Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <f6efb2f36fc1f40eb22df027e6bc956cac71745e.1615498565.git.andreyknvl@google.com>
 <c0f6a95b0fa59ce0ef502f4ea11522141e3c8faf.1615498565.git.andreyknvl@google.com>
 <YEtKVYVeUycUKySP@elver.google.com>
In-Reply-To: <YEtKVYVeUycUKySP@elver.google.com>
From:   Andrey Konovalov <andreyknvl@google.com>
Date:   Fri, 12 Mar 2021 15:07:30 +0100
Message-ID: <CAAeHK+w3C+Umd9j__P=97KHQ-AEqS10gi-5DA5tc0Yav5zzWEA@mail.gmail.com>
Subject: Re: [PATCH 10/11] kasan: docs: update ignoring accesses section
To:     Marco Elver <elver@google.com>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Alexander Potapenko <glider@google.com>,
        Andrey Ryabinin <aryabinin@virtuozzo.com>,
        Dmitry Vyukov <dvyukov@google.com>,
        kasan-dev <kasan-dev@googlegroups.com>,
        Linux Memory Management List <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Mar 12, 2021 at 12:02 PM Marco Elver <elver@google.com> wrote:
>
> On Thu, Mar 11, 2021 at 10:37PM +0100, Andrey Konovalov wrote:
> [...]
> > +Other parts of the kernel might access metadata for allocated objects. Normally,
> > +KASAN detects and reports such accesses, but in certain cases (e.g., in memory
> > +allocators) these accesses are valid. Disabling instrumentation for memory
> > +allocators files helps with accesses that happen directly in that code for
> > +software KASAN modes. But it does not help when the accesses happen indirectly
> > +(through generic function calls) or with the hardware tag-based mode that does
> > +not use compiler instrumentation.
> > +
> > +To disable KASAN reports in a certain part of the kernel code:
> > +
> > +- For software modes, add a
> > +  ``kasan_disable_current()``/``kasan_enable_current()`` critical section.
>
> Should we mention function attribute __no_sanitize_address (and noinstr,
> which just applies to any kind of instrumentation) here? Perhaps with
> the note that called functions may still be instrumented, and in such
> cases would require combining with kasan_{disable,enable}_current().

Indeed, forgot about the attributes.. Will add, thank you!