On Tue, Apr 02, 2019 at 06:12:11AM +0100, Lee Jones wrote:
> On Mon, 18 Mar 2019, Andy Shevchenko wrote:
>
> > Add an mfd driver for Intel Merrifield Basin Cove PMIC.
>
> Nit: s/mfd/MFD/
Noted. And changed for v2.
> > +static const struct mfd_cell bcove_dev[] = {
> > + {
> > + .name = "mrfld_bcove_pwrbtn",
> > + .num_resources = 1,
> > + .resources = &irq_level2_resources[0],
> > + }, {
> > + .name = "mrfld_bcove_tmu",
> > + .num_resources = 1,
> > + .resources = &irq_level2_resources[1],
> > + }, {
> > + .name = "mrfld_bcove_thermal",
> > + .num_resources = 1,
> > + .resources = &irq_level2_resources[2],
> > + }, {
> > + .name = "mrfld_bcove_bcu",
> > + .num_resources = 1,
> > + .resources = &irq_level2_resources[3],
> > + }, {
> > + .name = "mrfld_bcove_adc",
> > + .num_resources = 1,
> > + .resources = &irq_level2_resources[4],
> > + }, {
> > + .name = "mrfld_bcove_charger",
> > + .num_resources = 1,
> > + .resources = &irq_level2_resources[5],
> > + }, {
> > + .name = "mrfld_bcove_extcon",
> > + .num_resources = 1,
> > + .resources = &irq_level2_resources[5],
> > + }, {
> > + .name = "mrfld_bcove_gpio",
> > + .num_resources = 1,
> > + .resources = &irq_level2_resources[6],
> > + },
> > + { .name = "mrfld_bcove_region", },
> > +};
> > +static int regmap_ipc_byte_reg_read(void *context, unsigned int reg,
>
> Prefixing these with regmap is pretty confusing, since this it not
> part of the Regmap API. Better to provide them with local names
> instead.
>
> bcove_ipc_byte_reg_read()
Good point. And changed for v2.
> > + for (i = 0; i < ARRAY_SIZE(irq_level2_resources); i++) {
> > + ret = platform_get_irq(pdev, i);
> > + if (ret < 0)
> > + return ret;
> > +
> > + irq_level2_resources[i].start = ret;
> > + irq_level2_resources[i].end = ret;
> > + }
>
> Although succinct, dragging values from one platform device into
> another doesn't sound that neat.
So, how to split resources given in one _physical_ multi-functional device to
several of them? Isn't it what MFD framework for?
Any other approach here? I'm all ears!
> Also, since the ordering of the
> devices is critical in this implementation, it also comes across as
> fragile.
How fragile? In ACPI we don't have IRQ labeling scheme. Index is used for that.
> Any reason why ACPI can't register all of the child devices, or for
> the child devices to obtain their IRQ directly from the tables?
And how are we supposed to enumerated them taking into consideration single
ACPI ID given?
--
With Best Regards,
Andy Shevchenko
