Re: [PATCH 2/7] stacktrace,sched: Make stack_trace_save_tsk() more robust

[Peter Zijlstra]

On Fri, Oct 22, 2021 at 05:09:35PM +0200, Peter Zijlstra wrote:
> --- a/kernel/stacktrace.c
> +++ b/kernel/stacktrace.c
> @@ -123,6 +123,13 @@ unsigned int stack_trace_save(unsigned l
>  }
>  EXPORT_SYMBOL_GPL(stack_trace_save);
>  
> +static int try_arch_stack_walk_tsk(struct task_struct *tsk, void *arg)
> +{
> +	stack_trace_consume_fn consume_entry = stack_trace_consume_entry_nosched;
> +	arch_stack_walk(consume_entry, arg, tsk, NULL);
> +	return 0;
> +}
> +
>  /**
>   * stack_trace_save_tsk - Save a task stack trace into a storage array
>   * @task:	The task to examine
> @@ -135,7 +142,6 @@ EXPORT_SYMBOL_GPL(stack_trace_save);
>  unsigned int stack_trace_save_tsk(struct task_struct *tsk, unsigned long *store,
>  				  unsigned int size, unsigned int skipnr)
>  {
> -	stack_trace_consume_fn consume_entry = stack_trace_consume_entry_nosched;
>  	struct stacktrace_cookie c = {
>  		.store	= store,
>  		.size	= size,
> @@ -143,11 +149,8 @@ unsigned int stack_trace_save_tsk(struct
>  		.skip	= skipnr + (current == tsk),
>  	};
>  
> -	if (!try_get_task_stack(tsk))
> -		return 0;

So I took that out because task_try_func() pins the task, except now
I see that _reliable() has a comment about zombies, which I suppose is
equally applicable to here and wchan.

Alternative to failing try_get_task_stack() is checking PF_EXITING in
try_arch_stack_walk_tsk(), which seems more consistent behaviour since
it doesn't rely on CONFIG_THREAD_INFO_IN_TASK.

> +	task_try_func(tsk, try_arch_stack_walk_tsk, &c);
>  
> -	arch_stack_walk(consume_entry, &c, tsk, NULL);
> -	put_task_stack(tsk);
>  	return c.len;
>  }
>  
> 
>