linux-mips.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 1/1] irqchip/mips-gic: Fix local interrupts
@ 2017-01-25 14:08 Marcin Nowakowski
  2017-01-25 14:08 ` Marcin Nowakowski
  2017-02-08 14:16 ` Marc Zyngier
  0 siblings, 2 replies; 4+ messages in thread
From: Marcin Nowakowski @ 2017-01-25 14:08 UTC (permalink / raw)
  To: linux-kernel
  Cc: Paul Burton, Thomas Gleixner, Jason Cooper, Marc Zyngier, linux-mips

Some local interrupts are not initialised properly at the moment and
cannot be used since the domain's alloc method is never called for them.

This has been observed earlier and partially fixed in commit
e875bd66dfb ("irqchip/mips-gic: Fix local interrupts"), but that change
still relied on the interrupt to be requested by an external driver (eg.
drivers/clocksource/mips-gic-timer.c).

This does however not solve the issue for interrupts that are not
referenced by any driver through the device tree and results in
request_irq() calls returning -ENOSYS. It can be observed when attempting
to use perf tool to access hardware performance counters.

Fix this by explicitly calling irq_create_fwspec_mapping() for local
interrupts.

Fixes: e875bd66dfb ("irqchip/mips-gic: Fix local interrupts")
Signed-off-by: Marcin Nowakowski <marcin.nowakowski@imgtec.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Jason Cooper <jason@lakedaemon.net>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Cc: linux-mips@linux-mips.org
---
 drivers/irqchip/irq-mips-gic.c | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)

diff --git a/drivers/irqchip/irq-mips-gic.c b/drivers/irqchip/irq-mips-gic.c
index c01c09e..11d12bc 100644
--- a/drivers/irqchip/irq-mips-gic.c
+++ b/drivers/irqchip/irq-mips-gic.c
@@ -968,6 +968,34 @@ static struct irq_domain_ops gic_ipi_domain_ops = {
 	.match = gic_ipi_domain_match,
 };
 
+static void __init gic_map_single_int(struct device_node *node,
+				      unsigned int irq)
+{
+	unsigned int linux_irq;
+	struct irq_fwspec local_int_fwspec = {
+		.fwnode         = &node->fwnode,
+		.param_count    = 3,
+		.param          = {
+			[0]     = GIC_LOCAL,
+			[1]     = irq,
+			[2]     = IRQ_TYPE_NONE,
+		},
+	};
+
+	if (!gic_local_irq_is_routable(irq))
+		return;
+
+	linux_irq = irq_create_fwspec_mapping(&local_int_fwspec);
+	WARN_ON(!linux_irq);
+}
+
+static void __init gic_map_interrupts(struct device_node *node)
+{
+	gic_map_single_int(node, GIC_LOCAL_INT_TIMER);
+	gic_map_single_int(node, GIC_LOCAL_INT_PERFCTR);
+	gic_map_single_int(node, GIC_LOCAL_INT_FDC);
+}
+
 static void __init __gic_init(unsigned long gic_base_addr,
 			      unsigned long gic_addrspace_size,
 			      unsigned int cpu_vec, unsigned int irqbase,
@@ -1067,6 +1095,7 @@ static void __init __gic_init(unsigned long gic_base_addr,
 	}
 
 	gic_basic_init();
+	gic_map_interrupts(node);
 }
 
 void __init gic_init(unsigned long gic_base_addr,
-- 
2.7.4

^ permalink raw reply	[flat|nested] 4+ messages in thread

* [PATCH 1/1] irqchip/mips-gic: Fix local interrupts
  2017-01-25 14:08 [PATCH 1/1] irqchip/mips-gic: Fix local interrupts Marcin Nowakowski
@ 2017-01-25 14:08 ` Marcin Nowakowski
  2017-02-08 14:16 ` Marc Zyngier
  1 sibling, 0 replies; 4+ messages in thread
From: Marcin Nowakowski @ 2017-01-25 14:08 UTC (permalink / raw)
  To: linux-kernel
  Cc: Paul Burton, Thomas Gleixner, Jason Cooper, Marc Zyngier, linux-mips

Some local interrupts are not initialised properly at the moment and
cannot be used since the domain's alloc method is never called for them.

This has been observed earlier and partially fixed in commit
e875bd66dfb ("irqchip/mips-gic: Fix local interrupts"), but that change
still relied on the interrupt to be requested by an external driver (eg.
drivers/clocksource/mips-gic-timer.c).

This does however not solve the issue for interrupts that are not
referenced by any driver through the device tree and results in
request_irq() calls returning -ENOSYS. It can be observed when attempting
to use perf tool to access hardware performance counters.

Fix this by explicitly calling irq_create_fwspec_mapping() for local
interrupts.

Fixes: e875bd66dfb ("irqchip/mips-gic: Fix local interrupts")
Signed-off-by: Marcin Nowakowski <marcin.nowakowski@imgtec.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Jason Cooper <jason@lakedaemon.net>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Cc: linux-mips@linux-mips.org
---
 drivers/irqchip/irq-mips-gic.c | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)

diff --git a/drivers/irqchip/irq-mips-gic.c b/drivers/irqchip/irq-mips-gic.c
index c01c09e..11d12bc 100644
--- a/drivers/irqchip/irq-mips-gic.c
+++ b/drivers/irqchip/irq-mips-gic.c
@@ -968,6 +968,34 @@ static struct irq_domain_ops gic_ipi_domain_ops = {
 	.match = gic_ipi_domain_match,
 };
 
+static void __init gic_map_single_int(struct device_node *node,
+				      unsigned int irq)
+{
+	unsigned int linux_irq;
+	struct irq_fwspec local_int_fwspec = {
+		.fwnode         = &node->fwnode,
+		.param_count    = 3,
+		.param          = {
+			[0]     = GIC_LOCAL,
+			[1]     = irq,
+			[2]     = IRQ_TYPE_NONE,
+		},
+	};
+
+	if (!gic_local_irq_is_routable(irq))
+		return;
+
+	linux_irq = irq_create_fwspec_mapping(&local_int_fwspec);
+	WARN_ON(!linux_irq);
+}
+
+static void __init gic_map_interrupts(struct device_node *node)
+{
+	gic_map_single_int(node, GIC_LOCAL_INT_TIMER);
+	gic_map_single_int(node, GIC_LOCAL_INT_PERFCTR);
+	gic_map_single_int(node, GIC_LOCAL_INT_FDC);
+}
+
 static void __init __gic_init(unsigned long gic_base_addr,
 			      unsigned long gic_addrspace_size,
 			      unsigned int cpu_vec, unsigned int irqbase,
@@ -1067,6 +1095,7 @@ static void __init __gic_init(unsigned long gic_base_addr,
 	}
 
 	gic_basic_init();
+	gic_map_interrupts(node);
 }
 
 void __init gic_init(unsigned long gic_base_addr,
-- 
2.7.4

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH 1/1] irqchip/mips-gic: Fix local interrupts
  2017-01-25 14:08 [PATCH 1/1] irqchip/mips-gic: Fix local interrupts Marcin Nowakowski
  2017-01-25 14:08 ` Marcin Nowakowski
@ 2017-02-08 14:16 ` Marc Zyngier
  2017-02-08 14:16   ` Marc Zyngier
  1 sibling, 1 reply; 4+ messages in thread
From: Marc Zyngier @ 2017-02-08 14:16 UTC (permalink / raw)
  To: Marcin Nowakowski
  Cc: linux-kernel, Paul Burton, Thomas Gleixner, Jason Cooper, linux-mips

On Wed, Jan 25 2017 at  2:08:25 pm GMT, Marcin Nowakowski <marcin.nowakowski@imgtec.com> wrote:
> Some local interrupts are not initialised properly at the moment and
> cannot be used since the domain's alloc method is never called for them.
>
> This has been observed earlier and partially fixed in commit
> e875bd66dfb ("irqchip/mips-gic: Fix local interrupts"), but that change
> still relied on the interrupt to be requested by an external driver (eg.
> drivers/clocksource/mips-gic-timer.c).
>
> This does however not solve the issue for interrupts that are not
> referenced by any driver through the device tree and results in
> request_irq() calls returning -ENOSYS. It can be observed when attempting
> to use perf tool to access hardware performance counters.
>
> Fix this by explicitly calling irq_create_fwspec_mapping() for local
> interrupts.
>
> Fixes: e875bd66dfb ("irqchip/mips-gic: Fix local interrupts")
> Signed-off-by: Marcin Nowakowski <marcin.nowakowski@imgtec.com>
> Cc: Paul Burton <paul.burton@imgtec.com>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Jason Cooper <jason@lakedaemon.net>
> Cc: Marc Zyngier <marc.zyngier@arm.com>
> Cc: linux-mips@linux-mips.org

Queued, thanks.

	M.
-- 
Jazz is not dead, it just smell funny.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH 1/1] irqchip/mips-gic: Fix local interrupts
  2017-02-08 14:16 ` Marc Zyngier
@ 2017-02-08 14:16   ` Marc Zyngier
  0 siblings, 0 replies; 4+ messages in thread
From: Marc Zyngier @ 2017-02-08 14:16 UTC (permalink / raw)
  To: Marcin Nowakowski
  Cc: linux-kernel, Paul Burton, Thomas Gleixner, Jason Cooper, linux-mips

On Wed, Jan 25 2017 at  2:08:25 pm GMT, Marcin Nowakowski <marcin.nowakowski@imgtec.com> wrote:
> Some local interrupts are not initialised properly at the moment and
> cannot be used since the domain's alloc method is never called for them.
>
> This has been observed earlier and partially fixed in commit
> e875bd66dfb ("irqchip/mips-gic: Fix local interrupts"), but that change
> still relied on the interrupt to be requested by an external driver (eg.
> drivers/clocksource/mips-gic-timer.c).
>
> This does however not solve the issue for interrupts that are not
> referenced by any driver through the device tree and results in
> request_irq() calls returning -ENOSYS. It can be observed when attempting
> to use perf tool to access hardware performance counters.
>
> Fix this by explicitly calling irq_create_fwspec_mapping() for local
> interrupts.
>
> Fixes: e875bd66dfb ("irqchip/mips-gic: Fix local interrupts")
> Signed-off-by: Marcin Nowakowski <marcin.nowakowski@imgtec.com>
> Cc: Paul Burton <paul.burton@imgtec.com>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Jason Cooper <jason@lakedaemon.net>
> Cc: Marc Zyngier <marc.zyngier@arm.com>
> Cc: linux-mips@linux-mips.org

Queued, thanks.

	M.
-- 
Jazz is not dead, it just smell funny.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2017-02-08 14:17 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-01-25 14:08 [PATCH 1/1] irqchip/mips-gic: Fix local interrupts Marcin Nowakowski
2017-01-25 14:08 ` Marcin Nowakowski
2017-02-08 14:16 ` Marc Zyngier
2017-02-08 14:16   ` Marc Zyngier

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).