android_kernel_xiaomi_sm7250/arch/sparc/kernel/nmi.c

281 lines
6.0 KiB
C
Raw Normal View History

/* Pseudo NMI support on sparc64 systems.
*
* Copyright (C) 2009 David S. Miller <davem@davemloft.net>
*
* The NMI watchdog support and infrastructure is based almost
* entirely upon the x86 NMI support code.
*/
#include <linux/kernel.h>
#include <linux/param.h>
#include <linux/init.h>
#include <linux/percpu.h>
#include <linux/nmi.h>
#include <linux/module.h>
#include <linux/kprobes.h>
#include <linux/kernel_stat.h>
#include <linux/reboot.h>
#include <linux/slab.h>
#include <linux/kdebug.h>
#include <linux/delay.h>
#include <linux/smp.h>
#include <asm/perf_counter.h>
#include <asm/ptrace.h>
#include <asm/local.h>
#include <asm/pcr.h>
/* We don't have a real NMI on sparc64, but we can fake one
* up using profiling counter overflow interrupts and interrupt
* levels.
*
* The profile overflow interrupts at level 15, so we use
* level 14 as our IRQ off level.
*/
static int panic_on_timeout;
/* nmi_active:
* >0: the NMI watchdog is active, but can be disabled
* <0: the NMI watchdog has not been set up, and cannot be enabled
* 0: the NMI watchdog is disabled, but can be enabled
*/
atomic_t nmi_active = ATOMIC_INIT(0); /* oprofile uses this */
EXPORT_SYMBOL(nmi_active);
static unsigned int nmi_hz = HZ;
static DEFINE_PER_CPU(short, wd_enabled);
static int endflag __initdata;
static DEFINE_PER_CPU(unsigned int, last_irq_sum);
static DEFINE_PER_CPU(local_t, alert_counter);
static DEFINE_PER_CPU(int, nmi_touch);
void touch_nmi_watchdog(void)
{
if (atomic_read(&nmi_active)) {
int cpu;
for_each_present_cpu(cpu) {
if (per_cpu(nmi_touch, cpu) != 1)
per_cpu(nmi_touch, cpu) = 1;
}
}
touch_softlockup_watchdog();
}
EXPORT_SYMBOL(touch_nmi_watchdog);
static void die_nmi(const char *str, struct pt_regs *regs, int do_panic)
{
if (notify_die(DIE_NMIWATCHDOG, str, regs, 0,
pt_regs_trap_type(regs), SIGINT) == NOTIFY_STOP)
return;
console_verbose();
bust_spinlocks(1);
printk(KERN_EMERG "%s", str);
printk(" on CPU%d, ip %08lx, registers:\n",
smp_processor_id(), regs->tpc);
show_regs(regs);
dump_stack();
bust_spinlocks(0);
if (do_panic || panic_on_oops)
panic("Non maskable interrupt");
nmi_exit();
local_irq_enable();
do_exit(SIGBUS);
}
notrace __kprobes void perfctr_irq(int irq, struct pt_regs *regs)
{
unsigned int sum, touched = 0;
int cpu = smp_processor_id();
clear_softint(1 << irq);
pcr_ops->write(PCR_PIC_PRIV);
local_cpu_data().__nmi_count++;
nmi_enter();
if (notify_die(DIE_NMI, "nmi", regs, 0,
pt_regs_trap_type(regs), SIGINT) == NOTIFY_STOP)
touched = 1;
sum = kstat_irqs_cpu(0, cpu);
if (__get_cpu_var(nmi_touch)) {
__get_cpu_var(nmi_touch) = 0;
touched = 1;
}
if (!touched && __get_cpu_var(last_irq_sum) == sum) {
local_inc(&__get_cpu_var(alert_counter));
sparc64: Kill spurious NMI watchdog triggers by increasing limit to 30 seconds. This is a compromise and a temporary workaround for bootup NMI watchdog triggers some people see with qla2xxx devices present. This happens when, for example: CPU 0 is in the driver init and looping submitting mailbox commands to load the firmware, then waiting for completion. CPU 1 is receiving the device interrupts. CPU 1 is where the NMI watchdog triggers. CPU 0 is submitting mailbox commands fast enough that by the time CPU 1 returns from the device interrupt handler, a new one is pending. This sequence runs for more than 5 seconds. The problematic case is CPU 1's timer interrupt running when the barrage of device interrupts begin. Then we have: timer interrupt return for softirq checking pending, thus enable interrupts qla2xxx interrupt return qla2xxx interrupt return ... 5+ seconds pass final qla2xxx interrupt for fw load return run timer softirq return At some point in the multi-second qla2xxx interrupt storm we trigger the NMI watchdog on CPU 1 from the NMI interrupt handler. The timer softirq, once we get back to running it, is smart enough to run the timer work enough times to make up for the missed timer interrupts. However, the NMI watchdogs (both x86 and sparc) use the timer interrupt count to notice the cpu is wedged. But in the above scenerio we'll receive only one such timer interrupt even if we last all the way back to running the timer softirq. The default watchdog trigger point is only 5 seconds, which is pretty low (the softwatchdog triggers at 60 seconds). So increase it to 30 seconds for now. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-03 11:35:20 +02:00
if (local_read(&__get_cpu_var(alert_counter)) == 30 * nmi_hz)
die_nmi("BUG: NMI Watchdog detected LOCKUP",
regs, panic_on_timeout);
} else {
__get_cpu_var(last_irq_sum) = sum;
local_set(&__get_cpu_var(alert_counter), 0);
}
if (__get_cpu_var(wd_enabled)) {
write_pic(picl_value(nmi_hz));
pcr_ops->write(pcr_enable);
}
nmi_exit();
}
static inline unsigned int get_nmi_count(int cpu)
{
return cpu_data(cpu).__nmi_count;
}
static __init void nmi_cpu_busy(void *data)
{
local_irq_enable_in_hardirq();
while (endflag == 0)
mb();
}
static void report_broken_nmi(int cpu, 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);
}
void stop_nmi_watchdog(void *unused)
{
pcr_ops->write(PCR_PIC_PRIV);
__get_cpu_var(wd_enabled) = 0;
atomic_dec(&nmi_active);
}
static int __init check_nmi_watchdog(void)
{
unsigned int *prev_nmi_count;
int cpu, err;
if (!atomic_read(&nmi_active))
return 0;
prev_nmi_count = kmalloc(nr_cpu_ids * sizeof(unsigned int), GFP_KERNEL);
if (!prev_nmi_count) {
err = -ENOMEM;
goto error;
}
printk(KERN_INFO "Testing NMI watchdog ... ");
smp_call_function(nmi_cpu_busy, (void *)&endflag, 0);
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);
err = -ENODEV;
goto error;
}
printk("OK.\n");
nmi_hz = 1;
kfree(prev_nmi_count);
return 0;
error:
on_each_cpu(stop_nmi_watchdog, NULL, 1);
return err;
}
void start_nmi_watchdog(void *unused)
{
__get_cpu_var(wd_enabled) = 1;
atomic_inc(&nmi_active);
pcr_ops->write(PCR_PIC_PRIV);
write_pic(picl_value(nmi_hz));
pcr_ops->write(pcr_enable);
}
static void nmi_adjust_hz_one(void *unused)
{
if (!__get_cpu_var(wd_enabled))
return;
pcr_ops->write(PCR_PIC_PRIV);
write_pic(picl_value(nmi_hz));
pcr_ops->write(pcr_enable);
}
void nmi_adjust_hz(unsigned int new_hz)
{
nmi_hz = new_hz;
on_each_cpu(nmi_adjust_hz_one, NULL, 1);
}
EXPORT_SYMBOL_GPL(nmi_adjust_hz);
static int nmi_shutdown(struct notifier_block *nb, unsigned long cmd, void *p)
{
on_each_cpu(stop_nmi_watchdog, NULL, 1);
return 0;
}
static struct notifier_block nmi_reboot_notifier = {
.notifier_call = nmi_shutdown,
};
int __init nmi_init(void)
{
int err;
on_each_cpu(start_nmi_watchdog, NULL, 1);
err = check_nmi_watchdog();
if (!err) {
err = register_reboot_notifier(&nmi_reboot_notifier);
if (err) {
on_each_cpu(stop_nmi_watchdog, NULL, 1);
atomic_set(&nmi_active, -1);
}
}
if (!err)
init_hw_perf_counters();
return err;
}
static int __init setup_nmi_watchdog(char *str)
{
if (!strncmp(str, "panic", 5))
panic_on_timeout = 1;
return 0;
}
__setup("nmi_watchdog=", setup_nmi_watchdog);