[PATCH 2/4] mfd: mt6360: Remove handle_post_irq callback function
Gene Chen
gene.chen.richtek at gmail.com
Fri Jul 17 05:35:50 EDT 2020
Matthias Brugger <matthias.bgg at gmail.com> 於 2020年7月10日 週五 下午10:22寫道:
>
>
>
> On 07/07/2020 12:30, Gene Chen wrote:
> > From: Gene Chen <gene_chen at richtek.com>
> >
> > Remove handle_post_irq which is used to retrigger irq.
> > Set irq level low trigger to keep irq always be handled.
> >
> > Signed-off-by: Gene Chen <gene_chen at richtek.com>
> > ---
> > drivers/mfd/mt6360-core.c | 25 ++++++++-----------------
> > include/linux/mfd/mt6360.h | 6 +++---
> > 2 files changed, 11 insertions(+), 20 deletions(-)
> >
> > diff --git a/drivers/mfd/mt6360-core.c b/drivers/mfd/mt6360-core.c
> > index 5dfc13e..2dd5918 100644
> > --- a/drivers/mfd/mt6360-core.c
> > +++ b/drivers/mfd/mt6360-core.c
> > @@ -208,24 +208,16 @@ static const struct regmap_irq mt6360_irqs[] = {
> > REGMAP_IRQ_REG_LINE(MT6360_LDO7_PGB_EVT, 8),
> > };
> >
> > -static int mt6360_pmu_handle_post_irq(void *irq_drv_data)
> > -{
> > - struct mt6360_data *data = irq_drv_data;
> > -
> > - return regmap_update_bits(data->regmap,
> > - MT6360_PMU_IRQ_SET, MT6360_IRQ_RETRIG, MT6360_IRQ_RETRIG);
> > -}
> > -
> > -static struct regmap_irq_chip mt6360_irq_chip = {
> > +static const struct regmap_irq_chip mt6360_irq_chip = {
> > + .name = "mt6360_irqs",
> > .irqs = mt6360_irqs,
> > .num_irqs = ARRAY_SIZE(mt6360_irqs),
> > - .num_regs = MT6360_PMU_IRQ_REGNUM,
> > - .mask_base = MT6360_PMU_CHG_MASK1,
> > - .status_base = MT6360_PMU_CHG_IRQ1,
> > - .ack_base = MT6360_PMU_CHG_IRQ1,
> > + .num_regs = MT6360_IRQ_REGNUM,
> > + .mask_base = MT6360_REG_PMU_CHGMASK1,
> > + .status_base = MT6360_REG_PMU_CHGIRQ1,
> > + .ack_base = MT6360_REG_PMU_CHGIRQ1,
>
> Non relevant patches for the commit. Please drop this.
>
ACK
> > .init_ack_masked = true,
> > .use_ack = true,
> > - .handle_post_irq = mt6360_pmu_handle_post_irq,
> > };
> >
> > static const struct regmap_config mt6360_pmu_regmap_config = {
> > @@ -339,10 +331,9 @@ static int mt6360_probe(struct i2c_client *client)
> > return -ENODEV;
> > }
> >
> > - mt6360_irq_chip.irq_drv_data = data;
> > ret = devm_regmap_add_irq_chip(&client->dev, data->regmap, client->irq,
> > - IRQF_TRIGGER_FALLING, 0,
> > - &mt6360_irq_chip, &data->irq_data);
> > + IRQF_ONESHOT, 0, &mt6360_irq_chip,
>
> From what I see IRQF_ONESHOT is set implicitly in regmap_add_irq_chip_fwnode.
> Actually I think what you want to add is IRQF_TRIGGER_LOW.
>
ACK.
I will modify argument "irqflags" to 0, because irq trigger type is
set by of_irq_get function
> > + &data->irq_data);
> > if (ret) {
> > dev_err(&client->dev, "Failed to add Regmap IRQ Chip\n");
> > return ret;
> > diff --git a/include/linux/mfd/mt6360.h b/include/linux/mfd/mt6360.h
> > index 76077e4..9fc6718 100644
> > --- a/include/linux/mfd/mt6360.h
> > +++ b/include/linux/mfd/mt6360.h
> > @@ -179,7 +179,7 @@ struct mt6360_data {
> > #define MT6360_PMU_SPARE2 (0xA0)
> > #define MT6360_PMU_SPARE3 (0xB0)
> > #define MT6360_PMU_SPARE4 (0xC0)
> > -#define MT6360_PMU_CHG_IRQ1 (0xD0)
> > +#define MT6360_REG_PMU_CHGIRQ1 (0xD0)
> > #define MT6360_PMU_CHG_IRQ2 (0xD1)
> > #define MT6360_PMU_CHG_IRQ3 (0xD2)
> > #define MT6360_PMU_CHG_IRQ4 (0xD3)
> > @@ -211,7 +211,7 @@ struct mt6360_data {
> > #define MT6360_PMU_BUCK2_STAT (0xED)
> > #define MT6360_PMU_LDO_STAT1 (0xEE)
> > #define MT6360_PMU_LDO_STAT2 (0xEF)
> > -#define MT6360_PMU_CHG_MASK1 (0xF0)
> > +#define MT6360_REG_PMU_CHGMASK1 (0xF0)
> > #define MT6360_PMU_CHG_MASK2 (0xF1)
> > #define MT6360_PMU_CHG_MASK3 (0xF2)
> > #define MT6360_PMU_CHG_MASK4 (0xF3)
> > @@ -230,7 +230,7 @@ struct mt6360_data {
> > #define MT6360_PMU_MAXREG (MT6360_PMU_LDO_MASK2)
> >
> > /* MT6360_PMU_IRQ_SET */
> > -#define MT6360_PMU_IRQ_REGNUM (MT6360_PMU_LDO_IRQ2 - MT6360_PMU_CHG_IRQ1 + 1)
>
> Not relevant for the commit. But why did you use this strange computation in the
> first place?
>
I wanted to emphasize irq range with start and end address.
> > +#define MT6360_IRQ_REGNUM 16
> > #define MT6360_IRQ_RETRIG BIT(2)
> >
> > #define CHIP_VEN_MASK (0xF0)
> >
More information about the Linux-mediatek
mailing list