linux-kernel - Re: [PATCH -rcu/kcsan 04/23] kcsan: Add core support for a subset of weak memory modeling

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]

Date:   Tue, 5 Oct 2021 15:13:25 +0200
From:   Marco Elver <elver@...gle.com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     "Paul E . McKenney" <paulmck@...nel.org>,
        Alexander Potapenko <glider@...gle.com>,
        Boqun Feng <boqun.feng@...il.com>,
        Borislav Petkov <bp@...en8.de>,
        Dmitry Vyukov <dvyukov@...gle.com>,
        Ingo Molnar <mingo@...nel.org>,
        Josh Poimboeuf <jpoimboe@...hat.com>,
        Mark Rutland <mark.rutland@....com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Waiman Long <longman@...hat.com>,
        Will Deacon <will@...nel.org>, kasan-dev@...glegroups.com,
        linux-arch@...r.kernel.org, linux-doc@...r.kernel.org,
        linux-kbuild@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-mm@...ck.org, x86@...nel.org
Subject: Re: [PATCH -rcu/kcsan 04/23] kcsan: Add core support for a subset of
 weak memory modeling

On Tue, 5 Oct 2021 at 14:53, Peter Zijlstra <peterz@...radead.org> wrote:
> On Tue, Oct 05, 2021 at 12:58:46PM +0200, Marco Elver wrote:
> > +#if !defined(CONFIG_ARCH_WANTS_NO_INSTR) || defined(CONFIG_STACK_VALIDATION)
> > +/*
> > + * Arch does not rely on noinstr, or objtool will remove memory barrier
> > + * instrumentation, and no instrumentation of noinstr code is expected.
> > + */
> > +#define kcsan_noinstr
>
> I think this still wants to be at the very least:
>
> #define kcsan_noinstr noinline notrace
>
> without noinline it is possible LTO (or similarly daft things) will end
> up inlining the calls, and since we rely on objtool to NOP out CALLs
> this must not happen.

Good point about noinline, will add.

> And since you want to mark these functions as uaccess_safe, there must
> not be any tracing on, hence notrace.

In the Makefile we've relied on:

  CFLAGS_REMOVE_core.o = $(CC_FLAGS_FTRACE)

just to disable it for all code here. That should be enough, right?