Re: [PATCH v4] tests/generic: check log recovery with readonly mount

[Zorro Lang <zlang@redhat.com>]

On Tue, Aug 24, 2021 at 05:06:38PM +0800, Murphy Zhou wrote:
> And followed by a rw mount. After log recovery by these 2 mount, the
> filesystem should be in a consistent state.
> 
> Suggested-by:  Donald Douwsma <ddouwsma@redhat.com>
> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
> Signed-off-by: Murphy Zhou <xzhou@redhat.com>
> ---
> 
> Thanks Eryu and Zorro!
> 
> v2:
>    Add explanation of the issue
>    add xfs_force_bdev data $SCRATCH_MNT
>    use DF_PROG
>    Re numbered this test
> v3:
>    Add _require_scratch_shutdown
>    Use _get_available_space
>    Explain why does not use _scratch_mount
> v4:
>    Add to recoveryloop group
>    Move to generic as there are no xfs specific operations
>    Remove all operations after 2 cycle mounts, let the fsck in fstests
> to check the filesystem consistency
>    Use _scratch_shutdown, MOUNT_PROG
>    Remove unnecessary comments
> 
>  tests/generic/999     | 45 +++++++++++++++++++++++++++++++++++++++++++
>  tests/generic/999.out |  2 ++
>  2 files changed, 47 insertions(+)
>  create mode 100755 tests/generic/999
>  create mode 100644 tests/generic/999.out
> 
> diff --git a/tests/generic/999 b/tests/generic/999
> new file mode 100755
> index 00000000..9685488b
> --- /dev/null
> +++ b/tests/generic/999
> @@ -0,0 +1,45 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2021 RedHat All Rights Reserved.
> +#
> +# FS QA Test 999
> +#
> +# Testcase for kernel commit:
> +#   50d25484bebe xfs: sync lazy sb accounting on quiesce of read-only mounts
> +#
> +# After shutdown and readonly mount, a following read-write mount would
> +# get wrong number of available blocks. This is caused by unmounting the log
> +# on a readonly filesystem doesn't log the sb counters.
> +#
> +. ./common/preamble
> +_begin_fstest auto quick recoveryloop shutdown
> +
> +# real QA test starts here
> +
> +_require_scratch
> +_require_scratch_shutdown
> +_scratch_mkfs > $seqres.full 2>&1
> +
> +# Don't use _scratch_mount because we need to mount without SELinux context
> +# to reproduce this issue. If we mount with _scratch_mount, SELinux context
> +# maybe applied and this testcase is not reproducing the original issue.
> +$MOUNT_PROG $SCRATCH_DEV $SCRATCH_MNT

This's weird, I think the SELinux mount option only can make below testfile
have no selinux label (xattr). But I don't think it's related with this bug.
So I tried to use "_scratch_mount", then still can reproduce this bug. May
you try it again?

# git diff
diff --git a/tests/generic/999 b/tests/generic/999
index 9685488b..a196e65d 100755
--- a/tests/generic/999
+++ b/tests/generic/999
@@ -23,8 +23,11 @@ _scratch_mkfs > $seqres.full 2>&1
 # Don't use _scratch_mount because we need to mount without SELinux context
 # to reproduce this issue. If we mount with _scratch_mount, SELinux context
 # maybe applied and this testcase is not reproducing the original issue.
-$MOUNT_PROG $SCRATCH_DEV $SCRATCH_MNT
-_xfs_force_bdev data $SCRATCH_MNT
+#$MOUNT_PROG $SCRATCH_DEV $SCRATCH_MNT
+_scratch_mount
+if [ "$FSTYP" = "xfs" ];then
+       _xfs_force_bdev data $SCRATCH_MNT
+fi

# ./check generic/999
FSTYP         -- xfs (non-debug)
PLATFORM      -- Linux/x86_64 hp-dl380pg8-01 4.18.0-xxx.el8.x86_64+debug #1 SMP Tue Jun 15 10:43:04 EDT 2021
MKFS_OPTIONS  -- -f /dev/mapper/rhel_hp--dl380pg8--01-xfscratch
MOUNT_OPTIONS -- -o context=system_u:object_r:root_t:s0 /dev/mapper/rhel_hp--dl380pg8--01-xfscratch /mnt/scratch

generic/999 16s ... _check_xfs_filesystem: filesystem on /dev/mapper/rhel_hp--dl380pg8--01-xfscratch is inconsistent (r)
(see /home/xfstests-dev/results//generic/999.full for details)

Ran: generic/999
Failures: generic/999
Failed 1 of 1 tests

> +_xfs_force_bdev data $SCRATCH_MNT

This's a xfs only helper, for a generic case, better to:

if [ "$FSTYP" = "xfs" ];then
	_xfs_force_bdev data $SCRATCH_MNT
fi

> +
> +echo Testing > $SCRATCH_MNT/testfile
> +
> +# -f is required to reproduce
> +_scratch_shutdown -f
> +
> +_scratch_cycle_mount ro
> +_scratch_cycle_mount
> +
> +# These 2 mount should have the log fully recovered. Exit here and let the
> +# fsck operation to check the consistence of the tested filesystem. On the
> +# buggy kernel, this testcase reports filesystem is in inconsistent state.
> +# On the fixed kernel, testcase pass.
> +
> +# success, all done
> +echo "Silence is golden"
> +status=0
> +exit
> diff --git a/tests/generic/999.out b/tests/generic/999.out
> new file mode 100644
> index 00000000..3b276ca8
> --- /dev/null
> +++ b/tests/generic/999.out
> @@ -0,0 +1,2 @@
> +QA output created by 999
> +Silence is golden
> -- 
> 2.20.1
>