1 files changed, 127 insertions, 87 deletions

diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index 18d7fd1a5064..0bb66927e3d0 100644
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -17,7 +17,7 @@ int z_erofs_fill_inode(struct inode *inode)
 	struct erofs_sb_info *sbi = EROFS_SB(inode->i_sb);
 
 	if (!erofs_sb_has_big_pcluster(sbi) &&
-	    !erofs_sb_has_ztailpacking(sbi) &&
+	    !erofs_sb_has_ztailpacking(sbi) && !erofs_sb_has_fragments(sbi) &&
 	    vi->datalayout == EROFS_INODE_FLAT_COMPRESSION_LEGACY) {
 		vi->z_advise = 0;
 		vi->z_algorithmtype[0] = 0;
@@ -55,20 +55,25 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
 	if (test_bit(EROFS_I_Z_INITED_BIT, &vi->flags))
 		goto out_unlock;
 
-	DBG_BUGON(!erofs_sb_has_big_pcluster(EROFS_SB(sb)) &&
-		  !erofs_sb_has_ztailpacking(EROFS_SB(sb)) &&
-		  vi->datalayout == EROFS_INODE_FLAT_COMPRESSION_LEGACY);
-
 	pos = ALIGN(iloc(EROFS_SB(sb), vi->nid) + vi->inode_isize +
 		    vi->xattr_isize, 8);
-	kaddr = erofs_read_metabuf(&buf, sb, erofs_blknr(pos),
-				   EROFS_KMAP_ATOMIC);
+	kaddr = erofs_read_metabuf(&buf, sb, erofs_blknr(pos), EROFS_KMAP);
 	if (IS_ERR(kaddr)) {
 		err = PTR_ERR(kaddr);
 		goto out_unlock;
 	}
 
 	h = kaddr + erofs_blkoff(pos);
+	/*
+	 * if the highest bit of the 8-byte map header is set, the whole file
+	 * is stored in the packed inode. The rest bits keeps z_fragmentoff.
+	 */
+	if (h->h_clusterbits >> Z_EROFS_FRAGMENT_INODE_BIT) {
+		vi->z_advise = Z_EROFS_ADVISE_FRAGMENT_PCLUSTER;
+		vi->z_fragmentoff = le64_to_cpu(*(__le64 *)h) ^ (1ULL << 63);
+		vi->z_tailextent_headlcn = 0;
+		goto done;
+	}
 	vi->z_advise = le16_to_cpu(h->h_advise);
 	vi->z_algorithmtype[0] = h->h_algorithmtype & 15;
 	vi->z_algorithmtype[1] = h->h_algorithmtype >> 4;
@@ -79,7 +84,7 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
 		erofs_err(sb, "unknown HEAD%u format %u for nid %llu, please upgrade kernel",
 			  headnr + 1, vi->z_algorithmtype[headnr], vi->nid);
 		err = -EOPNOTSUPP;
-		goto unmap_done;
+		goto out_put_metabuf;
 	}
 
 	vi->z_logical_clusterbits = LOG_BLOCK_SIZE + (h->h_clusterbits & 7);
@@ -89,7 +94,7 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
 		erofs_err(sb, "per-inode big pcluster without sb feature for nid %llu",
 			  vi->nid);
 		err = -EFSCORRUPTED;
-		goto unmap_done;
+		goto out_put_metabuf;
 	}
 	if (vi->datalayout == EROFS_INODE_FLAT_COMPRESSION &&
 	    !(vi->z_advise & Z_EROFS_ADVISE_BIG_PCLUSTER_1) ^
@@ -97,12 +102,8 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
 		erofs_err(sb, "big pcluster head1/2 of compact indexes should be consistent for nid %llu",
 			  vi->nid);
 		err = -EFSCORRUPTED;
-		goto unmap_done;
+		goto out_put_metabuf;
 	}
-unmap_done:
-	erofs_put_metabuf(&buf);
-	if (err)
-		goto out_unlock;
 
 	if (vi->z_advise & Z_EROFS_ADVISE_INLINE_PCLUSTER) {
 		struct erofs_map_blocks map = {
@@ -121,11 +122,28 @@ unmap_done:
 			err = -EFSCORRUPTED;
 		}
 		if (err < 0)
-			goto out_unlock;
+			goto out_put_metabuf;
 	}
+
+	if (vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER &&
+	    !(h->h_clusterbits >> Z_EROFS_FRAGMENT_INODE_BIT)) {
+		struct erofs_map_blocks map = {
+			.buf = __EROFS_BUF_INITIALIZER
+		};
+
+		vi->z_fragmentoff = le32_to_cpu(h->h_fragmentoff);
+		err = z_erofs_do_map_blocks(inode, &map,
+					    EROFS_GET_BLOCKS_FINDTAIL);
+		erofs_put_metabuf(&map.buf);
+		if (err < 0)
+			goto out_put_metabuf;
+	}
+done:
 	/* paired with smp_mb() at the beginning of the function */
 	smp_mb();
 	set_bit(EROFS_I_Z_INITED_BIT, &vi->flags);
+out_put_metabuf:
+	erofs_put_metabuf(&buf);
 out_unlock:
 	clear_and_wake_up_bit(EROFS_I_BL_Z_BIT, &vi->flags);
 	return err;
@@ -141,22 +159,11 @@ struct z_erofs_maprecorder {
 	u8  type, headtype;
 	u16 clusterofs;
 	u16 delta[2];
-	erofs_blk_t pblk, compressedlcs;
+	erofs_blk_t pblk, compressedblks;
 	erofs_off_t nextpackoff;
+	bool partialref;
 };
 
-static int z_erofs_reload_indexes(struct z_erofs_maprecorder *m,
-				  erofs_blk_t eblk)
-{
-	struct super_block *const sb = m->inode->i_sb;
-
-	m->kaddr = erofs_read_metabuf(&m->map->buf, sb, eblk,
-				      EROFS_KMAP_ATOMIC);
-	if (IS_ERR(m->kaddr))
-		return PTR_ERR(m->kaddr);
-	return 0;
-}
-
 static int legacy_load_cluster_from_disk(struct z_erofs_maprecorder *m,
 					 unsigned long lcn)
 {
@@ -169,11 +176,11 @@ static int legacy_load_cluster_from_disk(struct z_erofs_maprecorder *m,
 		lcn * sizeof(struct z_erofs_vle_decompressed_index);
 	struct z_erofs_vle_decompressed_index *di;
 	unsigned int advise, type;
-	int err;
 
-	err = z_erofs_reload_indexes(m, erofs_blknr(pos));
-	if (err)
-		return err;
+	m->kaddr = erofs_read_metabuf(&m->map->buf, inode->i_sb,
+				      erofs_blknr(pos), EROFS_KMAP_ATOMIC);
+	if (IS_ERR(m->kaddr))
+		return PTR_ERR(m->kaddr);
 
 	m->nextpackoff = pos + sizeof(struct z_erofs_vle_decompressed_index);
 	m->lcn = lcn;
@@ -192,7 +199,7 @@ static int legacy_load_cluster_from_disk(struct z_erofs_maprecorder *m,
 				DBG_BUGON(1);
 				return -EFSCORRUPTED;
 			}
-			m->compressedlcs = m->delta[0] &
+			m->compressedblks = m->delta[0] &
 				~Z_EROFS_VLE_DI_D0_CBLKCNT;
 			m->delta[0] = 1;
 		}
@@ -201,6 +208,8 @@ 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;
@@ -293,7 +302,7 @@ static int unpack_compacted_index(struct z_erofs_maprecorder *m,
 				DBG_BUGON(1);
 				return -EFSCORRUPTED;
 			}
-			m->compressedlcs = lo & ~Z_EROFS_VLE_DI_D0_CBLKCNT;
+			m->compressedblks = lo & ~Z_EROFS_VLE_DI_D0_CBLKCNT;
 			m->delta[0] = 1;
 			return 0;
 		} else if (i + 1 != (int)vcnt) {
@@ -370,7 +379,6 @@ static int compacted_load_cluster_from_disk(struct z_erofs_maprecorder *m,
 	unsigned int compacted_4b_initial, compacted_2b;
 	unsigned int amortizedshift;
 	erofs_off_t pos;
-	int err;
 
 	if (lclusterbits != 12)
 		return -EOPNOTSUPP;
@@ -407,9 +415,10 @@ static int compacted_load_cluster_from_disk(struct z_erofs_maprecorder *m,
 	amortizedshift = 2;
 out:
 	pos += lcn * (1 << amortizedshift);
-	err = z_erofs_reload_indexes(m, erofs_blknr(pos));
-	if (err)
-		return err;
+	m->kaddr = erofs_read_metabuf(&m->map->buf, inode->i_sb,
+				      erofs_blknr(pos), EROFS_KMAP_ATOMIC);
+	if (IS_ERR(m->kaddr))
+		return PTR_ERR(m->kaddr);
 	return unpack_compacted_index(m, amortizedshift, pos, lookahead);
 }
 
@@ -431,48 +440,47 @@ static int z_erofs_extent_lookback(struct z_erofs_maprecorder *m,
 				   unsigned int lookback_distance)
 {
 	struct erofs_inode *const vi = EROFS_I(m->inode);
-	struct erofs_map_blocks *const map = m->map;
 	const unsigned int lclusterbits = vi->z_logical_clusterbits;
-	unsigned long lcn = m->lcn;
-	int err;
 
-	if (lcn < lookback_distance) {
-		erofs_err(m->inode->i_sb,
-			  "bogus lookback distance @ nid %llu", vi->nid);
-		DBG_BUGON(1);
-		return -EFSCORRUPTED;
-	}
+	while (m->lcn >= lookback_distance) {
+		unsigned long lcn = m->lcn - lookback_distance;
+		int err;
 
-	/* load extent head logical cluster if needed */
-	lcn -= lookback_distance;
-	err = z_erofs_load_cluster_from_disk(m, lcn, false);
-	if (err)
-		return err;
+		/* load extent head logical cluster if needed */
+		err = z_erofs_load_cluster_from_disk(m, lcn, false);
+		if (err)
+			return err;
 
-	switch (m->type) {
-	case Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD:
-		if (!m->delta[0]) {
+		switch (m->type) {
+		case Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD:
+			if (!m->delta[0]) {
+				erofs_err(m->inode->i_sb,
+					  "invalid lookback distance 0 @ nid %llu",
+					  vi->nid);
+				DBG_BUGON(1);
+				return -EFSCORRUPTED;
+			}
+			lookback_distance = m->delta[0];
+			continue;
+		case Z_EROFS_VLE_CLUSTER_TYPE_PLAIN:
+		case Z_EROFS_VLE_CLUSTER_TYPE_HEAD1:
+		case Z_EROFS_VLE_CLUSTER_TYPE_HEAD2:
+			m->headtype = m->type;
+			m->map->m_la = (lcn << lclusterbits) | m->clusterofs;
+			return 0;
+		default:
 			erofs_err(m->inode->i_sb,
-				  "invalid lookback distance 0 @ nid %llu",
-				  vi->nid);
+				  "unknown type %u @ lcn %lu of nid %llu",
+				  m->type, lcn, vi->nid);
 			DBG_BUGON(1);
-			return -EFSCORRUPTED;
+			return -EOPNOTSUPP;
 		}
-		return z_erofs_extent_lookback(m, m->delta[0]);
-	case Z_EROFS_VLE_CLUSTER_TYPE_PLAIN:
-	case Z_EROFS_VLE_CLUSTER_TYPE_HEAD1:
-	case Z_EROFS_VLE_CLUSTER_TYPE_HEAD2:
-		m->headtype = m->type;
-		map->m_la = (lcn << lclusterbits) | m->clusterofs;
-		break;
-	default:
-		erofs_err(m->inode->i_sb,
-			  "unknown type %u @ lcn %lu of nid %llu",
-			  m->type, lcn, vi->nid);
-		DBG_BUGON(1);
-		return -EOPNOTSUPP;
 	}
-	return 0;
+
+	erofs_err(m->inode->i_sb, "bogus lookback distance @ nid %llu",
+		  vi->nid);
+	DBG_BUGON(1);
+	return -EFSCORRUPTED;
 }
 
 static int z_erofs_get_extent_compressedlen(struct z_erofs_maprecorder *m,
@@ -494,11 +502,11 @@ static int z_erofs_get_extent_compressedlen(struct z_erofs_maprecorder *m,
 	     !(vi->z_advise & Z_EROFS_ADVISE_BIG_PCLUSTER_1)) ||
 	    ((m->headtype == Z_EROFS_VLE_CLUSTER_TYPE_HEAD2) &&
 	     !(vi->z_advise & Z_EROFS_ADVISE_BIG_PCLUSTER_2))) {
-		map->m_plen = 1 << lclusterbits;
+		map->m_plen = 1ULL << lclusterbits;
 		return 0;
 	}
 	lcn = m->lcn + 1;
-	if (m->compressedlcs)
+	if (m->compressedblks)
 		goto out;
 
 	err = z_erofs_load_cluster_from_disk(m, lcn, false);
@@ -507,7 +515,7 @@ static int z_erofs_get_extent_compressedlen(struct z_erofs_maprecorder *m,
 
 	/*
 	 * If the 1st NONHEAD lcluster has already been handled initially w/o
-	 * valid compressedlcs, which means at least it mustn't be CBLKCNT, or
+	 * valid compressedblks, which means at least it mustn't be CBLKCNT, or
 	 * an internal implemenatation error is detected.
 	 *
 	 * The following code can also handle it properly anyway, but let's
@@ -524,12 +532,12 @@ static int z_erofs_get_extent_compressedlen(struct z_erofs_maprecorder *m,
 		 * if the 1st NONHEAD lcluster is actually PLAIN or HEAD type
 		 * rather than CBLKCNT, it's a 1 lcluster-sized pcluster.
 		 */
-		m->compressedlcs = 1;
+		m->compressedblks = 1 << (lclusterbits - LOG_BLOCK_SIZE);
 		break;
 	case Z_EROFS_VLE_CLUSTER_TYPE_NONHEAD:
 		if (m->delta[0] != 1)
 			goto err_bonus_cblkcnt;
-		if (m->compressedlcs)
+		if (m->compressedblks)
 			break;
 		fallthrough;
 	default:
@@ -540,7 +548,7 @@ static int z_erofs_get_extent_compressedlen(struct z_erofs_maprecorder *m,
 		return -EFSCORRUPTED;
 	}
 out:
-	map->m_plen = m->compressedlcs << lclusterbits;
+	map->m_plen = (u64)m->compressedblks << LOG_BLOCK_SIZE;
 	return 0;
 err_bonus_cblkcnt:
 	erofs_err(m->inode->i_sb,
@@ -599,6 +607,7 @@ static int z_erofs_do_map_blocks(struct inode *inode,
 {
 	struct erofs_inode *const vi = EROFS_I(inode);
 	bool ztailpacking = vi->z_advise & Z_EROFS_ADVISE_INLINE_PCLUSTER;
+	bool fragment = vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER;
 	struct z_erofs_maprecorder m = {
 		.inode = inode,
 		.map = map,
@@ -630,6 +639,13 @@ static int z_erofs_do_map_blocks(struct inode *inode,
 		if (endoff >= m.clusterofs) {
 			m.headtype = m.type;
 			map->m_la = (m.lcn << lclusterbits) | m.clusterofs;
+			/*
+			 * For ztailpacking files, in order to inline data more
+			 * effectively, special EOF lclusters are now supported
+			 * which can have three parts at most.
+			 */
+			if (ztailpacking && end > inode->i_size)
+				end = inode->i_size;
 			break;
 		}
 		/* m.lcn should be >= 1 if endoff < m.clusterofs */
@@ -657,15 +673,23 @@ 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)
+	if (flags & EROFS_GET_BLOCKS_FINDTAIL) {
 		vi->z_tailextent_headlcn = m.lcn;
+		/* for non-compact indexes, fragmentoff is 64 bits */
+		if (fragment &&
+		    vi->datalayout == EROFS_INODE_FLAT_COMPRESSION_LEGACY)
+			vi->z_fragmentoff |= (u64)m.pblk << 32;
+	}
 	if (ztailpacking && m.lcn == vi->z_tailextent_headlcn) {
 		map->m_flags |= EROFS_MAP_META;
 		map->m_pa = vi->z_idataoff;
 		map->m_plen = vi->z_idata_size;
+	} else if (fragment && m.lcn == vi->z_tailextent_headlcn) {
+		map->m_flags |= EROFS_MAP_FRAGMENT;
 	} else {
 		map->m_pa = blknr_to_addr(m.pblk);
 		err = z_erofs_get_extent_compressedlen(&m, initial_lcn);
@@ -673,12 +697,18 @@ static int z_erofs_do_map_blocks(struct inode *inode,
 			goto out;
 	}
 
-	if (m.headtype == Z_EROFS_VLE_CLUSTER_TYPE_PLAIN)
-		map->m_algorithmformat = Z_EROFS_COMPRESSION_SHIFTED;
-	else if (m.headtype == Z_EROFS_VLE_CLUSTER_TYPE_HEAD2)
+	if (m.headtype == Z_EROFS_VLE_CLUSTER_TYPE_PLAIN) {
+		if (vi->z_advise & Z_EROFS_ADVISE_INTERLACED_PCLUSTER)
+			map->m_algorithmformat =
+				Z_EROFS_COMPRESSION_INTERLACED;
+		else
+			map->m_algorithmformat =
+				Z_EROFS_COMPRESSION_SHIFTED;
+	} else if (m.headtype == Z_EROFS_VLE_CLUSTER_TYPE_HEAD2) {
 		map->m_algorithmformat = vi->z_algorithmtype[1];
-	else
+	} else {
 		map->m_algorithmformat = vi->z_algorithmtype[0];
+	}
 
 	if ((flags & EROFS_GET_BLOCKS_FIEMAP) ||
 	    ((flags & EROFS_GET_BLOCKS_READMORE) &&
@@ -699,10 +729,10 @@ out:
 	return err;
 }
 
-int z_erofs_map_blocks_iter(struct inode *inode,
-			    struct erofs_map_blocks *map,
+int z_erofs_map_blocks_iter(struct inode *inode, struct erofs_map_blocks *map,
 			    int flags)
 {
+	struct erofs_inode *const vi = EROFS_I(inode);
 	int err = 0;
 
 	trace_z_erofs_map_blocks_iter_enter(inode, map, flags);
@@ -719,6 +749,15 @@ int z_erofs_map_blocks_iter(struct inode *inode,
 	if (err)
 		goto out;
 
+	if ((vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER) &&
+	    !vi->z_tailextent_headlcn) {
+		map->m_la = 0;
+		map->m_llen = inode->i_size;
+		map->m_flags = EROFS_MAP_MAPPED | EROFS_MAP_FULL_MAPPED |
+				EROFS_MAP_FRAGMENT;
+		goto out;
+	}
+
 	err = z_erofs_do_map_blocks(inode, map, flags);
 out:
 	trace_z_erofs_map_blocks_iter_exit(inode, map, flags, err);
@@ -745,7 +784,8 @@ static int z_erofs_iomap_begin_report(struct inode *inode, loff_t offset,
 	iomap->length = map.m_llen;
 	if (map.m_flags & EROFS_MAP_MAPPED) {
 		iomap->type = IOMAP_MAPPED;
-		iomap->addr = map.m_pa;
+		iomap->addr = map.m_flags & EROFS_MAP_FRAGMENT ?
+			      IOMAP_NULL_ADDR : map.m_pa;
 	} else {
 		iomap->type = IOMAP_HOLE;
 		iomap->addr = IOMAP_NULL_ADDR;