2015-07-20 16:46:39

by Felipe Balbi

[permalink] [raw]
Subject: [PATCH RESEND] irqchip: omap-intc: improve IRQ handler

as it turns out the current IRQ number will
*always* be available from SIR register which
renders the reads of PENDING registers as plain
unnecessary overhead.

In order to catch any situation where SIR reads
as zero, we're adding a WARN() to turn it into
a very verbose error and users actually report
it.

With this patch average running time of
omap_intc_handle_irq() reduced from about 28.5us
to 19.8us as measured by the kernel function
profiler.

Tested with BeagleBoneBlack Rev A5C.

Tested-by: Tony Lindgren <[email protected]>
Signed-off-by: Felipe Balbi <[email protected]>
---
drivers/irqchip/irq-omap-intc.c | 35 +++++------------------------------
1 file changed, 5 insertions(+), 30 deletions(-)

diff --git a/drivers/irqchip/irq-omap-intc.c b/drivers/irqchip/irq-omap-intc.c
index a569c6dbd1d1..0706db624213 100644
--- a/drivers/irqchip/irq-omap-intc.c
+++ b/drivers/irqchip/irq-omap-intc.c
@@ -331,37 +331,12 @@ static int __init omap_init_irq(u32 base, struct device_node *node)
static asmlinkage void __exception_irq_entry
omap_intc_handle_irq(struct pt_regs *regs)
{
- u32 irqnr = 0;
- int handled_irq = 0;
- int i;
-
- do {
- for (i = 0; i < omap_nr_pending; i++) {
- irqnr = intc_readl(INTC_PENDING_IRQ0 + (0x20 * i));
- if (irqnr)
- goto out;
- }
-
-out:
- if (!irqnr)
- break;
-
- irqnr = intc_readl(INTC_SIR);
- irqnr &= ACTIVEIRQ_MASK;
+ u32 irqnr;

- if (irqnr) {
- handle_domain_irq(domain, irqnr, regs);
- handled_irq = 1;
- }
- } while (irqnr);
-
- /*
- * If an irq is masked or deasserted while active, we will
- * keep ending up here with no irq handled. So remove it from
- * the INTC with an ack.
- */
- if (!handled_irq)
- omap_ack_irq(NULL);
+ irqnr = intc_readl(INTC_SIR);
+ irqnr &= ACTIVEIRQ_MASK;
+ WARN(!irqnr, "Spurious IRQ ?\n");
+ handle_domain_irq(domain, irqnr, regs);
}

void __init omap3_init_irq(void)
--
2.4.4


2015-07-20 20:44:25

by Thomas Gleixner

[permalink] [raw]
Subject: Re: [PATCH RESEND] irqchip: omap-intc: improve IRQ handler

On Mon, 20 Jul 2015, Felipe Balbi wrote:
> + irqnr = intc_readl(INTC_SIR);
> + irqnr &= ACTIVEIRQ_MASK;
> + WARN(!irqnr, "Spurious IRQ ?\n");

Shouldn't that be WARN_ONCE?

Thanks,

tglx

2015-07-20 20:49:46

by Felipe Balbi

[permalink] [raw]
Subject: Re: [PATCH RESEND] irqchip: omap-intc: improve IRQ handler

On Mon, Jul 20, 2015 at 10:44:08PM +0200, Thomas Gleixner wrote:
> On Mon, 20 Jul 2015, Felipe Balbi wrote:
> > + irqnr = intc_readl(INTC_SIR);
> > + irqnr &= ACTIVEIRQ_MASK;
> > + WARN(!irqnr, "Spurious IRQ ?\n");
>
> Shouldn't that be WARN_ONCE?

Sure, why not ?

8<--------------------------------------------------------------------

From a71590d725e54b888db278de6c278d2a106fc76d Mon Sep 17 00:00:00 2001
From: Felipe Balbi <[email protected]>
Date: Fri, 2 Jan 2015 16:18:54 -0600
Subject: [PATCH RESEND v2] irqchip: omap-intc: improve IRQ handler

as it turns out the current IRQ number will
*always* be available from SIR register which
renders the reads of PENDING registers as plain
unnecessary overhead.

In order to catch any situation where SIR reads
as zero, we're adding a WARN() to turn it into
a very verbose error and users actually report
it.

With this patch average running time of
omap_intc_handle_irq() reduced from about 28.5us
to 19.8us as measured by the kernel function
profiler.

Tested with BeagleBoneBlack Rev A5C.

Tested-by: Tony Lindgren <[email protected]>
Signed-off-by: Felipe Balbi <[email protected]>
---

Changes from V1:

- replace WARN() with WARN_ONCE()

drivers/irqchip/irq-omap-intc.c | 35 +++++------------------------------
1 file changed, 5 insertions(+), 30 deletions(-)

diff --git a/drivers/irqchip/irq-omap-intc.c b/drivers/irqchip/irq-omap-intc.c
index a569c6dbd1d1..e15f10746310 100644
--- a/drivers/irqchip/irq-omap-intc.c
+++ b/drivers/irqchip/irq-omap-intc.c
@@ -331,37 +331,12 @@ static int __init omap_init_irq(u32 base, struct device_node *node)
static asmlinkage void __exception_irq_entry
omap_intc_handle_irq(struct pt_regs *regs)
{
- u32 irqnr = 0;
- int handled_irq = 0;
- int i;
-
- do {
- for (i = 0; i < omap_nr_pending; i++) {
- irqnr = intc_readl(INTC_PENDING_IRQ0 + (0x20 * i));
- if (irqnr)
- goto out;
- }
-
-out:
- if (!irqnr)
- break;
-
- irqnr = intc_readl(INTC_SIR);
- irqnr &= ACTIVEIRQ_MASK;
+ u32 irqnr;

- if (irqnr) {
- handle_domain_irq(domain, irqnr, regs);
- handled_irq = 1;
- }
- } while (irqnr);
-
- /*
- * If an irq is masked or deasserted while active, we will
- * keep ending up here with no irq handled. So remove it from
- * the INTC with an ack.
- */
- if (!handled_irq)
- omap_ack_irq(NULL);
+ irqnr = intc_readl(INTC_SIR);
+ irqnr &= ACTIVEIRQ_MASK;
+ WARN_ONCE(!irqnr, "Spurious IRQ ?\n");
+ handle_domain_irq(domain, irqnr, regs);
}

void __init omap3_init_irq(void)
--
2.4.4


--
balbi


Attachments:
(No filename) (2.45 kB)
signature.asc (819.00 B)
Digital signature
Download all attachments
Subject: [tip:irq/core] irqchip: omap-intc: Improve IRQ handler

Commit-ID: 6ed3464897cc825a75218653c710d673282dfcf8
Gitweb: http://git.kernel.org/tip/6ed3464897cc825a75218653c710d673282dfcf8
Author: Felipe Balbi <[email protected]>
AuthorDate: Fri, 2 Jan 2015 16:18:54 -0600
Committer: Thomas Gleixner <[email protected]>
CommitDate: Wed, 22 Jul 2015 18:37:42 +0200

irqchip: omap-intc: Improve IRQ handler

As it turns out the current IRQ number will *always* be available from
SIR register which renders the reads of PENDING registers as plain
unnecessary overhead.

In order to catch any situation where SIR reads as zero, we're adding
a WARN() to turn it into a very verbose error and users actually
report it.

With this patch average running time of omap_intc_handle_irq() reduced
from about 28.5us to 19.8us as measured by the kernel function
profiler.

Tested with BeagleBoneBlack Rev A5C.

Tested-by: Tony Lindgren <[email protected]>
Signed-off-by: Felipe Balbi <[email protected]>
Cc: Linux ARM Kernel Mailing List <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Thomas Gleixner <[email protected]>
---
drivers/irqchip/irq-omap-intc.c | 35 +++++------------------------------
1 file changed, 5 insertions(+), 30 deletions(-)

diff --git a/drivers/irqchip/irq-omap-intc.c b/drivers/irqchip/irq-omap-intc.c
index 6cc0ad5..8587d0f 100644
--- a/drivers/irqchip/irq-omap-intc.c
+++ b/drivers/irqchip/irq-omap-intc.c
@@ -330,37 +330,12 @@ static int __init omap_init_irq(u32 base, struct device_node *node)
static asmlinkage void __exception_irq_entry
omap_intc_handle_irq(struct pt_regs *regs)
{
- u32 irqnr = 0;
- int handled_irq = 0;
- int i;
-
- do {
- for (i = 0; i < omap_nr_pending; i++) {
- irqnr = intc_readl(INTC_PENDING_IRQ0 + (0x20 * i));
- if (irqnr)
- goto out;
- }
-
-out:
- if (!irqnr)
- break;
-
- irqnr = intc_readl(INTC_SIR);
- irqnr &= ACTIVEIRQ_MASK;
+ u32 irqnr;

- if (irqnr) {
- handle_domain_irq(domain, irqnr, regs);
- handled_irq = 1;
- }
- } while (irqnr);
-
- /*
- * If an irq is masked or deasserted while active, we will
- * keep ending up here with no irq handled. So remove it from
- * the INTC with an ack.
- */
- if (!handled_irq)
- omap_ack_irq(NULL);
+ irqnr = intc_readl(INTC_SIR);
+ irqnr &= ACTIVEIRQ_MASK;
+ WARN_ONCE(!irqnr, "Spurious IRQ ?\n");
+ handle_domain_irq(domain, irqnr, regs);
}

void __init omap3_init_irq(void)