| Message ID | 20220830081212.164709-1-manivannan.sadhasivam@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | [v2] pinctrl: qcom: spmi-gpio: Make IRQCHIP immutable | expand |
On 2022-08-30 09:12, Manivannan Sadhasivam wrote: > The IRQCHIP implementation used inside the gpiochips are not supposed > to lower case > be changed during runtime. So let's make the one inside the spmi-gpio > gpiochip immutable. > > This fixes the below warning during boot: > gpio gpiochip0: (c440000.spmi:pmic@0:gpio@c000): not an immutable > chip, please consider fixing it! > > Separate callbacks need to be provided for irq_{mask/unmask} pointers > since > the callbacks are supposed to mask/unmask the corresponding parent IRQ > in > addition to changing the gpio_desc flags. This is all part of the existing documentation, so I don't think this is really needed. > > Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> With the above fixed: Acked-by: Marc Zyngier <maz@kernel.org> M.
On Tue, Aug 30, 2022 at 09:26:51AM +0100, Marc Zyngier wrote: > On 2022-08-30 09:12, Manivannan Sadhasivam wrote: > > The IRQCHIP implementation used inside the gpiochips are not supposed to > > lower case > > > be changed during runtime. So let's make the one inside the spmi-gpio > > gpiochip immutable. > > > > This fixes the below warning during boot: > > gpio gpiochip0: (c440000.spmi:pmic@0:gpio@c000): not an immutable > > chip, please consider fixing it! > > > > Separate callbacks need to be provided for irq_{mask/unmask} pointers > > since > > the callbacks are supposed to mask/unmask the corresponding parent IRQ > > in > > addition to changing the gpio_desc flags. > > This is all part of the existing documentation, so I don't think > this is really needed. > Yes it is documented, but developers usually refer the commits doing the similar thing while doing these kind of conversions. For them, this text serves as a quick documentation. So I prefer to keep it in the commit message. Thanks, Mani > > > > Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> > > With the above fixed: > > Acked-by: Marc Zyngier <maz@kernel.org> > > M. > -- > Jazz is not dead. It just smells funny...
On 2022-08-30 09:44, Manivannan Sadhasivam wrote: > On Tue, Aug 30, 2022 at 09:26:51AM +0100, Marc Zyngier wrote: >> On 2022-08-30 09:12, Manivannan Sadhasivam wrote: >> > The IRQCHIP implementation used inside the gpiochips are not supposed to >> >> lower case >> >> > be changed during runtime. So let's make the one inside the spmi-gpio >> > gpiochip immutable. >> > >> > This fixes the below warning during boot: >> > gpio gpiochip0: (c440000.spmi:pmic@0:gpio@c000): not an immutable >> > chip, please consider fixing it! >> > >> > Separate callbacks need to be provided for irq_{mask/unmask} pointers >> > since >> > the callbacks are supposed to mask/unmask the corresponding parent IRQ >> > in >> > addition to changing the gpio_desc flags. >> >> This is all part of the existing documentation, so I don't think >> this is really needed. >> > > Yes it is documented, but developers usually refer the commits doing > the similar > thing while doing these kind of conversions. For them, this text serves > as a > quick documentation. If they can't be bothered to read the documentation, why would they consider reading unrelated commits? > > So I prefer to keep it in the commit message. I still think this is pointless. M.
On Tue, Aug 30, 2022 at 09:48:28AM +0100, Marc Zyngier wrote: > On 2022-08-30 09:44, Manivannan Sadhasivam wrote: > > On Tue, Aug 30, 2022 at 09:26:51AM +0100, Marc Zyngier wrote: > > > On 2022-08-30 09:12, Manivannan Sadhasivam wrote: > > > > The IRQCHIP implementation used inside the gpiochips are not supposed to > > > > > > lower case > > > > > > > be changed during runtime. So let's make the one inside the spmi-gpio > > > > gpiochip immutable. > > > > > > > > This fixes the below warning during boot: > > > > gpio gpiochip0: (c440000.spmi:pmic@0:gpio@c000): not an immutable > > > > chip, please consider fixing it! > > > > > > > > Separate callbacks need to be provided for irq_{mask/unmask} pointers > > > > since > > > > the callbacks are supposed to mask/unmask the corresponding parent IRQ > > > > in > > > > addition to changing the gpio_desc flags. > > > > > > This is all part of the existing documentation, so I don't think > > > this is really needed. > > > > > > > Yes it is documented, but developers usually refer the commits doing the > > similar > > thing while doing these kind of conversions. For them, this text serves > > as a > > quick documentation. > > If they can't be bothered to read the documentation, why would they > consider reading unrelated commits? > > > > > So I prefer to keep it in the commit message. > > I still think this is pointless. > Okay. I'll remove it in v3. Thanks, Mani > M. > -- > Jazz is not dead. It just smells funny...
diff --git a/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c b/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c index ccaf40a9c0e6..d40e3b2d38cc 100644 --- a/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c +++ b/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c @@ -171,7 +171,6 @@ struct pmic_gpio_state { struct regmap *map; struct pinctrl_dev *ctrl; struct gpio_chip chip; - struct irq_chip irq; u8 usid; u8 pid_base; }; @@ -985,6 +984,33 @@ static int pmic_gpio_populate_parent_fwspec(struct gpio_chip *chip, return 0; } +static void pmic_gpio_irq_mask(struct irq_data *data) +{ + struct gpio_chip *gc = irq_data_get_irq_chip_data(data); + + irq_chip_mask_parent(data); + gpiochip_disable_irq(gc, data->hwirq); +} + +static void pmic_gpio_irq_unmask(struct irq_data *data) +{ + struct gpio_chip *gc = irq_data_get_irq_chip_data(data); + + irq_chip_unmask_parent(data); + gpiochip_enable_irq(gc, data->hwirq); +} + +static const struct irq_chip spmi_gpio_irq_chip = { + .name = "spmi-gpio", + .irq_ack = irq_chip_ack_parent, + .irq_mask = pmic_gpio_irq_mask, + .irq_unmask = pmic_gpio_irq_unmask, + .irq_set_type = irq_chip_set_type_parent, + .irq_set_wake = irq_chip_set_wake_parent, + .flags = IRQCHIP_IMMUTABLE | IRQCHIP_MASK_ON_SUSPEND, + GPIOCHIP_IRQ_RESOURCE_HELPERS, +}; + static int pmic_gpio_probe(struct platform_device *pdev) { struct irq_domain *parent_domain; @@ -1078,16 +1104,8 @@ static int pmic_gpio_probe(struct platform_device *pdev) if (!parent_domain) return -ENXIO; - state->irq.name = "spmi-gpio", - state->irq.irq_ack = irq_chip_ack_parent, - state->irq.irq_mask = irq_chip_mask_parent, - state->irq.irq_unmask = irq_chip_unmask_parent, - state->irq.irq_set_type = irq_chip_set_type_parent, - state->irq.irq_set_wake = irq_chip_set_wake_parent, - state->irq.flags = IRQCHIP_MASK_ON_SUSPEND, - girq = &state->chip.irq; - girq->chip = &state->irq; + gpio_irq_chip_set_chip(girq, &spmi_gpio_irq_chip); girq->default_type = IRQ_TYPE_NONE; girq->handler = handle_level_irq; girq->fwnode = of_node_to_fwnode(state->dev->of_node);
The IRQCHIP implementation used inside the gpiochips are not supposed to be changed during runtime. So let's make the one inside the spmi-gpio gpiochip immutable. This fixes the below warning during boot: gpio gpiochip0: (c440000.spmi:pmic@0:gpio@c000): not an immutable chip, please consider fixing it! Separate callbacks need to be provided for irq_{mask/unmask} pointers since the callbacks are supposed to mask/unmask the corresponding parent IRQ in addition to changing the gpio_desc flags. Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> --- Changes in v2: * Added new callbacks for irq_{mask/unmask} for masking/unmasking the parent IRQ that was missing in v1. (Thanks to Marc) * Modified the commit message accordingly. drivers/pinctrl/qcom/pinctrl-spmi-gpio.c | 38 +++++++++++++++++------- 1 file changed, 28 insertions(+), 10 deletions(-)