Re: [PATCH v4 3/5] drm: add SimpleDRM driver

[Noralf TrÃnnes]

Den 25.08.2016 15:09, skrev Rob Herring:
> On Mon, Aug 22, 2016 at 3:25 PM, Noralf TrÃnnes <noralf@xxxxxxxxxxx> wrote:
> > The SimpleDRM driver binds to simple-framebuffer devices and provides a
> > DRM/KMS API. It provides only a single CRTC+encoder+connector combination
> > plus one initial mode.
> >
> > Userspace can create dumb-buffers which can be blit into the real
> > framebuffer similar to UDL. No access to the real framebuffer is allowed
> > (compared to earlier version of this driver) to avoid security issues.
> > Furthermore, this way we can support arbitrary modes as long as we have a
> > conversion-helper.
> >
> > The driver was originally written by David Herrmann in 2014.
> > My main contribution is to make use of drm_simple_kms_helper and
> > rework the probe path to avoid use of the deprecated drm_platform_init()
> > and drm_driver.{load,unload}().
> > Additions have also been made for later changes to the Device Tree binding
> > document, like support for clocks, regulators and having the node under
> > /chosen. This code was lifted verbatim from simplefb.c.
> >
> > Cc: dh.herrmann@xxxxxxxxx
> > Cc: libv@xxxxxxxxx
> > Signed-off-by: Noralf TrÃnnes <noralf@xxxxxxxxxxx>
> [...]
>
> > +       /* Count the number of regulator supplies */
> > +       for_each_property_of_node(np, prop) {
> > +               p = strstr(prop->name, SUPPLY_SUFFIX);
> > +               if (p && p != prop->name)
> > +                       count++;
> > +       }
> The regulator API should have functions for this rather than open coding it.

I couldn't find anything that matches this usecase. There are regulator
bulk functions, but they require the names to be known in advance.
And I don't know how many regulators there are nor their names.

> From Documentation/devicetree/bindings/display/simple-framebuffer.txt:

- *-supply : Any number of regulators used by the framebuffer. These should
         be named according to the names in the device's design.

> > +
> > +       if (!count)
> > +               return 0;
> > +
> > +       sdrm->regulators = devm_kcalloc(&pdev->dev, count,
> > +                                       sizeof(struct regulator *),
> > +                                       GFP_KERNEL);
> > +       if (!sdrm->regulators)
> > +               return -ENOMEM;
> > +
> > +       /* Get all the regulators */
> > +       for_each_property_of_node(np, prop) {
> > +               char name[32]; /* 32 is max size of property name */
> > +
> > +               p = strstr(prop->name, SUPPLY_SUFFIX);
> > +               if (!p || p == prop->name)
> > +                       continue;
> This too.
>
> > +
> > +               strlcpy(name, prop->name,
> > +                       strlen(prop->name) - strlen(SUPPLY_SUFFIX) + 1);
> > +               regulator = devm_regulator_get_optional(&pdev->dev, name);
> [...]
>
> > +       if (IS_ENABLED(CONFIG_OF_ADDRESS) && of_chosen) {
> > +               struct device_node *np;
> > +
> > +               for_each_child_of_node(of_chosen, np) {
> > +                       if (of_device_is_compatible(np, "simple-framebuffer"))
> Rather than exporting of_chosen, this whole chunk can be replaced with
> a of_find_compatible_node call. Yes, that would match if
> simple-framebuffer exists somewhere else in the DT, but it is not the
> kernel's job to do DT validation.

This seems to do the job:

    /*
     * The binding doc states that simplefb nodes should be sub-nodes of
     * chosen and that older devicetree files can have them in a different
     * place. of_platform_device_create() just returns NULL if a device
     * has already been created for the node.
     */
    for_each_compatible_node(np, NULL, "simple-framebuffer")
        of_platform_device_create(np, NULL, NULL);


Noralf.

> > +                               of_platform_device_create(np, NULL, NULL);
> > +               }
> > +       }
> Rob