Re: [PATCH v10 4/6] x86/sgx: Define error codes for use by ENCLS[EUPDATESVN]

[Dave Hansen <dave.hansen@intel.com>]

On 8/1/25 04:25, Elena Reshetova wrote:
> Add error codes for ENCLS[EUPDATESVN], then SGX CPUSVN update
> process can know the execution state of EUPDATESVN and notify
> userspace.
> 
> Signed-off-by: Elena Reshetova <elena.reshetova@intel.com>
> ---
>  arch/x86/include/asm/sgx.h | 37 ++++++++++++++++++++++---------------
>  1 file changed, 22 insertions(+), 15 deletions(-)
> 
> diff --git a/arch/x86/include/asm/sgx.h b/arch/x86/include/asm/sgx.h
> index 6a0069761508..1abf1461fab6 100644
> --- a/arch/x86/include/asm/sgx.h
> +++ b/arch/x86/include/asm/sgx.h
> @@ -28,21 +28,22 @@
>  #define SGX_CPUID_EPC_MASK	GENMASK(3, 0)
>  
>  enum sgx_encls_function {
> -	ECREATE	= 0x00,
> -	EADD	= 0x01,
> -	EINIT	= 0x02,
> -	EREMOVE	= 0x03,
> -	EDGBRD	= 0x04,
> -	EDGBWR	= 0x05,
> -	EEXTEND	= 0x06,
> -	ELDU	= 0x08,
> -	EBLOCK	= 0x09,
> -	EPA	= 0x0A,
> -	EWB	= 0x0B,
> -	ETRACK	= 0x0C,
> -	EAUG	= 0x0D,
> -	EMODPR	= 0x0E,
> -	EMODT	= 0x0F,
> +	ECREATE		= 0x00,
> +	EADD		= 0x01,
> +	EINIT		= 0x02,
> +	EREMOVE		= 0x03,
> +	EDGBRD		= 0x04,
> +	EDGBWR		= 0x05,
> +	EEXTEND		= 0x06,
> +	ELDU		= 0x08,
> +	EBLOCK		= 0x09,
> +	EPA		= 0x0A,
> +	EWB		= 0x0B,
> +	ETRACK		= 0x0C,
> +	EAUG		= 0x0D,
> +	EMODPR		= 0x0E,
> +	EMODT		= 0x0F,
> +	EUPDATESVN	= 0x18,
>  };

This update is not consistent with the changelog nor the patch subject.

>  /**
> @@ -73,6 +74,10 @@ enum sgx_encls_function {
>   *				public key does not match IA32_SGXLEPUBKEYHASH.
>   * %SGX_PAGE_NOT_MODIFIABLE:	The EPC page cannot be modified because it
>   *				is in the PENDING or MODIFIED state.
> + * %SGX_INSUFFICIENT_ENTROPY:	Insufficient entropy in RNG.
> + * %SGX_NO_UPDATE:		EUPDATESVN was successful, but CPUSVN was not
> + *				updated because current SVN was not newer than
> + *				CPUSVN.

This comment bothers me. This is an *ERROR* code. It means that
EUPDATESVN was *NOT* successful. It failed. It didn't do an update.

Now, it's not a _bad_ error code. It's kinda like read() returning 0.
It's a "no harm no foul" kind of thing. But it's *NOT* success.

Ideally, we find a way to relay this in a very succinct way.