[U-Boot] [PATCH 1/2] fdt: Add new fdt_set_node_status & fdt_set_status_by_alias helpers

[Kumar Gala]

On Oct 23, 2011, at 3:49 AM, Wolfgang Denk wrote:

> Dear Kumar Gala,
> 
> In message <1318619444-2059-1-git-send-email-galak at kernel.crashing.org> you wrote:
>> From: Shengzhou Liu <Shengzhou.Liu at freescale.com>
>> 
>> Add common function fdt_set_node_status() to assist in various locations
>> that we set a nodes status.  This function utilizes the status values
>> that are part of the EPAPR spec (on power.org).
>> 
>> fdt_set_status_by_alias() is based on fdt_set_node_status() but uses an
>> alias string to identify the node to update.
>> 
>> We also add some shortcut functions to help the common cases of setting
>> "okay" and "disabled":
>> 
>> 	fdt_status_okay()
>> 	fdt_status_disabled()
>> 	fdt_status_okay_by_alias()
>> 	fdt_status_disabled_by_alias()
>> 
>> Finally, we fixup the corenet_ds ethernet code which previously had
>> a function by the same name that can be replaced with the new helpers.
>> 
>> Signed-off-by: Shengzhou Liu <Shengzhou.Liu at freescale.com>
>> Signed-off-by: Kumar Gala <galak at kernel.crashing.org>
> 
> This patch breaks a number of boards by growing the code size even for
> boards which never make use of this new stuff.
> 
> Affected boards: TQM8555 TQM8541
> 
> [Building with ELDK 4.2]:
> 
> + ./MAKEALL TQM8555
> Configuring for TQM8555 - Board: TQM85xx, Options: MPC8555,TQM8555=y,HOSTNAME=tqm8555,BOARDNAME="TQM8555"
> ppc_6xx-ld: warning: dot moved backwards before `.bss'
> ppc_6xx-ld: warning: dot moved backwards before `.bss'
> ppc_6xx-ld: u-boot: section .text lma 0xfffc0000 overlaps previous sections
> ppc_6xx-ld: u-boot: section .rodata lma 0xfffef388 overlaps previous sections
> ppc_6xx-ld: u-boot: section .reloc lma 0xffffa400 overlaps previous sections
> ppc_6xx-ld: u-boot: section .data lma 0xffffcd3c overlaps previous sections
> ppc_6xx-ld: u-boot: section .u_boot_cmd lma 0xffffea68 overlaps previous sections
> ppc_6xx-ld: u-boot: section .bootpg lma 0xfffff0dc overlaps previous sections


Is part of our issue with older gcc-4.2 toolchain that the linker doesn't garbage collect unused functions?

- k