From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1751670AbdKLQI3 (ORCPT <rfc822;w@1wt.eu>);
        Sun, 12 Nov 2017 11:08:29 -0500
Received: from mail2-relais-roc.national.inria.fr ([192.134.164.83]:65529 "EHLO
        mail2-relais-roc.national.inria.fr" rhost-flags-OK-OK-OK-OK)
        by vger.kernel.org with ESMTP id S1751317AbdKLQI2 (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Sun, 12 Nov 2017 11:08:28 -0500
X-IronPort-AV: E=Sophos;i="5.44,384,1505772000"; 
   d="scan'208";a="300445841"
Date: Sun, 12 Nov 2017 17:08:26 +0100 (CET)
From: Julia Lawall <julia.lawall@lip6.fr>
X-X-Sender: jll@hadrien
To: Masahiro Yamada <yamada.masahiro@socionext.com>
cc: linux-kernel@vger.kernel.org, Michal Marek <michal.lkml@markovi.net>,
        Nicolas Palix <nicolas.palix@imag.fr>, cocci@systeme.lip6.fr
Subject: [PATCH] coccinelle: flags.cocci: reorganize patterns to improve
 efficiency
Message-ID: <alpine.DEB.2.20.1711121704350.10412@hadrien>
User-Agent: Alpine 2.20 (DEB 67 2015-01-07)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

An initial rule just matching a function call makes it possible to focus
on one all at a time, which improves performance in complex functions with
multiple locking calls.  Using a constraint expression lock2 != pre.lock1;
makes it possible to eliminate the double matching of the lock, lock
sequence.

Signed-off-by: Julia Lawall <julia.lawall@lip6.fr>

---

diff --git a/scripts/coccinelle/locks/flags.cocci b/scripts/coccinelle/locks/flags.cocci
index debd70e..33f29a5 100644
--- a/scripts/coccinelle/locks/flags.cocci
+++ b/scripts/coccinelle/locks/flags.cocci
@@ -12,9 +12,9 @@ virtual context
 virtual org
 virtual report

-@r exists@
-expression lock1,lock2,flags;
-position p1,p2;
+@pre exists@
+expression lock1,flags;
+position p1;
 @@

 (
@@ -24,25 +24,12 @@ read_lock_irqsave@p1(lock1,flags)
 |
 write_lock_irqsave@p1(lock1,flags)
 )
-... when != flags
-(
-spin_lock_irqsave(lock1,flags)
-|
-read_lock_irqsave(lock1,flags)
-|
-write_lock_irqsave(lock1,flags)
-|
-spin_lock_irqsave@p2(lock2,flags)
-|
-read_lock_irqsave@p2(lock2,flags)
-|
-write_lock_irqsave@p2(lock2,flags)
-)

-@d exists@
-expression f <= r.flags;
-expression lock1,lock2,flags;
-position r.p1, r.p2;
+@r exists@
+expression lock2 != pre.lock1;
+expression f <= pre.flags;
+expression pre.lock1,pre.flags;
+position pre.p1,p2;
 @@

 (
@@ -63,16 +50,16 @@ position r.p1, r.p2;

 // ----------------------------------------------------------------------

-@script:python depends on d && org@
-p1 << r.p1;
+@script:python depends on org@
+p1 << pre.p1;
 p2 << r.p2;
 @@

 cocci.print_main("original lock",p1)
 cocci.print_secs("nested lock+irqsave that reuses flags",p2)

-@script:python depends on d && report@
-p1 << r.p1;
+@script:python depends on report@
+p1 << pre.p1;
 p2 << r.p2;
 @@