Re: [PATCH v8 1/3] tracing: Fix subbuf resize races with trace_pipe_raw readers

On August 28, 2026 4:24:36 AM EDT, Vincent Donnefort [off-list ref] wrote:

On Thu, Aug 27, 2026 at 03:33:14PM -0400, Steven Rostedt wrote:

quoted

On Thu, 27 Aug 2026 17:21:39 +0100
Vincent Donnefort [off-list ref] wrote:

quoted

On Thu, Aug 27, 2026 at 09:15:41AM -0400, Steven Rostedt wrote:

quoted

On Thu, 27 Aug 2026 07:31:23 +0100
Vincent Donnefort [off-list ref] wrote:
  

quoted

Looking at the function again, shouldn't we add !rb_is_static(cpu_buffer) as
well to the full page test?  

Why? 

-- Steve  

So that the splice can make progress when the ring-buffer is presistent or
user-mapped. Or do we say that's expected behaviour? 

Ah yeah, I see. Hmm, looking at the code to see how this would work, I
think I found another bug. Let me look deeper into it.

-- Steve

A bug if we allow rb_is_static()?

No, something else.

If not, shall I add it to my series?

Yeah, go ahead and update it.

Thanks,

-- Steve