Hi Marc
Thanks for your suggestions, I'm working on v4 and I would like to ask
if it is better to have a driver for only one irqchip
and create dt nodes for each chip, or just register all the chips in a
single driver with only one dt node.
在 2019/1/24 下午5:54, Marc Zyngier 写道:
On Thu, 24 Jan 2019 03:27:29 +0000,
Jiaxun Yang <jiaxun.yang@xxxxxxxxxxx> wrote:
This controller appeared on Loongson-1 family MCUs
including Loongson-1B and Loongson-1C.
Signed-off-by: Jiaxun Yang <jiaxun.yang@xxxxxxxxxxx>
---
drivers/irqchip/Kconfig | 9 ++
drivers/irqchip/Makefile | 1 +
drivers/irqchip/irq-ls1x.c | 176 +++++++++++++++++++++++++++++++++++++
3 files changed, 186 insertions(+)
create mode 100644 drivers/irqchip/irq-ls1x.c
diff --git a/drivers/irqchip/Kconfig b/drivers/irqchip/Kconfig
index 3d1e60779078..5dcb5456cd14 100644
--- a/drivers/irqchip/Kconfig
+++ b/drivers/irqchip/Kconfig
@@ -406,6 +406,15 @@ config IMX_IRQSTEER
help
Support for the i.MX IRQSTEER interrupt multiplexer/remapper.
+config LS1X_IRQ
+ bool "Loongson-1 Interrupt Controller"
+ depends on MACH_LOONGSON32
+ default y
+ select IRQ_DOMAIN
+ select GENERIC_IRQ_CHIP
+ help
+ Support for the Loongson-1 platform Interrupt Controller.
+
endmenu
config SIFIVE_PLIC
diff --git a/drivers/irqchip/Makefile b/drivers/irqchip/Makefile
index c93713d24b86..7acd0e36d0b4 100644
--- a/drivers/irqchip/Makefile
+++ b/drivers/irqchip/Makefile
@@ -94,3 +94,4 @@ obj-$(CONFIG_CSKY_APB_INTC) += irq-csky-apb-intc.o
obj-$(CONFIG_SIFIVE_PLIC) += irq-sifive-plic.o
obj-$(CONFIG_IMX_IRQSTEER) += irq-imx-irqsteer.o
obj-$(CONFIG_MADERA_IRQ) += irq-madera.o
+obj-$(CONFIG_LS1X_IRQ) += irq-ls1x.o
diff --git a/drivers/irqchip/irq-ls1x.c b/drivers/irqchip/irq-ls1x.c
new file mode 100644
index 000000000000..de92ca04cf9f
--- /dev/null
+++ b/drivers/irqchip/irq-ls1x.c
@@ -0,0 +1,176 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2019, Jiaxun Yang <jiaxun.yang@xxxxxxxxxxx>
+ * Loongson-1 platform IRQ support
+ */
+
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/interrupt.h>
+#include <linux/ioport.h>
+#include <linux/irqchip.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/io.h>
+#include <linux/irqchip/chained_irq.h>
+
+#include <asm/mach-loongson32/irq.h>
+
+
+#define MAX_CHIPS 5
+#define LS_REG_INTC_STATUS 0x00
+#define LS_REG_INTC_EN 0x04
+#define LS_REG_INTC_SET 0x08
+#define LS_REG_INTC_CLR 0x0c
+#define LS_REG_INTC_POL 0x10
+#define LS_REG_INTC_EDGE 0x14
+#define CHIP_SIZE 0x18
+
+static void ls_chained_handle_irq(struct irq_desc *desc)
+{
+ struct irq_chip_generic *gc = irq_desc_get_handler_data(desc);
+ struct irq_chip *chip = irq_desc_get_chip(desc);
+ u32 pending;
+
+ chained_irq_enter(chip, desc);
+ pending = readl(gc->reg_base + LS_REG_INTC_STATUS) &
+ readl(gc->reg_base + LS_REG_INTC_EN);
+
+ if (!pending) {
+ spurious_interrupt();
+ chained_irq_exit(chip, desc);
+ return;
+ }
Given the context, this is the same as writing:
if (!pending)
spurious_interrupt();
and let it go through.
+
+ while (pending) {
+ int bit = __ffs(pending);
+
+ generic_handle_irq(gc->irq_base + bit);
+ pending &= ~BIT(bit);
+ }
+
+ chained_irq_exit(chip, desc);
+}
+
+static void ls_intc_set_bit(struct irq_chip_generic *gc,
+ unsigned int offset,
+ u32 mask, bool set)
+{
+ if (set)
+ writel(readl(gc->reg_base + offset) | mask,
+ gc->reg_base + offset);
+ else
+ writel(readl(gc->reg_base + offset) & ~mask,
+ gc->reg_base + offset);
+}
+
+static int ls_intc_set_type(struct irq_data *data, unsigned int type)
+{
+ struct irq_chip_generic *gc = irq_data_get_irq_chip_data(data);
+ u32 mask = data->mask;
+
+ switch (type) {
+ case IRQ_TYPE_LEVEL_HIGH:
+ ls_intc_set_bit(gc, LS_REG_INTC_EDGE, mask, false);
+ ls_intc_set_bit(gc, LS_REG_INTC_POL, mask, true);
+ break;
+ case IRQ_TYPE_LEVEL_LOW:
+ ls_intc_set_bit(gc, LS_REG_INTC_EDGE, mask, false);
+ ls_intc_set_bit(gc, LS_REG_INTC_POL, mask, false);
+ break;
+ case IRQ_TYPE_EDGE_RISING:
+ ls_intc_set_bit(gc, LS_REG_INTC_EDGE, mask, true);
+ ls_intc_set_bit(gc, LS_REG_INTC_POL, mask, true);
+ break;
+ case IRQ_TYPE_EDGE_FALLING:
+ ls_intc_set_bit(gc, LS_REG_INTC_EDGE, mask, true);
+ ls_intc_set_bit(gc, LS_REG_INTC_POL, mask, false);
+ break;
+ case IRQ_TYPE_NONE:
+ break;
+ default:
+ return -EINVAL;
+ }
I just realised that in your DT binding, you define the interrupt
specifier as having a single cell. Which means that you never describe
the interrupt trigger there. Why?
Does it mean that all the drivers have to configure their interrupt
trigger themselves, and cannot rely on DT to do it? This seems quite
backward.
My fault, will fix it later.
+
+ return 0;
+}
+
+
+static int __init ls_intc_of_init(struct device_node *node,
+ struct device_node *parent)
+{
+ struct irq_chip_generic *gc[MAX_CHIPS];
+ struct irq_chip_type *ct;
+ struct irq_domain *domain;
+ void __iomem *base;
+ int parent_irq[MAX_CHIPS], err = 0;
+ unsigned int i = 0;
+
+ unsigned int num_chips = of_irq_count(node);
+
+ base = of_iomap(node, 0);
+ if (!base)
+ return -ENODEV;
+
+ for (i = 0; i < num_chips; i++) {
+ /* Mask all irqs */
+ writel(0x0, base + (i * CHIP_SIZE) +
+ LS_REG_INTC_EN);
+
+ /* Set all irqs to high level triggered */
+ writel(0xffffffff, base + (i * CHIP_SIZE) +
+ LS_REG_INTC_POL);
+
+ parent_irq[i] = irq_of_parse_and_map(node, i);
+
+ if (!parent_irq[i]) {
+ pr_warn("unable to get parent irq for irqchip %u\n", i);
+ goto out_err;
+ }
+
+ gc[i] = irq_alloc_generic_chip("INTC", 1,
+ LS1X_IRQ_BASE + (i * 32),
+ base + (i * CHIP_SIZE),
+ handle_level_irq);
+
+ ct = gc[i]->chip_types;
+ ct->regs.mask = LS_REG_INTC_EN;
+ ct->regs.ack = LS_REG_INTC_CLR;
+ ct->chip.irq_unmask = irq_gc_mask_set_bit;
+ ct->chip.irq_mask = irq_gc_mask_clr_bit;
+ ct->chip.irq_ack = irq_gc_ack_set_bit;
+ ct->chip.irq_set_type = ls_intc_set_type;
+
+ irq_setup_generic_chip(gc[i], IRQ_MSK(32), 0, 0,
+ IRQ_NOPROBE | IRQ_LEVEL);
+ }
+
+ domain = irq_domain_add_legacy(node, num_chips * 32, LS1X_IRQ_BASE, 0,
+ &irq_domain_simple_ops, NULL);
Why a legacy domain? This is usually reserved to old drivers that are
converted to a new infrastructure, while needing some form of platform
hacks. I don't see this being the case here.
It is also worrying that although you have up to 5 irqchips, they all
share a single domain. What does this mean? each irqchip is expected
to have its own domain.
Yes, I do like this for backward compatible reason. I'm turning
a legacy platform device mach(arch/mips/loongson32) in to
dt based generic mach and I would like to do it step by step rather than
one time.
So I use legacy domain in order to keep IRQ same with the
old driver exist on arch/mips/loongson32/common/irq.c
+ if (!domain) {
+ pr_warn("unable to register IRQ domain\n");
+ err = -EINVAL;
+ goto out_err;
+ }
+
+ for (i = 0; i < num_chips; i++)
+ irq_set_chained_handler_and_data(parent_irq[i],
+ ls_chained_handle_irq, gc[i]);
+
+ pr_info("ls1x-irq: %u chips registered\n", num_chips);
+
+ return 0;
+
+out_err:
+ for (i = 0; i < MAX_CHIPS; i++) {
+ if (gc[i])
But you've never initialised gc[], nor parent_irq[]. So you'll get
whatever the stack had at this point. Not great.
+ irq_destroy_generic_chip(gc[i], IRQ_MSK(32),
+ IRQ_NOPROBE | IRQ_LEVEL, 0);
+ if (parent_irq[i])
+ irq_dispose_mapping(parent_irq[i]);
+ }
+ return err;
+}
+
+IRQCHIP_DECLARE(ls1x_intc, "loongson,ls1x-intc", ls_intc_of_init);
--
2.20.1
Thanks,
M.
