Re: [RFC PATCH v5 2/2] MIPS: optional floating point support

[David Daney <ddaney.cavm@gmail.com>]

On 04/07/2014 10:53 AM, Manuel Lauss wrote:
[...]
> +#else	/* no CONFIG_MIPS_FPU_SUPPORT */
> +static inline int do_dsemulret(struct pt_regs *xcp)
> +{
> +	return 0;	/* 0 means error, should never get here anyway */
> +}
> +
> +static inline int fpu_emulator_cop1Handler(struct pt_regs *xcp,
> +				struct mips_fpu_struct *ctx, int has_fpu,
> +				void *__user *fault_addr)
> +{
> +	return SIGILL;	/* we don't speak MIPS FPU */

In a message in the other branch, Ralf suggested emitting very ugly 
messages in the case that something tries to use the FPU.

This code doesn't quite satisfy that request.


> +}
> +#endif	/* CONFIG_MIPS_FPU_SUPPORT */
> +
>   int process_fpemu_return(int sig, void __user *fault_addr);
>
>   /*
>