Merge tag 'erofs-for-6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs

Pull erofs updates from Gao Xiang:
 "There is no outstanding feature for this cycle. The most useful
  changes are SEEK_{DATA,HOLE} support and some decompression
  micro-optimization. Other than those, there are some bugfixes and
  cleanups as usual:

   - Add SEEK_{DATA,HOLE} support

   - Free redundant pclusters if no cached compressed data is valid

   - Add sysfs entry to drop internal caches

   - Several bugfixes & cleanups"

* tag 'erofs-for-6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs:
  erofs: handle NONHEAD !delta[1] lclusters gracefully
  erofs: clarify direct I/O support
  erofs: fix blksize < PAGE_SIZE for file-backed mounts
  erofs: get rid of `buf->kmap_type`
  erofs: fix file-backed mounts over FUSE
  erofs: simplify definition of the log functions
  erofs: add sysfs node to drop internal caches
  erofs: free pclusters if no cached folio is attached
  erofs: sunset `struct erofs_workgroup`
  erofs: move erofs_workgroup operations into zdata.c
  erofs: get rid of erofs_{find,insert}_workgroup
  erofs: add SEEK_{DATA,HOLE} support

9 files changed, 276 insertions, 296 deletions

diff --git a/Documentation/ABI/testing/sysfs-fs-erofs b/Documentation/ABI/testing/sysfs-fs-erofs
index 284224d1b56f..b134146d735b 100644
--- a/Documentation/ABI/testing/sysfs-fs-erofs
+++ b/Documentation/ABI/testing/sysfs-fs-erofs
@@ -16,3 +16,14 @@ Description:	Control strategy of sync decompression:
 		  readahead on atomic contexts only.
 		- 1 (force on): enable for readpage and readahead.
 		- 2 (force off): disable for all situations.
+
+What:		/sys/fs/erofs/<disk>/drop_caches
+Date:		November 2024
+Contact:	"Guo Chunhai" <guochunhai@vivo.com>
+Description:	Writing to this will drop compression-related caches,
+		currently used to drop in-memory pclusters and cached
+		compressed folios:
+
+		- 1 : invalidate cached compressed folios
+		- 2 : drop in-memory pclusters
+		- 3 : drop in-memory pclusters and cached compressed folios
diff --git a/fs/erofs/data.c b/fs/erofs/data.c
index 61debd799cf9..1c49f8962021 100644
--- a/fs/erofs/data.c
+++ b/fs/erofs/data.c
@@ -10,10 +10,10 @@
 
 void erofs_unmap_metabuf(struct erofs_buf *buf)
 {
-	if (buf->kmap_type == EROFS_KMAP)
-		kunmap_local(buf->base);
+	if (!buf->base)
+		return;
+	kunmap_local(buf->base);
 	buf->base = NULL;
-	buf->kmap_type = EROFS_NO_KMAP;
 }
 
 void erofs_put_metabuf(struct erofs_buf *buf)
@@ -38,20 +38,13 @@ void *erofs_bread(struct erofs_buf *buf, erofs_off_t offset,
 	}
 	if (!folio || !folio_contains(folio, index)) {
 		erofs_put_metabuf(buf);
-		folio = read_mapping_folio(buf->mapping, index, NULL);
+		folio = read_mapping_folio(buf->mapping, index, buf->file);
 		if (IS_ERR(folio))
 			return folio;
 	}
 	buf->page = folio_file_page(folio, index);
-
-	if (buf->kmap_type == EROFS_NO_KMAP) {
-		if (type == EROFS_KMAP)
-			buf->base = kmap_local_page(buf->page);
-		buf->kmap_type = type;
-	} else if (buf->kmap_type != type) {
-		DBG_BUGON(1);
-		return ERR_PTR(-EFAULT);
-	}
+	if (!buf->base && type == EROFS_KMAP)
+		buf->base = kmap_local_page(buf->page);
 	if (type == EROFS_NO_KMAP)
 		return NULL;
 	return buf->base + (offset & ~PAGE_MASK);
@@ -61,9 +54,11 @@ void erofs_init_metabuf(struct erofs_buf *buf, struct super_block *sb)
 {
 	struct erofs_sb_info *sbi = EROFS_SB(sb);
 
-	if (erofs_is_fileio_mode(sbi))
-		buf->mapping = file_inode(sbi->fdev)->i_mapping;
-	else if (erofs_is_fscache_mode(sb))
+	buf->file = NULL;
+	if (erofs_is_fileio_mode(sbi)) {
+		buf->file = sbi->fdev;		/* some fs like FUSE needs it */
+		buf->mapping = buf->file->f_mapping;
+	} else if (erofs_is_fscache_mode(sb))
 		buf->mapping = sbi->s_fscache->inode->i_mapping;
 	else
 		buf->mapping = sb->s_bdev->bd_mapping;
@@ -350,7 +345,6 @@ static int erofs_iomap_end(struct inode *inode, loff_t pos, loff_t length,
 		struct erofs_buf buf = {
 			.page = kmap_to_page(ptr),
 			.base = ptr,
-			.kmap_type = EROFS_KMAP,
 		};
 
 		DBG_BUGON(iomap->type != IOMAP_INLINE);
@@ -411,22 +405,9 @@ static ssize_t erofs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
 	if (IS_DAX(inode))
 		return dax_iomap_rw(iocb, to, &erofs_iomap_ops);
 #endif
-	if (iocb->ki_flags & IOCB_DIRECT) {
-		struct block_device *bdev = inode->i_sb->s_bdev;
-		unsigned int blksize_mask;
-
-		if (bdev)
-			blksize_mask = bdev_logical_block_size(bdev) - 1;
-		else
-			blksize_mask = i_blocksize(inode) - 1;
-
-		if ((iocb->ki_pos | iov_iter_count(to) |
-		     iov_iter_alignment(to)) & blksize_mask)
-			return -EINVAL;
-
+	if ((iocb->ki_flags & IOCB_DIRECT) && inode->i_sb->s_bdev)
 		return iomap_dio_rw(iocb, to, &erofs_iomap_ops,
 				    NULL, 0, NULL, 0);
-	}
 	return filemap_read(iocb, to, 0);
 }
 
@@ -473,8 +454,32 @@ static int erofs_file_mmap(struct file *file, struct vm_area_struct *vma)
 #define erofs_file_mmap	generic_file_readonly_mmap
 #endif
 
+static loff_t erofs_file_llseek(struct file *file, loff_t offset, int whence)
+{
+	struct inode *inode = file->f_mapping->host;
+	const struct iomap_ops *ops = &erofs_iomap_ops;
+
+	if (erofs_inode_is_data_compressed(EROFS_I(inode)->datalayout))
+#ifdef CONFIG_EROFS_FS_ZIP
+		ops = &z_erofs_iomap_report_ops;
+#else
+		return generic_file_llseek(file, offset, whence);
+#endif
+
+	if (whence == SEEK_HOLE)
+		offset = iomap_seek_hole(inode, offset, ops);
+	else if (whence == SEEK_DATA)
+		offset = iomap_seek_data(inode, offset, ops);
+	else
+		return generic_file_llseek(file, offset, whence);
+
+	if (offset < 0)
+		return offset;
+	return vfs_setpos(file, offset, inode->i_sb->s_maxbytes);
+}
+
 const struct file_operations erofs_file_fops = {
-	.llseek		= generic_file_llseek,
+	.llseek		= erofs_file_llseek,
 	.read_iter	= erofs_file_read_iter,
 	.mmap		= erofs_file_mmap,
 	.get_unmapped_area = thp_get_unmapped_area,
diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
index db29190656eb..d4b89407822a 100644
--- a/fs/erofs/inode.c
+++ b/fs/erofs/inode.c
@@ -318,6 +318,7 @@ int erofs_getattr(struct mnt_idmap *idmap, const struct path *path,
 		  unsigned int query_flags)
 {
 	struct inode *const inode = d_inode(path->dentry);
+	struct block_device *bdev = inode->i_sb->s_bdev;
 	bool compressed =
 		erofs_inode_is_data_compressed(EROFS_I(inode)->datalayout);
 
@@ -330,15 +331,14 @@ int erofs_getattr(struct mnt_idmap *idmap, const struct path *path,
 	/*
 	 * Return the DIO alignment restrictions if requested.
 	 *
-	 * In EROFS, STATX_DIOALIGN is not supported in ondemand mode and
-	 * compressed files, so in these cases we report no DIO support.
+	 * In EROFS, STATX_DIOALIGN is only supported in bdev-based mode
+	 * and uncompressed inodes, otherwise we report no DIO support.
 	 */
 	if ((request_mask & STATX_DIOALIGN) && S_ISREG(inode->i_mode)) {
 		stat->result_mask |= STATX_DIOALIGN;
-		if (!erofs_is_fscache_mode(inode->i_sb) && !compressed) {
-			stat->dio_mem_align =
-				bdev_logical_block_size(inode->i_sb->s_bdev);
-			stat->dio_offset_align = stat->dio_mem_align;
+		if (bdev && !compressed) {
+			stat->dio_mem_align = bdev_dma_alignment(bdev) + 1;
+			stat->dio_offset_align = bdev_logical_block_size(bdev);
 		}
 	}
 	generic_fillattr(idmap, request_mask, inode, stat);
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index 4efd578d7c62..1c847c30a918 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -20,18 +20,12 @@
 #include <linux/iomap.h>
 #include "erofs_fs.h"
 
-/* redefine pr_fmt "erofs: " */
-#undef pr_fmt
-#define pr_fmt(fmt) "erofs: " fmt
-
-__printf(3, 4) void _erofs_err(struct super_block *sb,
-			       const char *function, const char *fmt, ...);
+__printf(2, 3) void _erofs_printk(struct super_block *sb, const char *fmt, ...);
 #define erofs_err(sb, fmt, ...)	\
-	_erofs_err(sb, __func__, fmt "\n", ##__VA_ARGS__)
-__printf(3, 4) void _erofs_info(struct super_block *sb,
-			       const char *function, const char *fmt, ...);
+	_erofs_printk(sb, KERN_ERR fmt "\n", ##__VA_ARGS__)
 #define erofs_info(sb, fmt, ...) \
-	_erofs_info(sb, __func__, fmt "\n", ##__VA_ARGS__)
+	_erofs_printk(sb, KERN_INFO fmt "\n", ##__VA_ARGS__)
+
 #ifdef CONFIG_EROFS_FS_DEBUG
 #define DBG_BUGON               BUG_ON
 #else
@@ -208,12 +202,6 @@ enum {
 	EROFS_ZIP_CACHE_READAROUND
 };
 
-/* basic unit of the workstation of a super_block */
-struct erofs_workgroup {
-	pgoff_t index;
-	struct lockref lockref;
-};
-
 enum erofs_kmap_type {
 	EROFS_NO_KMAP,		/* don't map the buffer */
 	EROFS_KMAP,		/* use kmap_local_page() to map the buffer */
@@ -221,9 +209,9 @@ enum erofs_kmap_type {
 
 struct erofs_buf {
 	struct address_space *mapping;
+	struct file *file;
 	struct page *page;
 	void *base;
-	enum erofs_kmap_type kmap_type;
 };
 #define __EROFS_BUF_INITIALIZER	((struct erofs_buf){ .page = NULL })
 
@@ -456,20 +444,17 @@ static inline void erofs_pagepool_add(struct page **pagepool, struct page *page)
 void erofs_release_pages(struct page **pagepool);
 
 #ifdef CONFIG_EROFS_FS_ZIP
-void erofs_workgroup_put(struct erofs_workgroup *grp);
-struct erofs_workgroup *erofs_find_workgroup(struct super_block *sb,
-					     pgoff_t index);
-struct erofs_workgroup *erofs_insert_workgroup(struct super_block *sb,
-					       struct erofs_workgroup *grp);
-void erofs_workgroup_free_rcu(struct erofs_workgroup *grp);
+#define MNGD_MAPPING(sbi)	((sbi)->managed_cache->i_mapping)
+
+extern atomic_long_t erofs_global_shrink_cnt;
 void erofs_shrinker_register(struct super_block *sb);
 void erofs_shrinker_unregister(struct super_block *sb);
 int __init erofs_init_shrinker(void);
 void erofs_exit_shrinker(void);
 int __init z_erofs_init_subsystem(void);
 void z_erofs_exit_subsystem(void);
-int erofs_try_to_free_all_cached_folios(struct erofs_sb_info *sbi,
-					struct erofs_workgroup *egrp);
+unsigned long z_erofs_shrink_scan(struct erofs_sb_info *sbi,
+				  unsigned long nr_shrink);
 int z_erofs_map_blocks_iter(struct inode *inode, struct erofs_map_blocks *map,
 			    int flags);
 void *z_erofs_get_gbuf(unsigned int requiredpages);
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index bed3dbe5b7cb..c235a8e4315e 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -18,37 +18,22 @@
 
 static struct kmem_cache *erofs_inode_cachep __read_mostly;
 
-void _erofs_err(struct super_block *sb, const char *func, const char *fmt, ...)
+void _erofs_printk(struct super_block *sb, const char *fmt, ...)
 {
 	struct va_format vaf;
 	va_list args;
+	int level;
 
 	va_start(args, fmt);
 
-	vaf.fmt = fmt;
+	level = printk_get_level(fmt);
+	vaf.fmt = printk_skip_level(fmt);
 	vaf.va = &args;
-
 	if (sb)
-		pr_err("(device %s): %s: %pV", sb->s_id, func, &vaf);
+		printk("%c%cerofs (device %s): %pV",
+				KERN_SOH_ASCII, level, sb->s_id, &vaf);
 	else
-		pr_err("%s: %pV", func, &vaf);
-	va_end(args);
-}
-
-void _erofs_info(struct super_block *sb, const char *func, const char *fmt, ...)
-{
-	struct va_format vaf;
-	va_list args;
-
-	va_start(args, fmt);
-
-	vaf.fmt = fmt;
-	vaf.va = &args;
-
-	if (sb)
-		pr_info("(device %s): %pV", sb->s_id, &vaf);
-	else
-		pr_info("%pV", &vaf);
+		printk("%c%cerofs: %pV", KERN_SOH_ASCII, level, &vaf);
 	va_end(args);
 }
 
@@ -631,7 +616,11 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
 			errorfc(fc, "unsupported blksize for fscache mode");
 			return -EINVAL;
 		}
-		if (!sb_set_blocksize(sb, 1 << sbi->blkszbits)) {
+
+		if (erofs_is_fileio_mode(sbi)) {
+			sb->s_blocksize = 1 << sbi->blkszbits;
+			sb->s_blocksize_bits = sbi->blkszbits;
+		} else if (!sb_set_blocksize(sb, 1 << sbi->blkszbits)) {
 			errorfc(fc, "failed to set erofs blksize");
 			return -EINVAL;
 		}
diff --git a/fs/erofs/sysfs.c b/fs/erofs/sysfs.c
index 63cffd0fd261..19d586273b70 100644
--- a/fs/erofs/sysfs.c
+++ b/fs/erofs/sysfs.c
@@ -10,6 +10,7 @@
 
 enum {
 	attr_feature,
+	attr_drop_caches,
 	attr_pointer_ui,
 	attr_pointer_bool,
 };
@@ -57,11 +58,13 @@ static struct erofs_attr erofs_attr_##_name = {			\
 
 #ifdef CONFIG_EROFS_FS_ZIP
 EROFS_ATTR_RW_UI(sync_decompress, erofs_mount_opts);
+EROFS_ATTR_FUNC(drop_caches, 0200);
 #endif
 
 static struct attribute *erofs_attrs[] = {
 #ifdef CONFIG_EROFS_FS_ZIP
 	ATTR_LIST(sync_decompress),
+	ATTR_LIST(drop_caches),
 #endif
 	NULL,
 };
@@ -163,6 +166,20 @@ static ssize_t erofs_attr_store(struct kobject *kobj, struct attribute *attr,
 			return -EINVAL;
 		*(bool *)ptr = !!t;
 		return len;
+#ifdef CONFIG_EROFS_FS_ZIP
+	case attr_drop_caches:
+		ret = kstrtoul(skip_spaces(buf), 0, &t);
+		if (ret)
+			return ret;
+		if (t < 1 || t > 3)
+			return -EINVAL;
+
+		if (t & 2)
+			z_erofs_shrink_scan(sbi, ~0UL);
+		if (t & 1)
+			invalidate_mapping_pages(MNGD_MAPPING(sbi), 0, -1);
+		return len;
+#endif
 	}
 	return 0;
 }
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index a569ff9dfd04..01f147505487 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -44,12 +44,15 @@ __Z_EROFS_BVSET(z_erofs_bvset_inline, Z_EROFS_INLINE_BVECS);
  * A: Field should be accessed / updated in atomic for parallelized code.
  */
 struct z_erofs_pcluster {
-	struct erofs_workgroup obj;
 	struct mutex lock;
+	struct lockref lockref;
 
 	/* A: point to next chained pcluster or TAILs */
 	z_erofs_next_pcluster_t next;
 
+	/* I: start block address of this pcluster */
+	erofs_off_t index;
+
 	/* L: the maximum decompression size of this round */
 	unsigned int length;
 
@@ -108,7 +111,7 @@ struct z_erofs_decompressqueue {
 
 static inline bool z_erofs_is_inline_pcluster(struct z_erofs_pcluster *pcl)
 {
-	return !pcl->obj.index;
+	return !pcl->index;
 }
 
 static inline unsigned int z_erofs_pclusterpages(struct z_erofs_pcluster *pcl)
@@ -116,7 +119,6 @@ static inline unsigned int z_erofs_pclusterpages(struct z_erofs_pcluster *pcl)
 	return PAGE_ALIGN(pcl->pclustersize) >> PAGE_SHIFT;
 }
 
-#define MNGD_MAPPING(sbi)	((sbi)->managed_cache->i_mapping)
 static bool erofs_folio_is_managed(struct erofs_sb_info *sbi, struct folio *fo)
 {
 	return fo->mapping == MNGD_MAPPING(sbi);
@@ -548,7 +550,7 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
 		if (READ_ONCE(pcl->compressed_bvecs[i].page))
 			continue;
 
-		page = find_get_page(mc, pcl->obj.index + i);
+		page = find_get_page(mc, pcl->index + i);
 		if (!page) {
 			/* I/O is needed, no possible to decompress directly */
 			standalone = false;
@@ -564,13 +566,13 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
 				continue;
 			set_page_private(newpage, Z_EROFS_PREALLOCATED_PAGE);
 		}
-		spin_lock(&pcl->obj.lockref.lock);
+		spin_lock(&pcl->lockref.lock);
 		if (!pcl->compressed_bvecs[i].page) {
 			pcl->compressed_bvecs[i].page = page ? page : newpage;
-			spin_unlock(&pcl->obj.lockref.lock);
+			spin_unlock(&pcl->lockref.lock);
 			continue;
 		}
-		spin_unlock(&pcl->obj.lockref.lock);
+		spin_unlock(&pcl->lockref.lock);
 
 		if (page)
 			put_page(page);
@@ -587,11 +589,9 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
 }
 
 /* (erofs_shrinker) disconnect cached encoded data with pclusters */
-int erofs_try_to_free_all_cached_folios(struct erofs_sb_info *sbi,
-					struct erofs_workgroup *grp)
+static int erofs_try_to_free_all_cached_folios(struct erofs_sb_info *sbi,
+					       struct z_erofs_pcluster *pcl)
 {
-	struct z_erofs_pcluster *const pcl =
-		container_of(grp, struct z_erofs_pcluster, obj);
 	unsigned int pclusterpages = z_erofs_pclusterpages(pcl);
 	struct folio *folio;
 	int i;
@@ -626,8 +626,8 @@ static bool z_erofs_cache_release_folio(struct folio *folio, gfp_t gfp)
 		return true;
 
 	ret = false;
-	spin_lock(&pcl->obj.lockref.lock);
-	if (pcl->obj.lockref.count <= 0) {
+	spin_lock(&pcl->lockref.lock);
+	if (pcl->lockref.count <= 0) {
 		DBG_BUGON(z_erofs_is_inline_pcluster(pcl));
 		for (; bvec < end; ++bvec) {
 			if (bvec->page && page_folio(bvec->page) == folio) {
@@ -638,7 +638,7 @@ static bool z_erofs_cache_release_folio(struct folio *folio, gfp_t gfp)
 			}
 		}
 	}
-	spin_unlock(&pcl->obj.lockref.lock);
+	spin_unlock(&pcl->lockref.lock);
 	return ret;
 }
 
@@ -689,15 +689,15 @@ static int z_erofs_attach_page(struct z_erofs_decompress_frontend *fe,
 
 	if (exclusive) {
 		/* give priority for inplaceio to use file pages first */
-		spin_lock(&pcl->obj.lockref.lock);
+		spin_lock(&pcl->lockref.lock);
 		while (fe->icur > 0) {
 			if (pcl->compressed_bvecs[--fe->icur].page)
 				continue;
 			pcl->compressed_bvecs[fe->icur] = *bvec;
-			spin_unlock(&pcl->obj.lockref.lock);
+			spin_unlock(&pcl->lockref.lock);
 			return 0;
 		}
-		spin_unlock(&pcl->obj.lockref.lock);
+		spin_unlock(&pcl->lockref.lock);
 
 		/* otherwise, check if it can be used as a bvpage */
 		if (fe->mode >= Z_EROFS_PCLUSTER_FOLLOWED &&
@@ -710,13 +710,30 @@ static int z_erofs_attach_page(struct z_erofs_decompress_frontend *fe,
 	return ret;
 }
 
+static bool z_erofs_get_pcluster(struct z_erofs_pcluster *pcl)
+{
+	if (lockref_get_not_zero(&pcl->lockref))
+		return true;
+
+	spin_lock(&pcl->lockref.lock);
+	if (__lockref_is_dead(&pcl->lockref)) {
+		spin_unlock(&pcl->lockref.lock);
+		return false;
+	}
+
+	if (!pcl->lockref.count++)
+		atomic_long_dec(&erofs_global_shrink_cnt);
+	spin_unlock(&pcl->lockref.lock);
+	return true;
+}
+
 static int z_erofs_register_pcluster(struct z_erofs_decompress_frontend *fe)
 {
 	struct erofs_map_blocks *map = &fe->map;
 	struct super_block *sb = fe->inode->i_sb;
+	struct erofs_sb_info *sbi = EROFS_SB(sb);
 	bool ztailpacking = map->m_flags & EROFS_MAP_META;
-	struct z_erofs_pcluster *pcl;
-	struct erofs_workgroup *grp;
+	struct z_erofs_pcluster *pcl, *pre;
 	int err;
 
 	if (!(map->m_flags & EROFS_MAP_ENCODED) ||
@@ -730,8 +747,8 @@ static int z_erofs_register_pcluster(struct z_erofs_decompress_frontend *fe)
 	if (IS_ERR(pcl))
 		return PTR_ERR(pcl);
 
-	spin_lock_init(&pcl->obj.lockref.lock);
-	pcl->obj.lockref.count = 1;	/* one ref for this request */
+	spin_lock_init(&pcl->lockref.lock);
+	pcl->lockref.count = 1;		/* one ref for this request */
 	pcl->algorithmformat = map->m_algorithmformat;
 	pcl->length = 0;
 	pcl->partial = true;
@@ -749,19 +766,26 @@ static int z_erofs_register_pcluster(struct z_erofs_decompress_frontend *fe)
 	DBG_BUGON(!mutex_trylock(&pcl->lock));
 
 	if (ztailpacking) {
-		pcl->obj.index = 0;	/* which indicates ztailpacking */
+		pcl->index = 0;		/* which indicates ztailpacking */
 	} else {
-		pcl->obj.index = erofs_blknr(sb, map->m_pa);
-
-		grp = erofs_insert_workgroup(fe->inode->i_sb, &pcl->obj);
-		if (IS_ERR(grp)) {
-			err = PTR_ERR(grp);
-			goto err_out;
+		pcl->index = erofs_blknr(sb, map->m_pa);
+		while (1) {
+			xa_lock(&sbi->managed_pslots);
+			pre = __xa_cmpxchg(&sbi->managed_pslots, pcl->index,
+					   NULL, pcl, GFP_KERNEL);
+			if (!pre || xa_is_err(pre) || z_erofs_get_pcluster(pre)) {
+				xa_unlock(&sbi->managed_pslots);
+				break;
+			}
+			/* try to legitimize the current in-tree one */
+			xa_unlock(&sbi->managed_pslots);
+			cond_resched();
 		}
-
-		if (grp != &pcl->obj) {
-			fe->pcl = container_of(grp,
-					struct z_erofs_pcluster, obj);
+		if (xa_is_err(pre)) {
+			err = xa_err(pre);
+			goto err_out;
+		} else if (pre) {
+			fe->pcl = pre;
 			err = -EEXIST;
 			goto err_out;
 		}
@@ -781,7 +805,7 @@ static int z_erofs_pcluster_begin(struct z_erofs_decompress_frontend *fe)
 	struct erofs_map_blocks *map = &fe->map;
 	struct super_block *sb = fe->inode->i_sb;
 	erofs_blk_t blknr = erofs_blknr(sb, map->m_pa);
-	struct erofs_workgroup *grp = NULL;
+	struct z_erofs_pcluster *pcl = NULL;
 	int ret;
 
 	DBG_BUGON(fe->pcl);
@@ -789,14 +813,23 @@ static int z_erofs_pcluster_begin(struct z_erofs_decompress_frontend *fe)
 	DBG_BUGON(fe->owned_head == Z_EROFS_PCLUSTER_NIL);
 
 	if (!(map->m_flags & EROFS_MAP_META)) {
-		grp = erofs_find_workgroup(sb, blknr);
+		while (1) {
+			rcu_read_lock();
+			pcl = xa_load(&EROFS_SB(sb)->managed_pslots, blknr);
+			if (!pcl || z_erofs_get_pcluster(pcl)) {
+				DBG_BUGON(pcl && blknr != pcl->index);
+				rcu_read_unlock();
+				break;
+			}
+			rcu_read_unlock();
+		}
 	} else if ((map->m_pa & ~PAGE_MASK) + map->m_plen > PAGE_SIZE) {
 		DBG_BUGON(1);
 		return -EFSCORRUPTED;
 	}
 
-	if (grp) {
-		fe->pcl = container_of(grp, struct z_erofs_pcluster, obj);
+	if (pcl) {
+		fe->pcl = pcl;
 		ret = -EEXIST;
 	} else {
 		ret = z_erofs_register_pcluster(fe);
@@ -851,12 +884,87 @@ static void z_erofs_rcu_callback(struct rcu_head *head)
 			struct z_erofs_pcluster, rcu));
 }
 
-void erofs_workgroup_free_rcu(struct erofs_workgroup *grp)
+static bool __erofs_try_to_release_pcluster(struct erofs_sb_info *sbi,
+					  struct z_erofs_pcluster *pcl)
 {
-	struct z_erofs_pcluster *const pcl =
-		container_of(grp, struct z_erofs_pcluster, obj);
+	if (pcl->lockref.count)
+		return false;
 
-	call_rcu(&pcl->rcu, z_erofs_rcu_callback);
+	/*
+	 * Note that all cached folios should be detached before deleted from
+	 * the XArray.  Otherwise some folios could be still attached to the
+	 * orphan old pcluster when the new one is available in the tree.
+	 */
+	if (erofs_try_to_free_all_cached_folios(sbi, pcl))
+		return false;
+
+	/*
+	 * It's impossible to fail after the pcluster is freezed, but in order
+	 * to avoid some race conditions, add a DBG_BUGON to observe this.
+	 */
+	DBG_BUGON(__xa_erase(&sbi->managed_pslots, pcl->index) != pcl);
+
+	lockref_mark_dead(&pcl->lockref);
+	return true;
+}
+
+static bool erofs_try_to_release_pcluster(struct erofs_sb_info *sbi,
+					  struct z_erofs_pcluster *pcl)
+{
+	bool free;
+
+	spin_lock(&pcl->lockref.lock);
+	free = __erofs_try_to_release_pcluster(sbi, pcl);
+	spin_unlock(&pcl->lockref.lock);
+	if (free) {
+		atomic_long_dec(&erofs_global_shrink_cnt);
+		call_rcu(&pcl->rcu, z_erofs_rcu_callback);
+	}
+	return free;
+}
+
+unsigned long z_erofs_shrink_scan(struct erofs_sb_info *sbi,
+				  unsigned long nr_shrink)
+{
+	struct z_erofs_pcluster *pcl;
+	unsigned int freed = 0;
+	unsigned long index;
+
+	xa_lock(&sbi->managed_pslots);
+	xa_for_each(&sbi->managed_pslots, index, pcl) {
+		/* try to shrink each valid pcluster */
+		if (!erofs_try_to_release_pcluster(sbi, pcl))
+			continue;
+		xa_unlock(&sbi->managed_pslots);
+
+		++freed;
+		if (!--nr_shrink)
+			return freed;
+		xa_lock(&sbi->managed_pslots);
+	}
+	xa_unlock(&sbi->managed_pslots);
+	return freed;
+}
+
+static void z_erofs_put_pcluster(struct erofs_sb_info *sbi,
+		struct z_erofs_pcluster *pcl, bool try_free)
+{
+	bool free = false;
+
+	if (lockref_put_or_lock(&pcl->lockref))
+		return;
+
+	DBG_BUGON(__lockref_is_dead(&pcl->lockref));
+	if (!--pcl->lockref.count) {
+		if (try_free && xa_trylock(&sbi->managed_pslots)) {
+			free = __erofs_try_to_release_pcluster(sbi, pcl);
+			xa_unlock(&sbi->managed_pslots);
+		}
+		atomic_long_add(!free, &erofs_global_shrink_cnt);
+	}
+	spin_unlock(&pcl->lockref.lock);
+	if (free)
+		call_rcu(&pcl->rcu, z_erofs_rcu_callback);
 }
 
 static void z_erofs_pcluster_end(struct z_erofs_decompress_frontend *fe)
@@ -877,7 +985,7 @@ static void z_erofs_pcluster_end(struct z_erofs_decompress_frontend *fe)
 	 * any longer if the pcluster isn't hosted by ourselves.
 	 */
 	if (fe->mode < Z_EROFS_PCLUSTER_FOLLOWED_NOINPLACE)
-		erofs_workgroup_put(&pcl->obj);
+		z_erofs_put_pcluster(EROFS_I_SB(fe->inode), pcl, false);
 
 	fe->pcl = NULL;
 }
@@ -1179,6 +1287,7 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
 	int i, j, jtop, err2;
 	struct page *page;
 	bool overlapped;
+	bool try_free = true;
 
 	mutex_lock(&pcl->lock);
 	be->nr_pages = PAGE_ALIGN(pcl->length + pcl->pageofs_out) >> PAGE_SHIFT;
@@ -1236,9 +1345,12 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
 		/* managed folios are still left in compressed_bvecs[] */
 		for (i = 0; i < pclusterpages; ++i) {
 			page = be->compressed_pages[i];
-			if (!page ||
-			    erofs_folio_is_managed(sbi, page_folio(page)))
+			if (!page)
+				continue;
+			if (erofs_folio_is_managed(sbi, page_folio(page))) {
+				try_free = false;
 				continue;
+			}
 			(void)z_erofs_put_shortlivedpage(be->pagepool, page);
 			WRITE_ONCE(pcl->compressed_bvecs[i].page, NULL);
 		}
@@ -1284,6 +1396,11 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
 	/* pcluster lock MUST be taken before the following line */
 	WRITE_ONCE(pcl->next, Z_EROFS_PCLUSTER_NIL);
 	mutex_unlock(&pcl->lock);
+
+	if (z_erofs_is_inline_pcluster(pcl))
+		z_erofs_free_pcluster(pcl);
+	else
+		z_erofs_put_pcluster(sbi, pcl, try_free);
 	return err;
 }
 
@@ -1306,10 +1423,6 @@ static int z_erofs_decompress_queue(const struct z_erofs_decompressqueue *io,
 		owned = READ_ONCE(be.pcl->next);
 
 		err = z_erofs_decompress_pcluster(&be, err) ?: err;
-		if (z_erofs_is_inline_pcluster(be.pcl))
-			z_erofs_free_pcluster(be.pcl);
-		else
-			erofs_workgroup_put(&be.pcl->obj);
 	}
 	return err;
 }
@@ -1391,9 +1504,9 @@ static void z_erofs_fill_bio_vec(struct bio_vec *bvec,
 	bvec->bv_offset = 0;
 	bvec->bv_len = PAGE_SIZE;
 repeat:
-	spin_lock(&pcl->obj.lockref.lock);
+	spin_lock(&pcl->lockref.lock);
 	zbv = pcl->compressed_bvecs[nr];
-	spin_unlock(&pcl->obj.lockref.lock);
+	spin_unlock(&pcl->lockref.lock);
 	if (!zbv.page)
 		goto out_allocfolio;
 
@@ -1455,23 +1568,23 @@ repeat:
 	folio_put(folio);
 out_allocfolio:
 	page = __erofs_allocpage(&f->pagepool, gfp, true);
-	spin_lock(&pcl->obj.lockref.lock);
+	spin_lock(&pcl->lockref.lock);
 	if (unlikely(pcl->compressed_bvecs[nr].page != zbv.page)) {
 		if (page)
 			erofs_pagepool_add(&f->pagepool, page);
-		spin_unlock(&pcl->obj.lockref.lock);
+		spin_unlock(&pcl->lockref.lock);
 		cond_resched();
 		goto repeat;
 	}
 	pcl->compressed_bvecs[nr].page = page ? page : ERR_PTR(-ENOMEM);
-	spin_unlock(&pcl->obj.lockref.lock);
+	spin_unlock(&pcl->lockref.lock);
 	bvec->bv_page = page;
 	if (!page)
 		return;
 	folio = page_folio(page);
 out_tocache:
 	if (!tocache || bs != PAGE_SIZE ||
-	    filemap_add_folio(mc, folio, pcl->obj.index + nr, gfp)) {
+	    filemap_add_folio(mc, folio, pcl->index + nr, gfp)) {
 		/* turn into a temporary shortlived folio (1 ref) */
 		folio->private = (void *)Z_EROFS_SHORTLIVED_PAGE;
 		return;
@@ -1603,7 +1716,7 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
 
 		/* no device id here, thus it will always succeed */
 		mdev = (struct erofs_map_dev) {
-			.m_pa = erofs_pos(sb, pcl->obj.index),
+			.m_pa = erofs_pos(sb, pcl->index),
 		};
 		(void)erofs_map_dev(sb, &mdev);
 
diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index a076cca1f547..4535f2f0a014 100644
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -219,7 +219,7 @@ static int z_erofs_load_compact_lcluster(struct z_erofs_maprecorder *m,
 	unsigned int amortizedshift;
 	erofs_off_t pos;
 
-	if (lcn >= totalidx)
+	if (lcn >= totalidx || vi->z_logical_clusterbits > 14)
 		return -EINVAL;
 
 	m->lcn = lcn;
@@ -390,7 +390,7 @@ static int z_erofs_get_extent_decompressedlen(struct z_erofs_maprecorder *m)
 	u64 lcn = m->lcn, headlcn = map->m_la >> lclusterbits;
 	int err;
 
-	do {
+	while (1) {
 		/* handle the last EOF pcluster (no next HEAD lcluster) */
 		if ((lcn << lclusterbits) >= inode->i_size) {
 			map->m_llen = inode->i_size - map->m_la;
@@ -402,14 +402,16 @@ static int z_erofs_get_extent_decompressedlen(struct z_erofs_maprecorder *m)
 			return err;
 
 		if (m->type == Z_EROFS_LCLUSTER_TYPE_NONHEAD) {
-			DBG_BUGON(!m->delta[1] &&
-				  m->clusterofs != 1 << lclusterbits);
+			/* work around invalid d1 generated by pre-1.0 mkfs */
+			if (unlikely(!m->delta[1])) {
+				m->delta[1] = 1;
+				DBG_BUGON(1);
+			}
 		} else if (m->type == Z_EROFS_LCLUSTER_TYPE_PLAIN ||
 			   m->type == Z_EROFS_LCLUSTER_TYPE_HEAD1 ||
 			   m->type == Z_EROFS_LCLUSTER_TYPE_HEAD2) {
-			/* go on until the next HEAD lcluster */
 			if (lcn != headlcn)
-				break;
+				break;	/* ends at the next HEAD lcluster */
 			m->delta[1] = 1;
 		} else {
 			erofs_err(inode->i_sb, "unknown type %u @ lcn %llu of nid %llu",
@@ -418,8 +420,7 @@ static int z_erofs_get_extent_decompressedlen(struct z_erofs_maprecorder *m)
 			return -EOPNOTSUPP;
 		}
 		lcn += m->delta[1];
-	} while (m->delta[1]);
-
+	}
 	map->m_llen = (lcn << lclusterbits) + m->clusterofs - map->m_la;
 	return 0;
 }
diff --git a/fs/erofs/zutil.c b/fs/erofs/zutil.c
index 37afe2024840..75704f58ecfa 100644
--- a/fs/erofs/zutil.c
+++ b/fs/erofs/zutil.c
@@ -2,6 +2,7 @@
 /*
  * Copyright (C) 2018 HUAWEI, Inc.
  *             https://www.huawei.com/
+ * Copyright (C) 2024 Alibaba Cloud
  */
 #include "internal.h"
 
@@ -19,13 +20,12 @@ static unsigned int z_erofs_gbuf_count, z_erofs_gbuf_nrpages,
 module_param_named(global_buffers, z_erofs_gbuf_count, uint, 0444);
 module_param_named(reserved_pages, z_erofs_rsv_nrpages, uint, 0444);
 
-static atomic_long_t erofs_global_shrink_cnt;	/* for all mounted instances */
-/* protected by 'erofs_sb_list_lock' */
-static unsigned int shrinker_run_no;
+atomic_long_t erofs_global_shrink_cnt;	/* for all mounted instances */
 
-/* protects the mounted 'erofs_sb_list' */
+/* protects `erofs_sb_list_lock` and the mounted `erofs_sb_list` */
 static DEFINE_SPINLOCK(erofs_sb_list_lock);
 static LIST_HEAD(erofs_sb_list);
+static unsigned int shrinker_run_no;
 static struct shrinker *erofs_shrinker_info;
 
 static unsigned int z_erofs_gbuf_id(void)
@@ -214,145 +214,6 @@ void erofs_release_pages(struct page **pagepool)
 	}
 }
 
-static bool erofs_workgroup_get(struct erofs_workgroup *grp)
-{
-	if (lockref_get_not_zero(&grp->lockref))
-		return true;
-
-	spin_lock(&grp->lockref.lock);
-	if (__lockref_is_dead(&grp->lockref)) {
-		spin_unlock(&grp->lockref.lock);
-		return false;
-	}
-
-	if (!grp->lockref.count++)
-		atomic_long_dec(&erofs_global_shrink_cnt);
-	spin_unlock(&grp->lockref.lock);
-	return true;
-}
-
-struct erofs_workgroup *erofs_find_workgroup(struct super_block *sb,
-					     pgoff_t index)
-{
-	struct erofs_sb_info *sbi = EROFS_SB(sb);
-	struct erofs_workgroup *grp;
-
-repeat:
-	rcu_read_lock();
-	grp = xa_load(&sbi->managed_pslots, index);
-	if (grp) {
-		if (!erofs_workgroup_get(grp)) {
-			/* prefer to relax rcu read side */
-			rcu_read_unlock();
-			goto repeat;
-		}
-
-		DBG_BUGON(index != grp->index);
-	}
-	rcu_read_unlock();
-	return grp;
-}
-
-struct erofs_workgroup *erofs_insert_workgroup(struct super_block *sb,
-					       struct erofs_workgroup *grp)
-{
-	struct erofs_sb_info *const sbi = EROFS_SB(sb);
-	struct erofs_workgroup *pre;
-
-	DBG_BUGON(grp->lockref.count < 1);
-repeat:
-	xa_lock(&sbi->managed_pslots);
-	pre = __xa_cmpxchg(&sbi->managed_pslots, grp->index,
-			   NULL, grp, GFP_KERNEL);
-	if (pre) {
-		if (xa_is_err(pre)) {
-			pre = ERR_PTR(xa_err(pre));
-		} else if (!erofs_workgroup_get(pre)) {
-			/* try to legitimize the current in-tree one */
-			xa_unlock(&sbi->managed_pslots);
-			cond_resched();
-			goto repeat;
-		}
-		grp = pre;
-	}
-	xa_unlock(&sbi->managed_pslots);
-	return grp;
-}
-
-static void  __erofs_workgroup_free(struct erofs_workgroup *grp)
-{
-	atomic_long_dec(&erofs_global_shrink_cnt);
-	erofs_workgroup_free_rcu(grp);
-}
-
-void erofs_workgroup_put(struct erofs_workgroup *grp)
-{
-	if (lockref_put_or_lock(&grp->lockref))
-		return;
-
-	DBG_BUGON(__lockref_is_dead(&grp->lockref));
-	if (grp->lockref.count == 1)
-		atomic_long_inc(&erofs_global_shrink_cnt);
-	--grp->lockref.count;
-	spin_unlock(&grp->lockref.lock);
-}
-
-static bool erofs_try_to_release_workgroup(struct erofs_sb_info *sbi,
-					   struct erofs_workgroup *grp)
-{
-	int free = false;
-
-	spin_lock(&grp->lockref.lock);
-	if (grp->lockref.count)
-		goto out;
-
-	/*
-	 * Note that all cached pages should be detached before deleted from
-	 * the XArray. Otherwise some cached pages could be still attached to
-	 * the orphan old workgroup when the new one is available in the tree.
-	 */
-	if (erofs_try_to_free_all_cached_folios(sbi, grp))
-		goto out;
-
-	/*
-	 * It's impossible to fail after the workgroup is freezed,
-	 * however in order to avoid some race conditions, add a
-	 * DBG_BUGON to observe this in advance.
-	 */
-	DBG_BUGON(__xa_erase(&sbi->managed_pslots, grp->index) != grp);
-
-	lockref_mark_dead(&grp->lockref);
-	free = true;
-out:
-	spin_unlock(&grp->lockref.lock);
-	if (free)
-		__erofs_workgroup_free(grp);
-	return free;
-}
-
-static unsigned long erofs_shrink_workstation(struct erofs_sb_info *sbi,
-					      unsigned long nr_shrink)
-{
-	struct erofs_workgroup *grp;
-	unsigned int freed = 0;
-	unsigned long index;
-
-	xa_lock(&sbi->managed_pslots);
-	xa_for_each(&sbi->managed_pslots, index, grp) {
-		/* try to shrink each valid workgroup */
-		if (!erofs_try_to_release_workgroup(sbi, grp))
-			continue;
-		xa_unlock(&sbi->managed_pslots);
-
-		++freed;
-		if (!--nr_shrink)
-			return freed;
-		xa_lock(&sbi->managed_pslots);
-	}
-	xa_unlock(&sbi->managed_pslots);
-	return freed;
-}
-
 void erofs_shrinker_register(struct super_block *sb)
 {
 	struct erofs_sb_info *sbi = EROFS_SB(sb);
@@ -369,8 +230,8 @@ void erofs_shrinker_unregister(struct super_block *sb)
 	struct erofs_sb_info *const sbi = EROFS_SB(sb);
 
 	mutex_lock(&sbi->umount_mutex);
-	/* clean up all remaining workgroups in memory */
-	erofs_shrink_workstation(sbi, ~0UL);
+	/* clean up all remaining pclusters in memory */
+	z_erofs_shrink_scan(sbi, ~0UL);
 
 	spin_lock(&erofs_sb_list_lock);
 	list_del(&sbi->list);
@@ -418,9 +279,7 @@ static unsigned long erofs_shrink_scan(struct shrinker *shrink,
 
 		spin_unlock(&erofs_sb_list_lock);
 		sbi->shrinker_run_no = run_no;
-
-		freed += erofs_shrink_workstation(sbi, nr - freed);
-
+		freed += z_erofs_shrink_scan(sbi, nr - freed);
 		spin_lock(&erofs_sb_list_lock);
 		/* Get the next list element before we move this one */
 		p = p->next;