1 files changed, 67 insertions, 35 deletions

diff --git a/fs/ext4/file.c b/fs/ext4/file.c
index 54d6ff22585c..21df81347147 100644
--- a/fs/ext4/file.c
+++ b/fs/ext4/file.c
@@ -306,7 +306,7 @@ out:
 }
 
 static ssize_t ext4_handle_inode_extension(struct inode *inode, loff_t offset,
-					   ssize_t count)
+					   ssize_t written, ssize_t count)
 {
 	handle_t *handle;
 
@@ -315,7 +315,7 @@ static ssize_t ext4_handle_inode_extension(struct inode *inode, loff_t offset,
 	if (IS_ERR(handle))
 		return PTR_ERR(handle);
 
-	if (ext4_update_inode_size(inode, offset + count)) {
+	if (ext4_update_inode_size(inode, offset + written)) {
 		int ret = ext4_mark_inode_dirty(handle, inode);
 		if (unlikely(ret)) {
 			ext4_journal_stop(handle);
@@ -323,21 +323,21 @@ static ssize_t ext4_handle_inode_extension(struct inode *inode, loff_t offset,
 		}
 	}
 
-	if (inode->i_nlink)
+	if ((written == count) && inode->i_nlink)
 		ext4_orphan_del(handle, inode);
 	ext4_journal_stop(handle);
 
-	return count;
+	return written;
 }
 
 /*
  * Clean up the inode after DIO or DAX extending write has completed and the
  * inode size has been updated using ext4_handle_inode_extension().
  */
-static void ext4_inode_extension_cleanup(struct inode *inode, ssize_t count)
+static void ext4_inode_extension_cleanup(struct inode *inode, bool need_trunc)
 {
 	lockdep_assert_held_write(&inode->i_rwsem);
-	if (count < 0) {
+	if (need_trunc) {
 		ext4_truncate_failed_write(inode);
 		/*
 		 * If the truncate operation failed early, then the inode may
@@ -377,7 +377,12 @@ static int ext4_dio_write_end_io(struct kiocb *iocb, ssize_t size,
 	loff_t pos = iocb->ki_pos;
 	struct inode *inode = file_inode(iocb->ki_filp);
 
-	if (!error && size && flags & IOMAP_DIO_UNWRITTEN)
+
+	if (!error && size && (flags & IOMAP_DIO_UNWRITTEN) &&
+			(iocb->ki_flags & IOCB_ATOMIC))
+		error = ext4_convert_unwritten_extents_atomic(NULL, inode, pos,
+							      size);
+	else if (!error && size && flags & IOMAP_DIO_UNWRITTEN)
 		error = ext4_convert_unwritten_extents(NULL, inode, pos, size);
 	if (error)
 		return error;
@@ -392,8 +397,9 @@ static int ext4_dio_write_end_io(struct kiocb *iocb, ssize_t size,
 	 */
 	if (pos + size <= READ_ONCE(EXT4_I(inode)->i_disksize) &&
 	    pos + size <= i_size_read(inode))
-		return size;
-	return ext4_handle_inode_extension(inode, pos, size);
+		return 0;
+	error = ext4_handle_inode_extension(inode, pos, size, size);
+	return error < 0 ? error : 0;
 }
 
 static const struct iomap_dio_ops ext4_dio_write_ops = {
@@ -564,12 +570,9 @@ static ssize_t ext4_dio_write_iter(struct kiocb *iocb, struct iov_iter *from)
 		}
 
 		ret = ext4_orphan_add(handle, inode);
-		if (ret) {
-			ext4_journal_stop(handle);
-			goto out;
-		}
-
 		ext4_journal_stop(handle);
+		if (ret)
+			goto out;
 	}
 
 	if (ilock_shared && !unwritten)
@@ -586,7 +589,7 @@ static ssize_t ext4_dio_write_iter(struct kiocb *iocb, struct iov_iter *from)
 		 * writeback of delalloc blocks.
 		 */
 		WARN_ON_ONCE(ret == -EIOCBQUEUED);
-		ext4_inode_extension_cleanup(inode, ret);
+		ext4_inode_extension_cleanup(inode, ret < 0);
 	}
 
 out:
@@ -599,6 +602,13 @@ out:
 		ssize_t err;
 		loff_t endbyte;
 
+		/*
+		 * There is no support for atomic writes on buffered-io yet,
+		 * we should never fallback to buffered-io for DIO atomic
+		 * writes.
+		 */
+		WARN_ON_ONCE(iocb->ki_flags & IOCB_ATOMIC);
+
 		offset = iocb->ki_pos;
 		err = ext4_buffered_write_iter(iocb, from);
 		if (err < 0)
@@ -669,8 +679,8 @@ ext4_dax_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	ret = dax_iomap_rw(iocb, from, &ext4_iomap_ops);
 
 	if (extend) {
-		ret = ext4_handle_inode_extension(inode, offset, ret);
-		ext4_inode_extension_cleanup(inode, ret);
+		ret = ext4_handle_inode_extension(inode, offset, ret, count);
+		ext4_inode_extension_cleanup(inode, ret < (ssize_t)count);
 	}
 out:
 	inode_unlock(inode);
@@ -683,15 +693,30 @@ out:
 static ssize_t
 ext4_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 {
+	int ret;
 	struct inode *inode = file_inode(iocb->ki_filp);
 
-	if (unlikely(ext4_forced_shutdown(inode->i_sb)))
-		return -EIO;
+	ret = ext4_emergency_state(inode->i_sb);
+	if (unlikely(ret))
+		return ret;
 
 #ifdef CONFIG_FS_DAX
 	if (IS_DAX(inode))
 		return ext4_dax_write_iter(iocb, from);
 #endif
+
+	if (iocb->ki_flags & IOCB_ATOMIC) {
+		size_t len = iov_iter_count(from);
+
+		if (len < EXT4_SB(inode->i_sb)->s_awu_min ||
+		    len > EXT4_SB(inode->i_sb)->s_awu_max)
+			return -EINVAL;
+
+		ret = generic_atomic_write_valid(iocb, from);
+		if (ret)
+			return ret;
+	}
+
 	if (iocb->ki_flags & IOCB_DIRECT)
 		return ext4_dio_write_iter(iocb, from);
 	else
@@ -781,11 +806,16 @@ static const struct vm_operations_struct ext4_file_vm_ops = {
 
 static int ext4_file_mmap(struct file *file, struct vm_area_struct *vma)
 {
+	int ret;
 	struct inode *inode = file->f_mapping->host;
 	struct dax_device *dax_dev = EXT4_SB(inode->i_sb)->s_daxdev;
 
-	if (unlikely(ext4_forced_shutdown(inode->i_sb)))
-		return -EIO;
+	if (file->f_mode & FMODE_WRITE)
+		ret = ext4_emergency_state(inode->i_sb);
+	else
+		ret = ext4_forced_shutdown(inode->i_sb) ? -EIO : 0;
+	if (unlikely(ret))
+		return ret;
 
 	/*
 	 * We don't support synchronous mappings for non-DAX files and
@@ -816,7 +846,8 @@ static int ext4_sample_last_mounted(struct super_block *sb,
 	if (likely(ext4_test_mount_flag(sb, EXT4_MF_MNTDIR_SAMPLED)))
 		return 0;
 
-	if (sb_rdonly(sb) || !sb_start_intwrite_trylock(sb))
+	if (ext4_emergency_state(sb) || sb_rdonly(sb) ||
+	    !sb_start_intwrite_trylock(sb))
 		return 0;
 
 	ext4_set_mount_flag(sb, EXT4_MF_MNTDIR_SAMPLED);
@@ -844,8 +875,7 @@ static int ext4_sample_last_mounted(struct super_block *sb,
 	if (err)
 		goto out_journal;
 	lock_buffer(sbi->s_sbh);
-	strncpy(sbi->s_es->s_last_mounted, cp,
-		sizeof(sbi->s_es->s_last_mounted));
+	strtomem_pad(sbi->s_es->s_last_mounted, cp, 0);
 	ext4_superblock_csum_set(sb);
 	unlock_buffer(sbi->s_sbh);
 	ext4_handle_dirty_metadata(handle, NULL, sbi->s_sbh);
@@ -860,8 +890,12 @@ static int ext4_file_open(struct inode *inode, struct file *filp)
 {
 	int ret;
 
-	if (unlikely(ext4_forced_shutdown(inode->i_sb)))
-		return -EIO;
+	if (filp->f_mode & FMODE_WRITE)
+		ret = ext4_emergency_state(inode->i_sb);
+	else
+		ret = ext4_forced_shutdown(inode->i_sb) ? -EIO : 0;
+	if (unlikely(ret))
+		return ret;
 
 	ret = ext4_sample_last_mounted(inode->i_sb, filp->f_path.mnt);
 	if (ret)
@@ -885,8 +919,10 @@ static int ext4_file_open(struct inode *inode, struct file *filp)
 			return ret;
 	}
 
-	filp->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC |
-			FMODE_DIO_PARALLEL_WRITE;
+	if (ext4_inode_can_atomic_write(inode))
+		filp->f_mode |= FMODE_CAN_ATOMIC_WRITE;
+
+	filp->f_mode |= FMODE_NOWAIT | FMODE_CAN_ODIRECT;
 	return dquot_file_open(inode, filp);
 }
 
@@ -898,12 +934,7 @@ static int ext4_file_open(struct inode *inode, struct file *filp)
 loff_t ext4_llseek(struct file *file, loff_t offset, int whence)
 {
 	struct inode *inode = file->f_mapping->host;
-	loff_t maxbytes;
-
-	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
-		maxbytes = EXT4_SB(inode->i_sb)->s_bitmap_maxbytes;
-	else
-		maxbytes = inode->i_sb->s_maxbytes;
+	loff_t maxbytes = ext4_get_maxbytes(inode);
 
 	switch (whence) {
 	default:
@@ -938,7 +969,6 @@ const struct file_operations ext4_file_operations = {
 	.compat_ioctl	= ext4_compat_ioctl,
 #endif
 	.mmap		= ext4_file_mmap,
-	.mmap_supported_flags = MAP_SYNC,
 	.open		= ext4_file_open,
 	.release	= ext4_release_file,
 	.fsync		= ext4_sync_file,
@@ -946,6 +976,8 @@ const struct file_operations ext4_file_operations = {
 	.splice_read	= ext4_file_splice_read,
 	.splice_write	= iter_file_splice_write,
 	.fallocate	= ext4_fallocate,
+	.fop_flags	= FOP_MMAP_SYNC | FOP_BUFFER_RASYNC |
+			  FOP_DIO_PARALLEL_WRITE,
 };
 
 const struct inode_operations ext4_file_inode_operations = {