* [PATCH v4 1/5] watchdog: remove unused declaration
2017-06-16 6:57 [PATCH v4 0/5] Improve watchdog config for arch watchdogs Nicholas Piggin
@ 2017-06-16 6:57 ` Nicholas Piggin
2017-06-16 6:57 ` [PATCH v4 2/5] watchdog: introduce arch_touch_nmi_watchdog() Nicholas Piggin
` (4 subsequent siblings)
5 siblings, 0 replies; 13+ messages in thread
From: Nicholas Piggin @ 2017-06-16 6:57 UTC (permalink / raw)
To: Don Zickus; +Cc: Nicholas Piggin, Babu Moger, linux-kernel, linuxppc-dev
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
include/linux/nmi.h | 3 ---
1 file changed, 3 deletions(-)
diff --git a/include/linux/nmi.h b/include/linux/nmi.h
index aa3cd0878270..5e2e57536d98 100644
--- a/include/linux/nmi.h
+++ b/include/linux/nmi.h
@@ -12,9 +12,6 @@ extern void touch_softlockup_watchdog_sched(void);
extern void touch_softlockup_watchdog(void);
extern void touch_softlockup_watchdog_sync(void);
extern void touch_all_softlockup_watchdogs(void);
-extern int proc_dowatchdog_thresh(struct ctl_table *table, int write,
- void __user *buffer,
- size_t *lenp, loff_t *ppos);
extern unsigned int softlockup_panic;
extern unsigned int hardlockup_panic;
void lockup_detector_init(void);
--
2.11.0
^ permalink raw reply related [flat|nested] 13+ messages in thread* [PATCH v4 2/5] watchdog: introduce arch_touch_nmi_watchdog()
2017-06-16 6:57 [PATCH v4 0/5] Improve watchdog config for arch watchdogs Nicholas Piggin
2017-06-16 6:57 ` [PATCH v4 1/5] watchdog: remove unused declaration Nicholas Piggin
@ 2017-06-16 6:57 ` Nicholas Piggin
2017-06-16 18:21 ` Andrew Morton
2017-06-16 6:57 ` [PATCH v4 3/5] watchdog: split up config options Nicholas Piggin
` (3 subsequent siblings)
5 siblings, 1 reply; 13+ messages in thread
From: Nicholas Piggin @ 2017-06-16 6:57 UTC (permalink / raw)
To: Don Zickus; +Cc: Nicholas Piggin, Babu Moger, linux-kernel, linuxppc-dev
For architectures that define HAVE_NMI_WATCHDOG, instead of having
them provide the complete touch_nmi_watchdog() function, just have
them provide arch_touch_nmi_watchdog().
This gives the generic code more flexibility in implementing this
function, and arch implementations don't miss out on touching the
softlockup watchdog or other generic details.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/blackfin/include/asm/nmi.h | 2 ++
arch/blackfin/kernel/nmi.c | 2 +-
arch/mn10300/include/asm/nmi.h | 2 ++
arch/mn10300/kernel/mn10300-watchdog-low.S | 8 ++++----
arch/mn10300/kernel/mn10300-watchdog.c | 2 +-
arch/sparc/include/asm/nmi.h | 1 +
arch/sparc/kernel/nmi.c | 6 ++----
include/linux/nmi.h | 27 ++++++++++++++++-----------
kernel/watchdog_hld.c | 5 ++---
9 files changed, 31 insertions(+), 24 deletions(-)
diff --git a/arch/blackfin/include/asm/nmi.h b/arch/blackfin/include/asm/nmi.h
index b9caac4fcfd8..107d23705f46 100644
--- a/arch/blackfin/include/asm/nmi.h
+++ b/arch/blackfin/include/asm/nmi.h
@@ -9,4 +9,6 @@
#include <linux/nmi.h>
+extern void arch_touch_nmi_watchdog(void);
+
#endif
diff --git a/arch/blackfin/kernel/nmi.c b/arch/blackfin/kernel/nmi.c
index 633c37083e87..1e714329fe8a 100644
--- a/arch/blackfin/kernel/nmi.c
+++ b/arch/blackfin/kernel/nmi.c
@@ -190,7 +190,7 @@ static int __init init_nmi_wdt(void)
}
device_initcall(init_nmi_wdt);
-void touch_nmi_watchdog(void)
+void arch_touch_nmi_watchdog(void)
{
atomic_set(&nmi_touched[smp_processor_id()], 1);
}
diff --git a/arch/mn10300/include/asm/nmi.h b/arch/mn10300/include/asm/nmi.h
index f3671cbbc117..b05627597b1b 100644
--- a/arch/mn10300/include/asm/nmi.h
+++ b/arch/mn10300/include/asm/nmi.h
@@ -11,4 +11,6 @@
#ifndef _ASM_NMI_H
#define _ASM_NMI_H
+extern void arch_touch_nmi_watchdog(void);
+
#endif /* _ASM_NMI_H */
diff --git a/arch/mn10300/kernel/mn10300-watchdog-low.S b/arch/mn10300/kernel/mn10300-watchdog-low.S
index f2f5c9cfaabd..34f8773de7d0 100644
--- a/arch/mn10300/kernel/mn10300-watchdog-low.S
+++ b/arch/mn10300/kernel/mn10300-watchdog-low.S
@@ -50,9 +50,9 @@ watchdog_handler:
# we can't inline it)
#
###############################################################################
- .globl touch_nmi_watchdog
- .type touch_nmi_watchdog,@function
-touch_nmi_watchdog:
+ .globl arch_touch_nmi_watchdog
+ .type arch_touch_nmi_watchdog,@function
+arch_touch_nmi_watchdog:
clr d0
clr d1
mov watchdog_alert_counter, a0
@@ -63,4 +63,4 @@ touch_nmi_watchdog:
lne
ret [],0
- .size touch_nmi_watchdog,.-touch_nmi_watchdog
+ .size arch_touch_nmi_watchdog,.-arch_touch_nmi_watchdog
diff --git a/arch/mn10300/kernel/mn10300-watchdog.c b/arch/mn10300/kernel/mn10300-watchdog.c
index a2d8e6938d67..0d5641beadf5 100644
--- a/arch/mn10300/kernel/mn10300-watchdog.c
+++ b/arch/mn10300/kernel/mn10300-watchdog.c
@@ -31,7 +31,7 @@ static unsigned int watchdog;
static unsigned int watchdog_hz = 1;
unsigned int watchdog_alert_counter[NR_CPUS];
-EXPORT_SYMBOL(touch_nmi_watchdog);
+EXPORT_SYMBOL(arch_touch_nmi_watchdog);
/*
* the best way to detect whether a CPU has a 'hard lockup' problem
diff --git a/arch/sparc/include/asm/nmi.h b/arch/sparc/include/asm/nmi.h
index 26ad2b2607c6..284eac3ffaf2 100644
--- a/arch/sparc/include/asm/nmi.h
+++ b/arch/sparc/include/asm/nmi.h
@@ -7,6 +7,7 @@ void nmi_adjust_hz(unsigned int new_hz);
extern atomic_t nmi_active;
+void arch_touch_nmi_watchdog(void);
void start_nmi_watchdog(void *unused);
void stop_nmi_watchdog(void *unused);
diff --git a/arch/sparc/kernel/nmi.c b/arch/sparc/kernel/nmi.c
index 95e73c63c99d..048ad783ea3f 100644
--- a/arch/sparc/kernel/nmi.c
+++ b/arch/sparc/kernel/nmi.c
@@ -51,7 +51,7 @@ static DEFINE_PER_CPU(unsigned int, last_irq_sum);
static DEFINE_PER_CPU(long, alert_counter);
static DEFINE_PER_CPU(int, nmi_touch);
-void touch_nmi_watchdog(void)
+void arch_touch_nmi_watchdog(void)
{
if (atomic_read(&nmi_active)) {
int cpu;
@@ -61,10 +61,8 @@ void touch_nmi_watchdog(void)
per_cpu(nmi_touch, cpu) = 1;
}
}
-
- touch_softlockup_watchdog();
}
-EXPORT_SYMBOL(touch_nmi_watchdog);
+EXPORT_SYMBOL(arch_touch_nmi_watchdog);
static void die_nmi(const char *str, struct pt_regs *regs, int do_panic)
{
diff --git a/include/linux/nmi.h b/include/linux/nmi.h
index 5e2e57536d98..bd387ef8bccd 100644
--- a/include/linux/nmi.h
+++ b/include/linux/nmi.h
@@ -6,6 +6,9 @@
#include <linux/sched.h>
#include <asm/irq.h>
+#if defined(CONFIG_HAVE_NMI_WATCHDOG)
+#include <asm/nmi.h>
+#endif
#ifdef CONFIG_LOCKUP_DETECTOR
extern void touch_softlockup_watchdog_sched(void);
@@ -58,6 +61,18 @@ static inline void reset_hung_task_detector(void)
#define NMI_WATCHDOG_ENABLED (1 << NMI_WATCHDOG_ENABLED_BIT)
#define SOFT_WATCHDOG_ENABLED (1 << SOFT_WATCHDOG_ENABLED_BIT)
+#if defined(CONFIG_HARDLOCKUP_DETECTOR)
+extern void hardlockup_detector_disable(void);
+#else
+static inline void hardlockup_detector_disable(void) {}
+#endif
+
+#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG)
+extern void arch_touch_nmi_watchdog(void);
+#else
+static inline void arch_touch_nmi_watchdog(void) {}
+#endif
+
/**
* touch_nmi_watchdog - restart NMI watchdog timeout.
*
@@ -65,21 +80,11 @@ static inline void reset_hung_task_detector(void)
* may be used to reset the timeout - for code which intentionally
* disables interrupts for a long time. This call is stateless.
*/
-#if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR)
-#include <asm/nmi.h>
-extern void touch_nmi_watchdog(void);
-#else
static inline void touch_nmi_watchdog(void)
{
+ arch_touch_nmi_watchdog();
touch_softlockup_watchdog();
}
-#endif
-
-#if defined(CONFIG_HARDLOCKUP_DETECTOR)
-extern void hardlockup_detector_disable(void);
-#else
-static inline void hardlockup_detector_disable(void) {}
-#endif
/*
* Create trigger_all_cpu_backtrace() out of the arch-provided
diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c
index 54a427d1f344..90d688df6ce1 100644
--- a/kernel/watchdog_hld.c
+++ b/kernel/watchdog_hld.c
@@ -56,7 +56,7 @@ static int __init hardlockup_panic_setup(char *str)
}
__setup("nmi_watchdog=", hardlockup_panic_setup);
-void touch_nmi_watchdog(void)
+void arch_touch_nmi_watchdog(void)
{
/*
* Using __raw here because some code paths have
@@ -66,9 +66,8 @@ void touch_nmi_watchdog(void)
* going off.
*/
raw_cpu_write(watchdog_nmi_touch, true);
- touch_softlockup_watchdog();
}
-EXPORT_SYMBOL(touch_nmi_watchdog);
+EXPORT_SYMBOL(arch_touch_nmi_watchdog);
static struct perf_event_attr wd_hw_attr = {
.type = PERF_TYPE_HARDWARE,
--
2.11.0
^ permalink raw reply related [flat|nested] 13+ messages in thread* Re: [PATCH v4 2/5] watchdog: introduce arch_touch_nmi_watchdog()
2017-06-16 6:57 ` [PATCH v4 2/5] watchdog: introduce arch_touch_nmi_watchdog() Nicholas Piggin
@ 2017-06-16 18:21 ` Andrew Morton
2017-06-17 2:43 ` Nicholas Piggin
0 siblings, 1 reply; 13+ messages in thread
From: Andrew Morton @ 2017-06-16 18:21 UTC (permalink / raw)
To: Nicholas Piggin; +Cc: Don Zickus, Babu Moger, linux-kernel, linuxppc-dev
On Fri, 16 Jun 2017 16:57:12 +1000 Nicholas Piggin <npiggin@gmail.com> wrote:
> For architectures that define HAVE_NMI_WATCHDOG, instead of having
> them provide the complete touch_nmi_watchdog() function, just have
> them provide arch_touch_nmi_watchdog().
>
> This gives the generic code more flexibility in implementing this
> function, and arch implementations don't miss out on touching the
> softlockup watchdog or other generic details.
>
> ...
>
> --- a/arch/blackfin/include/asm/nmi.h
> +++ b/arch/blackfin/include/asm/nmi.h
> @@ -9,4 +9,6 @@
>
> #include <linux/nmi.h>
>
> +extern void arch_touch_nmi_watchdog(void);
Do we actually need to add this to the arch header files...
>
> ...
>
> --- a/include/linux/nmi.h
> +++ b/include/linux/nmi.h
> @@ -6,6 +6,9 @@
>
> #include <linux/sched.h>
> #include <asm/irq.h>
> +#if defined(CONFIG_HAVE_NMI_WATCHDOG)
> +#include <asm/nmi.h>
> +#endif
>
> #ifdef CONFIG_LOCKUP_DETECTOR
> extern void touch_softlockup_watchdog_sched(void);
> @@ -58,6 +61,18 @@ static inline void reset_hung_task_detector(void)
> #define NMI_WATCHDOG_ENABLED (1 << NMI_WATCHDOG_ENABLED_BIT)
> #define SOFT_WATCHDOG_ENABLED (1 << SOFT_WATCHDOG_ENABLED_BIT)
>
> +#if defined(CONFIG_HARDLOCKUP_DETECTOR)
> +extern void hardlockup_detector_disable(void);
> +#else
> +static inline void hardlockup_detector_disable(void) {}
> +#endif
> +
> +#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG)
> +extern void arch_touch_nmi_watchdog(void);
> +#else
> +static inline void arch_touch_nmi_watchdog(void) {}
> +#endif
> +
given that we have a global declaration here?
^ permalink raw reply [flat|nested] 13+ messages in thread* Re: [PATCH v4 2/5] watchdog: introduce arch_touch_nmi_watchdog()
2017-06-16 18:21 ` Andrew Morton
@ 2017-06-17 2:43 ` Nicholas Piggin
0 siblings, 0 replies; 13+ messages in thread
From: Nicholas Piggin @ 2017-06-17 2:43 UTC (permalink / raw)
To: Andrew Morton; +Cc: Don Zickus, Babu Moger, linux-kernel, linuxppc-dev
On Fri, 16 Jun 2017 11:21:17 -0700
Andrew Morton <akpm@linux-foundation.org> wrote:
> On Fri, 16 Jun 2017 16:57:12 +1000 Nicholas Piggin <npiggin@gmail.com> wrote:
>
> > For architectures that define HAVE_NMI_WATCHDOG, instead of having
> > them provide the complete touch_nmi_watchdog() function, just have
> > them provide arch_touch_nmi_watchdog().
> >
> > This gives the generic code more flexibility in implementing this
> > function, and arch implementations don't miss out on touching the
> > softlockup watchdog or other generic details.
> >
> > ...
> >
> > --- a/arch/blackfin/include/asm/nmi.h
> > +++ b/arch/blackfin/include/asm/nmi.h
> > @@ -9,4 +9,6 @@
> >
> > #include <linux/nmi.h>
> >
> > +extern void arch_touch_nmi_watchdog(void);
>
> Do we actually need to add this to the arch header files...
[snip]
> > +#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG)
> > +extern void arch_touch_nmi_watchdog(void);
> > +#else
> > +static inline void arch_touch_nmi_watchdog(void) {}
> > +#endif
> > +
>
> given that we have a global declaration here?
Probably not. I think it was a holdover from an earlier version where
I tried to let the arch declare it (one of the little embedded ones
had a comment somewhere saying it would be nice if they could make it
inline).
There was some difficulty with it, so yes let's remove these and do
that next time.
Thanks,
Nick
^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH v4 3/5] watchdog: split up config options
2017-06-16 6:57 [PATCH v4 0/5] Improve watchdog config for arch watchdogs Nicholas Piggin
2017-06-16 6:57 ` [PATCH v4 1/5] watchdog: remove unused declaration Nicholas Piggin
2017-06-16 6:57 ` [PATCH v4 2/5] watchdog: introduce arch_touch_nmi_watchdog() Nicholas Piggin
@ 2017-06-16 6:57 ` Nicholas Piggin
2017-06-17 12:35 ` Nicholas Piggin
2017-06-16 6:57 ` [PATCH v4 4/5] watchdog: provide watchdog_reconfigure() for arch watchdogs Nicholas Piggin
` (2 subsequent siblings)
5 siblings, 1 reply; 13+ messages in thread
From: Nicholas Piggin @ 2017-06-16 6:57 UTC (permalink / raw)
To: Don Zickus; +Cc: Nicholas Piggin, Babu Moger, linux-kernel, linuxppc-dev
Split SOFTLOCKUP_DETECTOR from LOCKUP_DETECTOR, and split
HARDLOCKUP_DETECTOR_PERF from HARDLOCKUP_DETECTOR.
LOCKUP_DETECTOR implies the general boot, sysctl, and programming
interfaces for the lockup detectors.
An architecture that wants to use a hard lockup detector must define
HAVE_HARDLOCKUP_DETECTOR_PERF or HAVE_HARDLOCKUP_DETECTOR_ARCH.
Alternatively an arch can define HAVE_NMI_WATCHDOG, which provides
the minimum arch_touch_nmi_watchdog, and it otherwise does its own
thing and does not implement the LOCKUP_DETECTOR interfaces.
sparc is unusual in that it has started to implement some of the
interfaces, but not fully yet. It should probably be converted to
a full HAVE_HARDLOCKUP_DETECTOR_ARCH.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/Kconfig | 23 ++++
arch/powerpc/Kconfig | 1 +
arch/powerpc/kernel/setup_64.c | 2 +-
arch/x86/Kconfig | 1 +
arch/x86/kernel/apic/hw_nmi.c | 2 +-
include/linux/nmi.h | 29 +++--
kernel/Makefile | 2 +-
kernel/sysctl.c | 31 +++---
kernel/watchdog.c | 243 +++++++++++++++++++++++++++--------------
kernel/watchdog_hld.c | 32 ------
lib/Kconfig.debug | 45 +++++---
11 files changed, 252 insertions(+), 159 deletions(-)
diff --git a/arch/Kconfig b/arch/Kconfig
index 6c00e5b00f8b..878addc6f141 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -288,6 +288,29 @@ config HAVE_PERF_EVENTS_NMI
subsystem. Also has support for calculating CPU cycle events
to determine how many clock cycles in a given period.
+
+config HAVE_HARDLOCKUP_DETECTOR_PERF
+ bool
+ depends on HAVE_PERF_EVENTS_NMI
+ help
+ The arch chooses to use the generic perf-NMI-based hardlockup
+ detector. Must define HAVE_PERF_EVENTS_NMI.
+
+config HAVE_NMI_WATCHDOG
+ bool
+ help
+ The arch provides a low level NMI watchdog. It provides
+ asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
+
+config HAVE_HARDLOCKUP_DETECTOR_ARCH
+ bool
+ select HAVE_NMI_WATCHDOG
+ help
+ The arch chooses to provide its own hardlockup detector, which is
+ a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
+ interfaces and parameters provided by hardlockup detector subsystem.
+
+
config HAVE_PERF_REGS
bool
help
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index f7c8f9972f61..7aba96ec3378 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -202,6 +202,7 @@ config PPC
select HAVE_OPTPROBES if PPC64
select HAVE_PERF_EVENTS
select HAVE_PERF_EVENTS_NMI if PPC64
+ select HAVE_HARDLOCKUP_DETECTOR_PERF if HAVE_PERF_EVENTS_NMI
select HAVE_PERF_REGS
select HAVE_PERF_USER_STACK_DUMP
select HAVE_RCU_TABLE_FREE if SMP
diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index f35ff9dea4fb..ab650905f75a 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -727,7 +727,7 @@ struct ppc_pci_io ppc_pci_io;
EXPORT_SYMBOL(ppc_pci_io);
#endif
-#ifdef CONFIG_HARDLOCKUP_DETECTOR
+#ifdef CONFIG_HARDLOCKUP_DETECTOR_PERF
u64 hw_nmi_get_sample_period(int watchdog_thresh)
{
return ppc_proc_freq * watchdog_thresh;
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 4ccfacc7232a..3c084149b5d1 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -157,6 +157,7 @@ config X86
select HAVE_PCSPKR_PLATFORM
select HAVE_PERF_EVENTS
select HAVE_PERF_EVENTS_NMI
+ select HAVE_HARDLOCKUP_DETECTOR_PERF if HAVE_PERF_EVENTS_NMI
select HAVE_PERF_REGS
select HAVE_PERF_USER_STACK_DUMP
select HAVE_REGS_AND_STACK_ACCESS_API
diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c
index c73c9fb281e1..d6f387780849 100644
--- a/arch/x86/kernel/apic/hw_nmi.c
+++ b/arch/x86/kernel/apic/hw_nmi.c
@@ -19,7 +19,7 @@
#include <linux/init.h>
#include <linux/delay.h>
-#ifdef CONFIG_HARDLOCKUP_DETECTOR
+#ifdef CONFIG_HARDLOCKUP_DETECTOR_PERF
u64 hw_nmi_get_sample_period(int watchdog_thresh)
{
return (u64)(cpu_khz) * 1000 * watchdog_thresh;
diff --git a/include/linux/nmi.h b/include/linux/nmi.h
index bd387ef8bccd..8aa01fd859fb 100644
--- a/include/linux/nmi.h
+++ b/include/linux/nmi.h
@@ -11,13 +11,21 @@
#endif
#ifdef CONFIG_LOCKUP_DETECTOR
+void lockup_detector_init(void);
+#else
+static inline void lockup_detector_init(void)
+{
+}
+#endif
+
+#ifdef CONFIG_SOFTLOCKUP_DETECTOR
extern void touch_softlockup_watchdog_sched(void);
extern void touch_softlockup_watchdog(void);
extern void touch_softlockup_watchdog_sync(void);
extern void touch_all_softlockup_watchdogs(void);
extern unsigned int softlockup_panic;
-extern unsigned int hardlockup_panic;
-void lockup_detector_init(void);
+extern int soft_watchdog_enabled;
+extern atomic_t watchdog_park_in_progress;
#else
static inline void touch_softlockup_watchdog_sched(void)
{
@@ -31,9 +39,6 @@ static inline void touch_softlockup_watchdog_sync(void)
static inline void touch_all_softlockup_watchdogs(void)
{
}
-static inline void lockup_detector_init(void)
-{
-}
#endif
#ifdef CONFIG_DETECT_HUNG_TASK
@@ -63,15 +68,18 @@ static inline void reset_hung_task_detector(void)
#if defined(CONFIG_HARDLOCKUP_DETECTOR)
extern void hardlockup_detector_disable(void);
+extern unsigned int hardlockup_panic;
#else
static inline void hardlockup_detector_disable(void) {}
#endif
-#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG)
+#if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF)
extern void arch_touch_nmi_watchdog(void);
#else
+#if !defined(CONFIG_HAVE_NMI_WATCHDOG)
static inline void arch_touch_nmi_watchdog(void) {}
#endif
+#endif
/**
* touch_nmi_watchdog - restart NMI watchdog timeout.
@@ -141,15 +149,18 @@ static inline bool trigger_single_cpu_backtrace(int cpu)
}
#endif
-#ifdef CONFIG_LOCKUP_DETECTOR
+#ifdef CONFIG_HARDLOCKUP_DETECTOR_PERF
u64 hw_nmi_get_sample_period(int watchdog_thresh);
+#endif
+
+#ifdef CONFIG_LOCKUP_DETECTOR
extern int nmi_watchdog_enabled;
-extern int soft_watchdog_enabled;
extern int watchdog_user_enabled;
extern int watchdog_thresh;
extern unsigned long watchdog_enabled;
+extern struct cpumask watchdog_cpumask;
extern unsigned long *watchdog_cpumask_bits;
-extern atomic_t watchdog_park_in_progress;
+extern int __read_mostly watchdog_suspended;
#ifdef CONFIG_SMP
extern int sysctl_softlockup_all_cpu_backtrace;
extern int sysctl_hardlockup_all_cpu_backtrace;
diff --git a/kernel/Makefile b/kernel/Makefile
index 72aa080f91f0..4cb8e8b23c6e 100644
--- a/kernel/Makefile
+++ b/kernel/Makefile
@@ -82,7 +82,7 @@ obj-$(CONFIG_KPROBES) += kprobes.o
obj-$(CONFIG_KGDB) += debug/
obj-$(CONFIG_DETECT_HUNG_TASK) += hung_task.o
obj-$(CONFIG_LOCKUP_DETECTOR) += watchdog.o
-obj-$(CONFIG_HARDLOCKUP_DETECTOR) += watchdog_hld.o
+obj-$(CONFIG_HARDLOCKUP_DETECTOR_PERF) += watchdog_hld.o
obj-$(CONFIG_SECCOMP) += seccomp.o
obj-$(CONFIG_RELAY) += relay.o
obj-$(CONFIG_SYSCTL) += utsname_sysctl.o
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index 4dfba1a76cc3..655686d546cb 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -880,6 +880,14 @@ static struct ctl_table kern_table[] = {
#endif
},
{
+ .procname = "watchdog_cpumask",
+ .data = &watchdog_cpumask_bits,
+ .maxlen = NR_CPUS,
+ .mode = 0644,
+ .proc_handler = proc_watchdog_cpumask,
+ },
+#ifdef CONFIG_SOFTLOCKUP_DETECTOR
+ {
.procname = "soft_watchdog",
.data = &soft_watchdog_enabled,
.maxlen = sizeof (int),
@@ -889,13 +897,6 @@ static struct ctl_table kern_table[] = {
.extra2 = &one,
},
{
- .procname = "watchdog_cpumask",
- .data = &watchdog_cpumask_bits,
- .maxlen = NR_CPUS,
- .mode = 0644,
- .proc_handler = proc_watchdog_cpumask,
- },
- {
.procname = "softlockup_panic",
.data = &softlockup_panic,
.maxlen = sizeof(int),
@@ -904,27 +905,29 @@ static struct ctl_table kern_table[] = {
.extra1 = &zero,
.extra2 = &one,
},
-#ifdef CONFIG_HARDLOCKUP_DETECTOR
+#ifdef CONFIG_SMP
{
- .procname = "hardlockup_panic",
- .data = &hardlockup_panic,
+ .procname = "softlockup_all_cpu_backtrace",
+ .data = &sysctl_softlockup_all_cpu_backtrace,
.maxlen = sizeof(int),
.mode = 0644,
.proc_handler = proc_dointvec_minmax,
.extra1 = &zero,
.extra2 = &one,
},
+#endif /* CONFIG_SMP */
#endif
-#ifdef CONFIG_SMP
+#ifdef CONFIG_HARDLOCKUP_DETECTOR
{
- .procname = "softlockup_all_cpu_backtrace",
- .data = &sysctl_softlockup_all_cpu_backtrace,
+ .procname = "hardlockup_panic",
+ .data = &hardlockup_panic,
.maxlen = sizeof(int),
.mode = 0644,
.proc_handler = proc_dointvec_minmax,
.extra1 = &zero,
.extra2 = &one,
},
+#ifdef CONFIG_SMP
{
.procname = "hardlockup_all_cpu_backtrace",
.data = &sysctl_hardlockup_all_cpu_backtrace,
@@ -936,6 +939,8 @@ static struct ctl_table kern_table[] = {
},
#endif /* CONFIG_SMP */
#endif
+#endif
+
#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
{
.procname = "unknown_nmi_panic",
diff --git a/kernel/watchdog.c b/kernel/watchdog.c
index 03e0b69bb5bf..1fba9c3d66dc 100644
--- a/kernel/watchdog.c
+++ b/kernel/watchdog.c
@@ -29,15 +29,58 @@
#include <linux/kvm_para.h>
#include <linux/kthread.h>
+/* Watchdog configuration */
static DEFINE_MUTEX(watchdog_proc_mutex);
-#if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR)
-unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED|NMI_WATCHDOG_ENABLED;
+int __read_mostly nmi_watchdog_enabled;
+
+#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG)
+unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED |
+ NMI_WATCHDOG_ENABLED;
#else
unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED;
#endif
-int __read_mostly nmi_watchdog_enabled;
+
+#ifdef CONFIG_HARDLOCKUP_DETECTOR
+/* boot commands */
+/*
+ * Should we panic when a soft-lockup or hard-lockup occurs:
+ */
+unsigned int __read_mostly hardlockup_panic =
+ CONFIG_BOOTPARAM_HARDLOCKUP_PANIC_VALUE;
+/*
+ * We may not want to enable hard lockup detection by default in all cases,
+ * for example when running the kernel as a guest on a hypervisor. In these
+ * cases this function can be called to disable hard lockup detection. This
+ * function should only be executed once by the boot processor before the
+ * kernel command line parameters are parsed, because otherwise it is not
+ * possible to override this in hardlockup_panic_setup().
+ */
+void hardlockup_detector_disable(void)
+{
+ watchdog_enabled &= ~NMI_WATCHDOG_ENABLED;
+}
+
+static int __init hardlockup_panic_setup(char *str)
+{
+ if (!strncmp(str, "panic", 5))
+ hardlockup_panic = 1;
+ else if (!strncmp(str, "nopanic", 7))
+ hardlockup_panic = 0;
+ else if (!strncmp(str, "0", 1))
+ watchdog_enabled &= ~NMI_WATCHDOG_ENABLED;
+ else if (!strncmp(str, "1", 1))
+ watchdog_enabled |= NMI_WATCHDOG_ENABLED;
+ return 1;
+}
+__setup("nmi_watchdog=", hardlockup_panic_setup);
+
+#endif
+
+#ifdef CONFIG_SOFTLOCKUP_DETECTOR
int __read_mostly soft_watchdog_enabled;
+#endif
+
int __read_mostly watchdog_user_enabled;
int __read_mostly watchdog_thresh = 10;
@@ -45,15 +88,9 @@ int __read_mostly watchdog_thresh = 10;
int __read_mostly sysctl_softlockup_all_cpu_backtrace;
int __read_mostly sysctl_hardlockup_all_cpu_backtrace;
#endif
-static struct cpumask watchdog_cpumask __read_mostly;
+struct cpumask watchdog_cpumask __read_mostly;
unsigned long *watchdog_cpumask_bits = cpumask_bits(&watchdog_cpumask);
-/* Helper for online, unparked cpus. */
-#define for_each_watchdog_cpu(cpu) \
- for_each_cpu_and((cpu), cpu_online_mask, &watchdog_cpumask)
-
-atomic_t watchdog_park_in_progress = ATOMIC_INIT(0);
-
/*
* The 'watchdog_running' variable is set to 1 when the watchdog threads
* are registered/started and is set to 0 when the watchdog threads are
@@ -72,7 +109,27 @@ static int __read_mostly watchdog_running;
* of 'watchdog_running' cannot change while the watchdog is deactivated
* temporarily (see related code in 'proc' handlers).
*/
-static int __read_mostly watchdog_suspended;
+int __read_mostly watchdog_suspended;
+
+/*
+ * These functions can be overridden if an architecture implements its
+ * own hardlockup detector.
+ */
+int __weak watchdog_nmi_enable(unsigned int cpu)
+{
+ return 0;
+}
+void __weak watchdog_nmi_disable(unsigned int cpu)
+{
+}
+
+#ifdef CONFIG_SOFTLOCKUP_DETECTOR
+
+/* Helper for online, unparked cpus. */
+#define for_each_watchdog_cpu(cpu) \
+ for_each_cpu_and((cpu), cpu_online_mask, &watchdog_cpumask)
+
+atomic_t watchdog_park_in_progress = ATOMIC_INIT(0);
static u64 __read_mostly sample_period;
@@ -120,6 +177,7 @@ static int __init softlockup_all_cpu_backtrace_setup(char *str)
return 1;
}
__setup("softlockup_all_cpu_backtrace=", softlockup_all_cpu_backtrace_setup);
+#ifdef CONFIG_HARDLOCKUP_DETECTOR
static int __init hardlockup_all_cpu_backtrace_setup(char *str)
{
sysctl_hardlockup_all_cpu_backtrace =
@@ -128,6 +186,7 @@ static int __init hardlockup_all_cpu_backtrace_setup(char *str)
}
__setup("hardlockup_all_cpu_backtrace=", hardlockup_all_cpu_backtrace_setup);
#endif
+#endif
/*
* Hard-lockup warnings should be triggered after just a few seconds. Soft-
@@ -213,18 +272,6 @@ void touch_softlockup_watchdog_sync(void)
__this_cpu_write(watchdog_touch_ts, 0);
}
-/* watchdog detector functions */
-bool is_hardlockup(void)
-{
- unsigned long hrint = __this_cpu_read(hrtimer_interrupts);
-
- if (__this_cpu_read(hrtimer_interrupts_saved) == hrint)
- return true;
-
- __this_cpu_write(hrtimer_interrupts_saved, hrint);
- return false;
-}
-
static int is_softlockup(unsigned long touch_ts)
{
unsigned long now = get_timestamp();
@@ -237,21 +284,21 @@ static int is_softlockup(unsigned long touch_ts)
return 0;
}
-static void watchdog_interrupt_count(void)
+/* watchdog detector functions */
+bool is_hardlockup(void)
{
- __this_cpu_inc(hrtimer_interrupts);
-}
+ unsigned long hrint = __this_cpu_read(hrtimer_interrupts);
-/*
- * These two functions are mostly architecture specific
- * defining them as weak here.
- */
-int __weak watchdog_nmi_enable(unsigned int cpu)
-{
- return 0;
+ if (__this_cpu_read(hrtimer_interrupts_saved) == hrint)
+ return true;
+
+ __this_cpu_write(hrtimer_interrupts_saved, hrint);
+ return false;
}
-void __weak watchdog_nmi_disable(unsigned int cpu)
+
+static void watchdog_interrupt_count(void)
{
+ __this_cpu_inc(hrtimer_interrupts);
}
static int watchdog_enable_all_cpus(void);
@@ -502,57 +549,6 @@ static void watchdog_unpark_threads(void)
kthread_unpark(per_cpu(softlockup_watchdog, cpu));
}
-/*
- * Suspend the hard and soft lockup detector by parking the watchdog threads.
- */
-int lockup_detector_suspend(void)
-{
- int ret = 0;
-
- get_online_cpus();
- mutex_lock(&watchdog_proc_mutex);
- /*
- * Multiple suspend requests can be active in parallel (counted by
- * the 'watchdog_suspended' variable). If the watchdog threads are
- * running, the first caller takes care that they will be parked.
- * The state of 'watchdog_running' cannot change while a suspend
- * request is active (see related code in 'proc' handlers).
- */
- if (watchdog_running && !watchdog_suspended)
- ret = watchdog_park_threads();
-
- if (ret == 0)
- watchdog_suspended++;
- else {
- watchdog_disable_all_cpus();
- pr_err("Failed to suspend lockup detectors, disabled\n");
- watchdog_enabled = 0;
- }
-
- mutex_unlock(&watchdog_proc_mutex);
-
- return ret;
-}
-
-/*
- * Resume the hard and soft lockup detector by unparking the watchdog threads.
- */
-void lockup_detector_resume(void)
-{
- mutex_lock(&watchdog_proc_mutex);
-
- watchdog_suspended--;
- /*
- * The watchdog threads are unparked if they were previously running
- * and if there is no more active suspend request.
- */
- if (watchdog_running && !watchdog_suspended)
- watchdog_unpark_threads();
-
- mutex_unlock(&watchdog_proc_mutex);
- put_online_cpus();
-}
-
static int update_watchdog_all_cpus(void)
{
int ret;
@@ -604,6 +600,81 @@ static void watchdog_disable_all_cpus(void)
}
}
+#else /* SOFTLOCKUP */
+static int watchdog_park_threads(void)
+{
+ return 0;
+}
+
+static void watchdog_unpark_threads(void)
+{
+}
+
+static int watchdog_enable_all_cpus(void)
+{
+ return 0;
+}
+
+static void watchdog_disable_all_cpus(void)
+{
+}
+
+static void set_sample_period(void)
+{
+}
+#endif /* SOFTLOCKUP */
+
+/*
+ * Suspend the hard and soft lockup detector by parking the watchdog threads.
+ */
+int lockup_detector_suspend(void)
+{
+ int ret = 0;
+
+ get_online_cpus();
+ mutex_lock(&watchdog_proc_mutex);
+ /*
+ * Multiple suspend requests can be active in parallel (counted by
+ * the 'watchdog_suspended' variable). If the watchdog threads are
+ * running, the first caller takes care that they will be parked.
+ * The state of 'watchdog_running' cannot change while a suspend
+ * request is active (see related code in 'proc' handlers).
+ */
+ if (watchdog_running && !watchdog_suspended)
+ ret = watchdog_park_threads();
+
+ if (ret == 0)
+ watchdog_suspended++;
+ else {
+ watchdog_disable_all_cpus();
+ pr_err("Failed to suspend lockup detectors, disabled\n");
+ watchdog_enabled = 0;
+ }
+
+ mutex_unlock(&watchdog_proc_mutex);
+
+ return ret;
+}
+
+/*
+ * Resume the hard and soft lockup detector by unparking the watchdog threads.
+ */
+void lockup_detector_resume(void)
+{
+ mutex_lock(&watchdog_proc_mutex);
+
+ watchdog_suspended--;
+ /*
+ * The watchdog threads are unparked if they were previously running
+ * and if there is no more active suspend request.
+ */
+ if (watchdog_running && !watchdog_suspended)
+ watchdog_unpark_threads();
+
+ mutex_unlock(&watchdog_proc_mutex);
+ put_online_cpus();
+}
+
#ifdef CONFIG_SYSCTL
/*
@@ -810,9 +881,11 @@ int proc_watchdog_cpumask(struct ctl_table *table, int write,
* a temporary cpumask, so we are likely not in a
* position to do much else to make things better.
*/
+#ifdef CONFIG_SOFTLOCKUP_DETECTOR
if (smpboot_update_cpumask_percpu_thread(
&watchdog_threads, &watchdog_cpumask) != 0)
pr_err("cpumask update failed\n");
+#endif
}
}
out:
diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c
index 90d688df6ce1..295a0d84934c 100644
--- a/kernel/watchdog_hld.c
+++ b/kernel/watchdog_hld.c
@@ -22,39 +22,7 @@ static DEFINE_PER_CPU(bool, hard_watchdog_warn);
static DEFINE_PER_CPU(bool, watchdog_nmi_touch);
static DEFINE_PER_CPU(struct perf_event *, watchdog_ev);
-/* boot commands */
-/*
- * Should we panic when a soft-lockup or hard-lockup occurs:
- */
-unsigned int __read_mostly hardlockup_panic =
- CONFIG_BOOTPARAM_HARDLOCKUP_PANIC_VALUE;
static unsigned long hardlockup_allcpu_dumped;
-/*
- * We may not want to enable hard lockup detection by default in all cases,
- * for example when running the kernel as a guest on a hypervisor. In these
- * cases this function can be called to disable hard lockup detection. This
- * function should only be executed once by the boot processor before the
- * kernel command line parameters are parsed, because otherwise it is not
- * possible to override this in hardlockup_panic_setup().
- */
-void hardlockup_detector_disable(void)
-{
- watchdog_enabled &= ~NMI_WATCHDOG_ENABLED;
-}
-
-static int __init hardlockup_panic_setup(char *str)
-{
- if (!strncmp(str, "panic", 5))
- hardlockup_panic = 1;
- else if (!strncmp(str, "nopanic", 7))
- hardlockup_panic = 0;
- else if (!strncmp(str, "0", 1))
- watchdog_enabled &= ~NMI_WATCHDOG_ENABLED;
- else if (!strncmp(str, "1", 1))
- watchdog_enabled |= NMI_WATCHDOG_ENABLED;
- return 1;
-}
-__setup("nmi_watchdog=", hardlockup_panic_setup);
void arch_touch_nmi_watchdog(void)
{
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index e4587ebe52c7..0f23f97ff6ba 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -778,34 +778,45 @@ config DEBUG_SHIRQ
menu "Debug Lockups and Hangs"
config LOCKUP_DETECTOR
- bool "Detect Hard and Soft Lockups"
+ bool
+
+config SOFTLOCKUP_DETECTOR
+ bool "Detect Soft Lockups"
depends on DEBUG_KERNEL && !S390
+ select LOCKUP_DETECTOR
help
Say Y here to enable the kernel to act as a watchdog to detect
- hard and soft lockups.
+ soft lockups.
Softlockups are bugs that cause the kernel to loop in kernel
mode for more than 20 seconds, without giving other tasks a
chance to run. The current stack trace is displayed upon
detection and the system will stay locked up.
+config HARDLOCKUP_DETECTOR_PERF
+ bool
+ select SOFTLOCKUP_DETECTOR
+
+#
+# arch/ can define HAVE_HARDLOCKUP_DETECTOR_ARCH to provide their own hard
+# lockup detector rather than the perf based detector.
+#
+config HARDLOCKUP_DETECTOR
+ bool "Detect Hard Lockups"
+ depends on DEBUG_KERNEL && !S390
+ depends on HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_ARCH
+ select LOCKUP_DETECTOR
+ select HARDLOCKUP_DETECTOR_PERF if HAVE_HARDLOCKUP_DETECTOR_PERF
+ select HARDLOCKUP_DETECTOR_ARCH if HAVE_HARDLOCKUP_DETECTOR_ARCH
+ help
+ Say Y here to enable the kernel to act as a watchdog to detect
+ hard lockups.
+
Hardlockups are bugs that cause the CPU to loop in kernel mode
for more than 10 seconds, without letting other interrupts have a
chance to run. The current stack trace is displayed upon detection
and the system will stay locked up.
- The overhead should be minimal. A periodic hrtimer runs to
- generate interrupts and kick the watchdog task every 4 seconds.
- An NMI is generated every 10 seconds or so to check for hardlockups.
-
- The frequency of hrtimer and NMI events and the soft and hard lockup
- thresholds can be controlled through the sysctl watchdog_thresh.
-
-config HARDLOCKUP_DETECTOR
- def_bool y
- depends on LOCKUP_DETECTOR && !HAVE_NMI_WATCHDOG
- depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
-
config BOOTPARAM_HARDLOCKUP_PANIC
bool "Panic (Reboot) On Hard Lockups"
depends on HARDLOCKUP_DETECTOR
@@ -826,7 +837,7 @@ config BOOTPARAM_HARDLOCKUP_PANIC_VALUE
config BOOTPARAM_SOFTLOCKUP_PANIC
bool "Panic (Reboot) On Soft Lockups"
- depends on LOCKUP_DETECTOR
+ depends on SOFTLOCKUP_DETECTOR
help
Say Y here to enable the kernel to panic on "soft lockups",
which are bugs that cause the kernel to loop in kernel
@@ -843,7 +854,7 @@ config BOOTPARAM_SOFTLOCKUP_PANIC
config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
int
- depends on LOCKUP_DETECTOR
+ depends on SOFTLOCKUP_DETECTOR
range 0 1
default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
@@ -851,7 +862,7 @@ config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
config DETECT_HUNG_TASK
bool "Detect Hung Tasks"
depends on DEBUG_KERNEL
- default LOCKUP_DETECTOR
+ default SOFTLOCKUP_DETECTOR
help
Say Y here to enable the kernel to detect "hung tasks",
which are bugs that cause the task to be stuck in
--
2.11.0
^ permalink raw reply related [flat|nested] 13+ messages in thread* Re: [PATCH v4 3/5] watchdog: split up config options
2017-06-16 6:57 ` [PATCH v4 3/5] watchdog: split up config options Nicholas Piggin
@ 2017-06-17 12:35 ` Nicholas Piggin
0 siblings, 0 replies; 13+ messages in thread
From: Nicholas Piggin @ 2017-06-17 12:35 UTC (permalink / raw)
To: Andrew Morton; +Cc: Don Zickus, Babu Moger, linux-kernel, linuxppc-dev
On Fri, 16 Jun 2017 16:57:13 +1000
Nicholas Piggin <npiggin@gmail.com> wrote:
> Split SOFTLOCKUP_DETECTOR from LOCKUP_DETECTOR, and split
> HARDLOCKUP_DETECTOR_PERF from HARDLOCKUP_DETECTOR.
>
> LOCKUP_DETECTOR implies the general boot, sysctl, and programming
> interfaces for the lockup detectors.
>
> An architecture that wants to use a hard lockup detector must define
> HAVE_HARDLOCKUP_DETECTOR_PERF or HAVE_HARDLOCKUP_DETECTOR_ARCH.
>
> Alternatively an arch can define HAVE_NMI_WATCHDOG, which provides
> the minimum arch_touch_nmi_watchdog, and it otherwise does its own
> thing and does not implement the LOCKUP_DETECTOR interfaces.
>
> sparc is unusual in that it has started to implement some of the
> interfaces, but not fully yet. It should probably be converted to
> a full HAVE_HARDLOCKUP_DETECTOR_ARCH.
>
> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
I haven't worked out the nm10300 Kconfig warning, but in the meantime
Andrew can you fold this patch in please?
Thanks,
Nick
---
arch/Kconfig | 6 +-----
1 file changed, 1 insertion(+), 5 deletions(-)
diff --git a/arch/Kconfig b/arch/Kconfig
index 878addc6f141..d1cd10f1c087 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -198,9 +198,6 @@ config HAVE_KPROBES_ON_FTRACE
config HAVE_NMI
bool
-config HAVE_NMI_WATCHDOG
- depends on HAVE_NMI
- bool
#
# An arch should select this if it provides all these things:
#
@@ -288,7 +285,6 @@ config HAVE_PERF_EVENTS_NMI
subsystem. Also has support for calculating CPU cycle events
to determine how many clock cycles in a given period.
-
config HAVE_HARDLOCKUP_DETECTOR_PERF
bool
depends on HAVE_PERF_EVENTS_NMI
@@ -297,6 +293,7 @@ config HAVE_HARDLOCKUP_DETECTOR_PERF
detector. Must define HAVE_PERF_EVENTS_NMI.
config HAVE_NMI_WATCHDOG
+ depends on HAVE_NMI
bool
help
The arch provides a low level NMI watchdog. It provides
@@ -310,7 +307,6 @@ config HAVE_HARDLOCKUP_DETECTOR_ARCH
a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
interfaces and parameters provided by hardlockup detector subsystem.
-
config HAVE_PERF_REGS
bool
help
--
2.11.0
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v4 4/5] watchdog: provide watchdog_reconfigure() for arch watchdogs
2017-06-16 6:57 [PATCH v4 0/5] Improve watchdog config for arch watchdogs Nicholas Piggin
` (2 preceding siblings ...)
2017-06-16 6:57 ` [PATCH v4 3/5] watchdog: split up config options Nicholas Piggin
@ 2017-06-16 6:57 ` Nicholas Piggin
2017-06-16 18:24 ` Andrew Morton
2017-06-16 6:57 ` [PATCH v4 5/5] powerpc/64s: implement arch-specific hardlockup watchdog Nicholas Piggin
2017-06-16 14:50 ` [PATCH v4 0/5] Improve watchdog config for arch watchdogs Don Zickus
5 siblings, 1 reply; 13+ messages in thread
From: Nicholas Piggin @ 2017-06-16 6:57 UTC (permalink / raw)
To: Don Zickus; +Cc: Nicholas Piggin, Babu Moger, linux-kernel, linuxppc-dev
After reconfiguring watchdog sysctls etc., architecture specific
watchdogs may not get all their parameters updated.
watchdog_reconfigure() can be implemented to pull the new values
in and set the arch NMI watchdog.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
kernel/watchdog.c | 29 +++++++++++++++++++++++++----
1 file changed, 25 insertions(+), 4 deletions(-)
diff --git a/kernel/watchdog.c b/kernel/watchdog.c
index 1fba9c3d66dc..06cd965f64d2 100644
--- a/kernel/watchdog.c
+++ b/kernel/watchdog.c
@@ -123,6 +123,11 @@ void __weak watchdog_nmi_disable(unsigned int cpu)
{
}
+void __weak watchdog_nmi_reconfigure(void)
+{
+}
+
+
#ifdef CONFIG_SOFTLOCKUP_DETECTOR
/* Helper for online, unparked cpus. */
@@ -600,6 +605,12 @@ static void watchdog_disable_all_cpus(void)
}
}
+static int watchdog_update_cpus(void)
+{
+ return smpboot_update_cpumask_percpu_thread(
+ &watchdog_threads, &watchdog_cpumask);
+}
+
#else /* SOFTLOCKUP */
static int watchdog_park_threads(void)
{
@@ -619,6 +630,11 @@ static void watchdog_disable_all_cpus(void)
{
}
+static int watchdog_update_cpus(void)
+{
+ return 0;
+}
+
static void set_sample_period(void)
{
}
@@ -651,6 +667,8 @@ int lockup_detector_suspend(void)
watchdog_enabled = 0;
}
+ watchdog_nmi_reconfigure();
+
mutex_unlock(&watchdog_proc_mutex);
return ret;
@@ -671,6 +689,8 @@ void lockup_detector_resume(void)
if (watchdog_running && !watchdog_suspended)
watchdog_unpark_threads();
+ watchdog_nmi_reconfigure();
+
mutex_unlock(&watchdog_proc_mutex);
put_online_cpus();
}
@@ -696,6 +716,8 @@ static int proc_watchdog_update(void)
else
watchdog_disable_all_cpus();
+ watchdog_nmi_reconfigure();
+
return err;
}
@@ -881,12 +903,11 @@ int proc_watchdog_cpumask(struct ctl_table *table, int write,
* a temporary cpumask, so we are likely not in a
* position to do much else to make things better.
*/
-#ifdef CONFIG_SOFTLOCKUP_DETECTOR
- if (smpboot_update_cpumask_percpu_thread(
- &watchdog_threads, &watchdog_cpumask) != 0)
+ if (watchdog_update_cpus() != 0)
pr_err("cpumask update failed\n");
-#endif
}
+
+ watchdog_nmi_reconfigure();
}
out:
mutex_unlock(&watchdog_proc_mutex);
--
2.11.0
^ permalink raw reply related [flat|nested] 13+ messages in thread* Re: [PATCH v4 4/5] watchdog: provide watchdog_reconfigure() for arch watchdogs
2017-06-16 6:57 ` [PATCH v4 4/5] watchdog: provide watchdog_reconfigure() for arch watchdogs Nicholas Piggin
@ 2017-06-16 18:24 ` Andrew Morton
2017-06-17 2:59 ` Nicholas Piggin
0 siblings, 1 reply; 13+ messages in thread
From: Andrew Morton @ 2017-06-16 18:24 UTC (permalink / raw)
To: Nicholas Piggin; +Cc: Don Zickus, Babu Moger, linux-kernel, linuxppc-dev
On Fri, 16 Jun 2017 16:57:14 +1000 Nicholas Piggin <npiggin@gmail.com> wrote:
> After reconfiguring watchdog sysctls etc., architecture specific
> watchdogs may not get all their parameters updated.
>
> watchdog_reconfigure() can be implemented to pull the new values
> in and set the arch NMI watchdog.
>
I'll update the title and changelog to say "watchdog_nmi_reconfigure".
> --- a/kernel/watchdog.c
> +++ b/kernel/watchdog.c
> @@ -123,6 +123,11 @@ void __weak watchdog_nmi_disable(unsigned int cpu)
> {
> }
>
> +void __weak watchdog_nmi_reconfigure(void)
> +{
> +}
Can we please get some documentation in here describing what it's for?
How arch maintainers might use this? When and why it is called, what
it must do? etc.
^ permalink raw reply [flat|nested] 13+ messages in thread* Re: [PATCH v4 4/5] watchdog: provide watchdog_reconfigure() for arch watchdogs
2017-06-16 18:24 ` Andrew Morton
@ 2017-06-17 2:59 ` Nicholas Piggin
0 siblings, 0 replies; 13+ messages in thread
From: Nicholas Piggin @ 2017-06-17 2:59 UTC (permalink / raw)
To: Andrew Morton; +Cc: Don Zickus, Babu Moger, linux-kernel, linuxppc-dev
On Fri, 16 Jun 2017 11:24:07 -0700
Andrew Morton <akpm@linux-foundation.org> wrote:
> On Fri, 16 Jun 2017 16:57:14 +1000 Nicholas Piggin <npiggin@gmail.com> wrote:
>
> > After reconfiguring watchdog sysctls etc., architecture specific
> > watchdogs may not get all their parameters updated.
> >
> > watchdog_reconfigure() can be implemented to pull the new values
> > in and set the arch NMI watchdog.
> >
>
> I'll update the title and changelog to say "watchdog_nmi_reconfigure".
Thanks.
> > --- a/kernel/watchdog.c
> > +++ b/kernel/watchdog.c
> > @@ -123,6 +123,11 @@ void __weak watchdog_nmi_disable(unsigned int cpu)
> > {
> > }
> >
> > +void __weak watchdog_nmi_reconfigure(void)
> > +{
> > +}
>
> Can we please get some documentation in here describing what it's for?
> How arch maintainers might use this? When and why it is called, what
> it must do? etc.
Good point, how's this?
---
kernel/watchdog.c | 15 +++++++++++++++
1 file changed, 15 insertions(+)
diff --git a/kernel/watchdog.c b/kernel/watchdog.c
index 06cd965f64d2..36531025496f 100644
--- a/kernel/watchdog.c
+++ b/kernel/watchdog.c
@@ -114,6 +114,10 @@ int __read_mostly watchdog_suspended;
/*
* These functions can be overridden if an architecture implements its
* own hardlockup detector.
+ *
+ * watchdog_nmi_enable/disable can be implemented to start and stop when
+ * softlockup watchdog threads start and stop. The arch must select the
+ * SOFTLOCKUP_DETECTOR Kconfig.
*/
int __weak watchdog_nmi_enable(unsigned int cpu)
{
@@ -123,6 +127,17 @@ void __weak watchdog_nmi_disable(unsigned int cpu)
{
}
+/*
+ * watchdog_nmi_reconfigure can be implemented to be notified after any
+ * watchdog configuration change. The arch hardlockup watchdog should
+ * respond to the following variables:
+ * - nmi_watchdog_enabled
+ * - watchdog_thresh
+ * - watchdog_cpumask
+ * - sysctl_hardlockup_all_cpu_backtrace
+ * - hardlockup_panic
+ * - watchdog_suspended
+ */
void __weak watchdog_nmi_reconfigure(void)
{
}
--
2.11.0
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v4 5/5] powerpc/64s: implement arch-specific hardlockup watchdog
2017-06-16 6:57 [PATCH v4 0/5] Improve watchdog config for arch watchdogs Nicholas Piggin
` (3 preceding siblings ...)
2017-06-16 6:57 ` [PATCH v4 4/5] watchdog: provide watchdog_reconfigure() for arch watchdogs Nicholas Piggin
@ 2017-06-16 6:57 ` Nicholas Piggin
2017-06-16 14:50 ` [PATCH v4 0/5] Improve watchdog config for arch watchdogs Don Zickus
5 siblings, 0 replies; 13+ messages in thread
From: Nicholas Piggin @ 2017-06-16 6:57 UTC (permalink / raw)
To: Don Zickus; +Cc: Nicholas Piggin, Babu Moger, linux-kernel, linuxppc-dev
Implement an arch-speicfic watchdog rather than use the perf-based
hardlockup detector.
The new watchdog takes the soft-NMI directly, rather than going
through perf. Perf interrupts are to be made maskable in future, so
that would prevent the perf detector from working in those regions.
Additionally, implement a SMP based detector where all CPUs watch one
another by pinging a shared cpumask. This is because powerpc Book3S
does not have a true periodic local NMI, but some platforms do
implement a true NMI IPI.
If a CPU is stuck with interrupts hard disabled, the soft-NMI watchdog
does not work, but the SMP watchdog will. Even on platforms without a
true NMI IPI to get a good trace from the stuck CPU, other CPUs will
notice the lockup sufficiently to report it and panic.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
arch/powerpc/Kconfig | 7 +-
arch/powerpc/include/asm/nmi.h | 11 +
arch/powerpc/include/asm/smp.h | 2 +
arch/powerpc/kernel/Makefile | 1 +
arch/powerpc/kernel/exceptions-64s.S | 30 ++-
arch/powerpc/kernel/kvm.c | 7 +
arch/powerpc/kernel/setup_64.c | 19 --
arch/powerpc/kernel/smp.c | 20 +-
arch/powerpc/kernel/watchdog.c | 388 +++++++++++++++++++++++++++++++++++
9 files changed, 460 insertions(+), 25 deletions(-)
create mode 100644 arch/powerpc/kernel/watchdog.c
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index 7aba96ec3378..6b6ff76b18de 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -82,7 +82,7 @@ config NR_IRQS
config NMI_IPI
bool
- depends on SMP && (DEBUGGER || KEXEC_CORE)
+ depends on SMP && (DEBUGGER || KEXEC_CORE || HARDLOCKUP_DETECTOR)
default y
config STACKTRACE_SUPPORT
@@ -197,12 +197,13 @@ config PPC
select HAVE_MEMBLOCK
select HAVE_MEMBLOCK_NODE_MAP
select HAVE_MOD_ARCH_SPECIFIC
- select HAVE_NMI if PERF_EVENTS
+ select HAVE_NMI if PERF_EVENTS || (PPC64 && PPC_BOOK3S)
+ select HAVE_HARDLOCKUP_DETECTOR_ARCH if (PPC64 && PPC_BOOK3S)
select HAVE_OPROFILE
select HAVE_OPTPROBES if PPC64
select HAVE_PERF_EVENTS
select HAVE_PERF_EVENTS_NMI if PPC64
- select HAVE_HARDLOCKUP_DETECTOR_PERF if HAVE_PERF_EVENTS_NMI
+ select HAVE_HARDLOCKUP_DETECTOR_PERF if HAVE_PERF_EVENTS_NMI && !HAVE_HARDLOCKUP_DETECTOR_ARCH
select HAVE_PERF_REGS
select HAVE_PERF_USER_STACK_DUMP
select HAVE_RCU_TABLE_FREE if SMP
diff --git a/arch/powerpc/include/asm/nmi.h b/arch/powerpc/include/asm/nmi.h
index ff1ccb375e60..6f8e79cd35d8 100644
--- a/arch/powerpc/include/asm/nmi.h
+++ b/arch/powerpc/include/asm/nmi.h
@@ -1,4 +1,15 @@
#ifndef _ASM_NMI_H
#define _ASM_NMI_H
+#ifdef CONFIG_HARDLOCKUP_DETECTOR
+extern void arch_touch_nmi_watchdog(void);
+
+extern void arch_trigger_cpumask_backtrace(const cpumask_t *mask,
+ bool exclude_self);
+#define arch_trigger_cpumask_backtrace arch_trigger_cpumask_backtrace
+
+#else
+static inline void arch_touch_nmi_watchdog(void) {}
+#endif
+
#endif /* _ASM_NMI_H */
diff --git a/arch/powerpc/include/asm/smp.h b/arch/powerpc/include/asm/smp.h
index ebddb2111d87..8ea98504f900 100644
--- a/arch/powerpc/include/asm/smp.h
+++ b/arch/powerpc/include/asm/smp.h
@@ -55,6 +55,8 @@ struct smp_ops_t {
int (*cpu_bootable)(unsigned int nr);
};
+extern void smp_flush_nmi_ipi(u64 delay_us);
+extern int smp_send_nmi_ipi(int cpu, void (*fn)(struct pt_regs *), u64 delay_us);
extern void smp_send_debugger_break(void);
extern void start_secondary_resume(void);
extern void smp_generic_give_timebase(void);
diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
index e132902e1f14..459c17bc5cf3 100644
--- a/arch/powerpc/kernel/Makefile
+++ b/arch/powerpc/kernel/Makefile
@@ -40,6 +40,7 @@ obj-$(CONFIG_PPC64) += setup_64.o sys_ppc32.o \
signal_64.o ptrace32.o \
paca.o nvram_64.o firmware.o
obj-$(CONFIG_VDSO32) += vdso32/
+obj-$(CONFIG_HARDLOCKUP_DETECTOR) += watchdog.o
obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o
obj-$(CONFIG_PPC_BOOK3S_64) += cpu_setup_ppc970.o cpu_setup_pa6t.o
obj-$(CONFIG_PPC_BOOK3S_64) += cpu_setup_power.o
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index ae418b85c17c..bde11f0dc69e 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -1273,6 +1273,31 @@ EXC_REAL_NONE(0x1800, 0x100)
EXC_VIRT_NONE(0x5800, 0x100)
#endif
+#if defined(CONFIG_HARDLOCKUP_DETECTOR) && defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_ARCH)
+
+#define MASKED_DEC_HANDLER_LABEL 3f
+
+#define MASKED_DEC_HANDLER(_H) \
+3: /* soft-nmi */ \
+ std r12,PACA_EXGEN+EX_R12(r13); \
+ GET_SCRATCH0(r10); \
+ std r10,PACA_EXGEN+EX_R13(r13); \
+ EXCEPTION_PROLOG_PSERIES_1(soft_nmi_common, _H)
+
+EXC_COMMON_BEGIN(soft_nmi_common)
+ mr r10,r1
+ ld r1,PACAEMERGSP(r13)
+ ld r1,PACA_NMI_EMERG_SP(r13)
+ subi r1,r1,INT_FRAME_SIZE
+ EXCEPTION_COMMON_NORET_STACK(PACA_EXGEN, 0x900,
+ system_reset, soft_nmi_interrupt,
+ ADD_NVGPRS;ADD_RECONCILE)
+ b ret_from_except
+
+#else
+#define MASKED_DEC_HANDLER_LABEL 2f /* normal return */
+#define MASKED_DEC_HANDLER(_H)
+#endif
/*
* An interrupt came in while soft-disabled. We set paca->irq_happened, then:
@@ -1295,7 +1320,7 @@ masked_##_H##interrupt: \
lis r10,0x7fff; \
ori r10,r10,0xffff; \
mtspr SPRN_DEC,r10; \
- b 2f; \
+ b MASKED_DEC_HANDLER_LABEL; \
1: cmpwi r10,PACA_IRQ_DBELL; \
beq 2f; \
cmpwi r10,PACA_IRQ_HMI; \
@@ -1310,7 +1335,8 @@ masked_##_H##interrupt: \
ld r11,PACA_EXGEN+EX_R11(r13); \
GET_SCRATCH0(r13); \
##_H##rfid; \
- b .
+ b .; \
+ MASKED_DEC_HANDLER(_H)
/*
* Real mode exceptions actually use this too, but alternate
diff --git a/arch/powerpc/kernel/kvm.c b/arch/powerpc/kernel/kvm.c
index 9ad37f827a97..1086ea37c832 100644
--- a/arch/powerpc/kernel/kvm.c
+++ b/arch/powerpc/kernel/kvm.c
@@ -25,6 +25,7 @@
#include <linux/kvm_para.h>
#include <linux/slab.h>
#include <linux/of.h>
+#include <linux/nmi.h> /* hardlockup_detector_disable() */
#include <asm/reg.h>
#include <asm/sections.h>
@@ -718,6 +719,12 @@ static __init void kvm_free_tmp(void)
static int __init kvm_guest_init(void)
{
+ /*
+ * The hardlockup detector is likely to get false positives in
+ * KVM guests, so disable it by default.
+ */
+ hardlockup_detector_disable();
+
if (!kvm_para_available())
goto free_tmp;
diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index ab650905f75a..9c15b3b8ba4d 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -726,22 +726,3 @@ unsigned long memory_block_size_bytes(void)
struct ppc_pci_io ppc_pci_io;
EXPORT_SYMBOL(ppc_pci_io);
#endif
-
-#ifdef CONFIG_HARDLOCKUP_DETECTOR_PERF
-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)
-{
- hardlockup_detector_disable();
-
- return 0;
-}
-early_initcall(disable_hardlockup_detector);
-#endif
diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
index df2a41647d8e..5b602a284084 100644
--- a/arch/powerpc/kernel/smp.c
+++ b/arch/powerpc/kernel/smp.c
@@ -433,13 +433,31 @@ static void do_smp_send_nmi_ipi(int cpu)
}
}
+void smp_flush_nmi_ipi(u64 delay_us)
+{
+ unsigned long flags;
+
+ nmi_ipi_lock_start(&flags);
+ while (nmi_ipi_busy_count) {
+ nmi_ipi_unlock_end(&flags);
+ udelay(1);
+ if (delay_us) {
+ delay_us--;
+ if (!delay_us)
+ return;
+ }
+ nmi_ipi_lock_start(&flags);
+ }
+ nmi_ipi_unlock_end(&flags);
+}
+
/*
* - cpu is the target CPU (must not be this CPU), or NMI_IPI_ALL_OTHERS.
* - fn is the target callback function.
* - delay_us > 0 is the delay before giving up waiting for targets to
* enter the handler, == 0 specifies indefinite delay.
*/
-static int smp_send_nmi_ipi(int cpu, void (*fn)(struct pt_regs *), u64 delay_us)
+int smp_send_nmi_ipi(int cpu, void (*fn)(struct pt_regs *), u64 delay_us)
{
unsigned long flags;
int me = raw_smp_processor_id();
diff --git a/arch/powerpc/kernel/watchdog.c b/arch/powerpc/kernel/watchdog.c
new file mode 100644
index 000000000000..4ed9cf37f369
--- /dev/null
+++ b/arch/powerpc/kernel/watchdog.c
@@ -0,0 +1,388 @@
+/*
+ * Watchdog support on powerpc systems.
+ *
+ * Copyright 2017, IBM Corporation.
+ *
+ * This uses code from arch/sparc/kernel/nmi.c and kernel/watchdog.c
+ */
+#include <linux/kernel.h>
+#include <linux/param.h>
+#include <linux/init.h>
+#include <linux/percpu.h>
+#include <linux/cpu.h>
+#include <linux/nmi.h>
+#include <linux/module.h>
+#include <linux/export.h>
+#include <linux/kprobes.h>
+#include <linux/hardirq.h>
+#include <linux/reboot.h>
+#include <linux/slab.h>
+#include <linux/kdebug.h>
+#include <linux/sched/debug.h>
+#include <linux/delay.h>
+#include <linux/smp.h>
+
+#include <asm/paca.h>
+
+/*
+ * The watchdog has a simple timer that runs on each CPU, once per timer
+ * period. This is the heartbeat.
+ *
+ * Then there are checks to see if the heartbeat has not triggered on a CPU
+ * for the panic timeout period. Currently the watchdog only supports an
+ * SMP check, so the heartbeat only turns on when we have 2 or more CPUs.
+ *
+ * This is not an NMI watchdog, but Linux uses that name for a generic
+ * watchdog in some cases, so NMI gets used in some places.
+ */
+
+static cpumask_t wd_cpus_enabled __read_mostly;
+
+static u64 wd_panic_timeout_tb __read_mostly; /* timebase ticks until panic */
+static u64 wd_smp_panic_timeout_tb __read_mostly; /* panic other CPUs */
+
+static u64 wd_timer_period_ms __read_mostly; /* interval between heartbeat */
+
+static DEFINE_PER_CPU(struct timer_list, wd_timer);
+static DEFINE_PER_CPU(u64, wd_timer_tb);
+
+/*
+ * These are for the SMP checker. CPUs clear their pending bit in their
+ * heartbeat. If the bitmask becomes empty, the time is noted and the
+ * bitmask is refilled.
+ *
+ * All CPUs clear their bit in the pending mask every timer period.
+ * Once all have cleared, the time is noted and the bits are reset.
+ * If the time since all clear was greater than the panic timeout,
+ * we can panic with the list of stuck CPUs.
+ *
+ * This will work best with NMI IPIs for crash code so the stuck CPUs
+ * can be pulled out to get their backtraces.
+ */
+static unsigned long __wd_smp_lock = 0;
+static cpumask_t wd_smp_cpus_pending;
+static cpumask_t wd_smp_cpus_stuck;
+static u64 wd_smp_last_reset_tb;
+
+static inline void wd_smp_lock(unsigned long *flags)
+{
+ /*
+ * Avoid locking layers if possible.
+ * This may be called from low level interrupt handlers at some
+ * point in future.
+ */
+ local_irq_save(*flags);
+ while (unlikely(test_and_set_bit_lock(0, &__wd_smp_lock)))
+ cpu_relax();
+}
+
+static inline void wd_smp_unlock(unsigned long *flags)
+{
+ clear_bit_unlock(0, &__wd_smp_lock);
+ local_irq_restore(*flags);
+}
+
+static void wd_lockup_ipi(struct pt_regs *regs)
+{
+ pr_emerg("Watchdog CPU:%d Hard LOCKUP\n", raw_smp_processor_id());
+ print_modules();
+ print_irqtrace_events(current);
+ if (regs)
+ show_regs(regs);
+ else
+ dump_stack();
+
+ if (hardlockup_panic)
+ nmi_panic(regs, "Hard LOCKUP");
+}
+
+static void set_cpu_stuck(int cpu, u64 tb)
+{
+ cpumask_set_cpu(cpu, &wd_smp_cpus_stuck);
+ cpumask_clear_cpu(cpu, &wd_smp_cpus_pending);
+ if (cpumask_empty(&wd_smp_cpus_pending)) {
+ wd_smp_last_reset_tb = tb;
+ cpumask_andnot(&wd_smp_cpus_pending,
+ &wd_cpus_enabled,
+ &wd_smp_cpus_stuck);
+ }
+}
+
+static void watchdog_smp_panic(int cpu, u64 tb)
+{
+ unsigned long flags;
+ int c;
+
+ wd_smp_lock(&flags);
+ /* Double check some things under lock */
+ if ((s64)(tb - wd_smp_last_reset_tb) < (s64)wd_smp_panic_timeout_tb)
+ goto out;
+ if (cpumask_test_cpu(cpu, &wd_smp_cpus_pending))
+ goto out;
+ if (cpumask_weight(&wd_smp_cpus_pending) == 0)
+ goto out;
+
+ pr_emerg("Watchdog CPU:%d detected Hard LOCKUP other CPUS:%*pbl\n",
+ cpu, cpumask_pr_args(&wd_smp_cpus_pending));
+
+ /*
+ * Try to trigger the stuck CPUs.
+ */
+ for_each_cpu(c, &wd_smp_cpus_pending) {
+ if (c == cpu)
+ continue;
+ smp_send_nmi_ipi(c, wd_lockup_ipi, 1000000);
+ }
+ smp_flush_nmi_ipi(1000000);
+
+ /* Take the stuck CPU out of the watch group */
+ for_each_cpu(c, &wd_smp_cpus_pending)
+ set_cpu_stuck(c, tb);
+
+out:
+ wd_smp_unlock(&flags);
+
+ printk_safe_flush();
+ /*
+ * printk_safe_flush() seems to require another print
+ * before anything actually goes out to console.
+ */
+ if (sysctl_hardlockup_all_cpu_backtrace)
+ trigger_allbutself_cpu_backtrace();
+
+ if (hardlockup_panic)
+ nmi_panic(NULL, "Hard LOCKUP");
+}
+
+static void wd_smp_clear_cpu_pending(int cpu, u64 tb)
+{
+ if (!cpumask_test_cpu(cpu, &wd_smp_cpus_pending)) {
+ if (unlikely(cpumask_test_cpu(cpu, &wd_smp_cpus_stuck))) {
+ unsigned long flags;
+
+ pr_emerg("Watchdog CPU:%d became unstuck\n", cpu);
+ wd_smp_lock(&flags);
+ cpumask_clear_cpu(cpu, &wd_smp_cpus_stuck);
+ wd_smp_unlock(&flags);
+ }
+ return;
+ }
+ cpumask_clear_cpu(cpu, &wd_smp_cpus_pending);
+ if (cpumask_empty(&wd_smp_cpus_pending)) {
+ unsigned long flags;
+
+ wd_smp_lock(&flags);
+ if (cpumask_empty(&wd_smp_cpus_pending)) {
+ wd_smp_last_reset_tb = tb;
+ cpumask_andnot(&wd_smp_cpus_pending,
+ &wd_cpus_enabled,
+ &wd_smp_cpus_stuck);
+ }
+ wd_smp_unlock(&flags);
+ }
+}
+
+static void watchdog_timer_interrupt(int cpu)
+{
+ u64 tb = get_tb();
+
+ per_cpu(wd_timer_tb, cpu) = tb;
+
+ wd_smp_clear_cpu_pending(cpu, tb);
+
+ if ((s64)(tb - wd_smp_last_reset_tb) >= (s64)wd_smp_panic_timeout_tb)
+ watchdog_smp_panic(cpu, tb);
+}
+
+void soft_nmi_interrupt(struct pt_regs *regs)
+{
+ unsigned long flags;
+ int cpu = raw_smp_processor_id();
+ u64 tb;
+
+ if (!cpumask_test_cpu(cpu, &wd_cpus_enabled))
+ return;
+
+ nmi_enter();
+ tb = get_tb();
+ if (tb - per_cpu(wd_timer_tb, cpu) >= wd_panic_timeout_tb) {
+ per_cpu(wd_timer_tb, cpu) = tb;
+
+ wd_smp_lock(&flags);
+ if (cpumask_test_cpu(cpu, &wd_smp_cpus_stuck)) {
+ wd_smp_unlock(&flags);
+ goto out;
+ }
+ set_cpu_stuck(cpu, tb);
+
+ pr_emerg("Watchdog CPU:%d Hard LOCKUP\n", cpu);
+ print_modules();
+ print_irqtrace_events(current);
+ if (regs)
+ show_regs(regs);
+ else
+ dump_stack();
+
+ wd_smp_unlock(&flags);
+
+ if (sysctl_hardlockup_all_cpu_backtrace)
+ trigger_allbutself_cpu_backtrace();
+
+ if (hardlockup_panic)
+ nmi_panic(regs, "Hard LOCKUP");
+ }
+ if (wd_panic_timeout_tb < 0x7fffffff)
+ mtspr(SPRN_DEC, wd_panic_timeout_tb);
+
+out:
+ nmi_exit();
+}
+
+static void wd_timer_reset(unsigned int cpu, struct timer_list *t)
+{
+ t->expires = jiffies + msecs_to_jiffies(wd_timer_period_ms);
+ if (wd_timer_period_ms > 1000)
+ t->expires = __round_jiffies_up(t->expires, cpu);
+ add_timer_on(t, cpu);
+}
+
+static void wd_timer_fn(unsigned long data)
+{
+ struct timer_list *t = this_cpu_ptr(&wd_timer);
+ int cpu = smp_processor_id();
+
+ watchdog_timer_interrupt(cpu);
+
+ wd_timer_reset(cpu, t);
+}
+
+void arch_touch_nmi_watchdog(void)
+{
+ int cpu = smp_processor_id();
+
+ watchdog_timer_interrupt(cpu);
+}
+EXPORT_SYMBOL(arch_touch_nmi_watchdog);
+
+static void start_watchdog_timer_on(unsigned int cpu)
+{
+ struct timer_list *t = per_cpu_ptr(&wd_timer, cpu);
+
+ per_cpu(wd_timer_tb, cpu) = get_tb();
+
+ setup_pinned_timer(t, wd_timer_fn, 0);
+ wd_timer_reset(cpu, t);
+}
+
+static void stop_watchdog_timer_on(unsigned int cpu)
+{
+ struct timer_list *t = per_cpu_ptr(&wd_timer, cpu);
+
+ del_timer_sync(t);
+}
+
+static int start_wd_on_cpu(unsigned int cpu)
+{
+ if (cpumask_test_cpu(cpu, &wd_cpus_enabled)) {
+ WARN_ON(1);
+ return 0;
+ }
+
+ if (!cpumask_test_cpu(cpu, &watchdog_cpumask))
+ return 0;
+
+ cpumask_set_cpu(cpu, &wd_cpus_enabled);
+ if (cpumask_weight(&wd_cpus_enabled) == 1) {
+ cpumask_set_cpu(cpu, &wd_smp_cpus_pending);
+ wd_smp_last_reset_tb = get_tb();
+ }
+ smp_wmb();
+ start_watchdog_timer_on(cpu);
+
+ return 0;
+}
+
+static int stop_wd_on_cpu(unsigned int cpu)
+{
+ if (!cpumask_test_cpu(cpu, &wd_cpus_enabled)) {
+ WARN_ON(1);
+ return 0;
+ }
+
+ stop_watchdog_timer_on(cpu);
+
+ cpumask_clear_cpu(cpu, &wd_cpus_enabled);
+ wd_smp_clear_cpu_pending(cpu, get_tb());
+
+ return 0;
+}
+
+static void watchdog_calc_timeouts(void)
+{
+ wd_panic_timeout_tb = watchdog_thresh * ppc_tb_freq;
+
+ /* Have the SMP detector trigger a bit later */
+ wd_smp_panic_timeout_tb = wd_panic_timeout_tb * 3 / 2;
+
+ /* 2/5 is the factor that the perf based detector uses */
+ wd_timer_period_ms = watchdog_thresh * 1000 * 2 / 5;
+}
+
+void watchdog_nmi_reconfigure(void)
+{
+ int cpu;
+
+ watchdog_calc_timeouts();
+
+ for_each_cpu(cpu, &wd_cpus_enabled)
+ stop_wd_on_cpu(cpu);
+
+ if (!(watchdog_enabled & NMI_WATCHDOG_ENABLED))
+ return;
+
+ if (watchdog_suspended)
+ return;
+
+ for_each_cpu_and(cpu, cpu_online_mask, &watchdog_cpumask)
+ start_wd_on_cpu(cpu);
+}
+
+/*
+ * This runs after lockup_detector_init() which sets up watchdog_cpumask.
+ */
+static int __init powerpc_watchdog_init(void)
+{
+ int err;
+
+ watchdog_calc_timeouts();
+
+ err = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "powerpc/watchdog:online",
+ start_wd_on_cpu, stop_wd_on_cpu);
+ if (err < 0)
+ pr_warning("Watchdog could not be initialized");
+
+ return 0;
+}
+arch_initcall(powerpc_watchdog_init);
+
+static void handle_backtrace_ipi(struct pt_regs *regs)
+{
+ nmi_cpu_backtrace(regs);
+}
+
+static void raise_backtrace_ipi(cpumask_t *mask)
+{
+ unsigned int cpu;
+
+ for_each_cpu(cpu, mask) {
+ if (cpu == smp_processor_id())
+ handle_backtrace_ipi(NULL);
+ else
+ smp_send_nmi_ipi(cpu, handle_backtrace_ipi, 1000000);
+ }
+}
+
+void arch_trigger_cpumask_backtrace(const cpumask_t *mask, bool exclude_self)
+{
+ nmi_trigger_cpumask_backtrace(mask, exclude_self, raise_backtrace_ipi);
+}
--
2.11.0
^ permalink raw reply related [flat|nested] 13+ messages in thread* Re: [PATCH v4 0/5] Improve watchdog config for arch watchdogs
2017-06-16 6:57 [PATCH v4 0/5] Improve watchdog config for arch watchdogs Nicholas Piggin
` (4 preceding siblings ...)
2017-06-16 6:57 ` [PATCH v4 5/5] powerpc/64s: implement arch-specific hardlockup watchdog Nicholas Piggin
@ 2017-06-16 14:50 ` Don Zickus
2017-06-16 17:16 ` Babu Moger
5 siblings, 1 reply; 13+ messages in thread
From: Don Zickus @ 2017-06-16 14:50 UTC (permalink / raw)
To: Nicholas Piggin, Andrew Morton; +Cc: Babu Moger, linux-kernel, linuxppc-dev
(adding Andrew)
On Fri, Jun 16, 2017 at 04:57:10PM +1000, Nicholas Piggin wrote:
> This is the latest series to make the hardlockup watchdog more
> easily replaceable by arch code. The last patch provides some
> justification for why we want to do this (existing sparc watchdog
> is another that could benefit).
>
> Since last time, we tidied up the Kconfig approach so it was not
> so jumbled -- Thanks Don and Babu for help and suggestions there.
Thanks Nick! I think this version is good enough for now.
>
> Since last posting the powerpc patch, I re-added the soft-NMI
> watchdog, but have it using the emergency stack so we can always
> avoid the process stacks when perf interrupts are soft-masked.
>
> I'm not sure of the best strategy to merge this. The powerpc
> watchdog is too much to carry anywhere but powerpc tree now, and
> watchdog patches seem to go via Andrew...
Andrew, suggestions here?
Reviewed-by: Don Zickus <dzickus@redhat.com>
Acked-by: Don Zickus <dzickus@redhat.com>
>
> Thanks,
> Nick
>
> Nicholas Piggin (5):
> watchdog: remove unused declaration
> watchdog: introduce arch_touch_nmi_watchdog()
> watchdog: split up config options
> watchdog: provide watchdog_reconfigure() for arch watchdogs
> powerpc/64s: implement arch-specific hardlockup watchdog
>
> arch/Kconfig | 23 ++
> arch/blackfin/include/asm/nmi.h | 2 +
> arch/blackfin/kernel/nmi.c | 2 +-
> arch/mn10300/include/asm/nmi.h | 2 +
> arch/mn10300/kernel/mn10300-watchdog-low.S | 8 +-
> arch/mn10300/kernel/mn10300-watchdog.c | 2 +-
> arch/powerpc/Kconfig | 6 +-
> arch/powerpc/include/asm/nmi.h | 11 +
> arch/powerpc/include/asm/smp.h | 2 +
> arch/powerpc/kernel/Makefile | 1 +
> arch/powerpc/kernel/exceptions-64s.S | 30 ++-
> arch/powerpc/kernel/kvm.c | 7 +
> arch/powerpc/kernel/setup_64.c | 19 --
> arch/powerpc/kernel/smp.c | 20 +-
> arch/powerpc/kernel/watchdog.c | 388 +++++++++++++++++++++++++++++
> arch/sparc/include/asm/nmi.h | 1 +
> arch/sparc/kernel/nmi.c | 6 +-
> arch/x86/Kconfig | 1 +
> arch/x86/kernel/apic/hw_nmi.c | 2 +-
> include/linux/nmi.h | 57 +++--
> kernel/Makefile | 2 +-
> kernel/sysctl.c | 31 ++-
> kernel/watchdog.c | 268 +++++++++++++-------
> kernel/watchdog_hld.c | 37 +--
> lib/Kconfig.debug | 45 ++--
> 25 files changed, 763 insertions(+), 210 deletions(-)
> create mode 100644 arch/powerpc/kernel/watchdog.c
>
> --
> 2.11.0
>
^ permalink raw reply [flat|nested] 13+ messages in thread* Re: [PATCH v4 0/5] Improve watchdog config for arch watchdogs
2017-06-16 14:50 ` [PATCH v4 0/5] Improve watchdog config for arch watchdogs Don Zickus
@ 2017-06-16 17:16 ` Babu Moger
0 siblings, 0 replies; 13+ messages in thread
From: Babu Moger @ 2017-06-16 17:16 UTC (permalink / raw)
To: Don Zickus, Nicholas Piggin, Andrew Morton; +Cc: linux-kernel, linuxppc-dev
Tested on sparc:
Tested-by: Babu Moger <babu.moger@oracle.com>
Reviewed patch #1, #2, #3
Reviewed-by: Babu Moger <babu.moger@oracle.com>
On 6/16/2017 9:50 AM, Don Zickus wrote:
> (adding Andrew)
>
> On Fri, Jun 16, 2017 at 04:57:10PM +1000, Nicholas Piggin wrote:
>> This is the latest series to make the hardlockup watchdog more
>> easily replaceable by arch code. The last patch provides some
>> justification for why we want to do this (existing sparc watchdog
>> is another that could benefit).
>>
>> Since last time, we tidied up the Kconfig approach so it was not
>> so jumbled -- Thanks Don and Babu for help and suggestions there.
> Thanks Nick! I think this version is good enough for now.
>
>> Since last posting the powerpc patch, I re-added the soft-NMI
>> watchdog, but have it using the emergency stack so we can always
>> avoid the process stacks when perf interrupts are soft-masked.
>>
>> I'm not sure of the best strategy to merge this. The powerpc
>> watchdog is too much to carry anywhere but powerpc tree now, and
>> watchdog patches seem to go via Andrew...
> Andrew, suggestions here?
>
>
> Reviewed-by: Don Zickus <dzickus@redhat.com>
> Acked-by: Don Zickus <dzickus@redhat.com>
>
>
>> Thanks,
>> Nick
>>
>> Nicholas Piggin (5):
>> watchdog: remove unused declaration
>> watchdog: introduce arch_touch_nmi_watchdog()
>> watchdog: split up config options
>> watchdog: provide watchdog_reconfigure() for arch watchdogs
>> powerpc/64s: implement arch-specific hardlockup watchdog
>>
>> arch/Kconfig | 23 ++
>> arch/blackfin/include/asm/nmi.h | 2 +
>> arch/blackfin/kernel/nmi.c | 2 +-
>> arch/mn10300/include/asm/nmi.h | 2 +
>> arch/mn10300/kernel/mn10300-watchdog-low.S | 8 +-
>> arch/mn10300/kernel/mn10300-watchdog.c | 2 +-
>> arch/powerpc/Kconfig | 6 +-
>> arch/powerpc/include/asm/nmi.h | 11 +
>> arch/powerpc/include/asm/smp.h | 2 +
>> arch/powerpc/kernel/Makefile | 1 +
>> arch/powerpc/kernel/exceptions-64s.S | 30 ++-
>> arch/powerpc/kernel/kvm.c | 7 +
>> arch/powerpc/kernel/setup_64.c | 19 --
>> arch/powerpc/kernel/smp.c | 20 +-
>> arch/powerpc/kernel/watchdog.c | 388 +++++++++++++++++++++++++++++
>> arch/sparc/include/asm/nmi.h | 1 +
>> arch/sparc/kernel/nmi.c | 6 +-
>> arch/x86/Kconfig | 1 +
>> arch/x86/kernel/apic/hw_nmi.c | 2 +-
>> include/linux/nmi.h | 57 +++--
>> kernel/Makefile | 2 +-
>> kernel/sysctl.c | 31 ++-
>> kernel/watchdog.c | 268 +++++++++++++-------
>> kernel/watchdog_hld.c | 37 +--
>> lib/Kconfig.debug | 45 ++--
>> 25 files changed, 763 insertions(+), 210 deletions(-)
>> create mode 100644 arch/powerpc/kernel/watchdog.c
>>
>> --
>> 2.11.0
>>
^ permalink raw reply [flat|nested] 13+ messages in thread