Re: [PATCH] tracing: Move histogram trigger variables from stack to per CPU structure

[Steven Rostedt]

On Fri, 4 Apr 2025 09:58:41 -0400
Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:

> Hmm, this is incorrect. hist_pads can be NULL or allocated, so the check
> should be:
> 
> 	if (!hist_pads)
> 		goto fail;
> 	hist_pad = per_cpu_ptr(hist_pads, smp_processor_id());
> 
> Will send v2.

In fact, if hist_pads fails to allocate, the triggers should not be enabled
(it would fail to enable). So I'll just add to the start of the function:

	if (WARN_ON_ONCE(!hist_pads))
		return NULL;

-- Steve