From mboxrd@z Thu Jan  1 00:00:00 1970
From: Scott Wood <scottwood@freescale.com>
Date: Fri, 7 Oct 2011 10:46:33 -0500
Subject: [U-Boot] [PATCH 2/2] nand: Merge changes from Linux nand driver
In-Reply-To: <4E8F14EA.9040902@aizo.com>
References: <CD1B567EF015FE41B29FF476B8600DDD26BBF25B89@vsrv-mail01.newtechgroup.local>
	<20111006210607.900BE140874A@gemini.denx.de>
	<4E8E1CAC.6000405@freescale.com> <4E8F14EA.9040902@aizo.com>
Message-ID: <4E8F1ED9.5040701@freescale.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

On 10/07/2011 10:04 AM, Christian Hitz wrote:
> Am 06.10.2011 23:25, schrieb Scott Wood:
>> What do the string changes have to do with synchronizing the NAND driver?
> 
> Strings read from ONFI are now sanitized. For that strim() is used which
> was not yet available in U-Boots string library. Of course I went
> overboard with syncing the whole string library. I'll submit a new
> version that only adds strim().

Please make it a separate patch.

>> How much does this update affect image size?  This may be a good time to
>> start ifdeffing portions of the NAND code based on what's actually
>> needed by the target.
> 
> before: u-boot.bin: 298112 bytes
> after: u-boot.bin: 301928 bytes
> (based on our at91sam9g20ek based design, ONFI enabled)
> 
> One option would be to ifdef the 1-bit software ECC. This would free
> aroud 2k (down to: 299744 bytes).

There are a bunch of things that could be ifdeffed, which gc-sections
doesn't help with because they are referenced with things like:

	if (!pointer)
		pointer = some_func;


Think large page versus small page, functions that the hw-specific
driver overrides, all the different types of ECC (ifdeffing out 1-bit
but not 4 or 8 won't keep U-Boot from growing on boards that use 1-bit),
ONFI, the read_page/write_page variants, etc.

-Scott