Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230328111714.2056-1-zhangqing@loongson.cn> <CA+fCnZevgYh7CzJ9gOWJ80SwY4Y9w8UO2ZiFAXEnAhQhFgrffA@mail.gmail.com>
 <dccfbff3-7bad-de33-4d96-248bdff44a8b@loongson.cn>
In-Reply-To: <dccfbff3-7bad-de33-4d96-248bdff44a8b@loongson.cn>
From:   Andrey Konovalov <andreyknvl@gmail.com>
Date:   Thu, 30 Mar 2023 04:55:57 +0200
Message-ID: <CA+fCnZddt50+10SZ+hZRKBudsmMF0W9XpsDG6=58p1ot62LjXQ@mail.gmail.com>
Subject: Re: [PATCH] LoongArch: Add kernel address sanitizer support
To:     Qing Zhang <zhangqing@loongson.cn>
Cc:     Jonathan Corbet <corbet@lwn.net>,
        Huacai Chen <chenhuacai@kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        Alexander Potapenko <glider@google.com>,
        Dmitry Vyukov <dvyukov@google.com>,
        Vincenzo Frascino <vincenzo.frascino@arm.com>,
        WANG Xuerui <kernel@xen0n.name>,
        Jiaxun Yang <jiaxun.yang@flygoat.com>,
        kasan-dev@googlegroups.com, linux-doc@vger.kernel.org,
        linux-mm@kvack.org, loongarch@lists.linux.dev,
        linux-kernel@vger.kernel.org, linux-hardening@vger.kernel.org,
        Andrey Ryabinin <ryabinin.a.a@gmail.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Mar 30, 2023 at 4:06=E2=80=AFAM Qing Zhang <zhangqing@loongson.cn> =
wrote:
>
> > But I don't think you need this check here at all: addr_has_metadata
> > already checks that shadow exists.
> >
> On LongArch, there's a lot of holes between different segments, so kasan
> shadow area is some different type of memory that we concatenate, we
> can't use if (unlikely((void *)addr <
> kasan_shadow_to_mem((void *)KASAN_SHADOW_START))) to determine the
> validity, and in arch/loongarch/include/asm/kasan.h I construct invalid
> NULL.

I get that, but you already added a special case for
__HAVE_ARCH_SHADOW_MAP to addr_has_metadata, so you can just call it?

> This is because in pagetable_init on loongarch/mips, we populate pmd/pud
> with invalid_pmd_table/invalid_pud_table,

I see. Please add this into the patch description for v2.

> So pmd_init/pud_init(p) is required, perhaps we define them as __weak in
> mm/kasan/init.c, like mm/sparse-vmemmap.c.

Yes, this makes sense to do, so that KASAN doesn't depend on
definitions from sparse-vmemmap.c.

Thank you!