[PATCH 07/25] tegra: Change #ifdef for nop

[Sean Anderson]

On 10/9/23 11:32, Simon Glass wrote:
> Hi Sean,
> 
> On Sat, 7 Oct 2023 at 17:21, Sean Anderson <seanga2 at gmail.com> wrote:
>>
>> On 10/7/23 19:10, Simon Glass wrote:
>>> Hi Tom.
>>>
>>> On Sun, 24 Sept 2023 at 18:43, Tom Rini <trini at konsulko.com> wrote:
>>>>
>>>> On Sun, Sep 24, 2023 at 02:39:25PM -0600, Simon Glass wrote:
>>>>
>>>>> This code is normally compiled for Tegra, but sandbox can also compile
>>>>> it. We should not use UNIT_TEST as a synonym for SANDBOX, since it is
>>>>> possible to disable UNIT_TEST for sandbox.
>>>>>
>>>>> Correct the condition.
>>>>>
>>>>> Signed-off-by: Simon Glass <sjg at chromium.org>
>>>>> ---
>>>>>
>>>>>    include/k210/pll.h | 2 +-
>>>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/include/k210/pll.h b/include/k210/pll.h
>>>>> index fd16a89cb203..6dd60b2eb4fc 100644
>>>>> --- a/include/k210/pll.h
>>>>> +++ b/include/k210/pll.h
>>>>> @@ -13,7 +13,7 @@ struct k210_pll_config {
>>>>>         u8 od;
>>>>>    };
>>>>>
>>>>> -#ifdef CONFIG_UNIT_TEST
>>>>> +#ifdef CONFIG_SANDBOX
>>>>>    TEST_STATIC int k210_pll_calc_config(u32 rate, u32 rate_in,
>>>>>                                      struct k210_pll_config *best);
>>>>>    #ifndef nop
>>>>
>>>> Tegra? Do you mean sifive?  That's where CLK_K210 stuff is... but it
>>>
>>> Oh yes, I got confused.
>>>
>>>> also seems wrong, you can run unit test on real hardware, and this is a
>>>> test that could (should?) be run on that platform.
>>>
>>> Only if it enables UNIT_TEST. You cannot run unit tests without that.
>>> The current tests are designed for sandbox.
>>
>> FWIW I have run this test on actual hardware. My intent here was to allow
>> unit tests to access functions which would otherwise be declared static.
> 
> Er, with or without UNIT_TEST enabled? How can it even build if this
> declaration is only for sandbox?

With UNIT_TEST of course. Although since this is a forward-declaration, the
UNIT_TEST ifdef isn't really even necessary. If it's on actual hardware, nop
should already be defined. So maybe this should be something like

#if CONFIG_SANDBOX
#define nop()
#endif

--Sean