[tip:branch?] perf_counters: powerpc: Add support for POWER7 processors

[tip-bot for Paul Mackerras <paulus@samba.org>]

Commit-ID:  526e9a09d2e2dcb6c75c68710465678bb647e5fe
Gitweb:     http://git.kernel.org/tip/526e9a09d2e2dcb6c75c68710465678bb647e5fe
Author:     Paul Mackerras <paulus@samba.org>
AuthorDate: Thu, 11 Jun 2009 14:54:01 +1000
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Thu, 11 Jun 2009 12:07:37 +0200

perf_counters: powerpc: Add support for POWER7 processors

This adds the back-end for the PMU on POWER7 processors.  POWER7
has 4 fully-programmable counters and two fixed-function counters
(which do respect the freeze conditions, can generate interrupts,
and are writable, unlike PMC5/6 on POWER5+/6).

Signed-off-by: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <18992.36329.189378.17992@drongo.ozlabs.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>


---
 arch/powerpc/kernel/Makefile       |    3 +-
 arch/powerpc/kernel/perf_counter.c |    4 +
 arch/powerpc/kernel/power7-pmu.c   |  316 ++++++++++++++++++++++++++++++++++++
 3 files changed, 322 insertions(+), 1 deletions(-)

diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
index 9ba1bb7..a2c6834 100644
--- a/arch/powerpc/kernel/Makefile
+++ b/arch/powerpc/kernel/Makefile
@@ -95,7 +95,8 @@ obj64-$(CONFIG_AUDIT)		+= compat_audit.o
 obj-$(CONFIG_DYNAMIC_FTRACE)	+= ftrace.o
 obj-$(CONFIG_FUNCTION_GRAPH_TRACER)	+= ftrace.o
 obj-$(CONFIG_PERF_COUNTERS)	+= perf_counter.o power4-pmu.o ppc970-pmu.o \
-				   power5-pmu.o power5+-pmu.o power6-pmu.o
+				   power5-pmu.o power5+-pmu.o power6-pmu.o \
+				   power7-pmu.o
 
 obj-$(CONFIG_8XX_MINIMAL_FPEMU) += softemu8xx.o
 
diff --git a/arch/powerpc/kernel/perf_counter.c b/arch/powerpc/kernel/perf_counter.c
index 4990ce2..5d12e68 100644
--- a/arch/powerpc/kernel/perf_counter.c
+++ b/arch/powerpc/kernel/perf_counter.c
@@ -1181,6 +1181,7 @@ extern struct power_pmu ppc970_pmu;
 extern struct power_pmu power5_pmu;
 extern struct power_pmu power5p_pmu;
 extern struct power_pmu power6_pmu;
+extern struct power_pmu power7_pmu;
 
 static int init_perf_counters(void)
 {
@@ -1207,6 +1208,9 @@ static int init_perf_counters(void)
 	case 0x3e:
 		ppmu = &power6_pmu;
 		break;
+	case 0x3f:
+		ppmu = &power7_pmu;
+		break;
 	}
 
 	/*
diff --git a/arch/powerpc/kernel/power7-pmu.c b/arch/powerpc/kernel/power7-pmu.c
new file mode 100644
index 0000000..dfac48d
--- /dev/null
+++ b/arch/powerpc/kernel/power7-pmu.c
@@ -0,0 +1,316 @@
+/*
+ * Performance counter support for POWER7 processors.
+ *
+ * Copyright 2009 Paul Mackerras, IBM Corporation.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+#include <linux/kernel.h>
+#include <linux/perf_counter.h>
+#include <asm/reg.h>
+
+/*
+ * Bits in event code for POWER7
+ */
+#define PM_PMC_SH	16	/* PMC number (1-based) for direct events */
+#define PM_PMC_MSK	0xf
+#define PM_PMC_MSKS	(PM_PMC_MSK << PM_PMC_SH)
+#define PM_UNIT_SH	12	/* TTMMUX number and setting - unit select */
+#define PM_UNIT_MSK	0xf
+#define PM_COMBINE_SH	11	/* Combined event bit */
+#define PM_COMBINE_MSK	1
+#define PM_COMBINE_MSKS	0x800
+#define PM_L2SEL_SH	8	/* L2 event select */
+#define PM_L2SEL_MSK	7
+#define PM_PMCSEL_MSK	0xff
+
+/*
+ * Bits in MMCR1 for POWER7
+ */
+#define MMCR1_TTM0SEL_SH	60
+#define MMCR1_TTM1SEL_SH	56
+#define MMCR1_TTM2SEL_SH	52
+#define MMCR1_TTM3SEL_SH	48
+#define MMCR1_TTMSEL_MSK	0xf
+#define MMCR1_L2SEL_SH		45
+#define MMCR1_L2SEL_MSK		7
+#define MMCR1_PMC1_COMBINE_SH	35
+#define MMCR1_PMC2_COMBINE_SH	34
+#define MMCR1_PMC3_COMBINE_SH	33
+#define MMCR1_PMC4_COMBINE_SH	32
+#define MMCR1_PMC1SEL_SH	24
+#define MMCR1_PMC2SEL_SH	16
+#define MMCR1_PMC3SEL_SH	8
+#define MMCR1_PMC4SEL_SH	0
+#define MMCR1_PMCSEL_SH(n)	(MMCR1_PMC1SEL_SH - (n) * 8)
+#define MMCR1_PMCSEL_MSK	0xff
+
+/*
+ * Bits in MMCRA
+ */
+
+/*
+ * Layout of constraint bits:
+ * 6666555555555544444444443333333333222222222211111111110000000000
+ * 3210987654321098765432109876543210987654321098765432109876543210
+ *                                                 [  ><><><><><><>
+ *                                                  NC P6P5P4P3P2P1
+ *
+ * NC - number of counters
+ *     15: NC error 0x8000
+ *     12-14: number of events needing PMC1-4 0x7000
+ *
+ * P6
+ *     11: P6 error 0x800
+ *     10-11: Count of events needing PMC6
+ *
+ * P1..P5
+ *     0-9: Count of events needing PMC1..PMC5
+ */
+
+static int power7_get_constraint(u64 event, u64 *maskp, u64 *valp)
+{
+	int pmc, sh;
+	u64 mask = 0, value = 0;
+
+	pmc = (event >> PM_PMC_SH) & PM_PMC_MSK;
+	if (pmc) {
+		if (pmc > 6)
+			return -1;
+		sh = (pmc - 1) * 2;
+		mask |= 2 << sh;
+		value |= 1 << sh;
+		if (pmc >= 5 && !(event == 0x500fa || event == 0x600f4))
+			return -1;
+	}
+	if (pmc < 5) {
+		/* need a counter from PMC1-4 set */
+		mask  |= 0x8000;
+		value |= 0x1000;
+	}
+	*maskp = mask;
+	*valp = value;
+	return 0;
+}
+
+#define MAX_ALT	2	/* at most 2 alternatives for any event */
+
+static const unsigned int event_alternatives[][MAX_ALT] = {
+	{ 0x200f2, 0x300f2 },		/* PM_INST_DISP */
+	{ 0x200f4, 0x600f4 },		/* PM_RUN_CYC */
+	{ 0x400fa, 0x500fa },		/* PM_RUN_INST_CMPL */
+};
+
+/*
+ * Scan the alternatives table for a match and return the
+ * index into the alternatives table if found, else -1.
+ */
+static int find_alternative(u64 event)
+{
+	int i, j;
+
+	for (i = 0; i < ARRAY_SIZE(event_alternatives); ++i) {
+		if (event < event_alternatives[i][0])
+			break;
+		for (j = 0; j < MAX_ALT && event_alternatives[i][j]; ++j)
+			if (event == event_alternatives[i][j])
+				return i;
+	}
+	return -1;
+}
+
+static s64 find_alternative_decode(u64 event)
+{
+	int pmc, psel;
+
+	/* this only handles the 4x decode events */
+	pmc = (event >> PM_PMC_SH) & PM_PMC_MSK;
+	psel = event & PM_PMCSEL_MSK;
+	if ((pmc == 2 || pmc == 4) && (psel & ~7) == 0x40)
+		return event - (1 << PM_PMC_SH) + 8;
+	if ((pmc == 1 || pmc == 3) && (psel & ~7) == 0x48)
+		return event + (1 << PM_PMC_SH) - 8;
+	return -1;
+}
+
+static int power7_get_alternatives(u64 event, unsigned int flags, u64 alt[])
+{
+	int i, j, nalt = 1;
+	s64 ae;
+
+	alt[0] = event;
+	nalt = 1;
+	i = find_alternative(event);
+	if (i >= 0) {
+		for (j = 0; j < MAX_ALT; ++j) {
+			ae = event_alternatives[i][j];
+			if (ae && ae != event)
+				alt[nalt++] = ae;
+		}
+	} else {
+		ae = find_alternative_decode(event);
+		if (ae > 0)
+			alt[nalt++] = ae;
+	}
+
+	if (flags & PPMU_ONLY_COUNT_RUN) {
+		/*
+		 * We're only counting in RUN state,
+		 * so PM_CYC is equivalent to PM_RUN_CYC
+		 * and PM_INST_CMPL === PM_RUN_INST_CMPL.
+		 * This doesn't include alternatives that don't provide
+		 * any extra flexibility in assigning PMCs.
+		 */
+		j = nalt;
+		for (i = 0; i < nalt; ++i) {
+			switch (alt[i]) {
+			case 0x1e:	/* PM_CYC */
+				alt[j++] = 0x600f4;	/* PM_RUN_CYC */
+				break;
+			case 0x600f4:	/* PM_RUN_CYC */
+				alt[j++] = 0x1e;
+				break;
+			case 0x2:	/* PM_PPC_CMPL */
+				alt[j++] = 0x500fa;	/* PM_RUN_INST_CMPL */
+				break;
+			case 0x500fa:	/* PM_RUN_INST_CMPL */
+				alt[j++] = 0x2;	/* PM_PPC_CMPL */
+				break;
+			}
+		}
+		nalt = j;
+	}
+
+	return nalt;
+}
+
+/*
+ * Returns 1 if event counts things relating to marked instructions
+ * and thus needs the MMCRA_SAMPLE_ENABLE bit set, or 0 if not.
+ */
+static int power7_marked_instr_event(u64 event)
+{
+	int pmc, psel;
+	int unit;
+
+	pmc = (event >> PM_PMC_SH) & PM_PMC_MSK;
+	unit = (event >> PM_UNIT_SH) & PM_UNIT_MSK;
+	psel = event & PM_PMCSEL_MSK & ~1;	/* trim off edge/level bit */
+	if (pmc >= 5)
+		return 0;
+
+	switch (psel >> 4) {
+	case 2:
+		return pmc == 2 || pmc == 4;
+	case 3:
+		if (psel == 0x3c)
+			return pmc == 1;
+		if (psel == 0x3e)
+			return pmc != 2;
+		return 1;
+	case 4:
+	case 5:
+		return unit == 0xd;
+	case 6:
+		if (psel == 0x64)
+			return pmc >= 3;
+	case 8:
+		return unit == 0xd;
+	}
+	return 0;
+}
+
+static int power7_compute_mmcr(u64 event[], int n_ev,
+			       unsigned int hwc[], u64 mmcr[])
+{
+	u64 mmcr1 = 0;
+	u64 mmcra = 0;
+	unsigned int pmc, unit, combine, l2sel, psel;
+	unsigned int pmc_inuse = 0;
+	int i;
+
+	/* First pass to count resource use */
+	for (i = 0; i < n_ev; ++i) {
+		pmc = (event[i] >> PM_PMC_SH) & PM_PMC_MSK;
+		if (pmc) {
+			if (pmc > 6)
+				return -1;
+			if (pmc_inuse & (1 << (pmc - 1)))
+				return -1;
+			pmc_inuse |= 1 << (pmc - 1);
+		}
+	}
+
+	/* Second pass: assign PMCs, set all MMCR1 fields */
+	for (i = 0; i < n_ev; ++i) {
+		pmc = (event[i] >> PM_PMC_SH) & PM_PMC_MSK;
+		unit = (event[i] >> PM_UNIT_SH) & PM_UNIT_MSK;
+		combine = (event[i] >> PM_COMBINE_SH) & PM_COMBINE_MSK;
+		l2sel = (event[i] >> PM_L2SEL_SH) & PM_L2SEL_MSK;
+		psel = event[i] & PM_PMCSEL_MSK;
+		if (!pmc) {
+			/* Bus event or any-PMC direct event */
+			for (pmc = 0; pmc < 4; ++pmc) {
+				if (!(pmc_inuse & (1 << pmc)))
+					break;
+			}
+			if (pmc >= 4)
+				return -1;
+			pmc_inuse |= 1 << pmc;
+		} else {
+			/* Direct or decoded event */
+			--pmc;
+		}
+		if (pmc <= 3) {
+			mmcr1 |= (u64) unit << (MMCR1_TTM0SEL_SH - 4 * pmc);
+			mmcr1 |= (u64) combine << (MMCR1_PMC1_COMBINE_SH - pmc);
+			mmcr1 |= psel << MMCR1_PMCSEL_SH(pmc);
+			if (unit == 6)	/* L2 events */
+				mmcr1 |= (u64) l2sel << MMCR1_L2SEL_SH;
+		}
+		if (power7_marked_instr_event(event[i]))
+			mmcra |= MMCRA_SAMPLE_ENABLE;
+		hwc[i] = pmc;
+	}
+
+	/* Return MMCRx values */
+	mmcr[0] = 0;
+	if (pmc_inuse & 1)
+		mmcr[0] = MMCR0_PMC1CE;
+	if (pmc_inuse & 0x3e)
+		mmcr[0] |= MMCR0_PMCjCE;
+	mmcr[1] = mmcr1;
+	mmcr[2] = mmcra;
+	return 0;
+}
+
+static void power7_disable_pmc(unsigned int pmc, u64 mmcr[])
+{
+	if (pmc <= 3)
+		mmcr[1] &= ~(0xffULL << MMCR1_PMCSEL_SH(pmc));
+}
+
+static int power7_generic_events[] = {
+	[PERF_COUNT_CPU_CYCLES] = 0x1e,
+	[PERF_COUNT_INSTRUCTIONS] = 2,
+	[PERF_COUNT_CACHE_REFERENCES] = 0xc880,		/* LD_REF_L1_LSU */
+	[PERF_COUNT_CACHE_MISSES] = 0x400f0,		/* LD_MISS_L1 */
+	[PERF_COUNT_BRANCH_INSTRUCTIONS] = 0x10068,	/* BRU_FIN */
+	[PERF_COUNT_BRANCH_MISSES] = 0x400f6,		/* BR_MPRED */
+};
+
+struct power_pmu power7_pmu = {
+	.n_counter = 6,
+	.max_alternatives = MAX_ALT + 1,
+	.add_fields = 0x1555ull,
+	.test_adder = 0x3000ull,
+	.compute_mmcr = power7_compute_mmcr,
+	.get_constraint = power7_get_constraint,
+	.get_alternatives = power7_get_alternatives,
+	.disable_pmc = power7_disable_pmc,
+	.n_generic = ARRAY_SIZE(power7_generic_events),
+	.generic_events = power7_generic_events,
+};