Skip to contentSkip to conversation

x86/spec-ctrl: Knobs for STIBP and PSFD, and follow hardware STIBP hint

PATCH In review Andrew Cooper 2 participants 2 messages Mar 16, 2022 – Mar 17, 2022

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86/spec-ctrl: Knobs for STIBP and PSFD, and follow hardware STIBP hint

STIBP and PSFD are slightly weird bits, because they're both implied by other bits in MSR_SPEC_CTRL. Add fine grain controls for them, and take the implications into account when setting IBRS/SSBD.

Rearrange the IBPB text/variables/logic to keep all the MSR_SPEC_CTRL bits together, for consistency.

However, AMD have a hardware hint CPUID bit recommending that STIBP be set uniaterally. This is advertised on Zen3, so follow the recommendation. This is the only default change.

---
CC: Jan Beulich <[email protected]>
CC: Roger Pau Monné <[email protected]>
CC: Wei Liu <[email protected]>

---
 docs/misc/xen-command-line.pandoc | 20 +++++++++----
 xen/arch/x86/spec_ctrl.c          | 59 ++++++++++++++++++++++++++++++++++++---
 2 files changed, 70 insertions(+), 9 deletions(-)
docs/misc/xen-command-line.pandoc +15 -5
index 1dc7e1ca0706..6fa1fe97aeeb 100644
@@ -2254,8 +2254,8 @@ By default SSBD will be mitigated at runtime (i.e `ssbd=runtime`).
 
 ### spec-ctrl (x86)
 > `= List of [ <bool>, xen=<bool>, {pv,hvm,msr-sc,rsb,md-clear}=<bool>,
->              bti-thunk=retpoline|lfence|jmp, {ibrs,ibpb,ssbd,eager-fpu,
->              l1d-flush,branch-harden,srb-lock}=<bool> ]`
+>              bti-thunk=retpoline|lfence|jmp, {ibrs,stibp,ssbd,psfd,ibpb,
+>              eager-fpu,l1d-flush,branch-harden,srb-lock}=<bool> ]`
 
 Controls for speculative execution sidechannel mitigations.  By default, Xen
 will pick the most appropriate mitigations based on compiled in support,
@@ -2306,9 +2306,10 @@ On hardware supporting IBRS (Indirect Branch Restricted Speculation), the
 If Xen is not using IBRS itself, functionality is still set up so IBRS can be
 virtualised for guests.
 
-On hardware supporting IBPB (Indirect Branch Prediction Barrier), the `ibpb=`
-option can be used to force (the default) or prevent Xen from issuing branch
-prediction barriers on vcpu context switches.
+On hardware supporting STIBP (Single Thread Indirect Branch Predictors), the
+`stibp=` option can be used to force or prevent Xen using the feature itself.
+By default, Xen will use STIBP when IBRS is in use (IBRS implies STIBP), and
+when hardware hints recommend using it as a blanket setting.
 
 On hardware supporting SSBD (Speculative Store Bypass Disable), the `ssbd=`
 option can be used to force or prevent Xen using the feature itself.  On AMD
@@ -2316,6 +2317,15 @@ hardware, this is a global option applied at boot, and not virtualised for
 guest use.  On Intel hardware, the feature is virtualised for guests,
 independently of Xen's choice of setting.
 
+On hardware supporting PSFD (Predictive Store Forwarding Disable), the `psfd=`
+option can be used to force or prevent Xen using the feature itself.  By
+default, Xen will not use PSFD.  PSFD is implied by SSBD, and SSBD is off by
+default.
+
+On hardware supporting IBPB (Indirect Branch Prediction Barrier), the `ibpb=`
+option can be used to force (the default) or prevent Xen from issuing branch
+prediction barriers on vcpu context switches.
+
 On all hardware, the `eager-fpu=` option can be used to force or prevent Xen
 from using fully eager FPU context switches.  This is currently implemented as
 a global control.  By default, Xen will choose to use fully eager context
xen/arch/x86/spec_ctrl.c +55 -4

101 lines not on this page · raw message

Signature · 1 line
2.11.0
Reviewed-by on v1 1/1
#1 lore raw
Quoted 4 lines · Andrew Cooper

STIBP and PSFD are slightly weird bits, because they're both implied by other bits in MSR_SPEC_CTRL. Add fine grain controls for them, and take the implications into account when setting IBRS/SSBD.

Quoted 8 lines

Rearrange the IBPB text/variables/logic to keep all the MSR_SPEC_CTRL bits together, for consistency.

However, AMD have a hardware hint CPUID bit recommending that STIBP be set uniaterally. This is advertised on Zen3, so follow the recommendation. This is the only default change.

In principle

but I have two comments:

Quoted 11 lines
xen/arch/x86/spec_ctrl.c +3 -2
@@ -170,12 +174,18 @@ static int __init cf_check parse_spec_ctrl(const char *s)
             else
                 rc = -EINVAL;
         }
+
         else if ( (val = parse_boolean("ibrs", s, ss)) >= 0 )
             opt_ibrs = val;
-        else if ( (val = parse_boolean("ibpb", s, ss)) >= 0 )
-            opt_ibpb = val;
+        else if ( (val = parse_boolean("stibp", s, ss)) >= 0 )
+            opt_stibp = val;

xen/arch/x86/spec_ctrl.c static int __init cf_check parse_spec_ctrl(const char *s) · line 191

Quoted 10 lines
xen/arch/x86/spec_ctrl.c +5 -0
         else if ( (val = parse_boolean("ssbd", s, ss)) >= 0 )
             opt_ssbd = val;
+        else if ( (val = parse_boolean("psfd", s, ss)) >= 0 )
+            opt_psfd = val;
+
+        else if ( (val = parse_boolean("ibpb", s, ss)) >= 0 )
+            opt_ibpb = val;
         else if ( (val = parse_boolean("eager-fpu", s, ss)) >= 0 )
             opt_eager_fpu = val;
         else if ( (val = parse_boolean("l1d-flush", s, ss)) >= 0 )

Personally I find blank lines ahead of "else if" misleading. Could I talk you into moving ibrs+stibp and ssbd+psfd close to the end of this (immediately ahead of "else"), prefixing each pair with a comment about one feature implying the other (and hence the comments replacing the blank lines)?

Otoh I notice that we already have blank lines elsewhere in the middle if this block of code, but at least there they're accompanied by a comment making more obvious why there is such separation. Which means as an intermediate approach I'd be okay with no re-ordering, but with comments added.

Quoted 12 lines
xen/arch/x86/spec_ctrl.c +7 -0
@@ -1070,12 +1083,50 @@ void __init init_speculation_mitigations(void)
 
     /* If we have IBRS available, see whether we should use it. */
     if ( has_spec_ctrl && ibrs )
+    {
+        /* IBRS implies STIBP.  */
+        if ( opt_stibp == -1 )
+            opt_stibp = 1;
+
         default_xen_spec_ctrl |= SPEC_CTRL_IBRS;
+    }
+

xen/arch/x86/spec_ctrl.c void __init init_speculation_mitigations(void) · line 1103

Quoted 10 lines
xen/arch/x86/spec_ctrl.c +10 -0
+    /* Use STIBP by default if the hardware hint is set. */
+    if ( opt_stibp == -1 && boot_cpu_has(X86_FEATURE_STIBP_ALWAYS) )
+        opt_stibp = 1;
+
+    /*
+     * Otherwise, don't use STIBP by default.  It has some severe performance
+     * implications on older hardware.
+     */
+    if ( opt_stibp == -1 )
+        opt_stibp = 0;

I'd find this easier to read if written along the lines of

    if ( opt_stibp == -1 )
    {
        /*
         * Use STIBP by default if the hardware hint is set.  Otherwise,
         * don't use STIBP by default.  It has some severe performance
         * implications on older hardware.
         */
        opt_stibp = !!boot_cpu_has(X86_FEATURE_STIBP_ALWAYS);
    }

FTAOD I'm not going to insist on either adjustment.

Jan