Merge remote-tracking branch 'ubifs/linux-next'

12 files changed, 336 insertions, 88 deletions

diff --git a/drivers/mtd/ubi/build.c b/drivers/mtd/ubi/build.c
index 5ebe280225d6..f38e8de81811 100644
--- a/drivers/mtd/ubi/build.c
+++ b/drivers/mtd/ubi/build.c
@@ -690,11 +690,25 @@ static int io_init(struct ubi_device *ubi)
 	ubi_assert(ubi->hdrs_min_io_size <= ubi->min_io_size);
 	ubi_assert(ubi->min_io_size % ubi->hdrs_min_io_size == 0);
 
+	ubi->max_write_size = ubi->mtd->writebufsize;
+	/*
+	 * Maximum write size has to be greater or equivalent to min. I/O
+	 * size, and be multiple of min. I/O size.
+	 */
+	if (ubi->max_write_size < ubi->min_io_size ||
+	    ubi->max_write_size % ubi->min_io_size ||
+	    !is_power_of_2(ubi->max_write_size)) {
+		ubi_err("bad write buffer size %d for %d min. I/O unit",
+			ubi->max_write_size, ubi->min_io_size);
+		return -EINVAL;
+	}
+
 	/* Calculate default aligned sizes of EC and VID headers */
 	ubi->ec_hdr_alsize = ALIGN(UBI_EC_HDR_SIZE, ubi->hdrs_min_io_size);
 	ubi->vid_hdr_alsize = ALIGN(UBI_VID_HDR_SIZE, ubi->hdrs_min_io_size);
 
 	dbg_msg("min_io_size      %d", ubi->min_io_size);
+	dbg_msg("max_write_size   %d", ubi->max_write_size);
 	dbg_msg("hdrs_min_io_size %d", ubi->hdrs_min_io_size);
 	dbg_msg("ec_hdr_alsize    %d", ubi->ec_hdr_alsize);
 	dbg_msg("vid_hdr_alsize   %d", ubi->vid_hdr_alsize);
diff --git a/drivers/mtd/ubi/kapi.c b/drivers/mtd/ubi/kapi.c
index 69fa4ef03c53..d39716e5b204 100644
--- a/drivers/mtd/ubi/kapi.c
+++ b/drivers/mtd/ubi/kapi.c
@@ -40,7 +40,9 @@ void ubi_do_get_device_info(struct ubi_device *ubi, struct ubi_device_info *di)
 {
 	di->ubi_num = ubi->ubi_num;
 	di->leb_size = ubi->leb_size;
+	di->leb_start = ubi->leb_start;
 	di->min_io_size = ubi->min_io_size;
+	di->max_write_size = ubi->max_write_size;
 	di->ro_mode = ubi->ro_mode;
 	di->cdev = ubi->cdev.dev;
 }
diff --git a/drivers/mtd/ubi/ubi.h b/drivers/mtd/ubi/ubi.h
index 0b0149c41fe3..b78994330ebc 100644
--- a/drivers/mtd/ubi/ubi.h
+++ b/drivers/mtd/ubi/ubi.h
@@ -381,6 +381,8 @@ struct ubi_wl_entry;
  * @bad_allowed: whether the MTD device admits of bad physical eraseblocks or
  *               not
  * @nor_flash: non-zero if working on top of NOR flash
+ * @max_write_size: maximum amount of bytes the underlying flash can write at a
+ *                  time (MTD write buffer size)
  * @mtd: MTD device descriptor
  *
  * @peb_buf1: a buffer of PEB size used for different purposes
@@ -464,6 +466,7 @@ struct ubi_device {
 	int vid_hdr_shift;
 	unsigned int bad_allowed:1;
 	unsigned int nor_flash:1;
+	int max_write_size;
 	struct mtd_info *mtd;
 
 	void *peb_buf1;
diff --git a/fs/ubifs/commit.c b/fs/ubifs/commit.c
index 02429d81ca33..b148fbc80f8d 100644
--- a/fs/ubifs/commit.c
+++ b/fs/ubifs/commit.c
@@ -48,6 +48,56 @@
 #include <linux/slab.h>
 #include "ubifs.h"
 
+/*
+ * nothing_to_commit - check if there is nothing to commit.
+ * @c: UBIFS file-system description object
+ *
+ * This is a helper function which checks if there is anything to commit. It is
+ * used as an optimization to avoid starting the commit if it is not really
+ * necessary. Indeed, the commit operation always assumes flash I/O (e.g.,
+ * writing the commit start node to the log), and it is better to avoid doing
+ * this unnecessarily. E.g., 'ubifs_sync_fs()' runs the commit, but if there is
+ * nothing to commit, it is more optimal to avoid any flash I/O.
+ *
+ * This function has to be called with @c->commit_sem locked for writing -
+ * this function does not take LPT/TNC locks because the @c->commit_sem
+ * guarantees that we have exclusive access to the TNC and LPT data structures.
+ *
+ * This function returns %1 if there is nothing to commit and %0 otherwise.
+ */
+static int nothing_to_commit(struct ubifs_info *c)
+{
+	/*
+	 * During mounting or remounting from R/O mode to R/W mode we may
+	 * commit for various recovery-related reasons.
+	 */
+	if (c->mounting || c->remounting_rw)
+		return 0;
+
+	/*
+	 * If the root TNC node is dirty, we definitely have something to
+	 * commit.
+	 */
+	if (c->zroot.znode && test_bit(DIRTY_ZNODE, &c->zroot.znode->flags))
+		return 0;
+
+	/*
+	 * Even though the TNC is clean, the LPT tree may have dirty nodes. For
+	 * example, this may happen if the budgeting subsystem invoked GC to
+	 * make some free space, and the GC found an LEB with only dirty and
+	 * free space. In this case GC would just change the lprops of this
+	 * LEB (by turning all space into free space) and unmap it.
+	 */
+	if (c->nroot && test_bit(DIRTY_CNODE, &c->nroot->flags))
+		return 0;
+
+	ubifs_assert(atomic_long_read(&c->dirty_zn_cnt) == 0);
+	ubifs_assert(c->dirty_pn_cnt == 0);
+	ubifs_assert(c->dirty_nn_cnt == 0);
+
+	return 1;
+}
+
 /**
  * do_commit - commit the journal.
  * @c: UBIFS file-system description object
@@ -70,6 +120,12 @@ static int do_commit(struct ubifs_info *c)
 		goto out_up;
 	}
 
+	if (nothing_to_commit(c)) {
+		up_write(&c->commit_sem);
+		err = 0;
+		goto out_cancel;
+	}
+
 	/* Sync all write buffers (necessary for recovery) */
 	for (i = 0; i < c->jhead_cnt; i++) {
 		err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
@@ -162,12 +218,12 @@ static int do_commit(struct ubifs_info *c)
 	if (err)
 		goto out;
 
+out_cancel:
 	spin_lock(&c->cs_lock);
 	c->cmt_state = COMMIT_RESTING;
 	wake_up(&c->cmt_wq);
 	dbg_cmt("commit end");
 	spin_unlock(&c->cs_lock);
-
 	return 0;
 
 out_up:
diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c
index 0bee4dbffc31..bcb1acb79263 100644
--- a/fs/ubifs/debug.c
+++ b/fs/ubifs/debug.c
@@ -2813,19 +2813,19 @@ int dbg_debugfs_init_fs(struct ubifs_info *c)
 	}
 
 	fname = "dump_lprops";
-	dent = debugfs_create_file(fname, S_IWUGO, d->dfs_dir, c, &dfs_fops);
+	dent = debugfs_create_file(fname, S_IWUSR, d->dfs_dir, c, &dfs_fops);
 	if (IS_ERR(dent))
 		goto out_remove;
 	d->dfs_dump_lprops = dent;
 
 	fname = "dump_budg";
-	dent = debugfs_create_file(fname, S_IWUGO, d->dfs_dir, c, &dfs_fops);
+	dent = debugfs_create_file(fname, S_IWUSR, d->dfs_dir, c, &dfs_fops);
 	if (IS_ERR(dent))
 		goto out_remove;
 	d->dfs_dump_budg = dent;
 
 	fname = "dump_tnc";
-	dent = debugfs_create_file(fname, S_IWUGO, d->dfs_dir, c, &dfs_fops);
+	dent = debugfs_create_file(fname, S_IWUSR, d->dfs_dir, c, &dfs_fops);
 	if (IS_ERR(dent))
 		goto out_remove;
 	d->dfs_dump_tnc = dent;
diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
index d82173182eeb..dfd168b7807e 100644
--- a/fs/ubifs/io.c
+++ b/fs/ubifs/io.c
@@ -31,6 +31,26 @@
  * buffer is full or when it is not used for some time (by timer). This is
  * similar to the mechanism is used by JFFS2.
  *
+ * UBIFS distinguishes between minimum write size (@c->min_io_size) and maximum
+ * write size (@c->max_write_size). The latter is the maximum amount of bytes
+ * the underlying flash is able to program at a time, and writing in
+ * @c->max_write_size units should presumably be faster. Obviously,
+ * @c->min_io_size <= @c->max_write_size. Write-buffers are of
+ * @c->max_write_size bytes in size for maximum performance. However, when a
+ * write-buffer is flushed, only the portion of it (aligned to @c->min_io_size
+ * boundary) which contains data is written, not the whole write-buffer,
+ * because this is more space-efficient.
+ *
+ * This optimization adds few complications to the code. Indeed, on the one
+ * hand, we want to write in optimal @c->max_write_size bytes chunks, which
+ * also means aligning writes at the @c->max_write_size bytes offsets. On the
+ * other hand, we do not want to waste space when synchronizing the write
+ * buffer, so during synchronization we writes in smaller chunks. And this makes
+ * the next write offset to be not aligned to @c->max_write_size bytes. So the
+ * have to make sure that the write-buffer offset (@wbuf->offs) becomes aligned
+ * to @c->max_write_size bytes again. We do this by temporarily shrinking
+ * write-buffer size (@wbuf->size).
+ *
  * Write-buffers are defined by 'struct ubifs_wbuf' objects and protected by
  * mutexes defined inside these objects. Since sometimes upper-level code
  * has to lock the write-buffer (e.g. journal space reservation code), many
@@ -46,8 +66,8 @@
  * UBIFS uses padding when it pads to the next min. I/O unit. In this case it
  * uses padding nodes or padding bytes, if the padding node does not fit.
  *
- * All UBIFS nodes are protected by CRC checksums and UBIFS checks all nodes
- * every time they are read from the flash media.
+ * All UBIFS nodes are protected by CRC checksums and UBIFS checks CRC when
+ * they are read from the flash media.
  */
 
 #include <linux/crc32.h>
@@ -88,8 +108,12 @@ void ubifs_ro_mode(struct ubifs_info *c, int err)
  * This function may skip data nodes CRC checking if @c->no_chk_data_crc is
  * true, which is controlled by corresponding UBIFS mount option. However, if
  * @must_chk_crc is true, then @c->no_chk_data_crc is ignored and CRC is
- * checked. Similarly, if @c->always_chk_crc is true, @c->no_chk_data_crc is
- * ignored and CRC is checked.
+ * checked. Similarly, if @c->mounting or @c->remounting_rw is true (we are
+ * mounting or re-mounting to R/W mode), @c->no_chk_data_crc is ignored and CRC
+ * is checked. This is because during mounting or re-mounting from R/O mode to
+ * R/W mode we may read journal nodes (when replying the journal or doing the
+ * recovery) and the journal nodes may potentially be corrupted, so checking is
+ * required.
  *
  * This function returns zero in case of success and %-EUCLEAN in case of bad
  * CRC or magic.
@@ -131,8 +155,8 @@ int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum,
 		   node_len > c->ranges[type].max_len)
 		goto out_len;
 
-	if (!must_chk_crc && type == UBIFS_DATA_NODE && !c->always_chk_crc &&
-	     c->no_chk_data_crc)
+	if (!must_chk_crc && type == UBIFS_DATA_NODE && !c->mounting &&
+	    !c->remounting_rw && c->no_chk_data_crc)
 		return 0;
 
 	crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
@@ -343,11 +367,17 @@ static void cancel_wbuf_timer_nolock(struct ubifs_wbuf *wbuf)
  *
  * This function synchronizes write-buffer @buf and returns zero in case of
  * success or a negative error code in case of failure.
+ *
+ * Note, although write-buffers are of @c->max_write_size, this function does
+ * not necessarily writes all @c->max_write_size bytes to the flash. Instead,
+ * if the write-buffer is only partially filled with data, only the used part
+ * of the write-buffer (aligned on @c->min_io_size boundary) is synchronized.
+ * This way we waste less space.
  */
 int ubifs_wbuf_sync_nolock(struct ubifs_wbuf *wbuf)
 {
 	struct ubifs_info *c = wbuf->c;
-	int err, dirt;
+	int err, dirt, sync_len;
 
 	cancel_wbuf_timer_nolock(wbuf);
 	if (!wbuf->used || wbuf->lnum == -1)
@@ -357,27 +387,53 @@ int ubifs_wbuf_sync_nolock(struct ubifs_wbuf *wbuf)
 	dbg_io("LEB %d:%d, %d bytes, jhead %s",
 	       wbuf->lnum, wbuf->offs, wbuf->used, dbg_jhead(wbuf->jhead));
 	ubifs_assert(!(wbuf->avail & 7));
-	ubifs_assert(wbuf->offs + c->min_io_size <= c->leb_size);
+	ubifs_assert(wbuf->offs + wbuf->size <= c->leb_size);
+	ubifs_assert(wbuf->size >= c->min_io_size);
+	ubifs_assert(wbuf->size <= c->max_write_size);
+	ubifs_assert(wbuf->size % c->min_io_size == 0);
 	ubifs_assert(!c->ro_media && !c->ro_mount);
+	if (c->leb_size - wbuf->offs >= c->max_write_size)
+		ubifs_assert(!((wbuf->offs + wbuf->size) % c->max_write_size ));
 
 	if (c->ro_error)
 		return -EROFS;
 
-	ubifs_pad(c, wbuf->buf + wbuf->used, wbuf->avail);
+	/*
+	 * Do not write whole write buffer but write only the minimum necessary
+	 * amount of min. I/O units.
+	 */
+	sync_len = ALIGN(wbuf->used, c->min_io_size);
+	dirt = sync_len - wbuf->used;
+	if (dirt)
+		ubifs_pad(c, wbuf->buf + wbuf->used, dirt);
 	err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf, wbuf->offs,
-			    c->min_io_size, wbuf->dtype);
+			    sync_len, wbuf->dtype);
 	if (err) {
 		ubifs_err("cannot write %d bytes to LEB %d:%d",
-			  c->min_io_size, wbuf->lnum, wbuf->offs);
+			  sync_len, wbuf->lnum, wbuf->offs);
 		dbg_dump_stack();
 		return err;
 	}
 
-	dirt = wbuf->avail;
-
 	spin_lock(&wbuf->lock);
-	wbuf->offs += c->min_io_size;
-	wbuf->avail = c->min_io_size;
+	wbuf->offs += sync_len;
+	/*
+	 * Now @wbuf->offs is not necessarily aligned to @c->max_write_size.
+	 * But our goal is to optimize writes and make sure we write in
+	 * @c->max_write_size chunks and to @c->max_write_size-aligned offset.
+	 * Thus, if @wbuf->offs is not aligned to @c->max_write_size now, make
+	 * sure that @wbuf->offs + @wbuf->size is aligned to
+	 * @c->max_write_size. This way we make sure that after next
+	 * write-buffer flush we are again at the optimal offset (aligned to
+	 * @c->max_write_size).
+	 */
+	if (c->leb_size - wbuf->offs < c->max_write_size)
+		wbuf->size = c->leb_size - wbuf->offs;
+	else if (wbuf->offs & (c->max_write_size - 1))
+		wbuf->size = ALIGN(wbuf->offs, c->max_write_size) - wbuf->offs;
+	else
+		wbuf->size = c->max_write_size;
+	wbuf->avail = wbuf->size;
 	wbuf->used = 0;
 	wbuf->next_ino = 0;
 	spin_unlock(&wbuf->lock);
@@ -420,7 +476,13 @@ int ubifs_wbuf_seek_nolock(struct ubifs_wbuf *wbuf, int lnum, int offs,
 	spin_lock(&wbuf->lock);
 	wbuf->lnum = lnum;
 	wbuf->offs = offs;
-	wbuf->avail = c->min_io_size;
+	if (c->leb_size - wbuf->offs < c->max_write_size)
+		wbuf->size = c->leb_size - wbuf->offs;
+	else if (wbuf->offs & (c->max_write_size - 1))
+		wbuf->size = ALIGN(wbuf->offs, c->max_write_size) - wbuf->offs;
+	else
+		wbuf->size = c->max_write_size;
+	wbuf->avail = wbuf->size;
 	wbuf->used = 0;
 	spin_unlock(&wbuf->lock);
 	wbuf->dtype = dtype;
@@ -500,8 +562,9 @@ out_timers:
  *
  * This function writes data to flash via write-buffer @wbuf. This means that
  * the last piece of the node won't reach the flash media immediately if it
- * does not take whole minimal I/O unit. Instead, the node will sit in RAM
- * until the write-buffer is synchronized (e.g., by timer).
+ * does not take whole max. write unit (@c->max_write_size). Instead, the node
+ * will sit in RAM until the write-buffer is synchronized (e.g., by timer, or
+ * because more data are appended to the write-buffer).
  *
  * This function returns zero in case of success and a negative error code in
  * case of failure. If the node cannot be written because there is no more
@@ -518,9 +581,14 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 	ubifs_assert(len > 0 && wbuf->lnum >= 0 && wbuf->lnum < c->leb_cnt);
 	ubifs_assert(wbuf->offs >= 0 && wbuf->offs % c->min_io_size == 0);
 	ubifs_assert(!(wbuf->offs & 7) && wbuf->offs <= c->leb_size);
-	ubifs_assert(wbuf->avail > 0 && wbuf->avail <= c->min_io_size);
+	ubifs_assert(wbuf->avail > 0 && wbuf->avail <= wbuf->size);
+	ubifs_assert(wbuf->size >= c->min_io_size);
+	ubifs_assert(wbuf->size <= c->max_write_size);
+	ubifs_assert(wbuf->size % c->min_io_size == 0);
 	ubifs_assert(mutex_is_locked(&wbuf->io_mutex));
 	ubifs_assert(!c->ro_media && !c->ro_mount);
+	if (c->leb_size - wbuf->offs >= c->max_write_size)
+		ubifs_assert(!((wbuf->offs + wbuf->size) % c->max_write_size ));
 
 	if (c->leb_size - wbuf->offs - wbuf->used < aligned_len) {
 		err = -ENOSPC;
@@ -543,14 +611,18 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 			dbg_io("flush jhead %s wbuf to LEB %d:%d",
 			       dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs);
 			err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf,
-					    wbuf->offs, c->min_io_size,
+					    wbuf->offs, wbuf->size,
 					    wbuf->dtype);
 			if (err)
 				goto out;
 
 			spin_lock(&wbuf->lock);
-			wbuf->offs += c->min_io_size;
-			wbuf->avail = c->min_io_size;
+			wbuf->offs += wbuf->size;
+			if (c->leb_size - wbuf->offs >= c->max_write_size)
+				wbuf->size = c->max_write_size;
+			else
+				wbuf->size = c->leb_size - wbuf->offs;
+			wbuf->avail = wbuf->size;
 			wbuf->used = 0;
 			wbuf->next_ino = 0;
 			spin_unlock(&wbuf->lock);
@@ -564,33 +636,57 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 		goto exit;
 	}
 
-	/*
-	 * The node is large enough and does not fit entirely within current
-	 * minimal I/O unit. We have to fill and flush write-buffer and switch
-	 * to the next min. I/O unit.
-	 */
-	dbg_io("flush jhead %s wbuf to LEB %d:%d",
-	       dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs);
-	memcpy(wbuf->buf + wbuf->used, buf, wbuf->avail);
-	err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf, wbuf->offs,
-			    c->min_io_size, wbuf->dtype);
-	if (err)
-		goto out;
+	offs = wbuf->offs;
+	written = 0;
 
-	offs = wbuf->offs + c->min_io_size;
-	len -= wbuf->avail;
-	aligned_len -= wbuf->avail;
-	written = wbuf->avail;
+	if (wbuf->used) {
+		/*
+		 * The node is large enough and does not fit entirely within
+		 * current available space. We have to fill and flush
+		 * write-buffer and switch to the next max. write unit.
+		 */
+		dbg_io("flush jhead %s wbuf to LEB %d:%d",
+		       dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs);
+		memcpy(wbuf->buf + wbuf->used, buf, wbuf->avail);
+		err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf, wbuf->offs,
+				    wbuf->size, wbuf->dtype);
+		if (err)
+			goto out;
+
+		offs += wbuf->size;
+		len -= wbuf->avail;
+		aligned_len -= wbuf->avail;
+		written += wbuf->avail;
+	} else if (wbuf->offs & (c->max_write_size - 1)) {
+		/*
+		 * The write-buffer offset is not aligned to
+		 * @c->max_write_size and @wbuf->size is less than
+		 * @c->max_write_size. Write @wbuf->size bytes to make sure the
+		 * following writes are done in optimal @c->max_write_size
+		 * chunks.
+		 */
+		dbg_io("write %d bytes to LEB %d:%d",
+		       wbuf->size, wbuf->lnum, wbuf->offs);
+		err = ubi_leb_write(c->ubi, wbuf->lnum, buf, wbuf->offs,
+				    wbuf->size, wbuf->dtype);
+		if (err)
+			goto out;
+
+		offs += wbuf->size;
+		len -= wbuf->size;
+		aligned_len -= wbuf->size;
+		written += wbuf->size;
+	}
 
 	/*
-	 * The remaining data may take more whole min. I/O units, so write the
-	 * remains multiple to min. I/O unit size directly to the flash media.
+	 * The remaining data may take more whole max. write units, so write the
+	 * remains multiple to max. write unit size directly to the flash media.
 	 * We align node length to 8-byte boundary because we anyway flash wbuf
 	 * if the remaining space is less than 8 bytes.
 	 */
-	n = aligned_len >> c->min_io_shift;
+	n = aligned_len >> c->max_write_shift;
 	if (n) {
-		n <<= c->min_io_shift;
+		n <<= c->max_write_shift;
 		dbg_io("write %d bytes to LEB %d:%d", n, wbuf->lnum, offs);
 		err = ubi_leb_write(c->ubi, wbuf->lnum, buf + written, offs, n,
 				    wbuf->dtype);
@@ -606,14 +702,18 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 	if (aligned_len)
 		/*
 		 * And now we have what's left and what does not take whole
-		 * min. I/O unit, so write it to the write-buffer and we are
+		 * max. write unit, so write it to the write-buffer and we are
 		 * done.
 		 */
 		memcpy(wbuf->buf, buf + written, len);
 
 	wbuf->offs = offs;
+	if (c->leb_size - wbuf->offs >= c->max_write_size)
+		wbuf->size = c->max_write_size;
+	else
+		wbuf->size = c->leb_size - wbuf->offs;
+	wbuf->avail = wbuf->size - aligned_len;
 	wbuf->used = aligned_len;
-	wbuf->avail = c->min_io_size - aligned_len;
 	wbuf->next_ino = 0;
 	spin_unlock(&wbuf->lock);
 
@@ -837,11 +937,11 @@ int ubifs_wbuf_init(struct ubifs_info *c, struct ubifs_wbuf *wbuf)
 {
 	size_t size;
 
-	wbuf->buf = kmalloc(c->min_io_size, GFP_KERNEL);
+	wbuf->buf = kmalloc(c->max_write_size, GFP_KERNEL);
 	if (!wbuf->buf)
 		return -ENOMEM;
 
-	size = (c->min_io_size / UBIFS_CH_SZ + 1) * sizeof(ino_t);
+	size = (c->max_write_size / UBIFS_CH_SZ + 1) * sizeof(ino_t);
 	wbuf->inodes = kmalloc(size, GFP_KERNEL);
 	if (!wbuf->inodes) {
 		kfree(wbuf->buf);
@@ -851,7 +951,14 @@ int ubifs_wbuf_init(struct ubifs_info *c, struct ubifs_wbuf *wbuf)
 
 	wbuf->used = 0;
 	wbuf->lnum = wbuf->offs = -1;
-	wbuf->avail = c->min_io_size;
+	/*
+	 * If the LEB starts at the max. write size aligned address, then
+	 * write-buffer size has to be set to @c->max_write_size. Otherwise,
+	 * set it to something smaller so that it ends at the closest max.
+	 * write size boundary.
+	 */
+	size = c->max_write_size - (c->leb_start % c->max_write_size);
+	wbuf->avail = wbuf->size = size;
 	wbuf->dtype = UBI_UNKNOWN;
 	wbuf->sync_callback = NULL;
 	mutex_init(&wbuf->io_mutex);
diff --git a/fs/ubifs/recovery.c b/fs/ubifs/recovery.c
index 77e9b874b6c2..936f2cbfe6b6 100644
--- a/fs/ubifs/recovery.c
+++ b/fs/ubifs/recovery.c
@@ -28,6 +28,23 @@
  * UBIFS always cleans away all remnants of an unclean un-mount, so that
  * errors do not accumulate. However UBIFS defers recovery if it is mounted
  * read-only, and the flash is not modified in that case.
+ *
+ * The general UBIFS approach to the recovery is that it recovers from
+ * corruptions which could be caused by power cuts, but it refuses to recover
+ * from corruption caused by other reasons. And UBIFS tries to distinguish
+ * between these 2 reasons of corruptions and silently recover in the former
+ * case and loudly complain in the latter case.
+ *
+ * UBIFS writes only to erased LEBs, so it writes only to the flash space
+ * containing only 0xFFs. UBIFS also always writes strictly from the beginning
+ * of the LEB to the end. And UBIFS assumes that the underlying flash media
+ * writes in @c->max_write_size bytes at a time.
+ *
+ * Hence, if UBIFS finds a corrupted node at offset X, it expects only the min.
+ * I/O unit corresponding to offset X to contain corrupted data, all the
+ * following min. I/O units have to contain empty space (all 0xFFs). If this is
+ * not true, the corruption cannot be the result of a power cut, and UBIFS
+ * refuses to mount.
  */
 
 #include <linux/crc32.h>
@@ -362,8 +379,9 @@ int ubifs_write_rcvrd_mst_node(struct ubifs_info *c)
  * @offs: offset to check
  *
  * This function returns %1 if @offs was in the last write to the LEB whose data
- * is in @buf, otherwise %0 is returned.  The determination is made by checking
- * for subsequent empty space starting from the next @c->min_io_size boundary.
+ * is in @buf, otherwise %0 is returned. The determination is made by checking
+ * for subsequent empty space starting from the next @c->max_write_size
+ * boundary.
  */
 static int is_last_write(const struct ubifs_info *c, void *buf, int offs)
 {
@@ -371,10 +389,10 @@ static int is_last_write(const struct ubifs_info *c, void *buf, int offs)
 	uint8_t *p;
 
 	/*
-	 * Round up to the next @c->min_io_size boundary i.e. @offs is in the
-	 * last wbuf written. After that should be empty space.
+	 * Round up to the next @c->max_write_size boundary i.e. @offs is in
+	 * the last wbuf written. After that should be empty space.
 	 */
-	empty_offs = ALIGN(offs + 1, c->min_io_size);
+	empty_offs = ALIGN(offs + 1, c->max_write_size);
 	check_len = c->leb_size - empty_offs;
 	p = buf + empty_offs - offs;
 	return is_empty(p, check_len);
@@ -429,7 +447,7 @@ static int no_more_nodes(const struct ubifs_info *c, void *buf, int len,
 	int skip, dlen = le32_to_cpu(ch->len);
 
 	/* Check for empty space after the corrupt node's common header */
-	skip = ALIGN(offs + UBIFS_CH_SZ, c->min_io_size) - offs;
+	skip = ALIGN(offs + UBIFS_CH_SZ, c->max_write_size) - offs;
 	if (is_empty(buf + skip, len - skip))
 		return 1;
 	/*
@@ -441,7 +459,7 @@ static int no_more_nodes(const struct ubifs_info *c, void *buf, int len,
 		return 0;
 	}
 	/* Now we know the corrupt node's length we can skip over it */
-	skip = ALIGN(offs + dlen, c->min_io_size) - offs;
+	skip = ALIGN(offs + dlen, c->max_write_size) - offs;
 	/* After which there should be empty space */
 	if (is_empty(buf + skip, len - skip))
 		return 1;
@@ -671,10 +689,14 @@ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
 		} else {
 			int corruption = first_non_ff(buf, len);
 
+			/*
+			 * See header comment for this file for more
+			 * explanations about the reasons we have this check.
+			 */
 			ubifs_err("corrupt empty space LEB %d:%d, corruption "
 				  "starts at %d", lnum, offs, corruption);
 			/* Make sure we dump interesting non-0xFF data */
-			offs = corruption;
+			offs += corruption;
 			buf += corruption;
 			goto corrupted;
 		}
@@ -836,12 +858,8 @@ struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum,
 static int recover_head(const struct ubifs_info *c, int lnum, int offs,
 			void *sbuf)
 {
-	int len, err;
+	int len = c->max_write_size, err;
 
-	if (c->min_io_size > 1)
-		len = c->min_io_size;
-	else
-		len = 512;
 	if (offs + len > c->leb_size)
 		len = c->leb_size - offs;
 
diff --git a/fs/ubifs/scan.c b/fs/ubifs/scan.c
index 3e1ee57dbeaa..36216b46f772 100644
--- a/fs/ubifs/scan.c
+++ b/fs/ubifs/scan.c
@@ -328,7 +328,7 @@ struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
 		if (!quiet)
 			ubifs_err("empty space starts at non-aligned offset %d",
 				  offs);
-		goto corrupted;;
+		goto corrupted;
 	}
 
 	ubifs_end_scan(c, sleb, lnum, offs);
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 6e11c2975dcf..c20c6d2a0779 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -512,9 +512,12 @@ static int init_constants_early(struct ubifs_info *c)
 
 	c->leb_cnt = c->vi.size;
 	c->leb_size = c->vi.usable_leb_size;
+	c->leb_start = c->di.leb_start;
 	c->half_leb_size = c->leb_size / 2;
 	c->min_io_size = c->di.min_io_size;
 	c->min_io_shift = fls(c->min_io_size) - 1;
+	c->max_write_size = c->di.max_write_size;
+	c->max_write_shift = fls(c->max_write_size) - 1;
 
 	if (c->leb_size < UBIFS_MIN_LEB_SZ) {
 		ubifs_err("too small LEBs (%d bytes), min. is %d bytes",
@@ -534,6 +537,18 @@ static int init_constants_early(struct ubifs_info *c)
 	}
 
 	/*
+	 * Maximum write size has to be greater or equivalent to min. I/O
+	 * size, and be multiple of min. I/O size.
+	 */
+	if (c->max_write_size < c->min_io_size ||
+	    c->max_write_size % c->min_io_size ||
+	    !is_power_of_2(c->max_write_size)) {
+		ubifs_err("bad write buffer size %d for %d min. I/O unit",
+			  c->max_write_size, c->min_io_size);
+		return -EINVAL;
+	}
+
+	/*
 	 * UBIFS aligns all node to 8-byte boundary, so to make function in
 	 * io.c simpler, assume minimum I/O unit size to be 8 bytes if it is
 	 * less than 8.
@@ -541,6 +556,10 @@ static int init_constants_early(struct ubifs_info *c)
 	if (c->min_io_size < 8) {
 		c->min_io_size = 8;
 		c->min_io_shift = 3;
+		if (c->max_write_size < c->min_io_size) {
+			c->max_write_size = c->min_io_size;
+			c->max_write_shift = c->min_io_shift;
+		}
 	}
 
 	c->ref_node_alsz = ALIGN(UBIFS_REF_NODE_SZ, c->min_io_size);
@@ -1202,11 +1221,7 @@ static int mount_ubifs(struct ubifs_info *c)
 	if (c->bulk_read == 1)
 		bu_init(c);
 
-	/*
-	 * We have to check all CRCs, even for data nodes, when we mount the FS
-	 * (specifically, when we are replaying).
-	 */
-	c->always_chk_crc = 1;
+	c->mounting = 1;
 
 	err = ubifs_read_superblock(c);
 	if (err)
@@ -1382,7 +1397,7 @@ static int mount_ubifs(struct ubifs_info *c)
 	if (err)
 		goto out_infos;
 
-	c->always_chk_crc = 0;
+	c->mounting = 0;
 
 	ubifs_msg("mounted UBI device %d, volume %d, name \"%s\"",
 		  c->vi.ubi_num, c->vi.vol_id, c->vi.name);
@@ -1403,6 +1418,7 @@ static int mount_ubifs(struct ubifs_info *c)
 
 	dbg_msg("compiled on:         " __DATE__ " at " __TIME__);
 	dbg_msg("min. I/O unit size:  %d bytes", c->min_io_size);
+	dbg_msg("max. write size:     %d bytes", c->max_write_size);
 	dbg_msg("LEB size:            %d bytes (%d KiB)",
 		c->leb_size, c->leb_size >> 10);
 	dbg_msg("data journal heads:  %d",
@@ -1543,7 +1559,6 @@ static int ubifs_remount_rw(struct ubifs_info *c)
 	mutex_lock(&c->umount_mutex);
 	dbg_save_space_info(c);
 	c->remounting_rw = 1;
-	c->always_chk_crc = 1;
 
 	err = check_free_space(c);
 	if (err)
@@ -1650,7 +1665,6 @@ static int ubifs_remount_rw(struct ubifs_info *c)
 	dbg_gen("re-mounted read-write");
 	c->ro_mount = 0;
 	c->remounting_rw = 0;
-	c->always_chk_crc = 0;
 	err = dbg_check_space_info(c);
 	mutex_unlock(&c->umount_mutex);
 	return err;
@@ -1667,7 +1681,6 @@ out:
 	c->ileb_buf = NULL;
 	ubifs_lpt_free(c, 1);
 	c->remounting_rw = 0;
-	c->always_chk_crc = 0;
 	mutex_unlock(&c->umount_mutex);
 	return err;
 }
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
index ad9cf0133622..de485979ca39 100644
--- a/fs/ubifs/tnc.c
+++ b/fs/ubifs/tnc.c
@@ -447,8 +447,11 @@ static int tnc_read_node_nm(struct ubifs_info *c, struct ubifs_zbranch *zbr,
  *
  * Note, this function does not check CRC of data nodes if @c->no_chk_data_crc
  * is true (it is controlled by corresponding mount option). However, if
- * @c->always_chk_crc is true, @c->no_chk_data_crc is ignored and CRC is always
- * checked.
+ * @c->mounting or @c->remounting_rw is true (we are mounting or re-mounting to
+ * R/W mode), @c->no_chk_data_crc is ignored and CRC is checked. This is
+ * because during mounting or re-mounting from R/O mode to R/W mode we may read
+ * journal nodes (when replying the journal or doing the recovery) and the
+ * journal nodes may potentially be corrupted, so checking is required.
  */
 static int try_read_node(const struct ubifs_info *c, void *buf, int type,
 			 int len, int lnum, int offs)
@@ -476,7 +479,8 @@ static int try_read_node(const struct ubifs_info *c, void *buf, int type,
 	if (node_len != len)
 		return 0;
 
-	if (type == UBIFS_DATA_NODE && !c->always_chk_crc && c->no_chk_data_crc)
+	if (type == UBIFS_DATA_NODE && c->no_chk_data_crc && !c->mounting &&
+	    !c->remounting_rw)
 		return 1;
 
 	crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index 381d6b207a52..362495078489 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -646,6 +646,7 @@ typedef int (*ubifs_lpt_scan_callback)(struct ubifs_info *c,
  * @offs: write-buffer offset in this logical eraseblock
  * @avail: number of bytes available in the write-buffer
  * @used:  number of used bytes in the write-buffer
+ * @size: write-buffer size (in [@c->min_io_size, @c->max_write_size] range)
  * @dtype: type of data stored in this LEB (%UBI_LONGTERM, %UBI_SHORTTERM,
  * %UBI_UNKNOWN)
  * @jhead: journal head the mutex belongs to (note, needed only to shut lockdep
@@ -680,6 +681,7 @@ struct ubifs_wbuf {
 	int offs;
 	int avail;
 	int used;
+	int size;
 	int dtype;
 	int jhead;
 	int (*sync_callback)(struct ubifs_info *c, int lnum, int free, int pad);
@@ -1024,7 +1026,12 @@ struct ubifs_debug_info;
  *
  * @min_io_size: minimal input/output unit size
  * @min_io_shift: number of bits in @min_io_size minus one
+ * @max_write_size: maximum amount of bytes the underlying flash can write at a
+ *                  time (MTD write buffer size)
+ * @max_write_shift: number of bits in @max_write_size minus one
  * @leb_size: logical eraseblock size in bytes
+ * @leb_start: starting offset of logical eraseblocks within physical
+ *             eraseblocks
  * @half_leb_size: half LEB size
  * @idx_leb_size: how many bytes of an LEB are effectively available when it is
  *                used to store indexing nodes (@leb_size - @max_idx_node_sz)
@@ -1166,22 +1173,21 @@ struct ubifs_debug_info;
  * @rp_uid: reserved pool user ID
  * @rp_gid: reserved pool group ID
  *
- * @empty: if the UBI device is empty
+ * @empty: %1 if the UBI device is empty
+ * @need_recovery: %1 if the file-system needs recovery
+ * @replaying: %1 during journal replay
+ * @mounting: %1 while mounting
+ * @remounting_rw: %1 while re-mounting from R/O mode to R/W mode
  * @replay_tree: temporary tree used during journal replay
  * @replay_list: temporary list used during journal replay
  * @replay_buds: list of buds to replay
  * @cs_sqnum: sequence number of first node in the log (commit start node)
  * @replay_sqnum: sequence number of node currently being replayed
- * @need_recovery: file-system needs recovery
- * @replaying: set to %1 during journal replay
  * @unclean_leb_list: LEBs to recover when re-mounting R/O mounted FS to R/W
  *                    mode
  * @rcvrd_mst_node: recovered master node to write when re-mounting R/O mounted
  *                  FS to R/W mode
  * @size_tree: inode size information for recovery
- * @remounting_rw: set while re-mounting from R/O mode to R/W mode
- * @always_chk_crc: always check CRCs (while mounting and remounting to R/W
- *                  mode)
  * @mount_opts: UBIFS-specific mount options
  *
  * @dbg: debugging-related information
@@ -1271,7 +1277,10 @@ struct ubifs_info {
 
 	int min_io_size;
 	int min_io_shift;
+	int max_write_size;
+	int max_write_shift;
 	int leb_size;
+	int leb_start;
 	int half_leb_size;
 	int idx_leb_size;
 	int leb_cnt;
@@ -1402,19 +1411,19 @@ struct ubifs_info {
 	gid_t rp_gid;
 
 	/* The below fields are used only during mounting and re-mounting */
-	int empty;
+	unsigned int empty:1;
+	unsigned int need_recovery:1;
+	unsigned int replaying:1;
+	unsigned int mounting:1;
+	unsigned int remounting_rw:1;
 	struct rb_root replay_tree;
 	struct list_head replay_list;
 	struct list_head replay_buds;
 	unsigned long long cs_sqnum;
 	unsigned long long replay_sqnum;
-	int need_recovery;
-	int replaying;
 	struct list_head unclean_leb_list;
 	struct ubifs_mst_node *rcvrd_mst_node;
 	struct rb_root size_tree;
-	int remounting_rw;
-	int always_chk_crc;
 	struct ubifs_mount_opts mount_opts;
 
 #ifdef CONFIG_UBIFS_FS_DEBUG
diff --git a/include/linux/mtd/ubi.h b/include/linux/mtd/ubi.h
index b31bd9e9bca3..84854edf4436 100644
--- a/include/linux/mtd/ubi.h
+++ b/include/linux/mtd/ubi.h
@@ -116,18 +116,40 @@ struct ubi_volume_info {
  * struct ubi_device_info - UBI device description data structure.
  * @ubi_num: ubi device number
  * @leb_size: logical eraseblock size on this UBI device
+ * @leb_start: starting offset of logical eraseblocks within physical
+ *             eraseblocks
  * @min_io_size: minimal I/O unit size
+ * @max_write_size: maximum amount of bytes the underlying flash can write at a
+ *                  time (MTD write buffer size)
  * @ro_mode: if this device is in read-only mode
  * @cdev: UBI character device major and minor numbers
  *
  * Note, @leb_size is the logical eraseblock size offered by the UBI device.
  * Volumes of this UBI device may have smaller logical eraseblock size if their
  * alignment is not equivalent to %1.
+ *
+ * The @max_write_size field describes flash write maximum write unit. For
+ * example, NOR flash allows for changing individual bytes, so @min_io_size is
+ * %1. However, it does not mean than NOR flash has to write data byte-by-byte.
+ * Instead, CFI NOR flashes have a write-buffer of, e.g., 64 bytes, and when
+ * writing large chunks of data, they write 64-bytes at a time. Obviously, this
+ * improves write throughput.
+ *
+ * Also, the MTD device may have N interleaved (striped) flash chips
+ * underneath, in which case @min_io_size can be physical min. I/O size of
+ * single flash chip, while @max_write_size can be N * @min_io_size.
+ *
+ * The @max_write_size field is always greater or equivalent to @min_io_size.
+ * E.g., some NOR flashes may have (@min_io_size = 1, @max_write_size = 64). In
+ * contrast, NAND flashes usually have @min_io_size = @max_write_size = NAND
+ * page size.
  */
 struct ubi_device_info {
 	int ubi_num;
 	int leb_size;
+	int leb_start;
 	int min_io_size;
+	int max_write_size;
 	int ro_mode;
 	dev_t cdev;
 };