[PATCH 1/2] oprofile: Disable oprofile NMI timer on ppc64

[PATCH 1/2] oprofile: Disable oprofile NMI timer on ppc64

[Anton Blanchard]

We want to enable the hard lockup detector on ppc64, but right now
that enables the oprofile NMI timer too.

We'd prefer not to enable the oprofile NMI timer, it adds another
element to our PMU testing and it requires us to increase our
exported symbols (eg cpu_khz).

Modify the config entry for OPROFILE_NMI_TIMER to disable it on PPC64.

Signed-off-by: Anton Blanchard <anton@samba.org>
---
 arch/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/Kconfig b/arch/Kconfig
index 05d7a8a..0cc605d 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -32,7 +32,7 @@ config HAVE_OPROFILE
 
 config OPROFILE_NMI_TIMER
 	def_bool y
-	depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
+	depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !PPC64
 
 config KPROBES
 	bool "Kprobes"
-- 
2.1.0

[PATCH 2/2] powerpc: Add ppc64 hard lockup detector support

[Anton Blanchard]

The hard lockup detector uses a PMU event as a periodic NMI to
detect if we are stuck (where stuck means no timer interrupts have
occurred).

Ben's rework of the ppc64 soft disable code has made ppc64 PMU
exceptions a partial NMI. They can get disabled if an external
interrupt comes in, but otherwise PMU interrupts will fire in
interrupt disabled regions.

We disable the hard lockup detector by default for a few reasons:

- It breaks userspace event based branches on POWER8.
- It is likely to produce false positives on KVM guests.
- Since PMCs can only count to 2^31, counting cycles means we might
  take multiple PMU exceptions per second per hardware thread even
  if our hard lockup timeout is 10 seconds.

It can be enabled via a boot option, or via procfs.

Signed-off-by: Anton Blanchard <anton@samba.org>
---
 arch/powerpc/Kconfig           |  1 +
 arch/powerpc/include/asm/nmi.h |  4 ++++
 arch/powerpc/kernel/setup_64.c | 20 ++++++++++++++++++++
 3 files changed, 25 insertions(+)
 create mode 100644 arch/powerpc/include/asm/nmi.h

diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index c102668..716c9e6 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -152,6 +152,7 @@ config PPC
 	select DCACHE_WORD_ACCESS if PPC64 && CPU_LITTLE_ENDIAN
 	select NO_BOOTMEM
 	select HAVE_GENERIC_RCU_GUP
+	select HAVE_PERF_EVENTS_NMI if PPC64
 
 config GENERIC_CSUM
 	def_bool CPU_LITTLE_ENDIAN
diff --git a/arch/powerpc/include/asm/nmi.h b/arch/powerpc/include/asm/nmi.h
new file mode 100644
index 0000000..ff1ccb3
--- /dev/null
+++ b/arch/powerpc/include/asm/nmi.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_NMI_H
+#define _ASM_NMI_H
+
+#endif /* _ASM_NMI_H */
diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index 300be7d..ae39ec6 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -37,6 +37,7 @@
 #include <linux/memblock.h>
 #include <linux/hugetlb.h>
 #include <linux/memory.h>
+#include <linux/nmi.h>
 
 #include <asm/io.h>
 #include <asm/kdump.h>
@@ -779,3 +780,22 @@ unsigned long memory_block_size_bytes(void)
 struct ppc_pci_io ppc_pci_io;
 EXPORT_SYMBOL(ppc_pci_io);
 #endif
+
+#ifdef CONFIG_HARDLOCKUP_DETECTOR
+u64 hw_nmi_get_sample_period(int watchdog_thresh)
+{
+	return ppc_proc_freq * watchdog_thresh;
+}
+
+/*
+ * The hardlockup detector breaks PMU event based branches and is likely
+ * to get false positives in KVM guests, so disable it by default.
+ */
+static int __init disable_hardlockup_detector(void)
+{
+	watchdog_enable_hardlockup_detector(false);
+
+	return 0;
+}
+early_initcall(disable_hardlockup_detector);
+#endif
-- 
2.1.0

Re: [PATCH 2/2] powerpc: Add ppc64 hard lockup detector support

[Scott Wood]

On Thu, 2015-04-09 at 12:52 +1000, Anton Blanchard wrote:
> The hard lockup detector uses a PMU event as a periodic NMI to
> detect if we are stuck (where stuck means no timer interrupts have
> occurred).
> 
> Ben's rework of the ppc64 soft disable code has made ppc64 PMU
> exceptions a partial NMI. They can get disabled if an external
> interrupt comes in, but otherwise PMU interrupts will fire in
> interrupt disabled regions.
> 
> We disable the hard lockup detector by default for a few reasons:
> 
> - It breaks userspace event based branches on POWER8.
> - It is likely to produce false positives on KVM guests.

What causes the false positives with KVM?  I'm wondering if it makes
sense to enable this by default for book3e.

> - Since PMCs can only count to 2^31, counting cycles means we might
>   take multiple PMU exceptions per second per hardware thread even
>   if our hard lockup timeout is 10 seconds.

It'd be nice if this could be used with some event other than
PERF_COUNT_HW_CPU_CYCLES, such as something based on the timebase -- or
using the actual watchdog on book3e which would not consume a perf
counter and would be a full NMI as long as MSR[CE] is left on.

-Scott

Re: [PATCH 1/2] oprofile: Disable oprofile NMI timer on ppc64

[Robert Richter]

On 09.04.15 12:52:55, Anton Blanchard wrote:
> We want to enable the hard lockup detector on ppc64, but right now
> that enables the oprofile NMI timer too.
> 
> We'd prefer not to enable the oprofile NMI timer, it adds another
> element to our PMU testing and it requires us to increase our
> exported symbols (eg cpu_khz).
> 
> Modify the config entry for OPROFILE_NMI_TIMER to disable it on PPC64.
> 
> Signed-off-by: Anton Blanchard <anton@samba.org>

Acked-by: Robert Richter <rric@kernel.org>

Please pass this patch along with patch 2 of this series (ppc tree or
so).

Thanks,

-Robert

> ---
>  arch/Kconfig | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/Kconfig b/arch/Kconfig
> index 05d7a8a..0cc605d 100644
> --- a/arch/Kconfig
> +++ b/arch/Kconfig
> @@ -32,7 +32,7 @@ config HAVE_OPROFILE
>  
>  config OPROFILE_NMI_TIMER
>  	def_bool y
> -	depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
> +	depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !PPC64
>  
>  config KPROBES
>  	bool "Kprobes"
> -- 
> 2.1.0
> 

[PATCH 1/2] oprofile: Add HAVE_OPROFILE_NMI_TIMER

[Anton Blanchard]

HAVE_PERF_EVENTS_NMI is used for two things - the oprofile NMI timer
and the hardlockup detector.

Create HAVE_OPROFILE_NMI_TIMER so an architecture can select them
separately. On ppc64 we want to add the hardlockup detector, but not
the oprofile NMI timer fallback.

Signed-off-by: Anton Blanchard <anton@samba.org>
---
 arch/Kconfig     | 5 ++++-
 arch/x86/Kconfig | 1 +
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/arch/Kconfig b/arch/Kconfig
index 05d7a8a..3d50e42 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -30,9 +30,12 @@ config OPROFILE_EVENT_MULTIPLEX
 config HAVE_OPROFILE
 	bool
 
+config HAVE_OPROFILE_NMI_TIMER
+	bool
+
 config OPROFILE_NMI_TIMER
 	def_bool y
-	depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
+	depends on PERF_EVENTS && HAVE_OPROFILE_NMI_TIMER
 
 config KPROBES
 	bool "Kprobes"
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index ba397bd..98d9c72 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -138,6 +138,7 @@ config X86
 	select HAVE_ACPI_APEI_NMI if ACPI
 	select ACPI_LEGACY_TABLES_LOOKUP if ACPI
 	select X86_FEATURE_NAMES if PROC_FS
+	select HAVE_OPROFILE_NMI_TIMER
 
 config INSTRUCTION_DECODER
 	def_bool y
-- 
2.1.0

[PATCH 2/2] powerpc: Add ppc64 hard lockup detector support

[Anton Blanchard]

The hard lockup detector uses a PMU event as a periodic NMI to
detect if we are stuck (where stuck means no timer interrupts have
occurred).

Ben's rework of the ppc64 soft disable code has made ppc64 PMU
exceptions a partial NMI. They can get disabled if an external
interrupt comes in, but otherwise PMU interrupts will fire in
interrupt disabled regions.

We disable the hard lockup detector by default for a few reasons:

- It breaks userspace event based branches on POWER8.
- It is likely to produce false positives on KVM guests.
- Since PMCs can only count to 2^31, counting cycles means we might
  take multiple PMU exceptions per second per hardware thread even
  if our hard lockup timeout is 10 seconds.

It can be enabled via a boot option, or via procfs.

Signed-off-by: Anton Blanchard <anton@samba.org>
---
 arch/powerpc/Kconfig           |  1 +
 arch/powerpc/include/asm/nmi.h |  4 ++++
 arch/powerpc/kernel/setup_64.c | 20 ++++++++++++++++++++
 3 files changed, 25 insertions(+)
 create mode 100644 arch/powerpc/include/asm/nmi.h

diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index a2a168e..f6f4734 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -152,6 +152,7 @@ config PPC
 	select DCACHE_WORD_ACCESS if PPC64 && CPU_LITTLE_ENDIAN
 	select NO_BOOTMEM
 	select HAVE_GENERIC_RCU_GUP
+	select HAVE_PERF_EVENTS_NMI if PPC64
 
 config GENERIC_CSUM
 	def_bool CPU_LITTLE_ENDIAN
diff --git a/arch/powerpc/include/asm/nmi.h b/arch/powerpc/include/asm/nmi.h
new file mode 100644
index 0000000..ff1ccb3
--- /dev/null
+++ b/arch/powerpc/include/asm/nmi.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_NMI_H
+#define _ASM_NMI_H
+
+#endif /* _ASM_NMI_H */
diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index 300be7d..ae39ec6 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -37,6 +37,7 @@
 #include <linux/memblock.h>
 #include <linux/hugetlb.h>
 #include <linux/memory.h>
+#include <linux/nmi.h>
 
 #include <asm/io.h>
 #include <asm/kdump.h>
@@ -779,3 +780,22 @@ unsigned long memory_block_size_bytes(void)
 struct ppc_pci_io ppc_pci_io;
 EXPORT_SYMBOL(ppc_pci_io);
 #endif
+
+#ifdef CONFIG_HARDLOCKUP_DETECTOR
+u64 hw_nmi_get_sample_period(int watchdog_thresh)
+{
+	return ppc_proc_freq * watchdog_thresh;
+}
+
+/*
+ * The hardlockup detector breaks PMU event based branches and is likely
+ * to get false positives in KVM guests, so disable it by default.
+ */
+static int __init disable_hardlockup_detector(void)
+{
+	watchdog_enable_hardlockup_detector(false);
+
+	return 0;
+}
+early_initcall(disable_hardlockup_detector);
+#endif
-- 
2.1.0

[PATCH 1/2] powerpc: Hard disable interrupts in xmon

[Anton Blanchard]

xmon only soft disables interrupts. This seems like a bad idea - we
certainly don't want decrementer and PMU exceptions going off when
we are debugging something inside xmon.

This issue was uncovered when the hard lockup detector went off
inside xmon. To ensure we wont get a spurious hard lockup warning,
I also call touch_nmi_watchdog() when exiting xmon.

Signed-off-by: Anton Blanchard <anton@samba.org>
---

Index: b/arch/powerpc/xmon/xmon.c
===================================================================
--- a/arch/powerpc/xmon/xmon.c
+++ b/arch/powerpc/xmon/xmon.c
@@ -24,6 +24,7 @@
 #include <linux/interrupt.h>
 #include <linux/irq.h>
 #include <linux/bug.h>
+#include <linux/nmi.h>
 
 #include <asm/ptrace.h>
 #include <asm/string.h>
@@ -374,6 +375,7 @@ static int xmon_core(struct pt_regs *reg
 #endif
 
 	local_irq_save(flags);
+	hard_irq_disable();
 
 	bp = in_breakpoint_table(regs->nip, &offset);
 	if (bp != NULL) {
@@ -558,6 +560,7 @@ static int xmon_core(struct pt_regs *reg
 #endif
 	insert_cpu_bpts();
 
+	touch_nmi_watchdog();
 	local_irq_restore(flags);
 
 	return cmd != 'X' && cmd != EOF;

[PATCH 2/2] powerpc: Add ppc64 hard lockup detector support

[Anton Blanchard]

The hard lockup detector uses a PMU event as a periodic NMI to
detect if we are stuck (where stuck means no timer interrupts have
occurred).

Ben's rework of the ppc64 soft disable code has made ppc64 PMU
exceptions a partial NMI. They can get disabled if an external interrupt
comes in, but otherwise PMU interrupts will fire in interrupt disabled
regions.

I wrote a kernel module to test this patch and noticed we sometimes
missed hard lockup warnings. The RCU code detected the stall first and
issued an IPI to backtrace all CPUs. Unfortunately an IPI is an external
interrupt and that will hard disable interrupts, preventing the hard
lockup detector from going off.

If I reduced the hard lockup threshold to 5 seconds:

echo 5 > /proc/sys/kernel/watchdog_thresh

Then it would beat the RCU code in detecting a stall and get a
correct backtrace out.

Another downside is that our PMCs can only count to 2^31, so even when
we ask for 10 seconds of processor cycles, we end up taking a couple
of PMU exceptions a second.

Signed-off-by: Anton Blanchard <anton@samba.org>
---

Index: b/arch/powerpc/Kconfig
===================================================================
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -145,6 +145,7 @@ config PPC
 	select HAVE_IRQ_EXIT_ON_IRQ_STACK
 	select ARCH_USE_CMPXCHG_LOCKREF if PPC64
 	select HAVE_ARCH_AUDITSYSCALL
+	select HAVE_PERF_EVENTS_NMI if PPC64
 
 config GENERIC_CSUM
 	def_bool CPU_LITTLE_ENDIAN
Index: b/arch/powerpc/include/asm/nmi.h
===================================================================
--- /dev/null
+++ b/arch/powerpc/include/asm/nmi.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_NMI_H
+#define _ASM_NMI_H
+
+#endif /* _ASM_NMI_H */
Index: b/arch/powerpc/kernel/setup_64.c
===================================================================
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -796,3 +796,10 @@ unsigned long memory_block_size_bytes(vo
 struct ppc_pci_io ppc_pci_io;
 EXPORT_SYMBOL(ppc_pci_io);
 #endif
+
+#ifdef CONFIG_HARDLOCKUP_DETECTOR
+u64 hw_nmi_get_sample_period(int watchdog_thresh)
+{
+	return ppc_proc_freq * watchdog_thresh;
+}
+#endif

[PATCH 2/2] powerpc: Add ppc64 hard lockup detector support

[Anton Blanchard]

The hard lockup detector uses a PMU event as a periodic NMI to
detect if we are stuck (where stuck means no timer interrupts have
occurred).

Ben's rework of the ppc64 soft disable code has made ppc64 PMU
exceptions a partial NMI. They can get disabled if an external interrupt
comes in, but otherwise PMU interrupts will fire in interrupt disabled
regions.

I wrote a kernel module to test this patch and noticed we sometimes
missed hard lockup warnings. The RCU code detected the stall first and
issued an IPI to backtrace all CPUs. Unfortunately an IPI is an external
interrupt and that will hard disable interrupts, preventing the hard
lockup detector from going off.

If I reduced the hard lockup threshold to 5 seconds:

echo 5 > /proc/sys/kernel/watchdog_thresh

Then it would beat the RCU code in detecting a stall and get a
correct backtrace out.

Another downside is that our PMCs can only count to 2^31, so even when
we ask for 10 seconds of processor cycles, we end up taking a couple
of PMU exceptions a second.

Signed-off-by: Anton Blanchard <anton@samba.org>
---

v2: Mikey noticed a build issue with oprofile. Since our NMI is just
the PMU hardware it doesn't make any sense for oprofile to try and
use it.

Index: b/arch/powerpc/Kconfig
===================================================================
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -145,6 +145,7 @@ config PPC
 	select HAVE_IRQ_EXIT_ON_IRQ_STACK
 	select ARCH_USE_CMPXCHG_LOCKREF if PPC64
 	select HAVE_ARCH_AUDITSYSCALL
+	select HAVE_PERF_EVENTS_NMI if PPC64
 
 config GENERIC_CSUM
 	def_bool CPU_LITTLE_ENDIAN
Index: b/arch/powerpc/include/asm/nmi.h
===================================================================
--- /dev/null
+++ b/arch/powerpc/include/asm/nmi.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_NMI_H
+#define _ASM_NMI_H
+
+#endif /* _ASM_NMI_H */
Index: b/arch/powerpc/kernel/setup_64.c
===================================================================
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -800,3 +800,10 @@ unsigned long memory_block_size_bytes(vo
 struct ppc_pci_io ppc_pci_io;
 EXPORT_SYMBOL(ppc_pci_io);
 #endif
+
+#ifdef CONFIG_HARDLOCKUP_DETECTOR
+u64 hw_nmi_get_sample_period(int watchdog_thresh)
+{
+	return ppc_proc_freq * watchdog_thresh;
+}
+#endif
Index: b/arch/Kconfig
===================================================================
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -32,7 +32,7 @@ config HAVE_OPROFILE
 
 config OPROFILE_NMI_TIMER
 	def_bool y
-	depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
+	depends on (PERF_EVENTS && HAVE_PERF_EVENTS_NMI) && !PPC
 
 config KPROBES
 	bool "Kprobes"

Re: [PATCH 2/2] powerpc: Add ppc64 hard lockup detector support

[Paul E. McKenney]

On Tue, Aug 12, 2014 at 09:31:37AM +1000, Anton Blanchard wrote:
> The hard lockup detector uses a PMU event as a periodic NMI to
> detect if we are stuck (where stuck means no timer interrupts have
> occurred).
> 
> Ben's rework of the ppc64 soft disable code has made ppc64 PMU
> exceptions a partial NMI. They can get disabled if an external interrupt
> comes in, but otherwise PMU interrupts will fire in interrupt disabled
> regions.
> 
> I wrote a kernel module to test this patch and noticed we sometimes
> missed hard lockup warnings. The RCU code detected the stall first and
> issued an IPI to backtrace all CPUs. Unfortunately an IPI is an external
> interrupt and that will hard disable interrupts, preventing the hard
> lockup detector from going off.

If it helps, commit bc1dce514e9b (rcu: Don't use NMIs to dump other
CPUs' stacks) makes RCU avoid this behavior.  It instead reads the
stacks out remotely when this commit is applied.  It is in -tip, and
should make mainline this merge window.  Corresponding patch below.

						Thanx, Paul

------------------------------------------------------------------------

rcu: Don't use NMIs to dump other CPUs' stacks

Although NMI-based stack dumps are in principle more accurate, they are
also more likely to trigger deadlocks.  This commit therefore replaces
all uses of trigger_all_cpu_backtrace() with rcu_dump_cpu_stacks(), so
that the CPU detecting an RCU CPU stall does the stack dumping.

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Lai Jiangshan <laijs@cn.fujitsu.com>

diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index 3f93033d3c61..8f3e4d43d736 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -1013,10 +1013,7 @@ static void record_gp_stall_check_time(struct rcu_state *rsp)
 }
 
 /*
- * Dump stacks of all tasks running on stalled CPUs.  This is a fallback
- * for architectures that do not implement trigger_all_cpu_backtrace().
- * The NMI-triggered stack traces are more accurate because they are
- * printed by the target CPU.
+ * Dump stacks of all tasks running on stalled CPUs.
  */
 static void rcu_dump_cpu_stacks(struct rcu_state *rsp)
 {
@@ -1094,7 +1091,7 @@ static void print_other_cpu_stall(struct rcu_state *rsp)
 	       (long)rsp->gpnum, (long)rsp->completed, totqlen);
 	if (ndetected == 0)
 		pr_err("INFO: Stall ended before state dump start\n");
-	else if (!trigger_all_cpu_backtrace())
+	else
 		rcu_dump_cpu_stacks(rsp);
 
 	/* Complain about tasks blocking the grace period. */
@@ -1125,8 +1122,7 @@ static void print_cpu_stall(struct rcu_state *rsp)
 	pr_cont(" (t=%lu jiffies g=%ld c=%ld q=%lu)\n",
 		jiffies - rsp->gp_start,
 		(long)rsp->gpnum, (long)rsp->completed, totqlen);
-	if (!trigger_all_cpu_backtrace())
-		dump_stack();
+	rcu_dump_cpu_stacks(rsp);
 
 	raw_spin_lock_irqsave(&rnp->lock, flags);
 	if (ULONG_CMP_GE(jiffies, ACCESS_ONCE(rsp->jiffies_stall)))