Re: [PATCH v5 01/10] ring-buffer: Free cpu_buffer::free_page with subbuf_order

On Fri, Aug 14, 2026 at 12:49:32AM +0900, Masami Hiramatsu wrote:
> On Thu, 13 Aug 2026 13:54:39 +0000
> [email protected] wrote:
> 
> > Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> > 
> > Pre-existing issues:
> > - [Critical] A data race and locking mismatch between `ring_buffer_subbuf_order_set()`, `ring_buffer_alloc_read_page()`, and `ring_buffer_free_read_page()` allows `cpu_buffer->free_page` to cache a page of an old order while `buffer->subbuf_order` reflects a new larger order. This leads to heap buffer overflow and buddy allocator corruption.
> > --
> > 
> > --- Patch [1]: [PATCH v5 01/10] ring-buffer: Free cpu_buffer::free_page with subbuf_order ---
> > commit f550c0bfeadcc22eca47316e1820b4d91f6ef261
> > Author: Vincent Donnefort <[email protected]>
> > 
> > ring-buffer: Free cpu_buffer::free_page with subbuf_order
> >     
> > When sub-buffers use an order greater than 0, cpu_buffer->free_page is
> > allocated with subbuf_order. Use the correct order for
> > cpu_buffer->free_page.
> > 
> > > diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
> > > index 2667992f0aa29..a3d28b2e2c94a 100644
> > > --- a/kernel/trace/ring_buffer.c
> > > +++ b/kernel/trace/ring_buffer.c
> > > @@ -2631,7 +2631,7 @@ static void rb_free_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer)
> > >  		free_buffer_page(bpage);
> > >  	}
> > >  
> > > -	free_page((unsigned long)cpu_buffer->free_page);
> > > +	free_pages((unsigned long)cpu_buffer->free_page, cpu_buffer->buffer->subbuf_order);
> > 
> > [Severity: Critical]
> > This is a pre-existing issue, but does this change expose a buddy allocator
> > corruption if cpu_buffer->free_page contains a page of a smaller order than
> > buffer->subbuf_order?
> 
> Hmm, I think you have to fix cpu_buffer->{free_page,subbuf_order} race issue at first.
> Let me try another series to fix it.
> 
> Thank you,

You mean this one?

https://lore.kernel.org/all/[email protected]/

> 
> -- 
> Masami Hiramatsu (Google) <[email protected]>

-- 
Vincent