Hi Rob,
On Tue, 29 Jul 2025 14:51:37 -0500
Rob Herring <robh@xxxxxxxxxx> wrote:
> On Fri, Jul 25, 2025 at 05:26:14PM +0200, Herve Codina wrote:
> > On the Renesas RZ/N1 SoC, GPIOs can generate interruptions. Those
> > interruption lines are multiplexed by the GPIO Interrupt Multiplexer in
> > order to map 32 * 3 GPIO interrupt lines to 8 GIC interrupt lines.
> >
> > The GPIO interrupt multiplexer IP does nothing but select 8 GPIO
> > IRQ lines out of the 96 available to wire them to the GIC input lines.
> >
> > Signed-off-by: Herve Codina <herve.codina@xxxxxxxxxxx>
> > ---
> > drivers/soc/renesas/Kconfig | 4 +
> > drivers/soc/renesas/Makefile | 1 +
> > drivers/soc/renesas/rzn1_irqmux.c | 169 ++++++++++++++++++++++++++++++
> > 3 files changed, 174 insertions(+)
> > create mode 100644 drivers/soc/renesas/rzn1_irqmux.c
> >
> > diff --git a/drivers/soc/renesas/Kconfig b/drivers/soc/renesas/Kconfig
> > index fbc3b69d21a7..9e8ac33052fb 100644
> > --- a/drivers/soc/renesas/Kconfig
> > +++ b/drivers/soc/renesas/Kconfig
> > @@ -58,6 +58,7 @@ config ARCH_RZN1
> > select PM
> > select PM_GENERIC_DOMAINS
> > select ARM_AMBA
> > + select RZN1_IRQMUX
> >
> > if ARM && ARCH_RENESAS
> >
> > @@ -435,6 +436,9 @@ config PWC_RZV2M
> > config RST_RCAR
> > bool "Reset Controller support for R-Car" if COMPILE_TEST
> >
> > +config RZN1_IRQMUX
> > + bool "Renesas RZ/N1 GPIO IRQ multiplexer support" if COMPILE_TEST
> > +
> > config SYSC_RZ
> > bool "System controller for RZ SoCs" if COMPILE_TEST
> >
> > diff --git a/drivers/soc/renesas/Makefile b/drivers/soc/renesas/Makefile
> > index 3bdcc6a395d5..daa932c7698d 100644
> > --- a/drivers/soc/renesas/Makefile
> > +++ b/drivers/soc/renesas/Makefile
> > @@ -14,4 +14,5 @@ obj-$(CONFIG_SYS_R9A09G057) += r9a09g057-sys.o
> > # Family
> > obj-$(CONFIG_PWC_RZV2M) += pwc-rzv2m.o
> > obj-$(CONFIG_RST_RCAR) += rcar-rst.o
> > +obj-$(CONFIG_RZN1_IRQMUX) += rzn1_irqmux.o
> > obj-$(CONFIG_SYSC_RZ) += rz-sysc.o
> > diff --git a/drivers/soc/renesas/rzn1_irqmux.c b/drivers/soc/renesas/rzn1_irqmux.c
> > new file mode 100644
> > index 000000000000..37e41c2b9104
> > --- /dev/null
> > +++ b/drivers/soc/renesas/rzn1_irqmux.c
> > @@ -0,0 +1,169 @@
> > +// SPDX-License-Identifier: GPL-2.0-only
> > +/*
> > + * RZ/N1 GPIO Interrupt Multiplexer
> > + *
> > + * Copyright 2025 Schneider Electric
> > + * Author: Herve Codina <herve.codina@xxxxxxxxxxx>
> > + */
> > +
> > +#include <linux/mod_devicetable.h>
> > +#include <linux/module.h>
> > +#include <linux/of.h>
> > +#include <linux/of_irq.h>
> > +#include <linux/platform_device.h>
> > +
> > +#define IRQMUX_MAX_IRQS 8
> > +
> > +static int irqmux_is_phandle_args_equal(const struct of_phandle_args *a,
> > + const struct of_phandle_args *b)
> > +{
> > + int i;
> > +
> > + if (a->np != b->np)
> > + return false;
> > +
> > + if (a->args_count != b->args_count)
> > + return false;
> > +
> > + for (i = 0; i < a->args_count; i++) {
> > + if (a->args[i] != b->args[i])
> > + return false;
> > + }
> > +
> > + return true;
> > +}
> > +
> > +static int irqmux_find_interrupt_index(struct device *dev, struct device_node *np,
> > + const struct of_phandle_args *expected_irq)
> > +{
> > + struct of_phandle_args out_irq;
> > + bool is_equal;
> > + int ret;
> > + int i;
> > +
> > + for (i = 0; i < IRQMUX_MAX_IRQS; i++) {
> > + ret = of_irq_parse_one(np, i, &out_irq);
>
> I don't really want more users of this... More below.
>
> > + if (ret)
> > + return ret;
> > +
> > + is_equal = irqmux_is_phandle_args_equal(expected_irq, &out_irq);
> > + of_node_put(out_irq.np);
> > + if (is_equal)
> > + return i;
> > + }
> > +
> > + return -ENOENT;
> > +}
> > +
> > +struct irqmux_cb_data {
> > + struct device_node *np;
> > + struct device *dev;
> > + u32 __iomem *regs;
> > +};
> > +
> > +static int irqmux_imap_cb(void *data, const __be32 *imap,
> > + const struct of_phandle_args *parent_args)
> > +{
> > + struct irqmux_cb_data *priv = data;
> > + u32 src_hwirq;
> > + int index;
> > +
> > + /*
> > + * The child #address-cells is 0. Already checked in irqmux_setup().
> > + * The first value in imap is the src_hwirq
> > + */
> > + src_hwirq = be32_to_cpu(*imap);
>
> The iterator should take care of the endianness conversion.
Ok, it will take care.
>
> > +
> > + /*
> > + * Get the index in our interrupt array that matches the parent in the
> > + * interrupt-map
> > + */
> > + index = irqmux_find_interrupt_index(priv->dev, priv->np, parent_args);
> > + if (index < 0)
> > + return dev_err_probe(priv->dev, index, "output interrupt not found\n");
> > +
> > + dev_info(priv->dev, "interrupt %u mapped to output interrupt[%u]\n",
> > + src_hwirq, index);
>
> Do you even need "interrupts"? Just make the "interrupt-map" index
> important and correspond to the hw index. That would greatly simplify
> all this.
I would like to avoid to be based on the interrupt-map index.
Indeed, IMHO, it is less robust. I don't thing that we can enforce the
interrupt-map items order. Based on interrupt-map index, we need to ensure
that the first item is related to GIC 103, the second one to GIC 104 and so
on.
Anyway, I can simplify the code relying on the interrupt-map index even if
it is less robust.
I will propose this rework in the next iteration.
Best regards,
Hervé
>
> > +
> > + /*
> > + * Our interrupt array items matches 1:1 the interrupt lines that could
> > + * be configured by registers (same order, same number).
> > + * Configure the related register with the src hwirq retrieved from the
> > + * interrupt-map.
> > + */
> > + writel(src_hwirq, priv->regs + index);
> > +
> > + return 0;
> > +}
