[PATCH 0/3] powerpc/irq: Use optimizations for /proc/interrupts

STALE110d

Revision v1 of 2 in this series.

8 messages, 2 authors, 2026-05-29 · open the first message on its own page

[PATCH 0/3] powerpc/irq: Use optimizations for /proc/interrupts

From: Shrikanth Hegde <sshegde@linux.ibm.com>
Date: 2026-05-23 17:41:02

This series is based on work by Thomas Gleixner[1]. 

There are several optimization which helps to bring the time taken to
read /proc/interrupts. Though it is not in fastpaths, still many tools
read it often and this reduction in time will save meaningful amount of
cycle over the long duration.

[1]: https://lore.kernel.org/all/20260517194421.705253664@kernel.org/

Bulk of the design/code is copied from Thomas's work on x86. powerpc had
exact pattern which made it quite easy to port.

Using micro-loops[2] of reading /proc/interrupts shows real benefit of
the series. Bulk of the gain is already achieved using Thomas's
series[1]. This series on top provide some more benefits. As the system
size goes up, benefits could be more.

[2]: https://lore.kernel.org/all/87jysxw65f.ffs@tglx/

======== Performance Data ==================
Time taken to read /proc/interrupts 1000 times[2]
Base		:  103us 
v6		:   63us
v6+patch 1+2	:   57us
v6+patch 1+2+3	:   54us

Base: tip/master at c991e3897ead ("Merge branch into tip/master: 'x86/tdx'")
Depends on v6[1] to be applied first.

Shrikanth Hegde (3):
  powerpc/irq: Move __softirq_pending out of irq_stat
  powerpc/irq: Make irqstats array based
  powerpc/irq: Suppress unlikely interrupt stats by default

 arch/powerpc/include/asm/hardirq.h |  31 ++++---
 arch/powerpc/kernel/dbell.c        |   2 +-
 arch/powerpc/kernel/irq.c          | 131 +++++++++++++++--------------
 arch/powerpc/kernel/time.c         |   6 +-
 arch/powerpc/kernel/traps.c        |  11 +--
 arch/powerpc/kernel/watchdog.c     |   2 +-
 6 files changed, 95 insertions(+), 88 deletions(-)

-- 
2.47.3

[PATCH 1/3] powerpc/irq: Move __softirq_pending out of irq_stat

From: Shrikanth Hegde <sshegde@linux.ibm.com>
Date: 2026-05-23 17:40:55

__softirq_pending isn't part of arch specific irq_stats. It is used
by softirq core for various decision making such as whether to kick off
ksoftirqd. 

Move it out of irq_cpustat_t. This makes it simple to make irq_cpustat_t
array based approach.

Signed-off-by: Shrikanth Hegde <sshegde@linux.ibm.com>
---
 arch/powerpc/include/asm/hardirq.h | 3 ++-
 arch/powerpc/kernel/irq.c          | 1 +
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/include/asm/hardirq.h
index f133b5930ae1..bf3926a0c69c 100644
--- a/arch/powerpc/include/asm/hardirq.h
+++ b/arch/powerpc/include/asm/hardirq.h

@@ -6,7 +6,6 @@
 #include <linux/irq.h>
 
 typedef struct {
-	unsigned int __softirq_pending;
 	unsigned int timer_irqs_event;
 	unsigned int broadcast_irqs_event;
 	unsigned int timer_irqs_others;

@@ -23,6 +22,8 @@ typedef struct {
 } ____cacheline_aligned irq_cpustat_t;
 
 DECLARE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat);
+DECLARE_PER_CPU(unsigned int, __softirq_pending);
+#define local_softirq_pending_ref       __softirq_pending
 
 #define __ARCH_IRQ_STAT
 #define __ARCH_IRQ_EXIT_IRQS_DISABLED

diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index a0e8b998c9b5..f33df5e5c23f 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c

@@ -73,6 +73,7 @@
 
 DEFINE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat);
 EXPORT_PER_CPU_SYMBOL(irq_stat);
+DEFINE_PER_CPU(unsigned int, __softirq_pending);
 
 #ifdef CONFIG_PPC32
 atomic_t ppc_n_lost_interrupts;

-- 
2.47.3

[PATCH 3/3] powerpc/irq: Suppress unlikely interrupt stats by default

From: Shrikanth Hegde <sshegde@linux.ibm.com>
Date: 2026-05-23 17:41:06

Some interrupts are always zero and that is expected since they occur
very rarely and are mostly error indications. Don't print them by
default.

"MCE" -  "Machine check exceptions"
"NMI" -  "System Reset interrupts"

Print them if they occur once. Maintain a bitmap to know which
interrupts are to be printed.

Time taken to read /proc/interrupts 1000 times.
Base and v6 details can be found in cover-letter.
Base            :  103us
v6              :   63us
v6+patch 1+2    :   57us
v6+patch 1+2+3  :   54us

Patch 3 shows an additional 5% gain compared to patch 1+2. So it does
make sense to print them only if they are ever set.

Note: Since /proc/interrupts depend on kconfig and arch dependent,
userspace tools don't make explicit assumptions.

Signed-off-by: Shrikanth Hegde <sshegde@linux.ibm.com>
---
 arch/powerpc/include/asm/hardirq.h |  1 +
 arch/powerpc/kernel/irq.c          | 37 +++++++++++++++++++++++++++---
 arch/powerpc/kernel/traps.c        |  4 ++--
 3 files changed, 37 insertions(+), 5 deletions(-)

diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/include/asm/hardirq.h
index 38098e35b241..be6cd5aab016 100644
--- a/arch/powerpc/include/asm/hardirq.h
+++ b/arch/powerpc/include/asm/hardirq.h

@@ -31,6 +31,7 @@ DECLARE_PER_CPU(unsigned int, __softirq_pending);
 #define local_softirq_pending_ref       __softirq_pending
 
 #define inc_irq_stat(index)	__this_cpu_inc(irq_stat.counts[IRQ_COUNT_##index])
+void inc_irq_stat_and_enable(enum irq_stat_counts which);
 
 #define __ARCH_IRQ_STAT
 #define __ARCH_IRQ_EXIT_IRQS_DISABLED

diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index e67a18f62142..048ddfa66fc4 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c

@@ -87,9 +87,13 @@ u32 tau_interrupts(unsigned long cpu);
 struct irq_stat_info {
 	const char	*symbol;
 	const char	*text;
+	int		skip;
 };
 
-#define ISE(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text = txt}
+/* ISE - IRQ STAT ENABLED, ISC - IRQ STAT CONDITIONAL */
+#define ISE(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text = txt, .skip = 0}
+#define ISC(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text = txt, .skip = 1}
+
 
 static struct irq_stat_info irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
 	ISE(LOC_TIMER,		"LOC", "  Local timer interrupts for timer event device\n"),

@@ -97,8 +101,8 @@ static struct irq_stat_info irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
 	ISE(OTHER_TIMER,	"LOC", "  Local timer interrupts for others\n"),
 	ISE(SPURIOUS,		"SPU", "  Spurious interrupts\n"),
 	ISE(PMI,		"PMI", "  Performance monitoring interrupts\n"),
-	ISE(MCE,		"MCE", "  Machine check exceptions\n"),
-	ISE(NMI_SRESET,		"NMI", "  System Reset interrupts\n"),
+	ISC(MCE,		"MCE", "  Machine check exceptions\n"),
+	ISC(NMI_SRESET,		"NMI", "  System Reset interrupts\n"),
 #ifdef CONFIG_PPC_WATCHDOG
 	ISE(WATCHDOG,		"WDG", "  Watchdog soft-NMI interrupts\n"),
 #endif

@@ -107,11 +111,25 @@ static struct irq_stat_info irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
 #endif
 };
 
+/*
+ * Used for default disabled counters to increment the stats and to enable the
+ * entry for /proc/interrupts output.
+ */
+static DECLARE_BITMAP(irq_stat_count_show, IRQ_COUNT_MAX) __read_mostly;
+void inc_irq_stat_and_enable(enum irq_stat_counts which)
+{
+	__this_cpu_inc(irq_stat.counts[which]);
+	set_bit(which, irq_stat_count_show);
+}
+
 int arch_show_interrupts(struct seq_file *p, int prec)
 {
 	const struct irq_stat_info *info = irq_stat_info;
 
 	for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++, info++) {
+		if (!test_bit(i, irq_stat_count_show))
+			continue;
+
 		seq_printf(p, "%*s:", prec, info->symbol);
 		irq_proc_emit_counts(p, &irq_stat.counts[i]);
 		seq_puts(p, info->text);

@@ -138,6 +156,19 @@ int arch_show_interrupts(struct seq_file *p, int prec)
 	return 0;
 }
 
+static int __init irq_init_stats(void)
+{
+	struct irq_stat_info *info = irq_stat_info;
+
+	for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++, info++) {
+		if (info->skip == 0)
+			set_bit(i, irq_stat_count_show);
+	}
+
+	return 0;
+}
+late_initcall(irq_init_stats);
+
 /*
  * /proc/stat helpers
  */

diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index a8f15154bd9a..3eacbd20fc80 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c

@@ -459,7 +459,7 @@ DEFINE_INTERRUPT_HANDLER_NMI(system_reset_exception)
 	}
 
 	hv_nmi_check_nonrecoverable(regs);
-	inc_irq_stat(NMI_SRESET);
+	inc_irq_stat_and_enable(IRQ_COUNT_NMI_SRESET);
 
 	/* See if any machine dependent calls */
 	if (ppc_md.system_reset_exception) {

@@ -816,7 +816,7 @@ static void __machine_check_exception(struct pt_regs *regs)
 {
 	int recover = 0;
 
-	inc_irq_stat(MCE);
+	inc_irq_stat_and_enable(IRQ_COUNT_MCE);
 
 	add_taint(TAINT_MACHINE_CHECK, LOCKDEP_NOW_UNRELIABLE);

-- 
2.47.3

[PATCH 2/3] powerpc/irq: Make irqstats array based

From: Shrikanth Hegde <sshegde@linux.ibm.com>
Date: 2026-05-23 17:41:16

Current irq_cpustat_t has separate member for handling each arch
specific interrupt type. The same can be achieved with array instead
indexed by corresponding irq counter type.

This helps to,

- Make it easy to integrate into genirq improvements by calling 
  genirq provided irq_proc_emit_counts. That speeds up quite a bit
  by printing all 0's once as much as possible.

- Adding a new vector or software counter only requires to update the table
  and everything just works

- Remove ifdef usage a bit.

- Instead of going through each member, it simply becomes an array
  traversal.

Time taken to read /proc/interrupts 1000 times.
Base and v6 details can be found in cover-letter.
Base		:  103us
v6		:   63us
v6+this_patch	:   57us

A Decent 10% reduction can be seen in a system 240 CPUs. As the system
size increases the gain would be more as emitting 0 would reduce more
and more. 

Signed-off-by: Shrikanth Hegde <sshegde@linux.ibm.com>
---
 arch/powerpc/include/asm/hardirq.h |  27 +++++---
 arch/powerpc/kernel/dbell.c        |   2 +-
 arch/powerpc/kernel/irq.c          | 107 ++++++++++-------------------
 arch/powerpc/kernel/time.c         |   6 +-
 arch/powerpc/kernel/traps.c        |  11 ++-
 arch/powerpc/kernel/watchdog.c     |   2 +-
 6 files changed, 64 insertions(+), 91 deletions(-)

diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/include/asm/hardirq.h
index bf3926a0c69c..38098e35b241 100644
--- a/arch/powerpc/include/asm/hardirq.h
+++ b/arch/powerpc/include/asm/hardirq.h

@@ -5,26 +5,33 @@
 #include <linux/threads.h>
 #include <linux/irq.h>
 
-typedef struct {
-	unsigned int timer_irqs_event;
-	unsigned int broadcast_irqs_event;
-	unsigned int timer_irqs_others;
-	unsigned int pmu_irqs;
-	unsigned int mce_exceptions;
-	unsigned int spurious_irqs;
-	unsigned int sreset_irqs;
+enum irq_stat_counts {
+	IRQ_COUNT_LOC_TIMER,
+	IRQ_COUNT_BCT_TIMER,
+	IRQ_COUNT_OTHER_TIMER,
+	IRQ_COUNT_SPURIOUS,
+	IRQ_COUNT_PMI,
+	IRQ_COUNT_MCE,
+	IRQ_COUNT_NMI_SRESET,
 #ifdef CONFIG_PPC_WATCHDOG
-	unsigned int soft_nmi_irqs;
+	IRQ_COUNT_WATCHDOG,
 #endif
 #ifdef CONFIG_PPC_DOORBELL
-	unsigned int doorbell_irqs;
+	IRQ_COUNT_DOORBELL,
 #endif
+	IRQ_COUNT_MAX,
+};
+
+typedef struct {
+	unsigned int counts[IRQ_COUNT_MAX];
 } ____cacheline_aligned irq_cpustat_t;
 
 DECLARE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat);
 DECLARE_PER_CPU(unsigned int, __softirq_pending);
 #define local_softirq_pending_ref       __softirq_pending
 
+#define inc_irq_stat(index)	__this_cpu_inc(irq_stat.counts[IRQ_COUNT_##index])
+
 #define __ARCH_IRQ_STAT
 #define __ARCH_IRQ_EXIT_IRQS_DISABLED

diff --git a/arch/powerpc/kernel/dbell.c b/arch/powerpc/kernel/dbell.c
index 5712dd846263..f5e298a4c4c0 100644
--- a/arch/powerpc/kernel/dbell.c
+++ b/arch/powerpc/kernel/dbell.c

@@ -31,7 +31,7 @@ DEFINE_INTERRUPT_HANDLER_ASYNC(doorbell_exception)
 		do_hard_irq_enable();
 
 	kvmppc_clear_host_ipi(smp_processor_id());
-	__this_cpu_inc(irq_stat.doorbell_irqs);
+	inc_irq_stat(DOORBELL);
 
 	smp_ipi_demux_relaxed(); /* already performed the barrier */

diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index f33df5e5c23f..e67a18f62142 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c

@@ -84,79 +84,57 @@ u32 tau_interrupts(unsigned long cpu);
 #endif
 #endif /* CONFIG_PPC32 */
 
+struct irq_stat_info {
+	const char	*symbol;
+	const char	*text;
+};
+
+#define ISE(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text = txt}
+
+static struct irq_stat_info irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
+	ISE(LOC_TIMER,		"LOC", "  Local timer interrupts for timer event device\n"),
+	ISE(BCT_TIMER,		"BCT", "  Broadcast timer interrupts for timer event device\n"),
+	ISE(OTHER_TIMER,	"LOC", "  Local timer interrupts for others\n"),
+	ISE(SPURIOUS,		"SPU", "  Spurious interrupts\n"),
+	ISE(PMI,		"PMI", "  Performance monitoring interrupts\n"),
+	ISE(MCE,		"MCE", "  Machine check exceptions\n"),
+	ISE(NMI_SRESET,		"NMI", "  System Reset interrupts\n"),
+#ifdef CONFIG_PPC_WATCHDOG
+	ISE(WATCHDOG,		"WDG", "  Watchdog soft-NMI interrupts\n"),
+#endif
+#ifdef CONFIG_PPC_DOORBELL
+	ISE(DOORBELL,		"DBL", "  Doorbell interrupts\n"),
+#endif
+};
+
 int arch_show_interrupts(struct seq_file *p, int prec)
 {
-	int j;
+	const struct irq_stat_info *info = irq_stat_info;
+
+	for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++, info++) {
+		seq_printf(p, "%*s:", prec, info->symbol);
+		irq_proc_emit_counts(p, &irq_stat.counts[i]);
+		seq_puts(p, info->text);
+	}
 
 #if defined(CONFIG_PPC32) && defined(CONFIG_TAU_INT)
 	if (tau_initialized) {
+		int j;
 		seq_printf(p, "%*s:", prec, "TAU");
 		for_each_online_cpu(j)
 			seq_put_decimal_ull_width(p, " ", tau_interrupts(j), 10);
 		seq_puts(p, "  PowerPC             Thermal Assist (cpu temp)\n");
 	}
 #endif /* CONFIG_PPC32 && CONFIG_TAU_INT */
-
-	seq_printf(p, "%*s:", prec, "LOC");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).timer_irqs_event, 10);
-        seq_printf(p, "  Local timer interrupts for timer event device\n");
-
-	seq_printf(p, "%*s:", prec, "BCT");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).broadcast_irqs_event, 10);
-	seq_printf(p, "  Broadcast timer interrupts for timer event device\n");
-
-	seq_printf(p, "%*s:", prec, "LOC");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).timer_irqs_others, 10);
-        seq_printf(p, "  Local timer interrupts for others\n");
-
-	seq_printf(p, "%*s:", prec, "SPU");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).spurious_irqs, 10);
-	seq_printf(p, "  Spurious interrupts\n");
-
-	seq_printf(p, "%*s:", prec, "PMI");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).pmu_irqs, 10);
-	seq_printf(p, "  Performance monitoring interrupts\n");
-
-	seq_printf(p, "%*s:", prec, "MCE");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).mce_exceptions, 10);
-	seq_printf(p, "  Machine check exceptions\n");
-
 #ifdef CONFIG_PPC_BOOK3S_64
 	if (cpu_has_feature(CPU_FTR_HVMODE)) {
+		int j;
 		seq_printf(p, "%*s:", prec, "HMI");
 		for_each_online_cpu(j)
 			seq_put_decimal_ull_width(p, " ", paca_ptrs[j]->hmi_irqs, 10);
 		seq_printf(p, "  Hypervisor Maintenance Interrupts\n");
 	}
 #endif
-
-	seq_printf(p, "%*s:", prec, "NMI");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).sreset_irqs, 10);
-	seq_printf(p, "  System Reset interrupts\n");
-
-#ifdef CONFIG_PPC_WATCHDOG
-	seq_printf(p, "%*s:", prec, "WDG");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).soft_nmi_irqs, 10);
-	seq_printf(p, "  Watchdog soft-NMI interrupts\n");
-#endif
-
-#ifdef CONFIG_PPC_DOORBELL
-	if (cpu_has_feature(CPU_FTR_DBELL)) {
-		seq_printf(p, "%*s:", prec, "DBL");
-		for_each_online_cpu(j)
-			seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).doorbell_irqs, 10);
-		seq_printf(p, "  Doorbell interrupts\n");
-	}
-#endif
-
 	return 0;
 }

@@ -165,24 +143,15 @@ int arch_show_interrupts(struct seq_file *p, int prec)
  */
 u64 arch_irq_stat_cpu(unsigned int cpu)
 {
-	u64 sum = per_cpu(irq_stat, cpu).timer_irqs_event;
+	irq_cpustat_t *p = per_cpu_ptr(&irq_stat, cpu);
+	u64 sum = 0;
+
+	for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++)
+		sum += p->counts[i];
 
-	sum += per_cpu(irq_stat, cpu).broadcast_irqs_event;
-	sum += per_cpu(irq_stat, cpu).pmu_irqs;
-	sum += per_cpu(irq_stat, cpu).mce_exceptions;
-	sum += per_cpu(irq_stat, cpu).spurious_irqs;
-	sum += per_cpu(irq_stat, cpu).timer_irqs_others;
 #ifdef CONFIG_PPC_BOOK3S_64
 	sum += paca_ptrs[cpu]->hmi_irqs;
 #endif
-	sum += per_cpu(irq_stat, cpu).sreset_irqs;
-#ifdef CONFIG_PPC_WATCHDOG
-	sum += per_cpu(irq_stat, cpu).soft_nmi_irqs;
-#endif
-#ifdef CONFIG_PPC_DOORBELL
-	sum += per_cpu(irq_stat, cpu).doorbell_irqs;
-#endif
-
 	return sum;
 }

@@ -248,7 +217,7 @@ static void __do_irq(struct pt_regs *regs, unsigned long oldsp)
 
 	/* And finally process it */
 	if (unlikely(!irq))
-		__this_cpu_inc(irq_stat.spurious_irqs);
+		inc_irq_stat(SPURIOUS);
 	else
 		generic_handle_irq(irq);

diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c
index 4bbeb8644d3d..44da7be36199 100644
--- a/arch/powerpc/kernel/time.c
+++ b/arch/powerpc/kernel/time.c

@@ -572,13 +572,13 @@ DEFINE_INTERRUPT_HANDLER_ASYNC(timer_interrupt)
 	now = get_tb();
 	if (now >= *next_tb) {
 		evt->event_handler(evt);
-		__this_cpu_inc(irq_stat.timer_irqs_event);
+		inc_irq_stat(LOC_TIMER);
 	} else {
 		now = *next_tb - now;
 		if (now > decrementer_max)
 			now = decrementer_max;
 		set_dec_or_work(now);
-		__this_cpu_inc(irq_stat.timer_irqs_others);
+		inc_irq_stat(OTHER_TIMER);
 	}
 
 	trace_timer_interrupt_exit(regs);

@@ -591,7 +591,7 @@ EXPORT_SYMBOL(timer_interrupt);
 void timer_broadcast_interrupt(void)
 {
 	tick_receive_broadcast();
-	__this_cpu_inc(irq_stat.broadcast_irqs_event);
+	inc_irq_stat(BCT_TIMER);
 }
 #endif

diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index cb8e9357383e..a8f15154bd9a 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c

@@ -459,8 +459,7 @@ DEFINE_INTERRUPT_HANDLER_NMI(system_reset_exception)
 	}
 
 	hv_nmi_check_nonrecoverable(regs);
-
-	__this_cpu_inc(irq_stat.sreset_irqs);
+	inc_irq_stat(NMI_SRESET);
 
 	/* See if any machine dependent calls */
 	if (ppc_md.system_reset_exception) {

@@ -817,7 +816,7 @@ static void __machine_check_exception(struct pt_regs *regs)
 {
 	int recover = 0;
 
-	__this_cpu_inc(irq_stat.mce_exceptions);
+	inc_irq_stat(MCE);
 
 	add_taint(TAINT_MACHINE_CHECK, LOCKDEP_NOW_UNRELIABLE);

@@ -1932,8 +1931,7 @@ DEFINE_INTERRUPT_HANDLER(vsx_unavailable_tm)
 DECLARE_INTERRUPT_HANDLER_NMI(performance_monitor_exception_nmi);
 DEFINE_INTERRUPT_HANDLER_NMI(performance_monitor_exception_nmi)
 {
-	__this_cpu_inc(irq_stat.pmu_irqs);
-
+	inc_irq_stat(PMI);
 	perf_irq(regs);
 
 	return 0;

@@ -1943,8 +1941,7 @@ DEFINE_INTERRUPT_HANDLER_NMI(performance_monitor_exception_nmi)
 DECLARE_INTERRUPT_HANDLER_ASYNC(performance_monitor_exception_async);
 DEFINE_INTERRUPT_HANDLER_ASYNC(performance_monitor_exception_async)
 {
-	__this_cpu_inc(irq_stat.pmu_irqs);
-
+	inc_irq_stat(PMI);
 	perf_irq(regs);
 }

diff --git a/arch/powerpc/kernel/watchdog.c b/arch/powerpc/kernel/watchdog.c
index 764001deb060..f516eeccc9f6 100644
--- a/arch/powerpc/kernel/watchdog.c
+++ b/arch/powerpc/kernel/watchdog.c

@@ -381,7 +381,7 @@ DEFINE_INTERRUPT_HANDLER_NMI(soft_nmi_interrupt)
 	if (!cpumask_test_cpu(cpu, &wd_cpus_enabled))
 		return 0;
 
-	__this_cpu_inc(irq_stat.soft_nmi_irqs);
+	inc_irq_stat(WATCHDOG);
 
 	tb = get_tb();
 	if (tb - per_cpu(wd_timer_tb, cpu) >= wd_panic_timeout_tb) {

-- 
2.47.3

Re: [PATCH 1/3] powerpc/irq: Move __softirq_pending out of irq_stat

From: "Christophe Leroy (CS GROUP)" <chleroy@kernel.org>
Date: 2026-05-29 07:43:22


Le 23/05/2026 à 19:40, Shrikanth Hegde a écrit :

__softirq_pending isn't part of arch specific irq_stats. It is used
by softirq core for various decision making such as whether to kick off
ksoftirqd.

Move it out of irq_cpustat_t. This makes it simple to make irq_cpustat_t
array based approach.

Some reference to commit 0fd7d8628529 ("softirq/core: Consolidate 
default local_softirq_pending() implementations") should be made I think 
for better understanding.

Signed-off-by: Shrikanth Hegde <sshegde@linux.ibm.com>

Reviewed-by: Christophe Leroy (CS GROUP) <chleroy@kernel.org>

quoted hunk

---
  arch/powerpc/include/asm/hardirq.h | 3 ++-
  arch/powerpc/kernel/irq.c          | 1 +
  2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/include/asm/hardirq.h
index f133b5930ae1..bf3926a0c69c 100644
--- a/arch/powerpc/include/asm/hardirq.h
+++ b/arch/powerpc/include/asm/hardirq.h

@@ -6,7 +6,6 @@
  #include <linux/irq.h>
  
  typedef struct {
-	unsigned int __softirq_pending;
  	unsigned int timer_irqs_event;
  	unsigned int broadcast_irqs_event;
  	unsigned int timer_irqs_others;

@@ -23,6 +22,8 @@ typedef struct {
  } ____cacheline_aligned irq_cpustat_t;
  
  DECLARE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat);
+DECLARE_PER_CPU(unsigned int, __softirq_pending);
+#define local_softirq_pending_ref       __softirq_pending
  
  #define __ARCH_IRQ_STAT
  #define __ARCH_IRQ_EXIT_IRQS_DISABLED

diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index a0e8b998c9b5..f33df5e5c23f 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c

@@ -73,6 +73,7 @@
  
  DEFINE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat);
  EXPORT_PER_CPU_SYMBOL(irq_stat);
+DEFINE_PER_CPU(unsigned int, __softirq_pending);
  
  #ifdef CONFIG_PPC32
  atomic_t ppc_n_lost_interrupts;

Re: [PATCH 2/3] powerpc/irq: Make irqstats array based

From: "Christophe Leroy (CS GROUP)" <chleroy@kernel.org>
Date: 2026-05-29 07:46:43


Le 23/05/2026 à 19:40, Shrikanth Hegde a écrit :

Current irq_cpustat_t has separate member for handling each arch
specific interrupt type. The same can be achieved with array instead
indexed by corresponding irq counter type.

This helps to,

- Make it easy to integrate into genirq improvements by calling
   genirq provided irq_proc_emit_counts. That speeds up quite a bit
   by printing all 0's once as much as possible.

- Adding a new vector or software counter only requires to update the table
   and everything just works

- Remove ifdef usage a bit.

- Instead of going through each member, it simply becomes an array
   traversal.

Time taken to read /proc/interrupts 1000 times.
Base and v6 details can be found in cover-letter.
Base		:  103us
v6		:   63us
v6+this_patch	:   57us

A Decent 10% reduction can be seen in a system 240 CPUs. As the system
size increases the gain would be more as emitting 0 would reduce more
and more.

Signed-off-by: Shrikanth Hegde <sshegde@linux.ibm.com>

Reviewed-by: Christophe Leroy (CS GROUP) <chleroy@kernel.org>

quoted hunk

---
  arch/powerpc/include/asm/hardirq.h |  27 +++++---
  arch/powerpc/kernel/dbell.c        |   2 +-
  arch/powerpc/kernel/irq.c          | 107 ++++++++++-------------------
  arch/powerpc/kernel/time.c         |   6 +-
  arch/powerpc/kernel/traps.c        |  11 ++-
  arch/powerpc/kernel/watchdog.c     |   2 +-
  6 files changed, 64 insertions(+), 91 deletions(-)

diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/include/asm/hardirq.h
index bf3926a0c69c..38098e35b241 100644
--- a/arch/powerpc/include/asm/hardirq.h
+++ b/arch/powerpc/include/asm/hardirq.h

@@ -5,26 +5,33 @@
  #include <linux/threads.h>
  #include <linux/irq.h>
  
-typedef struct {
-	unsigned int timer_irqs_event;
-	unsigned int broadcast_irqs_event;
-	unsigned int timer_irqs_others;
-	unsigned int pmu_irqs;
-	unsigned int mce_exceptions;
-	unsigned int spurious_irqs;
-	unsigned int sreset_irqs;
+enum irq_stat_counts {
+	IRQ_COUNT_LOC_TIMER,
+	IRQ_COUNT_BCT_TIMER,
+	IRQ_COUNT_OTHER_TIMER,
+	IRQ_COUNT_SPURIOUS,
+	IRQ_COUNT_PMI,
+	IRQ_COUNT_MCE,
+	IRQ_COUNT_NMI_SRESET,
  #ifdef CONFIG_PPC_WATCHDOG
-	unsigned int soft_nmi_irqs;
+	IRQ_COUNT_WATCHDOG,
  #endif
  #ifdef CONFIG_PPC_DOORBELL
-	unsigned int doorbell_irqs;
+	IRQ_COUNT_DOORBELL,
  #endif
+	IRQ_COUNT_MAX,
+};
+
+typedef struct {
+	unsigned int counts[IRQ_COUNT_MAX];
  } ____cacheline_aligned irq_cpustat_t;
  
  DECLARE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat);
  DECLARE_PER_CPU(unsigned int, __softirq_pending);
  #define local_softirq_pending_ref       __softirq_pending
  
+#define inc_irq_stat(index)	__this_cpu_inc(irq_stat.counts[IRQ_COUNT_##index])
+
  #define __ARCH_IRQ_STAT
  #define __ARCH_IRQ_EXIT_IRQS_DISABLED

diff --git a/arch/powerpc/kernel/dbell.c b/arch/powerpc/kernel/dbell.c
index 5712dd846263..f5e298a4c4c0 100644
--- a/arch/powerpc/kernel/dbell.c
+++ b/arch/powerpc/kernel/dbell.c

@@ -31,7 +31,7 @@ DEFINE_INTERRUPT_HANDLER_ASYNC(doorbell_exception)
  		do_hard_irq_enable();
  
  	kvmppc_clear_host_ipi(smp_processor_id());
-	__this_cpu_inc(irq_stat.doorbell_irqs);
+	inc_irq_stat(DOORBELL);
  
  	smp_ipi_demux_relaxed(); /* already performed the barrier */

diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index f33df5e5c23f..e67a18f62142 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c

@@ -84,79 +84,57 @@ u32 tau_interrupts(unsigned long cpu);
  #endif
  #endif /* CONFIG_PPC32 */
  
+struct irq_stat_info {
+	const char	*symbol;
+	const char	*text;
+};
+
+#define ISE(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text = txt}
+
+static struct irq_stat_info irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
+	ISE(LOC_TIMER,		"LOC", "  Local timer interrupts for timer event device\n"),
+	ISE(BCT_TIMER,		"BCT", "  Broadcast timer interrupts for timer event device\n"),
+	ISE(OTHER_TIMER,	"LOC", "  Local timer interrupts for others\n"),
+	ISE(SPURIOUS,		"SPU", "  Spurious interrupts\n"),
+	ISE(PMI,		"PMI", "  Performance monitoring interrupts\n"),
+	ISE(MCE,		"MCE", "  Machine check exceptions\n"),
+	ISE(NMI_SRESET,		"NMI", "  System Reset interrupts\n"),
+#ifdef CONFIG_PPC_WATCHDOG
+	ISE(WATCHDOG,		"WDG", "  Watchdog soft-NMI interrupts\n"),
+#endif
+#ifdef CONFIG_PPC_DOORBELL
+	ISE(DOORBELL,		"DBL", "  Doorbell interrupts\n"),
+#endif
+};
+
  int arch_show_interrupts(struct seq_file *p, int prec)
  {
-	int j;
+	const struct irq_stat_info *info = irq_stat_info;
+
+	for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++, info++) {
+		seq_printf(p, "%*s:", prec, info->symbol);
+		irq_proc_emit_counts(p, &irq_stat.counts[i]);
+		seq_puts(p, info->text);
+	}
  
  #if defined(CONFIG_PPC32) && defined(CONFIG_TAU_INT)
  	if (tau_initialized) {
+		int j;
  		seq_printf(p, "%*s:", prec, "TAU");
  		for_each_online_cpu(j)
  			seq_put_decimal_ull_width(p, " ", tau_interrupts(j), 10);
  		seq_puts(p, "  PowerPC             Thermal Assist (cpu temp)\n");
  	}
  #endif /* CONFIG_PPC32 && CONFIG_TAU_INT */
-
-	seq_printf(p, "%*s:", prec, "LOC");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).timer_irqs_event, 10);
-        seq_printf(p, "  Local timer interrupts for timer event device\n");
-
-	seq_printf(p, "%*s:", prec, "BCT");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).broadcast_irqs_event, 10);
-	seq_printf(p, "  Broadcast timer interrupts for timer event device\n");
-
-	seq_printf(p, "%*s:", prec, "LOC");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).timer_irqs_others, 10);
-        seq_printf(p, "  Local timer interrupts for others\n");
-
-	seq_printf(p, "%*s:", prec, "SPU");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).spurious_irqs, 10);
-	seq_printf(p, "  Spurious interrupts\n");
-
-	seq_printf(p, "%*s:", prec, "PMI");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).pmu_irqs, 10);
-	seq_printf(p, "  Performance monitoring interrupts\n");
-
-	seq_printf(p, "%*s:", prec, "MCE");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).mce_exceptions, 10);
-	seq_printf(p, "  Machine check exceptions\n");
-
  #ifdef CONFIG_PPC_BOOK3S_64
  	if (cpu_has_feature(CPU_FTR_HVMODE)) {
+		int j;
  		seq_printf(p, "%*s:", prec, "HMI");
  		for_each_online_cpu(j)
  			seq_put_decimal_ull_width(p, " ", paca_ptrs[j]->hmi_irqs, 10);
  		seq_printf(p, "  Hypervisor Maintenance Interrupts\n");
  	}
  #endif
-
-	seq_printf(p, "%*s:", prec, "NMI");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).sreset_irqs, 10);
-	seq_printf(p, "  System Reset interrupts\n");
-
-#ifdef CONFIG_PPC_WATCHDOG
-	seq_printf(p, "%*s:", prec, "WDG");
-	for_each_online_cpu(j)
-		seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).soft_nmi_irqs, 10);
-	seq_printf(p, "  Watchdog soft-NMI interrupts\n");
-#endif
-
-#ifdef CONFIG_PPC_DOORBELL
-	if (cpu_has_feature(CPU_FTR_DBELL)) {
-		seq_printf(p, "%*s:", prec, "DBL");
-		for_each_online_cpu(j)
-			seq_put_decimal_ull_width(p, " ", per_cpu(irq_stat, j).doorbell_irqs, 10);
-		seq_printf(p, "  Doorbell interrupts\n");
-	}
-#endif
-
  	return 0;
  }

@@ -165,24 +143,15 @@ int arch_show_interrupts(struct seq_file *p, int prec)
   */
  u64 arch_irq_stat_cpu(unsigned int cpu)
  {
-	u64 sum = per_cpu(irq_stat, cpu).timer_irqs_event;
+	irq_cpustat_t *p = per_cpu_ptr(&irq_stat, cpu);
+	u64 sum = 0;
+
+	for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++)
+		sum += p->counts[i];
  
-	sum += per_cpu(irq_stat, cpu).broadcast_irqs_event;
-	sum += per_cpu(irq_stat, cpu).pmu_irqs;
-	sum += per_cpu(irq_stat, cpu).mce_exceptions;
-	sum += per_cpu(irq_stat, cpu).spurious_irqs;
-	sum += per_cpu(irq_stat, cpu).timer_irqs_others;
  #ifdef CONFIG_PPC_BOOK3S_64
  	sum += paca_ptrs[cpu]->hmi_irqs;
  #endif
-	sum += per_cpu(irq_stat, cpu).sreset_irqs;
-#ifdef CONFIG_PPC_WATCHDOG
-	sum += per_cpu(irq_stat, cpu).soft_nmi_irqs;
-#endif
-#ifdef CONFIG_PPC_DOORBELL
-	sum += per_cpu(irq_stat, cpu).doorbell_irqs;
-#endif
-
  	return sum;
  }

@@ -248,7 +217,7 @@ static void __do_irq(struct pt_regs *regs, unsigned long oldsp)
  
  	/* And finally process it */
  	if (unlikely(!irq))
-		__this_cpu_inc(irq_stat.spurious_irqs);
+		inc_irq_stat(SPURIOUS);
  	else
  		generic_handle_irq(irq);

diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c
index 4bbeb8644d3d..44da7be36199 100644
--- a/arch/powerpc/kernel/time.c
+++ b/arch/powerpc/kernel/time.c

@@ -572,13 +572,13 @@ DEFINE_INTERRUPT_HANDLER_ASYNC(timer_interrupt)
  	now = get_tb();
  	if (now >= *next_tb) {
  		evt->event_handler(evt);
-		__this_cpu_inc(irq_stat.timer_irqs_event);
+		inc_irq_stat(LOC_TIMER);
  	} else {
  		now = *next_tb - now;
  		if (now > decrementer_max)
  			now = decrementer_max;
  		set_dec_or_work(now);
-		__this_cpu_inc(irq_stat.timer_irqs_others);
+		inc_irq_stat(OTHER_TIMER);
  	}
  
  	trace_timer_interrupt_exit(regs);

@@ -591,7 +591,7 @@ EXPORT_SYMBOL(timer_interrupt);
  void timer_broadcast_interrupt(void)
  {
  	tick_receive_broadcast();
-	__this_cpu_inc(irq_stat.broadcast_irqs_event);
+	inc_irq_stat(BCT_TIMER);
  }
  #endif

diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index cb8e9357383e..a8f15154bd9a 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c

@@ -459,8 +459,7 @@ DEFINE_INTERRUPT_HANDLER_NMI(system_reset_exception)
  	}
  
  	hv_nmi_check_nonrecoverable(regs);
-
-	__this_cpu_inc(irq_stat.sreset_irqs);
+	inc_irq_stat(NMI_SRESET);
  
  	/* See if any machine dependent calls */
  	if (ppc_md.system_reset_exception) {

@@ -817,7 +816,7 @@ static void __machine_check_exception(struct pt_regs *regs)
  {
  	int recover = 0;
  
-	__this_cpu_inc(irq_stat.mce_exceptions);
+	inc_irq_stat(MCE);
  
  	add_taint(TAINT_MACHINE_CHECK, LOCKDEP_NOW_UNRELIABLE);

@@ -1932,8 +1931,7 @@ DEFINE_INTERRUPT_HANDLER(vsx_unavailable_tm)
  DECLARE_INTERRUPT_HANDLER_NMI(performance_monitor_exception_nmi);
  DEFINE_INTERRUPT_HANDLER_NMI(performance_monitor_exception_nmi)
  {
-	__this_cpu_inc(irq_stat.pmu_irqs);
-
+	inc_irq_stat(PMI);
  	perf_irq(regs);
  
  	return 0;

@@ -1943,8 +1941,7 @@ DEFINE_INTERRUPT_HANDLER_NMI(performance_monitor_exception_nmi)
  DECLARE_INTERRUPT_HANDLER_ASYNC(performance_monitor_exception_async);
  DEFINE_INTERRUPT_HANDLER_ASYNC(performance_monitor_exception_async)
  {
-	__this_cpu_inc(irq_stat.pmu_irqs);
-
+	inc_irq_stat(PMI);
  	perf_irq(regs);
  }

diff --git a/arch/powerpc/kernel/watchdog.c b/arch/powerpc/kernel/watchdog.c
index 764001deb060..f516eeccc9f6 100644
--- a/arch/powerpc/kernel/watchdog.c
+++ b/arch/powerpc/kernel/watchdog.c

@@ -381,7 +381,7 @@ DEFINE_INTERRUPT_HANDLER_NMI(soft_nmi_interrupt)
  	if (!cpumask_test_cpu(cpu, &wd_cpus_enabled))
  		return 0;
  
-	__this_cpu_inc(irq_stat.soft_nmi_irqs);
+	inc_irq_stat(WATCHDOG);
  
  	tb = get_tb();
  	if (tb - per_cpu(wd_timer_tb, cpu) >= wd_panic_timeout_tb) {

Re: [PATCH 3/3] powerpc/irq: Suppress unlikely interrupt stats by default

From: "Christophe Leroy (CS GROUP)" <chleroy@kernel.org>
Date: 2026-05-29 07:51:39


Le 23/05/2026 à 19:40, Shrikanth Hegde a écrit :

Some interrupts are always zero and that is expected since they occur
very rarely and are mostly error indications. Don't print them by
default.

"MCE" -  "Machine check exceptions"
"NMI" -  "System Reset interrupts"

Print them if they occur once. Maintain a bitmap to know which
interrupts are to be printed.

Is that bitmap needed at all ? Can't we just print them as soon as they 
are not zero ?

quoted hunk

Time taken to read /proc/interrupts 1000 times.
Base and v6 details can be found in cover-letter.
Base            :  103us
v6              :   63us
v6+patch 1+2    :   57us
v6+patch 1+2+3  :   54us

Patch 3 shows an additional 5% gain compared to patch 1+2. So it does
make sense to print them only if they are ever set.

Note: Since /proc/interrupts depend on kconfig and arch dependent,
userspace tools don't make explicit assumptions.

Signed-off-by: Shrikanth Hegde <sshegde@linux.ibm.com>
---
  arch/powerpc/include/asm/hardirq.h |  1 +
  arch/powerpc/kernel/irq.c          | 37 +++++++++++++++++++++++++++---
  arch/powerpc/kernel/traps.c        |  4 ++--
  3 files changed, 37 insertions(+), 5 deletions(-)

diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/include/asm/hardirq.h
index 38098e35b241..be6cd5aab016 100644
--- a/arch/powerpc/include/asm/hardirq.h
+++ b/arch/powerpc/include/asm/hardirq.h

@@ -31,6 +31,7 @@ DECLARE_PER_CPU(unsigned int, __softirq_pending);
  #define local_softirq_pending_ref       __softirq_pending
  
  #define inc_irq_stat(index)	__this_cpu_inc(irq_stat.counts[IRQ_COUNT_##index])
+void inc_irq_stat_and_enable(enum irq_stat_counts which);
  
  #define __ARCH_IRQ_STAT
  #define __ARCH_IRQ_EXIT_IRQS_DISABLED

diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index e67a18f62142..048ddfa66fc4 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c

@@ -87,9 +87,13 @@ u32 tau_interrupts(unsigned long cpu);
  struct irq_stat_info {
  	const char	*symbol;
  	const char	*text;
+	int		skip;

I'd call it 'optional' instead, and then during the print, if value 0 
and optional then don't print.

quoted hunk

  };
  
-#define ISE(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text = txt}
+/* ISE - IRQ STAT ENABLED, ISC - IRQ STAT CONDITIONAL */
+#define ISE(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text = txt, .skip = 0}
+#define ISC(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text = txt, .skip = 1}
+
  
  static struct irq_stat_info irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
  	ISE(LOC_TIMER,		"LOC", "  Local timer interrupts for timer event device\n"),

@@ -97,8 +101,8 @@ static struct irq_stat_info irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
  	ISE(OTHER_TIMER,	"LOC", "  Local timer interrupts for others\n"),
  	ISE(SPURIOUS,		"SPU", "  Spurious interrupts\n"),
  	ISE(PMI,		"PMI", "  Performance monitoring interrupts\n"),
-	ISE(MCE,		"MCE", "  Machine check exceptions\n"),
-	ISE(NMI_SRESET,		"NMI", "  System Reset interrupts\n"),
+	ISC(MCE,		"MCE", "  Machine check exceptions\n"),
+	ISC(NMI_SRESET,		"NMI", "  System Reset interrupts\n"),
  #ifdef CONFIG_PPC_WATCHDOG
  	ISE(WATCHDOG,		"WDG", "  Watchdog soft-NMI interrupts\n"),
  #endif

@@ -107,11 +111,25 @@ static struct irq_stat_info irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
  #endif
  };
  
+/*
+ * Used for default disabled counters to increment the stats and to enable the
+ * entry for /proc/interrupts output.
+ */
+static DECLARE_BITMAP(irq_stat_count_show, IRQ_COUNT_MAX) __read_mostly;
+void inc_irq_stat_and_enable(enum irq_stat_counts which)
+{
+	__this_cpu_inc(irq_stat.counts[which]);
+	set_bit(which, irq_stat_count_show);
+}
+
  int arch_show_interrupts(struct seq_file *p, int prec)
  {
  	const struct irq_stat_info *info = irq_stat_info;
  
  	for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++, info++) {
+		if (!test_bit(i, irq_stat_count_show))
+			continue;
+
  		seq_printf(p, "%*s:", prec, info->symbol);
  		irq_proc_emit_counts(p, &irq_stat.counts[i]);
  		seq_puts(p, info->text);

@@ -138,6 +156,19 @@ int arch_show_interrupts(struct seq_file *p, int prec)
  	return 0;
  }
  
+static int __init irq_init_stats(void)
+{
+	struct irq_stat_info *info = irq_stat_info;
+
+	for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++, info++) {
+		if (info->skip == 0)
+			set_bit(i, irq_stat_count_show);
+	}
+
+	return 0;
+}
+late_initcall(irq_init_stats);
+
  /*
   * /proc/stat helpers
   */

diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index a8f15154bd9a..3eacbd20fc80 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c

@@ -459,7 +459,7 @@ DEFINE_INTERRUPT_HANDLER_NMI(system_reset_exception)
  	}
  
  	hv_nmi_check_nonrecoverable(regs);
-	inc_irq_stat(NMI_SRESET);
+	inc_irq_stat_and_enable(IRQ_COUNT_NMI_SRESET);
  
  	/* See if any machine dependent calls */
  	if (ppc_md.system_reset_exception) {

@@ -816,7 +816,7 @@ static void __machine_check_exception(struct pt_regs *regs)
  {
  	int recover = 0;
  
-	inc_irq_stat(MCE);
+	inc_irq_stat_and_enable(IRQ_COUNT_MCE);
  
  	add_taint(TAINT_MACHINE_CHECK, LOCKDEP_NOW_UNRELIABLE);

Re: [PATCH 3/3] powerpc/irq: Suppress unlikely interrupt stats by default

From: Shrikanth Hegde <sshegde@linux.ibm.com>
Date: 2026-05-29 08:09:32

Hi Christophe.
Thanks for reviewing the patch series.

On 5/29/26 1:21 PM, Christophe Leroy (CS GROUP) wrote:


Le 23/05/2026 à 19:40, Shrikanth Hegde a écrit :

quoted

Some interrupts are always zero and that is expected since they occur
very rarely and are mostly error indications. Don't print them by
default.

"MCE" -  "Machine check exceptions"
"NMI" -  "System Reset interrupts"

Print them if they occur once. Maintain a bitmap to know which
interrupts are to be printed.

Is that bitmap needed at all ? Can't we just print them as soon as they 
are not zero ?

I think Yes. Otherwise need to traverse all the per_cpu count and to see 
if it is zero or not.

quoted

Time taken to read /proc/interrupts 1000 times.
Base and v6 details can be found in cover-letter.
Base            :  103us
v6              :   63us
v6+patch 1+2    :   57us
v6+patch 1+2+3  :   54us

Patch 3 shows an additional 5% gain compared to patch 1+2. So it does
make sense to print them only if they are ever set.

Note: Since /proc/interrupts depend on kconfig and arch dependent,
userspace tools don't make explicit assumptions.

Signed-off-by: Shrikanth Hegde <sshegde@linux.ibm.com>
---
  arch/powerpc/include/asm/hardirq.h |  1 +
  arch/powerpc/kernel/irq.c          | 37 +++++++++++++++++++++++++++---
  arch/powerpc/kernel/traps.c        |  4 ++--
  3 files changed, 37 insertions(+), 5 deletions(-)

diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/

include/asm/hardirq.h
index 38098e35b241..be6cd5aab016 100644

--- a/arch/powerpc/include/asm/hardirq.h
+++ b/arch/powerpc/include/asm/hardirq.h

@@ -31,6 +31,7 @@ DECLARE_PER_CPU(unsigned int, __softirq_pending);

  #define local_softirq_pending_ref       __softirq_pending
  #define inc_irq_stat(index)    
__this_cpu_inc(irq_stat.counts[IRQ_COUNT_##index])
+void inc_irq_stat_and_enable(enum irq_stat_counts which);
  #define __ARCH_IRQ_STAT
  #define __ARCH_IRQ_EXIT_IRQS_DISABLED

diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index e67a18f62142..048ddfa66fc4 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c

@@ -87,9 +87,13 @@ u32 tau_interrupts(unsigned long cpu);

  struct irq_stat_info {
      const char    *symbol;
      const char    *text;
+    int        skip;

I'd call it 'optional' instead, and then during the print, if value 0 
and optional then don't print.

Ok. Makes sense. Will do it in v2.

Will add below way

         if (!(info->optional && test_bit(i, irq_stat_count_show)))
             continue;

quoted

  };
-#define ISE(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text 
= txt}
+/* ISE - IRQ STAT ENABLED, ISC - IRQ STAT CONDITIONAL */
+#define ISE(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text 
= txt, .skip = 0}
+#define ISC(idx, sym, txt)[IRQ_COUNT_##idx] = { .symbol = sym, .text 
= txt, .skip = 1}
+
  static struct irq_stat_info irq_stat_info[IRQ_COUNT_MAX] 
__ro_after_init = {
      ISE(LOC_TIMER,        "LOC", "  Local timer interrupts for timer 
event device\n"),

@@ -97,8 +101,8 @@ static struct irq_stat_info

irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
      ISE(OTHER_TIMER,    "LOC", "  Local timer interrupts for 
others\n"),
      ISE(SPURIOUS,        "SPU", "  Spurious interrupts\n"),
      ISE(PMI,        "PMI", "  Performance monitoring interrupts\n"),
-    ISE(MCE,        "MCE", "  Machine check exceptions\n"),
-    ISE(NMI_SRESET,        "NMI", "  System Reset interrupts\n"),
+    ISC(MCE,        "MCE", "  Machine check exceptions\n"),
+    ISC(NMI_SRESET,        "NMI", "  System Reset interrupts\n"),
  #ifdef CONFIG_PPC_WATCHDOG
      ISE(WATCHDOG,        "WDG", "  Watchdog soft-NMI interrupts\n"),
  #endif

@@ -107,11 +111,25 @@ static struct irq_stat_info

irq_stat_info[IRQ_COUNT_MAX] __ro_after_init = {
  #endif
  };
+/*
+ * Used for default disabled counters to increment the stats and to 
enable the
+ * entry for /proc/interrupts output.
+ */
+static DECLARE_BITMAP(irq_stat_count_show, IRQ_COUNT_MAX) __read_mostly;
+void inc_irq_stat_and_enable(enum irq_stat_counts which)
+{
+    __this_cpu_inc(irq_stat.counts[which]);
+    set_bit(which, irq_stat_count_show);
+}
+
  int arch_show_interrupts(struct seq_file *p, int prec)
  {
      const struct irq_stat_info *info = irq_stat_info;
      for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++, 
info++) {
+        if (!test_bit(i, irq_stat_count_show))
+            continue;
+
          seq_printf(p, "%*s:", prec, info->symbol);
          irq_proc_emit_counts(p, &irq_stat.counts[i]);
          seq_puts(p, info->text);

@@ -138,6 +156,19 @@ int arch_show_interrupts(struct seq_file *p, int

prec)
      return 0;
  }
+static int __init irq_init_stats(void)
+{
+    struct irq_stat_info *info = irq_stat_info;
+
+    for (unsigned int i = 0; i < ARRAY_SIZE(irq_stat_info); i++, 
info++) {
+        if (info->skip == 0)
+            set_bit(i, irq_stat_count_show);
+    }
+
+    return 0;
+}
+late_initcall(irq_init_stats);
+
  /*
   * /proc/stat helpers
   */

diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index a8f15154bd9a..3eacbd20fc80 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c

@@ -459,7 +459,7 @@ DEFINE_INTERRUPT_HANDLER_NMI(system_reset_exception)

      }
      hv_nmi_check_nonrecoverable(regs);
-    inc_irq_stat(NMI_SRESET);
+    inc_irq_stat_and_enable(IRQ_COUNT_NMI_SRESET);
      /* See if any machine dependent calls */
      if (ppc_md.system_reset_exception) {

@@ -816,7 +816,7 @@ static void __machine_check_exception(struct

pt_regs *regs)
  {
      int recover = 0;
-    inc_irq_stat(MCE);
+    inc_irq_stat_and_enable(IRQ_COUNT_MCE);
      add_taint(TAINT_MACHINE_CHECK, LOCKDEP_NOW_UNRELIABLE);

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help