static int __init check_nmi_watchdog(void)
{
unsigned int *prev_nmi_count;
int cpu;
if ((nmi_watchdog == NMI_NONE) || (nmi_watchdog == NMI_DISABLED))
return 0;
if (!atomic_read(&nmi_active))
return 0;
prev_nmi_count = kmalloc(NR_CPUS * sizeof(int), GFP_KERNEL);
if (!prev_nmi_count)
return -1;
printk(KERN_INFO "Testing NMI watchdog ... ");
if (nmi_watchdog == NMI_LOCAL_APIC)
smp_call_function(nmi_cpu_busy, (void *)&endflag, 0, 0);
for_each_possible_cpu(cpu)
prev_nmi_count[cpu] = per_cpu(irq_stat, cpu).__nmi_count;
local_irq_enable();
mdelay((20*1000)/nmi_hz); // wait 20 ticks
for_each_possible_cpu(cpu) {
#ifdef CONFIG_SMP
/* Check cpu_callin_map here because that is set
after the timer is started. */
if (!cpu_isset(cpu, cpu_callin_map))
continue;
#endif
if (!per_cpu(wd_enabled, cpu))
continue;
if (nmi_count(cpu) - prev_nmi_count[cpu] <= 5) {
printk(KERN_WARNING "WARNING: CPU#%d: NMI "
"appears to be stuck (%d->%d)!\n",
cpu,
prev_nmi_count[cpu],
nmi_count(cpu));
per_cpu(wd_enabled, cpu) = 0;
atomic_dec(&nmi_active);
}
}
endflag = 1;
if (!atomic_read(&nmi_active)) {
kfree(prev_nmi_count);
atomic_set(&nmi_active, -1);
return -1;
}
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;
}
static int __init check_nmi_watchdog(void)
{
volatile int endflag = 0;
unsigned int *prev_nmi_count;
int cpu;
if (nmi_watchdog == NMI_NONE)
return 0;
prev_nmi_count = kmalloc(NR_CPUS * sizeof(int), GFP_KERNEL);
if (!prev_nmi_count)
return -1;
printk(KERN_INFO "Testing NMI watchdog ... ");
if (nmi_watchdog == NMI_LOCAL_APIC)
smp_call_function(nmi_cpu_busy, (void *)&endflag, 0, 0);
for (cpu = 0; cpu < NR_CPUS; cpu++)
prev_nmi_count[cpu] = per_cpu(irq_stat, cpu).__nmi_count;
local_irq_enable();
mdelay((10*1000)/nmi_hz); // wait 10 ticks
for (cpu = 0; cpu < NR_CPUS; cpu++) {
#ifdef CONFIG_SMP
/* Check cpu_callin_map here because that is set
after the timer is started. */
if (!cpu_isset(cpu, cpu_callin_map))
continue;
#endif
if (nmi_count(cpu) - prev_nmi_count[cpu] <= 5) {
endflag = 1;
printk("CPU#%d: NMI appears to be stuck (%d->%d)!\n",
cpu,
prev_nmi_count[cpu],
nmi_count(cpu));
nmi_active = 0;
lapic_nmi_owner &= ~LAPIC_NMI_WATCHDOG;
kfree(prev_nmi_count);
return -1;
}
}
endflag = 1;
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 = 1;
kfree(prev_nmi_count);
return 0;
}
int __init check_watchdog(void)
{
irq_cpustat_t tmp[1];
printk(KERN_INFO "Testing Watchdog... ");
memcpy(tmp, irq_stat, sizeof(tmp));
local_irq_enable();
mdelay((10 * 1000) / watchdog_hz); /* wait 10 ticks */
local_irq_disable();
if (nmi_count(0) - tmp[0].__nmi_count <= 5) {
printk(KERN_WARNING "CPU#%d: Watchdog appears to be stuck!\n",
0);
return -1;
}
printk(KERN_INFO "OK.\n");
/* now that we know it works we can reduce NMI frequency to
* something more reasonable; makes a difference in some configs
*/
watchdog_hz = 1;
return 0;
}
int __init check_nmi_watchdog (void)
{
irq_cpustat_t tmp[NR_CPUS];
int j, cpu;
if (nmi_watchdog == NMI_LOCAL_APIC && !cpu_has_lapic())
return -1;
printk(KERN_INFO "testing NMI watchdog ... ");
memcpy(tmp, irq_stat, sizeof(tmp));
sti();
mdelay((10*1000)/nmi_hz); // wait 10 ticks
for (j = 0; j < smp_num_cpus; j++) {
cpu = cpu_logical_map(j);
if (nmi_count(cpu) - tmp[cpu].__nmi_count <= 5) {
printk("CPU#%d: NMI appears to be stuck!\n", cpu);
return -1;
}
}
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 = 1;
return 0;
}
asmlinkage
void watchdog_interrupt(struct pt_regs *regs, enum exception_code excep)
{
/*
* Since current-> is always on the stack, and we always switch
* the stack NMI-atomically, it's safe to use smp_processor_id().
*/
int sum, cpu = smp_processor_id();
u8 wdt, tmp;
wdt = WDCTR & ~WDCTR_WDCNE;
WDCTR = wdt;
tmp = WDCTR;
NMICR = NMICR_WDIF;
nmi_count(cpu)++;
kstat_this_cpu.irqs[NMIIRQ]++;
sum = irq_stat[cpu].__irq_count;
if (last_irq_sums[cpu] == sum) {
/*
* Ayiee, looks like this CPU is stuck ...
* wait a few IRQs (5 seconds) before doing the oops ...
*/
watchdog_alert_counter++;
if (watchdog_alert_counter == 5 * watchdog_hz) {
spin_lock(&watchdog_print_lock);
/*
* We are in trouble anyway, lets at least try
* to get a message out.
*/
bust_spinlocks(1);
printk(KERN_ERR
"NMI Watchdog detected LOCKUP on CPU%d,"
" pc %08lx, registers:\n",
cpu, regs->pc);
show_registers(regs);
printk("console shuts up ...\n");
console_silent();
spin_unlock(&watchdog_print_lock);
bust_spinlocks(0);
#ifdef CONFIG_GDBSTUB
if (gdbstub_busy)
gdbstub_exception(regs, excep);
else
gdbstub_intercept(regs, excep);
#endif
do_exit(SIGSEGV);
}
} else {
last_irq_sums[cpu] = sum;
watchdog_alert_counter = 0;
}
WDCTR = wdt | WDCTR_WDRST;
tmp = WDCTR;
WDCTR = wdt | WDCTR_WDCNE;
tmp = WDCTR;
}
int __init check_nmi_watchdog (void)
{
unsigned int prev_nmi_count[NR_CPUS];
int cpu;
printk(KERN_INFO "testing NMI watchdog ... ");
for (cpu = 0; cpu < NR_CPUS; cpu++)
prev_nmi_count[cpu] = irq_stat[cpu].__nmi_count;
local_irq_enable();
mdelay((10*1000)/nmi_hz); // wait 10 ticks
/* FIXME: Only boot CPU is online at this stage. Check CPUs
as they come up. */
for (cpu = 0; cpu < NR_CPUS; cpu++) {
if (!cpu_online(cpu))
continue;
if (nmi_count(cpu) - prev_nmi_count[cpu] <= 5) {
printk("CPU#%d: NMI appears to be stuck!\n", cpu);
nmi_active = 0;
lapic_nmi_owner &= ~LAPIC_NMI_WATCHDOG;
return -1;
}
}
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 = 1;
return 0;
}
int __init check_nmi_watchdog (void)
{
unsigned int prev_nmi_count[NR_CPUS];
int j, cpu;
printk(KERN_INFO "testing NMI watchdog ... ");
for (j = 0; j < smp_num_cpus; j++) {
cpu = cpu_logical_map(j);
prev_nmi_count[cpu] = irq_stat[cpu].__nmi_count;
}
sti();
mdelay((10*1000)/nmi_hz); // wait 10 ticks
for (j = 0; j < smp_num_cpus; j++) {
cpu = cpu_logical_map(j);
if (nmi_count(cpu) - prev_nmi_count[cpu] <= 5) {
printk("CPU#%d: NMI appears to be stuck!\n", cpu);
return -1;
}
}
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 = 1;
return 0;
}
int __init check_watchdog(void)
{
irq_cpustat_t tmp[1];
printk(KERN_INFO "Testing Watchdog... ");
memcpy(tmp, irq_stat, sizeof(tmp));
local_irq_enable();
mdelay((10 * 1000) / watchdog_hz); /* */
local_irq_disable();
if (nmi_count(0) - tmp[0].__nmi_count <= 5) {
printk(KERN_WARNING "CPU#%d: Watchdog appears to be stuck!\n",
0);
return -1;
}
printk(KERN_INFO "OK.\n");
/*
*/
watchdog_hz = 1;
return 0;
}
int __init check_nmi_watchdog (void)
{
static unsigned int __initdata prev_nmi_count[NR_CPUS];
int cpu;
bool_t ok = 1;
if ( !nmi_watchdog )
return 0;
printk("Testing NMI watchdog on all CPUs:");
for_each_online_cpu ( cpu )
prev_nmi_count[cpu] = nmi_count(cpu);
/* Wait for 10 ticks. Busy-wait on all CPUs: the LAPIC counter that
* the NMI watchdog uses only runs while the core's not halted */
if ( nmi_watchdog == NMI_LOCAL_APIC )
smp_call_function(wait_for_nmis, NULL, 0);
wait_for_nmis(NULL);
for_each_online_cpu ( cpu )
{
if ( nmi_count(cpu) - prev_nmi_count[cpu] <= 5 )
{
printk(" %d", cpu);
ok = 0;
}
}
printk(" %s\n", ok ? "ok" : "stuck");
/*
* Now that we know it works we can reduce NMI frequency to
* something more reasonable; makes a difference in some configs.
* There's a limit to how slow we can go because writing the perfctr
* MSRs only sets the low 32 bits, with the top 8 bits sign-extended
* from those, so it's not possible to set up a delay larger than
* 2^31 cycles and smaller than (2^40 - 2^31) cycles.
* (Intel SDM, section 18.22.2)
*/
if ( nmi_watchdog == NMI_LOCAL_APIC )
nmi_hz = max(1ul, cpu_khz >> 20);
return 0;
}
/*
* Display interrupt management information through /proc/interrupts
*/
int show_interrupts(struct seq_file *p, void *v)
{
int i = *(loff_t *) v, j, cpu;
struct irqaction *action;
unsigned long flags;
switch (i) {
/* display column title bar naming CPUs */
case 0:
seq_printf(p, " ");
for (j = 0; j < NR_CPUS; j++)
if (cpu_online(j))
seq_printf(p, "CPU%d ", j);
seq_putc(p, '\n');
break;
/* display information rows, one per active CPU */
case 1 ... NR_IRQS - 1:
spin_lock_irqsave(&irq_desc[i].lock, flags);
action = irq_desc[i].action;
if (action) {
seq_printf(p, "%3d: ", i);
for_each_present_cpu(cpu)
seq_printf(p, "%10u ", kstat_irqs_cpu(i, cpu));
seq_printf(p, " %14s.%u", irq_desc[i].chip->name,
(GxICR(i) & GxICR_LEVEL) >>
GxICR_LEVEL_SHIFT);
seq_printf(p, " %s", action->name);
for (action = action->next;
action;
action = action->next)
seq_printf(p, ", %s", action->name);
seq_putc(p, '\n');
}
spin_unlock_irqrestore(&irq_desc[i].lock, flags);
break;
/* polish off with NMI and error counters */
case NR_IRQS:
seq_printf(p, "NMI: ");
for (j = 0; j < NR_CPUS; j++)
if (cpu_online(j))
seq_printf(p, "%10u ", nmi_count(j));
seq_putc(p, '\n');
seq_printf(p, "ERR: %10u\n", atomic_read(&irq_err_count));
break;
}
return 0;
}
int __init check_nmi_watchdog (void)
{
static unsigned int __initdata prev_nmi_count[NR_CPUS];
int cpu;
if ( !nmi_watchdog )
return 0;
printk("Testing NMI watchdog --- ");
for_each_online_cpu ( cpu )
prev_nmi_count[cpu] = nmi_count(cpu);
local_irq_enable();
mdelay((10*1000)/nmi_hz); /* wait 10 ticks */
for_each_online_cpu ( cpu )
{
if ( nmi_count(cpu) - prev_nmi_count[cpu] <= 5 )
printk("CPU#%d stuck. ", cpu);
else
printk("CPU#%d okay. ", cpu);
}
printk("\n");
/*
* Now that we know it works we can reduce NMI frequency to
* something more reasonable; makes a difference in some configs.
* There's a limit to how slow we can go because writing the perfctr
* MSRs only sets the low 32 bits, with the top 8 bits sign-extended
* from those, so it's not possible to set up a delay larger than
* 2^31 cycles and smaller than (2^40 - 2^31) cycles.
* (Intel SDM, section 18.22.2)
*/
if ( nmi_watchdog == NMI_LOCAL_APIC )
nmi_hz = max(1ul, cpu_khz >> 20);
return 0;
}
dotraplinkage notrace __kprobes void
do_nmi(struct pt_regs *regs, long error_code)
{
nmi_enter();
#ifdef CONFIG_X86_32
{ int cpu; cpu = smp_processor_id(); ++nmi_count(cpu); }
#else
add_pda(__nmi_count, 1);
#endif
if (!ignore_nmis)
default_do_nmi(regs);
nmi_exit();
}
static void do_nmi_stats(unsigned char key)
{
int i;
struct domain *d;
struct vcpu *v;
printk("CPU\tNMI\n");
for_each_online_cpu ( i )
printk("%3d\t%3d\n", i, nmi_count(i));
if ( ((d = dom0) == NULL) || (d->vcpu == NULL) ||
((v = d->vcpu[0]) == NULL) )
return;
i = v->async_exception_mask & (1 << VCPU_TRAP_NMI);
if ( v->nmi_pending || i )
printk("dom0 vpu0: NMI %s%s\n",
v->nmi_pending ? "pending " : "",
i ? "masked " : "");
else
printk("dom0 vcpu0: NMI neither pending nor masked\n");
}
asmlinkage
void watchdog_interrupt(struct pt_regs *regs, enum exception_code excep)
{
/*
* Since current-> is always on the stack, and we always switch
* the stack NMI-atomically, it's safe to use smp_processor_id().
*/
int sum, cpu;
int irq = NMIIRQ;
u8 wdt, tmp;
wdt = WDCTR & ~WDCTR_WDCNE;
WDCTR = wdt;
tmp = WDCTR;
NMICR = NMICR_WDIF;
nmi_count(smp_processor_id())++;
kstat_incr_irqs_this_cpu(irq, irq_to_desc(irq));
for_each_online_cpu(cpu) {
sum = irq_stat[cpu].__irq_count;
if ((last_irq_sums[cpu] == sum)
#if defined(CONFIG_GDBSTUB) && defined(CONFIG_SMP)
&& !(CHK_GDBSTUB_BUSY()
|| atomic_read(&cpu_doing_single_step))
#endif
) {
/*
* Ayiee, looks like this CPU is stuck ...
* wait a few IRQs (5 seconds) before doing the oops ...
*/
watchdog_alert_counter[cpu]++;
if (watchdog_alert_counter[cpu] == 5 * watchdog_hz) {
spin_lock(&watchdog_print_lock);
/*
* We are in trouble anyway, lets at least try
* to get a message out.
*/
bust_spinlocks(1);
printk(KERN_ERR
"NMI Watchdog detected LOCKUP on CPU%d,"
" pc %08lx, registers:\n",
cpu, regs->pc);
#ifdef CONFIG_SMP
printk(KERN_ERR
"--- Register Dump (CPU%d) ---\n",
CPUID);
#endif
show_registers(regs);
#ifdef CONFIG_SMP
smp_nmi_call_function(watchdog_dump_register,
NULL, 1);
#endif
printk(KERN_NOTICE "console shuts up ...\n");
console_silent();
spin_unlock(&watchdog_print_lock);
bust_spinlocks(0);
#ifdef CONFIG_GDBSTUB
if (CHK_GDBSTUB_BUSY_AND_ACTIVE())
gdbstub_exception(regs, excep);
else
gdbstub_intercept(regs, excep);
#endif
do_exit(SIGSEGV);
}
} else {
last_irq_sums[cpu] = sum;
watchdog_alert_counter[cpu] = 0;
}
}
WDCTR = wdt | WDCTR_WDRST;
tmp = WDCTR;
WDCTR = wdt | WDCTR_WDCNE;
tmp = WDCTR;
}
asmlinkage
void watchdog_interrupt(struct pt_regs *regs, enum exception_code excep)
{
/*
*/
int sum, cpu;
int irq = NMIIRQ;
u8 wdt, tmp;
wdt = WDCTR & ~WDCTR_WDCNE;
WDCTR = wdt;
tmp = WDCTR;
NMICR = NMICR_WDIF;
nmi_count(smp_processor_id())++;
kstat_incr_irqs_this_cpu(irq, irq_to_desc(irq));
for_each_online_cpu(cpu) {
sum = irq_stat[cpu].__irq_count;
if ((last_irq_sums[cpu] == sum)
#if defined(CONFIG_GDBSTUB) && defined(CONFIG_SMP)
&& !(CHK_GDBSTUB_BUSY()
|| atomic_read(&cpu_doing_single_step))
#endif
) {
/*
*/
watchdog_alert_counter[cpu]++;
if (watchdog_alert_counter[cpu] == 5 * watchdog_hz) {
spin_lock(&watchdog_print_lock);
/*
*/
bust_spinlocks(1);
printk(KERN_ERR
"NMI Watchdog detected LOCKUP on CPU%d,"
" pc %08lx, registers:\n",
cpu, regs->pc);
#ifdef CONFIG_SMP
printk(KERN_ERR
"--- Register Dump (CPU%d) ---\n",
CPUID);
#endif
show_registers(regs);
#ifdef CONFIG_SMP
smp_nmi_call_function(watchdog_dump_register,
NULL, 1);
#endif
printk(KERN_NOTICE "console shuts up ...\n");
console_silent();
spin_unlock(&watchdog_print_lock);
bust_spinlocks(0);
#ifdef CONFIG_GDBSTUB
if (CHK_GDBSTUB_BUSY_AND_ACTIVE())
gdbstub_exception(regs, excep);
else
gdbstub_intercept(regs, excep);
#endif
do_exit(SIGSEGV);
}
} else {
last_irq_sums[cpu] = sum;
watchdog_alert_counter[cpu] = 0;
}
}
WDCTR = wdt | WDCTR_WDRST;
tmp = WDCTR;
WDCTR = wdt | WDCTR_WDCNE;
tmp = WDCTR;
}
