1 files changed, 439 insertions, 345 deletions

diff --git a/fs/ext4/indirect.c b/fs/ext4/indirect.c
index 87b30cd357e7..da76353b3a57 100644
--- a/fs/ext4/indirect.c
+++ b/fs/ext4/indirect.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: GPL-2.0
 /*
  *  linux/fs/ext4/indirect.c
  *
@@ -20,10 +21,10 @@
  *	(sct@redhat.com), 1993, 1998
  */
 
-#include <linux/aio.h>
 #include "ext4_jbd2.h"
 #include "truncate.h"
-#include "ext4_extents.h"	/* Needed for EXT_MAX_BLOCKS */
+#include <linux/dax.h>
+#include <linux/uio.h>
 
 #include <trace/events/ext4.h>
 
@@ -147,6 +148,7 @@ static Indirect *ext4_get_branch(struct inode *inode, int depth,
 	struct super_block *sb = inode->i_sb;
 	Indirect *p = chain;
 	struct buffer_head *bh;
+	unsigned int key;
 	int ret = -EIO;
 
 	*err = 0;
@@ -155,14 +157,20 @@ static Indirect *ext4_get_branch(struct inode *inode, int depth,
 	if (!p->key)
 		goto no_block;
 	while (--depth) {
-		bh = sb_getblk(sb, le32_to_cpu(p->key));
+		key = le32_to_cpu(p->key);
+		if (key > ext4_blocks_count(EXT4_SB(sb)->s_es)) {
+			/* the block was out of range */
+			ret = -EFSCORRUPTED;
+			goto failure;
+		}
+		bh = sb_getblk(sb, key);
 		if (unlikely(!bh)) {
 			ret = -ENOMEM;
 			goto failure;
 		}
 
 		if (!bh_uptodate_or_lock(bh)) {
-			if (bh_submit_read(bh) < 0) {
+			if (ext4_read_bh(bh, 0, NULL, false) < 0) {
 				put_bh(bh);
 				goto failure;
 			}
@@ -293,14 +301,12 @@ static int ext4_blks_to_allocate(Indirect *branch, int k, unsigned int blks,
 }
 
 /**
- *	ext4_alloc_branch - allocate and set up a chain of blocks.
- *	@handle: handle for this transaction
- *	@inode: owner
- *	@indirect_blks: number of allocated indirect blocks
- *	@blks: number of allocated direct blocks
- *	@goal: preferred place for allocation
- *	@offsets: offsets (in the blocks) to store the pointers to next.
- *	@branch: place to store the chain in.
+ * ext4_alloc_branch() - allocate and set up a chain of blocks
+ * @handle: handle for this transaction
+ * @ar: structure describing the allocation request
+ * @indirect_blks: number of allocated indirect blocks
+ * @offsets: offsets (in the blocks) to store the pointers to next.
+ * @branch: place to store the chain in.
  *
  *	This function allocates blocks, zeroes out all but the last one,
  *	links them into chain and (if we are synchronous) writes them to disk.
@@ -319,34 +325,27 @@ static int ext4_blks_to_allocate(Indirect *branch, int k, unsigned int blks,
  *	ext4_alloc_block() (normally -ENOSPC). Otherwise we set the chain
  *	as described above and return 0.
  */
-static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
-			     ext4_lblk_t iblock, int indirect_blks,
-			     int *blks, ext4_fsblk_t goal,
-			     ext4_lblk_t *offsets, Indirect *branch)
+static int ext4_alloc_branch(handle_t *handle,
+			     struct ext4_allocation_request *ar,
+			     int indirect_blks, ext4_lblk_t *offsets,
+			     Indirect *branch)
 {
-	struct ext4_allocation_request	ar;
 	struct buffer_head *		bh;
 	ext4_fsblk_t			b, new_blocks[4];
 	__le32				*p;
 	int				i, j, err, len = 1;
 
-	/*
-	 * Set up for the direct block allocation
-	 */
-	memset(&ar, 0, sizeof(ar));
-	ar.inode = inode;
-	ar.len = *blks;
-	ar.logical = iblock;
-	if (S_ISREG(inode->i_mode))
-		ar.flags = EXT4_MB_HINT_DATA;
-
 	for (i = 0; i <= indirect_blks; i++) {
 		if (i == indirect_blks) {
-			ar.goal = goal;
-			new_blocks[i] = ext4_mb_new_blocks(handle, &ar, &err);
-		} else
-			goal = new_blocks[i] = ext4_new_meta_blocks(handle, inode,
-							goal, 0, NULL, &err);
+			new_blocks[i] = ext4_mb_new_blocks(handle, ar, &err);
+		} else {
+			ar->goal = new_blocks[i] = ext4_new_meta_blocks(handle,
+					ar->inode, ar->goal,
+					ar->flags & EXT4_MB_DELALLOC_RESERVED,
+					NULL, &err);
+			/* Simplify error cleanup... */
+			branch[i+1].bh = NULL;
+		}
 		if (err) {
 			i--;
 			goto failed;
@@ -355,14 +354,15 @@ static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
 		if (i == 0)
 			continue;
 
-		bh = branch[i].bh = sb_getblk(inode->i_sb, new_blocks[i-1]);
+		bh = branch[i].bh = sb_getblk(ar->inode->i_sb, new_blocks[i-1]);
 		if (unlikely(!bh)) {
 			err = -ENOMEM;
 			goto failed;
 		}
 		lock_buffer(bh);
 		BUFFER_TRACE(bh, "call get_create_access");
-		err = ext4_journal_get_create_access(handle, bh);
+		err = ext4_journal_get_create_access(handle, ar->inode->i_sb,
+						     bh, EXT4_JTR_NONE);
 		if (err) {
 			unlock_buffer(bh);
 			goto failed;
@@ -373,7 +373,7 @@ static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
 		b = new_blocks[i];
 
 		if (i == indirect_blks)
-			len = ar.len;
+			len = ar->len;
 		for (j = 0; j < len; j++)
 			*p++ = cpu_to_le32(b++);
 
@@ -382,41 +382,49 @@ static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
 		unlock_buffer(bh);
 
 		BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
-		err = ext4_handle_dirty_metadata(handle, inode, bh);
+		err = ext4_handle_dirty_metadata(handle, ar->inode, bh);
 		if (err)
 			goto failed;
 	}
-	*blks = ar.len;
 	return 0;
 failed:
+	if (i == indirect_blks) {
+		/* Free data blocks */
+		ext4_free_blocks(handle, ar->inode, NULL, new_blocks[i],
+				 ar->len, 0);
+		i--;
+	}
 	for (; i >= 0; i--) {
-		if (i != indirect_blks && branch[i].bh)
-			ext4_forget(handle, 1, inode, branch[i].bh,
-				    branch[i].bh->b_blocknr);
-		ext4_free_blocks(handle, inode, NULL, new_blocks[i],
-				 (i == indirect_blks) ? ar.len : 1, 0);
+		/*
+		 * We want to ext4_forget() only freshly allocated indirect
+		 * blocks. Buffer for new_blocks[i] is at branch[i+1].bh
+		 * (buffer at branch[0].bh is indirect block / inode already
+		 * existing before ext4_alloc_branch() was called). Also
+		 * because blocks are freshly allocated, we don't need to
+		 * revoke them which is why we don't set
+		 * EXT4_FREE_BLOCKS_METADATA.
+		 */
+		ext4_free_blocks(handle, ar->inode, branch[i+1].bh,
+				 new_blocks[i], 1,
+				 branch[i+1].bh ? EXT4_FREE_BLOCKS_FORGET : 0);
 	}
 	return err;
 }
 
 /**
- * ext4_splice_branch - splice the allocated branch onto inode.
+ * ext4_splice_branch() - splice the allocated branch onto inode.
  * @handle: handle for this transaction
- * @inode: owner
- * @block: (logical) number of block we are adding
- * @chain: chain of indirect blocks (with a missing link - see
- *	ext4_alloc_branch)
+ * @ar: structure describing the allocation request
  * @where: location of missing link
  * @num:   number of indirect blocks we are adding
- * @blks:  number of direct blocks we are adding
  *
  * This function fills the missing link and does all housekeeping needed in
  * inode (->i_blocks, etc.). In case of success we end up with the full
  * chain to new block and return 0.
  */
-static int ext4_splice_branch(handle_t *handle, struct inode *inode,
-			      ext4_lblk_t block, Indirect *where, int num,
-			      int blks)
+static int ext4_splice_branch(handle_t *handle,
+			      struct ext4_allocation_request *ar,
+			      Indirect *where, int num)
 {
 	int i;
 	int err = 0;
@@ -429,7 +437,8 @@ static int ext4_splice_branch(handle_t *handle, struct inode *inode,
 	 */
 	if (where->bh) {
 		BUFFER_TRACE(where->bh, "get_write_access");
-		err = ext4_journal_get_write_access(handle, where->bh);
+		err = ext4_journal_get_write_access(handle, ar->inode->i_sb,
+						    where->bh, EXT4_JTR_NONE);
 		if (err)
 			goto err_out;
 	}
@@ -441,9 +450,9 @@ static int ext4_splice_branch(handle_t *handle, struct inode *inode,
 	 * Update the host buffer_head or inode to point to more just allocated
 	 * direct blocks blocks
 	 */
-	if (num == 0 && blks > 1) {
+	if (num == 0 && ar->len > 1) {
 		current_block = le32_to_cpu(where->key) + 1;
-		for (i = 1; i < blks; i++)
+		for (i = 1; i < ar->len; i++)
 			*(where->p + i) = cpu_to_le32(current_block++);
 	}
 
@@ -458,17 +467,19 @@ static int ext4_splice_branch(handle_t *handle, struct inode *inode,
 		 * the new i_size.  But that is not done here - it is done in
 		 * generic_commit_write->__mark_inode_dirty->ext4_dirty_inode.
 		 */
-		jbd_debug(5, "splicing indirect only\n");
+		ext4_debug("splicing indirect only\n");
 		BUFFER_TRACE(where->bh, "call ext4_handle_dirty_metadata");
-		err = ext4_handle_dirty_metadata(handle, inode, where->bh);
+		err = ext4_handle_dirty_metadata(handle, ar->inode, where->bh);
 		if (err)
 			goto err_out;
 	} else {
 		/*
 		 * OK, we spliced it into the inode itself on a direct block.
 		 */
-		ext4_mark_inode_dirty(handle, inode);
-		jbd_debug(5, "splicing direct\n");
+		err = ext4_mark_inode_dirty(handle, ar->inode);
+		if (unlikely(err))
+			goto err_out;
+		ext4_debug("splicing direct\n");
 	}
 	return err;
 
@@ -479,11 +490,11 @@ err_out:
 		 * need to revoke the block, which is why we don't
 		 * need to set EXT4_FREE_BLOCKS_METADATA.
 		 */
-		ext4_free_blocks(handle, inode, where[i].bh, 0, 1,
+		ext4_free_blocks(handle, ar->inode, where[i].bh, 0, 1,
 				 EXT4_FREE_BLOCKS_FORGET);
 	}
-	ext4_free_blocks(handle, inode, NULL, le32_to_cpu(where[num].key),
-			 blks, 0);
+	ext4_free_blocks(handle, ar->inode, NULL, le32_to_cpu(where[num].key),
+			 ar->len, 0);
 
 	return err;
 }
@@ -520,20 +531,20 @@ int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
 			struct ext4_map_blocks *map,
 			int flags)
 {
+	struct ext4_allocation_request ar;
 	int err = -EIO;
 	ext4_lblk_t offsets[4];
 	Indirect chain[4];
 	Indirect *partial;
-	ext4_fsblk_t goal;
 	int indirect_blks;
 	int blocks_to_boundary = 0;
 	int depth;
-	int count = 0;
+	u64 count = 0;
 	ext4_fsblk_t first_block = 0;
 
 	trace_ext4_ind_map_blocks_enter(inode, map->m_lblk, map->m_len, flags);
-	J_ASSERT(!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)));
-	J_ASSERT(handle != NULL || (flags & EXT4_GET_BLOCKS_CREATE) == 0);
+	ASSERT(!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)));
+	ASSERT(handle != NULL || (flags & EXT4_GET_BLOCKS_CREATE) == 0);
 	depth = ext4_block_to_path(inode, map->m_lblk, offsets,
 				   &blocks_to_boundary);
 
@@ -560,21 +571,53 @@ int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
 		goto got_it;
 	}
 
-	/* Next simple case - plain lookup or failed read of indirect block */
-	if ((flags & EXT4_GET_BLOCKS_CREATE) == 0 || err == -EIO)
+	/* Next simple case - plain lookup failed */
+	if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
+		unsigned epb = inode->i_sb->s_blocksize / sizeof(u32);
+		int i;
+
+		/*
+		 * Count number blocks in a subtree under 'partial'. At each
+		 * level we count number of complete empty subtrees beyond
+		 * current offset and then descend into the subtree only
+		 * partially beyond current offset.
+		 */
+		count = 0;
+		for (i = partial - chain + 1; i < depth; i++)
+			count = count * epb + (epb - offsets[i] - 1);
+		count++;
+		/* Fill in size of a hole we found */
+		map->m_pblk = 0;
+		map->m_len = umin(map->m_len, count);
+		goto cleanup;
+	}
+
+	/* Failed read of indirect block */
+	if (err == -EIO)
 		goto cleanup;
 
 	/*
 	 * Okay, we need to do block allocation.
 	*/
-	if (EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
-				       EXT4_FEATURE_RO_COMPAT_BIGALLOC)) {
+	if (ext4_has_feature_bigalloc(inode->i_sb)) {
 		EXT4_ERROR_INODE(inode, "Can't allocate blocks for "
 				 "non-extent mapped inodes with bigalloc");
-		return -ENOSPC;
+		err = -EFSCORRUPTED;
+		goto out;
 	}
 
-	goal = ext4_find_goal(inode, map->m_lblk, partial);
+	/* Set up for the direct block allocation */
+	memset(&ar, 0, sizeof(ar));
+	ar.inode = inode;
+	ar.logical = map->m_lblk;
+	if (S_ISREG(inode->i_mode))
+		ar.flags = EXT4_MB_HINT_DATA;
+	if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE)
+		ar.flags |= EXT4_MB_DELALLOC_RESERVED;
+	if (flags & EXT4_GET_BLOCKS_METADATA_NOFAIL)
+		ar.flags |= EXT4_MB_USE_RESERVED;
+
+	ar.goal = ext4_find_goal(inode, map->m_lblk, partial);
 
 	/* the number of blocks need to allocate for [d,t]indirect blocks */
 	indirect_blks = (chain + depth) - partial - 1;
@@ -583,13 +626,13 @@ int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
 	 * Next look up the indirect map to count the totoal number of
 	 * direct blocks to allocate for this branch.
 	 */
-	count = ext4_blks_to_allocate(partial, indirect_blks,
-				      map->m_len, blocks_to_boundary);
+	ar.len = ext4_blks_to_allocate(partial, indirect_blks,
+				       map->m_len, blocks_to_boundary);
+
 	/*
 	 * Block out ext4_truncate while we alter the tree
 	 */
-	err = ext4_alloc_branch(handle, inode, map->m_lblk, indirect_blks,
-				&count, goal,
+	err = ext4_alloc_branch(handle, &ar, indirect_blks,
 				offsets + (partial - chain), partial);
 
 	/*
@@ -600,14 +643,15 @@ int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
 	 * may need to return -EAGAIN upwards in the worst case.  --sct
 	 */
 	if (!err)
-		err = ext4_splice_branch(handle, inode, map->m_lblk,
-					 partial, indirect_blks, count);
+		err = ext4_splice_branch(handle, &ar, partial, indirect_blks);
 	if (err)
 		goto cleanup;
 
 	map->m_flags |= EXT4_MAP_NEW;
 
 	ext4_update_inode_fsync_trans(handle, inode, 1);
+	count = ar.len;
+
 got_it:
 	map->m_flags |= EXT4_MAP_MAPPED;
 	map->m_pblk = le32_to_cpu(chain[depth-1].key);
@@ -629,152 +673,6 @@ out:
 }
 
 /*
- * O_DIRECT for ext3 (or indirect map) based files
- *
- * If the O_DIRECT write will extend the file then add this inode to the
- * orphan list.  So recovery will truncate it back to the original size
- * if the machine crashes during the write.
- *
- * If the O_DIRECT write is intantiating holes inside i_size and the machine
- * crashes then stale disk data _may_ be exposed inside the file. But current
- * VFS code falls back into buffered path in that case so we are safe.
- */
-ssize_t ext4_ind_direct_IO(int rw, struct kiocb *iocb,
-			   const struct iovec *iov, loff_t offset,
-			   unsigned long nr_segs)
-{
-	struct file *file = iocb->ki_filp;
-	struct inode *inode = file->f_mapping->host;
-	struct ext4_inode_info *ei = EXT4_I(inode);
-	handle_t *handle;
-	ssize_t ret;
-	int orphan = 0;
-	size_t count = iov_length(iov, nr_segs);
-	int retries = 0;
-
-	if (rw == WRITE) {
-		loff_t final_size = offset + count;
-
-		if (final_size > inode->i_size) {
-			/* Credits for sb + inode write */
-			handle = ext4_journal_start(inode, EXT4_HT_INODE, 2);
-			if (IS_ERR(handle)) {
-				ret = PTR_ERR(handle);
-				goto out;
-			}
-			ret = ext4_orphan_add(handle, inode);
-			if (ret) {
-				ext4_journal_stop(handle);
-				goto out;
-			}
-			orphan = 1;
-			ei->i_disksize = inode->i_size;
-			ext4_journal_stop(handle);
-		}
-	}
-
-retry:
-	if (rw == READ && ext4_should_dioread_nolock(inode)) {
-		/*
-		 * Nolock dioread optimization may be dynamically disabled
-		 * via ext4_inode_block_unlocked_dio(). Check inode's state
-		 * while holding extra i_dio_count ref.
-		 */
-		atomic_inc(&inode->i_dio_count);
-		smp_mb();
-		if (unlikely(ext4_test_inode_state(inode,
-						    EXT4_STATE_DIOREAD_LOCK))) {
-			inode_dio_done(inode);
-			goto locked;
-		}
-		ret = __blockdev_direct_IO(rw, iocb, inode,
-				 inode->i_sb->s_bdev, iov,
-				 offset, nr_segs,
-				 ext4_get_block, NULL, NULL, 0);
-		inode_dio_done(inode);
-	} else {
-locked:
-		ret = blockdev_direct_IO(rw, iocb, inode, iov,
-				 offset, nr_segs, ext4_get_block);
-
-		if (unlikely((rw & WRITE) && ret < 0)) {
-			loff_t isize = i_size_read(inode);
-			loff_t end = offset + iov_length(iov, nr_segs);
-
-			if (end > isize)
-				ext4_truncate_failed_write(inode);
-		}
-	}
-	if (ret == -ENOSPC && ext4_should_retry_alloc(inode->i_sb, &retries))
-		goto retry;
-
-	if (orphan) {
-		int err;
-
-		/* Credits for sb + inode write */
-		handle = ext4_journal_start(inode, EXT4_HT_INODE, 2);
-		if (IS_ERR(handle)) {
-			/* This is really bad luck. We've written the data
-			 * but cannot extend i_size. Bail out and pretend
-			 * the write failed... */
-			ret = PTR_ERR(handle);
-			if (inode->i_nlink)
-				ext4_orphan_del(NULL, inode);
-
-			goto out;
-		}
-		if (inode->i_nlink)
-			ext4_orphan_del(handle, inode);
-		if (ret > 0) {
-			loff_t end = offset + ret;
-			if (end > inode->i_size) {
-				ei->i_disksize = end;
-				i_size_write(inode, end);
-				/*
-				 * We're going to return a positive `ret'
-				 * here due to non-zero-length I/O, so there's
-				 * no way of reporting error returns from
-				 * ext4_mark_inode_dirty() to userspace.  So
-				 * ignore it.
-				 */
-				ext4_mark_inode_dirty(handle, inode);
-			}
-		}
-		err = ext4_journal_stop(handle);
-		if (ret == 0)
-			ret = err;
-	}
-out:
-	return ret;
-}
-
-/*
- * Calculate the number of metadata blocks need to reserve
- * to allocate a new block at @lblocks for non extent file based file
- */
-int ext4_ind_calc_metadata_amount(struct inode *inode, sector_t lblock)
-{
-	struct ext4_inode_info *ei = EXT4_I(inode);
-	sector_t dind_mask = ~((sector_t)EXT4_ADDR_PER_BLOCK(inode->i_sb) - 1);
-	int blk_bits;
-
-	if (lblock < EXT4_NDIR_BLOCKS)
-		return 0;
-
-	lblock -= EXT4_NDIR_BLOCKS;
-
-	if (ei->i_da_metadata_calc_len &&
-	    (lblock & dind_mask) == ei->i_da_metadata_calc_last_lblock) {
-		ei->i_da_metadata_calc_len++;
-		return 0;
-	}
-	ei->i_da_metadata_calc_last_lblock = lblock & dind_mask;
-	ei->i_da_metadata_calc_len = 1;
-	blk_bits = order_base_2(lblock);
-	return (blk_bits / EXT4_ADDR_PER_BLOCK_BITS(inode->i_sb)) + 1;
-}
-
-/*
  * Calculate number of indirect blocks touched by mapping @nrblocks logically
  * contiguous blocks
  */
@@ -788,27 +686,64 @@ int ext4_ind_trans_blocks(struct inode *inode, int nrblocks)
 	return DIV_ROUND_UP(nrblocks, EXT4_ADDR_PER_BLOCK(inode->i_sb)) + 4;
 }
 
+static int ext4_ind_trunc_restart_fn(handle_t *handle, struct inode *inode,
+				     struct buffer_head *bh, int *dropped)
+{
+	int err;
+
+	if (bh) {
+		BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+		err = ext4_handle_dirty_metadata(handle, inode, bh);
+		if (unlikely(err))
+			return err;
+	}
+	err = ext4_mark_inode_dirty(handle, inode);
+	if (unlikely(err))
+		return err;
+	/*
+	 * Drop i_data_sem to avoid deadlock with ext4_map_blocks.  At this
+	 * moment, get_block can be called only for blocks inside i_size since
+	 * page cache has been already dropped and writes are blocked by
+	 * i_rwsem. So we can safely drop the i_data_sem here.
+	 */
+	BUG_ON(EXT4_JOURNAL(inode) == NULL);
+	ext4_discard_preallocations(inode);
+	up_write(&EXT4_I(inode)->i_data_sem);
+	*dropped = 1;
+	return 0;
+}
+
 /*
  * Truncate transactions can be complex and absolutely huge.  So we need to
- * be able to restart the transaction at a conventient checkpoint to make
+ * be able to restart the transaction at a convenient checkpoint to make
  * sure we don't overflow the journal.
  *
  * Try to extend this transaction for the purposes of truncation.  If
- * extend fails, we need to propagate the failure up and restart the
- * transaction in the top-level truncate loop. --sct
- *
- * Returns 0 if we managed to create more room.  If we can't create more
- * room, and the transaction must be restarted we return 1.
+ * extend fails, we restart transaction.
  */
-static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
+static int ext4_ind_truncate_ensure_credits(handle_t *handle,
+					    struct inode *inode,
+					    struct buffer_head *bh,
+					    int revoke_creds)
 {
-	if (!ext4_handle_valid(handle))
-		return 0;
-	if (ext4_handle_has_enough_credits(handle, EXT4_RESERVE_TRANS_BLOCKS+1))
-		return 0;
-	if (!ext4_journal_extend(handle, ext4_blocks_for_truncate(inode)))
-		return 0;
-	return 1;
+	int ret;
+	int dropped = 0;
+
+	ret = ext4_journal_ensure_credits_fn(handle, EXT4_RESERVE_TRANS_BLOCKS,
+			ext4_blocks_for_truncate(inode), revoke_creds,
+			ext4_ind_trunc_restart_fn(handle, inode, bh, &dropped));
+	if (dropped)
+		down_write(&EXT4_I(inode)->i_data_sem);
+	if (ret <= 0)
+		return ret;
+	if (bh) {
+		BUFFER_TRACE(bh, "retaking write access");
+		ret = ext4_journal_get_write_access(handle, inode->i_sb, bh,
+						    EXT4_JTR_NONE);
+		if (unlikely(ret))
+			return ret;
+	}
+	return 0;
 }
 
 /*
@@ -929,40 +864,23 @@ static int ext4_clear_blocks(handle_t *handle, struct inode *inode,
 	int	flags = EXT4_FREE_BLOCKS_VALIDATED;
 	int	err;
 
-	if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
+	if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode) ||
+	    ext4_test_inode_flag(inode, EXT4_INODE_EA_INODE))
 		flags |= EXT4_FREE_BLOCKS_FORGET | EXT4_FREE_BLOCKS_METADATA;
 	else if (ext4_should_journal_data(inode))
 		flags |= EXT4_FREE_BLOCKS_FORGET;
 
-	if (!ext4_data_block_valid(EXT4_SB(inode->i_sb), block_to_free,
-				   count)) {
+	if (!ext4_inode_block_valid(inode, block_to_free, count)) {
 		EXT4_ERROR_INODE(inode, "attempt to clear invalid "
 				 "blocks %llu len %lu",
 				 (unsigned long long) block_to_free, count);
 		return 1;
 	}
 
-	if (try_to_extend_transaction(handle, inode)) {
-		if (bh) {
-			BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
-			err = ext4_handle_dirty_metadata(handle, inode, bh);
-			if (unlikely(err))
-				goto out_err;
-		}
-		err = ext4_mark_inode_dirty(handle, inode);
-		if (unlikely(err))
-			goto out_err;
-		err = ext4_truncate_restart_trans(handle, inode,
-					ext4_blocks_for_truncate(inode));
-		if (unlikely(err))
-			goto out_err;
-		if (bh) {
-			BUFFER_TRACE(bh, "retaking write access");
-			err = ext4_journal_get_write_access(handle, bh);
-			if (unlikely(err))
-				goto out_err;
-		}
-	}
+	err = ext4_ind_truncate_ensure_credits(handle, inode, bh,
+				ext4_free_data_revoke_credits(inode, count));
+	if (err < 0)
+		goto out_err;
 
 	for (p = first; p < last; p++)
 		*p = 0;
@@ -1009,7 +927,8 @@ static void ext4_free_data(handle_t *handle, struct inode *inode,
 
 	if (this_bh) {				/* For indirect block */
 		BUFFER_TRACE(this_bh, "get_write_access");
-		err = ext4_journal_get_write_access(handle, this_bh);
+		err = ext4_journal_get_write_access(handle, inode->i_sb,
+						    this_bh, EXT4_JTR_NONE);
 		/* Important: if we can't update the indirect pointers
 		 * to the blocks, we can't free them. */
 		if (err)
@@ -1097,8 +1016,7 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 			if (!nr)
 				continue;		/* A hole */
 
-			if (!ext4_data_block_valid(EXT4_SB(inode->i_sb),
-						   nr, 1)) {
+			if (!ext4_inode_block_valid(inode, nr, 1)) {
 				EXT4_ERROR_INODE(inode,
 						 "invalid indirect mapped "
 						 "block %lu (level %d)",
@@ -1107,14 +1025,14 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 			}
 
 			/* Go read the buffer for the next level down */
-			bh = sb_bread(inode->i_sb, nr);
+			bh = ext4_sb_bread_nofail(inode->i_sb, nr);
 
 			/*
 			 * A read failure? Report error and clear slot
 			 * (should be rare).
 			 */
-			if (!bh) {
-				EXT4_ERROR_INODE_BLOCK(inode, nr,
+			if (IS_ERR(bh)) {
+				ext4_error_inode_block(inode, nr, -PTR_ERR(bh),
 						       "Read failure");
 				continue;
 			}
@@ -1128,7 +1046,7 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 			brelse(bh);
 
 			/*
-			 * Everything below this this pointer has been
+			 * Everything below this pointer has been
 			 * released.  Now let this top-of-subtree go.
 			 *
 			 * We want the freeing of this indirect block to be
@@ -1145,11 +1063,11 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 			 */
 			if (ext4_handle_is_aborted(handle))
 				return;
-			if (try_to_extend_transaction(handle, inode)) {
-				ext4_mark_inode_dirty(handle, inode);
-				ext4_truncate_restart_trans(handle, inode,
-					    ext4_blocks_for_truncate(inode));
-			}
+			if (ext4_ind_truncate_ensure_credits(handle, inode,
+					NULL,
+					ext4_free_metadata_revoke_credits(
+							inode->i_sb, 1)) < 0)
+				return;
 
 			/*
 			 * The forget flag here is critical because if
@@ -1173,7 +1091,8 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 				 */
 				BUFFER_TRACE(parent_bh, "get_write_access");
 				if (!ext4_journal_get_write_access(handle,
-								   parent_bh)){
+						inode->i_sb, parent_bh,
+						EXT4_JTR_NONE)) {
 					*p = 0;
 					BUFFER_TRACE(parent_bh,
 					"call ext4_handle_dirty_metadata");
@@ -1275,106 +1194,281 @@ do_indirects:
 			ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 1);
 			i_data[EXT4_IND_BLOCK] = 0;
 		}
+		fallthrough;
 	case EXT4_IND_BLOCK:
 		nr = i_data[EXT4_DIND_BLOCK];
 		if (nr) {
 			ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 2);
 			i_data[EXT4_DIND_BLOCK] = 0;
 		}
+		fallthrough;
 	case EXT4_DIND_BLOCK:
 		nr = i_data[EXT4_TIND_BLOCK];
 		if (nr) {
 			ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 3);
 			i_data[EXT4_TIND_BLOCK] = 0;
 		}
+		fallthrough;
 	case EXT4_TIND_BLOCK:
 		;
 	}
 }
 
-static int free_hole_blocks(handle_t *handle, struct inode *inode,
-			    struct buffer_head *parent_bh, __le32 *i_data,
-			    int level, ext4_lblk_t first,
-			    ext4_lblk_t count, int max)
+/**
+ *	ext4_ind_remove_space - remove space from the range
+ *	@handle: JBD handle for this transaction
+ *	@inode:	inode we are dealing with
+ *	@start:	First block to remove
+ *	@end:	One block after the last block to remove (exclusive)
+ *
+ *	Free the blocks in the defined range (end is exclusive endpoint of
+ *	range). This is used by ext4_punch_hole().
+ */
+int ext4_ind_remove_space(handle_t *handle, struct inode *inode,
+			  ext4_lblk_t start, ext4_lblk_t end)
 {
-	struct buffer_head *bh = NULL;
+	struct ext4_inode_info *ei = EXT4_I(inode);
+	__le32 *i_data = ei->i_data;
 	int addr_per_block = EXT4_ADDR_PER_BLOCK(inode->i_sb);
-	int ret = 0;
-	int i, inc;
-	ext4_lblk_t offset;
-	__le32 blk;
-
-	inc = 1 << ((EXT4_BLOCK_SIZE_BITS(inode->i_sb) - 2) * level);
-	for (i = 0, offset = 0; i < max; i++, i_data++, offset += inc) {
-		if (offset >= count + first)
-			break;
-		if (*i_data == 0 || (offset + inc) <= first)
-			continue;
-		blk = *i_data;
-		if (level > 0) {
-			ext4_lblk_t first2;
-			bh = sb_bread(inode->i_sb, le32_to_cpu(blk));
-			if (!bh) {
-				EXT4_ERROR_INODE_BLOCK(inode, le32_to_cpu(blk),
-						       "Read failure");
-				return -EIO;
+	ext4_lblk_t offsets[4], offsets2[4];
+	Indirect chain[4], chain2[4];
+	Indirect *partial, *partial2;
+	Indirect *p = NULL, *p2 = NULL;
+	ext4_lblk_t max_block;
+	__le32 nr = 0, nr2 = 0;
+	int n = 0, n2 = 0;
+	unsigned blocksize = inode->i_sb->s_blocksize;
+
+	max_block = (EXT4_SB(inode->i_sb)->s_bitmap_maxbytes + blocksize-1)
+					>> EXT4_BLOCK_SIZE_BITS(inode->i_sb);
+	if (end >= max_block)
+		end = max_block;
+	if ((start >= end) || (start > max_block))
+		return 0;
+
+	n = ext4_block_to_path(inode, start, offsets, NULL);
+	n2 = ext4_block_to_path(inode, end, offsets2, NULL);
+
+	BUG_ON(n > n2);
+
+	if ((n == 1) && (n == n2)) {
+		/* We're punching only within direct block range */
+		ext4_free_data(handle, inode, NULL, i_data + offsets[0],
+			       i_data + offsets2[0]);
+		return 0;
+	} else if (n2 > n) {
+		/*
+		 * Start and end are on a different levels so we're going to
+		 * free partial block at start, and partial block at end of
+		 * the range. If there are some levels in between then
+		 * do_indirects label will take care of that.
+		 */
+
+		if (n == 1) {
+			/*
+			 * Start is at the direct block level, free
+			 * everything to the end of the level.
+			 */
+			ext4_free_data(handle, inode, NULL, i_data + offsets[0],
+				       i_data + EXT4_NDIR_BLOCKS);
+			goto end_range;
+		}
+
+
+		partial = p = ext4_find_shared(inode, n, offsets, chain, &nr);
+		if (nr) {
+			if (partial == chain) {
+				/* Shared branch grows from the inode */
+				ext4_free_branches(handle, inode, NULL,
+					   &nr, &nr+1, (chain+n-1) - partial);
+				*partial->p = 0;
+			} else {
+				/* Shared branch grows from an indirect block */
+				BUFFER_TRACE(partial->bh, "get_write_access");
+				ext4_free_branches(handle, inode, partial->bh,
+					partial->p,
+					partial->p+1, (chain+n-1) - partial);
 			}
-			first2 = (first > offset) ? first - offset : 0;
-			ret = free_hole_blocks(handle, inode, bh,
-					       (__le32 *)bh->b_data, level - 1,
-					       first2, count - offset,
-					       inode->i_sb->s_blocksize >> 2);
-			if (ret) {
-				brelse(bh);
-				goto err;
+		}
+
+		/*
+		 * Clear the ends of indirect blocks on the shared branch
+		 * at the start of the range
+		 */
+		while (partial > chain) {
+			ext4_free_branches(handle, inode, partial->bh,
+				partial->p + 1,
+				(__le32 *)partial->bh->b_data+addr_per_block,
+				(chain+n-1) - partial);
+			partial--;
+		}
+
+end_range:
+		partial2 = p2 = ext4_find_shared(inode, n2, offsets2, chain2, &nr2);
+		if (nr2) {
+			if (partial2 == chain2) {
+				/*
+				 * Remember, end is exclusive so here we're at
+				 * the start of the next level we're not going
+				 * to free. Everything was covered by the start
+				 * of the range.
+				 */
+				goto do_indirects;
 			}
+		} else {
+			/*
+			 * ext4_find_shared returns Indirect structure which
+			 * points to the last element which should not be
+			 * removed by truncate. But this is end of the range
+			 * in punch_hole so we need to point to the next element
+			 */
+			partial2->p++;
 		}
-		if (level == 0 ||
-		    (bh && all_zeroes((__le32 *)bh->b_data,
-				      (__le32 *)bh->b_data + addr_per_block))) {
-			ext4_free_data(handle, inode, parent_bh, &blk, &blk+1);
-			*i_data = 0;
+
+		/*
+		 * Clear the ends of indirect blocks on the shared branch
+		 * at the end of the range
+		 */
+		while (partial2 > chain2) {
+			ext4_free_branches(handle, inode, partial2->bh,
+					   (__le32 *)partial2->bh->b_data,
+					   partial2->p,
+					   (chain2+n2-1) - partial2);
+			partial2--;
 		}
-		brelse(bh);
-		bh = NULL;
+		goto do_indirects;
 	}
 
-err:
-	return ret;
-}
+	/* Punch happened within the same level (n == n2) */
+	partial = p = ext4_find_shared(inode, n, offsets, chain, &nr);
+	partial2 = p2 = ext4_find_shared(inode, n2, offsets2, chain2, &nr2);
 
-int ext4_free_hole_blocks(handle_t *handle, struct inode *inode,
-			  ext4_lblk_t first, ext4_lblk_t stop)
-{
-	int addr_per_block = EXT4_ADDR_PER_BLOCK(inode->i_sb);
-	int level, ret = 0;
-	int num = EXT4_NDIR_BLOCKS;
-	ext4_lblk_t count, max = EXT4_NDIR_BLOCKS;
-	__le32 *i_data = EXT4_I(inode)->i_data;
-
-	count = stop - first;
-	for (level = 0; level < 4; level++, max *= addr_per_block) {
-		if (first < max) {
-			ret = free_hole_blocks(handle, inode, NULL, i_data,
-					       level, first, count, num);
-			if (ret)
-				goto err;
-			if (count > max - first)
-				count -= max - first;
-			else
+	/* Free top, but only if partial2 isn't its subtree. */
+	if (nr) {
+		int level = min(partial - chain, partial2 - chain2);
+		int i;
+		int subtree = 1;
+
+		for (i = 0; i <= level; i++) {
+			if (offsets[i] != offsets2[i]) {
+				subtree = 0;
 				break;
-			first = 0;
-		} else {
-			first -= max;
+			}
 		}
-		i_data += num;
-		if (level == 0) {
-			num = 1;
-			max = 1;
+
+		if (!subtree) {
+			if (partial == chain) {
+				/* Shared branch grows from the inode */
+				ext4_free_branches(handle, inode, NULL,
+						   &nr, &nr+1,
+						   (chain+n-1) - partial);
+				*partial->p = 0;
+			} else {
+				/* Shared branch grows from an indirect block */
+				BUFFER_TRACE(partial->bh, "get_write_access");
+				ext4_free_branches(handle, inode, partial->bh,
+						   partial->p,
+						   partial->p+1,
+						   (chain+n-1) - partial);
+			}
 		}
 	}
 
-err:
-	return ret;
-}
+	if (!nr2) {
+		/*
+		 * ext4_find_shared returns Indirect structure which
+		 * points to the last element which should not be
+		 * removed by truncate. But this is end of the range
+		 * in punch_hole so we need to point to the next element
+		 */
+		partial2->p++;
+	}
+
+	while (partial > chain || partial2 > chain2) {
+		int depth = (chain+n-1) - partial;
+		int depth2 = (chain2+n2-1) - partial2;
+
+		if (partial > chain && partial2 > chain2 &&
+		    partial->bh->b_blocknr == partial2->bh->b_blocknr) {
+			/*
+			 * We've converged on the same block. Clear the range,
+			 * then we're done.
+			 */
+			ext4_free_branches(handle, inode, partial->bh,
+					   partial->p + 1,
+					   partial2->p,
+					   (chain+n-1) - partial);
+			goto cleanup;
+		}
+
+		/*
+		 * The start and end partial branches may not be at the same
+		 * level even though the punch happened within one level. So, we
+		 * give them a chance to arrive at the same level, then walk
+		 * them in step with each other until we converge on the same
+		 * block.
+		 */
+		if (partial > chain && depth <= depth2) {
+			ext4_free_branches(handle, inode, partial->bh,
+					   partial->p + 1,
+					   (__le32 *)partial->bh->b_data+addr_per_block,
+					   (chain+n-1) - partial);
+			partial--;
+		}
+		if (partial2 > chain2 && depth2 <= depth) {
+			ext4_free_branches(handle, inode, partial2->bh,
+					   (__le32 *)partial2->bh->b_data,
+					   partial2->p,
+					   (chain2+n2-1) - partial2);
+			partial2--;
+		}
+	}
 
+cleanup:
+	while (p && p > chain) {
+		BUFFER_TRACE(p->bh, "call brelse");
+		brelse(p->bh);
+		p--;
+	}
+	while (p2 && p2 > chain2) {
+		BUFFER_TRACE(p2->bh, "call brelse");
+		brelse(p2->bh);
+		p2--;
+	}
+	return 0;
+
+do_indirects:
+	/* Kill the remaining (whole) subtrees */
+	switch (offsets[0]) {
+	default:
+		if (++n >= n2)
+			break;
+		nr = i_data[EXT4_IND_BLOCK];
+		if (nr) {
+			ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 1);
+			i_data[EXT4_IND_BLOCK] = 0;
+		}
+		fallthrough;
+	case EXT4_IND_BLOCK:
+		if (++n >= n2)
+			break;
+		nr = i_data[EXT4_DIND_BLOCK];
+		if (nr) {
+			ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 2);
+			i_data[EXT4_DIND_BLOCK] = 0;
+		}
+		fallthrough;
+	case EXT4_DIND_BLOCK:
+		if (++n >= n2)
+			break;
+		nr = i_data[EXT4_TIND_BLOCK];
+		if (nr) {
+			ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 3);
+			i_data[EXT4_TIND_BLOCK] = 0;
+		}
+		fallthrough;
+	case EXT4_TIND_BLOCK:
+		;
+	}
+	goto cleanup;
+}