MIME-Version: 1.0
References: <cover.1638308023.git.andreyknvl@google.com> <f405e36b20bd5d79dffef3f70b523885dcc6b163.1638308023.git.andreyknvl@google.com>
 <YajVYNBDOyI3hTx1@elver.google.com>
In-Reply-To: <YajVYNBDOyI3hTx1@elver.google.com>
From:   Andrey Konovalov <andreyknvl@gmail.com>
Date:   Mon, 6 Dec 2021 22:08:04 +0100
Message-ID: <CA+fCnZfuH6GNRQ7m-HU=MSrroe9BMounEuoFTSGJUGGk8=vKzQ@mail.gmail.com>
Subject: Re: [PATCH 20/31] kasan, vmalloc: reset tags in vmalloc functions
To:     Marco Elver <elver@google.com>
Cc:     andrey.konovalov@linux.dev,
        Alexander Potapenko <glider@google.com>,
        Vincenzo Frascino <vincenzo.frascino@arm.com>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Peter Collingbourne <pcc@google.com>,
        Dmitry Vyukov <dvyukov@google.com>,
        kasan-dev <kasan-dev@googlegroups.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Linux Memory Management List <linux-mm@kvack.org>,
        Will Deacon <will@kernel.org>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        Evgenii Stepanov <eugenis@google.com>,
        LKML <linux-kernel@vger.kernel.org>,
        Andrey Konovalov <andreyknvl@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Dec 2, 2021 at 3:17 PM Marco Elver <elver@google.com> wrote:
>
> On Tue, Nov 30, 2021 at 11:07PM +0100, andrey.konovalov@linux.dev wrote:
> > From: Andrey Konovalov <andreyknvl@google.com>
> >
> > In preparation for adding vmalloc support to SW/HW_TAGS KASAN,
> > reset pointer tags in functions that use pointer values in
> > range checks.
> >
> > vread() is a special case here. Resetting the pointer tag in its
> > prologue could technically lead to missing bad accesses to virtual
> > mappings in its implementation. However, vread() doesn't access the
> > virtual mappings cirectly. Instead, it recovers the physical address
>
> s/cirectly/directly/
>
> But this paragraph is a little confusing, because first you point out
> that vread() might miss bad accesses, but then say that it does checked
> accesses. I think to avoid confusing the reader, maybe just say that
> vread() is checked, but hypothetically, should its implementation change
> to directly access addr, invalid accesses might be missed.
>
> Did I get this right? Or am I still confused?

No, you got it right. Will reword in v2.

Thanks!