提交 5568b139 编写于 作者: S Steven Rostedt 提交者: Thomas Gleixner

ftrace: debug smp_processor_id, use notrace preempt disable

The debug smp_processor_id caused a recursive fault in debugging
the irqsoff tracer. The tracer used a smp_processor_id in the
ftrace callback, and this function called preempt_disable which
also is traced. This caused a recursive fault (stack overload).

Since using smp_processor_id without debugging on does not cause
faults with the tracer (even when the tracer is wrong), the
debug version should not cause a system reboot.

This changes the debug_smp_processor_id to use the notrace versions
of preempt_disable and enable.
Signed-off-by: NSteven Rostedt <srostedt@redhat.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
上级 f43fdad8
......@@ -37,7 +37,7 @@ notrace unsigned int debug_smp_processor_id(void)
/*
* Avoid recursion:
*/
preempt_disable();
preempt_disable_notrace();
if (!printk_ratelimit())
goto out_enable;
......@@ -49,7 +49,7 @@ notrace unsigned int debug_smp_processor_id(void)
dump_stack();
out_enable:
preempt_enable_no_resched();
preempt_enable_no_resched_notrace();
out:
return this_cpu;
}
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册