Re: [PATCH v2 1/2] x86/speculation: Allow enabling STIBP with legacy IBRS

[KP Singh]

On Tue, Feb 21, 2023 at 7:07 PM Pawan Gupta
<pawan.kumar.gupta@xxxxxxxxxxxxxxx> wrote:
>
> On Tue, Feb 21, 2023 at 07:49:07PM +0100, KP Singh wrote:
> > Setting the IBRS bit implicitly enables STIBP to protect against
> > cross-thread branch target injection. With enhanced IBRS, the bit it set
> > once and is not cleared again. However, on CPUs with just legacy IBRS,
> > IBRS bit set on user -> kernel and cleared on kernel -> user (a.k.a
> > KERNEL_IBRS). Clearing this bit also disables the implicitly enabled
> > STIBP, thus requiring some form of cross-thread protection in userspace.
> >
> > Enable STIBP, either opt-in via prctl or seccomp, or always on depending
> > on the choice of mitigation selected via spectre_v2_user.
> >
> > Reported-by: José Oliveira <joseloliveira11@xxxxxxxxx>
> > Reported-by: Rodrigo Branco <rodrigo@xxxxxxxxxxxxxxxxx>
> > Reviewed-by: Alexandra Sandulescu <aesa@xxxxxxxxxx>
> > Fixes: 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")
> > Cc: stable@xxxxxxxxxxxxxxx
> > Signed-off-by: KP Singh <kpsingh@xxxxxxxxxx>
> > ---
> >  arch/x86/kernel/cpu/bugs.c | 33 ++++++++++++++++++++++-----------
> >  1 file changed, 22 insertions(+), 11 deletions(-)
> >
> > diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
> > index 85168740f76a..5be6075d8e36 100644
> > --- a/arch/x86/kernel/cpu/bugs.c
> > +++ b/arch/x86/kernel/cpu/bugs.c
> > @@ -1124,14 +1124,30 @@ spectre_v2_parse_user_cmdline(void)
> >       return SPECTRE_V2_USER_CMD_AUTO;
> >  }
> >
> > -static inline bool spectre_v2_in_ibrs_mode(enum spectre_v2_mitigation mode)
> > +static inline bool spectre_v2_in_eibrs_mode(enum spectre_v2_mitigation mode)
> >  {
> > -     return mode == SPECTRE_V2_IBRS ||
> > -            mode == SPECTRE_V2_EIBRS ||
> > +     return mode == SPECTRE_V2_EIBRS ||
> >              mode == SPECTRE_V2_EIBRS_RETPOLINE ||
> >              mode == SPECTRE_V2_EIBRS_LFENCE;
> >  }
> >
> > +static inline bool spectre_v2_in_ibrs_mode(enum spectre_v2_mitigation mode)
> > +{
> > +     return spectre_v2_in_eibrs_mode(mode) || mode == SPECTRE_V2_IBRS;
> > +}
> > +
> > +static inline bool spectre_v2_user_needs_stibp(enum spectre_v2_mitigation mode)
> > +{
> > +     /*
> > +      * enhanced IBRS also protects against user-mode attacks as the IBRS bit
>
> Maybe:
>          * Enhanced IBRS mode also protects against cross-thread user-to-user
>          * attacks as the IBRS bit

updated, thanks!

>
> > +      * remains always set which implicitly enables cross-thread protections.
> > +      * However, In legacy IBRS mode, the IBRS bit is set only in kernel
> > +      * and cleared on return to userspace. This disables the implicit
> > +      * cross-thread protections and STIBP is needed.
> > +      */
> > +     return !spectre_v2_in_eibrs_mode(mode);
> > +}
> > +
> >  static void __init
> >  spectre_v2_user_select_mitigation(void)
> >  {
> > @@ -1193,13 +1209,8 @@ spectre_v2_user_select_mitigation(void)
> >                       "always-on" : "conditional");
> >       }
> >
> > -     /*
> > -      * If no STIBP, IBRS or enhanced IBRS is enabled, or SMT impossible,
> > -      * STIBP is not required.
> > -      */
> > -     if (!boot_cpu_has(X86_FEATURE_STIBP) ||
> > -         !smt_possible ||
> > -         spectre_v2_in_ibrs_mode(spectre_v2_enabled))
> > +     if (!boot_cpu_has(X86_FEATURE_STIBP) || !smt_possible ||
> > +         !spectre_v2_user_needs_stibp(spectre_v2_enabled))
>
> As pointed out in other discussions, it will be great if can get rid of
> eIBRS check, and do what the user asked for; or atleast print a warning

I think I will keep it as pr_info as, with eIBRS, the user does not
really need STIBP and the mitigation is still effective.

> about not setting STIBP bit explicitly.

That is a bit more complicated as, for now, the user is not really
exposed to STIBP explicitly yet.

{ "auto", SPECTRE_V2_USER_CMD_AUTO, false },
{ "off", SPECTRE_V2_USER_CMD_NONE, false },
{ "on", SPECTRE_V2_USER_CMD_FORCE, true },
{ "prctl", SPECTRE_V2_USER_CMD_PRCTL, false },
{ "prctl,ibpb", SPECTRE_V2_USER_CMD_PRCTL_IBPB, false },
{ "seccomp", SPECTRE_V2_USER_CMD_SECCOMP, false },
{ "seccomp,ibpb", SPECTRE_V2_USER_CMD_SECCOMP_IBPB, false },

I would prefer to do it as a follow up and fix this bug first.

It's a bit gnarly and I think we really need to think about the
options that are exposed to the user [especially in light of Intel /
AMD subtelties].

With the current patch the userspace is still getting working V2
mitigations on both dimensions time (Process A followed by Process B
where A does BTI on the subsequent B that are flushed via an IBPB) and
space (i.e. cross-thread branch target injection) whenever necessary.




>
> >               return;
> >
> >       /*
> > @@ -2327,7 +2338,7 @@ static ssize_t mmio_stale_data_show_state(char *buf)
> >
> >  static char *stibp_state(void)
> >  {
> > -     if (spectre_v2_in_ibrs_mode(spectre_v2_enabled))
> > +     if (!spectre_v2_user_needs_stibp(spectre_v2_enabled))
>
> Decoupling STIBP and eIBRS will also get rid of this check.
>
> >               return "";