Re: [workflows]Re: In defense of Link (was Re: [b4] initial "b4 dig" to supplant Link: trailers)

[Laurent Pinchart <laurent.pinchart@ideasonboard.com>]

On Mon, Oct 13, 2025 at 10:59:36AM -0400, Steven Rostedt wrote:
> On Mon, 13 Oct 2025 17:39:59 +0300 Laurent Pinchart wrote:
> 
> > Using Link: for this will not work well as soon as you need multiple
> > links. Compare the following (edited with made up information) commit
> > message with
> > 
> >     Remove the HAVE_FTRACE_MCOUNT config and use DYNAMIC_FTRACE directly where
> >     applicable. This has shown a 50% improvement in CPU usage.
> >     
> >     Link: https://lore.kernel.org/all/20250703154916.48e3ada7@gandalf.local.home/
> >     Link: https://lore.kernel.org/all/20250703140000.deadbeef@gandalf.local.home/
> > 
> > with
> > 
> >     The HAVE_FTRACE_MCOUNT config was reported as being reduntant ([1]).
> >     Remove it and use DYNAMIC_FTRACE directly where applicable. This has
> >     shown a 50% improvement in CPU usage ([2]).
> >     
> >     [1] https://lore.kernel.org/all/20250703154916.48e3ada7@gandalf.local.home/
> >     [2] https://lore.kernel.org/all/20250703140000.deadbeef@gandalf.local.home/
> 
> I'm fine with replacing link with normal reference tags for human
> discussion purposes. As I add them manually anyway, that doesn't make a
> difference to me.
> 
> But the main issue here is that if we get rid of the Link: tag to the
> submission, it will totally destroy my workflow. So much so, I'll just stop
> doing the daisy chain references.

My point was that, if we don't use a Link: trailer for human discussion
purpose, we could reclaim its usage for automation. The two problems are
somehow distinct but related, what triggered me was that a large part of
the discussion was based on the assumption Link: should be used for
human links, while I find that usage counterproductive.

-- 
Regards,

Laurent Pinchart