED (1 << BH_FullMapped)
/* Located in the special packed inode */
#define EROFS_MAP_FRAGMENT (1 << BH_Fragment)
/* The extent refers to partial decompressed data */
#define EROFS_MAP_PARTIAL_REF (1 << BH_Partialref)
struct erofs_map_blocks {
struct erofs_buf buf;

View File

@ -424,6 +424,8 @@ static int erofs_read_superblock(struct super_block *sb)
erofs_info(sb, "EXPERIMENTAL fscache-based on-demand read feature in use. Use at your own risk!");
if (erofs_sb_has_fragments(sbi))
erofs_info(sb, "EXPERIMENTAL compressed fragments feature in use. Use at your own risk!");
if (erofs_sb_has_dedupe(sbi))
erofs_info(sb, "EXPERIMENTAL global deduplication feature in use. Use at your own risk!");
out:
erofs_put_metabuf(&buf);
return ret;

View File

@ -77,6 +77,7 @@ EROFS_ATTR_FEATURE(compr_head2);
EROFS_ATTR_FEATURE(sb_chksum);
EROFS_ATTR_FEATURE(ztailpacking);
EROFS_ATTR_FEATURE(fragments);
EROFS_ATTR_FEATURE(dedupe);
static struct attribute *erofs_feat_attrs[] = {
ATTR_LIST(zero_padding),
@ -88,6 +89,7 @@ static struct attribute *erofs_feat_attrs[] = {
ATTR_LIST(sb_chksum),
ATTR_LIST(ztailpacking),
ATTR_LIST(fragments),
ATTR_LIST(dedupe),
NULL,
};
ATTRIBUTE_GROUPS(erofs_feat);

View File

@ -814,6 +814,7 @@ retry:
fe->pcl->multibases = true;
if ((map->m_flags & EROFS_MAP_FULL_MAPPED) &&
!(map->m_flags & EROFS_MAP_PARTIAL_REF) &&
fe->pcl->length == map->m_llen)
fe->pcl->partial = false;
if (fe->pcl->length < offset + end - map->m_la) {

View File

@ -163,6 +163,7 @@ struct z_erofs_maprecorder {
u16 delta[2];
erofs_blk_t pblk, compressedblks;
erofs_off_t nextpackoff;
bool partialref;
};
static int z_erofs_reload_indexes(struct z_erofs_maprecorder *m,
@ -221,6 +222,8 @@ static int legacy_load_cluster_from_disk(struct z_erofs_maprecorder *m,
case Z_EROFS_VLE_CLUSTER_TYPE_PLAIN:
case Z_EROFS_VLE_CLUSTER_TYPE_HEAD1:
case Z_EROFS_VLE_CLUSTER_TYPE_HEAD2:
if (advise & Z_EROFS_VLE_DI_PARTIAL_REF)
m->partialref = true;
m->clusterofs = le16_to_cpu(di->di_clusterofs);
m->pblk = le32_to_cpu(di->di_u.blkaddr);
break;
@ -684,7 +687,8 @@ static int z_erofs_do_map_blocks(struct inode *inode,
err = -EOPNOTSUPP;
goto unmap_out;
}
if (m.partialref)
map->m_flags |= EROFS_MAP_PARTIAL_REF;
map->m_llen = end - map->m_la;
if (flags & EROFS_GET_BLOCKS_FINDTAIL) {