linux-kernel - Re: [PATCH v4] lib/vsprintf: defer filling siphash key on RT

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

Message-ID: <CAHmME9qP2ZR6oQY3bf8W42RHW+-DamrzT=Gh1XJ3O7f72NU_bQ@mail.gmail.com>
Date:   Mon, 1 Aug 2022 15:55:17 +0200
From:   "Jason A. Donenfeld" <Jason@...c4.com>
To:     Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Cc:     linux-kernel@...r.kernel.org,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        John Ogness <john.ogness@...utronix.de>,
        Mike Galbraith <efault@....de>, Petr Mladek <pmladek@...e.com>,
        Rasmus Villemoes <linux@...musvillemoes.dk>,
        Sergey Senozhatsky <senozhatsky@...omium.org>,
        Steven Rostedt <rostedt@...dmis.org>,
        "Theodore Ts'o" <tytso@....edu>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH v4] lib/vsprintf: defer filling siphash key on RT

Hi Sebastian,

On 8/1/22, Sebastian Andrzej Siewior <bigeasy@...utronix.de> wrote:
> You do remember the warning that poped up in random core with
> CONFIG_PROVE_RAW_LOCK_NESTING enabled? Where I said it does not affect
> !RT it just points out a RT problem in a !RT config?
> If you fix this with one code path for RT and another one for !RT then
> you will have this warning _if_ the caller has a raw_spinlock_t
> acquired.

CONFIG_PROVE_RAW_LOCK_NESTING catches RT issues on non-RT. It doesn't
catch non-RT issues. So lockdep isn't actually warning about something
real as far as non-RT is concerned when CONFIG_PROVE_RAW_LOCK_NESTING
is enabled.

So probably this v4 patch should expand that condition to be:

IS_ENABLED(CONFIG_PROVE_RAW_LOCK_NESTING) || IS_ENABLED(CONFIG_PREEMPT_RT)

That's somewhat distasteful, I realize, but it *does* make the code
actually match reality, which is maybe better.

Jason