[1/6,v2] gpio: Add support for hierarchical IRQ domains
diff mbox series

Message ID 20190808123242.5359-1-linus.walleij@linaro.org
State New
Headers show
Series
  • [1/6,v2] gpio: Add support for hierarchical IRQ domains
Related show

Commit Message

Linus Walleij Aug. 8, 2019, 12:32 p.m. UTC
Hierarchical IRQ domains can be used to stack different IRQ
controllers on top of each other.

Bring hierarchical IRQ domains into the GPIOLIB core with the
following basic idea:

Drivers that need their interrupts handled hierarchically
specify a callback to translate the child hardware IRQ and
IRQ type for each GPIO offset to a parent hardware IRQ and
parent hardware IRQ type.

Users have to pass the callback, fwnode, and parent irqdomain
before calling gpiochip_irqchip_add().

We use the new method of just filling in the struct
gpio_irq_chip before adding the gpiochip for all hierarchical
irqchips of this type.

The code path for device tree is pretty straight-forward,
while the code path for old boardfiles or anything else will
be more convoluted requireing upfront allocation of the
interrupts when adding the chip.

One specific use-case where this can be useful is if a power
management controller has top-level controls for wakeup
interrupts. In such cases, the power management controller can
be a parent to other interrupt controllers and program
additional registers when an IRQ has its wake capability
enabled or disabled.

The hierarchical irqchip helper code will only be available
when IRQ_DOMAIN_HIERARCHY is selected to GPIO chips using
this should select or depend on that symbol. When using
hierarchical IRQs, the parent interrupt controller must
also be hierarchical all the way up to the top interrupt
controller wireing directly into the CPU, so on systems
that do not have this we can get rid of all the extra
code for supporting hierarchical irqs.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Cc: Lina Iyer <ilina@codeaurora.org>
Cc: Jon Hunter <jonathanh@nvidia.com>
Cc: Sowjanya Komatineni <skomatineni@nvidia.com>
Cc: Bitan Biswas <bbiswas@nvidia.com>
Cc: linux-tegra@vger.kernel.org
Cc: David Daney <david.daney@cavium.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Brian Masney <masneyb@onstation.org>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Brian Masney <masneyb@onstation.org>
Co-developed-by: Brian Masney <masneyb@onstation.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
ChangeLog v1->v2:
- Squash in changes contributed by Brian Masney to support
  customization and solve some bugs.
- Add gpiochip_populate_parent_fwspec_{two,four}cell functions
  so that irqchips using four cells can use this code.
- Pass girq->handler to the domain as default handler:
  there are apparently cases where handle_bad_irq will
  not quite cut it.
- Change "child_pin_to_irq" into "child_offset_to_irq" so
  we avoid confusion with pin control.
- Diet down the parent domain presence checks, one time
  is enough to check it.
- Change domain allocation error to -ENOMEM and do not
  print errors.
- Diet some chip_info() blather.
- Assume nr_irqs to be 1 in .translate() and warn if not.
ChangeLog RFC->v1:
- Tested on real hardware
- Incorporate Thierry's idea to have a translation callback.
  He was right about this approach, I was wrong in insisting
  on IRQ maps.
---
 Documentation/driver-api/gpio/driver.rst | 122 ++++++++-
 drivers/gpio/gpiolib.c                   | 318 ++++++++++++++++++++++-
 include/linux/gpio/driver.h              | 111 ++++++++
 3 files changed, 525 insertions(+), 26 deletions(-)

Comments

Linus Walleij Aug. 14, 2019, 8:18 a.m. UTC | #1
On Thu, Aug 8, 2019 at 2:32 PM Linus Walleij <linus.walleij@linaro.org> wrote:

> Hierarchical IRQ domains can be used to stack different IRQ
> controllers on top of each other.

This is now applied so we get a solid base for testing in linux-next.
Address further concerns with incremental patches, either written
by yourselves or by poking me to do them :)

Yours,
Linus Walleij
Brian Masney Aug. 16, 2019, 1:10 a.m. UTC | #2
Hi Linus,

On Thu, Aug 08, 2019 at 02:32:37PM +0200, Linus Walleij wrote:
> Hierarchical IRQ domains can be used to stack different IRQ
> controllers on top of each other.
> 
> Bring hierarchical IRQ domains into the GPIOLIB core with the
> following basic idea:
> 
> Drivers that need their interrupts handled hierarchically
> specify a callback to translate the child hardware IRQ and
> IRQ type for each GPIO offset to a parent hardware IRQ and
> parent hardware IRQ type.
> 
> Users have to pass the callback, fwnode, and parent irqdomain
> before calling gpiochip_irqchip_add().
> 
> We use the new method of just filling in the struct
> gpio_irq_chip before adding the gpiochip for all hierarchical
> irqchips of this type.
> 
> The code path for device tree is pretty straight-forward,
> while the code path for old boardfiles or anything else will
> be more convoluted requireing upfront allocation of the
> interrupts when adding the chip.
> 
> One specific use-case where this can be useful is if a power
> management controller has top-level controls for wakeup
> interrupts. In such cases, the power management controller can
> be a parent to other interrupt controllers and program
> additional registers when an IRQ has its wake capability
> enabled or disabled.
> 
> The hierarchical irqchip helper code will only be available
> when IRQ_DOMAIN_HIERARCHY is selected to GPIO chips using
> this should select or depend on that symbol. When using
> hierarchical IRQs, the parent interrupt controller must
> also be hierarchical all the way up to the top interrupt
> controller wireing directly into the CPU, so on systems
> that do not have this we can get rid of all the extra
> code for supporting hierarchical irqs.
> 
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Marc Zyngier <marc.zyngier@arm.com>
> Cc: Lina Iyer <ilina@codeaurora.org>
> Cc: Jon Hunter <jonathanh@nvidia.com>
> Cc: Sowjanya Komatineni <skomatineni@nvidia.com>
> Cc: Bitan Biswas <bbiswas@nvidia.com>
> Cc: linux-tegra@vger.kernel.org
> Cc: David Daney <david.daney@cavium.com>
> Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
> Cc: Brian Masney <masneyb@onstation.org>
> Signed-off-by: Thierry Reding <treding@nvidia.com>
> Signed-off-by: Brian Masney <masneyb@onstation.org>
> Co-developed-by: Brian Masney <masneyb@onstation.org>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> ---

[ snip ]


> @@ -1827,10 +2099,23 @@ EXPORT_SYMBOL_GPL(gpiochip_irq_domain_deactivate);
>  
>  static int gpiochip_to_irq(struct gpio_chip *chip, unsigned offset)

   ^^^^^^

I started to convert ssbi-gpio over to this and pm8xxx_gpio_to_irq() has
this little snippet that's different from spmi-gpio:

	[ fwspec mapping code ]

	/*
	 * Cache the IRQ since pm8xxx_gpio_get() needs this to get determine the
	 * line level.
	 */
	pin->irq = ret;

Here's the relevant code in pm8xxx_gpio_get():

	if (pin->mode == PM8XXX_GPIO_MODE_OUTPUT) {
		ret = pin->output_value;
	} else if (pin->irq >= 0) {
		ret = irq_get_irqchip_state(pin->irq, IRQCHIP_STATE_LINE_LEVEL, &state);
		...
	}

What do you think about using EXPORT_SYMBOL_GPL() for gpiochip_to_irq() so
that we can call it in pm8xxx_gpio_to_irq()? Or do you have any other
suggestions for how we can get rid of that IRQ cache?

I don't see any other issues for ssbi-gpio.

Brian
Linus Walleij Aug. 23, 2019, 8:24 a.m. UTC | #3
Hi Brian,

I tried to look into this ssbi-gpio problem...

Paging in Marc Z as well.

On Fri, Aug 16, 2019 at 3:10 AM Brian Masney <masneyb@onstation.org> wrote:

> I started to convert ssbi-gpio over to this and pm8xxx_gpio_to_irq() has
> this little snippet that's different from spmi-gpio:
>
>         [ fwspec mapping code ]
>
>         /*
>          * Cache the IRQ since pm8xxx_gpio_get() needs this to get determine the
>          * line level.
>          */
>         pin->irq = ret;
>
> Here's the relevant code in pm8xxx_gpio_get():
>
>         if (pin->mode == PM8XXX_GPIO_MODE_OUTPUT) {
>                 ret = pin->output_value;
>         } else if (pin->irq >= 0) {
>                 ret = irq_get_irqchip_state(pin->irq, IRQCHIP_STATE_LINE_LEVEL, &state);
>                 ...
>         }

It's a bit annoying that this API (irq_get_irqchip_state()) is relying on
the global irq numbers and the internal function using irqdesc
__irq_get_irqchip_state() is not exported.

We should be encouraged to operate on IRQ descriptors rather
than IRQ numbers when doing this kind of deep core work,
right?

> What do you think about using EXPORT_SYMBOL_GPL() for gpiochip_to_irq() so
> that we can call it in pm8xxx_gpio_to_irq()?

I would like to avoid that because people tend to abuse every
API I expose (leasson learnt: APIs shall be narrow and deep).

> Or do you have any other
> suggestions for how we can get rid of that IRQ cache?

So the issue is that moving this to the hierarchical helpers
rids pm8xxx_gpio_to_irq() and we needed that to map this.

I would rather just add a new driver API to wrap the irqchip
API:

ret = gpiochip_get_irq_state(offset, IRQCHIP_STATE_LINE_LEVEL, &state);

Where int gpiochip_get_own_irq_state(int offset, ...) will figure
out the gpio_desc of the offset and then call gpiod_to_irq()
and then use that irq number to call irq_get_irqchip_state()
and the goal is met.

This should work I think, and expose a more precise
API for what this driver wants.

> I don't see any other issues for ssbi-gpio.

That's good news!

Yours,
Linus Walleij
Marc Zyngier Aug. 23, 2019, 9:12 a.m. UTC | #4
On 23/08/2019 09:24, Linus Walleij wrote:
> Hi Brian,
> 
> I tried to look into this ssbi-gpio problem...
> 
> Paging in Marc Z as well.
> 
> On Fri, Aug 16, 2019 at 3:10 AM Brian Masney <masneyb@onstation.org> wrote:
> 
>> I started to convert ssbi-gpio over to this and pm8xxx_gpio_to_irq() has
>> this little snippet that's different from spmi-gpio:
>>
>>         [ fwspec mapping code ]
>>
>>         /*
>>          * Cache the IRQ since pm8xxx_gpio_get() needs this to get determine the
>>          * line level.
>>          */
>>         pin->irq = ret;
>>
>> Here's the relevant code in pm8xxx_gpio_get():
>>
>>         if (pin->mode == PM8XXX_GPIO_MODE_OUTPUT) {
>>                 ret = pin->output_value;
>>         } else if (pin->irq >= 0) {
>>                 ret = irq_get_irqchip_state(pin->irq, IRQCHIP_STATE_LINE_LEVEL, &state);
>>                 ...
>>         }
> 
> It's a bit annoying that this API (irq_get_irqchip_state()) is relying on
> the global irq numbers and the internal function using irqdesc
> __irq_get_irqchip_state() is not exported.
> 
> We should be encouraged to operate on IRQ descriptors rather
> than IRQ numbers when doing this kind of deep core work,
> right?

Certainly, I'd like to gradually move over from the IRQ number to an
irq_desc. In interrupt-heavy contexts, this ends up being quite time
consuming. I have an old patch series somewhere changing irq domains to
use irq_descs internally instead of IRQ numbers, which I should maybe
revive.

As for __irq_get_irqchip_state, the main issue is that it doesn't
perform any locking on its own, so you'd have to either provide your
own, or wrap it with something else.

Thanks,

	M.
Linus Walleij Aug. 23, 2019, 10:13 a.m. UTC | #5
On Fri, Aug 23, 2019 at 11:12 AM Marc Zyngier <marc.zyngier@arm.com> wrote:

> > We should be encouraged to operate on IRQ descriptors rather
> > than IRQ numbers when doing this kind of deep core work,
> > right?
>
> Certainly, I'd like to gradually move over from the IRQ number to an
> irq_desc. In interrupt-heavy contexts, this ends up being quite time
> consuming. I have an old patch series somewhere changing irq domains to
> use irq_descs internally instead of IRQ numbers, which I should maybe
> revive.

We currently interact most heavily with the irqchip by way of
mapping GPIO lines to the corresponding IRQs, so I suppose
the existing

int gpiod_to_irq(struct gpio_desc *desc);

Need a corresponding

struct irq_desc *gpiod_to_irq_desc(struct gpio_desc *desc);

at some point, and then we can start to de-pollute the kernel
from this for all drivers using the modern GPIO descriptor
API. It's a big task but can certainly be done with some
help.

On the driver back-end we are in surprisingly good shape
thanks to all the abstractions provided for dealing with IRQ
chip drivers.

Yours,
Linus Walleij

Patch
diff mbox series

diff --git a/Documentation/driver-api/gpio/driver.rst b/Documentation/driver-api/gpio/driver.rst
index 921c71a3d683..5e96dbc7f4e7 100644
--- a/Documentation/driver-api/gpio/driver.rst
+++ b/Documentation/driver-api/gpio/driver.rst
@@ -259,7 +259,7 @@  most often cascaded off a parent interrupt controller, and in some special
 cases the GPIO logic is melded with a SoC's primary interrupt controller.
 
 The IRQ portions of the GPIO block are implemented using an irq_chip, using
-the header <linux/irq.h>. So basically such a driver is utilizing two sub-
+the header <linux/irq.h>. So this combined driver is utilizing two sub-
 systems simultaneously: gpio and irq.
 
 It is legal for any IRQ consumer to request an IRQ from any irqchip even if it
@@ -391,25 +391,119 @@  Infrastructure helpers for GPIO irqchips
 ----------------------------------------
 
 To help out in handling the set-up and management of GPIO irqchips and the
-associated irqdomain and resource allocation callbacks, the gpiolib has
-some helpers that can be enabled by selecting the GPIOLIB_IRQCHIP Kconfig
-symbol:
-
-- gpiochip_irqchip_add(): adds a chained cascaded irqchip to a gpiochip. It
-  will pass the struct gpio_chip* for the chip to all IRQ callbacks, so the
-  callbacks need to embed the gpio_chip in its state container and obtain a
-  pointer to the container using container_of().
-  (See Documentation/driver-api/driver-model/design-patterns.rst)
+associated irqdomain and resource allocation callbacks. These are activated
+by selecting the Kconfig symbol GPIOLIB_IRQCHIP. If the symbol
+IRQ_DOMAIN_HIERARCHY is also selected, hierarchical helpers will also be
+provided. A big portion of overhead code will be managed by gpiolib,
+under the assumption that your interrupts are 1-to-1-mapped to the
+GPIO line index:
+
+  GPIO line offset   Hardware IRQ
+  0                  0
+  1                  1
+  2                  2
+  ...                ...
+  ngpio-1            ngpio-1
+
+If some GPIO lines do not have corresponding IRQs, the bitmask valid_mask
+and the flag need_valid_mask in gpio_irq_chip can be used to mask off some
+lines as invalid for associating with IRQs.
+
+The preferred way to set up the helpers is to fill in the
+struct gpio_irq_chip inside struct gpio_chip before adding the gpio_chip.
+If you do this, the additional irq_chip will be set up by gpiolib at the
+same time as setting up the rest of the GPIO functionality. The following
+is a typical example of a cascaded interrupt handler using gpio_irq_chip:
+
+  /* Typical state container with dynamic irqchip */
+  struct my_gpio {
+      struct gpio_chip gc;
+      struct irq_chip irq;
+  };
+
+  int irq; /* from platform etc */
+  struct my_gpio *g;
+  struct gpio_irq_chip *girq;
+
+  /* Set up the irqchip dynamically */
+  g->irq.name = "my_gpio_irq";
+  g->irq.irq_ack = my_gpio_ack_irq;
+  g->irq.irq_mask = my_gpio_mask_irq;
+  g->irq.irq_unmask = my_gpio_unmask_irq;
+  g->irq.irq_set_type = my_gpio_set_irq_type;
+
+  /* Get a pointer to the gpio_irq_chip */
+  girq = &g->gc.irq;
+  girq->chip = &g->irq;
+  girq->parent_handler = ftgpio_gpio_irq_handler;
+  girq->num_parents = 1;
+  girq->parents = devm_kcalloc(dev, 1, sizeof(*girq->parents),
+                               GFP_KERNEL);
+  if (!girq->parents)
+      return -ENOMEM;
+  girq->default_type = IRQ_TYPE_NONE;
+  girq->handler = handle_bad_irq;
+  girq->parents[0] = irq;
+
+  return devm_gpiochip_add_data(dev, &g->gc, g);
+
+The helper support using hierarchical interrupt controllers as well.
+In this case the typical set-up will look like this:
+
+  /* Typical state container with dynamic irqchip */
+  struct my_gpio {
+      struct gpio_chip gc;
+      struct irq_chip irq;
+      struct fwnode_handle *fwnode;
+  };
+
+  int irq; /* from platform etc */
+  struct my_gpio *g;
+  struct gpio_irq_chip *girq;
+
+  /* Set up the irqchip dynamically */
+  g->irq.name = "my_gpio_irq";
+  g->irq.irq_ack = my_gpio_ack_irq;
+  g->irq.irq_mask = my_gpio_mask_irq;
+  g->irq.irq_unmask = my_gpio_unmask_irq;
+  g->irq.irq_set_type = my_gpio_set_irq_type;
+
+  /* Get a pointer to the gpio_irq_chip */
+  girq = &g->gc.irq;
+  girq->chip = &g->irq;
+  girq->default_type = IRQ_TYPE_NONE;
+  girq->handler = handle_bad_irq;
+  girq->fwnode = g->fwnode;
+  girq->parent_domain = parent;
+  girq->child_to_parent_hwirq = my_gpio_child_to_parent_hwirq;
+
+  return devm_gpiochip_add_data(dev, &g->gc, g);
+
+As you can see pretty similar, but you do not supply a parent handler for
+the IRQ, instead a parent irqdomain, an fwnode for the hardware and
+a funcion .child_to_parent_hwirq() that has the purpose of looking up
+the parent hardware irq from a child (i.e. this gpio chip) hardware irq.
+As always it is good to look at examples in the kernel tree for advice
+on how to find the required pieces.
+
+The old way of adding irqchips to gpiochips after registration is also still
+available but we try to move away from this:
+
+- DEPRECATED: gpiochip_irqchip_add(): adds a chained cascaded irqchip to a
+  gpiochip. It will pass the struct gpio_chip* for the chip to all IRQ
+  callbacks, so the callbacks need to embed the gpio_chip in its state
+  container and obtain a pointer to the container using container_of().
+  (See Documentation/driver-model/design-patterns.txt)
 
 - gpiochip_irqchip_add_nested(): adds a nested cascaded irqchip to a gpiochip,
   as discussed above regarding different types of cascaded irqchips. The
   cascaded irq has to be handled by a threaded interrupt handler.
   Apart from that it works exactly like the chained irqchip.
 
-- gpiochip_set_chained_irqchip(): sets up a chained cascaded irq handler for a
-  gpio_chip from a parent IRQ and passes the struct gpio_chip* as handler
-  data. Notice that we pass is as the handler data, since the irqchip data is
-  likely used by the parent irqchip.
+- DEPRECATED: gpiochip_set_chained_irqchip(): sets up a chained cascaded irq
+  handler for a gpio_chip from a parent IRQ and passes the struct gpio_chip*
+  as handler data. Notice that we pass is as the handler data, since the
+  irqchip data is likely used by the parent irqchip.
 
 - gpiochip_set_nested_irqchip(): sets up a nested cascaded irq handler for a
   gpio_chip from a parent IRQ. As the parent IRQ has usually been
diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index 537a37a89891..4a71a58ce47f 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -1720,6 +1720,273 @@  void gpiochip_set_nested_irqchip(struct gpio_chip *gpiochip,
 }
 EXPORT_SYMBOL_GPL(gpiochip_set_nested_irqchip);
 
+#ifdef CONFIG_IRQ_DOMAIN_HIERARCHY
+
+/**
+ * gpiochip_set_hierarchical_irqchip() - connects a hierarchical irqchip
+ * to a gpiochip
+ * @gc: the gpiochip to set the irqchip hierarchical handler to
+ * @irqchip: the irqchip to handle this level of the hierarchy, the interrupt
+ * will then percolate up to the parent
+ */
+static void gpiochip_set_hierarchical_irqchip(struct gpio_chip *gc,
+					      struct irq_chip *irqchip)
+{
+	/* DT will deal with mapping each IRQ as we go along */
+	if (is_of_node(gc->irq.fwnode))
+		return;
+
+	/*
+	 * This is for legacy and boardfile "irqchip" fwnodes: allocate
+	 * irqs upfront instead of dynamically since we don't have the
+	 * dynamic type of allocation that hardware description languages
+	 * provide. Once all GPIO drivers using board files are gone from
+	 * the kernel we can delete this code, but for a transitional period
+	 * it is necessary to keep this around.
+	 */
+	if (is_fwnode_irqchip(gc->irq.fwnode)) {
+		int i;
+		int ret;
+
+		for (i = 0; i < gc->ngpio; i++) {
+			struct irq_fwspec fwspec;
+			unsigned int parent_hwirq;
+			unsigned int parent_type;
+			struct gpio_irq_chip *girq = &gc->irq;
+
+			/*
+			 * We call the child to parent translation function
+			 * only to check if the child IRQ is valid or not.
+			 * Just pick the rising edge type here as that is what
+			 * we likely need to support.
+			 */
+			ret = girq->child_to_parent_hwirq(gc, i,
+							  IRQ_TYPE_EDGE_RISING,
+							  &parent_hwirq,
+							  &parent_type);
+			if (ret) {
+				chip_err(gc, "skip set-up on hwirq %d\n",
+					 i);
+				continue;
+			}
+
+			fwspec.fwnode = gc->irq.fwnode;
+			/* This is the hwirq for the GPIO line side of things */
+			fwspec.param[0] = girq->child_offset_to_irq(gc, i);
+			/* Just pick something */
+			fwspec.param[1] = IRQ_TYPE_EDGE_RISING;
+			fwspec.param_count = 2;
+			ret = __irq_domain_alloc_irqs(gc->irq.domain,
+						      /* just pick something */
+						      -1,
+						      1,
+						      NUMA_NO_NODE,
+						      &fwspec,
+						      false,
+						      NULL);
+			if (ret < 0) {
+				chip_err(gc,
+					 "can not allocate irq for GPIO line %d parent hwirq %d in hierarchy domain: %d\n",
+					 i, parent_hwirq,
+					 ret);
+			}
+		}
+	}
+
+	chip_err(gc, "%s unknown fwnode type proceed anyway\n", __func__);
+
+	return;
+}
+
+static int gpiochip_hierarchy_irq_domain_translate(struct irq_domain *d,
+						   struct irq_fwspec *fwspec,
+						   unsigned long *hwirq,
+						   unsigned int *type)
+{
+	/* We support standard DT translation */
+	if (is_of_node(fwspec->fwnode) && fwspec->param_count == 2) {
+		return irq_domain_translate_twocell(d, fwspec, hwirq, type);
+	}
+
+	/* This is for board files and others not using DT */
+	if (is_fwnode_irqchip(fwspec->fwnode)) {
+		int ret;
+
+		ret = irq_domain_translate_twocell(d, fwspec, hwirq, type);
+		if (ret)
+			return ret;
+		WARN_ON(*type == IRQ_TYPE_NONE);
+		return 0;
+	}
+	return -EINVAL;
+}
+
+static int gpiochip_hierarchy_irq_domain_alloc(struct irq_domain *d,
+					       unsigned int irq,
+					       unsigned int nr_irqs,
+					       void *data)
+{
+	struct gpio_chip *gc = d->host_data;
+	irq_hw_number_t hwirq;
+	unsigned int type = IRQ_TYPE_NONE;
+	struct irq_fwspec *fwspec = data;
+	struct irq_fwspec parent_fwspec;
+	unsigned int parent_hwirq;
+	unsigned int parent_type;
+	struct gpio_irq_chip *girq = &gc->irq;
+	int ret;
+
+	/*
+	 * The nr_irqs parameter is always one except for PCI multi-MSI
+	 * so this should not happen.
+	 */
+	WARN_ON(nr_irqs != 1);
+
+	ret = gc->irq.child_irq_domain_ops.translate(d, fwspec, &hwirq, &type);
+	if (ret)
+		return ret;
+
+	chip_info(gc, "allocate IRQ %d, hwirq %lu\n", irq,  hwirq);
+
+	ret = girq->child_to_parent_hwirq(gc, hwirq, type,
+					  &parent_hwirq, &parent_type);
+	if (ret) {
+		chip_err(gc, "can't look up hwirq %lu\n", hwirq);
+		return ret;
+	}
+	chip_info(gc, "found parent hwirq %u\n", parent_hwirq);
+
+	/*
+	 * We set handle_bad_irq because the .set_type() should
+	 * always be invoked and set the right type of handler.
+	 */
+	irq_domain_set_info(d,
+			    irq,
+			    hwirq,
+			    gc->irq.chip,
+			    gc,
+			    girq->handler,
+			    NULL, NULL);
+	irq_set_probe(irq);
+
+	/*
+	 * Create a IRQ fwspec to send up to the parent irqdomain:
+	 * specify the hwirq we address on the parent and tie it
+	 * all together up the chain.
+	 */
+	parent_fwspec.fwnode = d->parent->fwnode;
+	/* This parent only handles asserted level IRQs */
+	girq->populate_parent_fwspec(gc, &parent_fwspec, parent_hwirq,
+				     parent_type);
+	chip_info(gc, "alloc_irqs_parent for %d parent hwirq %d\n",
+		  irq, parent_hwirq);
+	ret = irq_domain_alloc_irqs_parent(d, irq, 1, &parent_fwspec);
+	if (ret)
+		chip_err(gc,
+			 "failed to allocate parent hwirq %d for hwirq %lu\n",
+			 parent_hwirq, hwirq);
+
+	return ret;
+}
+
+static unsigned int gpiochip_child_offset_to_irq_noop(struct gpio_chip *chip,
+						      unsigned int offset)
+{
+	return offset;
+}
+
+static void gpiochip_hierarchy_setup_domain_ops(struct irq_domain_ops *ops)
+{
+	ops->activate = gpiochip_irq_domain_activate;
+	ops->deactivate = gpiochip_irq_domain_deactivate;
+	ops->alloc = gpiochip_hierarchy_irq_domain_alloc;
+	ops->free = irq_domain_free_irqs_common;
+
+	/*
+	 * We only allow overriding the translate() function for
+	 * hierarchical chips, and this should only be done if the user
+	 * really need something other than 1:1 translation.
+	 */
+	if (!ops->translate)
+		ops->translate = gpiochip_hierarchy_irq_domain_translate;
+}
+
+static int gpiochip_hierarchy_add_domain(struct gpio_chip *gc)
+{
+	if (!gc->irq.child_to_parent_hwirq ||
+	    !gc->irq.fwnode) {
+		chip_err(gc, "missing irqdomain vital data\n");
+		return -EINVAL;
+	}
+
+	if (!gc->irq.child_offset_to_irq)
+		gc->irq.child_offset_to_irq = gpiochip_child_offset_to_irq_noop;
+
+	if (!gc->irq.populate_parent_fwspec)
+		gc->irq.populate_parent_fwspec =
+			gpiochip_populate_parent_fwspec_twocell;
+
+	gpiochip_hierarchy_setup_domain_ops(&gc->irq.child_irq_domain_ops);
+
+	gc->irq.domain = irq_domain_create_hierarchy(
+		gc->irq.parent_domain,
+		0,
+		gc->ngpio,
+		gc->irq.fwnode,
+		&gc->irq.child_irq_domain_ops,
+		gc);
+
+	if (!gc->irq.domain)
+		return -ENOMEM;
+
+	gpiochip_set_hierarchical_irqchip(gc, gc->irq.chip);
+
+	return 0;
+}
+
+static bool gpiochip_hierarchy_is_hierarchical(struct gpio_chip *gc)
+{
+	return !!gc->irq.parent_domain;
+}
+
+void gpiochip_populate_parent_fwspec_twocell(struct gpio_chip *chip,
+					     struct irq_fwspec *fwspec,
+					     unsigned int parent_hwirq,
+					     unsigned int parent_type)
+{
+	fwspec->param_count = 2;
+	fwspec->param[0] = parent_hwirq;
+	fwspec->param[1] = parent_type;
+}
+EXPORT_SYMBOL_GPL(gpiochip_populate_parent_fwspec_twocell);
+
+void gpiochip_populate_parent_fwspec_fourcell(struct gpio_chip *chip,
+					      struct irq_fwspec *fwspec,
+					      unsigned int parent_hwirq,
+					      unsigned int parent_type)
+{
+	fwspec->param_count = 4;
+	fwspec->param[0] = 0;
+	fwspec->param[1] = parent_hwirq;
+	fwspec->param[2] = 0;
+	fwspec->param[3] = parent_type;
+}
+EXPORT_SYMBOL_GPL(gpiochip_populate_parent_fwspec_fourcell);
+
+#else
+
+static int gpiochip_hierarchy_add_domain(struct gpio_chip *gc)
+{
+	return -EINVAL;
+}
+
+static bool gpiochip_hierarchy_is_hierarchical(struct gpio_chip *gc)
+{
+	return false;
+}
+
+#endif /* CONFIG_IRQ_DOMAIN_HIERARCHY */
+
 /**
  * gpiochip_irq_map() - maps an IRQ into a GPIO irqchip
  * @d: the irqdomain used by this irqchip
@@ -1788,6 +2055,11 @@  static const struct irq_domain_ops gpiochip_domain_ops = {
 	.xlate	= irq_domain_xlate_twocell,
 };
 
+/*
+ * TODO: move these activate/deactivate in under the hierarchicial
+ * irqchip implementation as static once SPMI and SSBI (all external
+ * users) are phased over.
+ */
 /**
  * gpiochip_irq_domain_activate() - Lock a GPIO to be used as an IRQ
  * @domain: The IRQ domain used by this IRQ chip
@@ -1827,10 +2099,23 @@  EXPORT_SYMBOL_GPL(gpiochip_irq_domain_deactivate);
 
 static int gpiochip_to_irq(struct gpio_chip *chip, unsigned offset)
 {
+	struct irq_domain *domain = chip->irq.domain;
+
 	if (!gpiochip_irqchip_irq_valid(chip, offset))
 		return -ENXIO;
 
-	return irq_create_mapping(chip->irq.domain, offset);
+	if (irq_domain_is_hierarchy(domain)) {
+		struct irq_fwspec spec;
+
+		spec.fwnode = domain->fwnode;
+		spec.param_count = 2;
+		spec.param[0] = chip->irq.child_offset_to_irq(chip, offset);
+		spec.param[1] = IRQ_TYPE_NONE;
+
+		return irq_create_fwspec_mapping(&spec);
+	}
+
+	return irq_create_mapping(domain, offset);
 }
 
 static int gpiochip_irq_reqres(struct irq_data *d)
@@ -1907,7 +2192,7 @@  static int gpiochip_add_irqchip(struct gpio_chip *gpiochip,
 				struct lock_class_key *request_key)
 {
 	struct irq_chip *irqchip = gpiochip->irq.chip;
-	const struct irq_domain_ops *ops;
+	const struct irq_domain_ops *ops = NULL;
 	struct device_node *np;
 	unsigned int type;
 	unsigned int i;
@@ -1943,16 +2228,25 @@  static int gpiochip_add_irqchip(struct gpio_chip *gpiochip,
 	gpiochip->irq.lock_key = lock_key;
 	gpiochip->irq.request_key = request_key;
 
-	if (gpiochip->irq.domain_ops)
-		ops = gpiochip->irq.domain_ops;
-	else
-		ops = &gpiochip_domain_ops;
-
-	gpiochip->irq.domain = irq_domain_add_simple(np, gpiochip->ngpio,
-						     gpiochip->irq.first,
-						     ops, gpiochip);
-	if (!gpiochip->irq.domain)
-		return -EINVAL;
+	/* If a parent irqdomain is provided, let's build a hierarchy */
+	if (gpiochip_hierarchy_is_hierarchical(gpiochip)) {
+		int ret = gpiochip_hierarchy_add_domain(gpiochip);
+		if (ret)
+			return ret;
+	} else {
+		/* Some drivers provide custom irqdomain ops */
+		if (gpiochip->irq.domain_ops)
+			ops = gpiochip->irq.domain_ops;
+
+		if (!ops)
+			ops = &gpiochip_domain_ops;
+		gpiochip->irq.domain = irq_domain_add_simple(np,
+			gpiochip->ngpio,
+			gpiochip->irq.first,
+			ops, gpiochip);
+		if (!gpiochip->irq.domain)
+			return -EINVAL;
+	}
 
 	if (gpiochip->irq.parent_handler) {
 		void *data = gpiochip->irq.parent_handler_data ?: gpiochip;
diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h
index 6a0e420915a3..0e6d3b0c0211 100644
--- a/include/linux/gpio/driver.h
+++ b/include/linux/gpio/driver.h
@@ -23,6 +23,9 @@  enum gpio_lookup_flags;
 #ifdef CONFIG_GPIOLIB
 
 #ifdef CONFIG_GPIOLIB_IRQCHIP
+
+struct gpio_chip;
+
 /**
  * struct gpio_irq_chip - GPIO interrupt controller
  */
@@ -49,6 +52,84 @@  struct gpio_irq_chip {
 	 */
 	const struct irq_domain_ops *domain_ops;
 
+#ifdef CONFIG_IRQ_DOMAIN_HIERARCHY
+	/**
+	 * @fwnode:
+	 *
+	 * Firmware node corresponding to this gpiochip/irqchip, necessary
+	 * for hierarchical irqdomain support.
+	 */
+	struct fwnode_handle *fwnode;
+
+	/**
+	 * @parent_domain:
+	 *
+	 * If non-NULL, will be set as the parent of this GPIO interrupt
+	 * controller's IRQ domain to establish a hierarchical interrupt
+	 * domain. The presence of this will activate the hierarchical
+	 * interrupt support.
+	 */
+	struct irq_domain *parent_domain;
+
+	/**
+	 * @child_to_parent_hwirq:
+	 *
+	 * This callback translates a child hardware IRQ offset to a parent
+	 * hardware IRQ offset on a hierarchical interrupt chip. The child
+	 * hardware IRQs correspond to the GPIO index 0..ngpio-1 (see the
+	 * ngpio field of struct gpio_chip) and the corresponding parent
+	 * hardware IRQ and type (such as IRQ_TYPE_*) shall be returned by
+	 * the driver. The driver can calculate this from an offset or using
+	 * a lookup table or whatever method is best for this chip. Return
+	 * 0 on successful translation in the driver.
+	 *
+	 * If some ranges of hardware IRQs do not have a corresponding parent
+	 * HWIRQ, return -EINVAL, but also make sure to fill in @valid_mask and
+	 * @need_valid_mask to make these GPIO lines unavailable for
+	 * translation.
+	 */
+	int (*child_to_parent_hwirq)(struct gpio_chip *chip,
+				     unsigned int child_hwirq,
+				     unsigned int child_type,
+				     unsigned int *parent_hwirq,
+				     unsigned int *parent_type);
+
+	/**
+	 * @populate_parent_fwspec:
+	 *
+	 * This optional callback populates the &struct irq_fwspec for the
+	 * parent's IRQ domain. If this is not specified, then
+	 * &gpiochip_populate_parent_fwspec_twocell will be used. A four-cell
+	 * variant named &gpiochip_populate_parent_fwspec_fourcell is also
+	 * available.
+	 */
+	void (*populate_parent_fwspec)(struct gpio_chip *chip,
+				       struct irq_fwspec *fwspec,
+				       unsigned int parent_hwirq,
+				       unsigned int parent_type);
+
+	/**
+	 * @child_offset_to_irq:
+	 *
+	 * This optional callback is used to translate the child's GPIO line
+	 * offset on the GPIO chip to an IRQ number for the GPIO to_irq()
+	 * callback. If this is not specified, then a default callback will be
+	 * provided that returns the line offset.
+	 */
+	unsigned int (*child_offset_to_irq)(struct gpio_chip *chip,
+					    unsigned int pin);
+
+	/**
+	 * @child_irq_domain_ops:
+	 *
+	 * The IRQ domain operations that will be used for this GPIO IRQ
+	 * chip. If no operations are provided, then default callbacks will
+	 * be populated to setup the IRQ hierarchy. Some drivers need to
+	 * supply their own translate function.
+	 */
+	struct irq_domain_ops child_irq_domain_ops;
+#endif
+
 	/**
 	 * @handler:
 	 *
@@ -449,6 +530,36 @@  struct bgpio_pdata {
 	int ngpio;
 };
 
+#ifdef CONFIG_IRQ_DOMAIN_HIERARCHY
+
+void gpiochip_populate_parent_fwspec_twocell(struct gpio_chip *chip,
+					     struct irq_fwspec *fwspec,
+					     unsigned int parent_hwirq,
+					     unsigned int parent_type);
+void gpiochip_populate_parent_fwspec_fourcell(struct gpio_chip *chip,
+					      struct irq_fwspec *fwspec,
+					      unsigned int parent_hwirq,
+					      unsigned int parent_type);
+
+#else
+
+static void gpiochip_populate_parent_fwspec_twocell(struct gpio_chip *chip,
+						    struct irq_fwspec *fwspec,
+						    unsigned int parent_hwirq,
+						    unsigned int parent_type)
+{
+}
+
+static void gpiochip_populate_parent_fwspec_fourcell(struct gpio_chip *chip,
+						     struct irq_fwspec *fwspec,
+						     unsigned int parent_hwirq,
+						     unsigned int parent_type)
+{
+}
+
+#endif /* CONFIG_IRQ_DOMAIN_HIERARCHY */
+
+
 #if IS_ENABLED(CONFIG_GPIO_GENERIC)
 
 int bgpio_init(struct gpio_chip *gc, struct device *dev,