[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMe9rOrYY9eB0Je74iC0Yw5YwTaUPOqQ2fpr1RVPFTgyZaD0cA@mail.gmail.com>
Date: Sat, 15 Feb 2014 10:41:27 -0800
From: "H.J. Lu" <hjl.tools@...il.com>
To: Andi Kleen <andi@...stfloor.org>
Cc: Markus Trippelsdorf <markus@...ppelsdorf.de>,
LKML <linux-kernel@...r.kernel.org>, x86@...nel.org,
linux-kbuild@...r.kernel.org, Michal Marek <mmarek@...e.cz>
Subject: Re: Updated Link-Time-Optimization patchkit
On Sat, Feb 15, 2014 at 9:44 AM, Andi Kleen <andi@...stfloor.org> wrote:
> On Sat, Feb 15, 2014 at 02:38:14PM +0100, Markus Trippelsdorf wrote:
>> On 2014.02.14 at 22:21 +0100, Andi Kleen wrote:
>> > This is the updated LTO patchkit for 3.14-rc2. LTO allows
>> > the compiler to do global optimization over the whole kernel.
>>
>> It is mildly annoying that one couldn't use vanilla binutils. Have you
>> already opened bugs on sourceware.org/bugzilla/ to get this fixed for
>> gold and ld.bfd?
>
> The problem is supporting "pass through" of both pure (.S) assembler code and
> LTO through ld -r, because the kernel makefiles use ld -r heavily.
> Standard binutils would throw all the assembler away when in plugin LTO
> mode.
>
> HJ did the work (and fixed a couple other bugs that blocked kernel LTO),
> but the mainline binutils people rejected the patches. You can judge
> for yourself if their reasons make sense:
>
> https://sourceware.org/ml/binutils/2012-05/msg00057.html
>
> Noone has done the work for gold so far.
>
> There was some effort from the Fedora binutils maintainers to add the
> necessary patches to their version, but at least what is in FC20 doesn't
> seem to be enough.
There are some additional LTO bug fixes in my Linux
binutils.
--
H.J.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists