Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
In-Reply-To: <20180327023927.GA11454@guoren>
References: <cover.1521399976.git.ren_guo@c-sky.com> <c15b1fce07d0a19616026dfb927bfcbd88102b12.1521399976.git.ren_guo@c-sky.com>
 <CAK8P3a1rzS+MmubYqTE-8H89QnULcGbKth=eWTA5mJ5VLRNL0Q@mail.gmail.com>
 <20180320131342.GA31542@guoren> <CAK8P3a3pat6wQ+9gbf_wtp6+dur_h=Zp1896rogZHnehVnqVNA@mail.gmail.com>
 <20180321124137.GA21320@guoren> <CAK8P3a06MVAT35ZbG=gRkR1EDVdWPoT3bhwL+9fZSkmTHaDVBw@mail.gmail.com>
 <20180327023927.GA11454@guoren>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Tue, 27 Mar 2018 09:38:56 +0200
Message-ID: <CAK8P3a3A+BEmBnGCSiRdoaiqSJjhC5zaZebcq2GPVaLJatBk8Q@mail.gmail.com>
Subject: Re: [PATCH 15/19] csky: Build infrastructure
To:     Guo Ren <ren_guo@c-sky.com>
Cc:     linux-arch <linux-arch@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Daniel Lezcano <daniel.lezcano@linaro.org>,
        Jason Cooper <jason@lakedaemon.net>,
        c-sky_gcc_upstream@c-sky.com, gnu-csky@mentor.com,
        thomas.petazzoni@bootlin.com, wbx@uclibc-ng.org
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Mar 27, 2018 at 4:39 AM, Guo Ren <ren_guo@c-sky.com> wrote:
> On Mon, Mar 26, 2018 at 03:00:00PM +0200, Arnd Bergmann wrote:
>> Ok, I understand the part about ck610 being incompatible, but I'm
>> still not sure about the 8xx ones: Do you mean it's impossible to
>> have one kernel that runs across all of them for some other reason,
>> or is it something you haven't allowed because you see no use for it?
>
> Sorry, Csky gcc need "-mcpu=ck807" or "-mcpu=ck810" or "-mcpu=ck860" to
> determine the back-end policy. So I must seperate them with different vmlinux.

Usually the way gcc handles this, either each CPU is a strict superset
of another
one, so you just need to specify the one with the smallest instruction set,
or you have an option like -mcpu=generic that produces the common subset.

>> This is basically the same question as above: For c610, using the fixed
>> value is sufficient, because it's incompatible with the others. But if you want
>> to run the same kernel on both ck810 and ck860, then it needs some form
>> of runtime detection.
> Sorry, currently no runtime detection.
> But I agree with you that one vmlinux for all cpus is a good design for compat.
>
>> On other architectures, the L1_CACHE_BYTES constant is the maximum
>> possible cache line size, and the cache flush function uses the actual size
> The same with above, we don't detect cpus on runtime. So we just make it
> simple here.

Ok, fair enough. It's something that can always be added later, as long as
we keep this in mind early on.

       Arnd