lkml.org 
[lkml]   [2010]   [Mar]   [30]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: [PATCH 3/3] perf/core, x86: implement ARCH_PERFMON_EVENTSEL bit masks
On 29.03.10 18:48:58, Peter Zijlstra wrote:
> Could you fold this with your 2/3 and create x86_pmu_raw_event() which
> lives in arch/x86/kernel/cpu/perf_event.c, that's more consistent wrt
> the X86_RAW_EVENT_MASK name and that way you don't need to re-order the
> #include ""s either.

The patch below replaces patches 2 and 3.

-Robert

---
From 2d77650a4dc5ded763dc3c120381bdbe5a0be911 Mon Sep 17 00:00:00 2001
From: Robert Richter <robert.richter@amd.com>
Date: Wed, 17 Mar 2010 12:32:37 +0100
Subject: [PATCH] perf/core, x86: implement ARCH_PERFMON_EVENTSEL bit masks

ARCH_PERFMON_EVENTSEL bit masks are often used in the kernel. This
patch adds macros for the bit masks and removes local defines. The
function intel_pmu_raw_event() becomes x86_pmu_raw_event() which is
generic for x86 models and same also for p6. Duplicate code is
removed.

Signed-off-by: Robert Richter <robert.richter@amd.com>
---
arch/x86/include/asm/perf_event.h | 58 ++++++++++++++------------------
arch/x86/kernel/cpu/perf_event.c | 19 +++++++++--
arch/x86/kernel/cpu/perf_event_amd.c | 15 +--------
arch/x86/kernel/cpu/perf_event_intel.c | 22 +-----------
arch/x86/kernel/cpu/perf_event_p6.c | 20 +----------
5 files changed, 45 insertions(+), 89 deletions(-)
diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h
index 987bf67..f6d43db 100644
--- a/arch/x86/include/asm/perf_event.h
+++ b/arch/x86/include/asm/perf_event.h
@@ -18,39 +18,31 @@
#define MSR_ARCH_PERFMON_EVENTSEL0 0x186
#define MSR_ARCH_PERFMON_EVENTSEL1 0x187

-#define ARCH_PERFMON_EVENTSEL_ENABLE (1 << 22)
-#define ARCH_PERFMON_EVENTSEL_ANY (1 << 21)
-#define ARCH_PERFMON_EVENTSEL_INT (1 << 20)
-#define ARCH_PERFMON_EVENTSEL_OS (1 << 17)
-#define ARCH_PERFMON_EVENTSEL_USR (1 << 16)
-
-/*
- * Includes eventsel and unit mask as well:
- */
-
-
-#define INTEL_ARCH_EVTSEL_MASK 0x000000FFULL
-#define INTEL_ARCH_UNIT_MASK 0x0000FF00ULL
-#define INTEL_ARCH_EDGE_MASK 0x00040000ULL
-#define INTEL_ARCH_INV_MASK 0x00800000ULL
-#define INTEL_ARCH_CNT_MASK 0xFF000000ULL
-#define INTEL_ARCH_EVENT_MASK (INTEL_ARCH_UNIT_MASK|INTEL_ARCH_EVTSEL_MASK)
-
-/*
- * filter mask to validate fixed counter events.
- * the following filters disqualify for fixed counters:
- * - inv
- * - edge
- * - cnt-mask
- * The other filters are supported by fixed counters.
- * The any-thread option is supported starting with v3.
- */
-#define INTEL_ARCH_FIXED_MASK \
- (INTEL_ARCH_CNT_MASK| \
- INTEL_ARCH_INV_MASK| \
- INTEL_ARCH_EDGE_MASK|\
- INTEL_ARCH_UNIT_MASK|\
- INTEL_ARCH_EVENT_MASK)
+#define ARCH_PERFMON_EVENTSEL_EVENT 0x000000FFULL
+#define ARCH_PERFMON_EVENTSEL_UMASK 0x0000FF00ULL
+#define ARCH_PERFMON_EVENTSEL_USR (1ULL << 16)
+#define ARCH_PERFMON_EVENTSEL_OS (1ULL << 17)
+#define ARCH_PERFMON_EVENTSEL_EDGE (1ULL << 18)
+#define ARCH_PERFMON_EVENTSEL_INT (1ULL << 20)
+#define ARCH_PERFMON_EVENTSEL_ANY (1ULL << 21)
+#define ARCH_PERFMON_EVENTSEL_ENABLE (1ULL << 22)
+#define ARCH_PERFMON_EVENTSEL_INV (1ULL << 23)
+#define ARCH_PERFMON_EVENTSEL_CMASK 0xFF000000ULL
+
+#define AMD64_EVENTSEL_EVENT \
+ (ARCH_PERFMON_EVENTSEL_EVENT | (0x0FULL << 32))
+#define INTEL_ARCH_EVENT_MASK \
+ (ARCH_PERFMON_EVENTSEL_UMASK | ARCH_PERFMON_EVENTSEL_EVENT)
+
+#define X86_RAW_EVENT_MASK \
+ (ARCH_PERFMON_EVENTSEL_EVENT | \
+ ARCH_PERFMON_EVENTSEL_UMASK | \
+ ARCH_PERFMON_EVENTSEL_EDGE | \
+ ARCH_PERFMON_EVENTSEL_INV | \
+ ARCH_PERFMON_EVENTSEL_CMASK)
+#define AMD64_RAW_EVENT_MASK \
+ (X86_RAW_EVENT_MASK | \
+ AMD64_EVENTSEL_EVENT)

#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL 0x3c
#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK (0x00 << 8)
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index f2fc2d8..4e8d233 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -142,13 +142,21 @@ struct cpu_hw_events {
* Constraint on the Event code.
*/
#define INTEL_EVENT_CONSTRAINT(c, n) \
- EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVTSEL_MASK)
+ EVENT_CONSTRAINT(c, n, ARCH_PERFMON_EVENTSEL_EVENT)

/*
* Constraint on the Event code + UMask + fixed-mask
+ *
+ * filter mask to validate fixed counter events.
+ * the following filters disqualify for fixed counters:
+ * - inv
+ * - edge
+ * - cnt-mask
+ * The other filters are supported by fixed counters.
+ * The any-thread option is supported starting with v3.
*/
#define FIXED_EVENT_CONSTRAINT(c, n) \
- EVENT_CONSTRAINT(c, (1ULL << (32+n)), INTEL_ARCH_FIXED_MASK)
+ EVENT_CONSTRAINT(c, (1ULL << (32+n)), X86_RAW_EVENT_MASK)

/*
* Constraint on the Event code + UMask
@@ -436,6 +444,11 @@ static int x86_hw_config(struct perf_event_attr *attr, struct hw_perf_event *hwc
return 0;
}

+static u64 x86_pmu_raw_event(u64 hw_event)
+{
+ return hw_event & X86_RAW_EVENT_MASK;
+}
+
/*
* Setup the hardware configuration for a given attr_type
*/
@@ -1424,7 +1437,7 @@ void __init init_hw_perf_events(void)

if (x86_pmu.event_constraints) {
for_each_event_constraint(c, x86_pmu.event_constraints) {
- if (c->cmask != INTEL_ARCH_FIXED_MASK)
+ if (c->cmask != X86_RAW_EVENT_MASK)
continue;

c->idxmsk64 |= (1ULL << x86_pmu.num_counters) - 1;
diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c
index 9d363ce..30e799a 100644
--- a/arch/x86/kernel/cpu/perf_event_amd.c
+++ b/arch/x86/kernel/cpu/perf_event_amd.c
@@ -113,20 +113,7 @@ static u64 amd_pmu_event_map(int hw_event)

static u64 amd_pmu_raw_event(u64 hw_event)
{
-#define K7_EVNTSEL_EVENT_MASK 0xF000000FFULL
-#define K7_EVNTSEL_UNIT_MASK 0x00000FF00ULL
-#define K7_EVNTSEL_EDGE_MASK 0x000040000ULL
-#define K7_EVNTSEL_INV_MASK 0x000800000ULL
-#define K7_EVNTSEL_REG_MASK 0x0FF000000ULL
-
-#define K7_EVNTSEL_MASK \
- (K7_EVNTSEL_EVENT_MASK | \
- K7_EVNTSEL_UNIT_MASK | \
- K7_EVNTSEL_EDGE_MASK | \
- K7_EVNTSEL_INV_MASK | \
- K7_EVNTSEL_REG_MASK)
-
- return hw_event & K7_EVNTSEL_MASK;
+ return hw_event & AMD64_RAW_EVENT_MASK;
}

/*
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 0d68531..b65cfc3 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -452,24 +452,6 @@ static __initconst u64 atom_hw_cache_event_ids
},
};

-static u64 intel_pmu_raw_event(u64 hw_event)
-{
-#define CORE_EVNTSEL_EVENT_MASK 0x000000FFULL
-#define CORE_EVNTSEL_UNIT_MASK 0x0000FF00ULL
-#define CORE_EVNTSEL_EDGE_MASK 0x00040000ULL
-#define CORE_EVNTSEL_INV_MASK 0x00800000ULL
-#define CORE_EVNTSEL_REG_MASK 0xFF000000ULL
-
-#define CORE_EVNTSEL_MASK \
- (INTEL_ARCH_EVTSEL_MASK | \
- INTEL_ARCH_UNIT_MASK | \
- INTEL_ARCH_EDGE_MASK | \
- INTEL_ARCH_INV_MASK | \
- INTEL_ARCH_CNT_MASK)
-
- return hw_event & CORE_EVNTSEL_MASK;
-}
-
static void intel_pmu_disable_all(void)
{
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
@@ -788,7 +770,7 @@ static __initconst struct x86_pmu core_pmu = {
.eventsel = MSR_ARCH_PERFMON_EVENTSEL0,
.perfctr = MSR_ARCH_PERFMON_PERFCTR0,
.event_map = intel_pmu_event_map,
- .raw_event = intel_pmu_raw_event,
+ .raw_event = x86_pmu_raw_event,
.max_events = ARRAY_SIZE(intel_perfmon_event_map),
.apic = 1,
/*
@@ -827,7 +809,7 @@ static __initconst struct x86_pmu intel_pmu = {
.eventsel = MSR_ARCH_PERFMON_EVENTSEL0,
.perfctr = MSR_ARCH_PERFMON_PERFCTR0,
.event_map = intel_pmu_event_map,
- .raw_event = intel_pmu_raw_event,
+ .raw_event = x86_pmu_raw_event,
.max_events = ARRAY_SIZE(intel_perfmon_event_map),
.apic = 1,
/*
diff --git a/arch/x86/kernel/cpu/perf_event_p6.c b/arch/x86/kernel/cpu/perf_event_p6.c
index b26fbc7..03c139a 100644
--- a/arch/x86/kernel/cpu/perf_event_p6.c
+++ b/arch/x86/kernel/cpu/perf_event_p6.c
@@ -27,24 +27,6 @@ static u64 p6_pmu_event_map(int hw_event)
*/
#define P6_NOP_EVENT 0x0000002EULL

-static u64 p6_pmu_raw_event(u64 hw_event)
-{
-#define P6_EVNTSEL_EVENT_MASK 0x000000FFULL
-#define P6_EVNTSEL_UNIT_MASK 0x0000FF00ULL
-#define P6_EVNTSEL_EDGE_MASK 0x00040000ULL
-#define P6_EVNTSEL_INV_MASK 0x00800000ULL
-#define P6_EVNTSEL_REG_MASK 0xFF000000ULL
-
-#define P6_EVNTSEL_MASK \
- (P6_EVNTSEL_EVENT_MASK | \
- P6_EVNTSEL_UNIT_MASK | \
- P6_EVNTSEL_EDGE_MASK | \
- P6_EVNTSEL_INV_MASK | \
- P6_EVNTSEL_REG_MASK)
-
- return hw_event & P6_EVNTSEL_MASK;
-}
-
static struct event_constraint p6_event_constraints[] =
{
INTEL_EVENT_CONSTRAINT(0xc1, 0x1), /* FLOPS */
@@ -114,7 +96,7 @@ static __initconst struct x86_pmu p6_pmu = {
.eventsel = MSR_P6_EVNTSEL0,
.perfctr = MSR_P6_PERFCTR0,
.event_map = p6_pmu_event_map,
- .raw_event = p6_pmu_raw_event,
+ .raw_event = x86_pmu_raw_event,
.max_events = ARRAY_SIZE(p6_perfmon_event_map),
.apic = 1,
.max_period = (1ULL << 31) - 1,
--
1.7.0.3


--
Advanced Micro Devices, Inc.
Operating System Research Center
email: robert.richter@amd.com



\
 
 \ /
  Last update: 2010-03-30 11:31    [from the cache]
©2003-2011 Jasper Spaans