Re: [PATCH 1/4] KVM: s390: selftests: Use TAP interface in the memop test

[Janis Schoetterl-Glausch <scgl@linux.ibm.com>]

On 4/19/22 19:40, Thomas Huth wrote:
> On 14/04/2022 14.48, Janis Schoetterl-Glausch wrote:
>> On 4/14/22 12:53, Thomas Huth wrote:
>>> The memop test currently does not have any output (unless one of the
>>> TEST_ASSERT statement fails), so it's hard to say for a user whether
>>> a certain new sub-test has been included in the binary or not. Let's
>>> make this a little bit more user-friendly and include some TAP output
>>> via the kselftests.h interface.
>>>
>>> Signed-off-by: Thomas Huth <thuth@redhat.com>
>>> ---
>>>   tools/testing/selftests/kvm/s390x/memop.c | 90 ++++++++++++++++++-----
>>>   1 file changed, 73 insertions(+), 17 deletions(-)
>>>
>>> diff --git a/tools/testing/selftests/kvm/s390x/memop.c b/tools/testing/selftests/kvm/s390x/memop.c
>>> index b04c2c1b3c30..a2783d9afcac 100644
>>> --- a/tools/testing/selftests/kvm/s390x/memop.c
>>> +++ b/tools/testing/selftests/kvm/s390x/memop.c
>>> @@ -12,6 +12,7 @@
>>>     #include "test_util.h"
>>>   #include "kvm_util.h"
>>> +#include "kselftest.h"
>>>     enum mop_target {
>>>       LOGICAL,
>>> @@ -648,33 +649,88 @@ static void test_errors(void)
>>>       kvm_vm_free(t.kvm_vm);
>>>   }
>>>   +struct testdef {
>>> +    const char *name;
>>> +    void (*test)(void);
>>> +    bool needs_extension;
>>
>> Please make this numeric. You could also rename it to required_extension or similar.
> [...]
>>> +
>>> +    for (idx = 0; idx < ARRAY_SIZE(testlist); idx++) {
>>> +        if (!testlist[idx].needs_extension || extension_cap) {
>>
>> Then check here that extension_cap >= the required extension.
>> This way the test can easily be adapted in case of future extensions.
> 
> Not sure whether a ">=" will really be safe, since a future extension does not necessarily assert that previous extensions are available at the same time.

Hmm, I intend for that to hold. In any case, for the existing extension we have committed to it, e.g.
the documentation says:

Absolute accesses are permitted for the VM ioctl if KVM_CAP_S390_MEM_OP_EXTENSION
is > 0.

So, if we introduce an extension and allow for it to be removed with a higher extension number,
when we add testing support for that extension we'd have to change the capability check,
but the existing test case would not break.

I guess the most flexible way would be to initialize the array in the middle of main, then
you could do .skip = !extension_cap and in the future whatever expression makes sense, but
it's kinda ugly and should not be necessary anyway.
> 
> But I can still turn the bool into a numeric to make it a little bit more flexible for future use.
> 
>  Thomas
>