On Sunday 25 October 2009 17:55:07 Wolfgang Denk wrote: > Mike Frysinger wrote: > > i'm attaching two patches here. since we're past the merge window but > > before rc1, i dont know how invasive you want to get. > > > > the first one restores env_embedded.o building for certain config options > > (even though it'll only produce a 0 byte file). if you want to be > > cautious for this release, then i guess we can merge just this patch. > > > > the second one attempts to clean up env_embedded.o in all linker scripts > > where the board would only end up with a 0 byte file. obviously i cant > > test any of these since i dont have the hardware, but the logic seems > > straight forward. if you want to stay cautious, this would go into the > > next branch for start of next merge window. > > > > or just merge the 2nd patch only and assume that people who dont test the > > rc1+ are dead boards anyways. i got some build errors even after these > > fixes, but they seem unrelated to my env_embedded changes as they have to > > do with sections filling up & overflowing with my gcc-4.1.1 compiler. > > Hm... even though I like patch #2 more than the other one (which is > why I applied #2), there is a strange phenomenon remaining: > > > With patch #2 applied, I get these errors for the acadia_nand, > bamboo_nand, canyonlands_nand, and kilauea_nand configurations: > > -> ./MAKEALL kilauea_nand > Configuring for kilauea_nand board... > In file included from include/common.h:724: > /home/wd/git/u-boot/work/include/environment.h:117:1: warning: > "CONFIG_ENV_IS_EMBEDDED" redefined In file included from > /home/wd/git/u-boot/work/include/config.h:4, from include/common.h:37: > /home/wd/git/u-boot/work/include/configs/kilauea.h:127:1: warning: this > is > the location of the previous definition > > But when I remove the CONFIG_ENV_IS_EMBEDDED definition from line 127 > of include/configs/kilauea.h, then I get this: > > ./MAKEALL kilauea kilauea_nand > Configuring for kilauea board... > text data bss dec hex filename > 287816 24656 46548 359020 57a6c /work/wd/tmp-ppc/u-boot > Configuring for kilauea_nand board... > ppc_6xx-ld: cannot find common/env_embedded.o > make: *** [/work/wd/tmp-ppc/u-boot] Error 1 > > There seems to be still a logic error ...
should be harmless, but obviously it's ugly. the attached patch should fix things up. the related issue is the one i raised in the other env embedded thread wrt the meaning of "ENV_IS_EMBEDDED" ... -mike
From bce12bf90fdddaec153156ecf2ecd582faecb59b Mon Sep 17 00:00:00 2001 From: Mike Frysinger <[email protected]> Date: Mon, 26 Oct 2009 04:32:51 -0400 Subject: [PATCH] env: avoid redefining CONFIG_ENV_IS_EMBEDDED For NAND boards that embed the environment, they will often set up the CONFIG_ENV_IS_EMBEDDED manually. In that case, we should avoid redefining it for dependency purposes so we don't hit warnings like: Configuring for kilauea_nand board... In file included from include/common.h:724: /home/wd/git/u-boot/work/include/environment.h:117:1: warning: "CONFIG_ENV_IS_EMBEDDED" redefined In file included from /home/wd/git/u-boot/work/include/config.h:4, from include/common.h:37: /home/wd/git/u-boot/work/include/configs/kilauea.h:127:1: warning: this is the location of the previous definition Reported-by: Wolfgang Denk <[email protected]> Signed-off-by: Mike Frysinger <[email protected]> --- include/environment.h | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/include/environment.h b/include/environment.h index b9924fd..9832820 100644 --- a/include/environment.h +++ b/include/environment.h @@ -113,7 +113,7 @@ /* The build system likes to know if the env is embedded */ #ifdef DO_DEPS_ONLY -# ifdef ENV_IS_EMBEDDED +# if defined(ENV_IS_EMBEDDED) && !defined(CONFIG_ENV_IS_EMBEDDED) # define CONFIG_ENV_IS_EMBEDDED # endif #endif -- 1.6.5.1
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ U-Boot mailing list [email protected] http://lists.denx.de/mailman/listinfo/u-boot
