Re: [PATCH] memorder: Reconcile Figure 15.15 notation with W+RWC z-flow semantics

["Kunwu Chan" <kunwu.chan@linux.dev>]

March 11, 2026 at 10:39 PM, "Paul E. McKenney" <paulmck@kernel.org mailto:paulmck@kernel.org?to=%22Paul%20E.%20McKenney%22%20%3Cpaulmck%40kernel.org%3E > wrote:


> 
> On Wed, Mar 11, 2026 at 04:48:13PM +0800, Kunwu Chan wrote:
> 
> > 
> > State explicitly that Figure 15.15 uses x as a schematic variable name, while the surrounding W+RWC analysis in Listing 15.18 follows the concrete z-flow (P2 -> z -> P1). This eliminates notation drift across cross-references and preserves propagation-argument fidelity.
> >  
> >  Signed-off-by: Kunwu Chan <kunwu.chan@linux.dev>
> > 
> Again, good eyes, and thank you! I queued and pushed all three, and
> added a fourth commit that converted this to a footnote and mentioned
> both variables and the listing, shown below. Please check it to make
> sure that I have not messed something up.
> 
>  Thanx, Paul
> 
Hi Paul,
Checked — LGTM.
The footnote conversion is the right call. It cleanly resolves 
the Figure 15.15 x vs. W+RWC z notation mismatch.
Thanks for landing this promptly.

Best, Kunwu

> > 
> > ---
> >  memorder/memorder.tex | 2 ++
> >  1 file changed, 2 insertions(+)
> >  
> >  diff --git a/memorder/memorder.tex b/memorder/memorder.tex
> >  index 2357e7ea..8eb14a29 100644
> >  --- a/memorder/memorder.tex
> >  +++ b/memorder/memorder.tex
> >  @@ -2760,6 +2760,8 @@ that \co{P1()}'s read from \co{z} happens much later in time, but
> >  nevertheless still sees the old value of zero.
> >  This situation is depicted in
> >  \cref{fig:memorder:Load-to-Store is Counter-Temporal}:
> >  +The figure uses \co{x} as a generic variable name for the same
> >  +counter-temporal pattern.
> >  Just because a load sees the old value does \emph{not} mean that
> >  this load executed at an earlier time than did the store of the
> >  new value.
> >  -- 
> >  2.25.1
> > 
> ------------------------------------------------------------------------
> 
> commit d2c790e9077625e2e28ee919619b4d18175c3665
> Author: Paul E. McKenney <paulmck@kernel.org>
> Date: Wed Mar 11 07:34:07 2026 -0700
> 
>  memorder: Convert Figure 15.15 variable-x commentary to footnote
>  
>  Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
> 
> diff --git a/memorder/memorder.tex b/memorder/memorder.tex
> index 388903b6..4503c0a6 100644
> --- a/memorder/memorder.tex
> +++ b/memorder/memorder.tex
> @@ -2759,9 +2759,10 @@ It therefore takes time for the effect of the \co{P2()}'s store to
>  that \co{P1()}'s read from \co{z} happens much later in time, but
>  nevertheless still sees the old value of zero.
>  This situation is depicted in
> -\cref{fig:memorder:Load-to-Store is Counter-Temporal}:
> -The figure uses \co{x} as a generic variable name for the same
> -counter-temporal pattern.
> +\cref{fig:memorder:Load-to-Store is Counter-Temporal}:\footnote{
> + The figure uses \co{x} as a generic variable name for the same
> + counter-temporal pattern that uses the variable \co{z} in
> + \cref{lst:memorder:W+RWC Litmus Test With Release (No Ordering)}.}
>  Just because a load sees the old value does \emph{not} mean that
>  this load executed at an earlier time than did the store of the
>  new value.
>