[PATCH v2 1/2] fstests: introduce _require_no_mount_opts helper

[PATCH v2 1/2] fstests: introduce _require_no_mount_opts helper

[Eryu Guan]

Some tests require that there's no certain mount option, so
introduce a new helper _require_no_mount_opts() to do the check on
$MOUNT_OPTIONS.

Also convert generic/192 and xfs/134 to use this helper.

Signed-off-by: Eryu Guan <eguan@redhat.com>
---
v2:
- rebase on top of current master

 common/rc         |  8 ++++++++
 tests/generic/192 | 11 -----------
 tests/xfs/134     |  8 +++-----
 3 files changed, 11 insertions(+), 16 deletions(-)

diff --git a/common/rc b/common/rc
index 3a9c4d1..680ea2b 100644
--- a/common/rc
+++ b/common/rc
@@ -2926,8 +2926,16 @@ _require_cloner()
 		_notrun "cloner binary not present at $CLONER_PROG"
 }
 
+_require_no_mount_opts()
+{
+	if echo $MOUNT_OPTIONS | grep -q "$1"; then
+		_notrun "mount option \"$1\" not allowed in this test"
+	fi
+}
+
 _require_atime()
 {
+	_require_no_mount_opts "noatime"
 	if [ "$FSTYP" == "nfs" ]; then
 		_notrun "atime related mount options have no effect on NFS"
 	fi
diff --git a/tests/generic/192 b/tests/generic/192
index 6bbc87c..d3ed0ec 100755
--- a/tests/generic/192
+++ b/tests/generic/192
@@ -38,13 +38,6 @@ _access_time()
 	stat -c %X $1
 }
 
-is_noatime_set() {
-	case "$MOUNT_OPTIONS" in
-		*noatime*) return 0;;
-	esac
-	return 1
-}
-
 # get standard environment, filters and checks
 . ./common/rc
 . ./common/filter
@@ -61,10 +54,6 @@ _require_atime
 #delay=45
 delay=40
 
-if is_noatime_set; then
-	_notrun "this test makes no sense with noatime"
-fi
-
 testfile=$TEST_DIR/testfile
 rm -f $testfile
 rm -f $seqres.full
diff --git a/tests/xfs/134 b/tests/xfs/134
index cd47069..a7df9fe 100755
--- a/tests/xfs/134
+++ b/tests/xfs/134
@@ -50,6 +50,9 @@ _supported_os Linux IRIX
 
 _require_test
 _require_xfs_quota
+# we can't run with group quotas
+_require_no_mount_opts "gquota"
+_require_no_mount_opts "grpquota"
 
 dir=$SCRATCH_MNT/project
 
@@ -70,11 +73,6 @@ _scratch_mkfs_xfs >/dev/null 2>&1
 
 #if pquota's already in mount options then we dont need to enable
 
-# we can't run with group quotas
-if ( `echo $MOUNT_OPTIONS | grep -q gquota` || `echo $MOUNT_OPTIONS | grep -q grpquota` )
-then
-    _notrun "Can't run with group quotas enabled"
-fi
 EXTRA_MOUNT_OPTIONS="-o pquota"
 
 if ! _scratch_mount "$EXTRA_MOUNT_OPTIONS" >$tmp.out 2>&1
-- 
2.5.5

[PATCH v2 2/2] ext4/271: _notrun if there are journal related mount options

[Eryu Guan]

ext4/271 runs in no journal mode (-onoload), so running test with
journal related mount options makes no sense, and test fails after
kernel commit 1e381f60dad9 ("ext4: do not allow journal_opts for fs
w/o journal"), journal related mount options are not allowed in no
journal mode. So _notrun if there're journal related mount options.

Signed-off-by: Eryu Guan <eguan@redhat.com>
---
v2:
- update comments in the code
- rebase on top of current master

 tests/ext4/271 | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/tests/ext4/271 b/tests/ext4/271
index d68c271..ca344d9 100755
--- a/tests/ext4/271
+++ b/tests/ext4/271
@@ -39,6 +39,12 @@ trap "rm -f $tmp.*; exit \$status" 0 1 2 3 15
 _supported_fs ext4
 _supported_os Linux
 _require_scratch
+# this test needs no journal to be loaded, skip on journal related mount
+# options, otherwise mount would fail with "-o noload" mount option
+_require_no_mount_opts "data="
+_require_no_mount_opts "commit="
+_require_no_mount_opts "journal_checksum"
+_require_no_mount_opts "journal_async_commit"
 
 rm -f $seqres.full
 _scratch_mkfs_sized $((128 * 1024 * 1024)) >> $seqres.full 2>&1
-- 
2.5.5

Re: [PATCH v2 1/2] fstests: introduce _require_no_mount_opts helper

[Dave Chinner]

On Tue, Jun 21, 2016 at 06:40:28PM +0800, Eryu Guan wrote:
> Some tests require that there's no certain mount option, so
> introduce a new helper _require_no_mount_opts() to do the check on
> $MOUNT_OPTIONS.

I think this is fine, except for the name. It's more of an exclude
rule rather than a "require" rule. i.e.  _exclude_mount_option() is
closer to it's purpose.

The only other question I have is that mount options can be
different between test and scratch devices - the test device mount
options can be set via TEST_FS_MOUNT_OPTS, as well as via
MOUNT_OPTIONS. Does this rule need to handle that?

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

Re: [PATCH v2 1/2] fstests: introduce _require_no_mount_opts helper

[Eryu Guan]

On Wed, Jun 22, 2016 at 09:42:02AM +1000, Dave Chinner wrote:
> On Tue, Jun 21, 2016 at 06:40:28PM +0800, Eryu Guan wrote:
> > Some tests require that there's no certain mount option, so
> > introduce a new helper _require_no_mount_opts() to do the check on
> > $MOUNT_OPTIONS.
> 
> I think this is fine, except for the name. It's more of an exclude
> rule rather than a "require" rule. i.e.  _exclude_mount_option() is
> closer to it's purpose.

This does look better to me, thanks!

> 
> The only other question I have is that mount options can be
> different between test and scratch devices - the test device mount
> options can be set via TEST_FS_MOUNT_OPTS, as well as via
> MOUNT_OPTIONS. Does this rule need to handle that?

I didn't think about TEST_FS_MOUNT_OPTS. Currently there's no need to
handle it, MOUNT_OPTIONS is sufficient I think.

How about I rename it to _exclude_scratch_mount_option()? And we can
always add another _exclude_test_mount_option() if needed in future.

Thanks,
Eryu

Re: [PATCH v2 1/2] fstests: introduce _require_no_mount_opts helper

[Dave Chinner]

On Wed, Jun 22, 2016 at 11:23:51AM +0800, Eryu Guan wrote:
> On Wed, Jun 22, 2016 at 09:42:02AM +1000, Dave Chinner wrote:
> > On Tue, Jun 21, 2016 at 06:40:28PM +0800, Eryu Guan wrote:
> > > Some tests require that there's no certain mount option, so
> > > introduce a new helper _require_no_mount_opts() to do the check on
> > > $MOUNT_OPTIONS.
> > 
> > I think this is fine, except for the name. It's more of an exclude
> > rule rather than a "require" rule. i.e.  _exclude_mount_option() is
> > closer to it's purpose.
> 
> This does look better to me, thanks!
> 
> > 
> > The only other question I have is that mount options can be
> > different between test and scratch devices - the test device mount
> > options can be set via TEST_FS_MOUNT_OPTS, as well as via
> > MOUNT_OPTIONS. Does this rule need to handle that?
> 
> I didn't think about TEST_FS_MOUNT_OPTS. Currently there's no need to
> handle it, MOUNT_OPTIONS is sufficient I think.
> 
> How about I rename it to _exclude_scratch_mount_option()? And we can
> always add another _exclude_test_mount_option() if needed in future.

Sounds good.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

Re: [PATCH v2 1/2] fstests: introduce _require_no_mount_opts helper

[Eryu Guan]

On Thu, Jun 23, 2016 at 09:06:50AM +1000, Dave Chinner wrote:
> On Wed, Jun 22, 2016 at 11:23:51AM +0800, Eryu Guan wrote:
> > On Wed, Jun 22, 2016 at 09:42:02AM +1000, Dave Chinner wrote:
> > > On Tue, Jun 21, 2016 at 06:40:28PM +0800, Eryu Guan wrote:
> > > > Some tests require that there's no certain mount option, so
> > > > introduce a new helper _require_no_mount_opts() to do the check on
> > > > $MOUNT_OPTIONS.
> > > 
> > > I think this is fine, except for the name. It's more of an exclude
> > > rule rather than a "require" rule. i.e.  _exclude_mount_option() is
> > > closer to it's purpose.
> > 
> > This does look better to me, thanks!
> > 
> > > 
> > > The only other question I have is that mount options can be
> > > different between test and scratch devices - the test device mount
> > > options can be set via TEST_FS_MOUNT_OPTS, as well as via
> > > MOUNT_OPTIONS. Does this rule need to handle that?
> > 
> > I didn't think about TEST_FS_MOUNT_OPTS. Currently there's no need to
> > handle it, MOUNT_OPTIONS is sufficient I think.
> > 
> > How about I rename it to _exclude_scratch_mount_option()? And we can
> > always add another _exclude_test_mount_option() if needed in future.
> 
> Sounds good.

Hi Dave,

My v4 patches simply rename the helper to this
"_exclude_scratch_mount_option", no function change. Then I take your
reviews as a "Reviewed-by" tag based on "I think this is fine, except for
the name." and the above "Sounds good.". So I can queue them for next
pull request and start release testing, and don't have to bother you
providing an explicit reviewed-by.

Please let me know if you have different thoughts.

Thanks,
Eryu

Re: [PATCH v2 1/2] fstests: introduce _require_no_mount_opts helper

[Dave Chinner]

On Fri, Jun 24, 2016 at 12:50:19PM +0800, Eryu Guan wrote:
> On Thu, Jun 23, 2016 at 09:06:50AM +1000, Dave Chinner wrote:
> > On Wed, Jun 22, 2016 at 11:23:51AM +0800, Eryu Guan wrote:
> > > On Wed, Jun 22, 2016 at 09:42:02AM +1000, Dave Chinner wrote:
> > > > On Tue, Jun 21, 2016 at 06:40:28PM +0800, Eryu Guan wrote:
> > > > > Some tests require that there's no certain mount option, so
> > > > > introduce a new helper _require_no_mount_opts() to do the check on
> > > > > $MOUNT_OPTIONS.
> > > > 
> > > > I think this is fine, except for the name. It's more of an exclude
> > > > rule rather than a "require" rule. i.e.  _exclude_mount_option() is
> > > > closer to it's purpose.
> > > 
> > > This does look better to me, thanks!
> > > 
> > > > 
> > > > The only other question I have is that mount options can be
> > > > different between test and scratch devices - the test device mount
> > > > options can be set via TEST_FS_MOUNT_OPTS, as well as via
> > > > MOUNT_OPTIONS. Does this rule need to handle that?
> > > 
> > > I didn't think about TEST_FS_MOUNT_OPTS. Currently there's no need to
> > > handle it, MOUNT_OPTIONS is sufficient I think.
> > > 
> > > How about I rename it to _exclude_scratch_mount_option()? And we can
> > > always add another _exclude_test_mount_option() if needed in future.
> > 
> > Sounds good.
> 
> Hi Dave,
> 
> My v4 patches simply rename the helper to this
> "_exclude_scratch_mount_option", no function change. Then I take your
> reviews as a "Reviewed-by" tag based on "I think this is fine, except for
> the name." and the above "Sounds good.". So I can queue them for next
> pull request and start release testing, and don't have to bother you
> providing an explicit reviewed-by.

Yup - it's better to ask thn assume, though, so:

Reviewed-by: Dave Chinner <dchinner@redhat.com>

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com