Implement rdtscp emulation and rdtscp_aux "support"

The rdtscp instruction (and the associated TSC_AUX
msr) are present on most recent AMD processors,
and on the Nehalem and future Intel processors.
Cpuid has a bit to detect the presence of this feature.

Xen intentionally does not expose the cpuid rdtscp bit
to PV OS's or to HVM guests, but PV apps can see this
bit and, as a result, may choose to use the rdtscp
instruction.  When a PV guest with such an app is migrated
to a machine that does not have rdtscp support, the
app will get killed due to an invalid op.  Fix this
by emulating the rdtscp instruction.  We also need
to emulate rdtscp in the case where the machine has
rdtscp support, but rdtsc emulation is enabled (which
is unfortunately a different path: a privileged op).

The rdtscp instruction reads the TSC_AUX register which
presumably is set by the OS (and, in the case of
tsc_mode==pvrdtscp, will be set by Xen).  HV Linux
and PV Linux will not set TSC_AUX because the
cpuid rdtscp bit is not propogated by Xen; I'm told that
Windows always sets TSC_AUX to zero.  So for PV guests
running on rdtscp-capable hardware (that don't use
tsc_mode==pvrdtscp), always set TSC_AUX to zero.

Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com>

diff --git a/xen/arch/x86/time.c b/xen/arch/x86/time.c
index 34d3f8350a71e858a98805da043784ae62eadcbd..ffe1ab8187da1636ab329267ec9fe35d4127a60c 100644 (file)
--- a/xen/arch/x86/time.c
+++ b/xen/arch/x86/time.c
@@ -851,9 +851,13 @@ static void __update_vcpu_system_time(struct vcpu *v, int force)
     else
         tsc_stamp = t->local_tsc_stamp;
 
-    if ( d->arch.tsc_mode ==  TSC_MODE_PVRDTSCP &&
-              boot_cpu_has(X86_FEATURE_RDTSCP) )
-        write_rdtscp_aux(d->arch.incarnation);
+    if ( boot_cpu_has(X86_FEATURE_RDTSCP) )
+    {
+        if ( d->arch.tsc_mode ==  TSC_MODE_PVRDTSCP )
+            write_rdtscp_aux(d->arch.incarnation);
+        else
+            write_rdtscp_aux(0);
+    }
 
     /* Don't bother unless timestamps have changed or we are forced. */
     if ( !force && (u->tsc_timestamp == tsc_stamp) )
@@ -1608,7 +1612,7 @@ void tsc_check_reliability(void)
  * PV SoftTSC Emulation.
  */
 
-void pv_soft_rdtsc(struct vcpu *v, struct cpu_user_regs *regs)
+void pv_soft_rdtsc(struct vcpu *v, struct cpu_user_regs *regs, int rdtscp)
 {
     s_time_t now = get_s_time();
     struct domain *d = v->domain;
@@ -1633,6 +1637,10 @@ void pv_soft_rdtsc(struct vcpu *v, struct cpu_user_regs *regs)
 
     regs->eax = (uint32_t)now;
     regs->edx = (uint32_t)(now >> 32);
+
+    if ( rdtscp )
+         regs->ecx =
+             (d->arch.tsc_mode == TSC_MODE_PVRDTSCP) ?  d->arch.incarnation : 0;
 }
 
 static int host_tsc_is_safe(void)
@@ -1826,7 +1834,7 @@ static void dump_softtsc(unsigned char key)
             printk(",khz=%"PRIu32"",d->arch.tsc_khz);
         if ( d->arch.incarnation )
             printk(",inc=%"PRIu32"",d->arch.incarnation);
-        if ( !d->arch.vtsc )
+        if ( !(d->arch.vtsc_kerncount | d->arch.vtsc_usercount) )
         {
             printk("\n");
             continue;

diff --git a/xen/arch/x86/traps.c b/xen/arch/x86/traps.c
index 174dc25af41205125e1aa6884442e1d4020b2413..9ccb769748ea82eee8dc9d64806ea2045c823e25 100644 (file)
--- a/xen/arch/x86/traps.c
+++ b/xen/arch/x86/traps.c
@@ -831,6 +831,26 @@ static void pv_cpuid(struct cpu_user_regs *regs)
     regs->edx = d;
 }
 
+static int emulate_invalid_rdtscp(struct cpu_user_regs *regs)
+{
+    char opcode[3];
+    unsigned long eip, rc;
+    struct vcpu *v = current;
+
+    eip = regs->eip;
+    if ( (rc = copy_from_user(opcode, (char *)eip, sizeof(opcode))) != 0 )
+    {
+        propagate_page_fault(eip + sizeof(opcode) - rc, 0);
+        return EXCRET_fault_fixed;
+    }
+    if ( memcmp(opcode, "\xf\x1\xf9", sizeof(opcode)) )
+        return 0;
+    eip += sizeof(opcode);
+    pv_soft_rdtsc(v, regs, 1);
+    instruction_done(regs, eip, 0);
+    return EXCRET_fault_fixed;
+}
+
 static int emulate_forced_invalid_op(struct cpu_user_regs *regs)
 {
     char sig[5], instr[2];
@@ -879,7 +899,8 @@ asmlinkage void do_invalid_op(struct cpu_user_regs *regs)
 
     if ( likely(guest_mode(regs)) )
     {
-        if ( !emulate_forced_invalid_op(regs) )
+        if ( !emulate_invalid_rdtscp(regs) &&
+             !emulate_forced_invalid_op(regs) )
             do_guest_trap(TRAP_invalid_op, regs, 0);
         return;
     }
@@ -2009,11 +2030,12 @@ static int emulate_privileged_op(struct cpu_user_regs *regs)
 
  twobyte_opcode:
     /*
-     * All two-byte opcodes, except RDTSC (0x31) are executable only from
-     * guest kernel mode (virtual ring 0).
+     * All 2 and 3 byte opcodes, except RDTSC (0x31) and RDTSCP (0x1,0xF9)
+     * are executable only from guest kernel mode (virtual ring 0).
      */
     opcode = insn_fetch(u8, code_base, eip, code_limit);
     if ( !guest_kernel_mode(v, regs) &&
+         (opcode != 0x1) && /* always emulate rdtscp */
          !((opcode == 0x31) && v->domain->arch.vtsc) )
         goto fail;
 
@@ -2021,6 +2043,12 @@ static int emulate_privileged_op(struct cpu_user_regs *regs)
         goto fail;
     switch ( opcode )
     {
+    case 0x1: /* RDTSCP */
+        if ( insn_fetch(u8, code_base, eip, code_limit) != 0xf9 )
+            goto fail;
+        pv_soft_rdtsc(v, regs, 1);
+        break;
+
     case 0x06: /* CLTS */
         (void)do_fpu_taskswitch(0);
         break;
@@ -2269,7 +2297,7 @@ static int emulate_privileged_op(struct cpu_user_regs *regs)
     }
 
     case 0x31: /* RDTSC */
-        pv_soft_rdtsc(v, regs);
+        pv_soft_rdtsc(v, regs, 0);
         break;
 
     case 0x32: /* RDMSR */

diff --git a/xen/include/asm-x86/time.h b/xen/include/asm-x86/time.h
index 6dd071d7266327f8d871efc7c5f0ad0156240eaa..6a9b9df671b2a61b6bbf66a047e39c1185419b92 100644 (file)
--- a/xen/include/asm-x86/time.h
+++ b/xen/include/asm-x86/time.h
@@ -59,7 +59,7 @@ int pit_broadcast_is_available(void);
 uint64_t acpi_pm_tick_to_ns(uint64_t ticks);
 uint64_t ns_to_acpi_pm_tick(uint64_t ns);
 
-void pv_soft_rdtsc(struct vcpu *v, struct cpu_user_regs *regs);
+void pv_soft_rdtsc(struct vcpu *v, struct cpu_user_regs *regs, int rdtscp);
 
 void tsc_set_info(struct domain *d, uint32_t tsc_mode, uint64_t elapsed_nsec,
                   uint32_t gtsc_khz, uint32_t incarnation);