[PATCH v9 1/4] nmi_backtrace: add more trigger_*_cpu_backtrace() methods

[PATCH v9 1/4] nmi_backtrace: add more trigger_*_cpu_backtrace() methods

[Chris Metcalf]

Currently you can only request a backtrace of either all cpus, or
all cpus but yourself.  It can also be helpful to request a remote
backtrace of a single cpu, and since we want that, the logical
extension is to support a cpumask as the underlying primitive.

This change modifies the existing lib/nmi_backtrace.c code to take
a cpumask as its basic primitive, and modifies the linux/nmi.h code
to use the new "cpumask" method instead.

The existing clients of nmi_backtrace (arm and x86) are converted
to using the new cpumask approach in this change.

The other users of the backtracing API (sparc64 and mips) are converted
to use the cpumask approach rather than the all/allbutself approach.
The mips code ignored the "include_self" boolean but with this change
it will now also dump a local backtrace if requested.

Signed-off-by: Chris Metcalf <cmetcalf@mellanox.com>
Reviewed-by: Aaron Tomlin <atomlin@redhat.com>
Reviewed-by: Petr Mladek <pmladek@suse.com>
Tested-by: Daniel Thompson <daniel.thompson@linaro.org> [arm]
---
 arch/arm/include/asm/irq.h      |  5 +++--
 arch/arm/kernel/smp.c           |  4 ++--
 arch/mips/include/asm/irq.h     |  5 +++--
 arch/mips/kernel/process.c      | 11 +++++++++--
 arch/sparc/include/asm/irq_64.h |  5 +++--
 arch/sparc/kernel/process_64.c  | 10 +++++-----
 arch/x86/include/asm/irq.h      |  5 +++--
 arch/x86/kernel/apic/hw_nmi.c   | 18 +++++++++---------
 include/linux/nmi.h             | 31 ++++++++++++++++++++++++++-----
 lib/nmi_backtrace.c             | 17 +++++++++--------
 10 files changed, 72 insertions(+), 39 deletions(-)

diff --git a/arch/arm/include/asm/irq.h b/arch/arm/include/asm/irq.h
index 1bd9510de1b9..e53638c8ed8a 100644
--- a/arch/arm/include/asm/irq.h
+++ b/arch/arm/include/asm/irq.h
@@ -36,8 +36,9 @@ extern void set_handle_irq(void (*handle_irq)(struct pt_regs *));
 #endif
 
 #ifdef CONFIG_SMP
-extern void arch_trigger_all_cpu_backtrace(bool);
-#define arch_trigger_all_cpu_backtrace(x) arch_trigger_all_cpu_backtrace(x)
+extern void arch_trigger_cpumask_backtrace(const cpumask_t *mask,
+					   bool exclude_self);
+#define arch_trigger_cpumask_backtrace arch_trigger_cpumask_backtrace
 #endif
 
 static inline int nr_legacy_irqs(void)
diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c
index 861521606c6d..2a367b2de0e1 100644
--- a/arch/arm/kernel/smp.c
+++ b/arch/arm/kernel/smp.c
@@ -760,7 +760,7 @@ static void raise_nmi(cpumask_t *mask)
 	smp_cross_call(mask, IPI_CPU_BACKTRACE);
 }
 
-void arch_trigger_all_cpu_backtrace(bool include_self)
+void arch_trigger_cpumask_backtrace(const cpumask_t *mask, bool exclude_self)
 {
-	nmi_trigger_all_cpu_backtrace(include_self, raise_nmi);
+	nmi_trigger_cpumask_backtrace(mask, exclude_self, raise_nmi);
 }
diff --git a/arch/mips/include/asm/irq.h b/arch/mips/include/asm/irq.h
index 15e0fecbc300..6bf10e796553 100644
--- a/arch/mips/include/asm/irq.h
+++ b/arch/mips/include/asm/irq.h
@@ -51,7 +51,8 @@ extern int cp0_fdc_irq;
 
 extern int get_c0_fdc_int(void);
 
-void arch_trigger_all_cpu_backtrace(bool);
-#define arch_trigger_all_cpu_backtrace arch_trigger_all_cpu_backtrace
+void arch_trigger_cpumask_backtrace(const struct cpumask *mask,
+				    bool exclude_self);
+#define arch_trigger_cpumask_backtrace arch_trigger_cpumask_backtrace
 
 #endif /* _ASM_IRQ_H */
diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c
index 7429ad09fbe3..fea1fa7726e3 100644
--- a/arch/mips/kernel/process.c
+++ b/arch/mips/kernel/process.c
@@ -569,9 +569,16 @@ static void arch_dump_stack(void *info)
 	dump_stack();
 }
 
-void arch_trigger_all_cpu_backtrace(bool include_self)
+void arch_trigger_cpumask_backtrace(const cpumask_t *mask, bool exclude_self)
 {
-	smp_call_function(arch_dump_stack, NULL, 1);
+	long this_cpu = get_cpu();
+
+	if (cpumask_test_cpu(this_cpu, mask) && !exclude_self)
+		dump_stack();
+
+	smp_call_function_many(mask, arch_dump_stack, NULL, 1);
+
+	put_cpu();
 }
 
 int mips_get_process_fp_mode(struct task_struct *task)
diff --git a/arch/sparc/include/asm/irq_64.h b/arch/sparc/include/asm/irq_64.h
index 3f70f900e834..1d51a11fb261 100644
--- a/arch/sparc/include/asm/irq_64.h
+++ b/arch/sparc/include/asm/irq_64.h
@@ -86,8 +86,9 @@ static inline unsigned long get_softint(void)
 	return retval;
 }
 
-void arch_trigger_all_cpu_backtrace(bool);
-#define arch_trigger_all_cpu_backtrace arch_trigger_all_cpu_backtrace
+void arch_trigger_cpumask_backtrace(const struct cpumask *mask,
+				    bool exclude_self);
+#define arch_trigger_cpumask_backtrace arch_trigger_cpumask_backtrace
 
 extern void *hardirq_stack[NR_CPUS];
 extern void *softirq_stack[NR_CPUS];
diff --git a/arch/sparc/kernel/process_64.c b/arch/sparc/kernel/process_64.c
index fa14402b33f9..47ff5588e521 100644
--- a/arch/sparc/kernel/process_64.c
+++ b/arch/sparc/kernel/process_64.c
@@ -239,7 +239,7 @@ static void __global_reg_poll(struct global_reg_snapshot *gp)
 	}
 }
 
-void arch_trigger_all_cpu_backtrace(bool include_self)
+void arch_trigger_cpumask_backtrace(const cpumask_t *mask, bool exclude_self)
 {
 	struct thread_info *tp = current_thread_info();
 	struct pt_regs *regs = get_irq_regs();
@@ -255,15 +255,15 @@ void arch_trigger_all_cpu_backtrace(bool include_self)
 
 	memset(global_cpu_snapshot, 0, sizeof(global_cpu_snapshot));
 
-	if (include_self)
+	if (cpumask_test_cpu(this_cpu, mask) && !exclude_self)
 		__global_reg_self(tp, regs, this_cpu);
 
 	smp_fetch_global_regs();
 
-	for_each_online_cpu(cpu) {
+	for_each_cpu(cpu, mask) {
 		struct global_reg_snapshot *gp;
 
-		if (!include_self && cpu == this_cpu)
+		if (exclude_self && cpu == this_cpu)
 			continue;
 
 		gp = &global_cpu_snapshot[cpu].reg;
@@ -300,7 +300,7 @@ void arch_trigger_all_cpu_backtrace(bool include_self)
 
 static void sysrq_handle_globreg(int key)
 {
-	arch_trigger_all_cpu_backtrace(true);
+	trigger_all_cpu_backtrace();
 }
 
 static struct sysrq_key_op sparc_globalreg_op = {
diff --git a/arch/x86/include/asm/irq.h b/arch/x86/include/asm/irq.h
index e7de5c9a4fbd..16d3fa211962 100644
--- a/arch/x86/include/asm/irq.h
+++ b/arch/x86/include/asm/irq.h
@@ -50,8 +50,9 @@ extern int vector_used_by_percpu_irq(unsigned int vector);
 extern void init_ISA_irqs(void);
 
 #ifdef CONFIG_X86_LOCAL_APIC
-void arch_trigger_all_cpu_backtrace(bool);
-#define arch_trigger_all_cpu_backtrace arch_trigger_all_cpu_backtrace
+void arch_trigger_cpumask_backtrace(const struct cpumask *mask,
+				    bool exclude_self);
+#define arch_trigger_cpumask_backtrace arch_trigger_cpumask_backtrace
 #endif
 
 #endif /* _ASM_X86_IRQ_H */
diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c
index f29501e1a5c1..c73c9fb281e1 100644
--- a/arch/x86/kernel/apic/hw_nmi.c
+++ b/arch/x86/kernel/apic/hw_nmi.c
@@ -26,32 +26,32 @@ u64 hw_nmi_get_sample_period(int watchdog_thresh)
 }
 #endif
 
-#ifdef arch_trigger_all_cpu_backtrace
+#ifdef arch_trigger_cpumask_backtrace
 static void nmi_raise_cpu_backtrace(cpumask_t *mask)
 {
 	apic->send_IPI_mask(mask, NMI_VECTOR);
 }
 
-void arch_trigger_all_cpu_backtrace(bool include_self)
+void arch_trigger_cpumask_backtrace(const cpumask_t *mask, bool exclude_self)
 {
-	nmi_trigger_all_cpu_backtrace(include_self, nmi_raise_cpu_backtrace);
+	nmi_trigger_cpumask_backtrace(mask, exclude_self,
+				      nmi_raise_cpu_backtrace);
 }
 
-static int
-arch_trigger_all_cpu_backtrace_handler(unsigned int cmd, struct pt_regs *regs)
+static int nmi_cpu_backtrace_handler(unsigned int cmd, struct pt_regs *regs)
 {
 	if (nmi_cpu_backtrace(regs))
 		return NMI_HANDLED;
 
 	return NMI_DONE;
 }
-NOKPROBE_SYMBOL(arch_trigger_all_cpu_backtrace_handler);
+NOKPROBE_SYMBOL(nmi_cpu_backtrace_handler);
 
-static int __init register_trigger_all_cpu_backtrace(void)
+static int __init register_nmi_cpu_backtrace_handler(void)
 {
-	register_nmi_handler(NMI_LOCAL, arch_trigger_all_cpu_backtrace_handler,
+	register_nmi_handler(NMI_LOCAL, nmi_cpu_backtrace_handler,
 				0, "arch_bt");
 	return 0;
 }
-early_initcall(register_trigger_all_cpu_backtrace);
+early_initcall(register_nmi_cpu_backtrace_handler);
 #endif
diff --git a/include/linux/nmi.h b/include/linux/nmi.h
index 4630eeae18e0..a78c35cff1ae 100644
--- a/include/linux/nmi.h
+++ b/include/linux/nmi.h
@@ -35,21 +35,34 @@ static inline void hardlockup_detector_disable(void) {}
  * base function. Return whether such support was available,
  * to allow calling code to fall back to some other mechanism:
  */
-#ifdef arch_trigger_all_cpu_backtrace
+#ifdef arch_trigger_cpumask_backtrace
 static inline bool trigger_all_cpu_backtrace(void)
 {
-	arch_trigger_all_cpu_backtrace(true);
-
+	arch_trigger_cpumask_backtrace(cpu_online_mask, false);
 	return true;
 }
+
 static inline bool trigger_allbutself_cpu_backtrace(void)
 {
-	arch_trigger_all_cpu_backtrace(false);
+	arch_trigger_cpumask_backtrace(cpu_online_mask, true);
+	return true;
+}
+
+static inline bool trigger_cpumask_backtrace(struct cpumask *mask)
+{
+	arch_trigger_cpumask_backtrace(mask, false);
+	return true;
+}
+
+static inline bool trigger_single_cpu_backtrace(int cpu)
+{
+	arch_trigger_cpumask_backtrace(cpumask_of(cpu), false);
 	return true;
 }
 
 /* generic implementation */
-void nmi_trigger_all_cpu_backtrace(bool include_self,
+void nmi_trigger_cpumask_backtrace(const cpumask_t *mask,
+				   bool exclude_self,
 				   void (*raise)(cpumask_t *mask));
 bool nmi_cpu_backtrace(struct pt_regs *regs);
 
@@ -62,6 +75,14 @@ static inline bool trigger_allbutself_cpu_backtrace(void)
 {
 	return false;
 }
+static inline bool trigger_cpumask_backtrace(struct cpumask *mask)
+{
+	return false;
+}
+static inline bool trigger_single_cpu_backtrace(int cpu)
+{
+	return false;
+}
 #endif
 
 #ifdef CONFIG_LOCKUP_DETECTOR
diff --git a/lib/nmi_backtrace.c b/lib/nmi_backtrace.c
index 26caf51cc238..df347e355267 100644
--- a/lib/nmi_backtrace.c
+++ b/lib/nmi_backtrace.c
@@ -17,20 +17,21 @@
 #include <linux/kprobes.h>
 #include <linux/nmi.h>
 
-#ifdef arch_trigger_all_cpu_backtrace
+#ifdef arch_trigger_cpumask_backtrace
 /* For reliability, we're prepared to waste bits here. */
 static DECLARE_BITMAP(backtrace_mask, NR_CPUS) __read_mostly;
 
-/* "in progress" flag of arch_trigger_all_cpu_backtrace */
+/* "in progress" flag of arch_trigger_cpumask_backtrace */
 static unsigned long backtrace_flag;
 
 /*
- * When raise() is called it will be is passed a pointer to the
+ * When raise() is called it will be passed a pointer to the
  * backtrace_mask. Architectures that call nmi_cpu_backtrace()
  * directly from their raise() functions may rely on the mask
  * they are passed being updated as a side effect of this call.
  */
-void nmi_trigger_all_cpu_backtrace(bool include_self,
+void nmi_trigger_cpumask_backtrace(const cpumask_t *mask,
+				   bool exclude_self,
 				   void (*raise)(cpumask_t *mask))
 {
 	int i, this_cpu = get_cpu();
@@ -44,13 +45,13 @@ void nmi_trigger_all_cpu_backtrace(bool include_self,
 		return;
 	}
 
-	cpumask_copy(to_cpumask(backtrace_mask), cpu_online_mask);
-	if (!include_self)
+	cpumask_copy(to_cpumask(backtrace_mask), mask);
+	if (exclude_self)
 		cpumask_clear_cpu(this_cpu, to_cpumask(backtrace_mask));
 
 	if (!cpumask_empty(to_cpumask(backtrace_mask))) {
-		pr_info("Sending NMI to %s CPUs:\n",
-			(include_self ? "all" : "other"));
+		pr_info("Sending NMI from CPU %d to CPUs %*pbl:\n",
+			this_cpu, nr_cpumask_bits, to_cpumask(backtrace_mask));
 		raise(to_cpumask(backtrace_mask));
 	}
 
-- 
2.7.2

[PATCH v9 2/4] nmi_backtrace: do a local dump_stack() instead of a self-NMI

[Chris Metcalf]

Currently on arm there is code that checks whether it should call
dump_stack() explicitly, to avoid trying to raise an NMI when the
current context is not preemptible by the backtrace IPI.  Similarly,
the forthcoming arch/tile support uses an IPI mechanism that does
not support generating an NMI to self.

Accordingly, move the code that guards this case into the generic
mechanism, and invoke it unconditionally whenever we want a
backtrace of the current cpu.  It seems plausible that in all cases,
dump_stack() will generate better information than generating a
stack from the NMI handler.  The register state will be missing,
but that state is likely not particularly helpful in any case.

Or, if we think it is helpful, we should be capturing and emitting
the current register state in all cases when regs == NULL is passed
to nmi_cpu_backtrace().

Signed-off-by: Chris Metcalf <cmetcalf@mellanox.com>
Acked-by: Aaron Tomlin <atomlin@redhat.com>
Reviewed-by: Petr Mladek <pmladek@suse.com>
Tested-by: Daniel Thompson <daniel.thompson@linaro.org> [arm]
---
 arch/arm/kernel/smp.c | 9 ---------
 lib/nmi_backtrace.c   | 9 +++++++++
 2 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c
index 2a367b2de0e1..173ad30c20e0 100644
--- a/arch/arm/kernel/smp.c
+++ b/arch/arm/kernel/smp.c
@@ -748,15 +748,6 @@ core_initcall(register_cpufreq_notifier);
 
 static void raise_nmi(cpumask_t *mask)
 {
-	/*
-	 * Generate the backtrace directly if we are running in a calling
-	 * context that is not preemptible by the backtrace IPI. Note
-	 * that nmi_cpu_backtrace() automatically removes the current cpu
-	 * from mask.
-	 */
-	if (cpumask_test_cpu(smp_processor_id(), mask) && irqs_disabled())
-		nmi_cpu_backtrace(NULL);
-
 	smp_cross_call(mask, IPI_CPU_BACKTRACE);
 }
 
diff --git a/lib/nmi_backtrace.c b/lib/nmi_backtrace.c
index df347e355267..393a3cca1f47 100644
--- a/lib/nmi_backtrace.c
+++ b/lib/nmi_backtrace.c
@@ -49,6 +49,15 @@ void nmi_trigger_cpumask_backtrace(const cpumask_t *mask,
 	if (exclude_self)
 		cpumask_clear_cpu(this_cpu, to_cpumask(backtrace_mask));
 
+	/*
+	 * Don't try to send an NMI to this cpu; it may work on some
+	 * architectures, but on others it may not, and we'll get
+	 * information at least as useful just by doing a dump_stack() here.
+	 * Note that nmi_cpu_backtrace(NULL) will clear the cpu bit.
+	 */
+	if (cpumask_test_cpu(this_cpu, to_cpumask(backtrace_mask)))
+		nmi_cpu_backtrace(NULL);
+
 	if (!cpumask_empty(to_cpumask(backtrace_mask))) {
 		pr_info("Sending NMI from CPU %d to CPUs %*pbl:\n",
 			this_cpu, nr_cpumask_bits, to_cpumask(backtrace_mask));
-- 
2.7.2

[PATCH v9 3/4] arch/tile: adopt the new nmi_backtrace framework

[Chris Metcalf]

Previously tile was rolling its own method of capturing backtrace
data in the NMI handlers, but it was relying on running printk()
from the NMI handler, which is not always safe.  So adopt the
nmi_backtrace model (with the new cpumask extension) instead.

So we can call the nmi_backtrace code directly from the nmi handler,
move the nmi_enter()/exit() into the top-level tile NMI handler.

The semantics of the routine change slightly since it is now
synchronous with the remote cores completing the backtraces.
Previously it was asynchronous, but with protection to avoid starting
a new remote backtrace if the old one was still in progress.

Signed-off-by: Chris Metcalf <cmetcalf@mellanox.com>
---
 arch/tile/include/asm/irq.h |  5 ++--
 arch/tile/kernel/pmc.c      |  3 --
 arch/tile/kernel/process.c  | 73 +++++++++++----------------------------------
 arch/tile/kernel/traps.c    |  9 ++++--
 4 files changed, 27 insertions(+), 63 deletions(-)

diff --git a/arch/tile/include/asm/irq.h b/arch/tile/include/asm/irq.h
index 84a924034bdb..1fa1f2544ff9 100644
--- a/arch/tile/include/asm/irq.h
+++ b/arch/tile/include/asm/irq.h
@@ -79,8 +79,9 @@ void tile_irq_activate(unsigned int irq, int tile_irq_type);
 void setup_irq_regs(void);
 
 #ifdef __tilegx__
-void arch_trigger_all_cpu_backtrace(bool self);
-#define arch_trigger_all_cpu_backtrace arch_trigger_all_cpu_backtrace
+void arch_trigger_cpumask_backtrace(const struct cpumask *mask,
+				    bool exclude_self);
+#define arch_trigger_cpumask_backtrace arch_trigger_cpumask_backtrace
 #endif
 
 #endif /* _ASM_TILE_IRQ_H */
diff --git a/arch/tile/kernel/pmc.c b/arch/tile/kernel/pmc.c
index db62cc34b955..81cf8743a3f3 100644
--- a/arch/tile/kernel/pmc.c
+++ b/arch/tile/kernel/pmc.c
@@ -16,7 +16,6 @@
 #include <linux/spinlock.h>
 #include <linux/module.h>
 #include <linux/atomic.h>
-#include <linux/interrupt.h>
 
 #include <asm/processor.h>
 #include <asm/pmc.h>
@@ -29,9 +28,7 @@ int handle_perf_interrupt(struct pt_regs *regs, int fault)
 	if (!perf_irq)
 		panic("Unexpected PERF_COUNT interrupt %d\n", fault);
 
-	nmi_enter();
 	retval = perf_irq(regs, fault);
-	nmi_exit();
 	return retval;
 }
 
diff --git a/arch/tile/kernel/process.c b/arch/tile/kernel/process.c
index a465d8372edd..9f37106ef93a 100644
--- a/arch/tile/kernel/process.c
+++ b/arch/tile/kernel/process.c
@@ -22,7 +22,7 @@
 #include <linux/init.h>
 #include <linux/mm.h>
 #include <linux/compat.h>
-#include <linux/hardirq.h>
+#include <linux/nmi.h>
 #include <linux/syscalls.h>
 #include <linux/kernel.h>
 #include <linux/tracehook.h>
@@ -594,66 +594,18 @@ void show_regs(struct pt_regs *regs)
 	tile_show_stack(&kbt);
 }
 
-/* To ensure stack dump on tiles occurs one by one. */
-static DEFINE_SPINLOCK(backtrace_lock);
-/* To ensure no backtrace occurs before all of the stack dump are done. */
-static atomic_t backtrace_cpus;
-/* The cpu mask to avoid reentrance. */
-static struct cpumask backtrace_mask;
-
-void do_nmi_dump_stack(struct pt_regs *regs)
-{
-	int is_idle = is_idle_task(current) && !in_interrupt();
-	int cpu;
-
-	nmi_enter();
-	cpu = smp_processor_id();
-	if (WARN_ON_ONCE(!cpumask_test_and_clear_cpu(cpu, &backtrace_mask)))
-		goto done;
-
-	spin_lock(&backtrace_lock);
-	if (is_idle)
-		pr_info("CPU: %d idle\n", cpu);
-	else
-		show_regs(regs);
-	spin_unlock(&backtrace_lock);
-	atomic_dec(&backtrace_cpus);
-done:
-	nmi_exit();
-}
-
 #ifdef __tilegx__
-void arch_trigger_all_cpu_backtrace(bool self)
+void nmi_raise_cpu_backtrace(struct cpumask *in_mask)
 {
 	struct cpumask mask;
 	HV_Coord tile;
 	unsigned int timeout;
 	int cpu;
-	int ongoing;
 	HV_NMI_Info info[NR_CPUS];
 
-	ongoing = atomic_cmpxchg(&backtrace_cpus, 0, num_online_cpus() - 1);
-	if (ongoing != 0) {
-		pr_err("Trying to do all-cpu backtrace.\n");
-		pr_err("But another all-cpu backtrace is ongoing (%d cpus left)\n",
-		       ongoing);
-		if (self) {
-			pr_err("Reporting the stack on this cpu only.\n");
-			dump_stack();
-		}
-		return;
-	}
-
-	cpumask_copy(&mask, cpu_online_mask);
-	cpumask_clear_cpu(smp_processor_id(), &mask);
-	cpumask_copy(&backtrace_mask, &mask);
-
-	/* Backtrace for myself first. */
-	if (self)
-		dump_stack();
-
 	/* Tentatively dump stack on remote tiles via NMI. */
 	timeout = 100;
+	cpumask_copy(&mask, in_mask);
 	while (!cpumask_empty(&mask) && timeout) {
 		for_each_cpu(cpu, &mask) {
 			tile.x = cpu_x(cpu);
@@ -664,12 +616,17 @@ void arch_trigger_all_cpu_backtrace(bool self)
 		}
 
 		mdelay(10);
+		touch_softlockup_watchdog();
 		timeout--;
 	}
 
-	/* Warn about cpus stuck in ICS and decrement their counts here. */
+	/* Warn about cpus stuck in ICS. */
 	if (!cpumask_empty(&mask)) {
 		for_each_cpu(cpu, &mask) {
+
+			/* Clear the bit as if nmi_cpu_backtrace() ran. */
+			cpumask_clear_cpu(cpu, in_mask);
+
 			switch (info[cpu].result) {
 			case HV_NMI_RESULT_FAIL_ICS:
 				pr_warn("Skipping stack dump of cpu %d in ICS at pc %#llx\n",
@@ -680,16 +637,20 @@ void arch_trigger_all_cpu_backtrace(bool self)
 					cpu);
 				break;
 			case HV_ENOSYS:
-				pr_warn("Hypervisor too old to allow remote stack dumps.\n");
-				goto skip_for_each;
+				WARN_ONCE(1, "Hypervisor too old to allow remote stack dumps.\n");
+				break;
 			default:  /* should not happen */
 				pr_warn("Skipping stack dump of cpu %d [%d,%#llx]\n",
 					cpu, info[cpu].result, info[cpu].pc);
 				break;
 			}
 		}
-skip_for_each:
-		atomic_sub(cpumask_weight(&mask), &backtrace_cpus);
 	}
 }
+
+void arch_trigger_cpumask_backtrace(const cpumask_t *mask, bool exclude_self)
+{
+	nmi_trigger_cpumask_backtrace(mask, exclude_self,
+				      nmi_raise_cpu_backtrace);
+}
 #endif /* __tilegx_ */
diff --git a/arch/tile/kernel/traps.c b/arch/tile/kernel/traps.c
index 4d9651c5b1ad..39f427bb0de2 100644
--- a/arch/tile/kernel/traps.c
+++ b/arch/tile/kernel/traps.c
@@ -20,6 +20,8 @@
 #include <linux/reboot.h>
 #include <linux/uaccess.h>
 #include <linux/ptrace.h>
+#include <linux/hardirq.h>
+#include <linux/nmi.h>
 #include <asm/stack.h>
 #include <asm/traps.h>
 #include <asm/setup.h>
@@ -392,14 +394,17 @@ void __kprobes do_trap(struct pt_regs *regs, int fault_num,
 
 void do_nmi(struct pt_regs *regs, int fault_num, unsigned long reason)
 {
+	nmi_enter();
 	switch (reason) {
+#ifdef arch_trigger_cpumask_backtrace
 	case TILE_NMI_DUMP_STACK:
-		do_nmi_dump_stack(regs);
+		nmi_cpu_backtrace(regs);
 		break;
+#endif
 	default:
 		panic("Unexpected do_nmi type %ld", reason);
-		return;
 	}
+	nmi_exit();
 }
 
 /* Deprecated function currently only used here. */
-- 
2.7.2

[PATCH v9 4/4] nmi_backtrace: generate one-line reports for idle cpus

[Chris Metcalf]

When doing an nmi backtrace of many cores, most of which are idle,
the output is a little overwhelming and very uninformative.  Suppress
messages for cpus that are idling when they are interrupted and just
emit one line, "NMI backtrace for N skipped: idling at pc 0xNNN".

We do this by grouping all the cpuidle code together into a new
.cpuidle.text section, and then checking the address of the
interrupted PC to see if it lies within that section.

This commit suitably tags x86 and tile idle routines, and only
adds in the minimal framework for other architectures.

Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Daniel Thompson <daniel.thompson@linaro.org> [arm]
Signed-off-by: Chris Metcalf <cmetcalf@mellanox.com>
---
 arch/alpha/kernel/vmlinux.lds.S      |  1 +
 arch/arc/kernel/vmlinux.lds.S        |  1 +
 arch/arm/kernel/vmlinux-xip.lds.S    |  1 +
 arch/arm/kernel/vmlinux.lds.S        |  1 +
 arch/arm64/kernel/vmlinux.lds.S      |  1 +
 arch/avr32/kernel/vmlinux.lds.S      |  1 +
 arch/blackfin/kernel/vmlinux.lds.S   |  1 +
 arch/c6x/kernel/vmlinux.lds.S        |  1 +
 arch/cris/kernel/vmlinux.lds.S       |  1 +
 arch/frv/kernel/vmlinux.lds.S        |  1 +
 arch/h8300/kernel/vmlinux.lds.S      |  1 +
 arch/hexagon/kernel/vmlinux.lds.S    |  1 +
 arch/ia64/kernel/vmlinux.lds.S       |  1 +
 arch/m32r/kernel/vmlinux.lds.S       |  1 +
 arch/m68k/kernel/vmlinux-nommu.lds   |  1 +
 arch/m68k/kernel/vmlinux-std.lds     |  1 +
 arch/m68k/kernel/vmlinux-sun3.lds    |  1 +
 arch/metag/kernel/vmlinux.lds.S      |  1 +
 arch/microblaze/kernel/vmlinux.lds.S |  1 +
 arch/mips/kernel/vmlinux.lds.S       |  1 +
 arch/mn10300/kernel/vmlinux.lds.S    |  1 +
 arch/nios2/kernel/vmlinux.lds.S      |  1 +
 arch/openrisc/kernel/vmlinux.lds.S   |  1 +
 arch/parisc/kernel/vmlinux.lds.S     |  1 +
 arch/powerpc/kernel/vmlinux.lds.S    |  1 +
 arch/s390/kernel/vmlinux.lds.S       |  1 +
 arch/score/kernel/vmlinux.lds.S      |  1 +
 arch/sh/kernel/vmlinux.lds.S         |  1 +
 arch/sparc/kernel/vmlinux.lds.S      |  1 +
 arch/tile/kernel/entry.S             |  2 +-
 arch/tile/kernel/vmlinux.lds.S       |  1 +
 arch/um/kernel/dyn.lds.S             |  1 +
 arch/um/kernel/uml.lds.S             |  1 +
 arch/unicore32/kernel/vmlinux.lds.S  |  1 +
 arch/x86/include/asm/irqflags.h      | 12 ++++++++----
 arch/x86/kernel/acpi/cstate.c        |  2 +-
 arch/x86/kernel/process.c            |  4 ++--
 arch/x86/kernel/vmlinux.lds.S        |  1 +
 arch/xtensa/kernel/vmlinux.lds.S     |  3 +++
 drivers/acpi/processor_idle.c        |  5 +++--
 drivers/cpuidle/driver.c             |  5 +++--
 drivers/idle/intel_idle.c            |  4 ++--
 include/asm-generic/vmlinux.lds.h    |  6 ++++++
 include/linux/cpu.h                  |  5 +++++
 kernel/sched/idle.c                  | 13 +++++++++++--
 lib/nmi_backtrace.c                  | 16 +++++++++++-----
 scripts/mod/modpost.c                |  2 +-
 scripts/recordmcount.c               |  1 +
 scripts/recordmcount.pl              |  1 +
 49 files changed, 93 insertions(+), 22 deletions(-)

diff --git a/arch/alpha/kernel/vmlinux.lds.S b/arch/alpha/kernel/vmlinux.lds.S
index 647b84c15382..cebecfb76fbf 100644
--- a/arch/alpha/kernel/vmlinux.lds.S
+++ b/arch/alpha/kernel/vmlinux.lds.S
@@ -22,6 +22,7 @@ SECTIONS
 		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		*(.fixup)
 		*(.gnu.warning)
diff --git a/arch/arc/kernel/vmlinux.lds.S b/arch/arc/kernel/vmlinux.lds.S
index 894e696bddaa..65652160cfda 100644
--- a/arch/arc/kernel/vmlinux.lds.S
+++ b/arch/arc/kernel/vmlinux.lds.S
@@ -97,6 +97,7 @@ SECTIONS
 		_text = .;
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		*(.fixup)
diff --git a/arch/arm/kernel/vmlinux-xip.lds.S b/arch/arm/kernel/vmlinux-xip.lds.S
index cba1ec899a69..7fa487ef7e2f 100644
--- a/arch/arm/kernel/vmlinux-xip.lds.S
+++ b/arch/arm/kernel/vmlinux-xip.lds.S
@@ -98,6 +98,7 @@ SECTIONS
 			IRQENTRY_TEXT
 			TEXT_TEXT
 			SCHED_TEXT
+			CPUIDLE_TEXT
 			LOCK_TEXT
 			KPROBES_TEXT
 			*(.gnu.warning)
diff --git a/arch/arm/kernel/vmlinux.lds.S b/arch/arm/kernel/vmlinux.lds.S
index d24e5dd2aa7a..f7f55df0bf7b 100644
--- a/arch/arm/kernel/vmlinux.lds.S
+++ b/arch/arm/kernel/vmlinux.lds.S
@@ -111,6 +111,7 @@ SECTIONS
 			SOFTIRQENTRY_TEXT
 			TEXT_TEXT
 			SCHED_TEXT
+			CPUIDLE_TEXT
 			LOCK_TEXT
 			HYPERVISOR_TEXT
 			KPROBES_TEXT
diff --git a/arch/arm64/kernel/vmlinux.lds.S b/arch/arm64/kernel/vmlinux.lds.S
index 659963d40bb4..fe7f93b7b11b 100644
--- a/arch/arm64/kernel/vmlinux.lds.S
+++ b/arch/arm64/kernel/vmlinux.lds.S
@@ -122,6 +122,7 @@ SECTIONS
 			ENTRY_TEXT
 			TEXT_TEXT
 			SCHED_TEXT
+			CPUIDLE_TEXT
 			LOCK_TEXT
 			KPROBES_TEXT
 			HYPERVISOR_TEXT
diff --git a/arch/avr32/kernel/vmlinux.lds.S b/arch/avr32/kernel/vmlinux.lds.S
index a4589176bed5..17f2730eb497 100644
--- a/arch/avr32/kernel/vmlinux.lds.S
+++ b/arch/avr32/kernel/vmlinux.lds.S
@@ -52,6 +52,7 @@ SECTIONS
 		KPROBES_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		*(.fixup)
 		*(.gnu.warning)
diff --git a/arch/blackfin/kernel/vmlinux.lds.S b/arch/blackfin/kernel/vmlinux.lds.S
index d920b959ff3a..68069a120055 100644
--- a/arch/blackfin/kernel/vmlinux.lds.S
+++ b/arch/blackfin/kernel/vmlinux.lds.S
@@ -33,6 +33,7 @@ SECTIONS
 #ifndef CONFIG_SCHEDULE_L1
 		SCHED_TEXT
 #endif
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		IRQENTRY_TEXT
 		SOFTIRQENTRY_TEXT
diff --git a/arch/c6x/kernel/vmlinux.lds.S b/arch/c6x/kernel/vmlinux.lds.S
index 50bc10f97bcb..a1a5c166bc9b 100644
--- a/arch/c6x/kernel/vmlinux.lds.S
+++ b/arch/c6x/kernel/vmlinux.lds.S
@@ -70,6 +70,7 @@ SECTIONS
 		_stext = .;
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		IRQENTRY_TEXT
 		SOFTIRQENTRY_TEXT
diff --git a/arch/cris/kernel/vmlinux.lds.S b/arch/cris/kernel/vmlinux.lds.S
index 7552c2557506..979586261520 100644
--- a/arch/cris/kernel/vmlinux.lds.S
+++ b/arch/cris/kernel/vmlinux.lds.S
@@ -43,6 +43,7 @@ SECTIONS
 		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		*(.fixup)
 		*(.text.__*)
diff --git a/arch/frv/kernel/vmlinux.lds.S b/arch/frv/kernel/vmlinux.lds.S
index 7e958d829ec9..aa6e573d57da 100644
--- a/arch/frv/kernel/vmlinux.lds.S
+++ b/arch/frv/kernel/vmlinux.lds.S
@@ -63,6 +63,7 @@ SECTIONS
 	*(.text..tlbmiss)
 	TEXT_TEXT
 	SCHED_TEXT
+	CPUIDLE_TEXT
 	LOCK_TEXT
 #ifdef CONFIG_DEBUG_INFO
 	INIT_TEXT
diff --git a/arch/h8300/kernel/vmlinux.lds.S b/arch/h8300/kernel/vmlinux.lds.S
index cb5dfb02c88d..7f11da1b895e 100644
--- a/arch/h8300/kernel/vmlinux.lds.S
+++ b/arch/h8300/kernel/vmlinux.lds.S
@@ -29,6 +29,7 @@ SECTIONS
 	_stext = . ;
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 #if defined(CONFIG_ROMKERNEL)
 		*(.int_redirect)
diff --git a/arch/hexagon/kernel/vmlinux.lds.S b/arch/hexagon/kernel/vmlinux.lds.S
index 5f268c1071b3..ec87e67feb19 100644
--- a/arch/hexagon/kernel/vmlinux.lds.S
+++ b/arch/hexagon/kernel/vmlinux.lds.S
@@ -50,6 +50,7 @@ SECTIONS
 		_text = .;
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		*(.fixup)
diff --git a/arch/ia64/kernel/vmlinux.lds.S b/arch/ia64/kernel/vmlinux.lds.S
index dc506b05ffbd..f89d20c97412 100644
--- a/arch/ia64/kernel/vmlinux.lds.S
+++ b/arch/ia64/kernel/vmlinux.lds.S
@@ -46,6 +46,7 @@ SECTIONS {
 		__end_ivt_text = .;
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		*(.gnu.linkonce.t*)
diff --git a/arch/m32r/kernel/vmlinux.lds.S b/arch/m32r/kernel/vmlinux.lds.S
index 018e4a711d79..ad1fe56455aa 100644
--- a/arch/m32r/kernel/vmlinux.lds.S
+++ b/arch/m32r/kernel/vmlinux.lds.S
@@ -31,6 +31,7 @@ SECTIONS
 	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
+	CPUIDLE_TEXT
 	LOCK_TEXT
 	*(.fixup)
 	*(.gnu.warning)
diff --git a/arch/m68k/kernel/vmlinux-nommu.lds b/arch/m68k/kernel/vmlinux-nommu.lds
index 06a763f49fd3..d2c8abf1c8c4 100644
--- a/arch/m68k/kernel/vmlinux-nommu.lds
+++ b/arch/m68k/kernel/vmlinux-nommu.lds
@@ -45,6 +45,7 @@ SECTIONS {
 		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		*(.fixup)
 		. = ALIGN(16);
diff --git a/arch/m68k/kernel/vmlinux-std.lds b/arch/m68k/kernel/vmlinux-std.lds
index d0993594f558..5b5ce1e4d1ed 100644
--- a/arch/m68k/kernel/vmlinux-std.lds
+++ b/arch/m68k/kernel/vmlinux-std.lds
@@ -16,6 +16,7 @@ SECTIONS
 	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
+	CPUIDLE_TEXT
 	LOCK_TEXT
 	*(.fixup)
 	*(.gnu.warning)
diff --git a/arch/m68k/kernel/vmlinux-sun3.lds b/arch/m68k/kernel/vmlinux-sun3.lds
index 8080469ee6c1..fe5ea1974b16 100644
--- a/arch/m68k/kernel/vmlinux-sun3.lds
+++ b/arch/m68k/kernel/vmlinux-sun3.lds
@@ -16,6 +16,7 @@ SECTIONS
 	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
+	CPUIDLE_TEXT
 	LOCK_TEXT
 	*(.fixup)
 	*(.gnu.warning)
diff --git a/arch/metag/kernel/vmlinux.lds.S b/arch/metag/kernel/vmlinux.lds.S
index 150ace92c7ad..e6c700eaf207 100644
--- a/arch/metag/kernel/vmlinux.lds.S
+++ b/arch/metag/kernel/vmlinux.lds.S
@@ -21,6 +21,7 @@ SECTIONS
   .text : {
 	TEXT_TEXT
 	SCHED_TEXT
+	CPUIDLE_TEXT
 	LOCK_TEXT
 	KPROBES_TEXT
 	IRQENTRY_TEXT
diff --git a/arch/microblaze/kernel/vmlinux.lds.S b/arch/microblaze/kernel/vmlinux.lds.S
index 0a47f0410554..289d0e7f3e3a 100644
--- a/arch/microblaze/kernel/vmlinux.lds.S
+++ b/arch/microblaze/kernel/vmlinux.lds.S
@@ -33,6 +33,7 @@ SECTIONS {
 		EXIT_TEXT
 		EXIT_CALL
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		IRQENTRY_TEXT
diff --git a/arch/mips/kernel/vmlinux.lds.S b/arch/mips/kernel/vmlinux.lds.S
index a82c178d0bb9..d5de67591735 100644
--- a/arch/mips/kernel/vmlinux.lds.S
+++ b/arch/mips/kernel/vmlinux.lds.S
@@ -55,6 +55,7 @@ SECTIONS
 	.text : {
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		IRQENTRY_TEXT
diff --git a/arch/mn10300/kernel/vmlinux.lds.S b/arch/mn10300/kernel/vmlinux.lds.S
index 13c4814c29f8..2d5f1c3f1afb 100644
--- a/arch/mn10300/kernel/vmlinux.lds.S
+++ b/arch/mn10300/kernel/vmlinux.lds.S
@@ -30,6 +30,7 @@ SECTIONS
 	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
+	CPUIDLE_TEXT
 	LOCK_TEXT
 	KPROBES_TEXT
 	*(.fixup)
diff --git a/arch/nios2/kernel/vmlinux.lds.S b/arch/nios2/kernel/vmlinux.lds.S
index e23e89539967..6a8045bb1a77 100644
--- a/arch/nios2/kernel/vmlinux.lds.S
+++ b/arch/nios2/kernel/vmlinux.lds.S
@@ -37,6 +37,7 @@ SECTIONS
 	.text : {
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		IRQENTRY_TEXT
 		SOFTIRQENTRY_TEXT
diff --git a/arch/openrisc/kernel/vmlinux.lds.S b/arch/openrisc/kernel/vmlinux.lds.S
index d936de4c07ca..d68b9ede8423 100644
--- a/arch/openrisc/kernel/vmlinux.lds.S
+++ b/arch/openrisc/kernel/vmlinux.lds.S
@@ -47,6 +47,7 @@ SECTIONS
           _stext = .;
 	  TEXT_TEXT
 	  SCHED_TEXT
+	  CPUIDLE_TEXT
 	  LOCK_TEXT
 	  KPROBES_TEXT
 	  IRQENTRY_TEXT
diff --git a/arch/parisc/kernel/vmlinux.lds.S b/arch/parisc/kernel/vmlinux.lds.S
index f3ead0b6ce46..9ec8ec075dae 100644
--- a/arch/parisc/kernel/vmlinux.lds.S
+++ b/arch/parisc/kernel/vmlinux.lds.S
@@ -69,6 +69,7 @@ SECTIONS
 	.text ALIGN(PAGE_SIZE) : {
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		IRQENTRY_TEXT
diff --git a/arch/powerpc/kernel/vmlinux.lds.S b/arch/powerpc/kernel/vmlinux.lds.S
index b5fba689fca6..7ed59f0d947f 100644
--- a/arch/powerpc/kernel/vmlinux.lds.S
+++ b/arch/powerpc/kernel/vmlinux.lds.S
@@ -52,6 +52,7 @@ SECTIONS
 		/* careful! __ftr_alt_* sections need to be close to .text */
 		*(.text .fixup __ftr_alt_* .ref.text)
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		IRQENTRY_TEXT
diff --git a/arch/s390/kernel/vmlinux.lds.S b/arch/s390/kernel/vmlinux.lds.S
index 429bfd111961..000e6e91f6a0 100644
--- a/arch/s390/kernel/vmlinux.lds.S
+++ b/arch/s390/kernel/vmlinux.lds.S
@@ -35,6 +35,7 @@ SECTIONS
 		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		IRQENTRY_TEXT
diff --git a/arch/score/kernel/vmlinux.lds.S b/arch/score/kernel/vmlinux.lds.S
index 7274b5c4287e..4117890b1db1 100644
--- a/arch/score/kernel/vmlinux.lds.S
+++ b/arch/score/kernel/vmlinux.lds.S
@@ -40,6 +40,7 @@ SECTIONS
 		_text = .;	/* Text and read-only data */
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		*(.text.*)
diff --git a/arch/sh/kernel/vmlinux.lds.S b/arch/sh/kernel/vmlinux.lds.S
index 235a4101999f..5b9a3cc90c58 100644
--- a/arch/sh/kernel/vmlinux.lds.S
+++ b/arch/sh/kernel/vmlinux.lds.S
@@ -36,6 +36,7 @@ SECTIONS
 		TEXT_TEXT
 		EXTRA_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		IRQENTRY_TEXT
diff --git a/arch/sparc/kernel/vmlinux.lds.S b/arch/sparc/kernel/vmlinux.lds.S
index d79b3b734245..572db686f845 100644
--- a/arch/sparc/kernel/vmlinux.lds.S
+++ b/arch/sparc/kernel/vmlinux.lds.S
@@ -49,6 +49,7 @@ SECTIONS
 		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		IRQENTRY_TEXT
diff --git a/arch/tile/kernel/entry.S b/arch/tile/kernel/entry.S
index 670a3569450f..101de132e363 100644
--- a/arch/tile/kernel/entry.S
+++ b/arch/tile/kernel/entry.S
@@ -50,7 +50,7 @@ STD_ENTRY(smp_nap)
  * When interrupted at _cpu_idle_nap, we bump the PC forward 8, and
  * as a result return to the function that called _cpu_idle().
  */
-STD_ENTRY(_cpu_idle)
+STD_ENTRY_SECTION(_cpu_idle, .cpuidle.text)
 	movei r1, 1
 	IRQ_ENABLE_LOAD(r2, r3)
 	mtspr INTERRUPT_CRITICAL_SECTION, r1
diff --git a/arch/tile/kernel/vmlinux.lds.S b/arch/tile/kernel/vmlinux.lds.S
index 9d449caf8910..e1baf094fba4 100644
--- a/arch/tile/kernel/vmlinux.lds.S
+++ b/arch/tile/kernel/vmlinux.lds.S
@@ -42,6 +42,7 @@ SECTIONS
   .text : AT (ADDR(.text) - LOAD_OFFSET) {
     HEAD_TEXT
     SCHED_TEXT
+    CPUIDLE_TEXT
     LOCK_TEXT
     KPROBES_TEXT
     IRQENTRY_TEXT
diff --git a/arch/um/kernel/dyn.lds.S b/arch/um/kernel/dyn.lds.S
index adde088aeeff..4fdbcf958cd5 100644
--- a/arch/um/kernel/dyn.lds.S
+++ b/arch/um/kernel/dyn.lds.S
@@ -68,6 +68,7 @@ SECTIONS
     _stext = .;
     TEXT_TEXT
     SCHED_TEXT
+    CPUIDLE_TEXT
     LOCK_TEXT
     *(.fixup)
     *(.stub .text.* .gnu.linkonce.t.*)
diff --git a/arch/um/kernel/uml.lds.S b/arch/um/kernel/uml.lds.S
index 6899195602b7..1840f55ed042 100644
--- a/arch/um/kernel/uml.lds.S
+++ b/arch/um/kernel/uml.lds.S
@@ -28,6 +28,7 @@ SECTIONS
     _stext = .;
     TEXT_TEXT
     SCHED_TEXT
+    CPUIDLE_TEXT
     LOCK_TEXT
     *(.fixup)
     /* .gnu.warning sections are handled specially by elf32.em.  */
diff --git a/arch/unicore32/kernel/vmlinux.lds.S b/arch/unicore32/kernel/vmlinux.lds.S
index 77e407e49a63..56e788e8ee83 100644
--- a/arch/unicore32/kernel/vmlinux.lds.S
+++ b/arch/unicore32/kernel/vmlinux.lds.S
@@ -37,6 +37,7 @@ SECTIONS
 	.text : {		/* Real text segment */
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 
 		*(.fixup)
diff --git a/arch/x86/include/asm/irqflags.h b/arch/x86/include/asm/irqflags.h
index b77f5edb03b0..ac7692dcfa2e 100644
--- a/arch/x86/include/asm/irqflags.h
+++ b/arch/x86/include/asm/irqflags.h
@@ -4,6 +4,10 @@
 #include <asm/processor-flags.h>
 
 #ifndef __ASSEMBLY__
+
+/* Provide __cpuidle; we can't safely include <linux/cpu.h> */
+#define __cpuidle __attribute__((__section__(".cpuidle.text")))
+
 /*
  * Interrupt control:
  */
@@ -44,12 +48,12 @@ static inline void native_irq_enable(void)
 	asm volatile("sti": : :"memory");
 }
 
-static inline void native_safe_halt(void)
+static inline __cpuidle void native_safe_halt(void)
 {
 	asm volatile("sti; hlt": : :"memory");
 }
 
-static inline void native_halt(void)
+static inline __cpuidle void native_halt(void)
 {
 	asm volatile("hlt": : :"memory");
 }
@@ -86,7 +90,7 @@ static inline notrace void arch_local_irq_enable(void)
  * Used in the idle loop; sti takes one instruction cycle
  * to complete:
  */
-static inline void arch_safe_halt(void)
+static inline __cpuidle void arch_safe_halt(void)
 {
 	native_safe_halt();
 }
@@ -95,7 +99,7 @@ static inline void arch_safe_halt(void)
  * Used when interrupts are already enabled or to
  * shutdown the processor:
  */
-static inline void halt(void)
+static inline __cpuidle void halt(void)
 {
 	native_halt();
 }
diff --git a/arch/x86/kernel/acpi/cstate.c b/arch/x86/kernel/acpi/cstate.c
index bdfad642123f..af15f4444330 100644
--- a/arch/x86/kernel/acpi/cstate.c
+++ b/arch/x86/kernel/acpi/cstate.c
@@ -152,7 +152,7 @@ int acpi_processor_ffh_cstate_probe(unsigned int cpu,
 }
 EXPORT_SYMBOL_GPL(acpi_processor_ffh_cstate_probe);
 
-void acpi_processor_ffh_cstate_enter(struct acpi_processor_cx *cx)
+void __cpuidle acpi_processor_ffh_cstate_enter(struct acpi_processor_cx *cx)
 {
 	unsigned int cpu = smp_processor_id();
 	struct cstate_entry *percpu_entry;
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index 62c0b0ea2ce4..c400e30831dc 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -301,7 +301,7 @@ void arch_cpu_idle(void)
 /*
  * We use this if we don't have any better idle routine..
  */
-void default_idle(void)
+void __cpuidle default_idle(void)
 {
 	trace_cpu_idle_rcuidle(1, smp_processor_id());
 	safe_halt();
@@ -416,7 +416,7 @@ static int prefer_mwait_c1_over_halt(const struct cpuinfo_x86 *c)
  * with interrupts enabled and no flags, which is backwards compatible with the
  * original MWAIT implementation.
  */
-static void mwait_idle(void)
+static __cpuidle void mwait_idle(void)
 {
 	if (!current_set_polling_and_test()) {
 		trace_cpu_idle_rcuidle(1, smp_processor_id());
diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S
index 9297a002d8e5..dbf67f64d5ec 100644
--- a/arch/x86/kernel/vmlinux.lds.S
+++ b/arch/x86/kernel/vmlinux.lds.S
@@ -97,6 +97,7 @@ SECTIONS
 		_stext = .;
 		TEXT_TEXT
 		SCHED_TEXT
+		CPUIDLE_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
 		ENTRY_TEXT
diff --git a/arch/xtensa/kernel/vmlinux.lds.S b/arch/xtensa/kernel/vmlinux.lds.S
index c417cbe4ec87..18a174c7fb87 100644
--- a/arch/xtensa/kernel/vmlinux.lds.S
+++ b/arch/xtensa/kernel/vmlinux.lds.S
@@ -93,6 +93,9 @@ SECTIONS
     VMLINUX_SYMBOL(__sched_text_start) = .;
     *(.sched.literal .sched.text)
     VMLINUX_SYMBOL(__sched_text_end) = .;
+    VMLINUX_SYMBOL(__cpuidle_text_start) = .;
+    *(.cpuidle.literal .cpuidle.text)
+    VMLINUX_SYMBOL(__cpuidle_text_end) = .;
     VMLINUX_SYMBOL(__lock_text_start) = .;
     *(.spinlock.literal .spinlock.text)
     VMLINUX_SYMBOL(__lock_text_end) = .;
diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
index cea52528aa18..2237d3f24f0e 100644
--- a/drivers/acpi/processor_idle.c
+++ b/drivers/acpi/processor_idle.c
@@ -31,6 +31,7 @@
 #include <linux/sched.h>       /* need_resched() */
 #include <linux/tick.h>
 #include <linux/cpuidle.h>
+#include <linux/cpu.h>
 #include <acpi/processor.h>
 
 /*
@@ -115,7 +116,7 @@ static const struct dmi_system_id processor_power_dmi_table[] = {
  * Callers should disable interrupts before the call and enable
  * interrupts after return.
  */
-static void acpi_safe_halt(void)
+static void __cpuidle acpi_safe_halt(void)
 {
 	if (!tif_need_resched()) {
 		safe_halt();
@@ -645,7 +646,7 @@ static int acpi_idle_bm_check(void)
  *
  * Caller disables interrupt before call and enables interrupt after return.
  */
-static void acpi_idle_do_entry(struct acpi_processor_cx *cx)
+static void __cpuidle acpi_idle_do_entry(struct acpi_processor_cx *cx)
 {
 	if (cx->entry_method == ACPI_CSTATE_FFH) {
 		/* Call into architectural FFH based C-state */
diff --git a/drivers/cpuidle/driver.c b/drivers/cpuidle/driver.c
index 389ade4572be..ab264d393233 100644
--- a/drivers/cpuidle/driver.c
+++ b/drivers/cpuidle/driver.c
@@ -14,6 +14,7 @@
 #include <linux/cpuidle.h>
 #include <linux/cpumask.h>
 #include <linux/tick.h>
+#include <linux/cpu.h>
 
 #include "cpuidle.h"
 
@@ -178,8 +179,8 @@ static void __cpuidle_driver_init(struct cpuidle_driver *drv)
 }
 
 #ifdef CONFIG_ARCH_HAS_CPU_RELAX
-static int poll_idle(struct cpuidle_device *dev,
-		struct cpuidle_driver *drv, int index)
+static int __cpuidle poll_idle(struct cpuidle_device *dev,
+			       struct cpuidle_driver *drv, int index)
 {
 	local_irq_enable();
 	if (!current_set_polling_and_test()) {
diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
index 67ec58f9ef99..4466a2f969d7 100644
--- a/drivers/idle/intel_idle.c
+++ b/drivers/idle/intel_idle.c
@@ -863,8 +863,8 @@ static struct cpuidle_state dnv_cstates[] = {
  *
  * Must be called under local_irq_disable().
  */
-static int intel_idle(struct cpuidle_device *dev,
-		struct cpuidle_driver *drv, int index)
+static __cpuidle int intel_idle(struct cpuidle_device *dev,
+				struct cpuidle_driver *drv, int index)
 {
 	unsigned long ecx = 1; /* break on interrupt flag */
 	struct cpuidle_state *state = &drv->states[index];
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index 24563970ff7b..3e42bcdd014b 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -454,6 +454,12 @@
 		*(.spinlock.text)					\
 		VMLINUX_SYMBOL(__lock_text_end) = .;
 
+#define CPUIDLE_TEXT							\
+		ALIGN_FUNCTION();					\
+		VMLINUX_SYMBOL(__cpuidle_text_start) = .;		\
+		*(.cpuidle.text)					\
+		VMLINUX_SYMBOL(__cpuidle_text_end) = .;
+
 #define KPROBES_TEXT							\
 		ALIGN_FUNCTION();					\
 		VMLINUX_SYMBOL(__kprobes_text_start) = .;		\
diff --git a/include/linux/cpu.h b/include/linux/cpu.h
index 797d9c8e9a1b..6babfa6db9d9 100644
--- a/include/linux/cpu.h
+++ b/include/linux/cpu.h
@@ -239,6 +239,11 @@ void cpu_startup_entry(enum cpuhp_state state);
 
 void cpu_idle_poll_ctrl(bool enable);
 
+/* Attach to any functions which should be considered cpuidle. */
+#define __cpuidle	__attribute__((__section__(".cpuidle.text")))
+
+bool cpu_in_idle(unsigned long pc);
+
 void arch_cpu_idle(void);
 void arch_cpu_idle_prepare(void);
 void arch_cpu_idle_enter(void);
diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c
index 9fb873cfc75c..1d8718d5300d 100644
--- a/kernel/sched/idle.c
+++ b/kernel/sched/idle.c
@@ -16,6 +16,9 @@
 
 #include "sched.h"
 
+/* Linker adds these: start and end of __cpuidle functions */
+extern char __cpuidle_text_start[], __cpuidle_text_end[];
+
 /**
  * sched_idle_set_state - Record idle state for the current CPU.
  * @idle_state: State to record.
@@ -53,7 +56,7 @@ static int __init cpu_idle_nopoll_setup(char *__unused)
 __setup("hlt", cpu_idle_nopoll_setup);
 #endif
 
-static inline int cpu_idle_poll(void)
+static noinline int __cpuidle cpu_idle_poll(void)
 {
 	rcu_idle_enter();
 	trace_cpu_idle_rcuidle(0, smp_processor_id());
@@ -84,7 +87,7 @@ void __weak arch_cpu_idle(void)
  *
  * To use when the cpuidle framework cannot be used.
  */
-void default_idle_call(void)
+void __cpuidle default_idle_call(void)
 {
 	if (current_clr_polling_and_test()) {
 		local_irq_enable();
@@ -271,6 +274,12 @@ static void cpu_idle_loop(void)
 	}
 }
 
+bool cpu_in_idle(unsigned long pc)
+{
+	return pc >= (unsigned long)__cpuidle_text_start &&
+		pc < (unsigned long)__cpuidle_text_end;
+}
+
 void cpu_startup_entry(enum cpuhp_state state)
 {
 	/*
diff --git a/lib/nmi_backtrace.c b/lib/nmi_backtrace.c
index 393a3cca1f47..75554754eadf 100644
--- a/lib/nmi_backtrace.c
+++ b/lib/nmi_backtrace.c
@@ -16,6 +16,7 @@
 #include <linux/delay.h>
 #include <linux/kprobes.h>
 #include <linux/nmi.h>
+#include <linux/cpu.h>
 
 #ifdef arch_trigger_cpumask_backtrace
 /* For reliability, we're prepared to waste bits here. */
@@ -87,11 +88,16 @@ bool nmi_cpu_backtrace(struct pt_regs *regs)
 	int cpu = smp_processor_id();
 
 	if (cpumask_test_cpu(cpu, to_cpumask(backtrace_mask))) {
-		pr_warn("NMI backtrace for cpu %d\n", cpu);
-		if (regs)
-			show_regs(regs);
-		else
-			dump_stack();
+		if (regs && cpu_in_idle(instruction_pointer(regs))) {
+			pr_warn("NMI backtrace for cpu %d skipped: idling at pc %#lx\n",
+				cpu, instruction_pointer(regs));
+		} else {
+			pr_warn("NMI backtrace for cpu %d\n", cpu);
+			if (regs)
+				show_regs(regs);
+			else
+				dump_stack();
+		}
 		cpumask_clear_cpu(cpu, to_cpumask(backtrace_mask));
 		return true;
 	}
diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
index 48958d3cec9e..bd8349759095 100644
--- a/scripts/mod/modpost.c
+++ b/scripts/mod/modpost.c
@@ -888,7 +888,7 @@ static void check_section(const char *modname, struct elf_info *elf,
 
 #define DATA_SECTIONS ".data", ".data.rel"
 #define TEXT_SECTIONS ".text", ".text.unlikely", ".sched.text", \
-		".kprobes.text"
+		".kprobes.text", ".cpuidle.text"
 #define OTHER_TEXT_SECTIONS ".ref.text", ".head.text", ".spinlock.text", \
 		".fixup", ".entry.text", ".exception.text", ".text.*", \
 		".coldtext"
diff --git a/scripts/recordmcount.c b/scripts/recordmcount.c
index 42396a74405d..c0222107cf58 100644
--- a/scripts/recordmcount.c
+++ b/scripts/recordmcount.c
@@ -364,6 +364,7 @@ is_mcounted_section_name(char const *const txtname)
 		strcmp(".spinlock.text", txtname) == 0 ||
 		strcmp(".irqentry.text", txtname) == 0 ||
 		strcmp(".kprobes.text", txtname) == 0 ||
+		strcmp(".cpuidle.text", txtname) == 0 ||
 		strcmp(".text.unlikely", txtname) == 0;
 }
 
diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl
index 96e2486a6fc4..29cecf9b504f 100755
--- a/scripts/recordmcount.pl
+++ b/scripts/recordmcount.pl
@@ -135,6 +135,7 @@ my %text_sections = (
      ".spinlock.text" => 1,
      ".irqentry.text" => 1,
      ".kprobes.text" => 1,
+     ".cpuidle.text" => 1,
      ".text.unlikely" => 1,
 );
 
-- 
2.7.2

Re: [PATCH v9 4/4] nmi_backtrace: generate one-line reports for idle cpus

[Petr Mladek]

On Mon 2016-08-29 12:12:49, Chris Metcalf wrote:
> When doing an nmi backtrace of many cores, most of which are idle,
> the output is a little overwhelming and very uninformative.  Suppress
> messages for cpus that are idling when they are interrupted and just
> emit one line, "NMI backtrace for N skipped: idling at pc 0xNNN".
> 
> We do this by grouping all the cpuidle code together into a new
> .cpuidle.text section, and then checking the address of the
> interrupted PC to see if it lies within that section.
> 
> This commit suitably tags x86 and tile idle routines, and only
> adds in the minimal framework for other architectures.
> 
> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> Tested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> Tested-by: Daniel Thompson <daniel.thompson@linaro.org> [arm]
> Signed-off-by: Chris Metcalf <cmetcalf@mellanox.com>

This version works even with my configuration. Feel free to
add

Tested-by: Petr Mladek <pmladek@suse.com>

Best Regards,
Petr