mbox series

[v3,0/2] Add IAX45 support for RZ/Five SoC

Message ID 20240422205053.496095-1-prabhakar.mahadev-lad.rj@bp.renesas.com
Headers show
Series Add IAX45 support for RZ/Five SoC | expand

Message

Lad, Prabhakar April 22, 2024, 8:50 p.m. UTC
From: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>

Hi All,

The IAX45 block on RZ/Five SoC is almost identical to the IRQC bock found
on the RZ/G2L family of SoCs.

IAX45 performs various interrupt controls including synchronization for the
external interrupts of NMI, IRQ, and GPIOINT and the interrupts of the
built-in peripheral interrupts output by each module. And it notifies the
interrupt to the PLIC.
- Select 32 TINT from 82 GPIOINT.
- Integration of bus error interrupts from system bus.
- Integration of ECC error interrupts from On-chip RAM.
- Indicate interrupt status. (NMI, IRQ, TINT, integrated bus error
  interrupt and integrated ECC error interrupt)
- Setting of interrupt detection method. (NMI, IRQ and TINT)
- All interrupts are masked by INTMASK.
- Mask function for NMI, IRQ and TINT

This patch series adds support for IAX45 in the IRQC driver and enables
this on RZ/Five SoC.

v2->v3
* DTS/I patches dropped from the series as they have been merged into
  renesas-soc tree
* Just using a const from compat string instead of having it in a items
* Added RZ/Five specific irqchip

v2: https://patchwork.kernel.org/project/linux-renesas-soc/cover/20240403203503.634465-1-prabhakar.mahadev-lad.rj@bp.renesas.com/
v1: https://patchwork.kernel.org/project/linux-renesas-soc/cover/20240129151618.90922-1-prabhakar.mahadev-lad.rj@bp.renesas.com/

Cheers,
Prabhakar

Lad Prabhakar (2):
  dt-bindings: interrupt-controller: renesas,rzg2l-irqc: Document
    RZ/Five SoC
  irqchip/renesas-rzg2l: Add support for RZ/Five SoC

 .../renesas,rzg2l-irqc.yaml                   |  17 +-
 drivers/irqchip/irq-renesas-rzg2l.c           | 150 +++++++++++++++++-
 2 files changed, 157 insertions(+), 10 deletions(-)

Comments

Geert Uytterhoeven April 24, 2024, 2:58 p.m. UTC | #1
Hi Prabhakar,

On Mon, Apr 22, 2024 at 10:51 PM Prabhakar <prabhakar.csengg@gmail.com> wrote:
> From: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
>
> The IX45 block on the RZ/Five SoC has additional mask registers
> (NMSK/IMSK/TMSK) compared to the RZ/G2L (family) SoC.
>
> A new rzfive_irqc_chip irq_chip is introduced for RZ/Five, where function
> pointers for irq_(un)mask and irq_(dis/en)able handle the (un)masking
> of the interrupts. The irq_chip pointer is now passed as an init callback
> and stored in the priv pointer to differentiate between RZ/G2L and RZ/Five.
>
> Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> ---
> v2->v3
> - Added RZ/Five specific irqchip instead of polluting the functions
> - Fixed review comments pointed by Biju and Geert
> - Updated commit message
> - moved locking respective read/write functions

Thanks for the update!

> --- a/drivers/irqchip/irq-renesas-rzg2l.c
> +++ b/drivers/irqchip/irq-renesas-rzg2l.c
> @@ -138,6 +142,113 @@ static void rzg2l_irqc_eoi(struct irq_data *d)
>         irq_chip_eoi_parent(d);
>  }
>
> +static void rzfive_irqc_mask_irq_interrupt(struct rzg2l_irqc_priv *priv,
> +                                          unsigned int hwirq)
> +{
> +       u32 bit = BIT(hwirq - IRQC_IRQ_START);
> +
> +       raw_spin_lock(&priv->lock);

I think you best move the locking to the callers that really need it...

> +       writel_relaxed(readl_relaxed(priv->base + IMSK) | bit, priv->base + IMSK);
> +       raw_spin_unlock(&priv->lock);
> +}

> +static void rzfive_tint_irq_endisable(struct irq_data *d, bool enable)
> +{
> +       struct rzg2l_irqc_priv *priv = irq_data_to_priv(d);
> +       unsigned int hwirq = irqd_to_hwirq(d);
> +
> +       if (hwirq >= IRQC_TINT_START && hwirq < IRQC_NUM_IRQ) {
> +               u32 offset = hwirq - IRQC_TINT_START;
> +               u32 tssr_offset = TSSR_OFFSET(offset);
> +               u8 tssr_index = TSSR_INDEX(offset);
> +               u32 reg;
> +
> +               if (enable)
> +                       rzfive_irqc_unmask_tint_interrupt(priv, hwirq);
> +               else
> +                       rzfive_irqc_mask_tint_interrupt(priv, hwirq);

... else you will do a lock/unlock here, followed by another one below.

> +               raw_spin_lock(&priv->lock);
> +               reg = readl_relaxed(priv->base + TSSR(tssr_index));
> +               if (enable)
> +                       reg |= TIEN << TSSEL_SHIFT(tssr_offset);
> +               else
> +                       reg &= ~(TIEN << TSSEL_SHIFT(tssr_offset));
> +               writel_relaxed(reg, priv->base + TSSR(tssr_index));
> +               raw_spin_unlock(&priv->lock);
> +       } else {
> +               if (enable)
> +                       rzfive_irqc_unmask_irq_interrupt(priv, hwirq);
> +               else
> +                       rzfive_irqc_mask_irq_interrupt(priv, hwirq);
> +       }
> +}

> @@ -401,7 +529,8 @@ static int rzg2l_irqc_parse_interrupts(struct rzg2l_irqc_priv *priv,
>         return 0;
>  }
>
> -static int rzg2l_irqc_init(struct device_node *node, struct device_node *parent)
> +static int rzg2l_irqc_init_helper(struct device_node *node, struct device_node *parent,

rzg2l_irqc_common_init()?

> +                                 const struct irq_chip *irq_chip)
>  {
>         struct irq_domain *irq_domain, *parent_domain;
>         struct platform_device *pdev;

The rest LGTM, so
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>

Gr{oetje,eeting}s,

                        Geert
Lad, Prabhakar April 25, 2024, 11:37 a.m. UTC | #2
Hi Geert,

Thank you for the review.

On Wed, Apr 24, 2024 at 3:59 PM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Hi Prabhakar,
>
> On Mon, Apr 22, 2024 at 10:51 PM Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > From: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > The IX45 block on the RZ/Five SoC has additional mask registers
> > (NMSK/IMSK/TMSK) compared to the RZ/G2L (family) SoC.
> >
> > A new rzfive_irqc_chip irq_chip is introduced for RZ/Five, where function
> > pointers for irq_(un)mask and irq_(dis/en)able handle the (un)masking
> > of the interrupts. The irq_chip pointer is now passed as an init callback
> > and stored in the priv pointer to differentiate between RZ/G2L and RZ/Five.
> >
> > Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > ---
> > v2->v3
> > - Added RZ/Five specific irqchip instead of polluting the functions
> > - Fixed review comments pointed by Biju and Geert
> > - Updated commit message
> > - moved locking respective read/write functions
>
> Thanks for the update!
>
> > --- a/drivers/irqchip/irq-renesas-rzg2l.c
> > +++ b/drivers/irqchip/irq-renesas-rzg2l.c
> > @@ -138,6 +142,113 @@ static void rzg2l_irqc_eoi(struct irq_data *d)
> >         irq_chip_eoi_parent(d);
> >  }
> >
> > +static void rzfive_irqc_mask_irq_interrupt(struct rzg2l_irqc_priv *priv,
> > +                                          unsigned int hwirq)
> > +{
> > +       u32 bit = BIT(hwirq - IRQC_IRQ_START);
> > +
> > +       raw_spin_lock(&priv->lock);
>
> I think you best move the locking to the callers that really need it...
>
Ok, will do.

> > +       writel_relaxed(readl_relaxed(priv->base + IMSK) | bit, priv->base + IMSK);
> > +       raw_spin_unlock(&priv->lock);
> > +}
>
> > +static void rzfive_tint_irq_endisable(struct irq_data *d, bool enable)
> > +{
> > +       struct rzg2l_irqc_priv *priv = irq_data_to_priv(d);
> > +       unsigned int hwirq = irqd_to_hwirq(d);
> > +
> > +       if (hwirq >= IRQC_TINT_START && hwirq < IRQC_NUM_IRQ) {
> > +               u32 offset = hwirq - IRQC_TINT_START;
> > +               u32 tssr_offset = TSSR_OFFSET(offset);
> > +               u8 tssr_index = TSSR_INDEX(offset);
> > +               u32 reg;
> > +
> > +               if (enable)
> > +                       rzfive_irqc_unmask_tint_interrupt(priv, hwirq);
> > +               else
> > +                       rzfive_irqc_mask_tint_interrupt(priv, hwirq);
>
> ... else you will do a lock/unlock here, followed by another one below.
>
and move the above code into the lock below.

Cheers,
Prabhakar