MIME-Version: 1.0
References: <20230202084238.2408516-1-chenhuacai@loongson.cn>
 <363cd09a5dcb4deab21f58c19025254f@AcuMS.aculab.com> <CAAhV-H7Mz1Z5Bo59tq5VRSUx-N39axeiG7xZs2Szn6nuOxgZfQ@mail.gmail.com>
 <9936da8f577842b8b5edafcdc69dc2d1@AcuMS.aculab.com> <560d73a8-2f2a-4844-44ff-afffad9c8694@loongson.cn>
 <379bcb55-f75d-02ce-a51b-467e21ade5a3@xen0n.name> <dcd17646-4b8d-447b-bd85-c66c4a7b2cf4@app.fastmail.com>
 <924aa802-410d-a85c-b623-7ca30d15c637@loongson.cn> <59745bb8-00d0-4fa3-b75c-511db310ac16@app.fastmail.com>
In-Reply-To: <59745bb8-00d0-4fa3-b75c-511db310ac16@app.fastmail.com>
From:   Huacai Chen <chenhuacai@kernel.org>
Date:   Wed, 8 Feb 2023 19:17:59 +0800
Message-ID: <CAAhV-H7X_T7t88TbEZfkrwUcDAoJy1X4KbzWw_KZaoXuq730RA@mail.gmail.com>
Subject: Re: [PATCH] LoongArch: Make -mstrict-align be configurable
To:     Arnd Bergmann <arnd@arndb.de>
Cc:     Jianmin Lv <lvjianmin@loongson.cn>,
        WANG Xuerui <kernel@xen0n.name>,
        David Laight <David.Laight@aculab.com>,
        Huacai Chen <chenhuacai@loongson.cn>,
        "loongarch@lists.linux.dev" <loongarch@lists.linux.dev>,
        Linux-Arch <linux-arch@vger.kernel.org>,
        Xuefeng Li <lixuefeng@loongson.cn>, guoren <guoren@kernel.org>,
        Jiaxun Yang <jiaxun.yang@flygoat.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hi, Arnd,

On Tue, Feb 7, 2023 at 10:11 PM Arnd Bergmann <arnd@arndb.de> wrote:
>
> On Tue, Feb 7, 2023, at 14:28, Jianmin Lv wrote:
> > On 2023/2/7 =E4=B8=8B=E5=8D=886:32, Arnd Bergmann wrote:
> >> I agree the default should always be to have a kernel that works on
> >> every machine that has been produced, but this also depends on which
> >> models specifically lack the unaligned access. If it's just about
> >> pre-production silicon that is now all but scrapped, things are differ=
ent
> >> from a situation where users may actually use them for normal workload=
s.
> >>
> >> Is there an overview of the available loongarch CPU cores that have
> >> been produced so far, and which ones support unaligned access?
> >
> > So far, produced CPUs based LoongArch include 3A5000, 3B5000, 3C5000L,
> > 3C5000, 2K2000, 2K1000LA and 2K0500, where 2K1000LA and 2K0500 are
> > unaligned-access-unsupported, and others are unaligned-access-supported=
.
>
> Ok, so these are actually some of the newer (though low-end)
> implementations that require the workaround, not the older chips.
>
> In this case, I think both the kernel and toolchain need to default
> to -mstrict-align, unless someone specifically asks for the variant
> that can support unaligned access. The kernel option could be
> guarded by 'depends on EXPERT' to ensure that this is not set by
> default.
>
> To be sure that this is set correctly, the
> arch/loongarch/kernel/unaligned.c file should also never be included
> when EFFICIENT_UNALIGNED_ACCESS is set, to ensure that any attempt
> to run such a non-portable kernel on 2K1000LA results in a
> a kernel panic rather than silently fixing up the unaligned accesses
> at a huge performance cost.
OK, sounds reasonable, I will send V2 for that.

Huacai
>
>     Arnd
>