From mboxrd@z Thu Jan  1 00:00:00 1970
From: Albert ARIBAUD <albert.aribaud@free.fr>
Date: Fri, 24 Sep 2010 18:08:44 +0200
Subject: [U-Boot] [PATCH 2/2] [NEXT] arm926ejs: reduce code size with
 -msingle-pic-base
In-Reply-To: <AANLkTinLAcn5VcmVwxCOHYd7wXbptPK1PR-M0UfG469u@mail.gmail.com>
References: <1285163872-20910-1-git-send-email-albert.aribaud@free.fr>	<1285163872-20910-2-git-send-email-albert.aribaud@free.fr>	<AANLkTinK1PWePkBaBHEnxcyV6eOwtjZb-7+kA-JhHM9Z@mail.gmail.com>	<4C9A540D.3020205@free.fr>	<AANLkTimxhwaUbauQYfnGA+un3Pma8k6mhMEw7iDsZvkR@mail.gmail.com>	<4C9B88B9.3060702@free.fr>	<AANLkTin1qG2QD5cWohv5FC-W5OfgMcXDNbT_6UT-JJ6i@mail.gmail.com>	<4C9BC710.7040502@free.fr>
	<AANLkTinLAcn5VcmVwxCOHYd7wXbptPK1PR-M0UfG469u@mail.gmail.com>
Message-ID: <4C9CCD0C.9030205@free.fr>
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

I think I have found the root cause of your issue.

I was actually lucky that it worked at all for me; and as Heiko's board 
skipped relocation, it would not trigger the bug either... My computing 
of r9/r10 in the relocate_code stage was plain wrong, based on the FLASH 
execution location, not the RAM one. :/

I'll post a V2 patch set within the next hour -- Heiko, this second 
patchset has your fixes except I mouved the r9/r10 recomputation inside 
the #ifndef CONFIG_SKIP_RELOCATE_UBOOT, the logic being that if we can 
skip relocation, then we are already running at the right location and 
we don't need fixing the pic base either.

Ben and Heiko, please (re-)test the patch when it's out -- sorry for the 
extra workload.

Amicalement,
-- 
Albert.