Merge tag 'f2fs-for-6.0' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs

Pull f2fs updates from Jaegeuk Kim:
"In this cycle, we mainly fixed some corner cases that manipulate a
per-file compression flag inappropriately. And, we found f2fs counted
valid blocks in a section incorrectly when zone capacity is set, and
thus, fixed it with additional sysfs entry to check it easily.

Lastly, this series includes several patches with respect to the new
atomic write support such as a couple of bug fixes and re-adding
atomic_write_abort support that we removed by mistake in the previous
release.

Enhancements:
- add sysfs entries to understand atomic write operations and zone
capacity
- introduce memory mode to get a hint for low-memory devices
- adjust the waiting time of foreground GC
- decompress clusters under softirq to avoid non-deterministic
latency
- do not skip updating inode when retrying to flush node page
- enforce single zone capacity

Bug fixes:
- set the compression/no-compression flags correctly
- revive F2FS_IOC_ABORT_VOLATILE_WRITE
- check inline_data during compressed inode conversion
- understand zone capacity when calculating valid block count

As usual, the series includes several minor clean-ups and sanity
checks"

* tag 'f2fs-for-6.0' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (29 commits)
f2fs: use onstack pages instead of pvec
f2fs: intorduce f2fs_all_cluster_page_ready
f2fs: clean up f2fs_abort_atomic_write()
f2fs: handle decompress only post processing in softirq
f2fs: do not allow to decompress files have FI_COMPRESS_RELEASED
f2fs: do not set compression bit if kernel doesn't support
f2fs: remove device type check for direct IO
f2fs: fix null-ptr-deref in f2fs_get_dnode_of_data
f2fs: revive F2FS_IOC_ABORT_VOLATILE_WRITE
f2fs: fix to do sanity check on segment type in build_sit_entries()
f2fs: obsolete unused MAX_DISCARD_BLOCKS
f2fs: fix to avoid use f2fs_bug_on() in f2fs_new_node_page()
f2fs: fix to remove F2FS_COMPR_FL and tag F2FS_NOCOMP_FL at the same time
f2fs: introduce sysfs atomic write statistics
f2fs: don't bother wait_ms by foreground gc
f2fs: invalidate meta pages only for post_read required inode
f2fs: allow compression of files without blocks
f2fs: fix to check inline_data during compressed inode conversion
f2fs: Delete f2fs_copy_page() and replace with memcpy_page()
f2fs: fix to invalidate META_MAPPING before DIO write
...

Documentation/ABI/testing/sysfs-fs-f2fs

···
 Contact:	"Jaegeuk Kim" <[e-mail]>
 Description:	Controls max # of node block writes to be used for roll forward
 		recovery. This can limit the roll forward recovery time.
+
+What:		/sys/fs/f2fs/<disk>/unusable_blocks_per_sec
+Date:		June 2022
+Contact:	"Jaegeuk Kim" <[e-mail]>
+Description:	Shows the number of unusable blocks in a section which was defined by
+		the zone capacity reported by underlying zoned device.
+
+What:		/sys/fs/f2fs/<disk>/current_atomic_write
+Date:		July 2022
+Contact:	"Daeho Jeong" <[e-mail]>
+Description:	Show the total current atomic write block count, which is not committed yet.
+		This is a read-only entry.
+
+What:		/sys/fs/f2fs/<disk>/peak_atomic_write
+Date:		July 2022
+Contact:	"Daeho Jeong" <[e-mail]>
+Description:	Show the peak value of total current atomic write block count after boot.
+		If you write "0" here, you can initialize to "0".
+
+What:		/sys/fs/f2fs/<disk>/committed_atomic_block
+Date:		July 2022
+Contact:	"Daeho Jeong" <[e-mail]>
+Description:	Show the accumulated total committed atomic write block count after boot.
+		If you write "0" here, you can initialize to "0".
+
+What:		/sys/fs/f2fs/<disk>/revoked_atomic_block
+Date:		July 2022
+Contact:	"Daeho Jeong" <[e-mail]>
+Description:	Show the accumulated total revoked atomic write block count after boot.
+		If you write "0" here, you can initialize to "0".

Documentation/filesystems/f2fs.rst

···
 			 default, it is helpful for large sized SMR or ZNS devices to
 			 reduce memory cost by getting rid of fs metadata supports small
 			 discard.
+memory=%s		 Control memory mode. This supports "normal" and "low" modes.
+			 "low" mode is introduced to support low memory devices.
+			 Because of the nature of low memory devices, in this mode, f2fs
+			 will try to save memory sometimes by sacrificing performance.
+			 "normal" mode is the default mode and same as before.
 ======================== ============================================================
 
 Debugfs Entries

fs/f2fs/compress.c

···
 	return ret;
 }
 
-void f2fs_decompress_cluster(struct decompress_io_ctx *dic)
+static int f2fs_prepare_decomp_mem(struct decompress_io_ctx *dic,
+		bool pre_alloc);
+static void f2fs_release_decomp_mem(struct decompress_io_ctx *dic,
+		bool bypass_destroy_callback, bool pre_alloc);
+
+void f2fs_decompress_cluster(struct decompress_io_ctx *dic, bool in_task)
 {
 	struct f2fs_sb_info *sbi = F2FS_I_SB(dic->inode);
 	struct f2fs_inode_info *fi = F2FS_I(dic->inode);
 	const struct f2fs_compress_ops *cops =
 			f2fs_cops[fi->i_compress_algorithm];
+	bool bypass_callback = false;
 	int ret;
-	int i;
 
 	trace_f2fs_decompress_pages_start(dic->inode, dic->cluster_idx,
 				dic->cluster_size, fi->i_compress_algorithm);
···
 		goto out_end_io;
 	}
 
-	dic->tpages = page_array_alloc(dic->inode, dic->cluster_size);
-	if (!dic->tpages) {
-		ret = -ENOMEM;
-		goto out_end_io;
-	}
-
-	for (i = 0; i < dic->cluster_size; i++) {
-		if (dic->rpages[i]) {
-			dic->tpages[i] = dic->rpages[i];
-			continue;
-		}
-
-		dic->tpages[i] = f2fs_compress_alloc_page();
-		if (!dic->tpages[i]) {
-			ret = -ENOMEM;
-			goto out_end_io;
-		}
-	}
-
-	if (cops->init_decompress_ctx) {
-		ret = cops->init_decompress_ctx(dic);
-		if (ret)
-			goto out_end_io;
-	}
-
-	dic->rbuf = f2fs_vmap(dic->tpages, dic->cluster_size);
-	if (!dic->rbuf) {
-		ret = -ENOMEM;
-		goto out_destroy_decompress_ctx;
-	}
-
-	dic->cbuf = f2fs_vmap(dic->cpages, dic->nr_cpages);
-	if (!dic->cbuf) {
-		ret = -ENOMEM;
-		goto out_vunmap_rbuf;
+	ret = f2fs_prepare_decomp_mem(dic, false);
+	if (ret) {
+		bypass_callback = true;
+		goto out_release;
 	}
 
 	dic->clen = le32_to_cpu(dic->cbuf->clen);
···
 
 	if (dic->clen > PAGE_SIZE * dic->nr_cpages - COMPRESS_HEADER_SIZE) {
 		ret = -EFSCORRUPTED;
-		goto out_vunmap_cbuf;
+		goto out_release;
 	}
 
 	ret = cops->decompress_pages(dic);
···
 		}
 	}
 
-out_vunmap_cbuf:
-	vm_unmap_ram(dic->cbuf, dic->nr_cpages);
-out_vunmap_rbuf:
-	vm_unmap_ram(dic->rbuf, dic->cluster_size);
-out_destroy_decompress_ctx:
-	if (cops->destroy_decompress_ctx)
-		cops->destroy_decompress_ctx(dic);
+out_release:
+	f2fs_release_decomp_mem(dic, bypass_callback, false);
+
 out_end_io:
 	trace_f2fs_decompress_pages_end(dic->inode, dic->cluster_idx,
 							dic->clen, ret);
-	f2fs_decompress_end_io(dic, ret);
+	f2fs_decompress_end_io(dic, ret, in_task);
 }
 
 /*
···
  * (or in the case of a failure, cleans up without actually decompressing).
  */
 void f2fs_end_read_compressed_page(struct page *page, bool failed,
-						block_t blkaddr)
+		block_t blkaddr, bool in_task)
 {
 	struct decompress_io_ctx *dic =
 			(struct decompress_io_ctx *)page_private(page);
···
 
 	if (failed)
 		WRITE_ONCE(dic->failed, true);
-	else if (blkaddr)
+	else if (blkaddr && in_task)
 		f2fs_cache_compressed_page(sbi, page,
 					dic->inode->i_ino, blkaddr);
 
 	if (atomic_dec_and_test(&dic->remaining_pages))
-		f2fs_decompress_cluster(dic);
+		f2fs_decompress_cluster(dic, in_task);
 }
 
 static bool is_page_in_cluster(struct compress_ctx *cc, pgoff_t index)
···
 	return is_page_in_cluster(cc, index);
 }
 
-bool f2fs_all_cluster_page_loaded(struct compress_ctx *cc, struct pagevec *pvec,
-				int index, int nr_pages)
+bool f2fs_all_cluster_page_ready(struct compress_ctx *cc, struct page **pages,
+				int index, int nr_pages, bool uptodate)
 {
-	unsigned long pgidx;
-	int i;
+	unsigned long pgidx = pages[index]->index;
+	int i = uptodate ? 0 : 1;
+
+	/*
+	 * when uptodate set to true, try to check all pages in cluster is
+	 * uptodate or not.
+	 */
+	if (uptodate && (pgidx % cc->cluster_size))
+		return false;
 
 	if (nr_pages - index < cc->cluster_size)
 		return false;
 
-	pgidx = pvec->pages[index]->index;
-
-	for (i = 1; i < cc->cluster_size; i++) {
-		if (pvec->pages[index + i]->index != pgidx + i)
+	for (; i < cc->cluster_size; i++) {
+		if (pages[index + i]->index != pgidx + i)
+			return false;
+		if (uptodate && !PageUptodate(pages[index + i]))
 			return false;
 	}
 
···
 	return err;
 }
 
-static void f2fs_free_dic(struct decompress_io_ctx *dic);
+static inline bool allow_memalloc_for_decomp(struct f2fs_sb_info *sbi,
+		bool pre_alloc)
+{
+	return pre_alloc ^ f2fs_low_mem_mode(sbi);
+}
+
+static int f2fs_prepare_decomp_mem(struct decompress_io_ctx *dic,
+		bool pre_alloc)
+{
+	const struct f2fs_compress_ops *cops =
+		f2fs_cops[F2FS_I(dic->inode)->i_compress_algorithm];
+	int i;
+
+	if (!allow_memalloc_for_decomp(F2FS_I_SB(dic->inode), pre_alloc))
+		return 0;
+
+	dic->tpages = page_array_alloc(dic->inode, dic->cluster_size);
+	if (!dic->tpages)
+		return -ENOMEM;
+
+	for (i = 0; i < dic->cluster_size; i++) {
+		if (dic->rpages[i]) {
+			dic->tpages[i] = dic->rpages[i];
+			continue;
+		}
+
+		dic->tpages[i] = f2fs_compress_alloc_page();
+		if (!dic->tpages[i])
+			return -ENOMEM;
+	}
+
+	dic->rbuf = f2fs_vmap(dic->tpages, dic->cluster_size);
+	if (!dic->rbuf)
+		return -ENOMEM;
+
+	dic->cbuf = f2fs_vmap(dic->cpages, dic->nr_cpages);
+	if (!dic->cbuf)
+		return -ENOMEM;
+
+	if (cops->init_decompress_ctx) {
+		int ret = cops->init_decompress_ctx(dic);
+
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static void f2fs_release_decomp_mem(struct decompress_io_ctx *dic,
+		bool bypass_destroy_callback, bool pre_alloc)
+{
+	const struct f2fs_compress_ops *cops =
+		f2fs_cops[F2FS_I(dic->inode)->i_compress_algorithm];
+
+	if (!allow_memalloc_for_decomp(F2FS_I_SB(dic->inode), pre_alloc))
+		return;
+
+	if (!bypass_destroy_callback && cops->destroy_decompress_ctx)
+		cops->destroy_decompress_ctx(dic);
+
+	if (dic->cbuf)
+		vm_unmap_ram(dic->cbuf, dic->nr_cpages);
+
+	if (dic->rbuf)
+		vm_unmap_ram(dic->rbuf, dic->cluster_size);
+}
+
+static void f2fs_free_dic(struct decompress_io_ctx *dic,
+		bool bypass_destroy_callback);
 
 struct decompress_io_ctx *f2fs_alloc_dic(struct compress_ctx *cc)
 {
 	struct decompress_io_ctx *dic;
 	pgoff_t start_idx = start_idx_of_cluster(cc);
-	int i;
+	struct f2fs_sb_info *sbi = F2FS_I_SB(cc->inode);
+	int i, ret;
 
-	dic = f2fs_kmem_cache_alloc(dic_entry_slab, GFP_F2FS_ZERO,
-					false, F2FS_I_SB(cc->inode));
+	dic = f2fs_kmem_cache_alloc(dic_entry_slab, GFP_F2FS_ZERO, false, sbi);
 	if (!dic)
 		return ERR_PTR(-ENOMEM);
 
···
 	dic->nr_rpages = cc->cluster_size;
 
 	dic->cpages = page_array_alloc(dic->inode, dic->nr_cpages);
-	if (!dic->cpages)
+	if (!dic->cpages) {
+		ret = -ENOMEM;
 		goto out_free;
+	}
 
 	for (i = 0; i < dic->nr_cpages; i++) {
 		struct page *page;
 
 		page = f2fs_compress_alloc_page();
-		if (!page)
+		if (!page) {
+			ret = -ENOMEM;
 			goto out_free;
+		}
 
 		f2fs_set_compressed_page(page, cc->inode,
 					start_idx + i + 1, dic);
 		dic->cpages[i] = page;
 	}
 
+	ret = f2fs_prepare_decomp_mem(dic, true);
+	if (ret)
+		goto out_free;
+
 	return dic;
 
 out_free:
-	f2fs_free_dic(dic);
-	return ERR_PTR(-ENOMEM);
+	f2fs_free_dic(dic, true);
+	return ERR_PTR(ret);
 }
 
-static void f2fs_free_dic(struct decompress_io_ctx *dic)
+static void f2fs_free_dic(struct decompress_io_ctx *dic,
+		bool bypass_destroy_callback)
 {
 	int i;
+
+	f2fs_release_decomp_mem(dic, bypass_destroy_callback, true);
 
 	if (dic->tpages) {
 		for (i = 0; i < dic->cluster_size; i++) {
···
 	kmem_cache_free(dic_entry_slab, dic);
 }
 
-static void f2fs_put_dic(struct decompress_io_ctx *dic)
+static void f2fs_late_free_dic(struct work_struct *work)
 {
-	if (refcount_dec_and_test(&dic->refcnt))
-		f2fs_free_dic(dic);
+	struct decompress_io_ctx *dic =
+		container_of(work, struct decompress_io_ctx, free_work);
+
+	f2fs_free_dic(dic, false);
+}
+
+static void f2fs_put_dic(struct decompress_io_ctx *dic, bool in_task)
+{
+	if (refcount_dec_and_test(&dic->refcnt)) {
+		if (in_task) {
+			f2fs_free_dic(dic, false);
+		} else {
+			INIT_WORK(&dic->free_work, f2fs_late_free_dic);
+			queue_work(F2FS_I_SB(dic->inode)->post_read_wq,
+					&dic->free_work);
+		}
+	}
 }
 
 /*
  * Update and unlock the cluster's pagecache pages, and release the reference to
  * the decompress_io_ctx that was being held for I/O completion.
  */
-static void __f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed)
+static void __f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed,
+				bool in_task)
 {
 	int i;
 
···
 		unlock_page(rpage);
 	}
 
-	f2fs_put_dic(dic);
+	f2fs_put_dic(dic, in_task);
 }
 
 static void f2fs_verify_cluster(struct work_struct *work)
···
 			SetPageError(rpage);
 	}
 
-	__f2fs_decompress_end_io(dic, false);
+	__f2fs_decompress_end_io(dic, false, true);
 }
 
 /*
  * This is called when a compressed cluster has been decompressed
  * (or failed to be read and/or decompressed).
  */
-void f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed)
+void f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed,
+				bool in_task)
 {
 	if (!failed && dic->need_verity) {
 		/*
···
 		INIT_WORK(&dic->verity_work, f2fs_verify_cluster);
 		fsverity_enqueue_verify_work(&dic->verity_work);
 	} else {
-		__f2fs_decompress_end_io(dic, failed);
+		__f2fs_decompress_end_io(dic, failed, in_task);
 	}
 }
 
···
  *
  * This is called when the page is no longer needed and can be freed.
  */
-void f2fs_put_page_dic(struct page *page)
+void f2fs_put_page_dic(struct page *page, bool in_task)
 {
 	struct decompress_io_ctx *dic =
 			(struct decompress_io_ctx *)page_private(page);
 
-	f2fs_put_dic(dic);
+	f2fs_put_dic(dic, in_task);
 }
 
 /*
···
 {
 	dev_t dev = sbi->sb->s_bdev->bd_dev;
 	char slab_name[32];
+
+	if (!f2fs_sb_has_compression(sbi))
+		return 0;
 
 	sprintf(slab_name, "f2fs_page_array_entry-%u:%u", MAJOR(dev), MINOR(dev));
 

fs/f2fs/data.c

···
 	block_t fs_blkaddr;
 };
 
-static void f2fs_finish_read_bio(struct bio *bio)
+static void f2fs_finish_read_bio(struct bio *bio, bool in_task)
 {
 	struct bio_vec *bv;
 	struct bvec_iter_all iter_all;
···
 
 		if (f2fs_is_compressed_page(page)) {
 			if (bio->bi_status)
-				f2fs_end_read_compressed_page(page, true, 0);
-			f2fs_put_page_dic(page);
+				f2fs_end_read_compressed_page(page, true, 0,
+							in_task);
+			f2fs_put_page_dic(page, in_task);
 			continue;
 		}
 
···
 		fsverity_verify_bio(bio);
 	}
 
-	f2fs_finish_read_bio(bio);
+	f2fs_finish_read_bio(bio, true);
 }
 
 /*
···
  * can involve reading verity metadata pages from the file, and these verity
  * metadata pages may be encrypted and/or compressed.
  */
-static void f2fs_verify_and_finish_bio(struct bio *bio)
+static void f2fs_verify_and_finish_bio(struct bio *bio, bool in_task)
 {
 	struct bio_post_read_ctx *ctx = bio->bi_private;
 
···
 		INIT_WORK(&ctx->work, f2fs_verify_bio);
 		fsverity_enqueue_verify_work(&ctx->work);
 	} else {
-		f2fs_finish_read_bio(bio);
+		f2fs_finish_read_bio(bio, in_task);
 	}
 }
 
···
  * that the bio includes at least one compressed page.  The actual decompression
  * is done on a per-cluster basis, not a per-bio basis.
  */
-static void f2fs_handle_step_decompress(struct bio_post_read_ctx *ctx)
+static void f2fs_handle_step_decompress(struct bio_post_read_ctx *ctx,
+		bool in_task)
 {
 	struct bio_vec *bv;
 	struct bvec_iter_all iter_all;
···
 		/* PG_error was set if decryption failed. */
 		if (f2fs_is_compressed_page(page))
 			f2fs_end_read_compressed_page(page, PageError(page),
-						blkaddr);
+						blkaddr, in_task);
 		else
 			all_compressed = false;
 
···
 		fscrypt_decrypt_bio(ctx->bio);
 
 	if (ctx->enabled_steps & STEP_DECOMPRESS)
-		f2fs_handle_step_decompress(ctx);
+		f2fs_handle_step_decompress(ctx, true);
 
-	f2fs_verify_and_finish_bio(ctx->bio);
+	f2fs_verify_and_finish_bio(ctx->bio, true);
 }
 
 static void f2fs_read_end_io(struct bio *bio)
 {
 	struct f2fs_sb_info *sbi = F2FS_P_SB(bio_first_page_all(bio));
 	struct bio_post_read_ctx *ctx;
+	bool intask = in_task();
 
 	iostat_update_and_unbind_ctx(bio, 0);
 	ctx = bio->bi_private;
···
 	}
 
 	if (bio->bi_status) {
-		f2fs_finish_read_bio(bio);
+		f2fs_finish_read_bio(bio, intask);
 		return;
 	}
 
-	if (ctx && (ctx->enabled_steps & (STEP_DECRYPT | STEP_DECOMPRESS))) {
-		INIT_WORK(&ctx->work, f2fs_post_read_work);
-		queue_work(ctx->sbi->post_read_wq, &ctx->work);
-	} else {
-		f2fs_verify_and_finish_bio(bio);
+	if (ctx) {
+		unsigned int enabled_steps = ctx->enabled_steps &
+					(STEP_DECRYPT | STEP_DECOMPRESS);
+
+		/*
+		 * If we have only decompression step between decompression and
+		 * decrypt, we don't need post processing for this.
+		 */
+		if (enabled_steps == STEP_DECOMPRESS &&
+				!f2fs_low_mem_mode(sbi)) {
+			f2fs_handle_step_decompress(ctx, intask);
+		} else if (enabled_steps) {
+			INIT_WORK(&ctx->work, f2fs_post_read_work);
+			queue_work(ctx->sbi->post_read_wq, &ctx->work);
+			return;
+		}
 	}
+
+	f2fs_verify_and_finish_bio(bio, intask);
 }
 
 static void f2fs_write_end_io(struct bio *bio)
···
 		 */
 		f2fs_wait_on_block_writeback_range(inode,
 						map->m_pblk, map->m_len);
-		invalidate_mapping_pages(META_MAPPING(sbi),
-						map->m_pblk, map->m_pblk);
 
 		if (map->m_multidev_dio) {
 			block_t blk_addr = map->m_pblk;
···
 
 		if (f2fs_load_compressed_page(sbi, page, blkaddr)) {
 			if (atomic_dec_and_test(&dic->remaining_pages))
-				f2fs_decompress_cluster(dic);
+				f2fs_decompress_cluster(dic, true);
 			continue;
 		}
 
···
 					page->index, for_write);
 			if (IS_ERR(bio)) {
 				ret = PTR_ERR(bio);
-				f2fs_decompress_end_io(dic, ret);
+				f2fs_decompress_end_io(dic, ret, true);
 				f2fs_put_dnode(&dn);
 				*bio_ret = NULL;
 				return ret;
···
 		.submitted = false,
 		.compr_blocks = compr_blocks,
 		.need_lock = LOCK_RETRY,
+		.post_read = f2fs_post_read_required(inode),
 		.io_type = io_type,
 		.io_wbc = wbc,
 		.bio = bio,
···
 {
 	int ret = 0;
 	int done = 0, retry = 0;
-	struct pagevec pvec;
+	struct page *pages[F2FS_ONSTACK_PAGES];
 	struct f2fs_sb_info *sbi = F2FS_M_SB(mapping);
 	struct bio *bio = NULL;
 	sector_t last_block;
···
 	int submitted = 0;
 	int i;
 
-	pagevec_init(&pvec);
-
 	if (get_dirty_pages(mapping->host) <=
 				SM_I(F2FS_M_SB(mapping))->min_hot_blocks)
 		set_inode_flag(mapping->host, FI_HOT_DATA);
···
 		tag_pages_for_writeback(mapping, index, end);
 	done_index = index;
 	while (!done && !retry && (index <= end)) {
-		nr_pages = pagevec_lookup_range_tag(&pvec, mapping, &index, end,
-				tag);
+		nr_pages = find_get_pages_range_tag(mapping, &index, end,
+				tag, F2FS_ONSTACK_PAGES, pages);
 		if (nr_pages == 0)
 			break;
 
 		for (i = 0; i < nr_pages; i++) {
-			struct page *page = pvec.pages[i];
+			struct page *page = pages[i];
 			bool need_readd;
 readd:
 			need_readd = false;
···
 				if (!f2fs_cluster_is_empty(&cc))
 					goto lock_page;
 
+				if (f2fs_all_cluster_page_ready(&cc,
+					pages, i, nr_pages, true))
+					goto lock_page;
+
 				ret2 = f2fs_prepare_compress_overwrite(
 							inode, &pagep,
 							page->index, &fsdata);
···
 				} else if (ret2 &&
 					(!f2fs_compress_write_end(inode,
 						fsdata, page->index, 1) ||
-					 !f2fs_all_cluster_page_loaded(&cc,
-						&pvec, i, nr_pages))) {
+					 !f2fs_all_cluster_page_ready(&cc,
+						pages, i, nr_pages, false))) {
 					retry = 1;
 					break;
 				}
···
 			if (need_readd)
 				goto readd;
 		}
-		pagevec_release(&pvec);
+		release_pages(pages, nr_pages);
 		cond_resched();
 	}
 #ifdef CONFIG_F2FS_FS_COMPRESSION
···
 	struct inode *cow_inode = F2FS_I(inode)->cow_inode;
 	pgoff_t index = page->index;
 	int err = 0;
-	block_t ori_blk_addr;
+	block_t ori_blk_addr = NULL_ADDR;
 
 	/* If pos is beyond the end of file, reserve a new block in COW inode */
 	if ((pos & PAGE_MASK) >= i_size_read(inode))
-		return __reserve_data_block(cow_inode, index, blk_addr,
-					node_changed);
+		goto reserve_block;
 
 	/* Look for the block in COW inode first */
 	err = __find_data_block(cow_inode, index, blk_addr);
···
 	if (err)
 		return err;
 
+reserve_block:
 	/* Finally, we should reserve a new block in COW inode for the update */
 	err = __reserve_data_block(cow_inode, index, blk_addr, node_changed);
 	if (err)
 		return err;
+	inc_atomic_write_cnt(inode);
 
 	if (ori_blk_addr != NULL_ADDR)
 		*blk_addr = ori_blk_addr;

fs/f2fs/debug.c

···
 
 	bimodal = 0;
 	total_vblocks = 0;
-	blks_per_sec = BLKS_PER_SEC(sbi);
+	blks_per_sec = CAP_BLKS_PER_SEC(sbi);
 	hblks_per_sec = blks_per_sec / 2;
 	for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) {
 		vblocks = get_valid_blocks(sbi, segno, true);

fs/f2fs/f2fs.h

···
 	int fsync_mode;			/* fsync policy */
 	int fs_mode;			/* fs mode: LFS or ADAPTIVE */
 	int bggc_mode;			/* bggc mode: off, on or sync */
+	int memory_mode;		/* memory mode */
 	int discard_unit;		/*
 					 * discard command's offset/size should
 					 * be aligned to this unit: block,
···
 #define CP_PAUSE	0x00000040
 #define CP_RESIZE 	0x00000080
 
-#define MAX_DISCARD_BLOCKS(sbi)		BLKS_PER_SEC(sbi)
 #define DEF_MAX_DISCARD_REQUEST		8	/* issue 8 discards per round */
 #define DEF_MIN_DISCARD_ISSUE_TIME	50	/* 50 ms, if exists */
 #define DEF_MID_DISCARD_ISSUE_TIME	500	/* 500 ms, if device busy */
···
 #define RECOVERY_MAX_RA_BLOCKS		BIO_MAX_VECS
 #define RECOVERY_MIN_RA_BLOCKS		1
 
+#define F2FS_ONSTACK_PAGES	16	/* nr of onstack pages */
+
 struct rb_entry {
 	struct rb_node rb_node;		/* rb node located in rb-tree */
 	union {
···
 	FI_ENABLE_COMPRESS,	/* enable compression in "user" compression mode */
 	FI_COMPRESS_RELEASED,	/* compressed blocks were released */
 	FI_ALIGNED_WRITE,	/* enable aligned write */
+	FI_COW_FILE,		/* indicate COW file */
 	FI_MAX,			/* max flag, never be used */
 };
 
···
 	unsigned char i_compress_level;		/* compress level (lz4hc,zstd) */
 	unsigned short i_compress_flag;		/* compress flag */
 	unsigned int i_cluster_size;		/* cluster size */
+
+	unsigned int atomic_write_cnt;
 };
 
 static inline void get_extent_info(struct extent_info *ext,
···
 	bool retry;		/* need to reallocate block address */
 	int compr_blocks;	/* # of compressed block addresses */
 	bool encrypted;		/* indicate file is encrypted */
+	bool post_read;		/* require post read */
 	enum iostat_type io_type;	/* io type */
 	struct writeback_control *io_wbc; /* writeback control */
 	struct bio **bio;		/* bio for ipu */
···
 #ifdef CONFIG_BLK_DEV_ZONED
 	unsigned int nr_blkz;		/* Total number of zones */
 	unsigned long *blkz_seq;	/* Bitmap indicating sequential zones */
-	block_t *zone_capacity_blocks;  /* Array of zone capacity in blks */
 #endif
 };
 
···
 	DISCARD_UNIT_SEGMENT,	/* basic discard unit is segment */
 	DISCARD_UNIT_SECTION,	/* basic discard unit is section */
 };
+
+enum {
+	MEMORY_MODE_NORMAL,	/* memory mode for normal devices */
+	MEMORY_MODE_LOW,	/* memory mode for low memry devices */
+};
+
+
 
 static inline int f2fs_test_bit(unsigned int nr, char *addr);
 static inline void f2fs_set_bit(unsigned int nr, char *addr);
···
 	void *private;			/* payload buffer for specified decompression algorithm */
 	void *private2;			/* extra payload buffer */
 	struct work_struct verity_work;	/* work to verify the decompressed pages */
+	struct work_struct free_work;	/* work for late free this structure itself */
 };
 
 #define NULL_CLUSTER			((unsigned int)(~0))
···
 	unsigned int meta_ino_num;		/* meta inode number*/
 	unsigned int log_blocks_per_seg;	/* log2 blocks per segment */
 	unsigned int blocks_per_seg;		/* blocks per segment */
+	unsigned int unusable_blocks_per_sec;	/* unusable blocks per section */
 	unsigned int segs_per_sec;		/* segments per section */
 	unsigned int secs_per_zone;		/* sections per zone */
 	unsigned int total_sections;		/* total section count */
···
 
 	int max_fragment_chunk;			/* max chunk size for block fragmentation mode */
 	int max_fragment_hole;			/* max hole size for block fragmentation mode */
+
+	/* For atomic write statistics */
+	atomic64_t current_atomic_write;
+	s64 peak_atomic_write;
+	u64 committed_atomic_block;
+	u64 revoked_atomic_block;
 
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 	struct kmem_cache *page_array_slab;	/* page array entry */
···
 		dec_page_count(F2FS_I_SB(inode), F2FS_DIRTY_QDATA);
 }
 
+static inline void inc_atomic_write_cnt(struct inode *inode)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct f2fs_inode_info *fi = F2FS_I(inode);
+	u64 current_write;
+
+	fi->atomic_write_cnt++;
+	atomic64_inc(&sbi->current_atomic_write);
+	current_write = atomic64_read(&sbi->current_atomic_write);
+	if (current_write > sbi->peak_atomic_write)
+		sbi->peak_atomic_write = current_write;
+}
+
+static inline void release_atomic_write_cnt(struct inode *inode)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct f2fs_inode_info *fi = F2FS_I(inode);
+
+	atomic64_sub(fi->atomic_write_cnt, &sbi->current_atomic_write);
+	fi->atomic_write_cnt = 0;
+}
+
 static inline s64 get_pages(struct f2fs_sb_info *sbi, int count_type)
 {
 	return atomic_read(&sbi->nr_pages[count_type]);
···
 	}
 
 	return pagecache_get_page(mapping, index, fgp_flags, gfp_mask);
-}
-
-static inline void f2fs_copy_page(struct page *src, struct page *dst)
-{
-	char *src_kaddr = kmap(src);
-	char *dst_kaddr = kmap(dst);
-
-	memcpy(dst_kaddr, src_kaddr, PAGE_SIZE);
-	kunmap(dst);
-	kunmap(src);
 }
 
 static inline void f2fs_put_page(struct page *page, int unlock)
···
 static inline bool f2fs_is_atomic_file(struct inode *inode)
 {
 	return is_inode_flag_set(inode, FI_ATOMIC_FILE);
+}
+
+static inline bool f2fs_is_cow_file(struct inode *inode)
+{
+	return is_inode_flag_set(inode, FI_COW_FILE);
 }
 
 static inline bool f2fs_is_first_block_written(struct inode *inode)
···
 bool f2fs_is_compress_backend_ready(struct inode *inode);
 int f2fs_init_compress_mempool(void);
 void f2fs_destroy_compress_mempool(void);
-void f2fs_decompress_cluster(struct decompress_io_ctx *dic);
+void f2fs_decompress_cluster(struct decompress_io_ctx *dic, bool in_task);
 void f2fs_end_read_compressed_page(struct page *page, bool failed,
-							block_t blkaddr);
+				block_t blkaddr, bool in_task);
 bool f2fs_cluster_is_empty(struct compress_ctx *cc);
 bool f2fs_cluster_can_merge_page(struct compress_ctx *cc, pgoff_t index);
-bool f2fs_all_cluster_page_loaded(struct compress_ctx *cc, struct pagevec *pvec,
-				int index, int nr_pages);
+bool f2fs_all_cluster_page_ready(struct compress_ctx *cc, struct page **pages,
+				int index, int nr_pages, bool uptodate);
 bool f2fs_sanity_check_cluster(struct dnode_of_data *dn);
 void f2fs_compress_ctx_add_page(struct compress_ctx *cc, struct page *page);
 int f2fs_write_multi_pages(struct compress_ctx *cc,
···
 				unsigned nr_pages, sector_t *last_block_in_bio,
 				bool is_readahead, bool for_write);
 struct decompress_io_ctx *f2fs_alloc_dic(struct compress_ctx *cc);
-void f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed);
-void f2fs_put_page_dic(struct page *page);
+void f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed,
+				bool in_task);
+void f2fs_put_page_dic(struct page *page, bool in_task);
 unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn);
 int f2fs_init_compress_ctx(struct compress_ctx *cc);
 void f2fs_destroy_compress_ctx(struct compress_ctx *cc, bool reuse);
···
 }
 static inline int f2fs_init_compress_mempool(void) { return 0; }
 static inline void f2fs_destroy_compress_mempool(void) { }
-static inline void f2fs_decompress_cluster(struct decompress_io_ctx *dic) { }
+static inline void f2fs_decompress_cluster(struct decompress_io_ctx *dic,
+				bool in_task) { }
 static inline void f2fs_end_read_compressed_page(struct page *page,
-						bool failed, block_t blkaddr)
+				bool failed, block_t blkaddr, bool in_task)
 {
 	WARN_ON_ONCE(1);
 }
-static inline void f2fs_put_page_dic(struct page *page)
+static inline void f2fs_put_page_dic(struct page *page, bool in_task)
 {
 	WARN_ON_ONCE(1);
 }
···
 				unsigned int c_len) { }
 #endif
 
-static inline void set_compress_context(struct inode *inode)
+static inline int set_compress_context(struct inode *inode)
 {
+#ifdef CONFIG_F2FS_FS_COMPRESSION
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 
 	F2FS_I(inode)->i_compress_algorithm =
···
 	stat_inc_compr_inode(inode);
 	inc_compr_inode_stat(inode);
 	f2fs_mark_inode_dirty_sync(inode, true);
+	return 0;
+#else
+	return -EOPNOTSUPP;
+#endif
 }
 
 static inline bool f2fs_disable_compressed_file(struct inode *inode)
···
 	return F2FS_OPTION(sbi).fs_mode == FS_MODE_LFS;
 }
 
+static inline bool f2fs_low_mem_mode(struct f2fs_sb_info *sbi)
+{
+	return F2FS_OPTION(sbi).memory_mode == MEMORY_MODE_LOW;
+}
+
 static inline bool f2fs_may_compress(struct inode *inode)
 {
 	if (IS_SWAPFILE(inode) || f2fs_is_pinned_file(inode) ||
-				f2fs_is_atomic_file(inode))
+		f2fs_is_atomic_file(inode) || f2fs_has_inline_data(inode))
 		return false;
 	return S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode);
 }
···
 	/* disallow direct IO if any of devices has unaligned blksize */
 	if (f2fs_is_multi_device(sbi) && !sbi->aligned_blksize)
 		return true;
-	/*
-	 * for blkzoned device, fallback direct IO to buffered IO, so
-	 * all IOs can be serialized by log-structured write.
-	 */
-	if (f2fs_sb_has_blkzoned(sbi))
-		return true;
+
 	if (f2fs_lfs_mode(sbi) && (rw == WRITE)) {
 		if (block_unaligned_IO(inode, iocb, iter))
 			return true;

fs/f2fs/file.c

···
 				f2fs_put_page(psrc, 1);
 				return PTR_ERR(pdst);
 			}
-			f2fs_copy_page(psrc, pdst);
+			memcpy_page(pdst, 0, psrc, 0, PAGE_SIZE);
 			set_page_dirty(pdst);
 			f2fs_put_page(pdst, 1);
 			f2fs_put_page(psrc, 1);
···
 		return 0;
 
 	if (f2fs_is_pinned_file(inode)) {
-		block_t sec_blks = BLKS_PER_SEC(sbi);
+		block_t sec_blks = CAP_BLKS_PER_SEC(sbi);
 		block_t sec_len = roundup(map.m_len, sec_blks);
 
 		map.m_len = sec_blks;
···
 			atomic_read(&inode->i_writecount) != 1)
 		return 0;
 
-	if (f2fs_is_atomic_file(inode))
-		f2fs_abort_atomic_write(inode, true);
+	f2fs_abort_atomic_write(inode, true);
 	return 0;
 }
 
···
 	 * until all the writers close its file. Since this should be done
 	 * before dropping file lock, it needs to do in ->flush.
 	 */
-	if (f2fs_is_atomic_file(inode) &&
-			F2FS_I(inode)->atomic_write_task == current)
+	if (F2FS_I(inode)->atomic_write_task == current)
 		f2fs_abort_atomic_write(inode, true);
 	return 0;
 }
···
 		if (masked_flags & F2FS_COMPR_FL) {
 			if (!f2fs_disable_compressed_file(inode))
 				return -EINVAL;
-		}
-		if (iflags & F2FS_NOCOMP_FL)
-			return -EINVAL;
-		if (iflags & F2FS_COMPR_FL) {
+		} else {
 			if (!f2fs_may_compress(inode))
 				return -EINVAL;
-			if (S_ISREG(inode->i_mode) && inode->i_size)
+			if (S_ISREG(inode->i_mode) && F2FS_HAS_BLOCKS(inode))
 				return -EINVAL;
-
-			set_compress_context(inode);
+			if (set_compress_context(inode))
+				return -EOPNOTSUPP;
 		}
-	}
-	if ((iflags ^ masked_flags) & F2FS_NOCOMP_FL) {
-		if (masked_flags & F2FS_COMPR_FL)
-			return -EINVAL;
 	}
 
 	fi->i_flags = iflags | (fi->i_flags & ~mask);
···
 	spin_unlock(&sbi->inode_lock[ATOMIC_FILE]);
 
 	set_inode_flag(inode, FI_ATOMIC_FILE);
-	set_inode_flag(fi->cow_inode, FI_ATOMIC_FILE);
+	set_inode_flag(fi->cow_inode, FI_COW_FILE);
 	clear_inode_flag(fi->cow_inode, FI_INLINE_DATA);
 	f2fs_up_write(&fi->i_gc_rwsem[WRITE]);
 
 	f2fs_update_time(sbi, REQ_TIME);
 	fi->atomic_write_task = current;
 	stat_update_max_atomic_write(inode);
+	fi->atomic_write_cnt = 0;
 out:
 	inode_unlock(inode);
 	mnt_drop_write_file(filp);
···
 unlock_out:
 	inode_unlock(inode);
 	mnt_drop_write_file(filp);
+	return ret;
+}
+
+static int f2fs_ioc_abort_atomic_write(struct file *filp)
+{
+	struct inode *inode = file_inode(filp);
+	struct user_namespace *mnt_userns = file_mnt_user_ns(filp);
+	int ret;
+
+	if (!inode_owner_or_capable(mnt_userns, inode))
+		return -EACCES;
+
+	ret = mnt_want_write_file(filp);
+	if (ret)
+		return ret;
+
+	inode_lock(inode);
+
+	f2fs_abort_atomic_write(inode, true);
+
+	inode_unlock(inode);
+
+	mnt_drop_write_file(filp);
+	f2fs_update_time(F2FS_I_SB(inode), REQ_TIME);
 	return ret;
 }
 
···
 			ret = -EAGAIN;
 		goto out;
 	}
-	range->start += BLKS_PER_SEC(sbi);
+	range->start += CAP_BLKS_PER_SEC(sbi);
 	if (range->start <= end)
 		goto do_more;
 out:
···
 		goto out;
 	}
 
-	sec_num = DIV_ROUND_UP(total, BLKS_PER_SEC(sbi));
+	sec_num = DIV_ROUND_UP(total, CAP_BLKS_PER_SEC(sbi));
 
 	/*
 	 * make sure there are enough free section for LFS allocation, this can
···
 
 	for (i = 0; i < page_len; i++, redirty_idx++) {
 		page = find_lock_page(mapping, redirty_idx);
-		if (!page) {
-			ret = -ENOMEM;
-			break;
-		}
+
+		/* It will never fail, when page has pinned above */
+		f2fs_bug_on(F2FS_I_SB(inode), !page);
+
 		set_page_dirty(page);
 		f2fs_put_page(page, 1);
 		f2fs_put_page(page, 0);
···
 
 	if (!f2fs_is_compress_backend_ready(inode)) {
 		ret = -EOPNOTSUPP;
+		goto out;
+	}
+
+	if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
+		ret = -EINVAL;
 		goto out;
 	}
 
···
 		goto out;
 	}
 
+	if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) {
+		ret = -EINVAL;
+		goto out;
+	}
+
 	ret = filemap_write_and_wait_range(inode->i_mapping, 0, LLONG_MAX);
 	if (ret)
 		goto out;
···
 		return f2fs_ioc_start_atomic_write(filp);
 	case F2FS_IOC_COMMIT_ATOMIC_WRITE:
 		return f2fs_ioc_commit_atomic_write(filp);
+	case F2FS_IOC_ABORT_ATOMIC_WRITE:
+		return f2fs_ioc_abort_atomic_write(filp);
 	case F2FS_IOC_START_VOLATILE_WRITE:
 	case F2FS_IOC_RELEASE_VOLATILE_WRITE:
-	case F2FS_IOC_ABORT_VOLATILE_WRITE:
 		return -EOPNOTSUPP;
 	case F2FS_IOC_SHUTDOWN:
 		return f2fs_ioc_shutdown(filp, arg);
···
 	case F2FS_IOC_COMMIT_ATOMIC_WRITE:
 	case F2FS_IOC_START_VOLATILE_WRITE:
 	case F2FS_IOC_RELEASE_VOLATILE_WRITE:
-	case F2FS_IOC_ABORT_VOLATILE_WRITE:
+	case F2FS_IOC_ABORT_ATOMIC_WRITE:
 	case F2FS_IOC_SHUTDOWN:
 	case FITRIM:
 	case FS_IOC_SET_ENCRYPTION_POLICY:

fs/f2fs/gc.c

···
 		gc_control.nr_free_secs = foreground ? 1 : 0;
 
 		/* if return value is not zero, no victim was selected */
-		if (f2fs_gc(sbi, &gc_control))
-			wait_ms = gc_th->no_gc_sleep_time;
+		if (f2fs_gc(sbi, &gc_control)) {
+			/* don't bother wait_ms by foreground gc */
+			if (!foreground)
+				wait_ms = gc_th->no_gc_sleep_time;
+		}
 
 		if (foreground)
 			wake_up_all(&gc_th->fggc_wq);
···
 	unsigned long long age, u, accu;
 	unsigned long long max_mtime = sit_i->dirty_max_mtime;
 	unsigned long long min_mtime = sit_i->dirty_min_mtime;
-	unsigned int sec_blocks = BLKS_PER_SEC(sbi);
+	unsigned int sec_blocks = CAP_BLKS_PER_SEC(sbi);
 	unsigned int vblocks;
 	unsigned int dirty_threshold = max(am->max_candidate_count,
 					am->candidate_ratio *
···
 		 */
 		if ((gc_type == BG_GC && has_not_enough_free_secs(sbi, 0, 0)) ||
 			(!force_migrate && get_valid_blocks(sbi, segno, true) ==
-							BLKS_PER_SEC(sbi)))
+							CAP_BLKS_PER_SEC(sbi)))
 			return submitted;
 
 		if (check_valid_map(sbi, segno, off) == 0)

fs/f2fs/gc.h

···
 	return free_blks - ovp_blks;
 }
 
-static inline block_t limit_invalid_user_blocks(struct f2fs_sb_info *sbi)
+static inline block_t limit_invalid_user_blocks(block_t user_block_count)
 {
-	return (long)(sbi->user_block_count * LIMIT_INVALID_BLOCK) / 100;
+	return (long)(user_block_count * LIMIT_INVALID_BLOCK) / 100;
 }
 
-static inline block_t limit_free_user_blocks(struct f2fs_sb_info *sbi)
+static inline block_t limit_free_user_blocks(block_t reclaimable_user_blocks)
 {
-	block_t reclaimable_user_blocks = sbi->user_block_count -
-		written_block_count(sbi);
 	return (long)(reclaimable_user_blocks * LIMIT_FREE_BLOCK) / 100;
 }
 
···
 
 static inline bool has_enough_invalid_blocks(struct f2fs_sb_info *sbi)
 {
-	block_t invalid_user_blocks = sbi->user_block_count -
-					written_block_count(sbi);
+	block_t user_block_count = sbi->user_block_count;
+	block_t invalid_user_blocks = user_block_count -
+		written_block_count(sbi);
 	/*
 	 * Background GC is triggered with the following conditions.
 	 * 1. There are a number of invalid blocks.
 	 * 2. There is not enough free space.
 	 */
-	if (invalid_user_blocks > limit_invalid_user_blocks(sbi) &&
-			free_user_blocks(sbi) < limit_free_user_blocks(sbi))
-		return true;
-	return false;
+	return (invalid_user_blocks >
+			limit_invalid_user_blocks(user_block_count) &&
+		free_user_blocks(sbi) <
+			limit_free_user_blocks(invalid_user_blocks));
 }

fs/f2fs/inode.c

···
 	nid_t xnid = F2FS_I(inode)->i_xattr_nid;
 	int err = 0;
 
-	if (f2fs_is_atomic_file(inode))
-		f2fs_abort_atomic_write(inode, true);
+	f2fs_abort_atomic_write(inode, true);
 
 	trace_f2fs_evict_inode(inode);
 	truncate_inode_pages_final(&inode->i_data);

fs/f2fs/node.c

···
 		dec_valid_node_count(sbi, dn->inode, !ofs);
 		goto fail;
 	}
-	f2fs_bug_on(sbi, new_ni.blk_addr != NULL_ADDR);
+	if (unlikely(new_ni.blk_addr != NULL_ADDR)) {
+		err = -EFSCORRUPTED;
+		set_sbi_flag(sbi, SBI_NEED_FSCK);
+		goto fail;
+	}
 #endif
 	new_ni.nid = dn->nid;
 	new_ni.ino = dn->inode->i_ino;
···
 		for (i = 0; i < nr_pages; i++) {
 			struct page *page = pvec.pages[i];
 			bool submitted = false;
-			bool may_dirty = true;
 
 			/* give a priority to WB_SYNC threads */
 			if (atomic_read(&sbi->wb_sync_req[NODE]) &&
···
 			}
 
 			/* flush dirty inode */
-			if (IS_INODE(page) && may_dirty) {
-				may_dirty = false;
-				if (flush_dirty_inode(page))
-					goto lock_node;
-			}
+			if (IS_INODE(page) && flush_dirty_inode(page))
+				goto lock_node;
 write_node:
 			f2fs_wait_on_page_writeback(page, NODE, true, true);
 

fs/f2fs/segment.c

···
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct f2fs_inode_info *fi = F2FS_I(inode);
 
-	if (f2fs_is_atomic_file(inode)) {
-		if (clean)
-			truncate_inode_pages_final(inode->i_mapping);
-		clear_inode_flag(fi->cow_inode, FI_ATOMIC_FILE);
-		iput(fi->cow_inode);
-		fi->cow_inode = NULL;
-		clear_inode_flag(inode, FI_ATOMIC_FILE);
+	if (!f2fs_is_atomic_file(inode))
+		return;
 
-		spin_lock(&sbi->inode_lock[ATOMIC_FILE]);
-		sbi->atomic_files--;
-		spin_unlock(&sbi->inode_lock[ATOMIC_FILE]);
-	}
+	if (clean)
+		truncate_inode_pages_final(inode->i_mapping);
+	clear_inode_flag(fi->cow_inode, FI_COW_FILE);
+	iput(fi->cow_inode);
+	fi->cow_inode = NULL;
+	release_atomic_write_cnt(inode);
+	clear_inode_flag(inode, FI_ATOMIC_FILE);
+
+	spin_lock(&sbi->inode_lock[ATOMIC_FILE]);
+	sbi->atomic_files--;
+	spin_unlock(&sbi->inode_lock[ATOMIC_FILE]);
 }
 
 static int __replace_atomic_write_block(struct inode *inode, pgoff_t index,
···
 	}
 
 out:
+	if (ret)
+		sbi->revoked_atomic_block += fi->atomic_write_cnt;
+	else
+		sbi->committed_atomic_block += fi->atomic_write_cnt;
+
 	__complete_revoke_list(inode, &revoke_list, ret ? true : false);
 
 	return ret;
···
 				get_valid_blocks(sbi, segno, true);
 
 			f2fs_bug_on(sbi, unlikely(!valid_blocks ||
-					valid_blocks == BLKS_PER_SEC(sbi)));
+					valid_blocks == CAP_BLKS_PER_SEC(sbi)));
 
 			if (!IS_CURSEC(sbi, secno))
 				set_bit(secno, dirty_i->dirty_secmap);
···
 			unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
 
 			if (!valid_blocks ||
-					valid_blocks == BLKS_PER_SEC(sbi)) {
+					valid_blocks == CAP_BLKS_PER_SEC(sbi)) {
 				clear_bit(secno, dirty_i->dirty_secmap);
 				return;
 			}
···
 			return CURSEG_COLD_DATA;
 		if (file_is_hot(inode) ||
 				is_inode_flag_set(inode, FI_HOT_DATA) ||
-				f2fs_is_atomic_file(inode))
+				f2fs_is_cow_file(inode))
 			return CURSEG_HOT_DATA;
 		return f2fs_rw_hint_to_seg_type(inode->i_write_hint);
 	} else {
···
 		goto drop_bio;
 	}
 
-	invalidate_mapping_pages(META_MAPPING(sbi),
+	if (fio->post_read)
+		invalidate_mapping_pages(META_MAPPING(sbi),
 				fio->new_blkaddr, fio->new_blkaddr);
 
 	stat_inc_inplace_blocks(fio->sbi);
···
 void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr,
 								block_t len)
 {
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	block_t i;
+
+	if (!f2fs_post_read_required(inode))
+		return;
 
 	for (i = 0; i < len; i++)
 		f2fs_wait_on_block_writeback(inode, blkaddr + i);
+
+	invalidate_mapping_pages(META_MAPPING(sbi), blkaddr, blkaddr + len - 1);
 }
 
 static int read_compacted_summaries(struct f2fs_sb_info *sbi)
···
 				return err;
 			seg_info_from_raw_sit(se, &sit);
 
+			if (se->type >= NR_PERSISTENT_LOG) {
+				f2fs_err(sbi, "Invalid segment type: %u, segno: %u",
+							se->type, start);
+				return -EFSCORRUPTED;
+			}
+
 			sit_valid_blocks[SE_PAGETYPE(se)] += se->valid_blocks;
 
 			if (f2fs_block_unit_discard(sbi)) {
···
 		if (err)
 			break;
 		seg_info_from_raw_sit(se, &sit);
+
+		if (se->type >= NR_PERSISTENT_LOG) {
+			f2fs_err(sbi, "Invalid segment type: %u, segno: %u",
+							se->type, start);
+			err = -EFSCORRUPTED;
+			break;
+		}
 
 		sit_valid_blocks[SE_PAGETYPE(se)] += se->valid_blocks;
 
···
 	struct free_segmap_info *free_i = FREE_I(sbi);
 	unsigned int segno = 0, offset = 0, secno;
 	block_t valid_blocks, usable_blks_in_seg;
-	block_t blks_per_sec = BLKS_PER_SEC(sbi);
 
 	while (1) {
 		/* find dirty segment based on free segmap */
···
 		valid_blocks = get_valid_blocks(sbi, segno, true);
 		secno = GET_SEC_FROM_SEG(sbi, segno);
 
-		if (!valid_blocks || valid_blocks == blks_per_sec)
+		if (!valid_blocks || valid_blocks == CAP_BLKS_PER_SEC(sbi))
 			continue;
 		if (IS_CURSEC(sbi, secno))
 			continue;
···
 static inline unsigned int f2fs_usable_zone_segs_in_sec(
 		struct f2fs_sb_info *sbi, unsigned int segno)
 {
-	unsigned int dev_idx, zone_idx, unusable_segs_in_sec;
+	unsigned int dev_idx, zone_idx;
 
 	dev_idx = f2fs_target_device_index(sbi, START_BLOCK(sbi, segno));
 	zone_idx = get_zone_idx(sbi, GET_SEC_FROM_SEG(sbi, segno), dev_idx);
···
 	if (is_conv_zone(sbi, zone_idx, dev_idx))
 		return sbi->segs_per_sec;
 
-	/*
-	 * If the zone_capacity_blocks array is NULL, then zone capacity
-	 * is equal to the zone size for all zones
-	 */
-	if (!FDEV(dev_idx).zone_capacity_blocks)
+	if (!sbi->unusable_blocks_per_sec)
 		return sbi->segs_per_sec;
 
 	/* Get the segment count beyond zone capacity block */
-	unusable_segs_in_sec = (sbi->blocks_per_blkz -
-				FDEV(dev_idx).zone_capacity_blocks[zone_idx]) >>
-				sbi->log_blocks_per_seg;
-	return sbi->segs_per_sec - unusable_segs_in_sec;
+	return sbi->segs_per_sec - (sbi->unusable_blocks_per_sec >>
+						sbi->log_blocks_per_seg);
 }
 
 /*
···
 	if (is_conv_zone(sbi, zone_idx, dev_idx))
 		return sbi->blocks_per_seg;
 
-	if (!FDEV(dev_idx).zone_capacity_blocks)
+	if (!sbi->unusable_blocks_per_sec)
 		return sbi->blocks_per_seg;
 
 	sec_start_blkaddr = START_BLOCK(sbi, GET_SEG_FROM_SEC(sbi, secno));
-	sec_cap_blkaddr = sec_start_blkaddr +
-				FDEV(dev_idx).zone_capacity_blocks[zone_idx];
+	sec_cap_blkaddr = sec_start_blkaddr + CAP_BLKS_PER_SEC(sbi);
 
 	/*
 	 * If segment starts before zone capacity and spans beyond

fs/f2fs/segment.h

···
 		GET_SEGNO_FROM_SEG0(sbi, blk_addr)))
 #define BLKS_PER_SEC(sbi)					\
 	((sbi)->segs_per_sec * (sbi)->blocks_per_seg)
+#define CAP_BLKS_PER_SEC(sbi)					\
+	((sbi)->segs_per_sec * (sbi)->blocks_per_seg -		\
+	 (sbi)->unusable_blocks_per_sec)
 #define GET_SEC_FROM_SEG(sbi, segno)				\
 	(((segno) == -1) ? -1: (segno) / (sbi)->segs_per_sec)
 #define GET_SEG_FROM_SEC(sbi, secno)				\
···
 					get_pages(sbi, F2FS_DIRTY_DENTS) +
 					get_pages(sbi, F2FS_DIRTY_IMETA);
 	unsigned int total_dent_blocks = get_pages(sbi, F2FS_DIRTY_DENTS);
-	unsigned int node_secs = total_node_blocks / BLKS_PER_SEC(sbi);
-	unsigned int dent_secs = total_dent_blocks / BLKS_PER_SEC(sbi);
-	unsigned int node_blocks = total_node_blocks % BLKS_PER_SEC(sbi);
-	unsigned int dent_blocks = total_dent_blocks % BLKS_PER_SEC(sbi);
+	unsigned int node_secs = total_node_blocks / CAP_BLKS_PER_SEC(sbi);
+	unsigned int dent_secs = total_dent_blocks / CAP_BLKS_PER_SEC(sbi);
+	unsigned int node_blocks = total_node_blocks % CAP_BLKS_PER_SEC(sbi);
+	unsigned int dent_blocks = total_dent_blocks % CAP_BLKS_PER_SEC(sbi);
 	unsigned int free, need_lower, need_upper;
 
 	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))

fs/f2fs/super.c

···
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/fs.h>
+#include <linux/fs_context.h>
 #include <linux/sched/mm.h>
 #include <linux/statfs.h>
 #include <linux/buffer_head.h>
···
 	Opt_gc_merge,
 	Opt_nogc_merge,
 	Opt_discard_unit,
+	Opt_memory_mode,
 	Opt_err,
 };
 
···
 	{Opt_gc_merge, "gc_merge"},
 	{Opt_nogc_merge, "nogc_merge"},
 	{Opt_discard_unit, "discard_unit=%s"},
+	{Opt_memory_mode, "memory=%s"},
 	{Opt_err, NULL},
 };
 
···
 					  bool is_remount)
 {
 	struct f2fs_sb_info *sbi = F2FS_SB(sb);
-#ifdef CONFIG_FS_ENCRYPTION
+	struct fs_parameter param = {
+		.type = fs_value_is_string,
+		.string = arg->from ? arg->from : "",
+	};
+	struct fscrypt_dummy_policy *policy =
+		&F2FS_OPTION(sbi).dummy_enc_policy;
 	int err;
+
+	if (!IS_ENABLED(CONFIG_FS_ENCRYPTION)) {
+		f2fs_warn(sbi, "test_dummy_encryption option not supported");
+		return -EINVAL;
+	}
 
 	if (!f2fs_sb_has_encrypt(sbi)) {
 		f2fs_err(sbi, "Encrypt feature is off");
···
 	 * needed to allow it to be set or changed during remount.  We do allow
 	 * it to be specified during remount, but only if there is no change.
 	 */
-	if (is_remount && !F2FS_OPTION(sbi).dummy_enc_policy.policy) {
+	if (is_remount && !fscrypt_is_dummy_policy_set(policy)) {
 		f2fs_warn(sbi, "Can't set test_dummy_encryption on remount");
 		return -EINVAL;
 	}
-	err = fscrypt_set_test_dummy_encryption(
-		sb, arg->from, &F2FS_OPTION(sbi).dummy_enc_policy);
+
+	err = fscrypt_parse_test_dummy_encryption(&param, policy);
 	if (err) {
 		if (err == -EEXIST)
 			f2fs_warn(sbi,
···
 				  opt, err);
 		return -EINVAL;
 	}
+	err = fscrypt_add_test_dummy_key(sb, policy);
+	if (err) {
+		f2fs_warn(sbi, "Error adding test dummy encryption key [%d]",
+			  err);
+		return err;
+	}
 	f2fs_warn(sbi, "Test dummy encryption mode enabled");
 	return 0;
-#else
-	f2fs_warn(sbi, "test_dummy_encryption option not supported");
-	return -EINVAL;
-#endif
 }
 
 #ifdef CONFIG_F2FS_FS_COMPRESSION
···
 			}
 			kfree(name);
 			break;
+		case Opt_memory_mode:
+			name = match_strdup(&args[0]);
+			if (!name)
+				return -ENOMEM;
+			if (!strcmp(name, "normal")) {
+				F2FS_OPTION(sbi).memory_mode =
+						MEMORY_MODE_NORMAL;
+			} else if (!strcmp(name, "low")) {
+				F2FS_OPTION(sbi).memory_mode =
+						MEMORY_MODE_LOW;
+			} else {
+				kfree(name);
+				return -EINVAL;
+			}
+			kfree(name);
+			break;
 		default:
 			f2fs_err(sbi, "Unrecognized mount option \"%s\" or missing value",
 				 p);
···
 			atomic_inc(&inode->i_count);
 			spin_unlock(&inode->i_lock);
 
-			if (f2fs_is_atomic_file(inode))
-				f2fs_abort_atomic_write(inode, true);
+			f2fs_abort_atomic_write(inode, true);
 
 			/* should remain fi->extent_tree for writepage */
 			f2fs_destroy_extent_node(inode);
···
 		blkdev_put(FDEV(i).bdev, FMODE_EXCL);
 #ifdef CONFIG_BLK_DEV_ZONED
 		kvfree(FDEV(i).blkz_seq);
-		kfree(FDEV(i).zone_capacity_blocks);
 #endif
 	}
 	kvfree(sbi->devs);
···
 	else if (F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_SECTION)
 		seq_printf(seq, ",discard_unit=%s", "section");
 
+	if (F2FS_OPTION(sbi).memory_mode == MEMORY_MODE_NORMAL)
+		seq_printf(seq, ",memory=%s", "normal");
+	else if (F2FS_OPTION(sbi).memory_mode == MEMORY_MODE_LOW)
+		seq_printf(seq, ",memory=%s", "low");
+
 	return 0;
 }
 
···
 	F2FS_OPTION(sbi).compress_ext_cnt = 0;
 	F2FS_OPTION(sbi).compress_mode = COMPR_MODE_FS;
 	F2FS_OPTION(sbi).bggc_mode = BGGC_MODE_ON;
+	F2FS_OPTION(sbi).memory_mode = MEMORY_MODE_NORMAL;
 
 	sbi->sb->s_flags &= ~SB_INLINECRYPT;
 
···
 	sbi->max_fragment_chunk = DEF_FRAGMENT_SIZE;
 	sbi->max_fragment_hole = DEF_FRAGMENT_SIZE;
 	spin_lock_init(&sbi->gc_urgent_high_lock);
+	atomic64_set(&sbi->current_atomic_write, 0);
 
 	sbi->dir_level = DEF_DIR_LEVEL;
 	sbi->interval_time[CP_TIME] = DEF_CP_INTERVAL;
···
 #ifdef CONFIG_BLK_DEV_ZONED
 
 struct f2fs_report_zones_args {
+	struct f2fs_sb_info *sbi;
 	struct f2fs_dev_info *dev;
-	bool zone_cap_mismatch;
 };
 
 static int f2fs_report_zone_cb(struct blk_zone *zone, unsigned int idx,
 			      void *data)
 {
 	struct f2fs_report_zones_args *rz_args = data;
+	block_t unusable_blocks = (zone->len - zone->capacity) >>
+					F2FS_LOG_SECTORS_PER_BLOCK;
 
 	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
 		return 0;
 
 	set_bit(idx, rz_args->dev->blkz_seq);
-	rz_args->dev->zone_capacity_blocks[idx] = zone->capacity >>
-						F2FS_LOG_SECTORS_PER_BLOCK;
-	if (zone->len != zone->capacity && !rz_args->zone_cap_mismatch)
-		rz_args->zone_cap_mismatch = true;
-
+	if (!rz_args->sbi->unusable_blocks_per_sec) {
+		rz_args->sbi->unusable_blocks_per_sec = unusable_blocks;
+		return 0;
+	}
+	if (rz_args->sbi->unusable_blocks_per_sec != unusable_blocks) {
+		f2fs_err(rz_args->sbi, "F2FS supports single zone capacity\n");
+		return -EINVAL;
+	}
 	return 0;
 }
 
···
 	if (!FDEV(devi).blkz_seq)
 		return -ENOMEM;
 
-	/* Get block zones type and zone-capacity */
-	FDEV(devi).zone_capacity_blocks = f2fs_kzalloc(sbi,
-					FDEV(devi).nr_blkz * sizeof(block_t),
-					GFP_KERNEL);
-	if (!FDEV(devi).zone_capacity_blocks)
-		return -ENOMEM;
-
+	rep_zone_arg.sbi = sbi;
 	rep_zone_arg.dev = &FDEV(devi);
-	rep_zone_arg.zone_cap_mismatch = false;
 
 	ret = blkdev_report_zones(bdev, 0, BLK_ALL_ZONES, f2fs_report_zone_cb,
 				  &rep_zone_arg);
 	if (ret < 0)
 		return ret;
-
-	if (!rep_zone_arg.zone_cap_mismatch) {
-		kfree(FDEV(devi).zone_capacity_blocks);
-		FDEV(devi).zone_capacity_blocks = NULL;
-	}
-
 	return 0;
 }
 #endif

fs/f2fs/sysfs.c

···
 			sbi->gc_reclaimed_segs[sbi->gc_segment_mode]);
 	}
 
+	if (!strcmp(a->attr.name, "current_atomic_write")) {
+		s64 current_write = atomic64_read(&sbi->current_atomic_write);
+
+		return sysfs_emit(buf, "%lld\n", current_write);
+	}
+
+	if (!strcmp(a->attr.name, "peak_atomic_write"))
+		return sysfs_emit(buf, "%lld\n", sbi->peak_atomic_write);
+
+	if (!strcmp(a->attr.name, "committed_atomic_block"))
+		return sysfs_emit(buf, "%llu\n", sbi->committed_atomic_block);
+
+	if (!strcmp(a->attr.name, "revoked_atomic_block"))
+		return sysfs_emit(buf, "%llu\n", sbi->revoked_atomic_block);
+
 	ui = (unsigned int *)(ptr + a->offset);
 
 	return sprintf(buf, "%u\n", *ui);
···
 		return count;
 	}
 
+	if (!strcmp(a->attr.name, "peak_atomic_write")) {
+		if (t != 0)
+			return -EINVAL;
+		sbi->peak_atomic_write = 0;
+		return count;
+	}
+
+	if (!strcmp(a->attr.name, "committed_atomic_block")) {
+		if (t != 0)
+			return -EINVAL;
+		sbi->committed_atomic_block = 0;
+		return count;
+	}
+
+	if (!strcmp(a->attr.name, "revoked_atomic_block")) {
+		if (t != 0)
+			return -EINVAL;
+		sbi->revoked_atomic_block = 0;
+		return count;
+	}
+
 	*ui = (unsigned int)t;
 
 	return count;
···
 	.offset = _offset					\
 }
 
+#define F2FS_RO_ATTR(struct_type, struct_name, name, elname)	\
+	F2FS_ATTR_OFFSET(struct_type, name, 0444,		\
+		f2fs_sbi_show, NULL,				\
+		offsetof(struct struct_name, elname))
+
 #define F2FS_RW_ATTR(struct_type, struct_name, name, elname)	\
 	F2FS_ATTR_OFFSET(struct_type, name, 0644,		\
 		f2fs_sbi_show, f2fs_sbi_store,			\
···
 #endif /* CONFIG_FS_ENCRYPTION */
 #ifdef CONFIG_BLK_DEV_ZONED
 F2FS_FEATURE_RO_ATTR(block_zoned);
+F2FS_RO_ATTR(F2FS_SBI, f2fs_sb_info, unusable_blocks_per_sec,
+					unusable_blocks_per_sec);
 #endif
 F2FS_FEATURE_RO_ATTR(atomic_write);
 F2FS_FEATURE_RO_ATTR(extra_attr);
···
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, gc_reclaimed_segments, gc_reclaimed_segs);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, max_fragment_chunk, max_fragment_chunk);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, max_fragment_hole, max_fragment_hole);
+
+/* For atomic write */
+F2FS_RO_ATTR(F2FS_SBI, f2fs_sb_info, current_atomic_write, current_atomic_write);
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, peak_atomic_write, peak_atomic_write);
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, committed_atomic_block, committed_atomic_block);
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, revoked_atomic_block, revoked_atomic_block);
 
 #define ATTR_LIST(name) (&f2fs_attr_##name.attr)
 static struct attribute *f2fs_attrs[] = {
···
 	ATTR_LIST(moved_blocks_background),
 	ATTR_LIST(avg_vblocks),
 #endif
+#ifdef CONFIG_BLK_DEV_ZONED
+	ATTR_LIST(unusable_blocks_per_sec),
+#endif
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 	ATTR_LIST(compr_written_block),
 	ATTR_LIST(compr_saved_block),
···
 	ATTR_LIST(gc_reclaimed_segments),
 	ATTR_LIST(max_fragment_chunk),
 	ATTR_LIST(max_fragment_hole),
+	ATTR_LIST(current_atomic_write),
+	ATTR_LIST(peak_atomic_write),
+	ATTR_LIST(committed_atomic_block),
+	ATTR_LIST(revoked_atomic_block),
 	NULL,
 };
 ATTRIBUTE_GROUPS(f2fs);

include/uapi/linux/f2fs.h

···
 #define F2FS_IOC_COMMIT_ATOMIC_WRITE	_IO(F2FS_IOCTL_MAGIC, 2)
 #define F2FS_IOC_START_VOLATILE_WRITE	_IO(F2FS_IOCTL_MAGIC, 3)
 #define F2FS_IOC_RELEASE_VOLATILE_WRITE	_IO(F2FS_IOCTL_MAGIC, 4)
-#define F2FS_IOC_ABORT_VOLATILE_WRITE	_IO(F2FS_IOCTL_MAGIC, 5)
+#define F2FS_IOC_ABORT_ATOMIC_WRITE	_IO(F2FS_IOCTL_MAGIC, 5)
 #define F2FS_IOC_GARBAGE_COLLECT	_IOW(F2FS_IOCTL_MAGIC, 6, __u32)
 #define F2FS_IOC_WRITE_CHECKPOINT	_IO(F2FS_IOCTL_MAGIC, 7)
 #define F2FS_IOC_DEFRAGMENT		_IOWR(F2FS_IOCTL_MAGIC, 8,	\