On Wed, Aug 14, 2013 at 04:54:55PM +0200, Florian Lobmaier wrote:
> +static struct regmap_irq_chip as3722_irq_chip = {
> + .name = "as3722",
> + .irqs = as3722_irqs,
> + .num_irqs = ARRAY_SIZE(as3722_irqs),
> + .num_regs = 4,
> + .status_base = AS3722_INTERRUPTSTATUS1_REG,
> + .mask_base = AS3722_INTERRUPTMASK1_REG,
> + .wake_base = 1,
> +};
wake_base looks wrong - that should be either absent or a register name.
> +static void as3722_reg_init(struct as3722 *as3722,
> + struct as3722_reg_init *reg_data)
> +{
> + int ret;
> +
> + while (reg_data->reg != AS3722_REG_INIT_TERMINATE) {
> + ret = as3722_reg_write(as3722, reg_data->reg, reg_data->val);
> + if (ret) {
> + dev_err(as3722->dev,
> + "reg setup failed: %d\n", ret);
> + return;
> + }
> + reg_data++;
> + }
> +}
This looks like it might be supposed to be a register patch?
> +int as3722_read_adc(struct as3722 *as3722,
> + enum as3722_adc_channel channel,
> + enum as3722_adc_source source,
> + enum as3722_adc_voltange_range voltage_range)
This should be moved over to IIO - the ADC code in MFDs predates IIO
being available.
> +static irqreturn_t as3722_onkey_press_irq(int irq, void *irq_data)
> +static irqreturn_t as3722_onkey_lpress_irq(int irq, void *irq_data)
These should be handled by an input driver.
> +static irqreturn_t as3722_temp_sd0_shutdown_irq(int irq, void *irq_data)
> +{
> + struct as3722 *as3722 = irq_data;
> +
> + dev_dbg(as3722->dev, "AS3722 temp SD0 shutdown triggered\n");
> + return IRQ_HANDLED;
> +}
You probably want these to complain loudly rather than as dev_dbg(), I'm
not sure what the status of the thermal framework is.
> + /* enable 32kHz clock output if required */
> + if (pdata->enable_clk32out_pin)
> + as3722_set_bits(as3722, AS3722_RTC_CONTROL_REG,
> + AS3722_CLK32OUT_ENABLE_MASK,
> + AS3722_CLK32OUT_ENABLE_ON);
> + else
> + as3722_set_bits(as3722, AS3722_RTC_CONTROL_REG,
> + AS3722_CLK32OUT_ENABLE_MASK,
> + AS3722_CLK32OUT_ENABLE_OFF);
This looks like a job for the clk API.
signature.asc
Description: Digital signature
