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;
Date:   Fri, 21 Jul 2023 11:53:34 +0100 (BST)
From:   "Maciej W. Rozycki" <macro@orcam.me.uk>
To:     Huacai Chen <chenhuacai@kernel.org>
cc:     Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
        Jan-Benedict Glaw <jbglaw@lug-owl.de>,
        Guillaume Tucker <guillaume.tucker@collabora.com>,
        Huacai Chen <chenhuacai@loongson.cn>,
        linux-mips@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 0/3] MIPS: Fix build issues from the introduction of
 `need-compiler'
In-Reply-To: <CAAhV-H7K2qM35GuGV+B2p=N_f30i98KazvwVNCuvvG9fYfrghw@mail.gmail.com>
Message-ID: <alpine.DEB.2.21.2307211149070.17594@angie.orcam.me.uk>
References: <alpine.DEB.2.21.2307180025120.62448@angie.orcam.me.uk> <CAAhV-H4ewzWuZRe0RX+j2x2R_fROVExHuogWNaWBmFA=Tgb=Og@mail.gmail.com> <alpine.DEB.2.21.2307191621450.58367@angie.orcam.me.uk>
 <CAAhV-H7K2qM35GuGV+B2p=N_f30i98KazvwVNCuvvG9fYfrghw@mail.gmail.com>
User-Agent: Alpine 2.21 (DEB 202 2017-01-01)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Precedence: bulk

Hi Huacai,

> >  Secondly:
> >
> > cflags-$(CONFIG_CC_IS_GCC) += -march=loongson3a
> > cflags-$(CONFIG_CC_IS_CLANG) += -march=mips64r2
> >
> > document compiler peculiarities.  Does Clang support, or intend to,
> > `-march=loongson3a'?  If so, what version can we expect this stuff in?
> > GCC has had it since 4.6 or Y2010, which is pretty long ago.
> GCC support loongson3a/mips64r2, Clang only support mips64r2. If we use
> $(call cc-option,-march=loongson3a,-march=mips64r2)
> both GCC and Clang can work and we don't need to care about the compiler.

 This may well be a change we desire, but it has to be made and reviewed 
on its own rather than being buried within a set of unrelated changes.  
Then the rationale has to be given in the change description and a comment 
put in code explaining that it's not the usual case of old/new compiler, 
so that we can catch it later and remove should Clang developers decide to 
include `-march=loongson3a' support and our version requirements catch up.

  Maciej