Re: [RFC][PATCH 03/17] x86/cpu: Make srso_untrain_ret consistent

[Borislav Petkov]

On Wed, Aug 09, 2023 at 09:12:21AM +0200, Peter Zijlstra wrote:
> This does change srso_untrain_ret a little to be more consistent with
> srso_alias_untrain_ret (and zen_untrain_ret). Specifically I made
> srso_untrain_ret tail-call the srso_return_thunk, instead of doing the
> call directly. This matches how srso_alias_untrain_ret amd
> zen_untrain_ret also tail-call their respective return_thunk.
> 
> If this is a problem this can be easily fixed and a comment added to
> explain -- but this way they all end with a tail-call to their own
> return-thunk, which is nice and consistent.
> 
> Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
> ---
>  arch/x86/lib/retpoline.S |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> --- a/arch/x86/lib/retpoline.S
> +++ b/arch/x86/lib/retpoline.S
> @@ -262,7 +262,7 @@ SYM_INNER_LABEL(srso_safe_ret, SYM_L_GLO
>  	int3
>  	/* end of movabs */
>  	lfence
> -	call srso_safe_ret
> +	jmp srso_return_thunk
>  	int3
>  SYM_CODE_END(srso_safe_ret)
>  SYM_FUNC_END(srso_untrain_ret)

I don't see a problem with it but I'd let David comment here.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette