On 7/14/21 4:43 PM, Markus Armbruster wrote:
> Philippe Mathieu-Daudé <phi...@redhat.com> writes:
>
>> +Markus
>>
>> On 7/12/21 5:47 PM, Stefan Berger wrote:
>>>
>>> On 7/12/21 11:29 AM, Philippe Mathieu-Daudé wrote:
>>>> Hi Stefan,
>>>>
>>>> On 7/12/21 5:09 PM, Stefan Berger wrote:
>>>>> This series of patches adds test case for TPM 1.2 ACPI tables.
>>>>>
>>>>> Stefan
>>>>>
>>>>> v3:
>>>>> - Define enum TPMVersion for when CONFIG_TPM is not defined
>>>>> affected patches 2 and 6
>>>> I think in 11fb99e6f48..e542b71805d we missed an extra patch
>>>> for qtests. Probably (untested):
>>>
>>> Shouldn't we have seen test compilation errors already?
>>>
>>> I didn't go down this route for the build system (as you show below)
>>> because in this series we are testing ACPI tables and I introduce the
>>> reference to enum TPMVersion here, which wasn't needed before. The
>>> alternative may be to go into 8/9 and eliminate all TPM code if
>>> CONFIG_TPM is not set. The introduction of the enum now passes the tests
>>> with --enable-tpm and --disable-tpm.
>>>
>>> Otherwise the BIOS test are skipped due to this here:
>>>
>>>
>>> static void test_acpi_tcg_tpm(const char *machine, const char *tpm_if,
>>> uint64_t base, enum TPMVersion tpm_version)
>>> {
>>> #ifdef CONFIG_TPM
>>> [...]
>>>
>>> #else
>>> g_test_skip("TPM disabled");
>>> #endif
>>> }
>>>
>>> So I didn't want to clutter this code with more #ifdef CONFIG_TPM but
>>> maybe that would be the right solution.
>>
>> IMO the "right" solution is to check via QMP if TMP is supported
>> or not. This is now doable since commit caff255a546 ("tpm: Return
>> QMP error when TPM is disabled in build").
>>
>> Long term we'd like to decouple the tests/ build from the various
>> QEMU configurations, and build the tests once.
>
> This argument applies only to macros from target-specific headers like
> $TARGET-config-target.h, not to macros from config-host.h. #ifdef
> CONFIG_TPM should be fine, shouldn't it?
Some definitions depend on the host (OS, libraries installed, ...),
others depend on the --enable/--disable ./configure options.
IMO it would be nice if we could get qtests independent of the latter.
I suppose config-host.h holds both kinds.
