* [PATCH 0/3 v2] x86: Remove old nmi_watchdog
@ 2010-11-23 19:32 Don Zickus
2010-11-23 19:32 ` [PATCH 1/3] x86, NMI: Remove do_nmi_callback logic Don Zickus
` (3 more replies)
0 siblings, 4 replies; 9+ messages in thread
From: Don Zickus @ 2010-11-23 19:32 UTC (permalink / raw)
To: Ingo Molnar; +Cc: fweisbec, gorcunov, LKML, yinghai, Don Zickus
Ingo asked me to remove the old nmi_watchdog to minimize the amount of
code duplication we have to do for recent fixes. All the functionality
needed by the old nmi watchdog should be available in perf now.
This is done in two parts. The first part is simply remove the old code
and leave stubs for the functions that were removed. This mimics the
behaviour of the new nmi watchdog.
The second parts removes all the stub functions, which is probably a
little more intrusive, but basically dead code anyway.
The only small piece left is the perf counter reservation code. That can't
be removed until oprofile is modified to sit on top of perf (if that is
possible).
Compile tested for 32/64 bit with the lockup detector turned on and off.
Tested on 32/64 bit machines to make sure they still boot properly.
NOTE: This patch will clash with my nmi rewrite patch series. I don't
which patch series will go in first. I can refresh either one once one
of them is merged.
Don Zickus (3):
x86, NMI: Remove do_nmi_callback logic
x86, nmi_watchdog: remove the old nmi_watchdog
x86, nmi_watchdog: remove all stub function calls from old
nmi_watchdog
Documentation/kernel-parameters.txt | 10 +-
arch/x86/include/asm/nmi.h | 51 ---
arch/x86/include/asm/smpboot_hooks.h | 1 -
arch/x86/include/asm/timer.h | 6 -
arch/x86/kernel/apic/Makefile | 5 +-
arch/x86/kernel/apic/apic.c | 15 +-
arch/x86/kernel/apic/hw_nmi.c | 27 +-
arch/x86/kernel/apic/io_apic.c | 46 ---
arch/x86/kernel/apic/nmi.c | 567 ----------------------------
arch/x86/kernel/cpu/perf_event.c | 9 -
arch/x86/kernel/cpu/perfctr-watchdog.c | 642 --------------------------------
arch/x86/kernel/smpboot.c | 11 -
arch/x86/kernel/time.c | 18 -
arch/x86/kernel/traps.c | 26 +-
arch/x86/oprofile/nmi_timer_int.c | 3 -
drivers/acpi/acpica/nsinit.c | 2 -
drivers/watchdog/hpwdt.c | 7 +-
include/linux/nmi.h | 4 -
kernel/sysctl.c | 9 +-
kernel/sysctl_binary.c | 1 -
kernel/watchdog.c | 2 +
21 files changed, 29 insertions(+), 1433 deletions(-)
delete mode 100644 arch/x86/kernel/apic/nmi.c
--
1.7.3.2
^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH 1/3] x86, NMI: Remove do_nmi_callback logic
2010-11-23 19:32 [PATCH 0/3 v2] x86: Remove old nmi_watchdog Don Zickus
@ 2010-11-23 19:32 ` Don Zickus
2010-11-23 19:32 ` [PATCH 2/3] x86, nmi_watchdog: remove the old nmi_watchdog Don Zickus
` (2 subsequent siblings)
3 siblings, 0 replies; 9+ messages in thread
From: Don Zickus @ 2010-11-23 19:32 UTC (permalink / raw)
To: Ingo Molnar; +Cc: fweisbec, gorcunov, LKML, yinghai, Don Zickus, Huang Ying
From: Huang Ying <ying.huang@intel.com>
do_nmi_callback related logic is removed, because it is useless, just
adds code complexity.
unknown_nmi_panic sysctl is reserved to keep kernel ABI unchanged.
[ I made a bunch of tweaks to deal with the this being a standalone
patch instead of later in a series -dcz ]
Signed-off-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Don Zickus <dzickus@redhat.com>
---
arch/x86/kernel/apic/hw_nmi.c | 1 -
arch/x86/kernel/apic/nmi.c | 27 ---------------------------
arch/x86/kernel/traps.c | 16 +++++++++++-----
3 files changed, 11 insertions(+), 33 deletions(-)
diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c
index cefd694..e66b16d 100644
--- a/arch/x86/kernel/apic/hw_nmi.c
+++ b/arch/x86/kernel/apic/hw_nmi.c
@@ -100,7 +100,6 @@ void acpi_nmi_disable(void) { return; }
#endif
atomic_t nmi_active = ATOMIC_INIT(0); /* oprofile uses this */
EXPORT_SYMBOL(nmi_active);
-int unknown_nmi_panic;
void cpu_nmi_set_wd_enabled(void) { return; }
void stop_apic_nmi_watchdog(void *unused) { return; }
void setup_apic_nmi_watchdog(void *unused) { return; }
diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c
index c90041c..7bab143 100644
--- a/arch/x86/kernel/apic/nmi.c
+++ b/arch/x86/kernel/apic/nmi.c
@@ -37,7 +37,6 @@
#include <asm/mach_traps.h>
-int unknown_nmi_panic;
int nmi_watchdog_enabled;
/* For reliability, we're prepared to waste bits here. */
@@ -483,23 +482,6 @@ static void disable_ioapic_nmi_watchdog(void)
on_each_cpu(stop_apic_nmi_watchdog, NULL, 1);
}
-static int __init setup_unknown_nmi_panic(char *str)
-{
- unknown_nmi_panic = 1;
- return 1;
-}
-__setup("unknown_nmi_panic", setup_unknown_nmi_panic);
-
-static int unknown_nmi_panic_callback(struct pt_regs *regs, int cpu)
-{
- unsigned char reason = get_nmi_reason();
- char buf[64];
-
- sprintf(buf, "NMI received for unknown reason %02x\n", reason);
- die_nmi(buf, regs, 1); /* Always panic here */
- return 0;
-}
-
/*
* proc handler for /proc/sys/kernel/nmi
*/
@@ -540,15 +522,6 @@ int proc_nmi_enabled(struct ctl_table *table, int write,
#endif /* CONFIG_SYSCTL */
-int do_nmi_callback(struct pt_regs *regs, int cpu)
-{
-#ifdef CONFIG_SYSCTL
- if (unknown_nmi_panic)
- return unknown_nmi_panic_callback(regs, cpu);
-#endif
- return 0;
-}
-
void arch_trigger_all_cpu_backtrace(void)
{
int i;
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index cb838ca..53ea71f 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -83,6 +83,8 @@ EXPORT_SYMBOL_GPL(used_vectors);
static int ignore_nmis;
+int unknown_nmi_panic;
+
static inline void conditional_sti(struct pt_regs *regs)
{
if (regs->flags & X86_EFLAGS_IF)
@@ -300,6 +302,13 @@ gp_in_kernel:
die("general protection fault", regs, error_code);
}
+static int __init setup_unknown_nmi_panic(char *str)
+{
+ unknown_nmi_panic = 1;
+ return 1;
+}
+__setup("unknown_nmi_panic", setup_unknown_nmi_panic);
+
static notrace __kprobes void
mem_parity_error(unsigned char reason, struct pt_regs *regs)
{
@@ -371,7 +380,7 @@ unknown_nmi_error(unsigned char reason, struct pt_regs *regs)
reason, smp_processor_id());
printk(KERN_EMERG "Do you have a strange power saving mode enabled?\n");
- if (panic_on_unrecovered_nmi)
+ if (unknown_nmi_panic || panic_on_unrecovered_nmi)
panic("NMI: Not continuing");
printk(KERN_EMERG "Dazed and confused, but trying to continue\n");
@@ -405,12 +414,9 @@ static notrace __kprobes void default_do_nmi(struct pt_regs *regs)
*/
if (nmi_watchdog_tick(regs, reason))
return;
- if (!do_nmi_callback(regs, cpu))
#endif /* !CONFIG_LOCKUP_DETECTOR */
- unknown_nmi_error(reason, regs);
-#else
- unknown_nmi_error(reason, regs);
#endif
+ unknown_nmi_error(reason, regs);
return;
}
--
1.7.3.2
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH 2/3] x86, nmi_watchdog: remove the old nmi_watchdog
2010-11-23 19:32 [PATCH 0/3 v2] x86: Remove old nmi_watchdog Don Zickus
2010-11-23 19:32 ` [PATCH 1/3] x86, NMI: Remove do_nmi_callback logic Don Zickus
@ 2010-11-23 19:32 ` Don Zickus
2010-11-24 3:23 ` Yinghai Lu
2010-11-23 19:32 ` [PATCH 3/3] x86, nmi_watchdog: remove all stub function calls from " Don Zickus
2010-11-23 21:49 ` [PATCH 0/3 v2] x86: Remove " Cyrill Gorcunov
3 siblings, 1 reply; 9+ messages in thread
From: Don Zickus @ 2010-11-23 19:32 UTC (permalink / raw)
To: Ingo Molnar; +Cc: fweisbec, gorcunov, LKML, yinghai, Don Zickus
Now that we have a new nmi_watchdog that is more generic and sits on
top of the perf subsystem, we really do not need the old nmi_watchdog
any more.
In addition, the old nmi_watchdog doesn't really work if you are using
the default clocksource, hpet. The old nmi_watchdog code relied on
local apic interrupts to determine if the cpu is still alive. With
hpet as the clocksource, these interrupts don't increment any more
and the old nmi_watchdog triggers false postives.
This piece removes the old nmi_watchdog code and stubs out any variables
and functions calls. The stubs are the same ones used by the new
nmi_watchdog code, so it should be well tested.
v2: fix sparc breakage by undoing linux/nmi.h changes
clean up documentation and sysctl stuff
Signed-off-by: Don Zickus <dzickus@redhat.com>
---
Documentation/kernel-parameters.txt | 10 +-
arch/x86/include/asm/nmi.h | 4 -
arch/x86/kernel/apic/Makefile | 5 +-
arch/x86/kernel/apic/hw_nmi.c | 14 +-
arch/x86/kernel/apic/nmi.c | 540 -----------------------------------
arch/x86/kernel/traps.c | 8 -
kernel/sysctl.c | 9 +-
kernel/sysctl_binary.c | 1 -
kernel/watchdog.c | 2 +
9 files changed, 16 insertions(+), 577 deletions(-)
delete mode 100644 arch/x86/kernel/apic/nmi.c
diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
index ed45e98..e48b93f 100644
--- a/Documentation/kernel-parameters.txt
+++ b/Documentation/kernel-parameters.txt
@@ -1579,20 +1579,12 @@ and is between 256 and 4096 characters. It is defined in the file
nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
Format: [panic,][num]
- Valid num: 0,1,2
+ Valid num: 0
0 - turn nmi_watchdog off
- 1 - use the IO-APIC timer for the NMI watchdog
- 2 - use the local APIC for the NMI watchdog using
- a performance counter. Note: This will use one
- performance counter and the local APIC's performance
- vector.
When panic is specified, panic when an NMI watchdog
timeout occurs.
This is useful when you use a panic=... timeout and
need the box quickly up again.
- Instead of 1 and 2 it is possible to use the following
- symbolic names: lapic and ioapic
- Example: nmi_watchdog=2 or nmi_watchdog=panic,lapic
netpoll.carrier_timeout=
[NET] Specifies amount of time (in seconds) that
diff --git a/arch/x86/include/asm/nmi.h b/arch/x86/include/asm/nmi.h
index 932f0f8..33292ec 100644
--- a/arch/x86/include/asm/nmi.h
+++ b/arch/x86/include/asm/nmi.h
@@ -17,9 +17,6 @@ int do_nmi_callback(struct pt_regs *regs, int cpu);
extern void die_nmi(char *str, struct pt_regs *regs, int do_panic);
extern int check_nmi_watchdog(void);
-#if !defined(CONFIG_LOCKUP_DETECTOR)
-extern int nmi_watchdog_enabled;
-#endif
extern int avail_to_resrv_perfctr_nmi_bit(unsigned int);
extern int reserve_perfctr_nmi(unsigned int);
extern void release_perfctr_nmi(unsigned int);
@@ -30,7 +27,6 @@ extern void setup_apic_nmi_watchdog(void *);
extern void stop_apic_nmi_watchdog(void *);
extern void disable_timer_nmi_watchdog(void);
extern void enable_timer_nmi_watchdog(void);
-extern int nmi_watchdog_tick(struct pt_regs *regs, unsigned reason);
extern void cpu_nmi_set_wd_enabled(void);
extern atomic_t nmi_active;
diff --git a/arch/x86/kernel/apic/Makefile b/arch/x86/kernel/apic/Makefile
index 910f20b..3966b56 100644
--- a/arch/x86/kernel/apic/Makefile
+++ b/arch/x86/kernel/apic/Makefile
@@ -3,10 +3,7 @@
#
obj-$(CONFIG_X86_LOCAL_APIC) += apic.o apic_noop.o probe_$(BITS).o ipi.o
-ifneq ($(CONFIG_HARDLOCKUP_DETECTOR),y)
-obj-$(CONFIG_X86_LOCAL_APIC) += nmi.o
-endif
-obj-$(CONFIG_HARDLOCKUP_DETECTOR) += hw_nmi.o
+obj-y += hw_nmi.o
obj-$(CONFIG_X86_IO_APIC) += io_apic.o
obj-$(CONFIG_SMP) += ipi.o
diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c
index e66b16d..4a57e60 100644
--- a/arch/x86/kernel/apic/hw_nmi.c
+++ b/arch/x86/kernel/apic/hw_nmi.c
@@ -20,12 +20,22 @@
/* For reliability, we're prepared to waste bits here. */
static DECLARE_BITMAP(backtrace_mask, NR_CPUS) __read_mostly;
+#ifdef CONFIG_HARDLOCKUP_DETECTOR
u64 hw_nmi_get_sample_period(void)
{
return (u64)(cpu_khz) * 1000 * 60;
}
+#endif
+
+#ifndef CONFIG_HARDLOCKUP_DETECTOR
+void touch_nmi_watchdog(void)
+{
+ touch_softlockup_watchdog();
+}
+EXPORT_SYMBOL(touch_nmi_watchdog);
+#endif
-#ifdef ARCH_HAS_NMI_WATCHDOG
+#ifdef arch_trigger_all_cpu_backtrace
void arch_trigger_all_cpu_backtrace(void)
{
int i;
@@ -95,8 +105,6 @@ early_initcall(register_trigger_all_cpu_backtrace);
#if defined(CONFIG_X86_LOCAL_APIC)
unsigned int nmi_watchdog = NMI_NONE;
EXPORT_SYMBOL(nmi_watchdog);
-void acpi_nmi_enable(void) { return; }
-void acpi_nmi_disable(void) { return; }
#endif
atomic_t nmi_active = ATOMIC_INIT(0); /* oprofile uses this */
EXPORT_SYMBOL(nmi_active);
diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c
deleted file mode 100644
index 7bab143..0000000
--- a/arch/x86/kernel/apic/nmi.c
+++ /dev/null
@@ -1,540 +0,0 @@
-/*
- * NMI watchdog support on APIC systems
- *
- * Started by Ingo Molnar <mingo@redhat.com>
- *
- * Fixes:
- * Mikael Pettersson : AMD K7 support for local APIC NMI watchdog.
- * Mikael Pettersson : Power Management for local APIC NMI watchdog.
- * Mikael Pettersson : Pentium 4 support for local APIC NMI watchdog.
- * Pavel Machek and
- * Mikael Pettersson : PM converted to driver model. Disable/enable API.
- */
-
-#include <asm/apic.h>
-
-#include <linux/nmi.h>
-#include <linux/mm.h>
-#include <linux/delay.h>
-#include <linux/interrupt.h>
-#include <linux/module.h>
-#include <linux/slab.h>
-#include <linux/sysdev.h>
-#include <linux/sysctl.h>
-#include <linux/percpu.h>
-#include <linux/kprobes.h>
-#include <linux/cpumask.h>
-#include <linux/kernel_stat.h>
-#include <linux/kdebug.h>
-#include <linux/smp.h>
-
-#include <asm/i8259.h>
-#include <asm/io_apic.h>
-#include <asm/proto.h>
-#include <asm/timer.h>
-
-#include <asm/mce.h>
-
-#include <asm/mach_traps.h>
-
-int nmi_watchdog_enabled;
-
-/* For reliability, we're prepared to waste bits here. */
-static DECLARE_BITMAP(backtrace_mask, NR_CPUS) __read_mostly;
-
-/* nmi_active:
- * >0: the lapic NMI watchdog is active, but can be disabled
- * <0: the lapic NMI watchdog has not been set up, and cannot
- * be enabled
- * 0: the lapic NMI watchdog is disabled, but can be enabled
- */
-atomic_t nmi_active = ATOMIC_INIT(0); /* oprofile uses this */
-EXPORT_SYMBOL(nmi_active);
-
-unsigned int nmi_watchdog = NMI_NONE;
-EXPORT_SYMBOL(nmi_watchdog);
-
-static int panic_on_timeout;
-
-static unsigned int nmi_hz = HZ;
-static DEFINE_PER_CPU(short, wd_enabled);
-static int endflag __initdata;
-
-static inline unsigned int get_nmi_count(int cpu)
-{
- return per_cpu(irq_stat, cpu).__nmi_count;
-}
-
-static inline int mce_in_progress(void)
-{
-#if defined(CONFIG_X86_MCE)
- return atomic_read(&mce_entry) > 0;
-#endif
- return 0;
-}
-
-/*
- * Take the local apic timer and PIT/HPET into account. We don't
- * know which one is active, when we have highres/dyntick on
- */
-static inline unsigned int get_timer_irqs(int cpu)
-{
- return per_cpu(irq_stat, cpu).apic_timer_irqs +
- per_cpu(irq_stat, cpu).irq0_irqs;
-}
-
-#ifdef CONFIG_SMP
-/*
- * The performance counters used by NMI_LOCAL_APIC don't trigger when
- * the CPU is idle. To make sure the NMI watchdog really ticks on all
- * CPUs during the test make them busy.
- */
-static __init void nmi_cpu_busy(void *data)
-{
- local_irq_enable_in_hardirq();
- /*
- * Intentionally don't use cpu_relax here. This is
- * to make sure that the performance counter really ticks,
- * even if there is a simulator or similar that catches the
- * pause instruction. On a real HT machine this is fine because
- * all other CPUs are busy with "useless" delay loops and don't
- * care if they get somewhat less cycles.
- */
- while (endflag == 0)
- mb();
-}
-#endif
-
-static void report_broken_nmi(int cpu, unsigned int *prev_nmi_count)
-{
- printk(KERN_CONT "\n");
-
- printk(KERN_WARNING
- "WARNING: CPU#%d: NMI appears to be stuck (%d->%d)!\n",
- cpu, prev_nmi_count[cpu], get_nmi_count(cpu));
-
- printk(KERN_WARNING
- "Please report this to bugzilla.kernel.org,\n");
- printk(KERN_WARNING
- "and attach the output of the 'dmesg' command.\n");
-
- per_cpu(wd_enabled, cpu) = 0;
- atomic_dec(&nmi_active);
-}
-
-static void __acpi_nmi_disable(void *__unused)
-{
- apic_write(APIC_LVT0, APIC_DM_NMI | APIC_LVT_MASKED);
-}
-
-int __init check_nmi_watchdog(void)
-{
- unsigned int *prev_nmi_count;
- int cpu;
-
- if (!nmi_watchdog_active() || !atomic_read(&nmi_active))
- return 0;
-
- prev_nmi_count = kmalloc(nr_cpu_ids * sizeof(int), GFP_KERNEL);
- if (!prev_nmi_count)
- goto error;
-
- printk(KERN_INFO "Testing NMI watchdog ... ");
-
-#ifdef CONFIG_SMP
- if (nmi_watchdog == NMI_LOCAL_APIC)
- smp_call_function(nmi_cpu_busy, (void *)&endflag, 0);
-#endif
-
- for_each_possible_cpu(cpu)
- prev_nmi_count[cpu] = get_nmi_count(cpu);
- local_irq_enable();
- mdelay((20 * 1000) / nmi_hz); /* wait 20 ticks */
-
- for_each_online_cpu(cpu) {
- if (!per_cpu(wd_enabled, cpu))
- continue;
- if (get_nmi_count(cpu) - prev_nmi_count[cpu] <= 5)
- report_broken_nmi(cpu, prev_nmi_count);
- }
- endflag = 1;
- if (!atomic_read(&nmi_active)) {
- kfree(prev_nmi_count);
- atomic_set(&nmi_active, -1);
- goto error;
- }
- printk("OK.\n");
-
- /*
- * now that we know it works we can reduce NMI frequency to
- * something more reasonable; makes a difference in some configs
- */
- if (nmi_watchdog == NMI_LOCAL_APIC)
- nmi_hz = lapic_adjust_nmi_hz(1);
-
- kfree(prev_nmi_count);
- return 0;
-error:
- if (nmi_watchdog == NMI_IO_APIC) {
- if (!timer_through_8259)
- legacy_pic->mask(0);
- on_each_cpu(__acpi_nmi_disable, NULL, 1);
- }
-
-#ifdef CONFIG_X86_32
- timer_ack = 0;
-#endif
- return -1;
-}
-
-static int __init setup_nmi_watchdog(char *str)
-{
- unsigned int nmi;
-
- if (!strncmp(str, "panic", 5)) {
- panic_on_timeout = 1;
- str = strchr(str, ',');
- if (!str)
- return 1;
- ++str;
- }
-
- if (!strncmp(str, "lapic", 5))
- nmi_watchdog = NMI_LOCAL_APIC;
- else if (!strncmp(str, "ioapic", 6))
- nmi_watchdog = NMI_IO_APIC;
- else {
- get_option(&str, &nmi);
- if (nmi >= NMI_INVALID)
- return 0;
- nmi_watchdog = nmi;
- }
-
- return 1;
-}
-__setup("nmi_watchdog=", setup_nmi_watchdog);
-
-/*
- * Suspend/resume support
- */
-#ifdef CONFIG_PM
-
-static int nmi_pm_active; /* nmi_active before suspend */
-
-static int lapic_nmi_suspend(struct sys_device *dev, pm_message_t state)
-{
- /* only CPU0 goes here, other CPUs should be offline */
- nmi_pm_active = atomic_read(&nmi_active);
- stop_apic_nmi_watchdog(NULL);
- BUG_ON(atomic_read(&nmi_active) != 0);
- return 0;
-}
-
-static int lapic_nmi_resume(struct sys_device *dev)
-{
- /* only CPU0 goes here, other CPUs should be offline */
- if (nmi_pm_active > 0) {
- setup_apic_nmi_watchdog(NULL);
- touch_nmi_watchdog();
- }
- return 0;
-}
-
-static struct sysdev_class nmi_sysclass = {
- .name = "lapic_nmi",
- .resume = lapic_nmi_resume,
- .suspend = lapic_nmi_suspend,
-};
-
-static struct sys_device device_lapic_nmi = {
- .id = 0,
- .cls = &nmi_sysclass,
-};
-
-static int __init init_lapic_nmi_sysfs(void)
-{
- int error;
-
- /*
- * should really be a BUG_ON but b/c this is an
- * init call, it just doesn't work. -dcz
- */
- if (nmi_watchdog != NMI_LOCAL_APIC)
- return 0;
-
- if (atomic_read(&nmi_active) < 0)
- return 0;
-
- error = sysdev_class_register(&nmi_sysclass);
- if (!error)
- error = sysdev_register(&device_lapic_nmi);
- return error;
-}
-
-/* must come after the local APIC's device_initcall() */
-late_initcall(init_lapic_nmi_sysfs);
-
-#endif /* CONFIG_PM */
-
-static void __acpi_nmi_enable(void *__unused)
-{
- apic_write(APIC_LVT0, APIC_DM_NMI);
-}
-
-/*
- * Enable timer based NMIs on all CPUs:
- */
-void acpi_nmi_enable(void)
-{
- if (atomic_read(&nmi_active) && nmi_watchdog == NMI_IO_APIC)
- on_each_cpu(__acpi_nmi_enable, NULL, 1);
-}
-
-/*
- * Disable timer based NMIs on all CPUs:
- */
-void acpi_nmi_disable(void)
-{
- if (atomic_read(&nmi_active) && nmi_watchdog == NMI_IO_APIC)
- on_each_cpu(__acpi_nmi_disable, NULL, 1);
-}
-
-/*
- * This function is called as soon the LAPIC NMI watchdog driver has everything
- * in place and it's ready to check if the NMIs belong to the NMI watchdog
- */
-void cpu_nmi_set_wd_enabled(void)
-{
- __get_cpu_var(wd_enabled) = 1;
-}
-
-void setup_apic_nmi_watchdog(void *unused)
-{
- if (__get_cpu_var(wd_enabled))
- return;
-
- /* cheap hack to support suspend/resume */
- /* if cpu0 is not active neither should the other cpus */
- if (smp_processor_id() != 0 && atomic_read(&nmi_active) <= 0)
- return;
-
- switch (nmi_watchdog) {
- case NMI_LOCAL_APIC:
- if (lapic_watchdog_init(nmi_hz) < 0) {
- __get_cpu_var(wd_enabled) = 0;
- return;
- }
- /* FALL THROUGH */
- case NMI_IO_APIC:
- __get_cpu_var(wd_enabled) = 1;
- atomic_inc(&nmi_active);
- }
-}
-
-void stop_apic_nmi_watchdog(void *unused)
-{
- /* only support LOCAL and IO APICs for now */
- if (!nmi_watchdog_active())
- return;
- if (__get_cpu_var(wd_enabled) == 0)
- return;
- if (nmi_watchdog == NMI_LOCAL_APIC)
- lapic_watchdog_stop();
- else
- __acpi_nmi_disable(NULL);
- __get_cpu_var(wd_enabled) = 0;
- atomic_dec(&nmi_active);
-}
-
-/*
- * the best way to detect whether a CPU has a 'hard lockup' problem
- * is to check it's local APIC timer IRQ counts. If they are not
- * changing then that CPU has some problem.
- *
- * as these watchdog NMI IRQs are generated on every CPU, we only
- * have to check the current processor.
- *
- * since NMIs don't listen to _any_ locks, we have to be extremely
- * careful not to rely on unsafe variables. The printk might lock
- * up though, so we have to break up any console locks first ...
- * [when there will be more tty-related locks, break them up here too!]
- */
-
-static DEFINE_PER_CPU(unsigned, last_irq_sum);
-static DEFINE_PER_CPU(long, alert_counter);
-static DEFINE_PER_CPU(int, nmi_touch);
-
-void touch_nmi_watchdog(void)
-{
- if (nmi_watchdog_active()) {
- unsigned cpu;
-
- /*
- * Tell other CPUs to reset their alert counters. We cannot
- * do it ourselves because the alert count increase is not
- * atomic.
- */
- for_each_present_cpu(cpu) {
- if (per_cpu(nmi_touch, cpu) != 1)
- per_cpu(nmi_touch, cpu) = 1;
- }
- }
-
- /*
- * Tickle the softlockup detector too:
- */
- touch_softlockup_watchdog();
-}
-EXPORT_SYMBOL(touch_nmi_watchdog);
-
-notrace __kprobes int
-nmi_watchdog_tick(struct pt_regs *regs, unsigned reason)
-{
- /*
- * Since current_thread_info()-> is always on the stack, and we
- * always switch the stack NMI-atomically, it's safe to use
- * smp_processor_id().
- */
- unsigned int sum;
- int touched = 0;
- int cpu = smp_processor_id();
- int rc = 0;
-
- sum = get_timer_irqs(cpu);
-
- if (__get_cpu_var(nmi_touch)) {
- __get_cpu_var(nmi_touch) = 0;
- touched = 1;
- }
-
- /* We can be called before check_nmi_watchdog, hence NULL check. */
- if (cpumask_test_cpu(cpu, to_cpumask(backtrace_mask))) {
- static DEFINE_RAW_SPINLOCK(lock); /* Serialise the printks */
-
- raw_spin_lock(&lock);
- printk(KERN_WARNING "NMI backtrace for cpu %d\n", cpu);
- show_regs(regs);
- dump_stack();
- raw_spin_unlock(&lock);
- cpumask_clear_cpu(cpu, to_cpumask(backtrace_mask));
-
- rc = 1;
- }
-
- /* Could check oops_in_progress here too, but it's safer not to */
- if (mce_in_progress())
- touched = 1;
-
- /* if the none of the timers isn't firing, this cpu isn't doing much */
- if (!touched && __get_cpu_var(last_irq_sum) == sum) {
- /*
- * Ayiee, looks like this CPU is stuck ...
- * wait a few IRQs (5 seconds) before doing the oops ...
- */
- __this_cpu_inc(alert_counter);
- if (__this_cpu_read(alert_counter) == 5 * nmi_hz)
- /*
- * die_nmi will return ONLY if NOTIFY_STOP happens..
- */
- die_nmi("BUG: NMI Watchdog detected LOCKUP",
- regs, panic_on_timeout);
- } else {
- __get_cpu_var(last_irq_sum) = sum;
- __this_cpu_write(alert_counter, 0);
- }
-
- /* see if the nmi watchdog went off */
- if (!__get_cpu_var(wd_enabled))
- return rc;
- switch (nmi_watchdog) {
- case NMI_LOCAL_APIC:
- rc |= lapic_wd_event(nmi_hz);
- break;
- case NMI_IO_APIC:
- /*
- * don't know how to accurately check for this.
- * just assume it was a watchdog timer interrupt
- * This matches the old behaviour.
- */
- rc = 1;
- break;
- }
- return rc;
-}
-
-#ifdef CONFIG_SYSCTL
-
-static void enable_ioapic_nmi_watchdog_single(void *unused)
-{
- __get_cpu_var(wd_enabled) = 1;
- atomic_inc(&nmi_active);
- __acpi_nmi_enable(NULL);
-}
-
-static void enable_ioapic_nmi_watchdog(void)
-{
- on_each_cpu(enable_ioapic_nmi_watchdog_single, NULL, 1);
- touch_nmi_watchdog();
-}
-
-static void disable_ioapic_nmi_watchdog(void)
-{
- on_each_cpu(stop_apic_nmi_watchdog, NULL, 1);
-}
-
-/*
- * proc handler for /proc/sys/kernel/nmi
- */
-int proc_nmi_enabled(struct ctl_table *table, int write,
- void __user *buffer, size_t *length, loff_t *ppos)
-{
- int old_state;
-
- nmi_watchdog_enabled = (atomic_read(&nmi_active) > 0) ? 1 : 0;
- old_state = nmi_watchdog_enabled;
- proc_dointvec(table, write, buffer, length, ppos);
- if (!!old_state == !!nmi_watchdog_enabled)
- return 0;
-
- if (atomic_read(&nmi_active) < 0 || !nmi_watchdog_active()) {
- printk(KERN_WARNING
- "NMI watchdog is permanently disabled\n");
- return -EIO;
- }
-
- if (nmi_watchdog == NMI_LOCAL_APIC) {
- if (nmi_watchdog_enabled)
- enable_lapic_nmi_watchdog();
- else
- disable_lapic_nmi_watchdog();
- } else if (nmi_watchdog == NMI_IO_APIC) {
- if (nmi_watchdog_enabled)
- enable_ioapic_nmi_watchdog();
- else
- disable_ioapic_nmi_watchdog();
- } else {
- printk(KERN_WARNING
- "NMI watchdog doesn't know what hardware to touch\n");
- return -EIO;
- }
- return 0;
-}
-
-#endif /* CONFIG_SYSCTL */
-
-void arch_trigger_all_cpu_backtrace(void)
-{
- int i;
-
- cpumask_copy(to_cpumask(backtrace_mask), cpu_online_mask);
-
- printk(KERN_INFO "sending NMI to all CPUs:\n");
- apic->send_IPI_all(NMI_VECTOR);
-
- /* Wait for up to 10 seconds for all CPUs to do the backtrace */
- for (i = 0; i < 10 * 1000; i++) {
- if (cpumask_empty(to_cpumask(backtrace_mask)))
- break;
- mdelay(1);
- }
-}
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 53ea71f..4b331ce 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -407,14 +407,6 @@ static notrace __kprobes void default_do_nmi(struct pt_regs *regs)
== NOTIFY_STOP)
return;
-#ifndef CONFIG_LOCKUP_DETECTOR
- /*
- * Ok, so this is none of the documented NMI sources,
- * so it must be the NMI watchdog.
- */
- if (nmi_watchdog_tick(regs, reason))
- return;
-#endif /* !CONFIG_LOCKUP_DETECTOR */
#endif
unknown_nmi_error(reason, regs);
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index c33a1ed..0daad4e 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -737,7 +737,7 @@ static struct ctl_table kern_table[] = {
.extra2 = &one,
},
#endif
-#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86) && !defined(CONFIG_LOCKUP_DETECTOR)
+#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
{
.procname = "unknown_nmi_panic",
.data = &unknown_nmi_panic,
@@ -745,13 +745,6 @@ static struct ctl_table kern_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec,
},
- {
- .procname = "nmi_watchdog",
- .data = &nmi_watchdog_enabled,
- .maxlen = sizeof (int),
- .mode = 0644,
- .proc_handler = proc_nmi_enabled,
- },
#endif
#if defined(CONFIG_X86)
{
diff --git a/kernel/sysctl_binary.c b/kernel/sysctl_binary.c
index 1357c57..4b2545a 100644
--- a/kernel/sysctl_binary.c
+++ b/kernel/sysctl_binary.c
@@ -136,7 +136,6 @@ static const struct bin_table bin_kern_table[] = {
{ CTL_INT, KERN_IA64_UNALIGNED, "ignore-unaligned-usertrap" },
{ CTL_INT, KERN_COMPAT_LOG, "compat-log" },
{ CTL_INT, KERN_MAX_LOCK_DEPTH, "max_lock_depth" },
- { CTL_INT, KERN_NMI_WATCHDOG, "nmi_watchdog" },
{ CTL_INT, KERN_PANIC_ON_NMI, "panic_on_unrecovered_nmi" },
{}
};
diff --git a/kernel/watchdog.c b/kernel/watchdog.c
index bafba68..3c5441f 100644
--- a/kernel/watchdog.c
+++ b/kernel/watchdog.c
@@ -57,6 +57,8 @@ static int __init hardlockup_panic_setup(char *str)
{
if (!strncmp(str, "panic", 5))
hardlockup_panic = 1;
+ else if (!strncmp(str, "0", 1))
+ no_watchdog = 1;
return 1;
}
__setup("nmi_watchdog=", hardlockup_panic_setup);
--
1.7.3.2
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH 3/3] x86, nmi_watchdog: remove all stub function calls from old nmi_watchdog
2010-11-23 19:32 [PATCH 0/3 v2] x86: Remove old nmi_watchdog Don Zickus
2010-11-23 19:32 ` [PATCH 1/3] x86, NMI: Remove do_nmi_callback logic Don Zickus
2010-11-23 19:32 ` [PATCH 2/3] x86, nmi_watchdog: remove the old nmi_watchdog Don Zickus
@ 2010-11-23 19:32 ` Don Zickus
2010-11-23 21:49 ` [PATCH 0/3 v2] x86: Remove " Cyrill Gorcunov
3 siblings, 0 replies; 9+ messages in thread
From: Don Zickus @ 2010-11-23 19:32 UTC (permalink / raw)
To: Ingo Molnar; +Cc: fweisbec, gorcunov, LKML, yinghai, Don Zickus
Now that the bulk of the old nmi_watchdog is gone, remove all the stub
variables and hooks associated with it.
This touches lots of files mainly because of how the io_apic nmi_watchdog
was implemented. Now that the io_apic nmi_watchdog is forever gone,
remove all its fingers.
Most of this code was not being exercised by virtue of nmi_watchdog !=
NMI_IO_APIC, so there shouldn't be anything to risky here.
Signed-off-by: Don Zickus <dzickus@redhat.com>
---
arch/x86/include/asm/nmi.h | 47 ---
arch/x86/include/asm/smpboot_hooks.h | 1 -
arch/x86/include/asm/timer.h | 6 -
arch/x86/kernel/apic/apic.c | 15 +-
arch/x86/kernel/apic/hw_nmi.c | 12 -
arch/x86/kernel/apic/io_apic.c | 46 ---
arch/x86/kernel/cpu/perf_event.c | 9 -
arch/x86/kernel/cpu/perfctr-watchdog.c | 642 --------------------------------
arch/x86/kernel/smpboot.c | 11 -
arch/x86/kernel/time.c | 18 -
arch/x86/kernel/traps.c | 2 -
arch/x86/oprofile/nmi_timer_int.c | 3 -
drivers/acpi/acpica/nsinit.c | 2 -
drivers/watchdog/hpwdt.c | 7 +-
include/linux/nmi.h | 4 -
15 files changed, 2 insertions(+), 823 deletions(-)
diff --git a/arch/x86/include/asm/nmi.h b/arch/x86/include/asm/nmi.h
index 33292ec..3545838 100644
--- a/arch/x86/include/asm/nmi.h
+++ b/arch/x86/include/asm/nmi.h
@@ -7,35 +7,13 @@
#ifdef ARCH_HAS_NMI_WATCHDOG
-/**
- * do_nmi_callback
- *
- * Check to see if a callback exists and execute it. Return 1
- * if the handler exists and was handled successfully.
- */
-int do_nmi_callback(struct pt_regs *regs, int cpu);
-
extern void die_nmi(char *str, struct pt_regs *regs, int do_panic);
-extern int check_nmi_watchdog(void);
extern int avail_to_resrv_perfctr_nmi_bit(unsigned int);
extern int reserve_perfctr_nmi(unsigned int);
extern void release_perfctr_nmi(unsigned int);
extern int reserve_evntsel_nmi(unsigned int);
extern void release_evntsel_nmi(unsigned int);
-extern void setup_apic_nmi_watchdog(void *);
-extern void stop_apic_nmi_watchdog(void *);
-extern void disable_timer_nmi_watchdog(void);
-extern void enable_timer_nmi_watchdog(void);
-extern void cpu_nmi_set_wd_enabled(void);
-
-extern atomic_t nmi_active;
-extern unsigned int nmi_watchdog;
-#define NMI_NONE 0
-#define NMI_IO_APIC 1
-#define NMI_LOCAL_APIC 2
-#define NMI_INVALID 3
-
struct ctl_table;
extern int proc_nmi_enabled(struct ctl_table *, int ,
void __user *, size_t *, loff_t *);
@@ -43,33 +21,8 @@ extern int unknown_nmi_panic;
void arch_trigger_all_cpu_backtrace(void);
#define arch_trigger_all_cpu_backtrace arch_trigger_all_cpu_backtrace
-
-static inline void localise_nmi_watchdog(void)
-{
- if (nmi_watchdog == NMI_IO_APIC)
- nmi_watchdog = NMI_LOCAL_APIC;
-}
-
-/* check if nmi_watchdog is active (ie was specified at boot) */
-static inline int nmi_watchdog_active(void)
-{
- /*
- * actually it should be:
- * return (nmi_watchdog == NMI_LOCAL_APIC ||
- * nmi_watchdog == NMI_IO_APIC)
- * but since they are power of two we could use a
- * cheaper way --cvg
- */
- return nmi_watchdog & (NMI_LOCAL_APIC | NMI_IO_APIC);
-}
#endif
-void lapic_watchdog_stop(void);
-int lapic_watchdog_init(unsigned nmi_hz);
-int lapic_wd_event(unsigned nmi_hz);
-unsigned lapic_adjust_nmi_hz(unsigned hz);
-void disable_lapic_nmi_watchdog(void);
-void enable_lapic_nmi_watchdog(void);
void stop_nmi(void);
void restart_nmi(void);
diff --git a/arch/x86/include/asm/smpboot_hooks.h b/arch/x86/include/asm/smpboot_hooks.h
index 1def601..6c22bf3 100644
--- a/arch/x86/include/asm/smpboot_hooks.h
+++ b/arch/x86/include/asm/smpboot_hooks.h
@@ -48,7 +48,6 @@ static inline void __init smpboot_setup_io_apic(void)
setup_IO_APIC();
else {
nr_ioapics = 0;
- localise_nmi_watchdog();
}
#endif
}
diff --git a/arch/x86/include/asm/timer.h b/arch/x86/include/asm/timer.h
index 5469630..fa7b917 100644
--- a/arch/x86/include/asm/timer.h
+++ b/arch/x86/include/asm/timer.h
@@ -10,12 +10,6 @@
unsigned long long native_sched_clock(void);
extern int recalibrate_cpu_khz(void);
-#if defined(CONFIG_X86_32) && defined(CONFIG_X86_IO_APIC)
-extern int timer_ack;
-#else
-# define timer_ack (0)
-#endif
-
extern int no_timer_check;
/* Accelerators for sched_clock()
diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
index 850657d..c4dafc0 100644
--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -31,7 +31,6 @@
#include <linux/init.h>
#include <linux/cpu.h>
#include <linux/dmi.h>
-#include <linux/nmi.h>
#include <linux/smp.h>
#include <linux/mm.h>
@@ -800,11 +799,7 @@ void __init setup_boot_APIC_clock(void)
* PIT/HPET going. Otherwise register lapic as a dummy
* device.
*/
- if (nmi_watchdog != NMI_IO_APIC)
- lapic_clockevent.features &= ~CLOCK_EVT_FEAT_DUMMY;
- else
- pr_warning("APIC timer registered as dummy,"
- " due to nmi_watchdog=%d!\n", nmi_watchdog);
+ lapic_clockevent.features &= ~CLOCK_EVT_FEAT_DUMMY;
/* Setup the lapic or request the broadcast */
setup_APIC_timer();
@@ -1388,7 +1383,6 @@ void __cpuinit end_local_APIC_setup(void)
}
#endif
- setup_apic_nmi_watchdog(NULL);
apic_pm_activate();
}
@@ -1751,17 +1745,10 @@ int __init APIC_init_uniprocessor(void)
setup_IO_APIC();
else {
nr_ioapics = 0;
- localise_nmi_watchdog();
}
-#else
- localise_nmi_watchdog();
#endif
x86_init.timers.setup_percpu_clockev();
-#ifdef CONFIG_X86_64
- check_nmi_watchdog();
-#endif
-
return 0;
}
diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c
index 4a57e60..b0eb24c 100644
--- a/arch/x86/kernel/apic/hw_nmi.c
+++ b/arch/x86/kernel/apic/hw_nmi.c
@@ -100,15 +100,3 @@ static int __init register_trigger_all_cpu_backtrace(void)
}
early_initcall(register_trigger_all_cpu_backtrace);
#endif
-
-/* STUB calls to mimic old nmi_watchdog behaviour */
-#if defined(CONFIG_X86_LOCAL_APIC)
-unsigned int nmi_watchdog = NMI_NONE;
-EXPORT_SYMBOL(nmi_watchdog);
-#endif
-atomic_t nmi_active = ATOMIC_INIT(0); /* oprofile uses this */
-EXPORT_SYMBOL(nmi_active);
-void cpu_nmi_set_wd_enabled(void) { return; }
-void stop_apic_nmi_watchdog(void *unused) { return; }
-void setup_apic_nmi_watchdog(void *unused) { return; }
-int __init check_nmi_watchdog(void) { return 0; }
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 7cc0a72..e4a040c 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -54,7 +54,6 @@
#include <asm/dma.h>
#include <asm/timer.h>
#include <asm/i8259.h>
-#include <asm/nmi.h>
#include <asm/msidef.h>
#include <asm/hypertransport.h>
#include <asm/setup.h>
@@ -2643,24 +2642,6 @@ static void lapic_register_intr(int irq)
"edge");
}
-static void __init setup_nmi(void)
-{
- /*
- * Dirty trick to enable the NMI watchdog ...
- * We put the 8259A master into AEOI mode and
- * unmask on all local APICs LVT0 as NMI.
- *
- * The idea to use the 8259A in AEOI mode ('8259A Virtual Wire')
- * is from Maciej W. Rozycki - so we do not have to EOI from
- * the NMI handler or the timer interrupt.
- */
- apic_printk(APIC_VERBOSE, KERN_INFO "activating NMI Watchdog ...");
-
- enable_NMI_through_LVT0();
-
- apic_printk(APIC_VERBOSE, " done.\n");
-}
-
/*
* This looks a bit hackish but it's about the only one way of sending
* a few INTA cycles to 8259As and any associated glue logic. ICR does
@@ -2766,15 +2747,6 @@ static inline void __init check_timer(void)
*/
apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT);
legacy_pic->init(1);
-#ifdef CONFIG_X86_32
- {
- unsigned int ver;
-
- ver = apic_read(APIC_LVR);
- ver = GET_APIC_VERSION(ver);
- timer_ack = (nmi_watchdog == NMI_IO_APIC && !APIC_INTEGRATED(ver));
- }
-#endif
pin1 = find_isa_irq_pin(0, mp_INT);
apic1 = find_isa_irq_apic(0, mp_INT);
@@ -2822,10 +2794,6 @@ static inline void __init check_timer(void)
unmask_ioapic(cfg);
}
if (timer_irq_works()) {
- if (nmi_watchdog == NMI_IO_APIC) {
- setup_nmi();
- legacy_pic->unmask(0);
- }
if (disable_timer_pin_1 > 0)
clear_IO_APIC_pin(0, pin1);
goto out;
@@ -2851,11 +2819,6 @@ static inline void __init check_timer(void)
if (timer_irq_works()) {
apic_printk(APIC_QUIET, KERN_INFO "....... works.\n");
timer_through_8259 = 1;
- if (nmi_watchdog == NMI_IO_APIC) {
- legacy_pic->mask(0);
- setup_nmi();
- legacy_pic->unmask(0);
- }
goto out;
}
/*
@@ -2867,15 +2830,6 @@ static inline void __init check_timer(void)
apic_printk(APIC_QUIET, KERN_INFO "....... failed.\n");
}
- if (nmi_watchdog == NMI_IO_APIC) {
- apic_printk(APIC_QUIET, KERN_WARNING "timer doesn't work "
- "through the IO-APIC - disabling NMI Watchdog!\n");
- nmi_watchdog = NMI_NONE;
- }
-#ifdef CONFIG_X86_32
- timer_ack = 0;
-#endif
-
apic_printk(APIC_QUIET, KERN_INFO
"...trying to set up timer as Virtual Wire IRQ...\n");
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index ed63101..1f129a1 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -330,9 +330,6 @@ static bool reserve_pmc_hardware(void)
{
int i;
- if (nmi_watchdog == NMI_LOCAL_APIC)
- disable_lapic_nmi_watchdog();
-
for (i = 0; i < x86_pmu.num_counters; i++) {
if (!reserve_perfctr_nmi(x86_pmu.perfctr + i))
goto perfctr_fail;
@@ -355,9 +352,6 @@ perfctr_fail:
for (i--; i >= 0; i--)
release_perfctr_nmi(x86_pmu.perfctr + i);
- if (nmi_watchdog == NMI_LOCAL_APIC)
- enable_lapic_nmi_watchdog();
-
return false;
}
@@ -369,9 +363,6 @@ static void release_pmc_hardware(void)
release_perfctr_nmi(x86_pmu.perfctr + i);
release_evntsel_nmi(x86_pmu.eventsel + i);
}
-
- if (nmi_watchdog == NMI_LOCAL_APIC)
- enable_lapic_nmi_watchdog();
}
#else
diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c
index d9f4ff8..14d4592 100644
--- a/arch/x86/kernel/cpu/perfctr-watchdog.c
+++ b/arch/x86/kernel/cpu/perfctr-watchdog.c
@@ -22,26 +22,6 @@
#include <asm/apic.h>
#include <asm/perf_event.h>
-struct nmi_watchdog_ctlblk {
- unsigned int cccr_msr;
- unsigned int perfctr_msr; /* the MSR to reset in NMI handler */
- unsigned int evntsel_msr; /* the MSR to select the events to handle */
-};
-
-/* Interface defining a CPU specific perfctr watchdog */
-struct wd_ops {
- int (*reserve)(void);
- void (*unreserve)(void);
- int (*setup)(unsigned nmi_hz);
- void (*rearm)(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz);
- void (*stop)(void);
- unsigned perfctr;
- unsigned evntsel;
- u64 checkbit;
-};
-
-static const struct wd_ops *wd_ops;
-
/*
* this number is calculated from Intel's MSR_P4_CRU_ESCR5 register and it's
* offset from MSR_P4_BSU_ESCR0.
@@ -60,8 +40,6 @@ static const struct wd_ops *wd_ops;
static DECLARE_BITMAP(perfctr_nmi_owner, NMI_MAX_COUNTER_BITS);
static DECLARE_BITMAP(evntsel_nmi_owner, NMI_MAX_COUNTER_BITS);
-static DEFINE_PER_CPU(struct nmi_watchdog_ctlblk, nmi_watchdog_ctlblk);
-
/* converts an msr to an appropriate reservation bit */
static inline unsigned int nmi_perfctr_msr_to_bit(unsigned int msr)
{
@@ -172,623 +150,3 @@ void release_evntsel_nmi(unsigned int msr)
clear_bit(counter, evntsel_nmi_owner);
}
EXPORT_SYMBOL(release_evntsel_nmi);
-
-void disable_lapic_nmi_watchdog(void)
-{
- BUG_ON(nmi_watchdog != NMI_LOCAL_APIC);
-
- if (atomic_read(&nmi_active) <= 0)
- return;
-
- on_each_cpu(stop_apic_nmi_watchdog, NULL, 1);
-
- if (wd_ops)
- wd_ops->unreserve();
-
- BUG_ON(atomic_read(&nmi_active) != 0);
-}
-
-void enable_lapic_nmi_watchdog(void)
-{
- BUG_ON(nmi_watchdog != NMI_LOCAL_APIC);
-
- /* are we already enabled */
- if (atomic_read(&nmi_active) != 0)
- return;
-
- /* are we lapic aware */
- if (!wd_ops)
- return;
- if (!wd_ops->reserve()) {
- printk(KERN_ERR "NMI watchdog: cannot reserve perfctrs\n");
- return;
- }
-
- on_each_cpu(setup_apic_nmi_watchdog, NULL, 1);
- touch_nmi_watchdog();
-}
-
-/*
- * Activate the NMI watchdog via the local APIC.
- */
-
-static unsigned int adjust_for_32bit_ctr(unsigned int hz)
-{
- u64 counter_val;
- unsigned int retval = hz;
-
- /*
- * On Intel CPUs with P6/ARCH_PERFMON only 32 bits in the counter
- * are writable, with higher bits sign extending from bit 31.
- * So, we can only program the counter with 31 bit values and
- * 32nd bit should be 1, for 33.. to be 1.
- * Find the appropriate nmi_hz
- */
- counter_val = (u64)cpu_khz * 1000;
- do_div(counter_val, retval);
- if (counter_val > 0x7fffffffULL) {
- u64 count = (u64)cpu_khz * 1000;
- do_div(count, 0x7fffffffUL);
- retval = count + 1;
- }
- return retval;
-}
-
-static void write_watchdog_counter(unsigned int perfctr_msr,
- const char *descr, unsigned nmi_hz)
-{
- u64 count = (u64)cpu_khz * 1000;
-
- do_div(count, nmi_hz);
- if (descr)
- pr_debug("setting %s to -0x%08Lx\n", descr, count);
- wrmsrl(perfctr_msr, 0 - count);
-}
-
-static void write_watchdog_counter32(unsigned int perfctr_msr,
- const char *descr, unsigned nmi_hz)
-{
- u64 count = (u64)cpu_khz * 1000;
-
- do_div(count, nmi_hz);
- if (descr)
- pr_debug("setting %s to -0x%08Lx\n", descr, count);
- wrmsr(perfctr_msr, (u32)(-count), 0);
-}
-
-/*
- * AMD K7/K8/Family10h/Family11h support.
- * AMD keeps this interface nicely stable so there is not much variety
- */
-#define K7_EVNTSEL_ENABLE (1 << 22)
-#define K7_EVNTSEL_INT (1 << 20)
-#define K7_EVNTSEL_OS (1 << 17)
-#define K7_EVNTSEL_USR (1 << 16)
-#define K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING 0x76
-#define K7_NMI_EVENT K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING
-
-static int setup_k7_watchdog(unsigned nmi_hz)
-{
- unsigned int perfctr_msr, evntsel_msr;
- unsigned int evntsel;
- struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
-
- perfctr_msr = wd_ops->perfctr;
- evntsel_msr = wd_ops->evntsel;
-
- wrmsrl(perfctr_msr, 0UL);
-
- evntsel = K7_EVNTSEL_INT
- | K7_EVNTSEL_OS
- | K7_EVNTSEL_USR
- | K7_NMI_EVENT;
-
- /* setup the timer */
- wrmsr(evntsel_msr, evntsel, 0);
- write_watchdog_counter(perfctr_msr, "K7_PERFCTR0", nmi_hz);
-
- /* initialize the wd struct before enabling */
- wd->perfctr_msr = perfctr_msr;
- wd->evntsel_msr = evntsel_msr;
- wd->cccr_msr = 0; /* unused */
-
- /* ok, everything is initialized, announce that we're set */
- cpu_nmi_set_wd_enabled();
-
- apic_write(APIC_LVTPC, APIC_DM_NMI);
- evntsel |= K7_EVNTSEL_ENABLE;
- wrmsr(evntsel_msr, evntsel, 0);
-
- return 1;
-}
-
-static void single_msr_stop_watchdog(void)
-{
- struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
-
- wrmsr(wd->evntsel_msr, 0, 0);
-}
-
-static int single_msr_reserve(void)
-{
- if (!reserve_perfctr_nmi(wd_ops->perfctr))
- return 0;
-
- if (!reserve_evntsel_nmi(wd_ops->evntsel)) {
- release_perfctr_nmi(wd_ops->perfctr);
- return 0;
- }
- return 1;
-}
-
-static void single_msr_unreserve(void)
-{
- release_evntsel_nmi(wd_ops->evntsel);
- release_perfctr_nmi(wd_ops->perfctr);
-}
-
-static void __kprobes
-single_msr_rearm(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz)
-{
- /* start the cycle over again */
- write_watchdog_counter(wd->perfctr_msr, NULL, nmi_hz);
-}
-
-static const struct wd_ops k7_wd_ops = {
- .reserve = single_msr_reserve,
- .unreserve = single_msr_unreserve,
- .setup = setup_k7_watchdog,
- .rearm = single_msr_rearm,
- .stop = single_msr_stop_watchdog,
- .perfctr = MSR_K7_PERFCTR0,
- .evntsel = MSR_K7_EVNTSEL0,
- .checkbit = 1ULL << 47,
-};
-
-/*
- * Intel Model 6 (PPro+,P2,P3,P-M,Core1)
- */
-#define P6_EVNTSEL0_ENABLE (1 << 22)
-#define P6_EVNTSEL_INT (1 << 20)
-#define P6_EVNTSEL_OS (1 << 17)
-#define P6_EVNTSEL_USR (1 << 16)
-#define P6_EVENT_CPU_CLOCKS_NOT_HALTED 0x79
-#define P6_NMI_EVENT P6_EVENT_CPU_CLOCKS_NOT_HALTED
-
-static int setup_p6_watchdog(unsigned nmi_hz)
-{
- unsigned int perfctr_msr, evntsel_msr;
- unsigned int evntsel;
- struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
-
- perfctr_msr = wd_ops->perfctr;
- evntsel_msr = wd_ops->evntsel;
-
- /* KVM doesn't implement this MSR */
- if (wrmsr_safe(perfctr_msr, 0, 0) < 0)
- return 0;
-
- evntsel = P6_EVNTSEL_INT
- | P6_EVNTSEL_OS
- | P6_EVNTSEL_USR
- | P6_NMI_EVENT;
-
- /* setup the timer */
- wrmsr(evntsel_msr, evntsel, 0);
- nmi_hz = adjust_for_32bit_ctr(nmi_hz);
- write_watchdog_counter32(perfctr_msr, "P6_PERFCTR0", nmi_hz);
-
- /* initialize the wd struct before enabling */
- wd->perfctr_msr = perfctr_msr;
- wd->evntsel_msr = evntsel_msr;
- wd->cccr_msr = 0; /* unused */
-
- /* ok, everything is initialized, announce that we're set */
- cpu_nmi_set_wd_enabled();
-
- apic_write(APIC_LVTPC, APIC_DM_NMI);
- evntsel |= P6_EVNTSEL0_ENABLE;
- wrmsr(evntsel_msr, evntsel, 0);
-
- return 1;
-}
-
-static void __kprobes p6_rearm(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz)
-{
- /*
- * P6 based Pentium M need to re-unmask
- * the apic vector but it doesn't hurt
- * other P6 variant.
- * ArchPerfom/Core Duo also needs this
- */
- apic_write(APIC_LVTPC, APIC_DM_NMI);
-
- /* P6/ARCH_PERFMON has 32 bit counter write */
- write_watchdog_counter32(wd->perfctr_msr, NULL, nmi_hz);
-}
-
-static const struct wd_ops p6_wd_ops = {
- .reserve = single_msr_reserve,
- .unreserve = single_msr_unreserve,
- .setup = setup_p6_watchdog,
- .rearm = p6_rearm,
- .stop = single_msr_stop_watchdog,
- .perfctr = MSR_P6_PERFCTR0,
- .evntsel = MSR_P6_EVNTSEL0,
- .checkbit = 1ULL << 39,
-};
-
-/*
- * Intel P4 performance counters.
- * By far the most complicated of all.
- */
-#define MSR_P4_MISC_ENABLE_PERF_AVAIL (1 << 7)
-#define P4_ESCR_EVENT_SELECT(N) ((N) << 25)
-#define P4_ESCR_OS (1 << 3)
-#define P4_ESCR_USR (1 << 2)
-#define P4_CCCR_OVF_PMI0 (1 << 26)
-#define P4_CCCR_OVF_PMI1 (1 << 27)
-#define P4_CCCR_THRESHOLD(N) ((N) << 20)
-#define P4_CCCR_COMPLEMENT (1 << 19)
-#define P4_CCCR_COMPARE (1 << 18)
-#define P4_CCCR_REQUIRED (3 << 16)
-#define P4_CCCR_ESCR_SELECT(N) ((N) << 13)
-#define P4_CCCR_ENABLE (1 << 12)
-#define P4_CCCR_OVF (1 << 31)
-
-#define P4_CONTROLS 18
-static unsigned int p4_controls[18] = {
- MSR_P4_BPU_CCCR0,
- MSR_P4_BPU_CCCR1,
- MSR_P4_BPU_CCCR2,
- MSR_P4_BPU_CCCR3,
- MSR_P4_MS_CCCR0,
- MSR_P4_MS_CCCR1,
- MSR_P4_MS_CCCR2,
- MSR_P4_MS_CCCR3,
- MSR_P4_FLAME_CCCR0,
- MSR_P4_FLAME_CCCR1,
- MSR_P4_FLAME_CCCR2,
- MSR_P4_FLAME_CCCR3,
- MSR_P4_IQ_CCCR0,
- MSR_P4_IQ_CCCR1,
- MSR_P4_IQ_CCCR2,
- MSR_P4_IQ_CCCR3,
- MSR_P4_IQ_CCCR4,
- MSR_P4_IQ_CCCR5,
-};
-/*
- * Set up IQ_COUNTER0 to behave like a clock, by having IQ_CCCR0 filter
- * CRU_ESCR0 (with any non-null event selector) through a complemented
- * max threshold. [IA32-Vol3, Section 14.9.9]
- */
-static int setup_p4_watchdog(unsigned nmi_hz)
-{
- unsigned int perfctr_msr, evntsel_msr, cccr_msr;
- unsigned int evntsel, cccr_val;
- unsigned int misc_enable, dummy;
- unsigned int ht_num;
- struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
-
- rdmsr(MSR_IA32_MISC_ENABLE, misc_enable, dummy);
- if (!(misc_enable & MSR_P4_MISC_ENABLE_PERF_AVAIL))
- return 0;
-
-#ifdef CONFIG_SMP
- /* detect which hyperthread we are on */
- if (smp_num_siblings == 2) {
- unsigned int ebx, apicid;
-
- ebx = cpuid_ebx(1);
- apicid = (ebx >> 24) & 0xff;
- ht_num = apicid & 1;
- } else
-#endif
- ht_num = 0;
-
- /*
- * performance counters are shared resources
- * assign each hyperthread its own set
- * (re-use the ESCR0 register, seems safe
- * and keeps the cccr_val the same)
- */
- if (!ht_num) {
- /* logical cpu 0 */
- perfctr_msr = MSR_P4_IQ_PERFCTR0;
- evntsel_msr = MSR_P4_CRU_ESCR0;
- cccr_msr = MSR_P4_IQ_CCCR0;
- cccr_val = P4_CCCR_OVF_PMI0 | P4_CCCR_ESCR_SELECT(4);
-
- /*
- * If we're on the kdump kernel or other situation, we may
- * still have other performance counter registers set to
- * interrupt and they'll keep interrupting forever because
- * of the P4_CCCR_OVF quirk. So we need to ACK all the
- * pending interrupts and disable all the registers here,
- * before reenabling the NMI delivery. Refer to p4_rearm()
- * about the P4_CCCR_OVF quirk.
- */
- if (reset_devices) {
- unsigned int low, high;
- int i;
-
- for (i = 0; i < P4_CONTROLS; i++) {
- rdmsr(p4_controls[i], low, high);
- low &= ~(P4_CCCR_ENABLE | P4_CCCR_OVF);
- wrmsr(p4_controls[i], low, high);
- }
- }
- } else {
- /* logical cpu 1 */
- perfctr_msr = MSR_P4_IQ_PERFCTR1;
- evntsel_msr = MSR_P4_CRU_ESCR0;
- cccr_msr = MSR_P4_IQ_CCCR1;
-
- /* Pentium 4 D processors don't support P4_CCCR_OVF_PMI1 */
- if (boot_cpu_data.x86_model == 4 && boot_cpu_data.x86_mask == 4)
- cccr_val = P4_CCCR_OVF_PMI0;
- else
- cccr_val = P4_CCCR_OVF_PMI1;
- cccr_val |= P4_CCCR_ESCR_SELECT(4);
- }
-
- evntsel = P4_ESCR_EVENT_SELECT(0x3F)
- | P4_ESCR_OS
- | P4_ESCR_USR;
-
- cccr_val |= P4_CCCR_THRESHOLD(15)
- | P4_CCCR_COMPLEMENT
- | P4_CCCR_COMPARE
- | P4_CCCR_REQUIRED;
-
- wrmsr(evntsel_msr, evntsel, 0);
- wrmsr(cccr_msr, cccr_val, 0);
- write_watchdog_counter(perfctr_msr, "P4_IQ_COUNTER0", nmi_hz);
-
- wd->perfctr_msr = perfctr_msr;
- wd->evntsel_msr = evntsel_msr;
- wd->cccr_msr = cccr_msr;
-
- /* ok, everything is initialized, announce that we're set */
- cpu_nmi_set_wd_enabled();
-
- apic_write(APIC_LVTPC, APIC_DM_NMI);
- cccr_val |= P4_CCCR_ENABLE;
- wrmsr(cccr_msr, cccr_val, 0);
- return 1;
-}
-
-static void stop_p4_watchdog(void)
-{
- struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
- wrmsr(wd->cccr_msr, 0, 0);
- wrmsr(wd->evntsel_msr, 0, 0);
-}
-
-static int p4_reserve(void)
-{
- if (!reserve_perfctr_nmi(MSR_P4_IQ_PERFCTR0))
- return 0;
-#ifdef CONFIG_SMP
- if (smp_num_siblings > 1 && !reserve_perfctr_nmi(MSR_P4_IQ_PERFCTR1))
- goto fail1;
-#endif
- if (!reserve_evntsel_nmi(MSR_P4_CRU_ESCR0))
- goto fail2;
- /* RED-PEN why is ESCR1 not reserved here? */
- return 1;
- fail2:
-#ifdef CONFIG_SMP
- if (smp_num_siblings > 1)
- release_perfctr_nmi(MSR_P4_IQ_PERFCTR1);
- fail1:
-#endif
- release_perfctr_nmi(MSR_P4_IQ_PERFCTR0);
- return 0;
-}
-
-static void p4_unreserve(void)
-{
-#ifdef CONFIG_SMP
- if (smp_num_siblings > 1)
- release_perfctr_nmi(MSR_P4_IQ_PERFCTR1);
-#endif
- release_evntsel_nmi(MSR_P4_CRU_ESCR0);
- release_perfctr_nmi(MSR_P4_IQ_PERFCTR0);
-}
-
-static void __kprobes p4_rearm(struct nmi_watchdog_ctlblk *wd, unsigned nmi_hz)
-{
- unsigned dummy;
- /*
- * P4 quirks:
- * - An overflown perfctr will assert its interrupt
- * until the OVF flag in its CCCR is cleared.
- * - LVTPC is masked on interrupt and must be
- * unmasked by the LVTPC handler.
- */
- rdmsrl(wd->cccr_msr, dummy);
- dummy &= ~P4_CCCR_OVF;
- wrmsrl(wd->cccr_msr, dummy);
- apic_write(APIC_LVTPC, APIC_DM_NMI);
- /* start the cycle over again */
- write_watchdog_counter(wd->perfctr_msr, NULL, nmi_hz);
-}
-
-static const struct wd_ops p4_wd_ops = {
- .reserve = p4_reserve,
- .unreserve = p4_unreserve,
- .setup = setup_p4_watchdog,
- .rearm = p4_rearm,
- .stop = stop_p4_watchdog,
- /* RED-PEN this is wrong for the other sibling */
- .perfctr = MSR_P4_BPU_PERFCTR0,
- .evntsel = MSR_P4_BSU_ESCR0,
- .checkbit = 1ULL << 39,
-};
-
-/*
- * Watchdog using the Intel architected PerfMon.
- * Used for Core2 and hopefully all future Intel CPUs.
- */
-#define ARCH_PERFMON_NMI_EVENT_SEL ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL
-#define ARCH_PERFMON_NMI_EVENT_UMASK ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK
-
-static struct wd_ops intel_arch_wd_ops;
-
-static int setup_intel_arch_watchdog(unsigned nmi_hz)
-{
- unsigned int ebx;
- union cpuid10_eax eax;
- unsigned int unused;
- unsigned int perfctr_msr, evntsel_msr;
- unsigned int evntsel;
- struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
-
- /*
- * Check whether the Architectural PerfMon supports
- * Unhalted Core Cycles Event or not.
- * NOTE: Corresponding bit = 0 in ebx indicates event present.
- */
- cpuid(10, &(eax.full), &ebx, &unused, &unused);
- if ((eax.split.mask_length <
- (ARCH_PERFMON_UNHALTED_CORE_CYCLES_INDEX+1)) ||
- (ebx & ARCH_PERFMON_UNHALTED_CORE_CYCLES_PRESENT))
- return 0;
-
- perfctr_msr = wd_ops->perfctr;
- evntsel_msr = wd_ops->evntsel;
-
- wrmsrl(perfctr_msr, 0UL);
-
- evntsel = ARCH_PERFMON_EVENTSEL_INT
- | ARCH_PERFMON_EVENTSEL_OS
- | ARCH_PERFMON_EVENTSEL_USR
- | ARCH_PERFMON_NMI_EVENT_SEL
- | ARCH_PERFMON_NMI_EVENT_UMASK;
-
- /* setup the timer */
- wrmsr(evntsel_msr, evntsel, 0);
- nmi_hz = adjust_for_32bit_ctr(nmi_hz);
- write_watchdog_counter32(perfctr_msr, "INTEL_ARCH_PERFCTR0", nmi_hz);
-
- wd->perfctr_msr = perfctr_msr;
- wd->evntsel_msr = evntsel_msr;
- wd->cccr_msr = 0; /* unused */
-
- /* ok, everything is initialized, announce that we're set */
- cpu_nmi_set_wd_enabled();
-
- apic_write(APIC_LVTPC, APIC_DM_NMI);
- evntsel |= ARCH_PERFMON_EVENTSEL_ENABLE;
- wrmsr(evntsel_msr, evntsel, 0);
- intel_arch_wd_ops.checkbit = 1ULL << (eax.split.bit_width - 1);
- return 1;
-}
-
-static struct wd_ops intel_arch_wd_ops __read_mostly = {
- .reserve = single_msr_reserve,
- .unreserve = single_msr_unreserve,
- .setup = setup_intel_arch_watchdog,
- .rearm = p6_rearm,
- .stop = single_msr_stop_watchdog,
- .perfctr = MSR_ARCH_PERFMON_PERFCTR1,
- .evntsel = MSR_ARCH_PERFMON_EVENTSEL1,
-};
-
-static void probe_nmi_watchdog(void)
-{
- switch (boot_cpu_data.x86_vendor) {
- case X86_VENDOR_AMD:
- if (boot_cpu_data.x86 == 6 ||
- (boot_cpu_data.x86 >= 0xf && boot_cpu_data.x86 <= 0x15))
- wd_ops = &k7_wd_ops;
- return;
- case X86_VENDOR_INTEL:
- /* Work around where perfctr1 doesn't have a working enable
- * bit as described in the following errata:
- * AE49 Core Duo and Intel Core Solo 65 nm
- * AN49 Intel Pentium Dual-Core
- * AF49 Dual-Core Intel Xeon Processor LV
- */
- if ((boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model == 14) ||
- ((boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model == 15 &&
- boot_cpu_data.x86_mask == 4))) {
- intel_arch_wd_ops.perfctr = MSR_ARCH_PERFMON_PERFCTR0;
- intel_arch_wd_ops.evntsel = MSR_ARCH_PERFMON_EVENTSEL0;
- }
- if (cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) {
- wd_ops = &intel_arch_wd_ops;
- break;
- }
- switch (boot_cpu_data.x86) {
- case 6:
- if (boot_cpu_data.x86_model > 13)
- return;
-
- wd_ops = &p6_wd_ops;
- break;
- case 15:
- wd_ops = &p4_wd_ops;
- break;
- default:
- return;
- }
- break;
- }
-}
-
-/* Interface to nmi.c */
-
-int lapic_watchdog_init(unsigned nmi_hz)
-{
- if (!wd_ops) {
- probe_nmi_watchdog();
- if (!wd_ops) {
- printk(KERN_INFO "NMI watchdog: CPU not supported\n");
- return -1;
- }
-
- if (!wd_ops->reserve()) {
- printk(KERN_ERR
- "NMI watchdog: cannot reserve perfctrs\n");
- return -1;
- }
- }
-
- if (!(wd_ops->setup(nmi_hz))) {
- printk(KERN_ERR "Cannot setup NMI watchdog on CPU %d\n",
- raw_smp_processor_id());
- return -1;
- }
-
- return 0;
-}
-
-void lapic_watchdog_stop(void)
-{
- if (wd_ops)
- wd_ops->stop();
-}
-
-unsigned lapic_adjust_nmi_hz(unsigned hz)
-{
- struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
- if (wd->perfctr_msr == MSR_P6_PERFCTR0 ||
- wd->perfctr_msr == MSR_ARCH_PERFMON_PERFCTR1)
- hz = adjust_for_32bit_ctr(hz);
- return hz;
-}
-
-int __kprobes lapic_wd_event(unsigned nmi_hz)
-{
- struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
- u64 ctr;
-
- rdmsrl(wd->perfctr_msr, ctr);
- if (ctr & wd_ops->checkbit) /* perfctr still running? */
- return 0;
-
- wd_ops->rearm(wd, nmi_hz);
- return 1;
-}
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 083e99d..f0a0624 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -316,12 +316,6 @@ notrace static void __cpuinit start_secondary(void *unused)
*/
check_tsc_sync_target();
- if (nmi_watchdog == NMI_IO_APIC) {
- legacy_pic->mask(0);
- enable_NMI_through_LVT0();
- legacy_pic->unmask(0);
- }
-
/* This must be done before setting cpu_online_mask */
set_cpu_sibling_map(raw_smp_processor_id());
wmb();
@@ -1061,8 +1055,6 @@ static int __init smp_sanity_check(unsigned max_cpus)
printk(KERN_INFO "SMP mode deactivated.\n");
smpboot_clear_io_apic();
- localise_nmi_watchdog();
-
connect_bsp_APIC();
setup_local_APIC();
end_local_APIC_setup();
@@ -1196,7 +1188,6 @@ void __init native_smp_cpus_done(unsigned int max_cpus)
#ifdef CONFIG_X86_IO_APIC
setup_ioapic_dest();
#endif
- check_nmi_watchdog();
mtrr_aps_init();
}
@@ -1341,8 +1332,6 @@ int native_cpu_disable(void)
if (cpu == 0)
return -EBUSY;
- if (nmi_watchdog == NMI_LOCAL_APIC)
- stop_apic_nmi_watchdog(NULL);
clear_local_APIC();
cpu_disable_common();
diff --git a/arch/x86/kernel/time.c b/arch/x86/kernel/time.c
index fb5cc5e1..25a28a2 100644
--- a/arch/x86/kernel/time.c
+++ b/arch/x86/kernel/time.c
@@ -22,10 +22,6 @@
#include <asm/hpet.h>
#include <asm/time.h>
-#if defined(CONFIG_X86_32) && defined(CONFIG_X86_IO_APIC)
-int timer_ack;
-#endif
-
#ifdef CONFIG_X86_64
volatile unsigned long __jiffies __section_jiffies = INITIAL_JIFFIES;
#endif
@@ -63,20 +59,6 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id)
/* Keep nmi watchdog up to date */
inc_irq_stat(irq0_irqs);
- /* Optimized out for !IO_APIC and x86_64 */
- if (timer_ack) {
- /*
- * Subtle, when I/O APICs are used we have to ack timer IRQ
- * manually to deassert NMI lines for the watchdog if run
- * on an 82489DX-based system.
- */
- raw_spin_lock(&i8259A_lock);
- outb(0x0c, PIC_MASTER_OCW3);
- /* Ack the IRQ; AEOI will end it automatically. */
- inb(PIC_MASTER_POLL);
- raw_spin_unlock(&i8259A_lock);
- }
-
global_clock_event->event_handler(global_clock_event);
/* MCA bus quirk: Acknowledge irq0 by setting bit 7 in port 0x61 */
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 4b331ce..c40e9a0 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -444,14 +444,12 @@ do_nmi(struct pt_regs *regs, long error_code)
void stop_nmi(void)
{
- acpi_nmi_disable();
ignore_nmis++;
}
void restart_nmi(void)
{
ignore_nmis--;
- acpi_nmi_enable();
}
/* May run on IST stack. */
diff --git a/arch/x86/oprofile/nmi_timer_int.c b/arch/x86/oprofile/nmi_timer_int.c
index e3ecb71..0636dd9 100644
--- a/arch/x86/oprofile/nmi_timer_int.c
+++ b/arch/x86/oprofile/nmi_timer_int.c
@@ -58,9 +58,6 @@ static void timer_stop(void)
int __init op_nmi_timer_init(struct oprofile_operations *ops)
{
- if ((nmi_watchdog != NMI_IO_APIC) || (atomic_read(&nmi_active) <= 0))
- return -ENODEV;
-
ops->start = timer_start;
ops->stop = timer_stop;
ops->cpu_type = "timer";
diff --git a/drivers/acpi/acpica/nsinit.c b/drivers/acpi/acpica/nsinit.c
index 660a272..0cac7ec 100644
--- a/drivers/acpi/acpica/nsinit.c
+++ b/drivers/acpi/acpica/nsinit.c
@@ -577,9 +577,7 @@ acpi_ns_init_one_device(acpi_handle obj_handle,
* as possible (without an NMI being received in the middle of
* this) - so disable NMIs and initialize the device:
*/
- acpi_nmi_disable();
status = acpi_ns_evaluate(info);
- acpi_nmi_enable();
if (ACPI_SUCCESS(status)) {
walk_info->num_INI++;
diff --git a/drivers/watchdog/hpwdt.c b/drivers/watchdog/hpwdt.c
index 3d77116..c19f4a2 100644
--- a/drivers/watchdog/hpwdt.c
+++ b/drivers/watchdog/hpwdt.c
@@ -649,12 +649,7 @@ static void __devinit hpwdt_check_nmi_decoding(struct pci_dev *dev)
* If nmi_watchdog is turned off then we can turn on
* our nmi decoding capability.
*/
- if (!nmi_watchdog_active())
- hpwdt_nmi_decoding = 1;
- else
- dev_warn(&dev->dev, "NMI decoding is disabled. To enable this "
- "functionality you must reboot with nmi_watchdog=0 "
- "and load the hpwdt driver with priority=1.\n");
+ hpwdt_nmi_decoding = 1;
}
#else
static void __devinit hpwdt_check_nmi_decoding(struct pci_dev *dev)
diff --git a/include/linux/nmi.h b/include/linux/nmi.h
index 06aab5e..17ccf44 100644
--- a/include/linux/nmi.h
+++ b/include/linux/nmi.h
@@ -17,8 +17,6 @@
#ifdef ARCH_HAS_NMI_WATCHDOG
#include <asm/nmi.h>
extern void touch_nmi_watchdog(void);
-extern void acpi_nmi_disable(void);
-extern void acpi_nmi_enable(void);
#else
#ifndef CONFIG_HARDLOCKUP_DETECTOR
static inline void touch_nmi_watchdog(void)
@@ -28,8 +26,6 @@ static inline void touch_nmi_watchdog(void)
#else
extern void touch_nmi_watchdog(void);
#endif
-static inline void acpi_nmi_disable(void) { }
-static inline void acpi_nmi_enable(void) { }
#endif
/*
--
1.7.3.2
^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH 0/3 v2] x86: Remove old nmi_watchdog
2010-11-23 19:32 [PATCH 0/3 v2] x86: Remove old nmi_watchdog Don Zickus
` (2 preceding siblings ...)
2010-11-23 19:32 ` [PATCH 3/3] x86, nmi_watchdog: remove all stub function calls from " Don Zickus
@ 2010-11-23 21:49 ` Cyrill Gorcunov
2010-11-24 14:03 ` Don Zickus
3 siblings, 1 reply; 9+ messages in thread
From: Cyrill Gorcunov @ 2010-11-23 21:49 UTC (permalink / raw)
To: Don Zickus; +Cc: Ingo Molnar, fweisbec, LKML, yinghai
On Tue, Nov 23, 2010 at 02:32:02PM -0500, Don Zickus wrote:
> Ingo asked me to remove the old nmi_watchdog to minimize the amount of
> code duplication we have to do for recent fixes. All the functionality
> needed by the old nmi watchdog should be available in perf now.
>
Hi Don, if I understand correctly not many things are changed from previous
time, some tune ups here and there, right?
Cyrill
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 2/3] x86, nmi_watchdog: remove the old nmi_watchdog
2010-11-23 19:32 ` [PATCH 2/3] x86, nmi_watchdog: remove the old nmi_watchdog Don Zickus
@ 2010-11-24 3:23 ` Yinghai Lu
2010-11-24 14:08 ` Don Zickus
0 siblings, 1 reply; 9+ messages in thread
From: Yinghai Lu @ 2010-11-24 3:23 UTC (permalink / raw)
To: Don Zickus; +Cc: Ingo Molnar, fweisbec, gorcunov, LKML
On 11/23/2010 11:32 AM, Don Zickus wrote:
> Now that we have a new nmi_watchdog that is more generic and sits on
> top of the perf subsystem, we really do not need the old nmi_watchdog
> any more.
>
> In addition, the old nmi_watchdog doesn't really work if you are using
> the default clocksource, hpet. The old nmi_watchdog code relied on
> local apic interrupts to determine if the cpu is still alive. With
> hpet as the clocksource, these interrupts don't increment any more
> and the old nmi_watchdog triggers false postives.
>
> This piece removes the old nmi_watchdog code and stubs out any variables
> and functions calls. The stubs are the same ones used by the new
> nmi_watchdog code, so it should be well tested.
>
> v2: fix sparc breakage by undoing linux/nmi.h changes
> clean up documentation and sysctl stuff
>
> Signed-off-by: Don Zickus <dzickus@redhat.com>
> ---
> Documentation/kernel-parameters.txt | 10 +-
> arch/x86/include/asm/nmi.h | 4 -
> arch/x86/kernel/apic/Makefile | 5 +-
> arch/x86/kernel/apic/hw_nmi.c | 14 +-
> arch/x86/kernel/apic/nmi.c | 540 -----------------------------------
> arch/x86/kernel/traps.c | 8 -
> kernel/sysctl.c | 9 +-
> kernel/sysctl_binary.c | 1 -
> kernel/watchdog.c | 2 +
> 9 files changed, 16 insertions(+), 577 deletions(-)
> delete mode 100644 arch/x86/kernel/apic/nmi.c
>
> diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
> index ed45e98..e48b93f 100644
> --- a/Documentation/kernel-parameters.txt
> +++ b/Documentation/kernel-parameters.txt
> @@ -1579,20 +1579,12 @@ and is between 256 and 4096 characters. It is defined in the file
>
> nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
> Format: [panic,][num]
> - Valid num: 0,1,2
> + Valid num: 0
> 0 - turn nmi_watchdog off
thanks for keeping nmi_watchdog=0
> - 1 - use the IO-APIC timer for the NMI watchdog
> - 2 - use the local APIC for the NMI watchdog using
> - a performance counter. Note: This will use one
> - performance counter and the local APIC's performance
> - vector.
> When panic is specified, panic when an NMI watchdog
> timeout occurs.
> This is useful when you use a panic=... timeout and
> need the box quickly up again.
> - Instead of 1 and 2 it is possible to use the following
> - symbolic names: lapic and ioapic
> - Example: nmi_watchdog=2 or nmi_watchdog=panic,lapic
>
> netpoll.carrier_timeout=
> [NET] Specifies amount of time (in seconds) that
...
> diff --git a/kernel/sysctl.c b/kernel/sysctl.c
> index c33a1ed..0daad4e 100644
> --- a/kernel/sysctl.c
> +++ b/kernel/sysctl.c
> @@ -737,7 +737,7 @@ static struct ctl_table kern_table[] = {
> .extra2 = &one,
> },
> #endif
> -#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86) && !defined(CONFIG_LOCKUP_DETECTOR)
> +#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86)
> {
> .procname = "unknown_nmi_panic",
> .data = &unknown_nmi_panic,
> @@ -745,13 +745,6 @@ static struct ctl_table kern_table[] = {
> .mode = 0644,
> .proc_handler = proc_dointvec,
> },
> - {
> - .procname = "nmi_watchdog",
> - .data = &nmi_watchdog_enabled,
> - .maxlen = sizeof (int),
> - .mode = 0644,
> - .proc_handler = proc_nmi_enabled,
> - },
wonder if you can keep nmi_watchdog in sysctl? So in run-time it could be disabled after it booted up.
> #endif
> #if defined(CONFIG_X86)
> {
> diff --git a/kernel/sysctl_binary.c b/kernel/sysctl_binary.c
> index 1357c57..4b2545a 100644
> --- a/kernel/sysctl_binary.c
> +++ b/kernel/sysctl_binary.c
> @@ -136,7 +136,6 @@ static const struct bin_table bin_kern_table[] = {
> { CTL_INT, KERN_IA64_UNALIGNED, "ignore-unaligned-usertrap" },
> { CTL_INT, KERN_COMPAT_LOG, "compat-log" },
> { CTL_INT, KERN_MAX_LOCK_DEPTH, "max_lock_depth" },
> - { CTL_INT, KERN_NMI_WATCHDOG, "nmi_watchdog" },
> { CTL_INT, KERN_PANIC_ON_NMI, "panic_on_unrecovered_nmi" },
> {}
> };
> diff --git a/kernel/watchdog.c b/kernel/watchdog.c
> index bafba68..3c5441f 100644
> --- a/kernel/watchdog.c
> +++ b/kernel/watchdog.c
> @@ -57,6 +57,8 @@ static int __init hardlockup_panic_setup(char *str)
> {
> if (!strncmp(str, "panic", 5))
> hardlockup_panic = 1;
> + else if (!strncmp(str, "0", 1))
> + no_watchdog = 1;
> return 1;
> }
> __setup("nmi_watchdog=", hardlockup_panic_setup);
Thanks
Yinghai
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 0/3 v2] x86: Remove old nmi_watchdog
2010-11-23 21:49 ` [PATCH 0/3 v2] x86: Remove " Cyrill Gorcunov
@ 2010-11-24 14:03 ` Don Zickus
0 siblings, 0 replies; 9+ messages in thread
From: Don Zickus @ 2010-11-24 14:03 UTC (permalink / raw)
To: Cyrill Gorcunov; +Cc: Ingo Molnar, fweisbec, LKML, yinghai
On Wed, Nov 24, 2010 at 12:49:57AM +0300, Cyrill Gorcunov wrote:
> On Tue, Nov 23, 2010 at 02:32:02PM -0500, Don Zickus wrote:
> > Ingo asked me to remove the old nmi_watchdog to minimize the amount of
> > code duplication we have to do for recent fixes. All the functionality
> > needed by the old nmi watchdog should be available in perf now.
> >
>
> Hi Don, if I understand correctly not many things are changed from previous
> time, some tune ups here and there, right?
Right. Just the addition of the first patch to preserve
unknown_nmi_panic, some compiler breakage fix, and documentation stuff.
Cheers,
Don
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 2/3] x86, nmi_watchdog: remove the old nmi_watchdog
2010-11-24 3:23 ` Yinghai Lu
@ 2010-11-24 14:08 ` Don Zickus
2010-11-24 18:12 ` Yinghai
0 siblings, 1 reply; 9+ messages in thread
From: Don Zickus @ 2010-11-24 14:08 UTC (permalink / raw)
To: Yinghai Lu; +Cc: Ingo Molnar, fweisbec, gorcunov, LKML
On Tue, Nov 23, 2010 at 07:23:39PM -0800, Yinghai Lu wrote:
> > - {
> > - .procname = "nmi_watchdog",
> > - .data = &nmi_watchdog_enabled,
> > - .maxlen = sizeof (int),
> > - .mode = 0644,
> > - .proc_handler = proc_nmi_enabled,
> > - },
>
> wonder if you can keep nmi_watchdog in sysctl? So in run-time it could be disabled after it booted up.
Well, with the new code I have watchdog_enabled which does the same thing.
With the nmi watchdog moving to generic code, the idea was to call it a
hardlockup detector. I guess I was trying to rename things to avoid using
nmi all over the place to reflect the idea that a hardlockup might be able
to be used by something other than an nmi.
I can probably wrap this around watchdog_enabled to preserve it I suppose.
Let me know.
Cheers,
Don
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 2/3] x86, nmi_watchdog: remove the old nmi_watchdog
2010-11-24 14:08 ` Don Zickus
@ 2010-11-24 18:12 ` Yinghai
0 siblings, 0 replies; 9+ messages in thread
From: Yinghai @ 2010-11-24 18:12 UTC (permalink / raw)
To: Don Zickus; +Cc: Ingo Molnar, fweisbec, gorcunov, LKML
On Nov 24, 2010, at 6:08 AM, Don Zickus <dzickus@redhat.com> wrote:
> On Tue, Nov 23, 2010 at 07:23:39PM -0800, Yinghai Lu wrote:
>>> - {
>>> - .procname = "nmi_watchdog",
>>> - .data = &nmi_watchdog_enabled,
>>> - .maxlen = sizeof (int),
>>> - .mode = 0644,
>>> - .proc_handler = proc_nmi_enabled,
>>> - },
>>
>> wonder if you can keep nmi_watchdog in sysctl? So in run-time it could be disabled after it booted up.
>
> Well, with the new code I have watchdog_enabled which does the same thing.
> With the nmi watchdog moving to generic code, the idea was to call it a
> hardlockup detector. I guess I was trying to rename things to avoid using
> nmi all over the place to reflect the idea that a hardlockup might be able
> to be used by something other than an nmi.
>
> I can probably wrap this around watchdog_enabled to preserve it I suppose.
> Let me know.
Yes that will be good
Thanks
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2010-11-24 18:13 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-11-23 19:32 [PATCH 0/3 v2] x86: Remove old nmi_watchdog Don Zickus
2010-11-23 19:32 ` [PATCH 1/3] x86, NMI: Remove do_nmi_callback logic Don Zickus
2010-11-23 19:32 ` [PATCH 2/3] x86, nmi_watchdog: remove the old nmi_watchdog Don Zickus
2010-11-24 3:23 ` Yinghai Lu
2010-11-24 14:08 ` Don Zickus
2010-11-24 18:12 ` Yinghai
2010-11-23 19:32 ` [PATCH 3/3] x86, nmi_watchdog: remove all stub function calls from " Don Zickus
2010-11-23 21:49 ` [PATCH 0/3 v2] x86: Remove " Cyrill Gorcunov
2010-11-24 14:03 ` Don Zickus
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).