Re: [PATCH] bitops: use safer link explaining the algorithm

From: Yury Norov <yury.norov@gmail.com>
To: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Wolfram Sang <wsa+renesas@sang-engineering.com>,
	linux-renesas-soc@vger.kernel.org, linux-i3c@lists.infradead.org,
	Alexandre Belloni <alexandre.belloni@bootlin.com>,
	Rasmus Villemoes <linux@rasmusvillemoes.dk>,
	Kuan-Wei Chiu <visitorckw@gmail.com>,
	linus@ucla.edu
Subject: Re: [PATCH] bitops: use safer link explaining the algorithm
Date: Sun, 26 Jan 2025 12:46:17 -0500	[thread overview]
Message-ID: <Z5Z06WowFspmXby9@thinkpad> (raw)
In-Reply-To: <CAMuHMdWmE2UiE7kwAyQKONSE_ytrNux+Wwu-v__K=jjpAutxqw@mail.gmail.com>

On Sat, Jan 25, 2025 at 04:10:36PM +0100, Geert Uytterhoeven wrote:
> Hi Wolfram,
> 
> CC Linus (the funloop one ;-)
> 
> On Sat, Jan 25, 2025 at 2:03 PM Wolfram Sang
> <wsa+renesas@sang-engineering.com> wrote:
> > During review, a concern was raised that the link explaining the
> > algorithm might get stale. Meanwhile, the site has been archived in the
> > WayBack machine. So, use their link which is hopefully more stable.
> >
> > Fixes: c320592f3f2a ("bitops: add generic parity calculation for u8")
> > Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
> 
> Thanks for your patch!
> 
> > --- a/include/linux/bitops.h
> > +++ b/include/linux/bitops.h
> > @@ -254,7 +254,7 @@ static inline int parity8(u8 val)
> >  {
> >         /*
> >          * One explanation of this algorithm:
> > -        * https://funloop.org/codex/problem/parity/README.html
> > +        * http://web.archive.org/web/20250105093316/https://funloop.org/codex/problem/parity/README.html
> 
> Is the plan to replace all weblinks by webarchive links as a precaution?
> Even websites backed by big companies may disappear[1]...
> Putting the webarchive link here also impacts the funloop.org server
> statistics, downplaying its relevance, and possibly even causing an
> earlier shutdown.
> The URL can always be updated when the original site disappears.
> 
> >          */
> >         val ^= val >> 4;
> >         return (0x6996 >> (val & 0xf)) & 1;
> 
> [1] The "LessWatts.org" T-shirt I got from Intel survived the
>     corresponding website by many years ;-)

That was actually my concern. I used to think that kernel repository
should better be a self-contained thing. And the reason is that kernel
sources should be understandable even on an Atlantic flight.

But this days airliners have pretty good internet access. So I don't
want to look like a guy pushing 80-chars line limit, while we forgot
even the meaning of VGA acronym maybe 15 years ago.

Quick grepping says that the include/linux directory already has 235
'http' links. Even after excluding compiler linkage, we have 149 of them.

So...

If you guys feel that you can explain the algorithm you're employing
in a half-VGA-screen comment, it would be the best choice, and that's
enough.

If it's impossible or you think that external reference is really needed,
I trust you. But please refer the original source. In this case, it's:

        Warren, H. S. (2013). Hacker’s Delight (2nd ed), page 97.

If you think that pointing to a web-page with nice summary to the topic 
is helpful - please do. And if you would like to cache the link - I've
got nothing against it - as soon as you commit to maintain those links
up-to-date in the kernel sources.

Thanks,
Yury

-- 
linux-i3c mailing list
linux-i3c@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-i3c