Date:   Sat, 4 Mar 2023 23:09:46 +0800
From:   Xu Yilun <yilun.xu@intel.com>
To:     Marco Pagani <marpagan@redhat.com>
Cc:     Moritz Fischer <mdf@kernel.org>, Wu Hao <hao.wu@intel.com>,
        Tom Rix <trix@redhat.com>, linux-kernel@vger.kernel.org,
        linux-fpga@vger.kernel.org
Subject: Re: [RFC PATCH 1/4] fpga: add initial KUnit test suite
Message-ID: <ZANfOj1MN+HHaF9n@yilunxu-OptiPlex-7050>
References: <20230203170653.414990-1-marpagan@redhat.com>
 <20230203170653.414990-2-marpagan@redhat.com>
 <Y/Chiq2kiAFGZpV6@yilunxu-OptiPlex-7050>
 <8c56c1a2-551d-d3ed-9e32-87ba9b9ea559@redhat.com>
 <Y/hV2F8+ZIcG2Bsq@yilunxu-OptiPlex-7050>
 <7a888be6-db0a-16dd-e0c0-73900ac139ae@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <7a888be6-db0a-16dd-e0c0-73900ac139ae@redhat.com>
Precedence: bulk

> >>>> +	ret = init_sgt_bit(&sgt_bit, fake_bit, FAKE_BIT_SIZE);
> >>>> +	KUNIT_ASSERT_EQ(test, ret, 0);
> >>>
> >>> This is not fpga function, do we need the ASSERT?
> >>>
> >>
> >> You're right. I'll change it to EXPECT.
> > 
> > Mm.. I think we may move the sgt initialization in .suite_init, and just
> > return ERROR for failure. Does it help to quickly find out this is an
> > ENV error, not a test case failure?
> 
> I looked through the documentation for guidelines on how to handle
> initialization errors, but found only the eeprom example where KUNIT_ASSERT
> is used to handle errors in eeprom_buffer_test_init(). Existing test suites
> seem to use different approaches to handle initialization errors. Some
> return an error code, while others use KUnit assertions.
> 
> I'm more inclined to follow the example in the documentation and use
> KUnit assertions. Does this approach work for you?

It's good to me.

> 
> 
> After some thought, I'm restructuring the code to test single components
> in isolation before testing them together. In this way, I think the test
> suite will be more in line with the unit testing methodology.
> 
> 
> Thanks,
> Marco
>