[PATCH 2/2] perf/x86: Add Intel Westmere-EX uncore support

["Yan, Zheng" <zheng.z.yan@intel.com>]

From: "Yan, Zheng" <zheng.z.yan@intel.com>

The Westmere-EX uncore is similar to the Nehalem-EX uncore. The
differences are:
 - Westmere-EX uncore has 10 instances of Cbox. The MSRs for Cbox8
   and Cbox9 in the Westmere-EX aren't contiguous with Cbox 0~7.
 - The fvid field in the ZDP_CTL_FVC register in the Mbox is
   different. It's 5 bits in the Nehalem-EX, 6 bits in the
   Westmere-EX.

Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
---
 arch/x86/kernel/cpu/perf_event_intel_uncore.c | 56 ++++++++++++++++++++++-----
 arch/x86/kernel/cpu/perf_event_intel_uncore.h | 45 ++++++++++-----------
 2 files changed, 68 insertions(+), 33 deletions(-)

diff --git a/arch/x86/kernel/cpu/perf_event_intel_uncore.c b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
index 5750142..740406f 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_uncore.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
@@ -901,16 +901,21 @@ static struct attribute_group nhmex_uncore_cbox_format_group = {
 	.attrs = nhmex_uncore_cbox_formats_attr,
 };
 
+/* msr offset for each instance of cbox */
+static unsigned nhmex_cbox_msr_offsets[] = {
+	0x0, 0x80, 0x40, 0xc0, 0x20, 0xa0, 0x60, 0xe0, 0x240, 0x2c0,
+};
+
 static struct intel_uncore_type nhmex_uncore_cbox = {
 	.name			= "cbox",
 	.num_counters		= 6,
-	.num_boxes		= 8,
+	.num_boxes		= 10,
 	.perf_ctr_bits		= 48,
 	.event_ctl		= NHMEX_C0_MSR_PMON_EV_SEL0,
 	.perf_ctr		= NHMEX_C0_MSR_PMON_CTR0,
 	.event_mask		= NHMEX_PMON_RAW_EVENT_MASK,
 	.box_ctl		= NHMEX_C0_MSR_PMON_GLOBAL_CTL,
-	.msr_offset		= NHMEX_C_MSR_OFFSET,
+	.msr_offsets		= nhmex_cbox_msr_offsets,
 	.pair_ctr_ctl		= 1,
 	.ops			= &nhmex_uncore_ops,
 	.format_group		= &nhmex_uncore_cbox_format_group
@@ -1138,6 +1143,9 @@ static struct extra_reg nhmex_uncore_mbox_extra_regs[] = {
 	EVENT_EXTRA_END
 };
 
+/* Nehalem-EX or Westmere-EX ? */
+bool uncore_nhmex;
+
 static bool nhmex_mbox_get_shared_reg(struct intel_uncore_box *box, int idx, u64 config)
 {
 	struct intel_uncore_extra_reg *er;
@@ -1167,18 +1175,29 @@ static bool nhmex_mbox_get_shared_reg(struct intel_uncore_box *box, int idx, u64
 		return false;
 
 	/* mask of the shared fields */
-	mask = NHMEX_M_PMON_ZDP_CTL_FVC_MASK;
+	if (uncore_nhmex)
+		mask = NHMEX_M_PMON_ZDP_CTL_FVC_MASK;
+	else
+		mask = WSMEX_M_PMON_ZDP_CTL_FVC_MASK;
 	er = &box->shared_regs[EXTRA_REG_NHMEX_M_ZDP_CTL_FVC];
 
 	raw_spin_lock_irqsave(&er->lock, flags);
 	/* add mask of the non-shared field if it's in use */
-	if (__BITS_VALUE(atomic_read(&er->ref), idx, 8))
-		mask |= NHMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(idx);
+	if (__BITS_VALUE(atomic_read(&er->ref), idx, 8)) {
+		if (uncore_nhmex)
+			mask |= NHMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(idx);
+		else
+			mask |= WSMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(idx);
+	}
 
 	if (!atomic_read(&er->ref) || !((er->config ^ config) & mask)) {
 		atomic_add(1 << (idx * 8), &er->ref);
-		mask = NHMEX_M_PMON_ZDP_CTL_FVC_MASK |
-			NHMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(idx);
+		if (uncore_nhmex)
+			mask = NHMEX_M_PMON_ZDP_CTL_FVC_MASK |
+				NHMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(idx);
+		else
+			mask = WSMEX_M_PMON_ZDP_CTL_FVC_MASK |
+				WSMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(idx);
 		er->config &= ~mask;
 		er->config |= (config & mask);
 		ret = true;
@@ -1212,7 +1231,10 @@ u64 nhmex_mbox_alter_er(struct perf_event *event, int new_idx, bool modify)
 
 	/* get the non-shared control bits and shift them */
 	idx = orig_idx - EXTRA_REG_NHMEX_M_ZDP_CTL_FVC;
-	config &= NHMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(idx);
+	if (uncore_nhmex)
+		config &= NHMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(idx);
+	else
+		config &= WSMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(idx);
 	if (new_idx > orig_idx) {
 		idx = new_idx - orig_idx;
 		config <<= 3 * idx;
@@ -1222,6 +1244,10 @@ u64 nhmex_mbox_alter_er(struct perf_event *event, int new_idx, bool modify)
 	}
 
 	/* add the shared control bits back */
+	if (uncore_nhmex)
+		config |= NHMEX_M_PMON_ZDP_CTL_FVC_MASK & reg1->config;
+	else
+		config |= WSMEX_M_PMON_ZDP_CTL_FVC_MASK & reg1->config;
 	config |= NHMEX_M_PMON_ZDP_CTL_FVC_MASK & reg1->config;
 	if (modify) {
 		/* adjust the main event selector */
@@ -1480,6 +1506,12 @@ static struct uncore_event_desc nhmex_uncore_mbox_events[] = {
 	{ /* end: all zeroes */ },
 };
 
+static struct uncore_event_desc wsmex_uncore_mbox_events[] = {
+	INTEL_UNCORE_EVENT_DESC(bbox_cmds_read, "inc_sel=0xd,fvc=0x5000"),
+	INTEL_UNCORE_EVENT_DESC(bbox_cmds_write, "inc_sel=0xd,fvc=0x5040"),
+	{ /* end: all zeroes */ },
+};
+
 static struct intel_uncore_ops nhmex_uncore_mbox_ops = {
 	NHMEX_UNCORE_OPS_COMMON_INIT(),
 	.enable_event	= nhmex_mbox_msr_enable_event,
@@ -2791,7 +2823,13 @@ static int __init uncore_cpu_init(void)
 			snbep_uncore_cbox.num_boxes = max_cores;
 		msr_uncores = snbep_msr_uncores;
 		break;
-	case 46:
+	case 46: /* Nehalem-EX */
+		uncore_nhmex = true;
+	case 47: /* Westmere-EX aka. Xeon E7 */
+		if (!uncore_nhmex)
+			nhmex_uncore_mbox.event_descs = wsmex_uncore_mbox_events;
+		if (nhmex_uncore_cbox.num_boxes > max_cores)
+			nhmex_uncore_cbox.num_boxes = max_cores;
 		msr_uncores = nhmex_msr_uncores;
 		break;
 	default:
diff --git a/arch/x86/kernel/cpu/perf_event_intel_uncore.h b/arch/x86/kernel/cpu/perf_event_intel_uncore.h
index a51456e..9ee0bbe 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_uncore.h
+++ b/arch/x86/kernel/cpu/perf_event_intel_uncore.h
@@ -276,18 +276,12 @@
 		 NHMEX_M_PMON_CTL_INC_SEL_MASK |	\
 		 NHMEX_M_PMON_CTL_SET_FLAG_SEL_MASK)
 
-
-#define NHMEX_M_PMON_ZDP_CTL_FVC_FVID_MASK	0x1f
-#define NHMEX_M_PMON_ZDP_CTL_FVC_BCMD_MASK	(0x7 << 5)
-#define NHMEX_M_PMON_ZDP_CTL_FVC_RSP_MASK	(0x7 << 8)
-#define NHMEX_M_PMON_ZDP_CTL_FVC_PBOX_INIT_ERR	(1 << 23)
-#define NHMEX_M_PMON_ZDP_CTL_FVC_MASK			\
-		(NHMEX_M_PMON_ZDP_CTL_FVC_FVID_MASK |	\
-		 NHMEX_M_PMON_ZDP_CTL_FVC_BCMD_MASK |	\
-		 NHMEX_M_PMON_ZDP_CTL_FVC_RSP_MASK  |	\
-		 NHMEX_M_PMON_ZDP_CTL_FVC_PBOX_INIT_ERR)
+#define NHMEX_M_PMON_ZDP_CTL_FVC_MASK		(((1 << 11) - 1) | (1 << 23))
 #define NHMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(n)	(0x7 << (11 + 3 * (n)))
 
+#define WSMEX_M_PMON_ZDP_CTL_FVC_MASK		(((1 << 12) - 1) | (1 << 24))
+#define WSMEX_M_PMON_ZDP_CTL_FVC_EVENT_MASK(n)	(0x7 << (12 + 3 * (n)))
+
 /*
  * use the 9~13 bits to select event If the 7th bit is not set,
  * otherwise use the 19~21 bits to select event.
@@ -369,6 +363,7 @@ struct intel_uncore_type {
 	unsigned num_shared_regs:8;
 	unsigned single_fixed:1;
 	unsigned pair_ctr_ctl:1;
+	unsigned *msr_offsets;
 	struct event_constraint unconstrainted;
 	struct event_constraint *constraints;
 	struct intel_uncore_pmu *pmus;
@@ -486,29 +481,31 @@ unsigned uncore_pci_perf_ctr(struct intel_uncore_box *box, int idx)
 	return idx * 8 + box->pmu->type->perf_ctr;
 }
 
-static inline
-unsigned uncore_msr_box_ctl(struct intel_uncore_box *box)
+static inline unsigned uncore_msr_box_offset(struct intel_uncore_box *box)
+{
+	struct intel_uncore_pmu *pmu = box->pmu;
+	return pmu->type->msr_offsets ?
+		pmu->type->msr_offsets[pmu->pmu_idx] :
+		pmu->type->msr_offset * pmu->pmu_idx;
+}
+
+static inline unsigned uncore_msr_box_ctl(struct intel_uncore_box *box)
 {
 	if (!box->pmu->type->box_ctl)
 		return 0;
-	return box->pmu->type->box_ctl +
-		box->pmu->type->msr_offset * box->pmu->pmu_idx;
+	return box->pmu->type->box_ctl + uncore_msr_box_offset(box);
 }
 
-static inline
-unsigned uncore_msr_fixed_ctl(struct intel_uncore_box *box)
+static inline unsigned uncore_msr_fixed_ctl(struct intel_uncore_box *box)
 {
 	if (!box->pmu->type->fixed_ctl)
 		return 0;
-	return box->pmu->type->fixed_ctl +
-		box->pmu->type->msr_offset * box->pmu->pmu_idx;
+	return box->pmu->type->fixed_ctl + uncore_msr_box_offset(box);
 }
 
-static inline
-unsigned uncore_msr_fixed_ctr(struct intel_uncore_box *box)
+static inline unsigned uncore_msr_fixed_ctr(struct intel_uncore_box *box)
 {
-	return box->pmu->type->fixed_ctr +
-		box->pmu->type->msr_offset * box->pmu->pmu_idx;
+	return box->pmu->type->fixed_ctr + uncore_msr_box_offset(box);
 }
 
 static inline
@@ -516,7 +513,7 @@ unsigned uncore_msr_event_ctl(struct intel_uncore_box *box, int idx)
 {
 	return box->pmu->type->event_ctl +
 		(box->pmu->type->pair_ctr_ctl ? 2 * idx : idx) +
-		box->pmu->type->msr_offset * box->pmu->pmu_idx;
+		uncore_msr_box_offset(box);
 }
 
 static inline
@@ -524,7 +521,7 @@ unsigned uncore_msr_perf_ctr(struct intel_uncore_box *box, int idx)
 {
 	return box->pmu->type->perf_ctr +
 		(box->pmu->type->pair_ctr_ctl ? 2 * idx : idx) +
-		box->pmu->type->msr_offset * box->pmu->pmu_idx;
+		uncore_msr_box_offset(box);
 }
 
 static inline
-- 
1.7.11.2