Merge tag 'md-next-20230729' of...

Merge tag 'md-next-20230729' of https://git.kernel.org/pub/scm/linux/kernel/git/song/md into for-6.6/block

Pull MD updates from Song:

"1. Deprecate bitmap file support, by Christoph Hellwig;
 2. Fix deadlock with md sync thread, by Yu Kuai;
 3. Refactor md io accounting, by Yu Kuai;
 4. Various non-urgent fixes by Li Nan, Yu Kuai, and Jack Wang."

* tag 'md-next-20230729' of https://git.kernel.org/pub/scm/linux/kernel/git/song/md: (36 commits)
  md/md-bitmap: hold 'reconfig_mutex' in backlog_store()
  md/md-bitmap: remove unnecessary local variable in backlog_store()
  md/raid10: use dereference_rdev_and_rrdev() to get devices
  md/raid10: factor out dereference_rdev_and_rrdev()
  md/raid10: check replacement and rdev to prevent submit the same io twice
  md/raid1: Avoid lock contention from wake_up()
  md: restore 'noio_flag' for the last mddev_resume()
  md: don't quiesce in mddev_suspend()
  md: remove redundant check in fix_read_error()
  md/raid10: optimize fix_read_error
  md/raid1: prioritize adding disk to 'removed' mirror
  md/md-faulty: enable io accounting
  md/md-linear: enable io accounting
  md/md-multipath: enable io accounting
  md/raid10: switch to use md_account_bio() for io accounting
  md/raid1: switch to use md_account_bio() for io accounting
  raid5: fix missing io accounting in raid5_align_endio()
  md: also clone new io if io accounting is disabled
  md: move initialization and destruction of 'io_acct_set' to md.c
  md: deprecate bitmap file support
  ...

drivers/md/Kconfig

@@ -50,6 +50,16 @@ config MD_AUTODETECT
 
 	  If unsure, say Y.
 
+config MD_BITMAP_FILE
+	bool "MD bitmap file support (deprecated)"
+	default y
+	help
+	  If you say Y here, support for write intent bitmaps in files on an
+	  external file system is enabled.  This is an alternative to the internal
+	  bitmaps near the MD superblock, and very problematic code that abuses
+	  various kernel APIs and can only work with files on a file system not
+	  actually sitting on the MD device.
+
 config MD_LINEAR
 	tristate "Linear (append) mode (deprecated)"
 	depends on BLK_DEV_MD

drivers/md/dm-raid.c

@@ -3725,7 +3725,6 @@ static int raid_message(struct dm_target *ti, unsigned int argc, char **argv,
 	if (!strcasecmp(argv[0], "idle") || !strcasecmp(argv[0], "frozen")) {
 		if (mddev->sync_thread) {
 			set_bit(MD_RECOVERY_INTR, &mddev->recovery);
-			md_unregister_thread(&mddev->sync_thread);
 			md_reap_sync_thread(mddev);
 		}
 	} else if (decipher_sync_action(mddev, mddev->recovery) != st_idle)

drivers/md/md-bitmap.c

@@ -139,29 +139,26 @@ static void md_bitmap_checkfree(struct bitmap_counts *bitmap, unsigned long page
  */
 
 /* IO operations when bitmap is stored near all superblocks */
+
+/* choose a good rdev and read the page from there */
 static int read_sb_page(struct mddev *mddev, loff_t offset,
-			struct page *page,
-			unsigned long index, int size)
+		struct page *page, unsigned long index, int size)
 {
-	/* choose a good rdev and read the page from there */
 
+	sector_t sector = mddev->bitmap_info.offset + offset +
+		index * (PAGE_SIZE / SECTOR_SIZE);
 	struct md_rdev *rdev;
-	sector_t target;
 
 	rdev_for_each(rdev, mddev) {
-		if (! test_bit(In_sync, &rdev->flags)
-		    || test_bit(Faulty, &rdev->flags)
-		    || test_bit(Bitmap_sync, &rdev->flags))
-			continue;
+		u32 iosize = roundup(size, bdev_logical_block_size(rdev->bdev));
 
-		target = offset + index * (PAGE_SIZE/512);
+		if (!test_bit(In_sync, &rdev->flags) ||
+		    test_bit(Faulty, &rdev->flags) ||
+		    test_bit(Bitmap_sync, &rdev->flags))
+			continue;
 
-		if (sync_page_io(rdev, target,
-				 roundup(size, bdev_logical_block_size(rdev->bdev)),
-				 page, REQ_OP_READ, true)) {
-			page->index = index;
+		if (sync_page_io(rdev, sector, iosize, page, REQ_OP_READ, true))
 			return 0;
-		}
 	}
 	return -EIO;
 }
@@ -225,18 +222,19 @@ static unsigned int bitmap_io_size(unsigned int io_size, unsigned int opt_size,
 }
 
 static int __write_sb_page(struct md_rdev *rdev, struct bitmap *bitmap,
-			   struct page *page)
+			   unsigned long pg_index, struct page *page)
 {
 	struct block_device *bdev;
 	struct mddev *mddev = bitmap->mddev;
 	struct bitmap_storage *store = &bitmap->storage;
 	loff_t sboff, offset = mddev->bitmap_info.offset;
-	sector_t ps, doff;
+	sector_t ps = pg_index * PAGE_SIZE / SECTOR_SIZE;
 	unsigned int size = PAGE_SIZE;
 	unsigned int opt_size = PAGE_SIZE;
+	sector_t doff;
 
 	bdev = (rdev->meta_bdev) ? rdev->meta_bdev : rdev->bdev;
-	if (page->index == store->file_pages - 1) {
+	if (pg_index == store->file_pages - 1) {
 		unsigned int last_page_size = store->bytes & (PAGE_SIZE - 1);
 
 		if (last_page_size == 0)
@@ -245,7 +243,6 @@ static int __write_sb_page(struct md_rdev *rdev, struct bitmap *bitmap,
 		opt_size = optimal_io_size(bdev, last_page_size, size);
 	}
 
-	ps = page->index * PAGE_SIZE / SECTOR_SIZE;
 	sboff = rdev->sb_start + offset;
 	doff = rdev->data_offset;
 
@@ -279,55 +276,41 @@ static int __write_sb_page(struct md_rdev *rdev, struct bitmap *bitmap,
 	return 0;
 }
 
-static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait)
+static void write_sb_page(struct bitmap *bitmap, unsigned long pg_index,
+			  struct page *page, bool wait)
 {
-	struct md_rdev *rdev;
 	struct mddev *mddev = bitmap->mddev;
-	int ret;
 
 	do {
-		rdev = NULL;
+		struct md_rdev *rdev = NULL;
+
 		while ((rdev = next_active_rdev(rdev, mddev)) != NULL) {
-			ret = __write_sb_page(rdev, bitmap, page);
-			if (ret)
-				return ret;
+			if (__write_sb_page(rdev, bitmap, pg_index, page) < 0) {
+				set_bit(BITMAP_WRITE_ERROR, &bitmap->flags);
+				return;
+			}
 		}
 	} while (wait && md_super_wait(mddev) < 0);
-
-	return 0;
 }
 
 static void md_bitmap_file_kick(struct bitmap *bitmap);
-/*
- * write out a page to a file
- */
-static void write_page(struct bitmap *bitmap, struct page *page, int wait)
-{
-	struct buffer_head *bh;
-
-	if (bitmap->storage.file == NULL) {
-		switch (write_sb_page(bitmap, page, wait)) {
-		case -EINVAL:
-			set_bit(BITMAP_WRITE_ERROR, &bitmap->flags);
-		}
-	} else {
-
-		bh = page_buffers(page);
 
-		while (bh && bh->b_blocknr) {
-			atomic_inc(&bitmap->pending_writes);
-			set_buffer_locked(bh);
-			set_buffer_mapped(bh);
-			submit_bh(REQ_OP_WRITE | REQ_SYNC, bh);
-			bh = bh->b_this_page;
-		}
+#ifdef CONFIG_MD_BITMAP_FILE
+static void write_file_page(struct bitmap *bitmap, struct page *page, int wait)
+{
+	struct buffer_head *bh = page_buffers(page);
 
-		if (wait)
-			wait_event(bitmap->write_wait,
-				   atomic_read(&bitmap->pending_writes)==0);
+	while (bh && bh->b_blocknr) {
+		atomic_inc(&bitmap->pending_writes);
+		set_buffer_locked(bh);
+		set_buffer_mapped(bh);
+		submit_bh(REQ_OP_WRITE | REQ_SYNC, bh);
+		bh = bh->b_this_page;
 	}
-	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
-		md_bitmap_file_kick(bitmap);
+
+	if (wait)
+		wait_event(bitmap->write_wait,
+			   atomic_read(&bitmap->pending_writes) == 0);
 }
 
 static void end_bitmap_write(struct buffer_head *bh, int uptodate)
@@ -364,10 +347,8 @@ static void free_buffers(struct page *page)
  * This usage is similar to how swap files are handled, and allows us
  * to write to a file with no concerns of memory allocation failing.
  */
-static int read_page(struct file *file, unsigned long index,
-		     struct bitmap *bitmap,
-		     unsigned long count,
-		     struct page *page)
+static int read_file_page(struct file *file, unsigned long index,
+		struct bitmap *bitmap, unsigned long count, struct page *page)
 {
 	int ret = 0;
 	struct inode *inode = file_inode(file);
@@ -415,7 +396,6 @@ static int read_page(struct file *file, unsigned long index,
 		blk_cur++;
 		bh = bh->b_this_page;
 	}
-	page->index = index;
 
 	wait_event(bitmap->write_wait,
 		   atomic_read(&bitmap->pending_writes)==0);
@@ -429,11 +409,45 @@ static int read_page(struct file *file, unsigned long index,
 		       ret);
 	return ret;
 }
+#else /* CONFIG_MD_BITMAP_FILE */
+static void write_file_page(struct bitmap *bitmap, struct page *page, int wait)
+{
+}
+static int read_file_page(struct file *file, unsigned long index,
+		struct bitmap *bitmap, unsigned long count, struct page *page)
+{
+	return -EIO;
+}
+static void free_buffers(struct page *page)
+{
+	put_page(page);
+}
+#endif /* CONFIG_MD_BITMAP_FILE */
 
 /*
  * bitmap file superblock operations
  */
 
+/*
+ * write out a page to a file
+ */
+static void filemap_write_page(struct bitmap *bitmap, unsigned long pg_index,
+			       bool wait)
+{
+	struct bitmap_storage *store = &bitmap->storage;
+	struct page *page = store->filemap[pg_index];
+
+	if (mddev_is_clustered(bitmap->mddev)) {
+		pg_index += bitmap->cluster_slot *
+			DIV_ROUND_UP(store->bytes, PAGE_SIZE);
+	}
+
+	if (store->file)
+		write_file_page(bitmap, page, wait);
+	else
+		write_sb_page(bitmap, pg_index, page, wait);
+}
+
 /*
  * md_bitmap_wait_writes() should be called before writing any bitmap
  * blocks, to ensure previous writes, particularly from
@@ -488,7 +502,12 @@ void md_bitmap_update_sb(struct bitmap *bitmap)
 	sb->sectors_reserved = cpu_to_le32(bitmap->mddev->
 					   bitmap_info.space);
 	kunmap_atomic(sb);
-	write_page(bitmap, bitmap->storage.sb_page, 1);
+
+	if (bitmap->storage.file)
+		write_file_page(bitmap, bitmap->storage.sb_page, 1);
+	else
+		write_sb_page(bitmap, bitmap->storage.sb_index,
+			      bitmap->storage.sb_page, 1);
 }
 EXPORT_SYMBOL(md_bitmap_update_sb);
 
@@ -540,7 +559,7 @@ static int md_bitmap_new_disk_sb(struct bitmap *bitmap)
 	bitmap->storage.sb_page = alloc_page(GFP_KERNEL | __GFP_ZERO);
 	if (bitmap->storage.sb_page == NULL)
 		return -ENOMEM;
-	bitmap->storage.sb_page->index = 0;
+	bitmap->storage.sb_index = 0;
 
 	sb = kmap_atomic(bitmap->storage.sb_page);
 
@@ -601,7 +620,7 @@ static int md_bitmap_read_sb(struct bitmap *bitmap)
 	unsigned long sectors_reserved = 0;
 	int err = -EINVAL;
 	struct page *sb_page;
-	loff_t offset = bitmap->mddev->bitmap_info.offset;
+	loff_t offset = 0;
 
 	if (!bitmap->storage.file && !bitmap->mddev->bitmap_info.offset) {
 		chunksize = 128 * 1024 * 1024;
@@ -628,7 +647,7 @@ static int md_bitmap_read_sb(struct bitmap *bitmap)
 		bm_blocks = ((bm_blocks+7) >> 3) + sizeof(bitmap_super_t);
 		/* to 4k blocks */
 		bm_blocks = DIV_ROUND_UP_SECTOR_T(bm_blocks, 4096);
-		offset = bitmap->mddev->bitmap_info.offset + (bitmap->cluster_slot * (bm_blocks << 3));
+		offset = bitmap->cluster_slot * (bm_blocks << 3);
 		pr_debug("%s:%d bm slot: %d offset: %llu\n", __func__, __LINE__,
 			bitmap->cluster_slot, offset);
 	}
@@ -637,13 +656,11 @@ static int md_bitmap_read_sb(struct bitmap *bitmap)
 		loff_t isize = i_size_read(bitmap->storage.file->f_mapping->host);
 		int bytes = isize > PAGE_SIZE ? PAGE_SIZE : isize;
 
-		err = read_page(bitmap->storage.file, 0,
+		err = read_file_page(bitmap->storage.file, 0,
 				bitmap, bytes, sb_page);
 	} else {
-		err = read_sb_page(bitmap->mddev,
-				   offset,
-				   sb_page,
-				   0, sizeof(bitmap_super_t));
+		err = read_sb_page(bitmap->mddev, offset, sb_page, 0,
+				   sizeof(bitmap_super_t));
 	}
 	if (err)
 		return err;
@@ -819,7 +836,7 @@ static int md_bitmap_storage_alloc(struct bitmap_storage *store,
 	if (store->sb_page) {
 		store->filemap[0] = store->sb_page;
 		pnum = 1;
-		store->sb_page->index = offset;
+		store->sb_index = offset;
 	}
 
 	for ( ; pnum < num_pages; pnum++) {
@@ -828,7 +845,6 @@ static int md_bitmap_storage_alloc(struct bitmap_storage *store,
 			store->file_pages = pnum;
 			return -ENOMEM;
 		}
-		store->filemap[pnum]->index = pnum + offset;
 	}
 	store->file_pages = pnum;
 
@@ -847,14 +863,10 @@ static int md_bitmap_storage_alloc(struct bitmap_storage *store,
 
 static void md_bitmap_file_unmap(struct bitmap_storage *store)
 {
-	struct page **map, *sb_page;
-	int pages;
-	struct file *file;
-
-	file = store->file;
-	map = store->filemap;
-	pages = store->file_pages;
-	sb_page = store->sb_page;
+	struct file *file = store->file;
+	struct page *sb_page = store->sb_page;
+	struct page **map = store->filemap;
+	int pages = store->file_pages;
 
 	while (pages--)
 		if (map[pages] != sb_page) /* 0 is sb_page, release it below */
@@ -879,21 +891,13 @@ static void md_bitmap_file_unmap(struct bitmap_storage *store)
  */
 static void md_bitmap_file_kick(struct bitmap *bitmap)
 {
-	char *path, *ptr = NULL;
-
 	if (!test_and_set_bit(BITMAP_STALE, &bitmap->flags)) {
 		md_bitmap_update_sb(bitmap);
 
 		if (bitmap->storage.file) {
-			path = kmalloc(PAGE_SIZE, GFP_KERNEL);
-			if (path)
-				ptr = file_path(bitmap->storage.file,
-					     path, PAGE_SIZE);
+			pr_warn("%s: kicking failed bitmap file %pD4 from array!\n",
+				bmname(bitmap), bitmap->storage.file);
 
-			pr_warn("%s: kicking failed bitmap file %s from array!\n",
-				bmname(bitmap), IS_ERR(ptr) ? "" : ptr);
-
-			kfree(path);
 		} else
 			pr_warn("%s: disabling internal bitmap due to errors\n",
 				bmname(bitmap));
@@ -945,6 +949,7 @@ static void md_bitmap_file_set_bit(struct bitmap *bitmap, sector_t block)
 	void *kaddr;
 	unsigned long chunk = block >> bitmap->counts.chunkshift;
 	struct bitmap_storage *store = &bitmap->storage;
+	unsigned long index = file_page_index(store, chunk);
 	unsigned long node_offset = 0;
 
 	if (mddev_is_clustered(bitmap->mddev))
@@ -962,9 +967,9 @@ static void md_bitmap_file_set_bit(struct bitmap *bitmap, sector_t block)
 	else
 		set_bit_le(bit, kaddr);
 	kunmap_atomic(kaddr);
-	pr_debug("set file bit %lu page %lu\n", bit, page->index);
+	pr_debug("set file bit %lu page %lu\n", bit, index);
 	/* record page number so it gets flushed to disk when unplug occurs */
-	set_page_attr(bitmap, page->index - node_offset, BITMAP_PAGE_DIRTY);
+	set_page_attr(bitmap, index - node_offset, BITMAP_PAGE_DIRTY);
 }
 
 static void md_bitmap_file_clear_bit(struct bitmap *bitmap, sector_t block)
@@ -974,6 +979,7 @@ static void md_bitmap_file_clear_bit(struct bitmap *bitmap, sector_t block)
 	void *paddr;
 	unsigned long chunk = block >> bitmap->counts.chunkshift;
 	struct bitmap_storage *store = &bitmap->storage;
+	unsigned long index = file_page_index(store, chunk);
 	unsigned long node_offset = 0;
 
 	if (mddev_is_clustered(bitmap->mddev))
@@ -989,8 +995,8 @@ static void md_bitmap_file_clear_bit(struct bitmap *bitmap, sector_t block)
 	else
 		clear_bit_le(bit, paddr);
 	kunmap_atomic(paddr);
-	if (!test_page_attr(bitmap, page->index - node_offset, BITMAP_PAGE_NEEDWRITE)) {
-		set_page_attr(bitmap, page->index - node_offset, BITMAP_PAGE_PENDING);
+	if (!test_page_attr(bitmap, index - node_offset, BITMAP_PAGE_NEEDWRITE)) {
+		set_page_attr(bitmap, index - node_offset, BITMAP_PAGE_PENDING);
 		bitmap->allclean = 0;
 	}
 }
@@ -1042,7 +1048,7 @@ void md_bitmap_unplug(struct bitmap *bitmap)
 							  "md bitmap_unplug");
 			}
 			clear_page_attr(bitmap, i, BITMAP_PAGE_PENDING);
-			write_page(bitmap, bitmap->storage.filemap[i], 0);
+			filemap_write_page(bitmap, i, false);
 			writing = 1;
 		}
 	}
@@ -1084,33 +1090,31 @@ void md_bitmap_unplug_async(struct bitmap *bitmap)
 EXPORT_SYMBOL(md_bitmap_unplug_async);
 
 static void md_bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int needed);
-/* * bitmap_init_from_disk -- called at bitmap_create time to initialize
- * the in-memory bitmap from the on-disk bitmap -- also, sets up the
- * memory mapping of the bitmap file
- * Special cases:
- *   if there's no bitmap file, or if the bitmap file had been
- *   previously kicked from the array, we mark all the bits as
- *   1's in order to cause a full resync.
+
+/*
+ * Initialize the in-memory bitmap from the on-disk bitmap and set up the memory
+ * mapping of the bitmap file.
+ *
+ * Special case: If there's no bitmap file, or if the bitmap file had been
+ * previously kicked from the array, we mark all the bits as 1's in order to
+ * cause a full resync.
  *
  * We ignore all bits for sectors that end earlier than 'start'.
- * This is used when reading an out-of-date bitmap...
+ * This is used when reading an out-of-date bitmap.
  */
 static int md_bitmap_init_from_disk(struct bitmap *bitmap, sector_t start)
 {
-	unsigned long i, chunks, index, oldindex, bit, node_offset = 0;
-	struct page *page = NULL;
-	unsigned long bit_cnt = 0;
-	struct file *file;
-	unsigned long offset;
-	int outofdate;
-	int ret = -ENOSPC;
-	void *paddr;
+	bool outofdate = test_bit(BITMAP_STALE, &bitmap->flags);
+	struct mddev *mddev = bitmap->mddev;
+	unsigned long chunks = bitmap->counts.chunks;
 	struct bitmap_storage *store = &bitmap->storage;
+	struct file *file = store->file;
+	unsigned long node_offset = 0;
+	unsigned long bit_cnt = 0;
+	unsigned long i;
+	int ret;
 
-	chunks = bitmap->counts.chunks;
-	file = store->file;
-
-	if (!file && !bitmap->mddev->bitmap_info.offset) {
+	if (!file && !mddev->bitmap_info.offset) {
 		/* No permanent bitmap - fill with '1s'. */
 		store->filemap = NULL;
 		store->file_pages = 0;
@@ -1125,77 +1129,79 @@ static int md_bitmap_init_from_disk(struct bitmap *bitmap, sector_t start)
 		return 0;
 	}
 
-	outofdate = test_bit(BITMAP_STALE, &bitmap->flags);
-	if (outofdate)
-		pr_warn("%s: bitmap file is out of date, doing full recovery\n", bmname(bitmap));
-
 	if (file && i_size_read(file->f_mapping->host) < store->bytes) {
 		pr_warn("%s: bitmap file too short %lu < %lu\n",
 			bmname(bitmap),
 			(unsigned long) i_size_read(file->f_mapping->host),
 			store->bytes);
+		ret = -ENOSPC;
 		goto err;
 	}
 
-	oldindex = ~0L;
-	offset = 0;
-	if (!bitmap->mddev->bitmap_info.external)
-		offset = sizeof(bitmap_super_t);
-
-	if (mddev_is_clustered(bitmap->mddev))
+	if (mddev_is_clustered(mddev))
 		node_offset = bitmap->cluster_slot * (DIV_ROUND_UP(store->bytes, PAGE_SIZE));
 
-	for (i = 0; i < chunks; i++) {
-		int b;
-		index = file_page_index(&bitmap->storage, i);
-		bit = file_page_offset(&bitmap->storage, i);
-		if (index != oldindex) { /* this is a new page, read it in */
-			int count;
-			/* unmap the old page, we're done with it */
-			if (index == store->file_pages-1)
-				count = store->bytes - index * PAGE_SIZE;
-			else
-				count = PAGE_SIZE;
-			page = store->filemap[index];
-			if (file)
-				ret = read_page(file, index, bitmap,
-						count, page);
-			else
-				ret = read_sb_page(
-					bitmap->mddev,
-					bitmap->mddev->bitmap_info.offset,
-					page,
-					index + node_offset, count);
+	for (i = 0; i < store->file_pages; i++) {
+		struct page *page = store->filemap[i];
+		int count;
 
-			if (ret)
-				goto err;
+		/* unmap the old page, we're done with it */
+		if (i == store->file_pages - 1)
+			count = store->bytes - i * PAGE_SIZE;
+		else
+			count = PAGE_SIZE;
 
-			oldindex = index;
+		if (file)
+			ret = read_file_page(file, i, bitmap, count, page);
+		else
+			ret = read_sb_page(mddev, 0, page, i + node_offset,
+					   count);
+		if (ret)
+			goto err;
+	}
 
-			if (outofdate) {
-				/*
-				 * if bitmap is out of date, dirty the
-				 * whole page and write it out
-				 */
-				paddr = kmap_atomic(page);
-				memset(paddr + offset, 0xff,
-				       PAGE_SIZE - offset);
-				kunmap_atomic(paddr);
-				write_page(bitmap, page, 1);
+	if (outofdate) {
+		pr_warn("%s: bitmap file is out of date, doing full recovery\n",
+			bmname(bitmap));
 
+		for (i = 0; i < store->file_pages; i++) {
+			struct page *page = store->filemap[i];
+			unsigned long offset = 0;
+			void *paddr;
+
+			if (i == 0 && !mddev->bitmap_info.external)
+				offset = sizeof(bitmap_super_t);
+
+			/*
+			 * If the bitmap is out of date, dirty the whole page
+			 * and write it out
+			 */
+			paddr = kmap_atomic(page);
+			memset(paddr + offset, 0xff, PAGE_SIZE - offset);
+			kunmap_atomic(paddr);
+
+			filemap_write_page(bitmap, i, true);
+			if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags)) {
 				ret = -EIO;
-				if (test_bit(BITMAP_WRITE_ERROR,
-					     &bitmap->flags))
-					goto err;
+				goto err;
 			}
 		}
+	}
+
+	for (i = 0; i < chunks; i++) {
+		struct page *page = filemap_get_page(&bitmap->storage, i);
+		unsigned long bit = file_page_offset(&bitmap->storage, i);
+		void *paddr;
+		bool was_set;
+
 		paddr = kmap_atomic(page);
 		if (test_bit(BITMAP_HOSTENDIAN, &bitmap->flags))
-			b = test_bit(bit, paddr);
+			was_set = test_bit(bit, paddr);
 		else
-			b = test_bit_le(bit, paddr);
+			was_set = test_bit_le(bit, paddr);
 		kunmap_atomic(paddr);
-		if (b) {
+
+		if (was_set) {
 			/* if the disk bit is set, set the memory bit */
 			int needed = ((sector_t)(i+1) << bitmap->counts.chunkshift
 				      >= start);
@@ -1204,7 +1210,6 @@ static int md_bitmap_init_from_disk(struct bitmap *bitmap, sector_t start)
 						  needed);
 			bit_cnt++;
 		}
-		offset = 0;
 	}
 
 	pr_debug("%s: bitmap initialized from disk: read %lu pages, set %lu of %lu bits\n",
@@ -1396,9 +1401,8 @@ void md_bitmap_daemon_work(struct mddev *mddev)
 			break;
 		if (bitmap->storage.filemap &&
 		    test_and_clear_page_attr(bitmap, j,
-					     BITMAP_PAGE_NEEDWRITE)) {
-			write_page(bitmap, bitmap->storage.filemap[j], 0);
-		}
+					     BITMAP_PAGE_NEEDWRITE))
+			filemap_write_page(bitmap, j, false);
 	}
 
  done:
@@ -2542,6 +2546,10 @@ backlog_store(struct mddev *mddev, const char *buf, size_t len)
 	if (backlog > COUNTER_MAX)
 		return -EINVAL;
 
+	rv = mddev_lock(mddev);
+	if (rv)
+		return rv;
+
 	/*
 	 * Without write mostly device, it doesn't make sense to set
 	 * backlog for max_write_behind.
@@ -2555,6 +2563,7 @@ backlog_store(struct mddev *mddev, const char *buf, size_t len)
 	if (!has_write_mostly) {
 		pr_warn_ratelimited("%s: can't set backlog, no write mostly device available\n",
 				    mdname(mddev));
+		mddev_unlock(mddev);
 		return -EINVAL;
 	}
 
@@ -2565,13 +2574,13 @@ backlog_store(struct mddev *mddev, const char *buf, size_t len)
 			mddev_destroy_serial_pool(mddev, NULL, false);
 	} else if (backlog && !mddev->serial_info_pool) {
 		/* serial_info_pool is needed since backlog is not zero */
-		struct md_rdev *rdev;
-
 		rdev_for_each(rdev, mddev)
 			mddev_create_serial_pool(mddev, rdev, false);
 	}
 	if (old_mwb != backlog)
 		md_bitmap_update_sb(mddev->bitmap);
+
+	mddev_unlock(mddev);
 	return len;
 }
 

drivers/md/md-bitmap.h

@@ -201,6 +201,7 @@ struct bitmap {
 		struct file *file;		/* backing disk file */
 		struct page *sb_page;		/* cached copy of the bitmap
 						 * file superblock */
+		unsigned long sb_index;
 		struct page **filemap;		/* list of cache pages for
 						 * the file */
 		unsigned long *filemap_attr;	/* attributes associated

drivers/md/md-faulty.c

@@ -204,6 +204,8 @@ static bool faulty_make_request(struct mddev *mddev, struct bio *bio)
 			failit = 1;
 		}
 	}
+
+	md_account_bio(mddev, &bio);
 	if (failit) {
 		struct bio *b = bio_alloc_clone(conf->rdev->bdev, bio, GFP_NOIO,
 						&mddev->bio_set);

drivers/md/md-linear.c

@@ -238,6 +238,7 @@ static bool linear_make_request(struct mddev *mddev, struct bio *bio)
 		bio = split;
 	}
 
+	md_account_bio(mddev, &bio);
 	bio_set_dev(bio, tmp_dev->rdev->bdev);
 	bio->bi_iter.bi_sector = bio->bi_iter.bi_sector -
 		start_sector + data_offset;

drivers/md/md-multipath.c

@@ -107,6 +107,7 @@ static bool multipath_make_request(struct mddev *mddev, struct bio * bio)
 	    && md_flush_request(mddev, bio))
 		return true;
 
+	md_account_bio(mddev, &bio);
 	mp_bh = mempool_alloc(&conf->pool, GFP_NOIO);
 
 	mp_bh->master_bio = bio;

drivers/md/md.c

@@ -453,7 +453,6 @@ void mddev_suspend(struct mddev *mddev)
 		mddev->pers->prepare_suspend(mddev);
 
 	wait_event(mddev->sb_wait, percpu_ref_is_zero(&mddev->active_io));
-	mddev->pers->quiesce(mddev, 1);
 	clear_bit_unlock(MD_ALLOW_SB_UPDATE, &mddev->flags);
 	wait_event(mddev->sb_wait, !test_bit(MD_UPDATING_SB, &mddev->flags));
 
@@ -465,14 +464,15 @@ EXPORT_SYMBOL_GPL(mddev_suspend);
 
 void mddev_resume(struct mddev *mddev)
 {
-	/* entred the memalloc scope from mddev_suspend() */
-	memalloc_noio_restore(mddev->noio_flag);
 	lockdep_assert_held(&mddev->reconfig_mutex);
 	if (--mddev->suspended)
 		return;
+
+	/* entred the memalloc scope from mddev_suspend() */
+	memalloc_noio_restore(mddev->noio_flag);
+
 	percpu_ref_resurrect(&mddev->active_io);
 	wake_up(&mddev->sb_wait);
-	mddev->pers->quiesce(mddev, 0);
 
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 	md_wakeup_thread(mddev->thread);
@@ -643,6 +643,7 @@ void mddev_init(struct mddev *mddev)
 {
 	mutex_init(&mddev->open_mutex);
 	mutex_init(&mddev->reconfig_mutex);
+	mutex_init(&mddev->sync_mutex);
 	mutex_init(&mddev->bitmap_info.mutex);
 	INIT_LIST_HEAD(&mddev->disks);
 	INIT_LIST_HEAD(&mddev->all_mddevs);
@@ -650,6 +651,7 @@ void mddev_init(struct mddev *mddev)
 	timer_setup(&mddev->safemode_timer, md_safemode_timeout, 0);
 	atomic_set(&mddev->active, 1);
 	atomic_set(&mddev->openers, 0);
+	atomic_set(&mddev->sync_seq, 0);
 	spin_lock_init(&mddev->lock);
 	atomic_set(&mddev->flush_pending, 0);
 	init_waitqueue_head(&mddev->sb_wait);
@@ -2304,7 +2306,7 @@ int md_integrity_register(struct mddev *mddev)
 	pr_debug("md: data integrity enabled on %s\n", mdname(mddev));
 	if (bioset_integrity_create(&mddev->bio_set, BIO_POOL_SIZE) ||
 	    (mddev->level != 1 && mddev->level != 10 &&
-	     bioset_integrity_create(&mddev->io_acct_set, BIO_POOL_SIZE))) {
+	     bioset_integrity_create(&mddev->io_clone_set, BIO_POOL_SIZE))) {
 		/*
 		 * No need to handle the failure of bioset_integrity_create,
 		 * because the function is called by md_run() -> pers->run(),
@@ -4747,6 +4749,62 @@ action_show(struct mddev *mddev, char *page)
 	return sprintf(page, "%s\n", type);
 }
 
+static void stop_sync_thread(struct mddev *mddev)
+{
+	if (!test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
+		return;
+
+	if (mddev_lock(mddev))
+		return;
+
+	/*
+	 * Check again in case MD_RECOVERY_RUNNING is cleared before lock is
+	 * held.
+	 */
+	if (!test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) {
+		mddev_unlock(mddev);
+		return;
+	}
+
+	if (work_pending(&mddev->del_work))
+		flush_workqueue(md_misc_wq);
+
+	set_bit(MD_RECOVERY_INTR, &mddev->recovery);
+	/*
+	 * Thread might be blocked waiting for metadata update which will now
+	 * never happen
+	 */
+	md_wakeup_thread_directly(mddev->sync_thread);
+
+	mddev_unlock(mddev);
+}
+
+static void idle_sync_thread(struct mddev *mddev)
+{
+	int sync_seq = atomic_read(&mddev->sync_seq);
+
+	mutex_lock(&mddev->sync_mutex);
+	clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
+	stop_sync_thread(mddev);
+
+	wait_event(resync_wait, sync_seq != atomic_read(&mddev->sync_seq) ||
+			!test_bit(MD_RECOVERY_RUNNING, &mddev->recovery));
+
+	mutex_unlock(&mddev->sync_mutex);
+}
+
+static void frozen_sync_thread(struct mddev *mddev)
+{
+	mutex_lock(&mddev->sync_mutex);
+	set_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
+	stop_sync_thread(mddev);
+
+	wait_event(resync_wait, mddev->sync_thread == NULL &&
+			!test_bit(MD_RECOVERY_RUNNING, &mddev->recovery));
+
+	mutex_unlock(&mddev->sync_mutex);
+}
+
 static ssize_t
 action_store(struct mddev *mddev, const char *page, size_t len)
 {
@@ -4754,35 +4812,11 @@ action_store(struct mddev *mddev, const char *page, size_t len)
 		return -EINVAL;
 
 
-	if (cmd_match(page, "idle") || cmd_match(page, "frozen")) {
-		if (cmd_match(page, "frozen"))
-			set_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
-		else
-			clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
-		if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) &&
-		    mddev_lock(mddev) == 0) {
-			if (work_pending(&mddev->del_work))
-				flush_workqueue(md_misc_wq);
-			if (mddev->sync_thread) {
-				sector_t save_rp = mddev->reshape_position;
-
-				mddev_unlock(mddev);
-				set_bit(MD_RECOVERY_INTR, &mddev->recovery);
-				md_unregister_thread(&mddev->sync_thread);
-				mddev_lock_nointr(mddev);
-				/*
-				 * set RECOVERY_INTR again and restore reshape
-				 * position in case others changed them after
-				 * got lock, eg, reshape_position_store and
-				 * md_check_recovery.
-				 */
-				mddev->reshape_position = save_rp;
-				set_bit(MD_RECOVERY_INTR, &mddev->recovery);
-				md_reap_sync_thread(mddev);
-			}
-			mddev_unlock(mddev);
-		}
-	} else if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
+	if (cmd_match(page, "idle"))
+		idle_sync_thread(mddev);
+	else if (cmd_match(page, "frozen"))
+		frozen_sync_thread(mddev);
+	else if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
 		return -EBUSY;
 	else if (cmd_match(page, "resync"))
 		clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
@@ -5842,6 +5876,13 @@ int md_run(struct mddev *mddev)
 			goto exit_bio_set;
 	}
 
+	if (!bioset_initialized(&mddev->io_clone_set)) {
+		err = bioset_init(&mddev->io_clone_set, BIO_POOL_SIZE,
+				  offsetof(struct md_io_clone, bio_clone), 0);
+		if (err)
+			goto exit_sync_set;
+	}
+
 	spin_lock(&pers_lock);
 	pers = find_pers(mddev->level, mddev->clevel);
 	if (!pers || !try_module_get(pers->owner)) {
@@ -6019,6 +6060,8 @@ int md_run(struct mddev *mddev)
 	module_put(pers->owner);
 	md_bitmap_destroy(mddev);
 abort:
+	bioset_exit(&mddev->io_clone_set);
+exit_sync_set:
 	bioset_exit(&mddev->sync_set);
 exit_bio_set:
 	bioset_exit(&mddev->bio_set);
@@ -6176,7 +6219,6 @@ static void __md_stop_writes(struct mddev *mddev)
 		flush_workqueue(md_misc_wq);
 	if (mddev->sync_thread) {
 		set_bit(MD_RECOVERY_INTR, &mddev->recovery);
-		md_unregister_thread(&mddev->sync_thread);
 		md_reap_sync_thread(mddev);
 	}
 
@@ -6243,6 +6285,7 @@ static void __md_stop(struct mddev *mddev)
 	percpu_ref_exit(&mddev->active_io);
 	bioset_exit(&mddev->bio_set);
 	bioset_exit(&mddev->sync_set);
+	bioset_exit(&mddev->io_clone_set);
 }
 
 void md_stop(struct mddev *mddev)
@@ -7010,6 +7053,15 @@ static int set_bitmap_file(struct mddev *mddev, int fd)
 
 		if (mddev->bitmap || mddev->bitmap_info.file)
 			return -EEXIST; /* cannot add when bitmap is present */
+
+		if (!IS_ENABLED(CONFIG_MD_BITMAP_FILE)) {
+			pr_warn("%s: bitmap files not supported by this kernel\n",
+				mdname(mddev));
+			return -EINVAL;
+		}
+		pr_warn("%s: using deprecated bitmap file support\n",
+			mdname(mddev));
+
 		f = fget(fd);
 
 		if (f == NULL) {
@@ -8599,62 +8651,44 @@ void md_submit_discard_bio(struct mddev *mddev, struct md_rdev *rdev,
 }
 EXPORT_SYMBOL_GPL(md_submit_discard_bio);
 
-int acct_bioset_init(struct mddev *mddev)
+static void md_end_clone_io(struct bio *bio)
 {
-	int err = 0;
-
-	if (!bioset_initialized(&mddev->io_acct_set))
-		err = bioset_init(&mddev->io_acct_set, BIO_POOL_SIZE,
-			offsetof(struct md_io_acct, bio_clone), 0);
-	return err;
-}
-EXPORT_SYMBOL_GPL(acct_bioset_init);
-
-void acct_bioset_exit(struct mddev *mddev)
-{
-	bioset_exit(&mddev->io_acct_set);
-}
-EXPORT_SYMBOL_GPL(acct_bioset_exit);
-
-static void md_end_io_acct(struct bio *bio)
-{
-	struct md_io_acct *md_io_acct = bio->bi_private;
-	struct bio *orig_bio = md_io_acct->orig_bio;
-	struct mddev *mddev = md_io_acct->mddev;
+	struct md_io_clone *md_io_clone = bio->bi_private;
+	struct bio *orig_bio = md_io_clone->orig_bio;
+	struct mddev *mddev = md_io_clone->mddev;
 
 	orig_bio->bi_status = bio->bi_status;
 
-	bio_end_io_acct(orig_bio, md_io_acct->start_time);
+	if (md_io_clone->start_time)
+		bio_end_io_acct(orig_bio, md_io_clone->start_time);
+
 	bio_put(bio);
 	bio_endio(orig_bio);
-
 	percpu_ref_put(&mddev->active_io);
 }
 
-/*
- * Used by personalities that don't already clone the bio and thus can't
- * easily add the timestamp to their extended bio structure.
- */
-void md_account_bio(struct mddev *mddev, struct bio **bio)
+static void md_clone_bio(struct mddev *mddev, struct bio **bio)
 {
 	struct block_device *bdev = (*bio)->bi_bdev;
-	struct md_io_acct *md_io_acct;
-	struct bio *clone;
-
-	if (!blk_queue_io_stat(bdev->bd_disk->queue))
-		return;
+	struct md_io_clone *md_io_clone;
+	struct bio *clone =
+		bio_alloc_clone(bdev, *bio, GFP_NOIO, &mddev->io_clone_set);
+
+	md_io_clone = container_of(clone, struct md_io_clone, bio_clone);
+	md_io_clone->orig_bio = *bio;
+	md_io_clone->mddev = mddev;
+	if (blk_queue_io_stat(bdev->bd_disk->queue))
+		md_io_clone->start_time = bio_start_io_acct(*bio);
+
+	clone->bi_end_io = md_end_clone_io;
+	clone->bi_private = md_io_clone;
+	*bio = clone;
+}
 
+void md_account_bio(struct mddev *mddev, struct bio **bio)
+{
 	percpu_ref_get(&mddev->active_io);
-
-	clone = bio_alloc_clone(bdev, *bio, GFP_NOIO, &mddev->io_acct_set);
-	md_io_acct = container_of(clone, struct md_io_acct, bio_clone);
-	md_io_acct->orig_bio = *bio;
-	md_io_acct->start_time = bio_start_io_acct(*bio);
-	md_io_acct->mddev = mddev;
-
-	clone->bi_end_io = md_end_io_acct;
-	clone->bi_private = md_io_acct;
-	*bio = clone;
+	md_clone_bio(mddev, bio);
 }
 EXPORT_SYMBOL_GPL(md_account_bio);
 
@@ -9327,7 +9361,6 @@ void md_check_recovery(struct mddev *mddev)
 			 * ->spare_active and clear saved_raid_disk
 			 */
 			set_bit(MD_RECOVERY_INTR, &mddev->recovery);
-			md_unregister_thread(&mddev->sync_thread);
 			md_reap_sync_thread(mddev);
 			clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
 			clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
@@ -9356,17 +9389,24 @@ void md_check_recovery(struct mddev *mddev)
 		if (mddev->sb_flags)
 			md_update_sb(mddev, 0);
 
-		if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) &&
-		    !test_bit(MD_RECOVERY_DONE, &mddev->recovery)) {
-			/* resync/recovery still happening */
-			clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
-			goto unlock;
-		}
-		if (mddev->sync_thread) {
-			md_unregister_thread(&mddev->sync_thread);
+		/*
+		 * Never start a new sync thread if MD_RECOVERY_RUNNING is
+		 * still set.
+		 */
+		if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) {
+			if (!test_bit(MD_RECOVERY_DONE, &mddev->recovery)) {
+				/* resync/recovery still happening */
+				clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
+				goto unlock;
+			}
+
+			if (WARN_ON_ONCE(!mddev->sync_thread))
+				goto unlock;
+
 			md_reap_sync_thread(mddev);
 			goto unlock;
 		}
+
 		/* Set RUNNING before clearing NEEDED to avoid
 		 * any transients in the value of "sync_action".
 		 */
@@ -9443,7 +9483,10 @@ void md_reap_sync_thread(struct mddev *mddev)
 	sector_t old_dev_sectors = mddev->dev_sectors;
 	bool is_reshaped = false;
 
-	/* sync_thread should be unregistered, collect result */
+	/* resync has finished, collect result */
+	md_unregister_thread(&mddev->sync_thread);
+	atomic_inc(&mddev->sync_seq);
+
 	if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery) &&
 	    !test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery) &&
 	    mddev->degraded != mddev->raid_disks) {
@@ -9488,7 +9531,6 @@ void md_reap_sync_thread(struct mddev *mddev)
 	if (mddev_is_clustered(mddev) && is_reshaped
 				      && !test_bit(MD_CLOSING, &mddev->flags))
 		md_cluster_ops->update_size(mddev, old_dev_sectors);
-	wake_up(&resync_wait);
 	/* flag recovery needed just to double check */
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 	sysfs_notify_dirent_safe(mddev->sysfs_completed);
@@ -9496,6 +9538,7 @@ void md_reap_sync_thread(struct mddev *mddev)
 	md_new_event();
 	if (mddev->event_work.func)
 		queue_work(md_misc_wq, &mddev->event_work);
+	wake_up(&resync_wait);
 }
 EXPORT_SYMBOL(md_reap_sync_thread);
 

drivers/md/md.h

@@ -510,7 +510,7 @@ struct mddev {
 	struct bio_set			sync_set; /* for sync operations like
 						   * metadata and bitmap writes
 						   */
-	struct bio_set			io_acct_set; /* for raid0 and raid5 io accounting */
+	struct bio_set			io_clone_set;
 
 	/* Generic flush handling.
 	 * The last to finish preflush schedules a worker to submit
@@ -535,6 +535,11 @@ struct mddev {
 	 */
 	struct list_head		deleting;
 
+	/* Used to synchronize idle and frozen for action_store() */
+	struct mutex			sync_mutex;
+	/* The sequence number for sync thread */
+	atomic_t sync_seq;
+
 	bool	has_superblocks:1;
 	bool	fail_last_dev:1;
 	bool	serialize_policy:1;
@@ -731,7 +736,7 @@ struct md_thread {
 	void			*private;
 };
 
-struct md_io_acct {
+struct md_io_clone {
 	struct mddev	*mddev;
 	struct bio	*orig_bio;
 	unsigned long	start_time;
@@ -769,8 +774,6 @@ extern void md_error(struct mddev *mddev, struct md_rdev *rdev);
 extern void md_finish_reshape(struct mddev *mddev);
 void md_submit_discard_bio(struct mddev *mddev, struct md_rdev *rdev,
 			struct bio *bio, sector_t start, sector_t size);
-int acct_bioset_init(struct mddev *mddev);
-void acct_bioset_exit(struct mddev *mddev);
 void md_account_bio(struct mddev *mddev, struct bio **bio);
 
 extern bool __must_check md_flush_request(struct mddev *mddev, struct bio *bio);

drivers/md/raid0.c

@@ -377,7 +377,6 @@ static void raid0_free(struct mddev *mddev, void *priv)
 	struct r0conf *conf = priv;
 
 	free_conf(mddev, conf);
-	acct_bioset_exit(mddev);
 }
 
 static int raid0_run(struct mddev *mddev)
@@ -392,16 +391,11 @@ static int raid0_run(struct mddev *mddev)
 	if (md_check_no_bitmap(mddev))
 		return -EINVAL;
 
-	if (acct_bioset_init(mddev)) {
-		pr_err("md/raid0:%s: alloc acct bioset failed.\n", mdname(mddev));
-		return -ENOMEM;
-	}
-
 	/* if private is not null, we are here after takeover */
 	if (mddev->private == NULL) {
 		ret = create_strip_zones(mddev, &conf);
 		if (ret < 0)
-			goto exit_acct_set;
+			return ret;
 		mddev->private = conf;
 	}
 	conf = mddev->private;
@@ -432,15 +426,9 @@ static int raid0_run(struct mddev *mddev)
 
 	ret = md_integrity_register(mddev);
 	if (ret)
-		goto free;
+		free_conf(mddev, conf);
 
 	return ret;
-
-free:
-	free_conf(mddev, conf);
-exit_acct_set:
-	acct_bioset_exit(mddev);
-	return ret;
 }
 
 /*

drivers/md/raid1.c

@@ -304,8 +304,6 @@ static void call_bio_endio(struct r1bio *r1_bio)
 	if (!test_bit(R1BIO_Uptodate, &r1_bio->state))
 		bio->bi_status = BLK_STS_IOERR;
 
-	if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
-		bio_end_io_acct(bio, r1_bio->start_time);
 	bio_endio(bio);
 }
 
@@ -791,11 +789,17 @@ static int read_balance(struct r1conf *conf, struct r1bio *r1_bio, int *max_sect
 	return best_disk;
 }
 
+static void wake_up_barrier(struct r1conf *conf)
+{
+	if (wq_has_sleeper(&conf->wait_barrier))
+		wake_up(&conf->wait_barrier);
+}
+
 static void flush_bio_list(struct r1conf *conf, struct bio *bio)
 {
 	/* flush any pending bitmap writes to disk before proceeding w/ I/O */
 	raid1_prepare_flush_writes(conf->mddev->bitmap);
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 
 	while (bio) { /* submit pending writes */
 		struct bio *next = bio->bi_next;
@@ -972,7 +976,7 @@ static bool _wait_barrier(struct r1conf *conf, int idx, bool nowait)
 	 * In case freeze_array() is waiting for
 	 * get_unqueued_pending() == extra
 	 */
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 	/* Wait for the barrier in same barrier unit bucket to drop. */
 
 	/* Return false when nowait flag is set */
@@ -1015,7 +1019,7 @@ static bool wait_read_barrier(struct r1conf *conf, sector_t sector_nr, bool nowa
 	 * In case freeze_array() is waiting for
 	 * get_unqueued_pending() == extra
 	 */
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 	/* Wait for array to be unfrozen */
 
 	/* Return false when nowait flag is set */
@@ -1044,7 +1048,7 @@ static bool wait_barrier(struct r1conf *conf, sector_t sector_nr, bool nowait)
 static void _allow_barrier(struct r1conf *conf, int idx)
 {
 	atomic_dec(&conf->nr_pending[idx]);
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 }
 
 static void allow_barrier(struct r1conf *conf, sector_t sector_nr)
@@ -1173,7 +1177,7 @@ static void raid1_unplug(struct blk_plug_cb *cb, bool from_schedule)
 		spin_lock_irq(&conf->device_lock);
 		bio_list_merge(&conf->pending_bio_list, &plug->pending);
 		spin_unlock_irq(&conf->device_lock);
-		wake_up(&conf->wait_barrier);
+		wake_up_barrier(conf);
 		md_wakeup_thread(mddev->thread);
 		kfree(plug);
 		return;
@@ -1303,10 +1307,10 @@ static void raid1_read_request(struct mddev *mddev, struct bio *bio,
 	}
 
 	r1_bio->read_disk = rdisk;
-
-	if (!r1bio_existed && blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
-		r1_bio->start_time = bio_start_io_acct(bio);
-
+	if (!r1bio_existed) {
+		md_account_bio(mddev, &bio);
+		r1_bio->master_bio = bio;
+	}
 	read_bio = bio_alloc_clone(mirror->rdev->bdev, bio, gfp,
 				   &mddev->bio_set);
 
@@ -1500,8 +1504,8 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio,
 		r1_bio->sectors = max_sectors;
 	}
 
-	if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
-		r1_bio->start_time = bio_start_io_acct(bio);
+	md_account_bio(mddev, &bio);
+	r1_bio->master_bio = bio;
 	atomic_set(&r1_bio->remaining, 1);
 	atomic_set(&r1_bio->behind_remaining, 0);
 
@@ -1576,7 +1580,7 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio,
 	r1_bio_write_done(r1_bio);
 
 	/* In case raid1d snuck in to freeze_array */
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 }
 
 static bool raid1_make_request(struct mddev *mddev, struct bio *bio)
@@ -1766,7 +1770,7 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 {
 	struct r1conf *conf = mddev->private;
 	int err = -EEXIST;
-	int mirror = 0;
+	int mirror = 0, repl_slot = -1;
 	struct raid1_info *p;
 	int first = 0;
 	int last = conf->raid_disks - 1;
@@ -1809,17 +1813,21 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 			break;
 		}
 		if (test_bit(WantReplacement, &p->rdev->flags) &&
-		    p[conf->raid_disks].rdev == NULL) {
-			/* Add this device as a replacement */
-			clear_bit(In_sync, &rdev->flags);
-			set_bit(Replacement, &rdev->flags);
-			rdev->raid_disk = mirror;
-			err = 0;
-			conf->fullsync = 1;
-			rcu_assign_pointer(p[conf->raid_disks].rdev, rdev);
-			break;
-		}
+		    p[conf->raid_disks].rdev == NULL && repl_slot < 0)
+			repl_slot = mirror;
 	}
+
+	if (err && repl_slot >= 0) {
+		/* Add this device as a replacement */
+		p = conf->mirrors + repl_slot;
+		clear_bit(In_sync, &rdev->flags);
+		set_bit(Replacement, &rdev->flags);
+		rdev->raid_disk = repl_slot;
+		err = 0;
+		conf->fullsync = 1;
+		rcu_assign_pointer(p[conf->raid_disks].rdev, rdev);
+	}
+
 	print_conf(conf);
 	return err;
 }
@@ -2299,7 +2307,7 @@ static void fix_read_error(struct r1conf *conf, int read_disk,
 			d++;
 			if (d == conf->raid_disks * 2)
 				d = 0;
-		} while (!success && d != read_disk);
+		} while (d != read_disk);
 
 		if (!success) {
 			/* Cannot read from anywhere - mark it bad */

drivers/md/raid1.h

@@ -157,7 +157,6 @@ struct r1bio {
 	sector_t		sector;
 	int			sectors;
 	unsigned long		state;
-	unsigned long		start_time;
 	struct mddev		*mddev;
 	/*
 	 * original bio going to /dev/mdx

drivers/md/raid10.c

@@ -325,8 +325,6 @@ static void raid_end_bio_io(struct r10bio *r10_bio)
 	if (!test_bit(R10BIO_Uptodate, &r10_bio->state))
 		bio->bi_status = BLK_STS_IOERR;
 
-	if (r10_bio->start_time)
-		bio_end_io_acct(bio, r10_bio->start_time);
 	bio_endio(bio);
 	/*
 	 * Wake up any possible resync thread that waits for the device
@@ -1172,7 +1170,7 @@ static bool regular_request_wait(struct mddev *mddev, struct r10conf *conf,
 }
 
 static void raid10_read_request(struct mddev *mddev, struct bio *bio,
-				struct r10bio *r10_bio)
+				struct r10bio *r10_bio, bool io_accounting)
 {
 	struct r10conf *conf = mddev->private;
 	struct bio *read_bio;
@@ -1243,9 +1241,10 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio,
 	}
 	slot = r10_bio->read_slot;
 
-	if (!r10_bio->start_time &&
-	    blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
-		r10_bio->start_time = bio_start_io_acct(bio);
+	if (io_accounting) {
+		md_account_bio(mddev, &bio);
+		r10_bio->master_bio = bio;
+	}
 	read_bio = bio_alloc_clone(rdev->bdev, bio, gfp, &mddev->bio_set);
 
 	r10_bio->devs[slot].bio = read_bio;
@@ -1322,6 +1321,25 @@ static void raid10_write_one_disk(struct mddev *mddev, struct r10bio *r10_bio,
 	}
 }
 
+static struct md_rdev *dereference_rdev_and_rrdev(struct raid10_info *mirror,
+						  struct md_rdev **prrdev)
+{
+	struct md_rdev *rdev, *rrdev;
+
+	rrdev = rcu_dereference(mirror->replacement);
+	/*
+	 * Read replacement first to prevent reading both rdev and
+	 * replacement as NULL during replacement replace rdev.
+	 */
+	smp_mb();
+	rdev = rcu_dereference(mirror->rdev);
+	if (rdev == rrdev)
+		rrdev = NULL;
+
+	*prrdev = rrdev;
+	return rdev;
+}
+
 static void wait_blocked_dev(struct mddev *mddev, struct r10bio *r10_bio)
 {
 	int i;
@@ -1332,11 +1350,9 @@ static void wait_blocked_dev(struct mddev *mddev, struct r10bio *r10_bio)
 	blocked_rdev = NULL;
 	rcu_read_lock();
 	for (i = 0; i < conf->copies; i++) {
-		struct md_rdev *rdev = rcu_dereference(conf->mirrors[i].rdev);
-		struct md_rdev *rrdev = rcu_dereference(
-			conf->mirrors[i].replacement);
-		if (rdev == rrdev)
-			rrdev = NULL;
+		struct md_rdev *rdev, *rrdev;
+
+		rdev = dereference_rdev_and_rrdev(&conf->mirrors[i], &rrdev);
 		if (rdev && unlikely(test_bit(Blocked, &rdev->flags))) {
 			atomic_inc(&rdev->nr_pending);
 			blocked_rdev = rdev;
@@ -1465,15 +1481,7 @@ static void raid10_write_request(struct mddev *mddev, struct bio *bio,
 		int d = r10_bio->devs[i].devnum;
 		struct md_rdev *rdev, *rrdev;
 
-		rrdev = rcu_dereference(conf->mirrors[d].replacement);
-		/*
-		 * Read replacement first to prevent reading both rdev and
-		 * replacement as NULL during replacement replace rdev.
-		 */
-		smp_mb();
-		rdev = rcu_dereference(conf->mirrors[d].rdev);
-		if (rdev == rrdev)
-			rrdev = NULL;
+		rdev = dereference_rdev_and_rrdev(&conf->mirrors[d], &rrdev);
 		if (rdev && (test_bit(Faulty, &rdev->flags)))
 			rdev = NULL;
 		if (rrdev && (test_bit(Faulty, &rrdev->flags)))
@@ -1543,8 +1551,8 @@ static void raid10_write_request(struct mddev *mddev, struct bio *bio,
 		r10_bio->master_bio = bio;
 	}
 
-	if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
-		r10_bio->start_time = bio_start_io_acct(bio);
+	md_account_bio(mddev, &bio);
+	r10_bio->master_bio = bio;
 	atomic_set(&r10_bio->remaining, 1);
 	md_bitmap_startwrite(mddev->bitmap, r10_bio->sector, r10_bio->sectors, 0);
 
@@ -1571,12 +1579,11 @@ static void __make_request(struct mddev *mddev, struct bio *bio, int sectors)
 	r10_bio->sector = bio->bi_iter.bi_sector;
 	r10_bio->state = 0;
 	r10_bio->read_slot = -1;
-	r10_bio->start_time = 0;
 	memset(r10_bio->devs, 0, sizeof(r10_bio->devs[0]) *
 			conf->geo.raid_disks);
 
 	if (bio_data_dir(bio) == READ)
-		raid10_read_request(mddev, bio, r10_bio);
+		raid10_read_request(mddev, bio, r10_bio, true);
 	else
 		raid10_write_request(mddev, bio, r10_bio);
 }
@@ -1780,10 +1787,9 @@ static int raid10_handle_discard(struct mddev *mddev, struct bio *bio)
 	 */
 	rcu_read_lock();
 	for (disk = 0; disk < geo->raid_disks; disk++) {
-		struct md_rdev *rdev = rcu_dereference(conf->mirrors[disk].rdev);
-		struct md_rdev *rrdev = rcu_dereference(
-			conf->mirrors[disk].replacement);
+		struct md_rdev *rdev, *rrdev;
 
+		rdev = dereference_rdev_and_rrdev(&conf->mirrors[disk], &rrdev);
 		r10_bio->devs[disk].bio = NULL;
 		r10_bio->devs[disk].repl_bio = NULL;
 
@@ -2720,10 +2726,10 @@ static int r10_sync_page_io(struct md_rdev *rdev, sector_t sector,
 static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10bio *r10_bio)
 {
 	int sect = 0; /* Offset from r10_bio->sector */
-	int sectors = r10_bio->sectors;
+	int sectors = r10_bio->sectors, slot = r10_bio->read_slot;
 	struct md_rdev *rdev;
 	int max_read_errors = atomic_read(&mddev->max_corr_read_errors);
-	int d = r10_bio->devs[r10_bio->read_slot].devnum;
+	int d = r10_bio->devs[slot].devnum;
 
 	/* still own a reference to this rdev, so it cannot
 	 * have been cleared recently.
@@ -2744,13 +2750,13 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10
 		pr_notice("md/raid10:%s: %pg: Failing raid device\n",
 			  mdname(mddev), rdev->bdev);
 		md_error(mddev, rdev);
-		r10_bio->devs[r10_bio->read_slot].bio = IO_BLOCKED;
+		r10_bio->devs[slot].bio = IO_BLOCKED;
 		return;
 	}
 
 	while(sectors) {
 		int s = sectors;
-		int sl = r10_bio->read_slot;
+		int sl = slot;
 		int success = 0;
 		int start;
 
@@ -2785,7 +2791,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10
 			sl++;
 			if (sl == conf->copies)
 				sl = 0;
-		} while (!success && sl != r10_bio->read_slot);
+		} while (sl != slot);
 		rcu_read_unlock();
 
 		if (!success) {
@@ -2793,16 +2799,16 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10
 			 * as bad on the first device to discourage future
 			 * reads.
 			 */
-			int dn = r10_bio->devs[r10_bio->read_slot].devnum;
+			int dn = r10_bio->devs[slot].devnum;
 			rdev = conf->mirrors[dn].rdev;
 
 			if (!rdev_set_badblocks(
 				    rdev,
-				    r10_bio->devs[r10_bio->read_slot].addr
+				    r10_bio->devs[slot].addr
 				    + sect,
 				    s, 0)) {
 				md_error(mddev, rdev);
-				r10_bio->devs[r10_bio->read_slot].bio
+				r10_bio->devs[slot].bio
 					= IO_BLOCKED;
 			}
 			break;
@@ -2811,7 +2817,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10
 		start = sl;
 		/* write it back and re-read */
 		rcu_read_lock();
-		while (sl != r10_bio->read_slot) {
+		while (sl != slot) {
 			if (sl==0)
 				sl = conf->copies;
 			sl--;
@@ -2845,7 +2851,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10
 			rcu_read_lock();
 		}
 		sl = start;
-		while (sl != r10_bio->read_slot) {
+		while (sl != slot) {
 			if (sl==0)
 				sl = conf->copies;
 			sl--;
@@ -2985,7 +2991,7 @@ static void handle_read_error(struct mddev *mddev, struct r10bio *r10_bio)
 
 	rdev_dec_pending(rdev, mddev);
 	r10_bio->state = 0;
-	raid10_read_request(mddev, r10_bio->master_bio, r10_bio);
+	raid10_read_request(mddev, r10_bio->master_bio, r10_bio, false);
 	/*
 	 * allow_barrier after re-submit to ensure no sync io
 	 * can be issued while regular io pending.

drivers/md/raid10.h

@@ -123,7 +123,6 @@ struct r10bio {
 	sector_t		sector;	/* virtual sector number */
 	int			sectors;
 	unsigned long		state;
-	unsigned long		start_time;
 	struct mddev		*mddev;
 	/*
 	 * original bio going to /dev/mdx

drivers/md/raid5.c

@@ -5468,26 +5468,17 @@ static struct bio *remove_bio_from_retry(struct r5conf *conf,
  */
 static void raid5_align_endio(struct bio *bi)
 {
-	struct md_io_acct *md_io_acct = bi->bi_private;
-	struct bio *raid_bi = md_io_acct->orig_bio;
-	struct mddev *mddev;
-	struct r5conf *conf;
-	struct md_rdev *rdev;
+	struct bio *raid_bi = bi->bi_private;
+	struct md_rdev *rdev = (void *)raid_bi->bi_next;
+	struct mddev *mddev = rdev->mddev;
+	struct r5conf *conf = mddev->private;
 	blk_status_t error = bi->bi_status;
-	unsigned long start_time = md_io_acct->start_time;
 
 	bio_put(bi);
-
-	rdev = (void*)raid_bi->bi_next;
 	raid_bi->bi_next = NULL;
-	mddev = rdev->mddev;
-	conf = mddev->private;
-
 	rdev_dec_pending(rdev, conf->mddev);
 
 	if (!error) {
-		if (blk_queue_io_stat(raid_bi->bi_bdev->bd_disk->queue))
-			bio_end_io_acct(raid_bi, start_time);
 		bio_endio(raid_bi);
 		if (atomic_dec_and_test(&conf->active_aligned_reads))
 			wake_up(&conf->wait_for_quiescent);
@@ -5506,7 +5497,6 @@ static int raid5_read_one_chunk(struct mddev *mddev, struct bio *raid_bio)
 	struct md_rdev *rdev;
 	sector_t sector, end_sector, first_bad;
 	int bad_sectors, dd_idx;
-	struct md_io_acct *md_io_acct;
 	bool did_inc;
 
 	if (!in_chunk_boundary(mddev, raid_bio)) {
@@ -5543,16 +5533,13 @@ static int raid5_read_one_chunk(struct mddev *mddev, struct bio *raid_bio)
 		return 0;
 	}
 
-	align_bio = bio_alloc_clone(rdev->bdev, raid_bio, GFP_NOIO,
-				    &mddev->io_acct_set);
-	md_io_acct = container_of(align_bio, struct md_io_acct, bio_clone);
+	md_account_bio(mddev, &raid_bio);
 	raid_bio->bi_next = (void *)rdev;
-	if (blk_queue_io_stat(raid_bio->bi_bdev->bd_disk->queue))
-		md_io_acct->start_time = bio_start_io_acct(raid_bio);
-	md_io_acct->orig_bio = raid_bio;
 
+	align_bio = bio_alloc_clone(rdev->bdev, raid_bio, GFP_NOIO,
+				    &mddev->bio_set);
 	align_bio->bi_end_io = raid5_align_endio;
-	align_bio->bi_private = md_io_acct;
+	align_bio->bi_private = raid_bio;
 	align_bio->bi_iter.bi_sector = sector;
 
 	/* No reshape active, so we can trust rdev->data_offset */
@@ -7787,19 +7774,12 @@ static int raid5_run(struct mddev *mddev)
 	struct md_rdev *rdev;
 	struct md_rdev *journal_dev = NULL;
 	sector_t reshape_offset = 0;
-	int i, ret = 0;
+	int i;
 	long long min_offset_diff = 0;
 	int first = 1;
 
-	if (acct_bioset_init(mddev)) {
-		pr_err("md/raid456:%s: alloc acct bioset failed.\n", mdname(mddev));
+	if (mddev_init_writes_pending(mddev) < 0)
 		return -ENOMEM;
-	}
-
-	if (mddev_init_writes_pending(mddev) < 0) {
-		ret = -ENOMEM;
-		goto exit_acct_set;
-	}
 
 	if (mddev->recovery_cp != MaxSector)
 		pr_notice("md/raid:%s: not clean -- starting background reconstruction\n",
@@ -7830,8 +7810,7 @@ static int raid5_run(struct mddev *mddev)
 	    (mddev->bitmap_info.offset || mddev->bitmap_info.file)) {
 		pr_notice("md/raid:%s: array cannot have both journal and bitmap\n",
 			  mdname(mddev));
-		ret = -EINVAL;
-		goto exit_acct_set;
+		return -EINVAL;
 	}
 
 	if (mddev->reshape_position != MaxSector) {
@@ -7856,15 +7835,13 @@ static int raid5_run(struct mddev *mddev)
 		if (journal_dev) {
 			pr_warn("md/raid:%s: don't support reshape with journal - aborting.\n",
 				mdname(mddev));
-			ret = -EINVAL;
-			goto exit_acct_set;
+			return -EINVAL;
 		}
 
 		if (mddev->new_level != mddev->level) {
 			pr_warn("md/raid:%s: unsupported reshape required - aborting.\n",
 				mdname(mddev));
-			ret = -EINVAL;
-			goto exit_acct_set;
+			return -EINVAL;
 		}
 		old_disks = mddev->raid_disks - mddev->delta_disks;
 		/* reshape_position must be on a new-stripe boundary, and one
@@ -7880,8 +7857,7 @@ static int raid5_run(struct mddev *mddev)
 		if (sector_div(here_new, chunk_sectors * new_data_disks)) {
 			pr_warn("md/raid:%s: reshape_position not on a stripe boundary\n",
 				mdname(mddev));
-			ret = -EINVAL;
-			goto exit_acct_set;
+			return -EINVAL;
 		}
 		reshape_offset = here_new * chunk_sectors;
 		/* here_new is the stripe we will write to */
@@ -7903,8 +7879,7 @@ static int raid5_run(struct mddev *mddev)
 			else if (mddev->ro == 0) {
 				pr_warn("md/raid:%s: in-place reshape must be started in read-only mode - aborting\n",
 					mdname(mddev));
-				ret = -EINVAL;
-				goto exit_acct_set;
+				return -EINVAL;
 			}
 		} else if (mddev->reshape_backwards
 		    ? (here_new * chunk_sectors + min_offset_diff <=
@@ -7914,8 +7889,7 @@ static int raid5_run(struct mddev *mddev)
 			/* Reading from the same stripe as writing to - bad */
 			pr_warn("md/raid:%s: reshape_position too early for auto-recovery - aborting.\n",
 				mdname(mddev));
-			ret = -EINVAL;
-			goto exit_acct_set;
+			return -EINVAL;
 		}
 		pr_debug("md/raid:%s: reshape will continue\n", mdname(mddev));
 		/* OK, we should be able to continue; */
@@ -7939,10 +7913,8 @@ static int raid5_run(struct mddev *mddev)
 	else
 		conf = mddev->private;
 
-	if (IS_ERR(conf)) {
-		ret = PTR_ERR(conf);
-		goto exit_acct_set;
-	}
+	if (IS_ERR(conf))
+		return PTR_ERR(conf);
 
 	if (test_bit(MD_HAS_JOURNAL, &mddev->flags)) {
 		if (!journal_dev) {
@@ -8140,10 +8112,7 @@ static int raid5_run(struct mddev *mddev)
 	free_conf(conf);
 	mddev->private = NULL;
 	pr_warn("md/raid:%s: failed to run raid set.\n", mdname(mddev));
-	ret = -EIO;
-exit_acct_set:
-	acct_bioset_exit(mddev);
-	return ret;
+	return -EIO;
 }
 
 static void raid5_free(struct mddev *mddev, void *priv)
@@ -8151,7 +8120,6 @@ static void raid5_free(struct mddev *mddev, void *priv)
 	struct r5conf *conf = priv;
 
 	free_conf(conf);
-	acct_bioset_exit(mddev);
 	mddev->to_remove = &raid5_attrs_group;
 }