On Sat, Jun 08, 2019 at 02:26:54AM +0100, Andre Przywara wrote: > So far we are required to always define the CONFIG_SYS_MMC_ENV_DEV > variable, even if a platform specific function overrides the weak > function that is using it. > > Check for the existence of this Kconfig variable, eliminating the need > to define a dummy value. > > Signed-off-by: Andre Przywara <andre.przyw...@arm.com> > --- > env/mmc.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/env/mmc.c b/env/mmc.c > index c3cf35d01b..122fec3af8 100644 > --- a/env/mmc.c > +++ b/env/mmc.c > @@ -124,7 +124,11 @@ __weak int mmc_get_env_addr(struct mmc *mmc, int copy, > u32 *env_addr) > > __weak int mmc_get_env_dev(void) > { > +#ifdef CONFIG_SYS_MMC_ENV_DEV > return CONFIG_SYS_MMC_ENV_DEV; > +#else > + return 0; > +#endif > } > > #ifdef CONFIG_SYS_MMC_ENV_PART
Since 0 is a valid device, I'm concerned this might lead to unintended behavior. Can we return some error code here and catch it later? -- Tom
signature.asc
Description: PGP signature
_______________________________________________ U-Boot mailing list U-Boot@lists.denx.de https://lists.denx.de/listinfo/u-boot