linux-kernel - Re: [PATCH] coccinelle: Improve setup

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAGXu5jLgzcO8D1N8mkjHygd2tfoptGMAd1vK5j-8uZ5_xY9Egg@mail.gmail.com>
Date:   Wed, 23 Aug 2017 11:57:12 -0700
From:   Kees Cook <keescook@...omium.org>
To:     Julia Lawall <julia.lawall@...6.fr>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        Vaishali Thakkar <vaishali.thakkar@...cle.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Christoph Hellwig <hch@...radead.org>,
        Gilles Muller <Gilles.Muller@...6.fr>,
        Nicolas Palix <nicolas.palix@...g.fr>,
        Michal Marek <mmarek@...e.com>, cocci@...teme.lip6.fr
Subject: Re: [PATCH] coccinelle: Improve setup_timer.cocci matching

On Wed, Aug 23, 2017 at 6:13 AM, Julia Lawall <julia.lawall@...6.fr> wrote:
>
>
> On Tue, 22 Aug 2017, Kees Cook wrote:
>
>> This improves the patch mode of setup_timer.cocci. Several patterns were
>> missing:
>>  - assignments-before-init_timer() cases
>>  - limiting the .data case removal to the struct timer_list instance
>>  - handling calls by dereference (timer->field vs timer.field)
>>
>> Running this on the current kernel tree produces a large diff that I
>> would like to get applied as the first step in removing the .data
>> field from struct timer_list:
>>
>>  208 files changed, 367 insertions(+), 757 deletions(-)
>>
>> Signed-off-by: Kees Cook <keescook@...omium.org>
>> ---
>>  scripts/coccinelle/api/setup_timer.cocci | 129 +++++++++++++++++++++++++------
>>  1 file changed, 105 insertions(+), 24 deletions(-)
>>
>> diff --git a/scripts/coccinelle/api/setup_timer.cocci b/scripts/coccinelle/api/setup_timer.cocci
>> index eb6bd9e4ab1a..bc6bd8f0b4bf 100644
>> --- a/scripts/coccinelle/api/setup_timer.cocci
>> +++ b/scripts/coccinelle/api/setup_timer.cocci
>> @@ -2,6 +2,7 @@
>>  /// and data fields
>>  // Confidence: High
>>  // Copyright: (C) 2016 Vaishali Thakkar, Oracle. GPLv2
>> +// Copyright: (C) 2017 Kees Cook, Google. GPLv2
>>  // Options: --no-includes --include-headers
>>  // Keywords: init_timer, setup_timer
>>
>> @@ -10,60 +11,123 @@ virtual context
>>  virtual org
>>  virtual report
>>
>> +// Match the common cases first to avoid Coccinelle parsing loops with
>> +// "... when" clauses.
>> +
>>  @match_immediate_function_data_after_init_timer
>>  depends on patch && !context && !org && !report@
>>  expression e, func, da;
>>  @@
>>
>> --init_timer (&e);
>> -+setup_timer (&e, func, da);
>> +-init_timer
>> ++setup_timer
>> + ( \(&e\|e\)
>> ++, func, da
>> + );
>> +(
>> +-\(e.function\|e->function\) = func;
>> +-\(e.data\|e->data\) = da;
>> +|
>> +-\(e.function\|e->function\) = func;
>> +-\(e.data\|e->data\) = da;
>
> Same thing twice; I think you want to invert the last two lines.
>
>> +)
>> +
>> +@...ch_immediate_function_data_before_init_timer
>> +depends on patch && !context && !org && !report@
>> +expression e, func, da;
>> +@@
>>
>>  (
>> +-\(e.function\|e->function\) = func;
>> +-\(e.data\|e->data\) = da;
>> +|
>> +-\(e.function\|e->function\) = func;
>> +-\(e.data\|e->data\) = da;
>
> Same as in the previous case.

Oops, thanks! AIUI, missing these cases must makes runtime slower.
I'll fix it up and resend.

-Kees

-- 
Kees Cook
Pixel Security