Re: [PATCH v4 3/3] arm: Enable FPU for M-profile

On 2026-06-24 21:58, Christophe Lyon wrote:
> Before using the FPU, we must grant access to it.  This helps when
> running the GCC testsuite on a simulator where FPU access is not
> granted by default.
> 
> This patch grants full access, we may want to consider granting only
> unprivileged access later.
> 
> While updating crt0.S, include arm-acle-compat.h and stop using
> __ARM_ARCH_7A__.
> 
> 2026-05-29  Christophe Lyon  <[email protected]>
> 
> 	* libgloss/arm/cpu-init/rdimon-aem-m.S: Enable full FPU access on
> 	M-profile.
> 	* libgloss/arm/crt0.S: Call _rdimon_hw_init_hook for M profile
> 	too.
> ---
>   libgloss/arm/cpu-init/rdimon-aem-m.S | 29 +++++++++++++++++++++++++++-
>   libgloss/arm/crt0.S                  |  9 +++------
>   2 files changed, 31 insertions(+), 7 deletions(-)
> 
> diff --git a/libgloss/arm/cpu-init/rdimon-aem-m.S b/libgloss/arm/cpu-init/rdimon-aem-m.S
> index 611849077..c495efc2d 100644
> --- a/libgloss/arm/cpu-init/rdimon-aem-m.S
> +++ b/libgloss/arm/cpu-init/rdimon-aem-m.S
> @@ -28,9 +28,36 @@
>      We define a hardfault handler, required to prevent the simulator
>      from entering into an infinite loop.  This handler is located at
>      address 0xc by default when linking with mprofile-mps2.specs which
> -   uses --section-start=.hardfault_handler_addr=0xc.  */
> +   uses --section-start=.hardfault_handler_addr=0xc.
> +
> +   Enable the FPU when __ARM_FP is set.  */
>   
>   #if  (__ARM_ARCH_PROFILE == 'M')
> +    .syntax    unified
> +    .thumb
> +
> +    @ CPU Initialisation
> +    .globl     _rdimon_hw_init_hook
> +    .type      _rdimon_hw_init_hook, %function
> +
> +_rdimon_hw_init_hook:
> +    /* Enable FPU for M-profile only .  */
> +#if __ARM_FP
> +    /* Read CPACR */
> +    ldr        r0, =0xE000ED88
> +    ldr        r1, [R0]
> +    /* Enable full access to CP10 and CP11.  */
> +    orr        r1, r1, #(0xFF << 20)
> +    str        r1, [R0]
> +#if __ARM_ARCH >= 800
> +    ldr        r0, =0xE002ED88
> +    str        r1, [R0]

Is this okay to do?
Are we sure that this rdimon-aem-m.S is never included in a non-secure application? What happens if it is included in a non-secure application, secure fault?

Kind regards,
Torbjörn

> +#endif
> +    dsb
> +    isb
> +#endif
> +    bx lr
> +
>       /* On M-profile, call abort in case of hardfault: this causes the
>          simulator to abort execution instead of going into an infinite
>          loop.  */
> diff --git a/libgloss/arm/crt0.S b/libgloss/arm/crt0.S
> index b9e768007..cad002d2b 100644
> --- a/libgloss/arm/crt0.S
> +++ b/libgloss/arm/crt0.S
> @@ -1,6 +1,7 @@
>   #include "newlib.h"
>   #include "arm.h"
>   #include "swi.h"
> +#include "arm-acle-compat.h"
>   
>   /* ANSI concatenation macros.  */
>   #define CONCAT(a, b) CONCAT2(a, b)
> @@ -250,12 +251,8 @@
>   	FUNC_START	_start
>   	FN_EH_START
>   
> -	/* __ARM_ARCH_PROFILE is defined from GCC 4.8 onwards, however __ARM_ARCH_7A
> -	has been defined since 4.2 onwards, which is when v7-a support was added
> -	and hence 'A' profile support was added in the compiler.  Allow for this
> -	file to be built with older compilers.  We only call this for A profile
> -	cores.  */
> -#if defined (__ARM_ARCH_7A__) || (__ARM_ARCH_PROFILE == 'A')
> +	/* We only call this for A and M profile cores.  */
> +#if (__ARM_ARCH_PROFILE == 'A') || (__ARM_ARCH_PROFILE == 'M')
>   /*  The init hook does not use the stack and is called before the stack has been set up.  */
>   #ifdef ARM_RDI_MONITOR
>   	bl	_rdimon_hw_init_hook