mirror of
https://github.com/torvalds/linux.git
synced 2024-12-06 19:11:31 +00:00
8d539b84f1
The APIs that allow backtracing across CPUs have always had a way to exclude the current CPU. This convenience means callers didn't need to find a place to allocate a CPU mask just to handle the common case. Let's extend the API to take a CPU ID to exclude instead of just a boolean. This isn't any more complex for the API to handle and allows the hardlockup detector to exclude a different CPU (the one it already did a trace for) without needing to find space for a CPU mask. Arguably, this new API also encourages safer behavior. Specifically if the caller wants to avoid tracing the current CPU (maybe because they already traced the current CPU) this makes it more obvious to the caller that they need to make sure that the current CPU ID can't change. [akpm@linux-foundation.org: fix trigger_allbutcpu_cpu_backtrace() stub] Link: https://lkml.kernel.org/r/20230804065935.v4.1.Ia35521b91fc781368945161d7b28538f9996c182@changeid Signed-off-by: Douglas Anderson <dianders@chromium.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: kernel test robot <lkp@intel.com> Cc: Lecopzer Chen <lecopzer.chen@mediatek.com> Cc: Petr Mladek <pmladek@suse.com> Cc: Pingfan Liu <kernelfans@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
124 lines
3.4 KiB
C
124 lines
3.4 KiB
C
// SPDX-License-Identifier: GPL-2.0
|
|
/*
|
|
* NMI backtrace support
|
|
*
|
|
* Gratuitously copied from arch/x86/kernel/apic/hw_nmi.c by Russell King,
|
|
* with the following header:
|
|
*
|
|
* 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/cpumask.h>
|
|
#include <linux/delay.h>
|
|
#include <linux/kprobes.h>
|
|
#include <linux/nmi.h>
|
|
#include <linux/cpu.h>
|
|
#include <linux/sched/debug.h>
|
|
|
|
#ifdef arch_trigger_cpumask_backtrace
|
|
/* For reliability, we're prepared to waste bits here. */
|
|
static DECLARE_BITMAP(backtrace_mask, NR_CPUS) __read_mostly;
|
|
|
|
/* "in progress" flag of arch_trigger_cpumask_backtrace */
|
|
static unsigned long backtrace_flag;
|
|
|
|
/*
|
|
* When raise() is called it will be passed a pointer to the
|
|
* backtrace_mask. Architectures that call nmi_cpu_backtrace()
|
|
* directly from their raise() functions may rely on the mask
|
|
* they are passed being updated as a side effect of this call.
|
|
*/
|
|
void nmi_trigger_cpumask_backtrace(const cpumask_t *mask,
|
|
int exclude_cpu,
|
|
void (*raise)(cpumask_t *mask))
|
|
{
|
|
int i, this_cpu = get_cpu();
|
|
|
|
if (test_and_set_bit(0, &backtrace_flag)) {
|
|
/*
|
|
* If there is already a trigger_all_cpu_backtrace() in progress
|
|
* (backtrace_flag == 1), don't output double cpu dump infos.
|
|
*/
|
|
put_cpu();
|
|
return;
|
|
}
|
|
|
|
cpumask_copy(to_cpumask(backtrace_mask), mask);
|
|
if (exclude_cpu != -1)
|
|
cpumask_clear_cpu(exclude_cpu, to_cpumask(backtrace_mask));
|
|
|
|
/*
|
|
* Don't try to send an NMI to this cpu; it may work on some
|
|
* architectures, but on others it may not, and we'll get
|
|
* information at least as useful just by doing a dump_stack() here.
|
|
* Note that nmi_cpu_backtrace(NULL) will clear the cpu bit.
|
|
*/
|
|
if (cpumask_test_cpu(this_cpu, to_cpumask(backtrace_mask)))
|
|
nmi_cpu_backtrace(NULL);
|
|
|
|
if (!cpumask_empty(to_cpumask(backtrace_mask))) {
|
|
pr_info("Sending NMI from CPU %d to CPUs %*pbl:\n",
|
|
this_cpu, nr_cpumask_bits, to_cpumask(backtrace_mask));
|
|
nmi_backtrace_stall_snap(to_cpumask(backtrace_mask));
|
|
raise(to_cpumask(backtrace_mask));
|
|
}
|
|
|
|
/* 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);
|
|
touch_softlockup_watchdog();
|
|
}
|
|
nmi_backtrace_stall_check(to_cpumask(backtrace_mask));
|
|
|
|
/*
|
|
* Force flush any remote buffers that might be stuck in IRQ context
|
|
* and therefore could not run their irq_work.
|
|
*/
|
|
printk_trigger_flush();
|
|
|
|
clear_bit_unlock(0, &backtrace_flag);
|
|
put_cpu();
|
|
}
|
|
|
|
// Dump stacks even for idle CPUs.
|
|
static bool backtrace_idle;
|
|
module_param(backtrace_idle, bool, 0644);
|
|
|
|
bool nmi_cpu_backtrace(struct pt_regs *regs)
|
|
{
|
|
int cpu = smp_processor_id();
|
|
unsigned long flags;
|
|
|
|
if (cpumask_test_cpu(cpu, to_cpumask(backtrace_mask))) {
|
|
/*
|
|
* Allow nested NMI backtraces while serializing
|
|
* against other CPUs.
|
|
*/
|
|
printk_cpu_sync_get_irqsave(flags);
|
|
if (!READ_ONCE(backtrace_idle) && regs && cpu_in_idle(instruction_pointer(regs))) {
|
|
pr_warn("NMI backtrace for cpu %d skipped: idling at %pS\n",
|
|
cpu, (void *)instruction_pointer(regs));
|
|
} else {
|
|
pr_warn("NMI backtrace for cpu %d\n", cpu);
|
|
if (regs)
|
|
show_regs(regs);
|
|
else
|
|
dump_stack();
|
|
}
|
|
printk_cpu_sync_put_irqrestore(flags);
|
|
cpumask_clear_cpu(cpu, to_cpumask(backtrace_mask));
|
|
return true;
|
|
}
|
|
|
|
return false;
|
|
}
|
|
NOKPROBE_SYMBOL(nmi_cpu_backtrace);
|
|
#endif
|