fuse: truncate pagecache on atomic_o_trunc

commit 7622435 upstream.

fuse_finish_open() will be called with FUSE_NOWRITE in case of atomic
O_TRUNC.  This can deadlock with fuse_wait_on_page_writeback() in
fuse_launder_page() triggered by invalidate_inode_pages2().

Fix by replacing invalidate_inode_pages2() in fuse_finish_open() with a
truncate_pagecache() call.  This makes sense regardless of FOPEN_KEEP_CACHE
or fc->writeback cache, so do it unconditionally.

Reported-by: Xie Yongji <xieyongji@bytedance.com>
Reported-and-tested-by: syzbot+bea44a5189836d956894@syzkaller.appspotmail.com
Fixes: e464830 ("fuse: truncate pending writes on O_TRUNC")
Cc: <stable@vger.kernel.org>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

fs/fuse/file.c

@@ -198,23 +198,26 @@ void fuse_finish_open(struct inode *inode, struct file *file)
 	struct fuse_file *ff = file->private_data;
 	struct fuse_conn *fc = get_fuse_conn(inode);
 
-	if (!(ff->open_flags & FOPEN_KEEP_CACHE))
-		invalidate_inode_pages2(inode->i_mapping);
 	if (ff->open_flags & FOPEN_STREAM)
 		stream_open(inode, file);
 	else if (ff->open_flags & FOPEN_NONSEEKABLE)
 		nonseekable_open(inode, file);
+
 	if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) {
 		struct fuse_inode *fi = get_fuse_inode(inode);
 
 		spin_lock(&fi->lock);
 		fi->attr_version = atomic64_inc_return(&fc->attr_version);
 		i_size_write(inode, 0);
 		spin_unlock(&fi->lock);
+		truncate_pagecache(inode, 0);
 		fuse_invalidate_attr(inode);
 		if (fc->writeback_cache)
 			file_update_time(file);
+	} else if (!(ff->open_flags & FOPEN_KEEP_CACHE)) {
+		invalidate_inode_pages2(inode->i_mapping);
 	}
+
 	if ((file->f_mode & FMODE_WRITE) && fc->writeback_cache)
 		fuse_link_write_file(file);
 }