Re: [PATCH] tools: memory-model: Make plain accesses carry dependencies

[Alan Stern <stern@rowland.harvard.edu>]

On Thu, Dec 01, 2022 at 05:21:45PM +0000, Jonas Oberhauser wrote:
> Thanks a lot for the helpful and detailed comments!
> Three minor points before I send a new patch:
> 
> > even if W' or R' (or both!) is plain.	
> 
> The "is" sounds slightly weird to me in the sentence because the last part I read is 
> "(or both!)", so I would slightly prefer "are" here.

People are pretty casual about subject-verb number agreement these days 
(there's a growing tendency in English for people to make the verb agree 
with the last noun occurring in the subject rather than the subject as a 
whole), so that should be okay.

> > On the other hand, if you change the second "dependencies" to "ones" and "unmarked" to "plain", maybe the whole thing will fit on one line.
> 
> It fits even if I changed the second dependencies to "those" instead of "ones", i.e.,
> (* Redefine dependencies to include those carried through plain accesses *)
> 
> which I would prefer.

Fine.

> > if you replaced the whole conditional with a simple
> >	WRITE_ONCE(*y, *z2);
> > then the litmus test would become an example of OOTA!
> 
> In my opinion it is already an example of OOTA, which I would define as an
>    rfi | ctrl | addr | data | fence
> cycle.

That's not an unreasonable point of view (if you put rfe rather than 
rfi), but to me OOTA suggests something more: a value arising as if by 
magic rather than as a result of a computation.  In your version of the 
litmus test there is WRITE_ONCE(*y, 1), so it's a little understandable 
that you could end up with 1 as the final values of x and y.  But in my 
version, no values get computed anywhere, so the final value of x and y 
might just as easily be 1 or 56789 -- it literally arises "out of thin 
air".

> Let me know if you agree with these deviations from your suggestion
> and have a great time,

Yes; with those changes you can add:

Reviewed-by: Alan Stern <stern@rowland.harvard.edu>

> jonas
> 
> PS: 
> > When a colon is followed by a clause (as opposed to a list), it is customary to capitalize the first letter of that clause, just like we capitalize the first letter at the start of a sentence.
> 
> In German, we also capitalize after a colon; but my English teachers used to deduct many points throughout my adolescent life whenever I capitalized like that. I still remember some of that red ink with near perfect clarity. So I eventually really took it to heart and started pedantically not-capitalizing after every colon.
> Now the only time it ever mattered in my adult life, I find that I should do it German Style (or, as I just learned, APA & AP Style).
> I suppose life is that way sometimes.

Indeed.

Alan