From mboxrd@z Thu Jan  1 00:00:00 1970
From: Ben Warren <biggerbadderben@gmail.com>
Date: Wed, 01 Sep 2010 12:13:37 -0700
Subject: [U-Boot] [PATCH NET 2/2] Net: clarify board/cpu_eth_init calls
In-Reply-To: <201009011304.53715.vapier@gentoo.org>
References: <1283321104-953-1-git-send-email-biggerbadderben@gmail.com>
	<1283321104-953-2-git-send-email-biggerbadderben@gmail.com>
	<201009011304.53715.vapier@gentoo.org>
Message-ID: <4C7EA5E1.1040105@gmail.com>
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

  Hi Mike,

On 9/1/2010 10:04 AM, Mike Frysinger wrote:
> On Wednesday, September 01, 2010 02:05:04 Ben Warren wrote:
>> +	if (board_eth_init != __def_eth_init) {
>> +	} else if (cpu_eth_init != __def_eth_init) {
> i'm not sure these changes are useful.  the resolution of the symbols happens
> at link time, so it isnt like gcc will be able to optimize away the default.
>
> if anything, it'd make more sense to declare the functions as external/weak,
> and then check that the pointer is not NULL.  that'd save on the overhead of
> having uncalled stub functions that merely return 0 in the final linked image.
> -mike
This did work as I hoped on my PPC eval board, but maybe not globally.  
I remember that initially, with the functions defined as weak but with 
no body, check for NULL didn't work.  I've never tried declaring them as 
external, though.

thanks,
Ben