x86/spec-ctrl: Introduce an option to control L1D_FLUSH for HVM HAP guests

This mitigation requires up-to-date microcode, and is enabled by default on
affected hardware if available, and is used for HVM guests

The default for SMT/Hyperthreading is far more complicated to reason about,
not least because we don't know if the user is going to want to run any HVM
guests to begin with.  If a explicit default isn't given, nag the user to
perform a risk assessment and choose an explicit default, and leave other
configuration to the toolstack.

This is part of XSA-273 / CVE-2018-3620.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 3bd36952dab60290f33d6791070b57920e10754b)

diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown
index 158b5bb919e8acd6d8161eff5504d68154f950cf..57ef18194ae9760bffd03ffb7a9a6b80061acfbd 100644 (file)
--- a/docs/misc/xen-command-line.markdown
+++ b/docs/misc/xen-command-line.markdown
@@ -1791,7 +1791,8 @@ false disable the quirk workaround, which is also the default.
 
 ### spec-ctrl (x86)
 > `= List of [ <bool>, xen=<bool>, {pv,hvm,msr-sc,rsb}=<bool>,
->              bti-thunk=retpoline|lfence|jmp, {ibrs,ibpb,ssbd,eager-fpu}=<bool> ]`
+>              bti-thunk=retpoline|lfence|jmp, {ibrs,ibpb,ssbd,eager-fpu,
+>              l1d-flush}=<bool> ]`
 
 Controls for speculative execution sidechannel mitigations.  By default, Xen
 will pick the most appropriate mitigations based on compiled in support,
@@ -1846,6 +1847,12 @@ 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
 switches on hardware believed to speculate past #NM exceptions.
 
+On hardware supporting L1D_FLUSH, the `l1d-flush=` option can be used to force
+or prevent Xen from issuing an L1 data cache flush on each VMEntry.
+Irrespective of Xen's setting, the feature is virtualised for HVM guests to
+use.  By default, Xen will enable this mitigation on hardware believed to be
+vulnerable to L1TF.
+
 ### sync\_console
 > `= <boolean>`
 

diff --git a/xen/arch/x86/hvm/vmx/vmcs.c b/xen/arch/x86/hvm/vmx/vmcs.c
index 30a33dd0bda5ff55bb7379ac2891849d4969dae7..2ba0c408086254be215d4aafca67b9b1146a4163 100644 (file)
--- a/xen/arch/x86/hvm/vmx/vmcs.c
+++ b/xen/arch/x86/hvm/vmx/vmcs.c
@@ -38,6 +38,7 @@
 #include <asm/flushtlb.h>
 #include <asm/monitor.h>
 #include <asm/shadow.h>
+#include <asm/spec_ctrl.h>
 #include <asm/tboot.h>
 #include <asm/apic.h>
 
@@ -1274,6 +1275,10 @@ static int construct_vmcs(struct vcpu *v)
 
     vmx_vlapic_msr_changed(v);
 
+    if ( opt_l1d_flush && paging_mode_hap(d) )
+        rc = vmx_add_msr(v, MSR_FLUSH_CMD, FLUSH_CMD_L1D,
+                         VMX_MSR_GUEST_LOADONLY);
+
  out:
     vmx_vmcs_exit(v);
 

diff --git a/xen/arch/x86/spec_ctrl.c b/xen/arch/x86/spec_ctrl.c
index 9bcc2b6adcb3d249800ccd864334060ebc506cfe..59baebb9595b9ff07d8a8fd8d5b6c61a8b2c88a9 100644 (file)
--- a/xen/arch/x86/spec_ctrl.c
+++ b/xen/arch/x86/spec_ctrl.c
@@ -19,11 +19,13 @@
 #include <xen/errno.h>
 #include <xen/init.h>
 #include <xen/lib.h>
+#include <xen/warning.h>
 
 #include <asm/microcode.h>
 #include <asm/msr.h>
 #include <asm/processor.h>
 #include <asm/pv/shim.h>
+#include <asm/setup.h>
 #include <asm/spec_ctrl.h>
 #include <asm/spec_ctrl_asm.h>
 
@@ -46,6 +48,7 @@ static int8_t __initdata opt_ibrs = -1;
 bool __read_mostly opt_ibpb = true;
 bool __read_mostly opt_ssbd = false;
 int8_t __read_mostly opt_eager_fpu = -1;
+int8_t __read_mostly opt_l1d_flush = -1;
 
 bool __initdata bsp_delay_spec_ctrl;
 uint8_t __read_mostly default_xen_spec_ctrl;
@@ -139,6 +142,7 @@ static int __init parse_spec_ctrl(const char *s)
             opt_ibrs = 0;
             opt_ibpb = false;
             opt_ssbd = false;
+            opt_l1d_flush = 0;
         }
         else if ( val > 0 )
             rc = -EINVAL;
@@ -194,6 +198,8 @@ static int __init parse_spec_ctrl(const char *s)
             opt_ssbd = 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 )
+            opt_l1d_flush = val;
         else
             rc = -EINVAL;
 
@@ -290,7 +296,7 @@ static void __init print_details(enum ind_thunk thunk, uint64_t caps)
                "\n");
 
     /* Settings for Xen's protection, irrespective of guests. */
-    printk("  Xen settings: BTI-Thunk %s, SPEC_CTRL: %s%s, Other:%s\n",
+    printk("  Xen settings: BTI-Thunk %s, SPEC_CTRL: %s%s, Other:%s%s\n",
            thunk == THUNK_NONE      ? "N/A" :
            thunk == THUNK_RETPOLINE ? "RETPOLINE" :
            thunk == THUNK_LFENCE    ? "LFENCE" :
@@ -299,7 +305,8 @@ static void __init print_details(enum ind_thunk thunk, uint64_t caps)
            (default_xen_spec_ctrl & SPEC_CTRL_IBRS)  ? "IBRS+" :  "IBRS-",
            !boot_cpu_has(X86_FEATURE_SSBD)           ? "" :
            (default_xen_spec_ctrl & SPEC_CTRL_SSBD)  ? " SSBD+" : " SSBD-",
-           opt_ibpb                                  ? " IBPB"  : "");
+           opt_ibpb                                  ? " IBPB"  : "",
+           opt_l1d_flush                             ? " L1D_FLUSH" : "");
 
     /* L1TF diagnostics, printed if vulnerable or PV shadowing is in use. */
     if ( cpu_has_bug_l1tf || opt_pv_l1tf )
@@ -871,6 +878,33 @@ void __init init_speculation_mitigations(void)
             opt_pv_l1tf = OPT_PV_L1TF_DOMU;
     }
 
+    /*
+     * By default, enable L1D_FLUSH on L1TF-vulnerable hardware, unless
+     * instructed to skip the flush on vmentry by our outer hypervisor.
+     */
+    if ( !boot_cpu_has(X86_FEATURE_L1D_FLUSH) )
+        opt_l1d_flush = 0;
+    else if ( opt_l1d_flush == -1 )
+        opt_l1d_flush = cpu_has_bug_l1tf && !(caps & ARCH_CAPS_SKIP_L1DFL);
+
+    /*
+     * We do not disable HT by default on affected hardware.
+     *
+     * Firstly, if the user intends to use exclusively PV, or HVM shadow
+     * guests, HT isn't a concern and should remain fully enabled.  Secondly,
+     * safety for HVM HAP guests can be arranged by the toolstack with core
+     * parking, pinning or cpupool configurations, including mixed setups.
+     *
+     * However, if we are on affected hardware, with HT enabled, and the user
+     * hasn't explicitly chosen whether to use HT or not, nag them to do so.
+     */
+    if ( opt_smt == -1 && cpu_has_bug_l1tf && !pv_shim &&
+         boot_cpu_data.x86_num_siblings > 1 )
+        warning_add(
+            "Booted on L1TF-vulnerable hardware with SMT/Hyperthreading\n"
+            "enabled.  Please assess your configuration and choose an\n"
+            "explicit 'smt=<bool>' setting.  See XSA-273.\n");
+
     print_details(thunk, caps);
 
     /*

diff --git a/xen/include/asm-x86/spec_ctrl.h b/xen/include/asm-x86/spec_ctrl.h
index cdf5737dc204470ff5f57f818d9f021b3d4d32d7..8f8aad40bbd8994d5e402bd9335c4e03ccb6d12e 100644 (file)
--- a/xen/include/asm-x86/spec_ctrl.h
+++ b/xen/include/asm-x86/spec_ctrl.h
@@ -29,6 +29,7 @@ void init_speculation_mitigations(void);
 extern bool opt_ibpb;
 extern bool opt_ssbd;
 extern int8_t opt_eager_fpu;
+extern int8_t opt_l1d_flush;
 
 extern bool bsp_delay_spec_ctrl;
 extern uint8_t default_xen_spec_ctrl;