linux-kernel - [PATCH] loop/010: Add test for mode 0 fallocate() on loop devices

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20231110010139.3901150-5-sarthakkukreti@chromium.org>
Date:   Thu,  9 Nov 2023 17:01:39 -0800
From:   Sarthak Kukreti <sarthakkukreti@...omium.org>
To:     dm-devel@...ts.linux.dev, linux-block@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org
Cc:     Jens Axboe <axboe@...nel.dk>, Mike Snitzer <snitzer@...nel.org>,
        "Darrick J . Wong" <djwong@...nel.org>,
        Christoph Hellwig <hch@....de>,
        Dave Chinner <david@...morbit.com>,
        Brian Foster <bfoster@...hat.com>,
        Sarthak Kukreti <sarthakkukreti@...omium.org>
Subject: [PATCH] loop/010: Add test for mode 0 fallocate() on loop devices

A recent patch series[1] adds support for calling fallocate() in mode 0
on block devices. This test adds a basic sanity test for loopback devices
setup on a sparse file and validates that writes to the loopback device
succeed, even when the underlying filesystem runs out of space.

Signed-off-by: Sarthak Kukreti <sarthakkukreti@...omium.org>
---
 tests/loop/010     | 60 ++++++++++++++++++++++++++++++++++++++++++++++
 tests/loop/010.out |  2 ++
 2 files changed, 62 insertions(+)
 create mode 100644 tests/loop/010
 create mode 100644 tests/loop/010.out

diff --git a/tests/loop/010 b/tests/loop/010
new file mode 100644
index 0000000..091be5e
--- /dev/null
+++ b/tests/loop/010
@@ -0,0 +1,60 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-3.0+
+# Copyright (C) 2023 Google LLC.
+# Author: sarthakkukret@...gle.com (Sarthak Kukreti)
+#
+# Test if fallocate() on a loopback device provisions space on the underlying
+# filesystem and writes on the loop device succeed, even if the lower
+# filesystem is filled up.
+
+. tests/loop/rc
+
+DESCRIPTION="Loop device fallocate() space provisioning"
+QUICK=1
+
+requires() {
+	_have_program mkfs.ext4
+}
+
+test() {
+	echo "Running ${TEST_NAME}"
+
+	local mount_dir="$TMPDIR/mnt"
+	mkdir -p ${mount_dir}
+
+	local image_file="$TMPDIR/img"
+	truncate -s 1G "${image_file}"
+
+	local loop_device
+	loop_device="$(losetup -P -f --show "${image_file}")"
+
+	mkfs.ext4 ${loop_device} &> /dev/null
+	mount -t ext4 ${loop_device} ${mount_dir}
+
+	local provisioned_file="${mount_dir}/provisioned"
+	truncate -s 200M "${provisioned_file}"
+
+	local provisioned_loop_device
+	provisioned_loop_device="$(losetup -P -f --show "${provisioned_file}")"
+
+	# Provision space for the file: without provisioning support, this fails
+	# with EOPNOTSUPP.
+	fallocate -l 200M "${provisioned_loop_device}"
+
+	# Fill the filesystem, this command will error out with ENOSPC.
+	local fs_fill_file="${mount_dir}/fill"
+	dd if=/dev/zero of="${fs_fill_file}" bs=1M count=1024 status=none &>/dev/null
+	sync
+
+	# Write to provisioned loop device, ensure that it does not run into ENOSPC.
+	dd if=/dev/zero of="${provisioned_loop_device}" bs=1M count=200 status=none
+	sync
+
+	# Cleanup.
+	losetup --detach "${provisioned_loop_device}"
+	umount "${mount_dir}"
+	losetup --detach "${loop_device}"
+	rm "${image_file}"
+
+	echo "Test complete"
+}
\ No newline at end of file
diff --git a/tests/loop/010.out b/tests/loop/010.out
new file mode 100644
index 0000000..068c489
--- /dev/null
+++ b/tests/loop/010.out
@@ -0,0 +1,2 @@
+Running loop/009
+Test complete
-- 
2.42.0.758.gaed0368e0e-goog