imx8mq+4.14.78+yocto:aarch64-poky-linux-gcc: error: unrecognized command line option ‘-mfpu=neon’

[传策薛]

Hi community:

 I use imx8mq+4.14.78+yocto bsp，and I "bitbake fsl-image-qt5-validation-imx -c populate_sdk"， get the sdk。now I want to compile a ".c" file with options "-mfpu=neon -mfloat-abi=hard"。

but I get follow error:

aarch64-poky-linux-gcc --sysroot=/opt/fsl-imx-wayland/4.14-sumo/sysroots/aarch64-poky-linux -mfpu=neon -mfloat-abi=soft -O3 test_vfp_hard.c -o TEST_VFP_SOFT
aarch64-poky-linux-gcc: error: unrecognized command line option ‘-mfpu=neon’
aarch64-poky-linux-gcc: error: unrecognized command line option ‘-mfloat-abi=soft’
make: *** [soft] Error 1

How to fix it?

Thanks a lot.

[diegoadrian]

Hello,

Apparently, you need to export some flags on your makefile:

export CFLAGS="-mcpu=cortex-a53 -mfloat-abi=hard -mfpu=neon-fp-armv8 -mneon-for-64bits"

You can find more information on the below web page:

https://gist.github.com/fm4dd/c663217935dc17f0fc73c9c81b0aa845

Hope this information can help you.

Best regards,

Diego.

[传策薛]

I tried follows:

It seems aarch64 does not support "-mfloat-abi=hard"，"-mfpu=neon-fp-armv8"， "-mneon-for-64bits".

Using the GNU Compiler Collection (GCC): AArch64 Options 

I have found some ARM64 information. The aarch64-poky-linux-xx compiler supports hardware floating point arithmetic by default. Is it correct?

[diegoadrian]

Hello,

Unfortunately, we are unable to assist you with information related to the ARM processor, mostly because we do not have access to that information. I have found the below, hope it can help you. I apologize for the inconveniences this could give you.  

https://developer.arm.com/docs/100748/latest/using-common-compiler-options/selecting-floating-point-...

Best regards,

Diego.