tests/generic/459

#! /bin/bash
# SPDX-License-Identifier: GPL-2.0
# Copyright (c) 2017 Red Hat, Inc. All Rights Reserved.
#
# FS QA Test 459
#
# Test buffer filesystem error recovery during a full overcommited dm-thin device.
#
# When a dm-thin device reaches its full capacity, but the virtual device still
# shows available space, the filesystem should be able to handle such cases
# failing its operation without locking up.
#
# This test has been created first to cover a XFS problem where it loops
# indefinitely in xfsaild due items still in AIL. The buffers containing such
# items couldn't be resubmitted because the items were flush locked.
# But, once this doesn't require any special filesystem feature to be executed,
# this has been integrated as a generic test.
#
# This test might hang the filesystem when ran on an unpatched kernel
#
. ./common/preamble
_begin_fstest auto freeze thin

# Override the default cleanup function.
_cleanup()
{
	# Make sure $SCRATCH_MNT is unfreezed
	xfs_freeze -u $SCRATCH_MNT 2>/dev/null
	cd /
	rm -f $tmp.*
	$UMOUNT_PROG $SCRATCH_MNT >>$seqres.full 2>&1
	$LVM_PROG vgremove -ff $vgname >>$seqres.full 2>&1
	$LVM_PROG pvremove -ff $SCRATCH_DEV >>$seqres.full 2>&1
	$UDEV_SETTLE_PROG
}

# Import common functions.


# This tests for filesystem lockup not consistency, so don't check for fs
# consistency after test
_require_scratch_nolvm
_require_dm_target thin-pool
_require_dm_target snapshot
_require_command $LVM_PROG lvm
_require_command "$THIN_CHECK_PROG" thin_check
_require_freeze
_require_odirect

vgname=vg_$seq
lvname=lv_$seq
poolname=pool_$seq
snapname=snap_$seq
origpsize=200
virtsize=300
newpsize=300

# Check whether the filesystem has shutdown or remounted read-only. Shutdown
# behavior can differ based on filesystem and configuration. Some fs' may not
# have remounted without an additional write while others may have shutdown but
# do not necessarily reflect read-only state in the mount options. Check both
# here by first trying a simple write and following with an explicit ro check.
is_shutdown_or_ro()
{
	ro=0

	# if the fs has not shutdown, this may help trigger a remount-ro
	touch $SCRATCH_MNT/newfile > /dev/null 2>&1 || ro=1

	_fs_options /dev/mapper/$vgname-$snapname | grep -w "ro" > /dev/null
	[ $? == 0 ] && ro=1

	echo $ro
}

# Ensure we have enough disk space
_scratch_mkfs_sized $((350 * 1024 * 1024)) >>$seqres.full 2>&1

# Create a 200MB dm-thin POOL
$LVM_PROG pvcreate -f $SCRATCH_DEV >>$seqres.full 2>&1
$LVM_PROG vgcreate -f $vgname $SCRATCH_DEV >>$seqres.full 2>&1

$LVM_PROG lvcreate  --thinpool $poolname  --errorwhenfull y \
		    --zero n -L $origpsize \
		    --poolmetadatasize 4M $vgname >>$seqres.full 2>&1

# Create a overprovisioned 300MB dm-thin virt. device
$LVM_PROG lvcreate  --virtualsize $virtsize \
		    -T $vgname/$poolname \
		    -n $lvname >>$seqres.full 2>&1

$UDEV_SETTLE_PROG &>/dev/null
_mkfs_dev /dev/mapper/$vgname-$lvname >>$seqres.full 2>&1

# Running the test over the original volume doesn't reproduce the problem
# reliably, while, running it over a snapshot, makes the problem 100%
# reproducible, so, create a snapshot and run the test over it.
$LVM_PROG lvcreate  -k n -s $vgname/$lvname \
		    -n $snapname >>$seqres.full 2>&1

# Catch mount failure so we don't blindly go an freeze the root filesystem
# instead of lvm volume.
_mount /dev/mapper/$vgname-$snapname $SCRATCH_MNT || _fail "mount failed"

# Consume all space available in the volume and freeze to ensure everything
# required to make the fs consistent is flushed to disk.
$XFS_IO_PROG -f -d -c 'pwrite -b 1m 0 220m' $SCRATCH_MNT/f1 >>$seqres.full 2>&1

# In XFS, this freeze will never complete until the dm-thin POOL device is
# extended. It is expected, and is only used so xfsaild is triggered to
# flush AIL items, other filesystems usually get remounted as read-only during
# the above write process.
xfs_freeze -f $SCRATCH_MNT >>$seqres.full 2>&1 &
freezeid=$!

# Wait enough so xfsaild can run
sleep 10

# Make some extra space available so the freeze above can proceed
$LVM_PROG lvextend -L $newpsize $vgname/$poolname >>$seqres.full 2>&1

wait $freezeid
ret=$?

# Different filesystems will handle the lack of real space in different ways,
# some will remount the filesystem in read-only mode, some will not. These tests
# will check if:
#	- The filesystem turns into Read-Only and reject further writes
#	- The filesystem stays in Read-Write mode, but can be frozen/thawed
#	  without getting stuck.
if [ $ret -ne 0 ]; then
	# freeze failed, filesystem should reject further writes
	ISRO=`is_shutdown_or_ro`
	if [ $ISRO == 1 ]; then
		echo "Test OK"
	else
		echo "Freeze failed and FS isn't Read-Only. Test Failed"
		status=1
		exit
	fi
else
	# Try to thaw the filesystem, and complete test if if succeed.
	# NOTE: This will hang on affected XFS filesystems.
	xfs_freeze -u $SCRATCH_MNT >>$seqres.full 2>&1
	echo "Test OK"
fi

status=0
exit