[PATCH 6/9] perf/x86: enforce HT bug workaround for SNB/IVB/HSW

[Stephane Eranian <eranian@google.com>]

From: Maria Dimakopoulou <maria.n.dimakopoulou@gmail.com>

This patches activates the HT bug workaround for the
SNB/IVB/HSW processors. This covers non-PEBS mode.
Activation is done thru the constraint tables.

Both client and server processors needs this workaround.

Reviewed-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Maria Dimakopoulou <maria.n.dimakopoulou@gmail.com>
---
 arch/x86/kernel/cpu/perf_event_intel.c |   53 ++++++++++++++++++++++++++------
 1 file changed, 44 insertions(+), 9 deletions(-)

diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 3af9745..b1a2684 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -113,6 +113,12 @@ static struct event_constraint intel_snb_event_constraints[] __read_mostly =
 	INTEL_EVENT_CONSTRAINT(0xcd, 0x8), /* MEM_TRANS_RETIRED.LOAD_LATENCY */
 	INTEL_UEVENT_CONSTRAINT(0x04a3, 0xf), /* CYCLE_ACTIVITY.CYCLES_NO_DISPATCH */
 	INTEL_UEVENT_CONSTRAINT(0x02a3, 0x4), /* CYCLE_ACTIVITY.CYCLES_L1D_PENDING */
+
+	INTEL_EXCLEVT_CONSTRAINT(0xd0, 0xf), /* MEM_UOPS_RETIRED.* */
+	INTEL_EXCLEVT_CONSTRAINT(0xd1, 0xf), /* MEM_LOAD_UOPS_RETIRED.* */
+	INTEL_EXCLEVT_CONSTRAINT(0xd2, 0xf), /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */
+	INTEL_EXCLEVT_CONSTRAINT(0xd3, 0xf), /* MEM_LOAD_UOPS_LLC_MISS_RETIRED.* */
+
 	EVENT_CONSTRAINT_END
 };
 
@@ -131,15 +137,12 @@ static struct event_constraint intel_ivb_event_constraints[] __read_mostly =
 	INTEL_UEVENT_CONSTRAINT(0x08a3, 0x4), /* CYCLE_ACTIVITY.CYCLES_L1D_PENDING */
 	INTEL_UEVENT_CONSTRAINT(0x0ca3, 0x4), /* CYCLE_ACTIVITY.STALLS_L1D_PENDING */
 	INTEL_UEVENT_CONSTRAINT(0x01c0, 0x2), /* INST_RETIRED.PREC_DIST */
-	/*
-	 * Errata BV98 -- MEM_*_RETIRED events can leak between counters of SMT
-	 * siblings; disable these events because they can corrupt unrelated
-	 * counters.
-	 */
-	INTEL_EVENT_CONSTRAINT(0xd0, 0x0), /* MEM_UOPS_RETIRED.* */
-	INTEL_EVENT_CONSTRAINT(0xd1, 0x0), /* MEM_LOAD_UOPS_RETIRED.* */
-	INTEL_EVENT_CONSTRAINT(0xd2, 0x0), /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */
-	INTEL_EVENT_CONSTRAINT(0xd3, 0x0), /* MEM_LOAD_UOPS_LLC_MISS_RETIRED.* */
+
+	INTEL_EXCLEVT_CONSTRAINT(0xd0, 0xf), /* MEM_UOPS_RETIRED.* */
+	INTEL_EXCLEVT_CONSTRAINT(0xd1, 0xf), /* MEM_LOAD_UOPS_RETIRED.* */
+	INTEL_EXCLEVT_CONSTRAINT(0xd2, 0xf), /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */
+	INTEL_EXCLEVT_CONSTRAINT(0xd3, 0xf), /* MEM_LOAD_UOPS_LLC_MISS_RETIRED.* */
+
 	EVENT_CONSTRAINT_END
 };
 
@@ -217,6 +220,12 @@ static struct event_constraint intel_hsw_event_constraints[] = {
 	INTEL_EVENT_CONSTRAINT(0x0ca3, 0x4),
 	/* CYCLE_ACTIVITY.CYCLES_NO_EXECUTE */
 	INTEL_EVENT_CONSTRAINT(0x04a3, 0xf),
+
+	INTEL_EXCLEVT_CONSTRAINT(0xd0, 0xf), /* MEM_UOPS_RETIRED.* */
+	INTEL_EXCLEVT_CONSTRAINT(0xd1, 0xf), /* MEM_LOAD_UOPS_RETIRED.* */
+	INTEL_EXCLEVT_CONSTRAINT(0xd2, 0xf), /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */
+	INTEL_EXCLEVT_CONSTRAINT(0xd3, 0xf), /* MEM_LOAD_UOPS_LLC_MISS_RETIRED.* */
+
 	EVENT_CONSTRAINT_END
 };
 
@@ -2584,6 +2593,27 @@ static __init void intel_nehalem_quirk(void)
 	}
 }
 
+/*
+ * enable software workaround for errata:
+ * SNB: BJ122
+ * IVB: BV98
+ * HSW: HSD29
+ *
+ * Only needed when HT is enabled. However detecting
+ * this is too difficult and model specific so we enable
+ * it even with HT off for now.
+ */
+static __init void intel_ht_bug(void)
+{
+	x86_pmu.flags |= PMU_FL_EXCL_CNTRS;
+
+	x86_pmu.commit_scheduling = intel_commit_scheduling;
+	x86_pmu.start_scheduling = intel_start_scheduling;
+	x86_pmu.stop_scheduling = intel_stop_scheduling;
+
+	pr_info("CPU erratum BJ122, BV98, HSD29 worked around\n");
+}
+
 EVENT_ATTR_STR(mem-loads,	mem_ld_hsw,	"event=0xcd,umask=0x1,ldlat=3");
 EVENT_ATTR_STR(mem-stores,	mem_st_hsw,	"event=0xd0,umask=0x82")
 
@@ -2796,6 +2826,7 @@ __init int intel_pmu_init(void)
 	case 42: /* SandyBridge */
 	case 45: /* SandyBridge, "Romely-EP" */
 		x86_add_quirk(intel_sandybridge_quirk);
+		x86_add_quirk(intel_ht_bug);
 		memcpy(hw_cache_event_ids, snb_hw_cache_event_ids,
 		       sizeof(hw_cache_event_ids));
 		memcpy(hw_cache_extra_regs, snb_hw_cache_extra_regs,
@@ -2810,6 +2841,8 @@ __init int intel_pmu_init(void)
 			x86_pmu.extra_regs = intel_snbep_extra_regs;
 		else
 			x86_pmu.extra_regs = intel_snb_extra_regs;
+
+
 		/* all extra regs are per-cpu when HT is on */
 		x86_pmu.flags |= PMU_FL_HAS_RSP_1;
 		x86_pmu.flags |= PMU_FL_NO_HT_SHARING;
@@ -2827,6 +2860,7 @@ __init int intel_pmu_init(void)
 		break;
 	case 58: /* IvyBridge */
 	case 62: /* IvyBridge EP */
+		x86_add_quirk(intel_ht_bug);
 		memcpy(hw_cache_event_ids, snb_hw_cache_event_ids,
 		       sizeof(hw_cache_event_ids));
 		memcpy(hw_cache_extra_regs, snb_hw_cache_extra_regs,
@@ -2860,6 +2894,7 @@ __init int intel_pmu_init(void)
 	case 71:
 	case 63:
 	case 69:
+		x86_add_quirk(intel_ht_bug);
 		x86_pmu.late_ack = true;
 		memcpy(hw_cache_event_ids, snb_hw_cache_event_ids, sizeof(hw_cache_event_ids));
 		memcpy(hw_cache_extra_regs, snb_hw_cache_extra_regs, sizeof(hw_cache_extra_regs));
-- 
1.7.9.5