On Wed, Nov 13, 2013 at 11:46:49AM +0000, Jonathan Austin wrote:

Hi Ming,

On 13/11/13 11:15, Ming Lei wrote:

quoted

From: Jonathan Austin <redacted>

Set CONFIG_PAGE_OFFSET as PHYS_OFFSET when !MMU so that
we can keep CONFIG_PAGE_OFFSET same with PAGE_OFFSET.

I'm not so sure about this commit message - I don't think
it makes sense...Should it not be more like:

------->8-------
Subject: ARM: set PAGE_OFFSET = PHYS_OFFSET if !MMU

With !MMU we still set PAGE_OFFSET to 0xC0000000, which
makes little sense and causes a regression on !MMU after
f6537f2f0eba4eba ("scripts/kallsyms: filter symbols not
kernel address space")

This patch sets PAGE_OFFSET to PHYS_OFFSET in the !MMU
case, fixing the regression (Tested on Cortex-R7/Vexpress)
--------8<---------

I like this commit log better, too.

As CONFIG_PAGE_OFFSET is only used in scripts/kallsyms.c and
scripts/link-vmlinux.sh for !MMU.

Booting on efm32 with this patch works just fine, so if you want add my
Tested-by. But I'm unsure how valuable it is as I don't see a breakage
without it.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-K?nig            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |