x86/tsx: Introduce tsx= to use MSR_TSX_CTRL when available

To protect against the TSX Async Abort speculative vulnerability, Intel have
released new microcode for affected parts which introduce the MSR_TSX_CTRL
control, which allows TSX to be turned off.  This will be architectural on
future parts.

Introduce tsx= to provide a global on/off for TSX, including its enumeration
via CPUID.  Provide stub virtualisation of this MSR, as it is not exposed to
guests at the moment.

VMs may have booted before microcode is loaded, or before hosts have rebooted,
and they still want to migrate freely.  A VM which booted seeing TSX can
migrate safely to hosts with TSX disabled - TSX will start unconditionally
aborting, but still behave in a manner compatible with the ABI.

The guest-visible behaviour is equivalent to late loading the microcode and
setting the RTM_DISABLE bit in the course of live patching.

This is part of XSA-305 / CVE-2019-11135

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>

diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown
index a3194cadc3bcc5ba75eb2c47a560b350a68c3928..0f1b6a176e637bba839a1107f06e068e28b4c0e1 100644 (file)
--- a/docs/misc/xen-command-line.markdown
+++ b/docs/misc/xen-command-line.markdown
@@ -1815,6 +1815,20 @@ pages) must also be specified via the tbuf\_size parameter.
 ### tsc
 > `= unstable | skewed | stable:socket`
 
+### tsx
+    = <bool>
+
+    Applicability: x86
+    Default: true
+
+Controls for the use of Transactional Synchronization eXtensions.
+
+On Intel parts released in Q3 2019 (with updated microcode), and future parts,
+a control has been introduced which allows TSX to be turned off.
+
+On systems with the ability to turn TSX off, this boolean offers system wide
+control of whether TSX is enabled or disabled.
+
 ### ucode
 > `= [<integer> | scan]`
 

diff --git a/xen/arch/x86/Makefile b/xen/arch/x86/Makefile
index 59cadb759d9da44d2d2471494c5f43feae2e12b7..041068173ca6502ea2574874e0a1eecfbcc43aa1 100644 (file)
--- a/xen/arch/x86/Makefile
+++ b/xen/arch/x86/Makefile
@@ -63,6 +63,7 @@ obj-y += sysctl.o
 obj-y += time.o
 obj-y += trace.o
 obj-y += traps.o
+obj-y += tsx.o
 obj-y += usercopy.o
 obj-y += x86_emulate.o
 obj-$(CONFIG_TBOOT) += tboot.o

diff --git a/xen/arch/x86/cpuid.c b/xen/arch/x86/cpuid.c
index 2201f8ac75ab2170905f6d498bd4c88e4d627329..9aaf8b8283b6dcb00f02368df04f1cf06c86d874 100644 (file)
--- a/xen/arch/x86/cpuid.c
+++ b/xen/arch/x86/cpuid.c
@@ -605,6 +605,20 @@ void recalculate_cpuid_policy(struct domain *d)
     if ( cpu_has_itsc && (d->disable_migrate || d->arch.vtsc) )
         __set_bit(X86_FEATURE_ITSC, max_fs);
 
+    /*
+     * On hardware with MSR_TSX_CTRL, the admin may have elected to disable
+     * TSX and hide the feature bits.  Migrating-in VMs may have been booted
+     * pre-mitigation when the TSX features were visbile.
+     *
+     * This situation is compatible (albeit with a perf hit to any TSX code in
+     * the guest), so allow the feature bits to remain set.
+     */
+    if ( cpu_has_tsx_ctrl )
+    {
+        __set_bit(X86_FEATURE_HLE, max_fs);
+        __set_bit(X86_FEATURE_RTM, max_fs);
+    }
+
     /* Clamp the toolstacks choices to reality. */
     for ( i = 0; i < ARRAY_SIZE(fs); i++ )
         fs[i] &= max_fs[i];

diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
index 0b05b0388c51e5baf5dd5fc7a5fd0b1f8eea4880..2aa9ac06a4aafd07c9a0a95c6567c24b790127e7 100644 (file)
--- a/xen/arch/x86/hvm/hvm.c
+++ b/xen/arch/x86/hvm/hvm.c
@@ -3444,6 +3444,7 @@ int hvm_msr_read_intercept(unsigned int msr, uint64_t *msr_content)
     case MSR_FLUSH_CMD:
         /* Write-only */
     case MSR_TSX_FORCE_ABORT:
+    case MSR_TSX_CTRL:
         /* Not offered to guests. */
         goto gp_fault;
 
@@ -3669,6 +3670,7 @@ int hvm_msr_write_intercept(unsigned int msr, uint64_t msr_content,
     case MSR_ARCH_CAPABILITIES:
         /* Read-only */
     case MSR_TSX_FORCE_ABORT:
+    case MSR_TSX_CTRL:
         /* Not offered to guests. */
         goto gp_fault;
 

diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index 40af7e65d4f7eeb6567a4de5fcf84b20afc461f9..48963b186b57f1b5269199c169d70917b063af13 100644 (file)
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -1490,6 +1490,8 @@ void __init noreturn __start_xen(unsigned long mbi_p)
 
     early_microcode_init();
 
+    tsx_init(); /* Needs microcode.  May change HLE/RTM feature bits. */
+
     identify_cpu(&boot_cpu_data);
 
     set_in_cr4(X86_CR4_OSFXSR | X86_CR4_OSXMMEXCPT);

diff --git a/xen/arch/x86/smpboot.c b/xen/arch/x86/smpboot.c
index 1fda6c507a700108139c6d1bc05a987ffea0950c..641f830cd1f799242f069f5fd0c5fb6691089d60 100644 (file)
--- a/xen/arch/x86/smpboot.c
+++ b/xen/arch/x86/smpboot.c
@@ -363,6 +363,8 @@ void start_secondary(void *unused)
     if ( boot_cpu_has(X86_FEATURE_IBRSB) )
         wrmsrl(MSR_SPEC_CTRL, default_xen_spec_ctrl);
 
+    tsx_init(); /* Needs microcode.  May change HLE/RTM feature bits. */
+
     smp_callin();
 
     init_percpu_time();

diff --git a/xen/arch/x86/traps.c b/xen/arch/x86/traps.c
index 53301574e2e55980d9c3987446ef9766a9e8eeda..9b4bb6a009b3fc3b0c6fedbe3766e480cd2570ea 100644 (file)
--- a/xen/arch/x86/traps.c
+++ b/xen/arch/x86/traps.c
@@ -2654,6 +2654,7 @@ static int priv_op_read_msr(unsigned int reg, uint64_t *val,
     case MSR_FLUSH_CMD:
         /* Write-only */
     case MSR_TSX_FORCE_ABORT:
+    case MSR_TSX_CTRL:
         /* Not offered to guests. */
         break;
 
@@ -2878,6 +2879,7 @@ static int priv_op_write_msr(unsigned int reg, uint64_t val,
     case MSR_ARCH_CAPABILITIES:
         /* The MSR is read-only. */
     case MSR_TSX_FORCE_ABORT:
+    case MSR_TSX_CTRL:
         /* Not offered to guests. */
         break;
 

diff --git a/xen/arch/x86/tsx.c b/xen/arch/x86/tsx.c
new file mode 100644 (file)
index 0000000..3a853d3
--- /dev/null
+++ b/xen/arch/x86/tsx.c
@@ -0,0 +1,74 @@
+#include <xen/init.h>
+#include <asm/msr.h>
+
+/*
+ * Valid values:
+ *   1 => Explicit tsx=1
+ *   0 => Explicit tsx=0
+ *  -1 => Default, implicit tsx=1
+ *
+ * This is arranged such that the bottom bit encodes whether TSX is actually
+ * disabled, while identifying various explicit (>=0) and implicit (<0)
+ * conditions.
+ */
+int8_t __read_mostly opt_tsx = -1;
+int8_t __read_mostly cpu_has_tsx_ctrl = -1;
+
+static int __init parse_tsx(const char *s)
+{
+    int rc = 0, val = parse_bool(s);
+
+    if ( val >= 0 )
+        opt_tsx = val;
+    else
+        rc = -EINVAL;
+
+    return rc;
+}
+custom_param("tsx", parse_tsx);
+
+void tsx_init(void)
+{
+    /*
+     * This function is first called between microcode being loaded, and CPUID
+     * being scanned generally.  Calculate from raw data whether MSR_TSX_CTRL
+     * is available.
+     */
+    if ( unlikely(cpu_has_tsx_ctrl < 0) )
+    {
+        uint64_t caps = 0;
+
+        if ( boot_cpu_data.cpuid_level >= 7 &&
+             (cpuid_count_edx(7, 0) & cpufeat_mask(X86_FEATURE_ARCH_CAPS)) )
+            rdmsrl(MSR_ARCH_CAPABILITIES, caps);
+
+        cpu_has_tsx_ctrl = !!(caps & ARCH_CAPS_TSX_CTRL);
+    }
+
+    if ( cpu_has_tsx_ctrl )
+    {
+        uint64_t val;
+
+        rdmsrl(MSR_TSX_CTRL, val);
+
+        val &= ~(TSX_CTRL_RTM_DISABLE | TSX_CTRL_CPUID_CLEAR);
+        /* Check bottom bit only.  Higher bits are various sentinals. */
+        if ( !(opt_tsx & 1) )
+            val |= TSX_CTRL_RTM_DISABLE | TSX_CTRL_CPUID_CLEAR;
+
+        wrmsrl(MSR_TSX_CTRL, val);
+    }
+    else if ( opt_tsx >= 0 )
+        printk_once(XENLOG_WARNING
+                    "MSR_TSX_CTRL not available - Ignoring tsx= setting\n");
+}
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */

diff --git a/xen/include/asm-x86/msr-index.h b/xen/include/asm-x86/msr-index.h
index 5ef894ff2971aa4ea4815ff765da77eafab73668..b7c1673488876c7c306abd9a5589385d5bf5ac21 100644 (file)
--- a/xen/include/asm-x86/msr-index.h
+++ b/xen/include/asm-x86/msr-index.h
@@ -55,6 +55,7 @@
 #define ARCH_CAPS_SSB_NO               (_AC(1, ULL) << 4)
 #define ARCH_CAPS_MDS_NO               (_AC(1, ULL) << 5)
 #define ARCH_CAPS_IF_PSCHANGE_MC_NO    (_AC(1, ULL) << 6)
+#define ARCH_CAPS_TSX_CTRL             (_AC(1, ULL) << 7)
 
 #define MSR_FLUSH_CMD                  0x0000010b
 #define FLUSH_CMD_L1D                  (_AC(1, ULL) << 0)
@@ -62,6 +63,10 @@
 #define MSR_TSX_FORCE_ABORT             0x0000010f
 #define TSX_FORCE_ABORT_RTM             (_AC(1, ULL) <<  0)
 
+#define MSR_TSX_CTRL                    0x00000122
+#define TSX_CTRL_RTM_DISABLE            (_AC(1, ULL) <<  0)
+#define TSX_CTRL_CPUID_CLEAR            (_AC(1, ULL) <<  1)
+
 /* Intel MSRs. Some also available on other CPUs */
 #define MSR_IA32_PERFCTR0              0x000000c1
 #define MSR_IA32_A_PERFCTR0            0x000004c1

diff --git a/xen/include/asm-x86/processor.h b/xen/include/asm-x86/processor.h
index 44873477131159a7b274eb79da8082023dbdca2a..d068a8710f3ea2eb88963436d4838b9c4ecdb5a2 100644 (file)
--- a/xen/include/asm-x86/processor.h
+++ b/xen/include/asm-x86/processor.h
@@ -265,6 +265,16 @@ static always_inline unsigned int cpuid_count_ebx(
     return ebx;
 }
 
+static always_inline unsigned int cpuid_count_edx(
+    unsigned int leaf, unsigned int subleaf)
+{
+    unsigned int edx, tmp;
+
+    cpuid_count(leaf, subleaf, &tmp, &tmp, &tmp, &edx);
+
+    return edx;
+}
+
 static inline unsigned long read_cr0(void)
 {
     unsigned long cr0;
@@ -632,6 +642,9 @@ static inline uint8_t get_cpu_family(uint32_t raw, uint8_t *model,
     return fam;
 }
 
+extern int8_t opt_tsx, cpu_has_tsx_ctrl;
+void tsx_init(void);
+
 #endif /* !__ASSEMBLY__ */
 
 #endif /* __ASM_X86_PROCESSOR_H */

diff --git a/xen/include/xen/lib.h b/xen/include/xen/lib.h
index 2829d478adeb30580eff39fa298223d9b217c71a..6d97373765d3df1a7c4aa4ac3bf490bdd28f32ad 100644 (file)
--- a/xen/include/xen/lib.h
+++ b/xen/include/xen/lib.h
@@ -108,6 +108,16 @@ extern int printk_ratelimit(void);
 #define gprintk(lvl, fmt, args...) \
     printk(XENLOG_GUEST lvl "%pv " fmt, current, ## args)
 
+#define printk_once(fmt, args...)               \
+({                                              \
+    static bool __read_mostly once_;            \
+    if ( unlikely(!once_) )                     \
+    {                                           \
+        once_ = true;                           \
+        printk(fmt, ## args);                   \
+    }                                           \
+})
+
 #ifdef NDEBUG
 
 static inline void