[PATCH 4/4] [RFC][powerpc] nmi_watchdog: support for powerpc

From: Don Zickus
Date: Fri Feb 12 2010 - 17:20:44 EST


In order to make sure other arches compiled properly, I used powerpc as a
test bed. It was just basic sanity checking code to see if the
infrastructure bits worked ok. The lock up detection logic is non-existant
and the sample_period is some large made up number.

The interesting piece of the patch is the change to ppc970-pmc.c. I had to
move the initcall from arch_init to early_init to allow the nmi_watchdog to
register with the perf_event subsystem as cpus were coming online.
Otherwise it failed with ENXIO.

This patch is just a conversation starter. I am not sure if powerpc has a
true NMI and if this is really needed. But since I spent some time making
it work, I thought I would throw it out there.

Signed-off-by: Don Zickus <dzickus@xxxxxxxxxx>
---
arch/powerpc/Kconfig | 1 +
arch/powerpc/kernel/Makefile | 1 +
arch/powerpc/kernel/hw_nmi.c | 45 ++++++++++++++++++++++++++++++++++++++
arch/powerpc/kernel/ppc970-pmu.c | 2 +-
4 files changed, 48 insertions(+), 1 deletions(-)
create mode 100644 arch/powerpc/kernel/hw_nmi.c

diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index ba3948c..146b0b5 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -140,6 +140,7 @@ config PPC
select HAVE_SYSCALL_WRAPPERS if PPC64
select GENERIC_ATOMIC64 if PPC32
select HAVE_PERF_EVENTS
+ select PERF_EVENTS_NMI

config EARLY_PRINTK
bool
diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
index c002b04..08e3d2d 100644
--- a/arch/powerpc/kernel/Makefile
+++ b/arch/powerpc/kernel/Makefile
@@ -86,6 +86,7 @@ obj-$(CONFIG_KPROBES) += kprobes.o
obj-$(CONFIG_PPC_UDBG_16550) += legacy_serial.o udbg_16550.o
obj-$(CONFIG_STACKTRACE) += stacktrace.o
obj-$(CONFIG_SWIOTLB) += dma-swiotlb.o
+obj-$(CONFIG_NMI_WATCHDOG) += hw_nmi.o

pci64-$(CONFIG_PPC64) += pci_dn.o isa-bridge.o
obj-$(CONFIG_PCI) += pci_$(CONFIG_WORD_SIZE).o $(pci64-y) \
diff --git a/arch/powerpc/kernel/hw_nmi.c b/arch/powerpc/kernel/hw_nmi.c
new file mode 100644
index 0000000..2313724
--- /dev/null
+++ b/arch/powerpc/kernel/hw_nmi.c
@@ -0,0 +1,45 @@
+/*
+ * HW NMI watchdog support
+ *
+ * started by Don Zickus, Copyright (C) 2010 Red Hat, Inc.
+ *
+ * Arch specific calls to support NMI watchdog
+ *
+ * Bits copied from original nmi.c file
+ *
+ */
+
+#include <linux/smp.h>
+//#include <linux/cpumask.h>
+//#include <linux/sched.h>
+#include <linux/percpu.h>
+#include <linux/kernel_stat.h>
+
+//#include <linux/nmi.h>
+
+static DEFINE_PER_CPU(unsigned, last_irq_sum);
+
+int hw_nmi_is_cpu_stuck(struct pt_regs *regs)
+{
+ unsigned int sum;
+ int cpu = smp_processor_id();
+
+ /* We determine if the cpu is stuck by checking whether any
+ * interrupts have happened since we last checked. Of course
+ * an nmi storm could create false positives, but the higher
+ * level logic should account for that
+ */
+ return 0;
+ sum = kstat_irqs_cpu(0, cpu);
+ if (__get_cpu_var(last_irq_sum) == sum) {
+ return 1;
+ } else {
+ __get_cpu_var(last_irq_sum) = sum;
+ return 0;
+ }
+}
+
+u64 hw_nmi_get_sample_period(void)
+{
+ return 10000 * 1000;
+}
diff --git a/arch/powerpc/kernel/ppc970-pmu.c b/arch/powerpc/kernel/ppc970-pmu.c
index 8eff48e..d91aaf1 100644
--- a/arch/powerpc/kernel/ppc970-pmu.c
+++ b/arch/powerpc/kernel/ppc970-pmu.c
@@ -492,4 +492,4 @@ static int init_ppc970_pmu(void)
return register_power_pmu(&ppc970_pmu);
}

-arch_initcall(init_ppc970_pmu);
+early_initcall(init_ppc970_pmu);
--
1.6.6.83.gc9a2

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/