Received-SPF: pass (google.com: domain of linux-kernel+bounces-203376-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Precedence: bulk
MIME-Version: 1.0
References: <20240605212840.3227774-1-olvaffe@gmail.com> <ZmDiYze9MEFxn50C@smile.fi.intel.com>
In-Reply-To: <ZmDiYze9MEFxn50C@smile.fi.intel.com>
From: Chia-I Wu <olvaffe@gmail.com>
Date: Wed, 5 Jun 2024 15:52:26 -0700
Message-ID: <CAPaKu7SJ5TVXsqnsqnucpE=wVKwGCdXyCTNBhr+rcJrmLw5GgQ@mail.gmail.com>
Subject: Re: [PATCH v4] resource: add a simple test for walk_iomem_res_desc()
To: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>, Dan Williams <dan.j.williams@intel.com>, 
	AKASHI Takahiro <takahiro.akashi@linaro.org>, Dave Jiang <dave.jiang@intel.com>, 
	Alison Schofield <alison.schofield@intel.com>, Baoquan He <bhe@redhat.com>, 
	linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, Jun 5, 2024 at 3:10=E2=80=AFPM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
>
> On Wed, Jun 05, 2024 at 02:28:26PM -0700, Chia-I Wu wrote:
> > This mainly tests that find_next_iomem_res() does not miss resources.
>
> ...
>
> > v2: update subject, use DEFINE_RES_NAMED and hardcoded offsets
> > v3: really hardcode offsets, with 4KB intervals since 0x1000 is easier
> >     to read than 0x400
>
> Right, but...
>
> > v4: use RESOURCE_SIZE_MAX, split allocate_resource and KUNIT_ASSERT_EQ,
> >     and other cosmetic changes
>
> ...
>
> > +     ret =3D allocate_resource(&iomem_resource, &root, 0x100000,
> > +                     0, RESOURCE_SIZE_MAX, 0x100000, NULL, NULL);
>
> Just double check that limits.h is included.
>
> ...
>
> > +     /* build the resource tree */
> > +     res[0] =3D DEFINE_RES_NAMED(root.start + 0x0000, 0x1000, "SYSRAM =
1",
> > +                     IORESOURCE_SYSTEM_RAM);
> > +     res[1] =3D DEFINE_RES_NAMED(root.start + 0x1000, 0x1000, "OTHER",=
 0);
> > +
> > +     res[2] =3D DEFINE_RES_NAMED(root.start + 0x3000, 0x1000, "NESTED"=
, 0);
> > +     res[3] =3D DEFINE_RES_NAMED(root.start + 0x3800, 0x0400, "SYSRAM =
2",
> > +                     IORESOURCE_SYSTEM_RAM);
>
> ...here is overlap with the previous resource.
>
> And here is the gap to the next one, in case we make that overlapping gon=
e.
>
> > +     res[4] =3D DEFINE_RES_NAMED(root.start + 0x4000, 0x1000, "SYSRAM =
3",
> > +                     IORESOURCE_SYSTEM_RAM);
>
> It wasn't the case in previous data. Please, elaborate what's going on he=
re?
The test data is chosen to be

  first interval: a matching resource (res[0])
  second interval: a non-matching resource (res[1])
  third interval: a hole
  fourth interval: a matching resource (res[3]) nested in a
non-matching resource (res[2])
  fifth interval: a matching resource (res[4])

The idea hasn't changed between revisions.

res[3] went from a half of res[2] to a quarter of res[2] in v4.  I
guess it causes confusion if it is not viewed as a nested resource.

>
> ...
>
> And rather sending one version per 12h, take your time and think more abo=
ut
> test data. What are we testing? Are the testing data correct? Shouldn't w=
e also
> have negative test cases?
The current choice of test data covers the most common patterns.  Do
you have other patterns you want to cover?  I am new to the resource
code and that's why I am largely reactive to review feedback.

>
> --
> With Best Regards,
> Andy Shevchenko
>
>