Date: Wed, 12 Jul 2017 13:34:46 -0400 (EDT)
From: Nicolas Pitre <nico@fluxnic.net>
To: Andi Kleen <andi@firstfloor.org>
cc: npiggin@gmail.com, linux-kernel@vger.kernel.org,
        linux-kbuild@vger.kernel.org
Subject: Re: [RFC PATCH] x86: enable dead code and data elimination (LTO)
In-Reply-To: <87a849aaa3.fsf@firstfloor.org>
Message-ID: <alpine.LFD.2.20.1707121331060.1959@knanqh.ubzr>
References: <20170709031333.29443-1-npiggin@gmail.com> <20170709090551.bm2c55ctt3togim7@gmail.com> <CAK7LNAR-VF2o3-qcCXnkath_B74KQodh_vezPRErGPhZSao2VA@mail.gmail.com> <alpine.LFD.2.20.1707090946090.1959@knanqh.ubzr> <20170710121329.77dd3364@roar.ozlabs.ibm.com>
 <87a849aaa3.fsf@firstfloor.org>
User-Agent: Alpine 2.20 (LFD 67 2015-01-07)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 914
Lines: 26

On Wed, 12 Jul 2017, Andi Kleen wrote:

> Nicholas Piggin <npiggin@gmail.com> writes:
> >> 
> >> I think we should aim for gc-sections to be used by default and have LTO 
> >> as a possible option only.
> >
> > I agree after it starts getting implemented and debugged by small
> > system users, we could make it default in the interest of sharing
> > testing and reducing combinations.
> 
> From what i understand the main drawback in the past was
> is that various linker versions become very slow with thousands of
> sections.
> 
> So it may cost you built time. For a special small build it's probably
> ok, but you wouldn't want to make it default.
> 
> Also usually it's only useful without modules because if you
> use modules EXPORT_SYMBOL pulls in a lot of unused functions.

I created CONFIG_TRIM_UNUSED_KSYMS mainly to avoid that issue. It is 
highly effective with either gc-sections and LTO.


Nicolas