From mboxrd@z Thu Jan  1 00:00:00 1970
From: Scott Wood <scottwood@freescale.com>
Date: Mon, 17 Sep 2012 12:13:20 -0500
Subject: [U-Boot] [PATCH v2 09/11] S3C24XX: Add NAND Flash driver
In-Reply-To: <50575901.5020903@ti.com> (from trini@ti.com on Mon Sep 17
	12:08:17 2012)
References: <1347643742-19966-1-git-send-email-jose.goncalves@inov.pt>
	<1347643742-19966-10-git-send-email-jose.goncalves@inov.pt>
	<201209142021.12033.marex@denx.de> <50537B54.2060903@inov.pt>
	<20120914190120.GL22028@bill-the-cat> <505598FF.7010108@inov.pt>
	<20120917165757.GA16252@bill-the-cat> <1347901417.19543.9@snotra>
	<50575901.5020903@ti.com>
Message-ID: <1347902000.19543.10@snotra>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

On 09/17/2012 12:08:17 PM, Tom Rini wrote:
> On 09/17/12 10:03, Scott Wood wrote:
> > Weak symbols are not OK for configuring printf out of the SPL, as  
> you'll
> > still have all the format strings and caller code in the binary.  It
> > should be a macro (or an inline function that replaces the standard
> > printf declaration), but it should be in a system header (not the  
> CPU
> > directory -- not sure what Marek meant there) and be based on an
> > appropriate CONFIG symbol.
> 
> I'm a little leery of adding #if ... into <common.h> around printf.   
> I'd
> like to not worry about the branch/return bytes until we really really
> have to again but yes, the strings are more of a concern since they
> won't be collected out.  Just top of my head thinking above.

Caller code won't be collected either.  It's not just branch/return  
bytes but argument preparation -- possibly significant chunks of code  
to calculate values to be displayed, that might otherwise be optimized  
out.

-Scott