summaryrefslogtreecommitdiffstats
path: root/arch/x86/include/asm/msr.h
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86/include/asm/msr.h')
-rw-r--r--arch/x86/include/asm/msr.h26
1 files changed, 26 insertions, 0 deletions
diff --git a/arch/x86/include/asm/msr.h b/arch/x86/include/asm/msr.h
index ff0c120dafe5..02bdd6c65017 100644
--- a/arch/x86/include/asm/msr.h
+++ b/arch/x86/include/asm/msr.h
@@ -127,6 +127,32 @@ static __always_inline unsigned long long rdtsc(void)
return EAX_EDX_VAL(val, low, high);
}
+/**
+ * rdtsc_ordered() - read the current TSC in program order
+ *
+ * rdtsc_ordered() returns the result of RDTSC as a 64-bit integer.
+ * It is ordered like a load to a global in-memory counter. It should
+ * be impossible to observe non-monotonic rdtsc_unordered() behavior
+ * across multiple CPUs as long as the TSC is synced.
+ */
+static __always_inline unsigned long long rdtsc_ordered(void)
+{
+ /*
+ * The RDTSC instruction is not ordered relative to memory
+ * access. The Intel SDM and the AMD APM are both vague on this
+ * point, but empirically an RDTSC instruction can be
+ * speculatively executed before prior loads. An RDTSC
+ * immediately after an appropriate barrier appears to be
+ * ordered as a normal load, that is, it provides the same
+ * ordering guarantees as reading from a global memory location
+ * that some other imaginary CPU is updating continuously with a
+ * time stamp.
+ */
+ alternative_2("", "mfence", X86_FEATURE_MFENCE_RDTSC,
+ "lfence", X86_FEATURE_LFENCE_RDTSC);
+ return rdtsc();
+}
+
static inline unsigned long long native_read_pmc(int counter)
{
DECLARE_ARGS(val, low, high);
OpenPOWER on IntegriCloud