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: <cover.1614989433.git.andreyknvl@google.com> <a7f1d687b0550182c7f5b4a47c277a61425af65f.1614989433.git.andreyknvl@google.com>
 <YEYMDn/9zQI8g+3o@elver.google.com>
In-Reply-To: <YEYMDn/9zQI8g+3o@elver.google.com>
From:   Andrey Konovalov <andreyknvl@google.com>
Date:   Mon, 8 Mar 2021 15:14:21 +0100
Message-ID: <CAAeHK+zGDDhpzHqBrKcYhY3UvDG8iXfkCBVQ-5Se0QyESpQ91Q@mail.gmail.com>
Subject: Re: [PATCH 3/5] kasan, mm: integrate page_alloc init with HW_TAGS
To:     Marco Elver <elver@google.com>
Cc:     Alexander Potapenko <glider@google.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will.deacon@arm.com>,
        Vincenzo Frascino <vincenzo.frascino@arm.com>,
        Dmitry Vyukov <dvyukov@google.com>,
        Andrey Ryabinin <aryabinin@virtuozzo.com>,
        Peter Collingbourne <pcc@google.com>,
        Evgenii Stepanov <eugenis@google.com>,
        Branislav Rankov <Branislav.Rankov@arm.com>,
        Kevin Brodsky <kevin.brodsky@arm.com>,
        kasan-dev <kasan-dev@googlegroups.com>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        Linux Memory Management List <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, Mar 8, 2021 at 12:35 PM Marco Elver <elver@google.com> wrote:
>
> > -     kasan_free_nondeferred_pages(page, order, fpi_flags);
> > +     init = want_init_on_free();
> > +     if (init && !IS_ENABLED(CONFIG_KASAN_HW_TAGS))
>
> Doing the !IS_ENABLED(CONFIG_KASAN_HW_TAGS) check is awkward, and
> assumes internal knowledge of the KASAN implementation and how all
> current and future architectures that support HW_TAGS work.
>
> Could we instead add a static inline helper to <linux/kasan.h>, e.g.
> kasan_supports_init() or so?
>
> That way, these checks won't grow uncontrollable if a future
> architecture implements HW_TAGS but not init.

Good idea, I'll add a helper in v2.

> Hmm, KASAN certainly "supports" memory initialization always. So maybe
> "kasan_has_accelerated_init()" is more accurate?  I leave it to you to
> decide what the best option is.

Let's call it kasan_has_integrated_init().

Thanks!