Re: [PATCH] xfs: test that default grace periods init on first mount

From: Zorro Lang <zlang@redhat.com>
To: Eric Sandeen <sandeen@redhat.com>
Cc: fstests <fstests@vger.kernel.org>, linux-xfs <linux-xfs@vger.kernel.org>
Subject: Re: [PATCH] xfs: test that default grace periods init on first mount
Date: Thu, 20 Feb 2020 12:50:50 +0800	[thread overview]
Message-ID: <20200220045050.GF14282@dhcp-12-102.nay.redhat.com> (raw)
In-Reply-To: <7c6b4646-d7c5-cc03-9c90-c17daa22071d@redhat.com>

On Wed, Feb 19, 2020 at 09:26:29PM -0600, Eric Sandeen wrote:
> There's currently a bug in how default grace periods get set up
> before the very first quotacheck runs; we try to read the quota
> inodes before they are populated, and so the grace periods remain
> empty.  The /next/ mount fills them in.  This is a regression test
> for that bug.
> 
> Signed-off-by: Eric Sandeen <sandeen@redhat.com>
> ---
> 
> diff --git a/tests/xfs/995 b/tests/xfs/995
> new file mode 100755
> index 00000000..477855b8
> --- /dev/null
> +++ b/tests/xfs/995
> @@ -0,0 +1,50 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2020 Red Hat, Inc.  All Rights Reserved.
> +#
> +# FS QA Test 995
> +#
> +# Regression test xfs flaw which did not report proper quota default
> +# grace periods until the 2nd mount of a new filesystem with quota.
> +#
> +seq=`basename $0`
> +seqres=$RESULT_DIR/$seq
> +echo "QA output created by $seq"
> +
> +here=`pwd`
> +tmp=/tmp/$$
> +status=1	# failure is the default!
> +trap "_cleanup; exit \$status" 0 1 2 3 15
> +
> +_cleanup()
> +{
> +	cd /
> +	rm -f $tmp.*
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/quota
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +_supported_fs generic
> +_supported_os Linux
> +_require_scratch
> +_require_quota

If this's xfs specified case, I think "_require_xfs_quota" might be better then
_require_quota, at least it checks "XFS_QUOTA_PROG" and some xfs specified
things.

BTW, do you think "_require_xfs_quota_foreign" can help this case to be a
generic case?

> +
> +_scratch_mkfs >$seqres.full 2>&1
> +_qmount_option "usrquota"
> +_qmount
> +
> +xfs_quota -x -c "state -u" $SCRATCH_MNT | grep "grace time"

I think $XFS_QUOTA_PROG would be better than using xfs_quota directly.

> +_scratch_unmount
> +_qmount
> +xfs_quota -x -c "state -u" $SCRATCH_MNT | grep "grace time"
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/xfs/995.out b/tests/xfs/995.out
> new file mode 100644
> index 00000000..d10017d1
> --- /dev/null
> +++ b/tests/xfs/995.out
> @@ -0,0 +1,7 @@
> +QA output created by 995
> +Blocks grace time: [7 days]
> +Inodes grace time: [7 days]
> +Realtime Blocks grace time: [7 days]
> +Blocks grace time: [7 days]
> +Inodes grace time: [7 days]
> +Realtime Blocks grace time: [7 days]

Hmm... but if the bug is on the default grace time itself, and different
filesystems have different default timers, then I think a xfs specified case
would be better. I don't have a better idea to make it a generic case.

The case from me:
  [PATCH 1/2] generic: per-type quota timers set/get test
Although it trys to test default quota time too, but I think it can't verify
this bug properly, especially if the default time isn't correctly at beginning.
Feel free to correct me, if you have a better idea:)

Thanks,
Zorro

> diff --git a/tests/xfs/group b/tests/xfs/group
> index 0cbd0647..235a2715 100644
> --- a/tests/xfs/group
> +++ b/tests/xfs/group
> @@ -511,3 +511,4 @@
>  511 auto quick quota
>  512 auto quick acl attr
>  513 auto mount
> +995 auto quota quick
> 