f2fs: support data compression

This patch tries to support compression in f2fs.

- New term named cluster is defined as basic unit of compression, file can
be divided into multiple clusters logically. One cluster includes 4 << n
(n >= 0) logical pages, compression size is also cluster size, each of
cluster can be compressed or not.

- In cluster metadata layout, one special flag is used to indicate cluster
is compressed one or normal one, for compressed cluster, following metadata
maps cluster to [1, 4 << n - 1] physical blocks, in where f2fs stores
data including compress header and compressed data.

- In order to eliminate write amplification during overwrite, F2FS only
support compression on write-once file, data can be compressed only when
all logical blocks in file are valid and cluster compress ratio is lower
than specified threshold.

- To enable compression on regular inode, there are three ways:
* chattr +c file
* chattr +c dir; touch dir/file
* mount w/ -o compress_extension=ext; touch file.ext

Compress metadata layout:
                             [Dnode Structure]
             +-----------------------------------------------+
             | cluster 1 | cluster 2 | ......... | cluster N |
             +-----------------------------------------------+
             .           .                       .           .
       .                       .                .                      .
  .         Compressed Cluster       .        .        Normal Cluster            .
+----------+---------+---------+---------+  +---------+---------+---------+---------+
|compr flag| block 1 | block 2 | block 3 |  | block 1 | block 2 | block 3 | block 4 |
+----------+---------+---------+---------+  +---------+---------+---------+---------+
           .                             .
         .                                           .
       .                                                           .
      +-------------+-------------+----------+----------------------------+
      | data length | data chksum | reserved |      compressed data       |
      +-------------+-------------+----------+----------------------------+

Changelog:

20190326:
- fix error handling of read_end_io().
- remove unneeded comments in f2fs_encrypt_one_page().

20190327:
- fix wrong use of f2fs_cluster_is_full() in f2fs_mpage_readpages().
- don't jump into loop directly to avoid uninitialized variables.
- add TODO tag in error path of f2fs_write_cache_pages().

20190328:
- fix wrong merge condition in f2fs_read_multi_pages().
- check compressed file in f2fs_post_read_required().

20190401
- allow overwrite on non-compressed cluster.
- check cluster meta before writing compressed data.

20190402
- don't preallocate blocks for compressed file.

- add lz4 compress algorithm
- process multiple post read works in one workqueue
  Now f2fs supports processing post read work in multiple workqueue,
  it shows low performance due to schedule overhead of multiple
  workqueue executing orderly.

20190921
- compress: support buffered overwrite
C: compress cluster flag
V: valid block address
N: NEW_ADDR

One cluster contain 4 blocks

 before overwrite   after overwrite

- VVVV		->	CVNN
- CVNN		->	VVVV

- CVNN		->	CVNN
- CVNN		->	CVVV

- CVVV		->	CVNN
- CVVV		->	CVVV

20191029
- add kconfig F2FS_FS_COMPRESSION to isolate compression related
codes, add kconfig F2FS_FS_{LZO,LZ4} to cover backend algorithm.
note that: will remove lzo backend if Jaegeuk agreed that too.
- update codes according to Eric's comments.

20191101
- apply fixes from Jaegeuk

20191113
- apply fixes from Jaegeuk
- split workqueue for fsverity

20191216
- apply fixes from Jaegeuk

20200117
- fix to avoid NULL pointer dereference

[Jaegeuk Kim]
- add tracepoint for f2fs_{,de}compress_pages()
- fix many bugs and add some compression stats
- fix overwrite/mmap bugs
- address 32bit build error, reported by Geert.
- bug fixes when handling errors and i_compressed_blocks

Reported-by: <noreply@ellerman.id.au>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>

1 files changed, 166 insertions, 20 deletions

diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 82df24f30c11..0dff22566a1d 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -51,7 +51,8 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf)
 	struct inode *inode = file_inode(vmf->vma->vm_file);
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct dnode_of_data dn;
-	int err;
+	bool need_alloc = true;
+	int err = 0;
 
 	if (unlikely(f2fs_cp_error(sbi))) {
 		err = -EIO;
@@ -63,8 +64,25 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf)
 		goto err;
 	}
 
+#ifdef CONFIG_F2FS_FS_COMPRESSION
+	if (f2fs_compressed_file(inode)) {
+		int ret = f2fs_is_compressed_cluster(inode, page->index);
+
+		if (ret < 0) {
+			err = ret;
+			goto err;
+		} else if (ret) {
+			if (ret < F2FS_I(inode)->i_cluster_size) {
+				err = -EAGAIN;
+				goto err;
+			}
+			need_alloc = false;
+		}
+	}
+#endif
 	/* should do out of any locked page */
-	f2fs_balance_fs(sbi, true);
+	if (need_alloc)
+		f2fs_balance_fs(sbi, true);
 
 	sb_start_pagefault(inode->i_sb);
 
@@ -81,15 +99,17 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf)
 		goto out_sem;
 	}
 
-	/* block allocation */
-	__do_map_lock(sbi, F2FS_GET_BLOCK_PRE_AIO, true);
-	set_new_dnode(&dn, inode, NULL, NULL, 0);
-	err = f2fs_get_block(&dn, page->index);
-	f2fs_put_dnode(&dn);
-	__do_map_lock(sbi, F2FS_GET_BLOCK_PRE_AIO, false);
-	if (err) {
-		unlock_page(page);
-		goto out_sem;
+	if (need_alloc) {
+		/* block allocation */
+		__do_map_lock(sbi, F2FS_GET_BLOCK_PRE_AIO, true);
+		set_new_dnode(&dn, inode, NULL, NULL, 0);
+		err = f2fs_get_block(&dn, page->index);
+		f2fs_put_dnode(&dn);
+		__do_map_lock(sbi, F2FS_GET_BLOCK_PRE_AIO, false);
+		if (err) {
+			unlock_page(page);
+			goto out_sem;
+		}
 	}
 
 	/* fill the page */
@@ -156,6 +176,8 @@ static inline enum cp_reason_type need_do_checkpoint(struct inode *inode)
 
 	if (!S_ISREG(inode->i_mode))
 		cp_reason = CP_NON_REGULAR;
+	else if (f2fs_compressed_file(inode))
+		cp_reason = CP_COMPRESSED;
 	else if (inode->i_nlink != 1)
 		cp_reason = CP_HARDLINK;
 	else if (is_sbi_flag_set(sbi, SBI_NEED_CP))
@@ -486,6 +508,9 @@ static int f2fs_file_mmap(struct file *file, struct vm_area_struct *vma)
 	if (unlikely(f2fs_cp_error(F2FS_I_SB(inode))))
 		return -EIO;
 
+	if (!f2fs_is_compress_backend_ready(inode))
+		return -EOPNOTSUPP;
+
 	/* we don't need to use inline_data strictly */
 	err = f2fs_convert_inline_inode(inode);
 	if (err)
@@ -493,6 +518,7 @@ static int f2fs_file_mmap(struct file *file, struct vm_area_struct *vma)
 
 	file_accessed(file);
 	vma->vm_ops = &f2fs_file_vm_ops;
+	set_inode_flag(inode, FI_MMAP_FILE);
 	return 0;
 }
 
@@ -503,6 +529,9 @@ static int f2fs_file_open(struct inode *inode, struct file *filp)
 	if (err)
 		return err;
 
+	if (!f2fs_is_compress_backend_ready(inode))
+		return -EOPNOTSUPP;
+
 	err = fsverity_file_open(inode, filp);
 	if (err)
 		return err;
@@ -519,6 +548,9 @@ void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count)
 	int nr_free = 0, ofs = dn->ofs_in_node, len = count;
 	__le32 *addr;
 	int base = 0;
+	bool compressed_cluster = false;
+	int cluster_index = 0, valid_blocks = 0;
+	int cluster_size = F2FS_I(dn->inode)->i_cluster_size;
 
 	if (IS_INODE(dn->node_page) && f2fs_has_extra_attr(dn->inode))
 		base = get_extra_isize(dn->inode);
@@ -526,26 +558,43 @@ void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count)
 	raw_node = F2FS_NODE(dn->node_page);
 	addr = blkaddr_in_node(raw_node) + base + ofs;
 
-	for (; count > 0; count--, addr++, dn->ofs_in_node++) {
+	/* Assumption: truncateion starts with cluster */
+	for (; count > 0; count--, addr++, dn->ofs_in_node++, cluster_index++) {
 		block_t blkaddr = le32_to_cpu(*addr);
 
+		if (f2fs_compressed_file(dn->inode) &&
+					!(cluster_index & (cluster_size - 1))) {
+			if (compressed_cluster)
+				f2fs_i_compr_blocks_update(dn->inode,
+							valid_blocks, false);
+			compressed_cluster = (blkaddr == COMPRESS_ADDR);
+			valid_blocks = 0;
+		}
+
 		if (blkaddr == NULL_ADDR)
 			continue;
 
 		dn->data_blkaddr = NULL_ADDR;
 		f2fs_set_data_blkaddr(dn);
 
-		if (__is_valid_data_blkaddr(blkaddr) &&
-			!f2fs_is_valid_blkaddr(sbi, blkaddr,
+		if (__is_valid_data_blkaddr(blkaddr)) {
+			if (!f2fs_is_valid_blkaddr(sbi, blkaddr,
 					DATA_GENERIC_ENHANCE))
-			continue;
+				continue;
+			if (compressed_cluster)
+				valid_blocks++;
+		}
 
-		f2fs_invalidate_blocks(sbi, blkaddr);
 		if (dn->ofs_in_node == 0 && IS_INODE(dn->node_page))
 			clear_inode_flag(dn->inode, FI_FIRST_BLOCK_WRITTEN);
+
+		f2fs_invalidate_blocks(sbi, blkaddr);
 		nr_free++;
 	}
 
+	if (compressed_cluster)
+		f2fs_i_compr_blocks_update(dn->inode, valid_blocks, false);
+
 	if (nr_free) {
 		pgoff_t fofs;
 		/*
@@ -588,6 +637,9 @@ static int truncate_partial_data_page(struct inode *inode, u64 from,
 		return 0;
 	}
 
+	if (f2fs_compressed_file(inode))
+		return 0;
+
 	page = f2fs_get_lock_data_page(inode, index, true);
 	if (IS_ERR(page))
 		return PTR_ERR(page) == -ENOENT ? 0 : PTR_ERR(page);
@@ -603,7 +655,7 @@ truncate_out:
 	return 0;
 }
 
-int f2fs_truncate_blocks(struct inode *inode, u64 from, bool lock)
+static int do_truncate_blocks(struct inode *inode, u64 from, bool lock)
 {
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct dnode_of_data dn;
@@ -668,6 +720,28 @@ free_partial:
 	return err;
 }
 
+int f2fs_truncate_blocks(struct inode *inode, u64 from, bool lock)
+{
+	u64 free_from = from;
+
+	/*
+	 * for compressed file, only support cluster size
+	 * aligned truncation.
+	 */
+	if (f2fs_compressed_file(inode)) {
+		size_t cluster_shift = PAGE_SHIFT +
+					F2FS_I(inode)->i_log_cluster_size;
+		size_t cluster_mask = (1 << cluster_shift) - 1;
+
+		free_from = from >> cluster_shift;
+		if (from & cluster_mask)
+			free_from++;
+		free_from <<= cluster_shift;
+	}
+
+	return do_truncate_blocks(inode, free_from, lock);
+}
+
 int f2fs_truncate(struct inode *inode)
 {
 	int err;
@@ -787,6 +861,10 @@ int f2fs_setattr(struct dentry *dentry, struct iattr *attr)
 	if (unlikely(f2fs_cp_error(F2FS_I_SB(inode))))
 		return -EIO;
 
+	if ((attr->ia_valid & ATTR_SIZE) &&
+		!f2fs_is_compress_backend_ready(inode))
+		return -EOPNOTSUPP;
+
 	err = setattr_prepare(dentry, attr);
 	if (err)
 		return err;
@@ -1027,8 +1105,8 @@ next_dnode:
 	} else if (ret == -ENOENT) {
 		if (dn.max_level == 0)
 			return -ENOENT;
-		done = min((pgoff_t)ADDRS_PER_BLOCK(inode) - dn.ofs_in_node,
-									len);
+		done = min((pgoff_t)ADDRS_PER_BLOCK(inode) -
+						dn.ofs_in_node, len);
 		blkaddr += done;
 		do_replace += done;
 		goto next;
@@ -1622,6 +1700,8 @@ static long f2fs_fallocate(struct file *file, int mode,
 		return -EIO;
 	if (!f2fs_is_checkpoint_ready(F2FS_I_SB(inode)))
 		return -ENOSPC;
+	if (!f2fs_is_compress_backend_ready(inode))
+		return -EOPNOTSUPP;
 
 	/* f2fs only support ->fallocate for regular file */
 	if (!S_ISREG(inode->i_mode))
@@ -1631,6 +1711,11 @@ static long f2fs_fallocate(struct file *file, int mode,
 		(mode & (FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_INSERT_RANGE)))
 		return -EOPNOTSUPP;
 
+	if (f2fs_compressed_file(inode) &&
+		(mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_COLLAPSE_RANGE |
+			FALLOC_FL_ZERO_RANGE | FALLOC_FL_INSERT_RANGE)))
+		return -EOPNOTSUPP;
+
 	if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE |
 			FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE |
 			FALLOC_FL_INSERT_RANGE))
@@ -1720,7 +1805,40 @@ static int f2fs_setflags_common(struct inode *inode, u32 iflags, u32 mask)
 			return -ENOTEMPTY;
 	}
 
+	if (iflags & (F2FS_COMPR_FL | F2FS_NOCOMP_FL)) {
+		if (!f2fs_sb_has_compression(F2FS_I_SB(inode)))
+			return -EOPNOTSUPP;
+		if ((iflags & F2FS_COMPR_FL) && (iflags & F2FS_NOCOMP_FL))
+			return -EINVAL;
+	}
+
+	if ((iflags ^ fi->i_flags) & F2FS_COMPR_FL) {
+		if (S_ISREG(inode->i_mode) &&
+			(fi->i_flags & F2FS_COMPR_FL || i_size_read(inode) ||
+						F2FS_HAS_BLOCKS(inode)))
+			return -EINVAL;
+		if (iflags & F2FS_NOCOMP_FL)
+			return -EINVAL;
+		if (iflags & F2FS_COMPR_FL) {
+			int err = f2fs_convert_inline_inode(inode);
+
+			if (err)
+				return err;
+
+			if (!f2fs_may_compress(inode))
+				return -EINVAL;
+
+			set_compress_context(inode);
+		}
+	}
+	if ((iflags ^ fi->i_flags) & F2FS_NOCOMP_FL) {
+		if (fi->i_flags & F2FS_COMPR_FL)
+			return -EINVAL;
+	}
+
 	fi->i_flags = iflags | (fi->i_flags & ~mask);
+	f2fs_bug_on(F2FS_I_SB(inode), (fi->i_flags & F2FS_COMPR_FL) &&
+					(fi->i_flags & F2FS_NOCOMP_FL));
 
 	if (fi->i_flags & F2FS_PROJINHERIT_FL)
 		set_inode_flag(inode, FI_PROJ_INHERIT);
@@ -1746,11 +1864,13 @@ static const struct {
 	u32 iflag;
 	u32 fsflag;
 } f2fs_fsflags_map[] = {
+	{ F2FS_COMPR_FL,	FS_COMPR_FL },
 	{ F2FS_SYNC_FL,		FS_SYNC_FL },
 	{ F2FS_IMMUTABLE_FL,	FS_IMMUTABLE_FL },
 	{ F2FS_APPEND_FL,	FS_APPEND_FL },
 	{ F2FS_NODUMP_FL,	FS_NODUMP_FL },
 	{ F2FS_NOATIME_FL,	FS_NOATIME_FL },
+	{ F2FS_NOCOMP_FL,	FS_NOCOMP_FL },
 	{ F2FS_INDEX_FL,	FS_INDEX_FL },
 	{ F2FS_DIRSYNC_FL,	FS_DIRSYNC_FL },
 	{ F2FS_PROJINHERIT_FL,	FS_PROJINHERIT_FL },
@@ -1758,11 +1878,13 @@ static const struct {
 };
 
 #define F2FS_GETTABLE_FS_FL (		\
+		FS_COMPR_FL |		\
 		FS_SYNC_FL |		\
 		FS_IMMUTABLE_FL |	\
 		FS_APPEND_FL |		\
 		FS_NODUMP_FL |		\
 		FS_NOATIME_FL |		\
+		FS_NOCOMP_FL |		\
 		FS_INDEX_FL |		\
 		FS_DIRSYNC_FL |		\
 		FS_PROJINHERIT_FL |	\
@@ -1773,11 +1895,13 @@ static const struct {
 		FS_CASEFOLD_FL)
 
 #define F2FS_SETTABLE_FS_FL (		\
+		FS_COMPR_FL |		\
 		FS_SYNC_FL |		\
 		FS_IMMUTABLE_FL |	\
 		FS_APPEND_FL |		\
 		FS_NODUMP_FL |		\
 		FS_NOATIME_FL |		\
+		FS_NOCOMP_FL |		\
 		FS_DIRSYNC_FL |		\
 		FS_PROJINHERIT_FL |	\
 		FS_CASEFOLD_FL)
@@ -1898,6 +2022,8 @@ static int f2fs_ioc_start_atomic_write(struct file *filp)
 
 	inode_lock(inode);
 
+	f2fs_disable_compressed_file(inode);
+
 	if (f2fs_is_atomic_file(inode)) {
 		if (is_inode_flag_set(inode, FI_ATOMIC_REVOKE_REQUEST))
 			ret = -EINVAL;
@@ -3098,10 +3224,16 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg)
 		ret = -EAGAIN;
 		goto out;
 	}
+
 	ret = f2fs_convert_inline_inode(inode);
 	if (ret)
 		goto out;
 
+	if (f2fs_disable_compressed_file(inode)) {
+		ret = -EOPNOTSUPP;
+		goto out;
+	}
+
 	set_inode_flag(inode, FI_PIN_FILE);
 	ret = F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN];
 done:
@@ -3350,6 +3482,17 @@ long f2fs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 	}
 }
 
+static ssize_t f2fs_file_read_iter(struct kiocb *iocb, struct iov_iter *iter)
+{
+	struct file *file = iocb->ki_filp;
+	struct inode *inode = file_inode(file);
+
+	if (!f2fs_is_compress_backend_ready(inode))
+		return -EOPNOTSUPP;
+
+	return generic_file_read_iter(iocb, iter);
+}
+
 static ssize_t f2fs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 {
 	struct file *file = iocb->ki_filp;
@@ -3361,6 +3504,9 @@ static ssize_t f2fs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 		goto out;
 	}
 
+	if (!f2fs_is_compress_backend_ready(inode))
+		return -EOPNOTSUPP;
+
 	if (iocb->ki_flags & IOCB_NOWAIT) {
 		if (!inode_trylock(inode)) {
 			ret = -EAGAIN;
@@ -3498,7 +3644,7 @@ long f2fs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 
 const struct file_operations f2fs_file_operations = {
 	.llseek		= f2fs_llseek,
-	.read_iter	= generic_file_read_iter,
+	.read_iter	= f2fs_file_read_iter,
 	.write_iter	= f2fs_file_write_iter,
 	.open		= f2fs_file_open,
 	.release	= f2fs_release_file,