Re: [PATCH v4 29/39] irqchip/atmel-aic5: Add support to get nirqs from DT for sam9x60 & sam9x7

From: Conor Dooley
Date: Fri Mar 08 2024 - 05:16:27 EST


On Fri, Mar 08, 2024 at 08:50:43AM +0000, Varshini.Rajendran@xxxxxxxxxxxxx wrote:
> On 03/03/24 5:51 pm, claudiu beznea wrote:
> > EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
> >
> > On 23.02.2024 19:29, Varshini Rajendran wrote:
> >> Add support to get number of IRQs from the respective DT node for sam9x60
> >> and sam9x7 devices. Since only this factor differs between the two SoCs,
> >> this patch adds support for the same. Adapt the sam9x60 dtsi
> >> accordingly.
> >>
> >> Signed-off-by: Varshini Rajendran <varshini.rajendran@xxxxxxxxxxxxx>
> >> ---
> >> Changes in v4:
> >> - Changed the implementation to fetch the NIRQs from DT as per the
> >> comment to avoid introducing a new compatible when this is the only
> >> difference between the SoCs related to this IP.
> >> ---
> >> arch/arm/boot/dts/microchip/sam9x60.dtsi | 1 +
> >> drivers/irqchip/irq-atmel-aic5.c | 11 ++++++++---

Driver and binding changes should be in different patches. Having them
in the same patch is usually a red flag for ABI breakage.

> >> 2 files changed, 9 insertions(+), 3 deletions(-)
> >>
> >> diff --git a/arch/arm/boot/dts/microchip/sam9x60.dtsi b/arch/arm/boot/dts/microchip/sam9x60.dtsi
> >> index 73d570a17269..e405f68c9f54 100644
> >> --- a/arch/arm/boot/dts/microchip/sam9x60.dtsi
> >> +++ b/arch/arm/boot/dts/microchip/sam9x60.dtsi
> >> @@ -1201,6 +1201,7 @@ aic: interrupt-controller@fffff100 {
> >> interrupt-controller;
> >> reg = <0xfffff100 0x100>;
> >> atmel,external-irqs = <31>;
> >> + microchip,nr-irqs = <50>;
> >> };
> >>
> >> dbgu: serial@fffff200 {
> >> diff --git a/drivers/irqchip/irq-atmel-aic5.c b/drivers/irqchip/irq-atmel-aic5.c
> >> index 145535bd7560..5d96ad8860d3 100644
> >> --- a/drivers/irqchip/irq-atmel-aic5.c
> >> +++ b/drivers/irqchip/irq-atmel-aic5.c
> >> @@ -398,11 +398,16 @@ static int __init sama5d4_aic5_of_init(struct device_node *node,
> >> }
> >> IRQCHIP_DECLARE(sama5d4_aic5, "atmel,sama5d4-aic", sama5d4_aic5_of_init);
> >>
> >> -#define NR_SAM9X60_IRQS 50
> >> -
> >> static int __init sam9x60_aic5_of_init(struct device_node *node,
> >> struct device_node *parent)
> >> {
> >> - return aic5_of_init(node, parent, NR_SAM9X60_IRQS);
> >> + int ret, nr_irqs;
> >> +
> >> + ret = of_property_read_u32(node, "microchip,nr-irqs", &nr_irqs);
> >> + if (ret) {
> >> + pr_err("Not found microchip,nr-irqs property\n");
> >
> > This breaks the ABI. You should ensure old device trees are still working
> > with this patch.
>
> The only older device that uses this API is sam9x60 and the newly added
> sam9x7. This change has been tested to be working fine in both the
> devices.

Does it still work for a sam9x60 that does not have "microchip,nr-irqs"?
I can't see how it would, because you remove the define and return an
error. That's a pretty clear ABI breakage to me and I don't think it is
justified.

> If you still want me to use the macros as a fallback in the
> failure case I can do it. But this change was proposed to avoid adding
> macros in the first place. I can remove the error check just like they
> do while getting other device tree properties. Or if this is just a
> concern of the old devices working with the new change, then sam9x60
> works. Please let me know how to proceed.

I just noticed that this property is not documented in a binding. The
first thing you would will be asked when trying to add that is "why can
this not be determined based on the compatible", which means back to
having a define in the driver.
That said, having specific $soc_aic5_of_init() functions for each SoC
seems silly when usually only the number of interrupts changes. The
number of IRQs could be in the match data and you could use
aic5_of_init in your IRQCHIP_DECLARE directly.

> >> + return ret;
> >> + }
> >> + return aic5_of_init(node, parent, nr_irqs);
> >> }
> >> IRQCHIP_DECLARE(sam9x60_aic5, "microchip,sam9x60-aic", sam9x60_aic5_of_init);
>
> --
> Thanks and Regards,
> Varshini Rajendran.
>

Attachment: signature.asc
Description: PGP signature