mail archive of the barebox mailing list
 help / color / mirror / Atom feed
From: Sascha Hauer <s.hauer@pengutronix.de>
To: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: barebox@lists.infradead.org
Subject: Re: [PATCH] ARM: socfpga: generate smaller images when multiple boards are selected
Date: Mon, 26 Nov 2018 10:46:24 +0100	[thread overview]
Message-ID: <20181126094624.l56oxn7kintdqc7s@pengutronix.de> (raw)
In-Reply-To: <CAK7LNATiis8ci+bczBsPWHfVSC=pePnpTNddQigJ9uVWOtqDww@mail.gmail.com>

On Fri, Nov 23, 2018 at 05:49:30PM +0900, Masahiro Yamada wrote:
> On Fri, Nov 16, 2018 at 11:31 PM Sascha Hauer <s.hauer@pengutronix.de> wrote:
> >
> > The socfpga xload images are limited to 64KiB. This doesn't fit if
> > multiple boards are selected. The reason is that we include huge
> > C files and arrays in the early init code which get compiled once
> > for each board. -ffunction-sections is without effect here since all
> > functions have the same name and hence we get the same function
> > multiple times in the same section.
> >
> > To overcome this we surround all function names with a SECT() macro which
> > is used to add a board specific prefix to the section names. This way
> > -ffunction-sections can now do its work and discard unused functions.
> >
> > Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
> 
> 
> I just accidentally stopped by this commit.
> 
> It is a good thing that you noticed dead code elimination
> (-ffunction-sections -fdata-sections, --gc-sections)
> does not work if the same symbol names are used
> in the conventional incremental linking
> Linux kernel used to use.
> 
> As you notice, this is a potential problem
> of CONFIG_PBL_MULTI_IMAGES,
> and I know the right solution.
> 
> 
> Since Linux v4.13 (commit 98ced886dd79a),
> Linux kernel build system uses thin archive,
> where Kbuild uses $(AR) rcSTP instead of $(LD) -r
> to combine built-in objects.
> 
> With this, the final link stage does the right thing to do
> for --gc-sections.
> 
> Besides, thin archive is faster,
> and saves disk spaces.
> 
> If you sync core Makefiles with Linux,
> your problem will be solved.

That's really nice, thanks for noting this.

I had a short look into it, unfortunately syncing the Makefile doesn't
appear to be an easy job. I think we stick to the current patch for now
until someone finds time to do it. It's good to know there's a proper
solution to the problem though

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

      reply	other threads:[~2018-11-26  9:46 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-11-16 14:31 Sascha Hauer
2018-11-19  8:30 ` Steffen Trumtrar
2018-11-23  8:49 ` Masahiro Yamada
2018-11-26  9:46   ` Sascha Hauer [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20181126094624.l56oxn7kintdqc7s@pengutronix.de \
    --to=s.hauer@pengutronix.de \
    --cc=barebox@lists.infradead.org \
    --cc=yamada.masahiro@socionext.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox