mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Julia Lawall <julia.lawall@inria.fr>
To: Denis Efremov <efremov@linux.com>
Cc: cocci@systeme.lip6.fr, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 RESEND] coccinelle: misc: add minmax script
Date: Sat, 6 Mar 2021 15:08:56 +0100 (CET)	[thread overview]
Message-ID: <alpine.DEB.2.22.394.2103061503290.2976@hadrien> (raw)
In-Reply-To: <20210219090655.8985-1-efremov@linux.com>



On Fri, 19 Feb 2021, Denis Efremov wrote:

> Check for opencoded min(), max() implementations.
>
> Signed-off-by: Denis Efremov <efremov@linux.com>
> ---
>
> Changes in v2:
>  - <... ...> instead of ... when any
>  - org mode reports fixed
>  - patch rule to drop excessive ()
>
>  scripts/coccinelle/misc/minmax.cocci | 224 +++++++++++++++++++++++++++
>  1 file changed, 224 insertions(+)
>  create mode 100644 scripts/coccinelle/misc/minmax.cocci
>
> diff --git a/scripts/coccinelle/misc/minmax.cocci b/scripts/coccinelle/misc/minmax.cocci
> new file mode 100644
> index 000000000000..61d6b61fd82c
> --- /dev/null
> +++ b/scripts/coccinelle/misc/minmax.cocci
> @@ -0,0 +1,224 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +///
> +/// Check for opencoded min(), max() implementations.
> +/// Generated patches sometimes require adding a cast to fix compile warning.
> +/// Warnings/patches scope intentionally limited to a function body.
> +///
> +// Confidence: Medium
> +// Copyright: (C) 2021 Denis Efremov ISPRAS
> +// Options: --no-includes --include-headers
> +//
> +// Keywords: min, max
> +//
> +
> +
> +virtual report
> +virtual org
> +virtual context
> +virtual patch
> +
> +@rmax depends on !patch@
> +identifier func;
> +expression x, y;
> +binary operator cmp = {>, >=};
> +position p;
> +@@
> +
> +func(...)
> +{
> +	<...
> +*	x cmp@p y ? x : y

The rule below indicated with FIXME is supposed to deal with the
possibility of () that are unnecessary when using min and max.  It doesn't
work, because <... P ...> allow P to match 0 or more times, and thus
func@p matches every function.

A simpler solution is to just allow arbitrary () in the pattern, eg:

  (x) cmp@p (y) ? (x) : (y)

That will allow each occurrence of x and y to occur with and without
parentheses.  In the submitted  semantic patch, the () issue was only
considered in the patch case.  But it actually affects the purely matching
cases too, because () can be used at one occurrence, but not the other.

> +@script:python depends on report@
> +p << rmax.p;
> +@@
> +
> +coccilib.report.print_report(p[0], "WARNING opportunity for max()")

p is an array because it can be bound to different positions on different
control-flow paths.  Notably this occurs with <... ...>.  If there are
multiple occurrences of the pattern, there will be one match that contains
all of them.  Thus the reporting code should be:

for p0 in p:
  coccilib.report.print_report(p0, "WARNING opportunity for max()")

julia

  reply	other threads:[~2021-03-06 14:09 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-02-16 16:03 [PATCH] " Denis Efremov
2021-02-17 21:17 ` Julia Lawall
2021-02-19  9:06 ` [PATCH v2 RESEND] " Denis Efremov
2021-03-06 14:08   ` Julia Lawall [this message]
2021-03-05 10:16 ` [PATCH v3] " Denis Efremov
2021-03-08  7:04 ` [PATCH v4] " Denis Efremov
2021-03-08 21:30   ` Julia Lawall
2021-03-09  6:39 ` [PATCH v5] " Denis Efremov
2021-03-20 21:31   ` Julia Lawall

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=alpine.DEB.2.22.394.2103061503290.2976@hadrien \
    --to=julia.lawall@inria.fr \
    --cc=cocci@systeme.lip6.fr \
    --cc=efremov@linux.com \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®