Re: [PATCH v7 16/18] seltests/landlock: add invalid input data test

["Konstantin Meskhidze (A)" <konstantin.meskhidze@huawei.com>]

10/10/2022 1:37 PM, Mickaël Salaün пишет:
> 
> On 12/09/2022 19:22, Mickaël Salaün wrote:
>> 
>> On 10/09/2022 22:51, Konstantin Meskhidze (A) wrote:
>>>
>>>
>>> 9/6/2022 11:09 AM, Mickaël Salaün пишет:
>>>>
>>>> On 29/08/2022 19:03, Konstantin Meskhidze wrote:
>>>>> This patch adds rules with invalid user space supplied data:
>>>>>        - out of range ruleset attribute;
>>>>>        - unhandled allowed access;
>>>>>        - zero port value;
>>>>>        - zero access value;
>>>>>
>>>>> Signed-off-by: Konstantin Meskhidze <konstantin.meskhidze@huawei.com>
>>>>> ---
>>>>>
>>>>> Changes since v6:
>>>>> * Adds invalid ruleset attribute test.
>>>>>
>>>>> Changes since v5:
>>>>> * Formats code with clang-format-14.
>>>>>
>>>>> Changes since v4:
>>>>> * Refactors code with self->port variable.
>>>>>
>>>>> Changes since v3:
>>>>> * Adds inval test.
>>>>>
>>>>> ---
>>>>>     tools/testing/selftests/landlock/net_test.c | 66 ++++++++++++++++++++-
>>>>>     1 file changed, 65 insertions(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/tools/testing/selftests/landlock/net_test.c b/tools/testing/selftests/landlock/net_test.c
>>>>> index a93224d1521b..067ba45f58a5 100644
>>>>> --- a/tools/testing/selftests/landlock/net_test.c
>>>>> +++ b/tools/testing/selftests/landlock/net_test.c
>>>>> @@ -26,9 +26,12 @@
>>>>>
>>>>>     #define IP_ADDRESS "127.0.0.1"
>>>>>
>>>>> -/* Number pending connections queue to be hold */
>>>>> +/* Number pending connections queue to be hold. */
>>>>
>>>> Patch of a previous patch?
>>>>
>>>>
>>>>>     #define BACKLOG 10
>>>>>
>>>>> +/* Invalid attribute, out of landlock network access range. */
>>>>> +#define LANDLOCK_INVAL_ATTR 7
>>>>> +
>>>>>     FIXTURE(socket)
>>>>>     {
>>>>>     	uint port[MAX_SOCKET_NUM];
>>>>> @@ -719,4 +722,65 @@ TEST_F(socket, ruleset_expanding)
>>>>>     	/* Closes socket 1. */
>>>>>     	ASSERT_EQ(0, close(sockfd_1));
>>>>>     }
>>>>> +
>>>>> +TEST_F(socket, inval)
>>>>> +{
>>>>> +	struct landlock_ruleset_attr ruleset_attr = {
>>>>> +		.handled_access_net = LANDLOCK_ACCESS_NET_BIND_TCP
>>>>> +	};
>>>>> +	struct landlock_ruleset_attr ruleset_attr_inval = {
>>>>> +		.handled_access_net = LANDLOCK_INVAL_ATTR
>>>>
>>>> Please add a test similar to TEST_F_FORK(layout1,
>>>> file_and_dir_access_rights) instead of explicitly defining and only
>>>> testing LANDLOCK_INVAL_ATTR.
>>>>
>>>      Do you want fs test to be in this commit or maybe its better to add
>>> it into "[PATCH v7 01/18] landlock: rename access mask" one.
> 
> Just to make it clear, I didn't suggested an FS test, but a new network
> test similar to layout1.file_and_dir_access_rights but only related to
> the network. It should replace/extend the content of this patch (16/18).
> 
  Ok. I will check out out "layout1.file_and_dir_access_rights" one.
But anyway we need some test like TEST_F_FORK(layout1, with_net) and
TEST_F_FORK(socket, with_fs) with mixed attributes as you suggested.

> 
>> 
>> You can squash all the new tests patches (except the "move helper
>> function").
> You should move most of your patch descriptions in a comment above the
> related tests. The commit message should list all the new tests and
> quickly explain which part of the kernel is covered (i.e. mostly the TCP
> part of Landlock). You can get some inspiration from
> https://git.kernel.org/mic/c/f4056b9266b571c63f30cda70c2d89f7b7e8bb7b
> 
> You need to rebase on top of my next branch (from today).

  Ok. Thank you. Sorry for the delay - I was under the snow with 
Business Trip to China preparings. So I've got out of 10 days quarantine 
now and continued working.
> .