RE: [PATCH v2] irqchip: xilinx: Add support for multiple instances

[Mubin Usman Sayyed <MUBINUSM@xilinx.com>]

Hi Marc,

> -----Original Message-----
> From: Marc Zyngier <maz@kernel.org>
> Sent: Wednesday, February 5, 2020 10:23 PM
> To: Mubin Usman Sayyed <MUBINUSM@xilinx.com>
> Cc: tglx@linutronix.de; jason@lakedaemon.net; Michal Simek
> <michals@xilinx.com>; linux-arm-kernel@lists.infradead.org; linux-
> kernel@vger.kernel.org; Siva Durga Prasad Paladugu <sivadur@xilinx.com>;
> Anirudha Sarangi <anirudh@xilinx.com>
> Subject: Re: [PATCH v2] irqchip: xilinx: Add support for multiple instances
> 
> On 2020-02-05 14:05, Mubin Usman Sayyed wrote:
> > From: Mubin Sayyed <mubin.usman.sayyed@xilinx.com>
> >
> > This patch adds support for multiple instances of xilinx interrupt
> > controller. Below configurations are supported by driver,
> >
> > - peripheral->xilinx-intc->xilinx-intc->gic
> > - peripheral->xilinx-intc->xilinx-intc
> >
> > Signed-off-by: Anirudha Sarangi <anirudha.sarangi@xilinx.com>
> > Signed-off-by: Mubin Sayyed <mubin.usman.sayyed@xilinx.com>
> > ---
> > Changes for v2:
> >         - Removed write_fn/read_fn hooks, used xintc_write/
> >           xintc_read directly
> >         - Moved primary_intc declaration after xintc_irq_chip
> > ---
> >  drivers/irqchip/irq-xilinx-intc.c | 121
> > +++++++++++++++++++++++---------------
> >  1 file changed, 73 insertions(+), 48 deletions(-)
> >
> > diff --git a/drivers/irqchip/irq-xilinx-intc.c
> > b/drivers/irqchip/irq-xilinx-intc.c
> > index e3043de..14cb630 100644
> > --- a/drivers/irqchip/irq-xilinx-intc.c
> > +++ b/drivers/irqchip/irq-xilinx-intc.c
> > @@ -38,29 +38,32 @@ struct xintc_irq_chip {
> >         void            __iomem *base;
> >         struct          irq_domain *root_domain;
> >         u32             intr_mask;
> > +       struct                  irq_chip *intc_dev;
> > +       u32                             nr_irq;
> >  };
> >
> > -static struct xintc_irq_chip *xintc_irqc;
> > +static struct xintc_irq_chip *primary_intc;
> >
> > -static void xintc_write(int reg, u32 data)
> > +static void xintc_write(void __iomem *addr, u32 data)
> >  {
> >         if (static_branch_unlikely(&xintc_is_be))
> > -               iowrite32be(data, xintc_irqc->base + reg);
> > +               iowrite32be(data, addr);
> >         else
> > -               iowrite32(data, xintc_irqc->base + reg);
> > +               iowrite32(data, addr);
> >  }
> >
> > -static unsigned int xintc_read(int reg)
> > +static unsigned int xintc_read(void __iomem *addr)
> 
> Since you are changing this, please change the return value to reflect the size
> of what you're returning (u32 instead of unsigned int).
[Mubin]: Changed in v3
> 
> >  {
> >         if (static_branch_unlikely(&xintc_is_be))
> > -               return ioread32be(xintc_irqc->base + reg);
> > +               return ioread32be(addr);
> >         else
> > -               return ioread32(xintc_irqc->base + reg);
> > +               return ioread32(addr);
> >  }
> >
> >  static void intc_enable_or_unmask(struct irq_data *d)  {
> >         unsigned long mask = 1 << d->hwirq;
> > +       struct xintc_irq_chip *local_intc =
> > irq_data_get_irq_chip_data(d);
> >
> >         pr_debug("irq-xilinx: enable_or_unmask: %ld\n", d->hwirq);
> >
> > @@ -69,47 +72,57 @@ static void intc_enable_or_unmask(struct irq_data
> > *d)
> >          * acks the irq before calling the interrupt handler
> >          */
> >         if (irqd_is_level_type(d))
> > -               xintc_write(IAR, mask);
> > +               xintc_write(local_intc->base + IAR, mask);
> >
> > -       xintc_write(SIE, mask);
> > +       xintc_write(local_intc->base + SIE, mask);
> >  }
> >
> >  static void intc_disable_or_mask(struct irq_data *d)  {
> > +       struct xintc_irq_chip *local_intc =
> > irq_data_get_irq_chip_data(d);
> > +
> >         pr_debug("irq-xilinx: disable: %ld\n", d->hwirq);
> > -       xintc_write(CIE, 1 << d->hwirq);
> > +       xintc_write(local_intc->base + CIE, 1 << d->hwirq);
> >  }
> >
> >  static void intc_ack(struct irq_data *d)  {
> > +       struct xintc_irq_chip *local_intc =
> > irq_data_get_irq_chip_data(d);
> > +
> >         pr_debug("irq-xilinx: ack: %ld\n", d->hwirq);
> > -       xintc_write(IAR, 1 << d->hwirq);
> > +       xintc_write(local_intc->base + IAR, 1 << d->hwirq);
> >  }
> >
> >  static void intc_mask_ack(struct irq_data *d)  {
> >         unsigned long mask = 1 << d->hwirq;
> > +       struct xintc_irq_chip *local_intc =
> > irq_data_get_irq_chip_data(d);
> >
> >         pr_debug("irq-xilinx: disable_and_ack: %ld\n", d->hwirq);
> > -       xintc_write(CIE, mask);
> > -       xintc_write(IAR, mask);
> > +       xintc_write(local_intc->base + CIE, mask);
> > +       xintc_write(local_intc->base + IAR, mask);
> >  }
> >
> > -static struct irq_chip intc_dev = {
> > -       .name = "Xilinx INTC",
> > -       .irq_unmask = intc_enable_or_unmask,
> > -       .irq_mask = intc_disable_or_mask,
> > -       .irq_ack = intc_ack,
> > -       .irq_mask_ack = intc_mask_ack,
> > -};
> > +static unsigned int xintc_get_irq_local(struct xintc_irq_chip
> > *local_intc)
> > +{
> > +       int hwirq, irq = -1;
> 
> Type consistency for hwirq.
[Mubin]: Fixed in v3
> 
> > +
> > +       hwirq = xintc_read(local_intc->base + IVR);
> > +       if (hwirq != -1U)
> > +               irq = irq_find_mapping(local_intc->root_domain,
> > + hwirq);
> > +
> > +       pr_debug("irq-xilinx: hwirq=%d, irq=%d\n", hwirq, irq);
> > +
> > +       return irq;
> 
> That now gives you both -1 and 0 for error values. Please stick with 0.
[Mubin]: Fixed in v3
> 
> > +}
> >
> >  unsigned int xintc_get_irq(void)
> >  {
> > -       unsigned int hwirq, irq = -1;
> > +       int hwirq, irq = -1;
> >
> > -       hwirq = xintc_read(IVR);
> > +       hwirq = xintc_read(primary_intc->base + IVR);
> >         if (hwirq != -1U)
> > -               irq = irq_find_mapping(xintc_irqc->root_domain, hwirq);
> > +               irq = irq_find_mapping(primary_intc->root_domain,
> > hwirq);
> >
> >         pr_debug("irq-xilinx: hwirq=%d, irq=%d\n", hwirq, irq);
> 
> I have the ugly feeling I'm reading the same code twice... Surely you can
> make these two functions common code.
[Mubin]:  Upcoming patchset from Michal would be removing xintc_get_irq,  instead of that  "handle_domain_irq()"  would be used independently.  I hope that is fine.
 
> >
> > @@ -118,15 +131,18 @@ unsigned int xintc_get_irq(void)
> >
> >  static int xintc_map(struct irq_domain *d, unsigned int irq,
> > irq_hw_number_t hw)  {
> > -       if (xintc_irqc->intr_mask & (1 << hw)) {
> > -               irq_set_chip_and_handler_name(irq, &intc_dev,
> > +       struct xintc_irq_chip *local_intc = d->host_data;
> > +
> > +       if (local_intc->intr_mask & (1 << hw)) {
> 
> BIT(hw)
[Mubin]: Fixed in v3
> 
> > +               irq_set_chip_and_handler_name(irq,
> > local_intc->intc_dev,
> >                                                 handle_edge_irq,
> > "edge");
> >                 irq_clear_status_flags(irq, IRQ_LEVEL);
> >         } else {
> > -               irq_set_chip_and_handler_name(irq, &intc_dev,
> > +               irq_set_chip_and_handler_name(irq,
> > local_intc->intc_dev,
> >                                                 handle_level_irq,
> > "level");
> >                 irq_set_status_flags(irq, IRQ_LEVEL);
> >         }
> > +       irq_set_chip_data(irq, local_intc);
> >         return 0;
> >  }
> >
> > @@ -138,11 +154,13 @@ static const struct irq_domain_ops
> > xintc_irq_domain_ops = {  static void xil_intc_irq_handler(struct
> > irq_desc *desc)  {
> >         struct irq_chip *chip = irq_desc_get_chip(desc);
> > +       struct xintc_irq_chip *local_intc =
> > +               irq_data_get_irq_handler_data(&desc->irq_data);
> >         u32 pending;
> >
> >         chained_irq_enter(chip, desc);
> >         do {
> > -               pending = xintc_get_irq();
> > +               pending = xintc_get_irq_local(local_intc);
> >                 if (pending == -1U)
> >                         break;
> >                 generic_handle_irq(pending); @@ -153,28 +171,20 @@
> > static void xil_intc_irq_handler(struct irq_desc
> > *desc)
> >  static int __init xilinx_intc_of_init(struct device_node *intc,
> >                                              struct device_node
> > *parent)
> >  {
> > -       u32 nr_irq;
> >         int ret, irq;
> >         struct xintc_irq_chip *irqc;
> > -
> > -       if (xintc_irqc) {
> > -               pr_err("irq-xilinx: Multiple instances aren't
> > supported\n");
> > -               return -EINVAL;
> > -       }
> > +       struct irq_chip *intc_dev;
> >
> >         irqc = kzalloc(sizeof(*irqc), GFP_KERNEL);
> >         if (!irqc)
> >                 return -ENOMEM;
> > -
> > -       xintc_irqc = irqc;
> > -
> >         irqc->base = of_iomap(intc, 0);
> >         BUG_ON(!irqc->base);
> >
> > -       ret = of_property_read_u32(intc, "xlnx,num-intr-inputs",
> > &nr_irq);
> > +       ret = of_property_read_u32(intc, "xlnx,num-intr-inputs",
> > &irqc->nr_irq);
> >         if (ret < 0) {
> >                 pr_err("irq-xilinx: unable to read
> > xlnx,num-intr-inputs\n");
> > -               goto err_alloc;
> > +               goto error;
> >         }
> >
> >         ret = of_property_read_u32(intc, "xlnx,kind-of-intr",
> > &irqc->intr_mask); @@ -183,30 +193,42 @@ static int __init
> > xilinx_intc_of_init(struct device_node *intc,
> >                 irqc->intr_mask = 0;
> >         }
> >
> > -       if (irqc->intr_mask >> nr_irq)
> > +       if (irqc->intr_mask >> irqc->nr_irq)
> >                 pr_warn("irq-xilinx: mismatch in kind-of-intr
> > param\n");
> >
> >         pr_info("irq-xilinx: %pOF: num_irq=%d, edge=0x%x\n",
> > -               intc, nr_irq, irqc->intr_mask);
> > +               intc, irqc->nr_irq, irqc->intr_mask);
> > +
> > +       intc_dev = kzalloc(sizeof(*intc_dev), GFP_KERNEL);
> > +       if (!intc_dev) {
> > +               ret = -ENOMEM;
> > +               goto error;
> > +       }
> >
> > +       intc_dev->name = intc->full_name;
> 
> No. The world doesn't need to see the OF path of your interrupt controller in
> /proc/cpuinfo.
> The name that was there before was perfectly descriptive, please stick to it.
[Mubin]: Reverted in v3
> 
> > +       intc_dev->irq_unmask = intc_enable_or_unmask,
> > +       intc_dev->irq_mask = intc_disable_or_mask,
> > +       intc_dev->irq_ack = intc_ack,
> > +       intc_dev->irq_mask_ack = intc_mask_ack,
> 
> And this structure should stay static, as it was before.
[Mubin]: Reverted in v3
> 
> > +       irqc->intc_dev = intc_dev;
> >
> >         /*
> >          * Disable all external interrupts until they are
> >          * explicity requested.
> >          */
> > -       xintc_write(IER, 0);
> > +       xintc_write(irqc->base + IER, 0);
> >
> >         /* Acknowledge any pending interrupts just in case. */
> > -       xintc_write(IAR, 0xffffffff);
> > +       xintc_write(irqc->base + IAR, 0xffffffff);
> >
> >         /* Turn on the Master Enable. */
> > -       xintc_write(MER, MER_HIE | MER_ME);
> > -       if (!(xintc_read(MER) & (MER_HIE | MER_ME))) {
> > +       xintc_write(irqc->base + MER, MER_HIE | MER_ME);
> > +       if (!(xintc_read(irqc->base + MER) & (MER_HIE | MER_ME))) {
> >                 static_branch_enable(&xintc_is_be);
> > -               xintc_write(MER, MER_HIE | MER_ME);
> > +               xintc_write(irqc->base + MER, MER_HIE | MER_ME);
> >         }
> >
> > -       irqc->root_domain = irq_domain_add_linear(intc, nr_irq,
> > +       irqc->root_domain = irq_domain_add_linear(intc, irqc->nr_irq,
> >
> > &xintc_irq_domain_ops, irqc);
> >         if (!irqc->root_domain) {
> >                 pr_err("irq-xilinx: Unable to create IRQ domain\n");
> > @@ -225,13 +247,16 @@ static int __init xilinx_intc_of_init(struct
> > device_node *intc,
> >                         goto err_alloc;
> >                 }
> >         } else {
> > -               irq_set_default_host(irqc->root_domain);
> > +               primary_intc = irqc;
> > +               irq_set_default_host(primary_intc->root_domain);
> >         }
> >
> >         return 0;
> >
> >  err_alloc:
> > -       xintc_irqc = NULL;
> > +       kfree(intc_dev);
> > +error:
> > +       iounmap(irqc->base);
> >         kfree(irqc);
> >         return ret;
> >
> > --
> > 2.7.4
> >
> > This email and any attachments are intended for the sole use of the
> > named recipient(s) and contain(s) confidential information that may be
> > proprietary, privileged or copyrighted under applicable law. If you
> > are not the intended recipient, do not read, copy, or forward this
> > email message or any attachments. Delete this email message and any
> > attachments immediately.
> 
> Please tell your employer to fix their email server.
[Mubin]: Fixed in v3

Thanks,
Mubin
> 
> Thanks,
> 
>          M.
> --
> Jazz is not dead. It just smells funny...