On Thu, Apr 03, 2014 at 09:12:56AM +0000, Gupta, Pekon wrote: > (re-sending as earlier mail got moderated). > > Hi Enric, > > >From: Enric Balletbo Serra [mailto:eballe...@gmail.com] > >2014-04-03 9:09 GMT+02:00 Pekon Gupta <pe...@ti.com>: > >> fixes commit e37e954eba3edb5015a0a02880d57517f57425d8 > >> OMAP3: igep00x0: Convert to ti_omap3_common.h. > >> > >> Above commit introduced compilation error when building igep0020 and > >> igep0030 > >> boards for NAND boot. It removed 'CONFIG_SYS_NAND_U_BOOT_START' which tells > >> SPL/MLO offset of u-boot.img in NAND. > >> > >> Signed-off-by: Pekon Gupta <pe...@ti.com> > >> --- > >> include/configs/omap3_igep00x0.h | 2 ++ > >> 1 file changed, 2 insertions(+) > >> > >> diff --git a/include/configs/omap3_igep00x0.h > >> b/include/configs/omap3_igep00x0.h > >> index 8b8583a..4e87d6b 100644 > >> --- a/include/configs/omap3_igep00x0.h > >> +++ b/include/configs/omap3_igep00x0.h > >> @@ -199,6 +199,8 @@ > >> #define CONFIG_SYS_NAND_ECCSIZE 512 > >> #define CONFIG_SYS_NAND_ECCBYTES 3 > >> #define CONFIG_NAND_OMAP_ECCSCHEME OMAP_ECC_HAM1_CODE_HW > >> +#define CONFIG_SYS_NAND_U_BOOT_START CONFIG_SYS_TEXT_BASE > >> +#define CONFIG_SYS_NAND_U_BOOT_OFFS 0x80000 > > > > > >These two defines are defined in ti_armv7_common.h. The > >ti_armv7_common.h is included by ti_omap3_common.h, and > >omap3_igep00x0.h includes this file. So should not be necessary define > >this. I tried to build current mainline using igep0020_nand_config and > >built without errors for me. Which branch are you using ? > > > My bad, I din't check that one of my local commits had removed above > two defines from "ti_armv7_common.h", because these are board-specific > configs, and should be in individual board-configs. > I got misled because one of your commits (e37e954) removes these two > defines and has title "OMAP3: igep00x0: Convert to ti_omap3_common.h." > > Sorry about noise, please ignore this patch. I'll merge these changes with my > other patch, where I move out board-specific configs from "ti_armv7_common.h"
Well hold up. Where things need to be in NAND depends on where the SoC is going to check for an initial loader, followed by where it's going to check for redundant copies (if applicable). So you'll have a bit of a job convincing me that we don't want to default to a likely right location (I forget off-hand if bigger NAND pages would push the location higher or not or if we'd just sacrifice total number of possible redundant copies). -- Tom
signature.asc
Description: Digital signature
_______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
