RE: [PATCH v6 2/5] irqchip/irq-goldfish-pic: Add Goldfish PIC driver
From: Miodrag Dinic
Date: Wed Nov 01 2017 - 10:37:05 EST
Hello Marc,
> > + gc = irq_alloc_generic_chip("GFPIC", 1, GFPIC_IRQ_BASE, gfpic->base,
> > + handle_level_irq);
> > +
> > + ct = gc->chip_types;
>
> And what if the allocation fails?
Thanks, it will be addressed in V7.
Kind regards,
Miodrag
________________________________________
From: Marc Zyngier [marc.zyngier@xxxxxxx]
Sent: Tuesday, October 31, 2017 3:26 AM
To: Aleksandar Markovic
Cc: linux-mips@xxxxxxxxxxxxxx; Miodrag Dinic; Goran Ferenc; Aleksandar Markovic; David S. Miller; Douglas Leung; Greg Kroah-Hartman; James Hogan; Jason Cooper; linux-kernel@xxxxxxxxxxxxxxx; Mauro Carvalho Chehab; Miodrag Dinic; Paul Burton; Petar Jovanovic; Raghu Gandham; Randy Dunlap; Thomas Gleixner
Subject: Re: [PATCH v6 2/5] irqchip/irq-goldfish-pic: Add Goldfish PIC driver
On Mon, Oct 30 2017 at 12:56:33 pm GMT, Aleksandar Markovic <aleksandar.markovic@xxxxxxxxx> wrote:
> From: Miodrag Dinic <miodrag.dinic@xxxxxxxx>
>
> Add device driver for a virtual programmable interrupt controller
>
> The virtual PIC is designed as a device tree-based interrupt controller.
>
> The compatible string used by OS for binding the driver is
> "google,goldfish-pic".
>
> Signed-off-by: Miodrag Dinic <miodrag.dinic@xxxxxxxx>
> Signed-off-by: Goran Ferenc <goran.ferenc@xxxxxxxx>
> Signed-off-by: Aleksandar Markovic <aleksandar.markovic@xxxxxxxx>
[...]
> diff --git a/drivers/irqchip/irq-goldfish-pic.c b/drivers/irqchip/irq-goldfish-pic.c
> new file mode 100644
> index 0000000..48fb773
> --- /dev/null
> +++ b/drivers/irqchip/irq-goldfish-pic.c
[...]
> +static int __init goldfish_pic_of_init(struct device_node *of_node,
> + struct device_node *parent)
> +{
> + struct goldfish_pic_data *gfpic;
> + struct irq_chip_generic *gc;
> + struct irq_chip_type *ct;
> + unsigned int parent_irq;
> + int ret = 0;
> +
> + gfpic = kzalloc(sizeof(*gfpic), GFP_KERNEL);
> + if (!gfpic) {
> + ret = -ENOMEM;
> + goto out_err;
> + }
> +
> + parent_irq = irq_of_parse_and_map(of_node, 0);
> + if (!parent_irq) {
> + pr_err("Failed to map Goldfish PIC parent IRQ\n");
> + ret = -EINVAL;
> + goto out_free;
> + }
> +
> + gfpic->base = of_iomap(of_node, 0);
> + if (!gfpic->base) {
> + pr_err("Failed to map Goldfish PIC base\n");
> + ret = -ENOMEM;
> + goto out_unmap_irq;
> + }
> +
> + /* Mask interrupts. */
> + writel(1, gfpic->base + GFPIC_REG_IRQ_DISABLE_ALL);
> +
> + gc = irq_alloc_generic_chip("GFPIC", 1, GFPIC_IRQ_BASE, gfpic->base,
> + handle_level_irq);
> +
> + ct = gc->chip_types;
And what if the allocation fails?
> + ct->regs.enable = GFPIC_REG_IRQ_ENABLE;
> + ct->regs.disable = GFPIC_REG_IRQ_DISABLE;
> + ct->chip.irq_unmask = irq_gc_unmask_enable_reg;
> + ct->chip.irq_mask = irq_gc_mask_disable_reg;
> +
> + irq_setup_generic_chip(gc, IRQ_MSK(GFPIC_NR_IRQS), 0, 0,
> + IRQ_NOPROBE | IRQ_LEVEL);
> +
> + gfpic->irq_domain = irq_domain_add_legacy(of_node, GFPIC_NR_IRQS,
> + GFPIC_IRQ_BASE, 0,
> + &goldfish_irq_domain_ops,
> + NULL);
> + if (!gfpic->irq_domain) {
> + pr_err("Failed to add irqdomain for Goldfish PIC\n");
> + ret = -EINVAL;
> + goto out_iounmap;
> + }
> +
> + irq_set_chained_handler_and_data(parent_irq,
> + goldfish_pic_cascade, gfpic);
> +
> + pr_info("Successfully registered Goldfish PIC\n");
> + return 0;
> +
> +out_iounmap:
> + iounmap(gfpic->base);
> +out_unmap_irq:
> + irq_dispose_mapping(parent_irq);
> +out_free:
> + kfree(gfpic);
> +out_err:
> + return ret;
> +}
> +
> +IRQCHIP_DECLARE(google_gf_pic, "google,goldfish-pic", goldfish_pic_of_init);
Thanks,
M.
--
Jazz is not dead. It just smells funny.