[PATCH] config/arm: fix 32-bit cross compile with latest GCC

[Wathsala Vithanage]

On 1/14/26 11:13, Bruce Richardson wrote:
> When building for 32-bit arm with GCC versions >=11 and using the
> DPDK-provided cross-file, the initial configuration step fails  The
> error given is:
>
>    config/meson.build:189:8: ERROR: Problem encountered: \
> 		Compiler does not support "armv8-a" arch flag.
>
> but the real problem is actually an fpu-related error which is explained
> in the meson log file:
>
>    cc1: error: '-mfloat-abi=hard': selected architecture lacks an FPU
>
> This error can be fixed by specifying an fpu explicitly, and a number of
> options will work, e.g. -mfpu=neon, -mfpu=fp-armv8. Choosing the latter
> here to avoid mandating neon by default.
>
> Fixes: e754875c296c ("config/arm: add aarch32 cross-compilation")
> Cc: stable at dpdk.org
>
> Signed-off-by: Bruce Richardson <bruce.richardson at intel.com>

Reviewed-by: Wathsala Vithanage <wathsala.vithanage at arm.com>

>
> ---
> This patch should fix errors seen in the CI when building for 32-bit arm
> on Ubuntu 24.04 and later. CI systems running older Ubuntu using GCC 9
> do not show this error, which is why it hasn't been an issue till
> recently.
> ---
>   config/arm/arm32_armv8_linux_gcc | 3 +++
>   1 file changed, 3 insertions(+)
>
> diff --git a/config/arm/arm32_armv8_linux_gcc b/config/arm/arm32_armv8_linux_gcc
> index abcb182b16..231c9ac7a7 100644
> --- a/config/arm/arm32_armv8_linux_gcc
> +++ b/config/arm/arm32_armv8_linux_gcc
> @@ -13,5 +13,8 @@ cpu_family = 'aarch32'
>   cpu = 'armv8-a'
>   endian = 'little'
>   
> +[built-in options]
> +c_args = '-mfpu=fp-armv8'
> +
>   [properties]
>   platform = 'generic_aarch32'