[PATCH 3/4] AVR32: remove sub-arch/variants selection from menuconfig

Hans-Christian Egtvedt hans-christian.egtvedt at atmel.com
Tue Jan 26 13:00:51 UTC 2010


On Mon, 25 Jan 2010 21:49:28 +0100
"Yann E. MORIN" <yann.morin.1998 at anciens.enib.fr> wrote:

> The precedently only supported AVR32 target did select ARCH_HAS_MMU.
> Which means that the noMMU AVR32 case is either not coded, or not
> tested. So this patch also forces ARCH_HAS_MMU when AVR32 is selected
> (the user is still is command whether to use or not to use the MMU).
> 
> Signed-off-by: "Yann E. MORIN" <yann.morin.1998 at anciens.enib.fr>
> ---
>  Rules.mak                  |    1 -
>  extra/Configs/Config.avr32 |   11 +----------
>  2 files changed, 1 insertions(+), 11 deletions(-)
> 
> diff --git a/Rules.mak b/Rules.mak
> index 859377f..bf9ba01 100644
> --- a/Rules.mak
> +++ b/Rules.mak
> @@ -437,7 +437,6 @@ ifeq ($(TARGET_ARCH),frv)
>  endif
>  
>  ifeq ($(strip $(TARGET_ARCH)),avr32)
> -       CPU_CFLAGS-$(CONFIG_AVR32_AP7)  += -march=ap

Any reason you nuke this compiler flag? I would like to keep it,
although it could be selected always as long as target arch is avr32.

>         CPU_CFLAGS-$(CONFIG_LINKRELAX)  += -mrelax
>         CPU_LDFLAGS-$(CONFIG_LINKRELAX) += --relax
>  endif
> diff --git a/extra/Configs/Config.avr32 b/extra/Configs/Config.avr32
> index cbadb4c..8e348c6 100644
> --- a/extra/Configs/Config.avr32
> +++ b/extra/Configs/Config.avr32
> @@ -10,19 +10,10 @@ config TARGET_ARCH
>  config FORCE_OPTIONS_FOR_ARCH
>  	bool
>  	default y
> +	select ARCH_HAS_MMU
>  	select ARCH_BIG_ENDIAN
>  	select FORCE_SHAREABLE_TEXT_SEGMENTS
>  
> -choice
> -	prompt "Target CPU Type"
> -	default CONFIG_AVR32_AP7
> -
> -config CONFIG_AVR32_AP7
> -	bool "AVR32 AP7"
> -	select ARCH_HAS_MMU
> -
> -endchoice
> -
>  config LINKRELAX
>  	bool "Enable linker optimizations"
>  	default y

This part is fine by me. There is an option of supporting UC3
architecture as well, but it is highly experimental, so lets just keep
it simple until there is something actually working.

-- 
Best regards,
Hans-Christian Egtvedt


More information about the uClibc mailing list