On Wed, Dec 6, 2017 at 2:59 AM, Chengguang Xu <cgxu519@xxxxxxxxxx> wrote:
>
>> 在 2017年12月6日，上午1:14，Eryu Guan <eguan@xxxxxxxxxx> 写道：
>>
>> On Tue, Dec 05, 2017 at 06:13:35PM +0200, Amir Goldstein wrote:
>>> On Tue, Dec 5, 2017 at 5:50 PM, Chengguang Xu <cgxu519@xxxxxxxxxx> wrote:
>>>> 1. Add a check case in _require_xfs_io_command() to support syncfs.
>>>> 2. Introduce a helper to support scratch shutdown for overlayfs.
>>>>
>>>> Signed-off-by: Chengguang Xu <cgxu519@xxxxxxxxxx>
>>>> ---
>>>> common/rc | 22 ++++++++++++++++++++--
>>>> 1 file changed, 20 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/common/rc b/common/rc
>>>> index 4c053a5..e36ee24 100644
>>>> --- a/common/rc
>>>> +++ b/common/rc
>>>> @@ -669,6 +669,21 @@ _scratch_cleanup_files()
>>>> esac
>>>> }
>>>>
>>>> +_scratch_shutdown()
>>>> +{
>>>> +
>>>> + case $FSTYP in
>>>> + overlay)
>>>
>>> Looks good,
>>> but first you need to check for [ -z "$OVL_BASE_SCRATCH_MNT" ]
>>> meaning that tester is using "old" overlay config and we are not allowed to
>>> mess with base fs.
>>
>> Agreed, we don't want to shutdown the root fs accidently.
>
> If "$OVL_BASE_SCRATCH_MNT” is nothing, godown command should fail for improper param, right?
> I don’t clearly know how it makes rootfs shutdown.
> you mean I need to check for [ "$OVL_BASE_SCRATCH_MNT” = “/“ ] ?
>
Sorry, I meant you need to check if OVL_BASE_SCRATCH_DEV is not empty.
OVL_BASE_SCRATCH_MNT will always be valid dir, but either user set it
with old configuration, e.g.:
SCRATCH_DEV=/my/ovl/root
FSTYP=overlay
OR xfstests mounted a base fs with new overlay configuration, e.g.:
SCRATCH_DEV=/dev/my-ovl-bdev
SCRATCH_MNT=/my/ovl/root
FSTYP=xfs
If you shutdown OVL_BASE_SCRATCH_MNT in the first case, you may
shutdown the fs of the host running the test.
Amir.
--
To unsubscribe from this list: send the line "unsubscribe linux-unionfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
