Re: [PATCH 3/7] x86/bugs: Fix BHI handling of RRSBA

[Andrew Cooper]

On 11/04/2024 6:40 am, Josh Poimboeuf wrote:
> The ARCH_CAP_RRSBA check isn't correct: RRSBA may have already been
> disabled by the Spectre v2 mitigation (or can otherwise be disabled by
> the BHI mitigation itself if needed).  In that case retpolines are fine.
>
> Fixes: ec9404e40e8f ("x86/bhi: Add BHI mitigation knob")
> Signed-off-by: Josh Poimboeuf <jpoimboe@xxxxxxxxxx>
> ---
>  arch/x86/kernel/cpu/bugs.c | 30 ++++++++++++++++++------------
>  1 file changed, 18 insertions(+), 12 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
> index 27d6d64eeec3..0755600d5d18 100644
> --- a/arch/x86/kernel/cpu/bugs.c
> +++ b/arch/x86/kernel/cpu/bugs.c
> @@ -1538,20 +1538,25 @@ static enum spectre_v2_mitigation __init spectre_v2_select_retpoline(void)
>  	return SPECTRE_V2_RETPOLINE;
>  }
>  
> +static bool __ro_after_init rrsba_disabled;
> +
>  /* Disable in-kernel use of non-RSB RET predictors */
>  static void __init spec_ctrl_disable_kernel_rrsba(void)
>  {
> -	u64 ia32_cap;
> +	if (rrsba_disabled)
> +		return;
> +
> +	if (!(ia32_cap & ARCH_CAP_RRSBA)) {
> +		rrsba_disabled = true;
> +		return;
> +	}

You'll take this path if you have out-of-date microcode.

RRSBA is only enumerated from September last year, IIRC.  (Definitely
from this point on some CPUs.)

When RRSBA was introduced, I was under the (false) impression that all
eIBRS systems suffered RRSBA, but it turns out that select parts
(ICX,TGL,RKL) are non-RRSBA despite not having RRSBA_CTRL.

~Andrew