The idea of this changes was inspired by discussion [1].
Now IRQ mappings are always created for all (allowed) GPIOs in GPIO irqchip in
gpiochip_irqchip_add_key() and this goes against the idea of SPARSE_IRQ and,
as result, leads to:
- increasing of memory consumption because of allocated IRQ descriptors most
of which will never ever be used (especially on platform with a
high number of GPIOs).
- imposibility to use GPIO irqchip APIs by gpio drivers when HW implements
GPIO IRQ functionality as IRQ crossbar/multiplexer which has only limited
number of IRQ outputs (example from [1], all GPIOs can be mapped on only 8
IRQs).
Hence, remove static IRQ mapping code from gpiochip_irqchip_add_key() and
instead replace irq_find_mapping() with irq_create_mapping() in
gpiochip_to_irq(). Also add additional gpiochip_irqchip_irq_valid() calls
in gpiochip_to_irq() and gpiochip_irq_map().
After this change gpio2irq mapping will happen the following way when GPIO
irqchip APIs are used by gpio driver:
- IRQ mappings will be created statically if driver passes first_irq>0
vlaue in gpiochip_irqchip_add_key().
- IRQ mappings will be created dynamically from gpio_to_irq() or
of_irq_get().
Tested on am335x-evm and dra72-evm-revc.
- dra72-evm-revc: number of created irq mappings decreased from 402 -> 135
- am335x-evm: number of created irq mappings decreased from 188 -> 63
It's good to have more testing feedbacks for these patches
(personally I do not expect regressions, but who knows). One possible side
effect is "Cannot allocate irq_descs @ IRQ%d, assuming pre-allocated\n" message
during boot, but in this case it might be good to take a look on such GPIO
drivers to understand why they need to allocate IRQ descriptors.
[1] https://www.mail-archive.com/[email protected]/msg1435847.html
Grygorii Strashko (3):
gpio: tegra: remove gpio_to_irq() from hw irq handlers
gpio: pxa: remove gpio_to_irq() from hw irq handlers
gpiolib: make gpio irqchip compatible with sparse_irq
drivers/gpio/gpio-pxa.c | 8 +++++---
drivers/gpio/gpio-tegra.c | 3 ++-
drivers/gpio/gpiolib.c | 29 ++++++-----------------------
include/linux/gpio/driver.h | 1 -
4 files changed, 13 insertions(+), 28 deletions(-)
--
2.10.1
Now IRQ mappings are always created for all (allowed) GPIOs in gpio irqchip in
gpiochip_irqchip_add_key() which goes against the idea of SPARSE_IRQ and,
as result, leads to:
- increasing of memory consumption because of allocated IRQ descriptors most
of which will never ever be used (especially on platform with a
high number of GPIOs)
- imposibility to use GPIO irqchip APIs by gpio drivers when HW implements
GPIO IRQ functionality as IRQ crossbar/multiplexer which has only limited
number of IRQ outputs (example from [1], all GPIOs can be mapped on only 8
IRQs).
Hence, remove static IRQ mapping code from gpiochip_irqchip_add_key() and
instead replace irq_find_mapping() with irq_create_mapping() in
gpiochip_to_irq(). Also add additional gpiochip_irqchip_irq_valid() calls
in gpiochip_to_irq() and gpiochip_irq_map().
After this change gpio2irq mapping will happen the following way when GPIO
irqchip APIs are used by gpio driver:
- IRQ mappings will be created statically if driver passes first_irq>0
vlaue in gpiochip_irqchip_add_key().
- IRQ mappings will be created dynamically from gpio_to_irq() or
of_irq_get().
[1] https://www.mail-archive.com/[email protected]/msg1435847.html
Signed-off-by: Grygorii Strashko <[email protected]>
---
drivers/gpio/gpiolib.c | 29 ++++++-----------------------
include/linux/gpio/driver.h | 1 -
2 files changed, 6 insertions(+), 24 deletions(-)
diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index 9568708..ee6ba07 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -1605,6 +1605,9 @@ static int gpiochip_irq_map(struct irq_domain *d, unsigned int irq,
{
struct gpio_chip *chip = d->host_data;
+ if (!gpiochip_irqchip_irq_valid(chip, hwirq))
+ return -ENXIO;
+
irq_set_chip_data(irq, chip);
/*
* This lock class tells lockdep that GPIO irqs are in a different
@@ -1671,7 +1674,9 @@ static void gpiochip_irq_relres(struct irq_data *d)
static int gpiochip_to_irq(struct gpio_chip *chip, unsigned offset)
{
- return irq_find_mapping(chip->irqdomain, offset);
+ if (!gpiochip_irqchip_irq_valid(chip, offset))
+ return -ENXIO;
+ return irq_create_mapping(chip->irqdomain, offset);
}
/**
@@ -1747,9 +1752,6 @@ int gpiochip_irqchip_add_key(struct gpio_chip *gpiochip,
struct lock_class_key *lock_key)
{
struct device_node *of_node;
- bool irq_base_set = false;
- unsigned int offset;
- unsigned irq_base = 0;
if (!gpiochip || !irqchip)
return -EINVAL;
@@ -1806,25 +1808,6 @@ int gpiochip_irqchip_add_key(struct gpio_chip *gpiochip,
irqchip->irq_release_resources = gpiochip_irq_relres;
}
- /*
- * Prepare the mapping since the irqchip shall be orthogonal to
- * any gpiochip calls. If the first_irq was zero, this is
- * necessary to allocate descriptors for all IRQs.
- */
- for (offset = 0; offset < gpiochip->ngpio; offset++) {
- if (!gpiochip_irqchip_irq_valid(gpiochip, offset))
- continue;
- irq_base = irq_create_mapping(gpiochip->irqdomain, offset);
- if (!irq_base_set) {
- /*
- * Store the base into the gpiochip to be used when
- * unmapping the irqs.
- */
- gpiochip->irq_base = irq_base;
- irq_base_set = true;
- }
- }
-
acpi_gpiochip_request_interrupts(gpiochip);
return 0;
diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h
index af20369..a3e2e3e 100644
--- a/include/linux/gpio/driver.h
+++ b/include/linux/gpio/driver.h
@@ -165,7 +165,6 @@ struct gpio_chip {
*/
struct irq_chip *irqchip;
struct irq_domain *irqdomain;
- unsigned int irq_base;
irq_flow_handler_t irq_handler;
unsigned int irq_default_type;
unsigned int irq_chained_parent;
--
2.10.1
gpio_to_irq() API expected to be used by GPIO consumers and
not drivers and there are no guarantee that its gpiolib implementation
is irq safe.
Signed-off-by: Grygorii Strashko <[email protected]>
---
drivers/gpio/gpio-pxa.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/gpio/gpio-pxa.c b/drivers/gpio/gpio-pxa.c
index 832f3e4..6029899 100644
--- a/drivers/gpio/gpio-pxa.c
+++ b/drivers/gpio/gpio-pxa.c
@@ -451,7 +451,9 @@ static irqreturn_t pxa_gpio_demux_handler(int in_irq, void *d)
for_each_set_bit(n, &gedr, BITS_PER_LONG) {
loop = 1;
- generic_handle_irq(gpio_to_irq(gpio + n));
+ generic_handle_irq(
+ irq_find_mapping(pchip->irqdomain,
+ gpio + n));
}
}
handled += loop;
@@ -465,9 +467,9 @@ static irqreturn_t pxa_gpio_direct_handler(int in_irq, void *d)
struct pxa_gpio_chip *pchip = d;
if (in_irq == pchip->irq0) {
- generic_handle_irq(gpio_to_irq(0));
+ generic_handle_irq(irq_find_mapping(pchip->irqdomain, 0));
} else if (in_irq == pchip->irq1) {
- generic_handle_irq(gpio_to_irq(1));
+ generic_handle_irq(irq_find_mapping(pchip->irqdomain, 1));
} else {
pr_err("%s() unknown irq %d\n", __func__, in_irq);
return IRQ_NONE;
--
2.10.1
gpio_to_irq() API expected to be used by GPIO consumers and
not drivers and there are no guarantee that its gpiolib implementation
is irq safe.
Signed-off-by: Grygorii Strashko <[email protected]>
---
drivers/gpio/gpio-tegra.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpio/gpio-tegra.c b/drivers/gpio/gpio-tegra.c
index 88529d3..31b9a86 100644
--- a/drivers/gpio/gpio-tegra.c
+++ b/drivers/gpio/gpio-tegra.c
@@ -389,7 +389,8 @@ static void tegra_gpio_irq_handler(struct irq_desc *desc)
chained_irq_exit(chip, desc);
}
- generic_handle_irq(gpio_to_irq(gpio + pin));
+ generic_handle_irq(irq_find_mapping(tgi->irq_domain,
+ gpio + pin));
}
}
--
2.10.1
On Sun, Jul 9, 2017 at 12:44 AM, Grygorii Strashko
<[email protected]> wrote:
> Now IRQ mappings are always created for all (allowed) GPIOs in gpio irqchip in
> gpiochip_irqchip_add_key() which goes against the idea of SPARSE_IRQ and,
> as result, leads to:
> - increasing of memory consumption because of allocated IRQ descriptors most
> of which will never ever be used (especially on platform with a
> high number of GPIOs)
> - imposibility to use GPIO irqchip APIs by gpio drivers when HW implements
> GPIO IRQ functionality as IRQ crossbar/multiplexer which has only limited
> number of IRQ outputs (example from [1], all GPIOs can be mapped on only 8
> IRQs).
>
> Hence, remove static IRQ mapping code from gpiochip_irqchip_add_key() and
> instead replace irq_find_mapping() with irq_create_mapping() in
> gpiochip_to_irq(). Also add additional gpiochip_irqchip_irq_valid() calls
> in gpiochip_to_irq() and gpiochip_irq_map().
>
> After this change gpio2irq mapping will happen the following way when GPIO
> irqchip APIs are used by gpio driver:
> - IRQ mappings will be created statically if driver passes first_irq>0
> vlaue in gpiochip_irqchip_add_key().
> - IRQ mappings will be created dynamically from gpio_to_irq() or
> of_irq_get().
>
> [1] https://www.mail-archive.com/[email protected]/msg1435847.html
> Signed-off-by: Grygorii Strashko <[email protected]>
This is obviously better and what I should have done from the beginning
had I only been smart enough, so what better to do than to throw this in
now for -next and see what happens :)
Patch applied.
Yours,
Linus Walleij
On Sun, Jul 9, 2017 at 12:44 AM, Grygorii Strashko
<[email protected]> wrote:
> gpio_to_irq() API expected to be used by GPIO consumers and
> not drivers and there are no guarantee that its gpiolib implementation
> is irq safe.
>
> Signed-off-by: Grygorii Strashko <[email protected]>
Patch applied.
This is the right thing to do.
PXA developers: please test.
Yours,
Linus Walleij
On Sun, Jul 9, 2017 at 12:44 AM, Grygorii Strashko
<[email protected]> wrote:
> gpio_to_irq() API expected to be used by GPIO consumers and
> not drivers and there are no guarantee that its gpiolib implementation
> is irq safe.
>
> Signed-off-by: Grygorii Strashko <[email protected]>
Patch applied.
This is obviously the right way to do it.
Tegra people: please test.
Yours,
Linus Walleij
Linus Walleij <[email protected]> writes:
> On Sun, Jul 9, 2017 at 12:44 AM, Grygorii Strashko
> <[email protected]> wrote:
>
>> gpio_to_irq() API expected to be used by GPIO consumers and
>> not drivers and there are no guarantee that its gpiolib implementation
>> is irq safe.
>>
>> Signed-off-by: Grygorii Strashko <[email protected]>
>
> Patch applied.
>
> This is the right thing to do.
>
> PXA developers: please test.
Hi Linus,
Currently pxa25x and pxa27x are broken because of a .data section misalignment
(issue cross arm architectures), so testing is a bit hard for me right now. I'll
wait to have a working again tree to test, and then report.
Cheers.
--
Robert