Skip to contentSkip to conversation

perf: add memory access sampling support

PATCH In review Stephane Eranian 10 participants 68 messages Jan 24, 2013 – Apr 2, 2013

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

v7 cover perf: add memory access sampling support

This patch series had a new feature to the kernel perf_events interface and corresponding user level tool, perf.

With this patch, it is possible to sample (not trace) memory accesses (load, store). For loads, the instruction and data addresses are captured along with the latency and data source. For stores, the instruction and data addresses are capture along with limited cache and TLB information.

For load data source, the memory hierarchy level, the tlb, snoop and lock information is captured.

Although the perf_event interface is extended in a generic manner, sampling memory accesses requires HW support. The current patches implement the feature on Intel processors starting with Nehalem. The patches leverage the PEBS Load Latency and Precise Store mechanisms. Precise Store is present only on Sandy Bridge and Ivy Bridge based processors.

The perf tool is extended to make capturing and analyzing the data easier with a new command: perf mem.

$ perf mem -t load rec triad
$ perf mem -t load rep --stdio
# Samples: 19K of event 'cpu/mem-loads/pp'
# Total weight: 1013994
# Sort order  : local_weight,mem,sym,dso,symbol_daddr,dso_daddr,snoop,tlb,locked
#
# Overhead    Samples  Local Weight  Memory access Symbol      Shared Obj  Data Symbol             Data Object Snoop  TLB access   Locked
# ........  .........  ............  ............. ..........  ........... ......................  ........... ...... ............ ......
     0.10%          1           986  LFB hit       [.] triad   triad       [.] 0x00007f67dffe8038  [unknown]    None  L1 or L2 hit  No    
     0.09%          1           890  LFB hit       [.] triad   triad       [.] 0x00007f67df91a750  [unknown]    None  L1 or L2 hit  No    
     0.08%          1           826  LFB hit       [.] triad   triad       [.] 0x00007f67e288fba8  [unknown]    None  L1 or L2 hit  No    
     0.08%          1           825  LFB hit       [.] triad   triad       [.] 0x00007f67dea28c80  [unknown]    None  L1 or L2 hit  No    
     0.08%          1           787  LFB hit       [.] triad   triad       [.] 0x00007f67df055a60  [unknown]    None  L1 or L2 hit  No    

The perf mem command is a wrapper around perf record/report. It passes the right options to the report and record commands. Note that the TUI mode is supported. Overhead is relative to total cost, which is sum of the weights of each sample.

One powerful feature of perf is that users can toy with the sort order to display the information in different formats or from a different angle. This is particularly useful with memory sampling:

$ perf mem -t load rep --sort=mem
# Samples: 19K of event 'cpu/mem-loads/pp'
# Total weight: 1013994
# Sort order  : mem
#
# Overhead      Samples             Memory access
# ........  ...........  ........................
#
    85.26%        10633  LFB hit                 
     7.35%         8151  L1 hit                  
     3.13%          383  L3 hit                  
     3.09%          195  Local RAM hit           
     1.16%          259  L2 hit                  
     0.00%            4  Uncached hit            

Or if one is interested in the data view:

$ perf mem -t load rep --sort=symbol_daddr,local_weight
# Samples: 19K of event 'cpu/mem-loads/pp'
# Total cost : 1013994
# Sort order : symbol_daddr,cost
#
# Overhead      Samples             Data Symbol  Local Weight
# ........  ...........  ......................  ............
#
     0.10%            1  [.] 0x00007f67dffe8038           986
     0.09%            1  [.] 0x00007f67df91a750           890
     0.08%            1  [.] 0x00007f67e288fba8           826

One note on the weight (cost) displayed: On Intel processors with PEBS Load Latency, as described in the SDM, the cost encompasses the number of cycles from dispatch to Globally Observable (GO) state. That means, that it includes OOO execution. It is not usual to see L1D Hits with a cost of > 100 cycles. Always look at the memory level for an approximation of the access penalty, then interpret the cost value accordingly.

Data symbolization is working for initialized global variables. Dynamically allocated data and bss symbolization is currently non-functional.

There is no weight associated with stores.

In v2, we leverage some of Andi Kleen's Haswell patches, namely the weighted samples and perf tool event parser fixes. We also introduce PERF_RECORD_MISC_DATA_MMAP to tag mmaps for data vs. code. This helps the perf tool distinguish data. vs. code mmaps (and therefore symbols). We have also integrated the feedback from v1. Note that in v2 data symbol resolution is not yet fully operational, but there is a slight improvement.

In v3, we rebased the patch on 3.7.0-rc6 which includes certain of Nahyung's patches.

In v4, we rebase to v3.7 tip and also included the fixes from Namhyung Kim for symbolization of data addresses. We now have accesses to global variables working.

In v5, we rebase to 3.8.0-rc1. We also updated the WEIGHT patches from Andi to fix a couple of issues. Integrated the feedback from jolsa@. Reintegrated the man page.

In v6, we rebased to 3.8.0-rc3 and fixed the issue reported by jolsa@ related to hist_entry->mem_info maps not being marked as referenced.

In v7, we rebased on 3.8.0-rc4 and also merged in more patches from Andi's HSW series. We now use his weigthened sample support completely.

---

Andi Kleen (3):
perf, x86: Support CPU specific sysfs events perf, core: Add a concept of a weightened sample v2 perf, tools: Add support for weight v7 (modified)

Namhyung Kim (2):
perf tools: Ignore ABS symbols when loading data maps perf tools: Fix output of symbol_daddr offset

Stephane Eranian (13):
perf/x86: improve sysfs event mapping with event string perf/x86: add flags to event constraints
perf: add support for PERF_SAMPLE_ADDR in dump_sampple()
perf: add generic memory sampling interface perf/x86: add memory profiling via PEBS Load Latency perf/x86: export PEBS load latency threshold register to sysfs perf/x86: add support for PEBS Precise Store
perf tools: add mem access sampling core support perf report: add support for mem access profiling perf record: add support for mem access profiling perf tools: add new mem command for memory access profiling
perf: add PERF_RECORD_MISC_MMAP_DATA to RECORD_MMAP perf tools: detect data vs. text mappings

 arch/x86/include/uapi/asm/msr-index.h         |    1 +
 arch/x86/kernel/cpu/perf_event.c              |   67 +++--
 arch/x86/kernel/cpu/perf_event.h              |   62 ++++-
 arch/x86/kernel/cpu/perf_event_intel.c        |   35 ++-
 arch/x86/kernel/cpu/perf_event_intel_ds.c     |  182 ++++++++++++-
 arch/x86/kernel/cpu/perf_event_intel_uncore.c |    2 +-
 include/linux/perf_event.h                    |    5 +
 include/uapi/linux/perf_event.h               |   71 ++++-
 kernel/events/core.c                          |   15 ++
 tools/perf/Documentation/perf-mem.txt         |   48 ++++
 tools/perf/Documentation/perf-record.txt      |    6 +
 tools/perf/Documentation/perf-report.txt      |    2 +-
 tools/perf/Documentation/perf-top.txt         |    2 +-
 tools/perf/Makefile                           |    1 +
 tools/perf/builtin-annotate.c                 |    2 +-
 tools/perf/builtin-diff.c                     |    7 +-
 tools/perf/builtin-mem.c                      |  242 +++++++++++++++++
 tools/perf/builtin-record.c                   |    2 +
 tools/perf/builtin-report.c                   |  145 ++++++++++-
 tools/perf/builtin-top.c                      |    5 +-
 tools/perf/builtin.h                          |    1 +
 tools/perf/command-list.txt                   |    1 +
 tools/perf/perf.c                             |    1 +
 tools/perf/perf.h                             |    1 +
 tools/perf/util/event.h                       |    2 +
 tools/perf/util/evsel.c                       |   19 ++
 tools/perf/util/hist.c                        |  101 +++++++-
 tools/perf/util/hist.h                        |   21 +-
 tools/perf/util/machine.c                     |   10 +-
 tools/perf/util/session.c                     |   44 ++++
 tools/perf/util/session.h                     |    4 +
 tools/perf/util/sort.c                        |  346 ++++++++++++++++++++++++-
 tools/perf/util/sort.h                        |   12 +-
 tools/perf/util/symbol-elf.c                  |    3 +
 tools/perf/util/symbol.h                      |    6 +
 35 files changed, 1406 insertions(+), 68 deletions(-)
 create mode 100644 tools/perf/Documentation/perf-mem.txt
 create mode 100644 tools/perf/builtin-mem.c
Signature · 1 line
1.7.9.5
#1 lore raw

PATCH v7 1/18 perf, x86: Support CPU specific sysfs events

From: Andi Kleen <[email protected]>

Add a way for the CPU initialization code to register additional events, and merge them into the events attribute directory. Used in the next patch.

---
 arch/x86/kernel/cpu/perf_event.c |   33 +++++++++++++++++++++++++++++++++
 arch/x86/kernel/cpu/perf_event.h |    1 +
 2 files changed, 34 insertions(+)
arch/x86/kernel/cpu/perf_event.c +33 -0
index 6774c17..015668c 100644
@@ -1335,6 +1335,30 @@ static void __init filter_events(struct attribute **attrs)
 	}
 }
 
+/* Merge two pointer arrays */
+static __init struct attribute **merge_attr(struct attribute **a,
+					    struct attribute **b)
+{
+	struct attribute **new;
+	int j, i;
+
+	for (j = 0; a[j]; j++)
+		;
+	for (i = 0; b[i]; i++)
+		j++;
+	j++;
+	new = kmalloc(sizeof(struct attribute *) * j, GFP_KERNEL);
+	if (!new)
+		return NULL;
+	j = 0;
+	for (i = 0; a[i]; i++)
+		new[j++] = a[i];
+	for (i = 0; b[i]; i++)
+		new[j++] = b[i];
+	new[j] = NULL;
+	return new;
+}
+
 static ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
 			  char *page)
 {
@@ -1476,6 +1500,15 @@ static int __init init_hw_perf_events(void)
 	else
 		filter_events(x86_pmu_events_group.attrs);
 
+	if (x86_pmu.cpu_events) {
+		struct attribute *tmp;
+
+		tmp = merge_attr(x86_pmu_events_group.attrs,
+				 x86_pmu.cpu_events);
+		if (!WARN_ON(!tmp))
+			x86_pmu_events_group.attrs = tmp;
+	}
+
 	pr_info("... version:                %d\n",     x86_pmu.version);
 	pr_info("... bit width:              %d\n",     x86_pmu.cntval_bits);
 	pr_info("... generic registers:      %d\n",     x86_pmu.num_counters);
arch/x86/kernel/cpu/perf_event.h +1 -0
index 115c1ea..4170043 100644
@@ -355,6 +355,7 @@ struct x86_pmu {
 	struct attribute **format_attrs;
 
 	ssize_t		(*events_sysfs_show)(char *page, u64 config);
+	struct attribute **cpu_events;
 
 	/*
 	 * CPU Hotplug hooks
Signature · 1 line
1.7.9.5
#2 lore raw

[tip:perf/x86] perf/x86: Support CPU specific sysfs events

Commit-ID:  4c38cd7b6641f3f2cdf9c4f450c2628e2e400c1e
Gitweb:     http://git.kernel.org/tip/4c38cd7b6641f3f2cdf9c4f450c2628e2e400c1e
Author:     Andi Kleen <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:25 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 25 Jan 2013 10:19:00 +0100

perf/x86: Support CPU specific sysfs events

Add a way for the CPU initialization code to register additional events, and merge them into the events attribute directory. Used in the next patch.

---
 arch/x86/kernel/cpu/perf_event.c | 34 ++++++++++++++++++++++++++++++++++
 arch/x86/kernel/cpu/perf_event.h |  1 +
 2 files changed, 35 insertions(+)
arch/x86/kernel/cpu/perf_event.c +34 -0
index 6774c17..f133091 100644
@@ -1335,6 +1335,32 @@ static void __init filter_events(struct attribute **attrs)
 	}
 }
 
+/* Merge two pointer arrays */
+static __init struct attribute **merge_attr(struct attribute **a, struct attribute **b)
+{
+	struct attribute **new;
+	int j, i;
+
+	for (j = 0; a[j]; j++)
+		;
+	for (i = 0; b[i]; i++)
+		j++;
+	j++;
+
+	new = kmalloc(sizeof(struct attribute *) * j, GFP_KERNEL);
+	if (!new)
+		return NULL;
+
+	j = 0;
+	for (i = 0; a[i]; i++)
+		new[j++] = a[i];
+	for (i = 0; b[i]; i++)
+		new[j++] = b[i];
+	new[j] = NULL;
+
+	return new;
+}
+
 static ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
 			  char *page)
 {
@@ -1476,6 +1502,14 @@ static int __init init_hw_perf_events(void)
 	else
 		filter_events(x86_pmu_events_group.attrs);
 
+	if (x86_pmu.cpu_events) {
+		struct attribute *tmp;
+
+		tmp = merge_attr(x86_pmu_events_group.attrs, x86_pmu.cpu_events);
+		if (!WARN_ON(!tmp))
+			x86_pmu_events_group.attrs = tmp;
+	}
+
 	pr_info("... version:                %d\n",     x86_pmu.version);
 	pr_info("... bit width:              %d\n",     x86_pmu.cntval_bits);
 	pr_info("... generic registers:      %d\n",     x86_pmu.num_counters);
arch/x86/kernel/cpu/perf_event.h +1 -0
index 115c1ea..4170043 100644
@@ -355,6 +355,7 @@ struct x86_pmu {
 	struct attribute **format_attrs;
 
 	ssize_t		(*events_sysfs_show)(char *page, u64 config);
+	struct attribute **cpu_events;
 
 	/*
 	 * CPU Hotplug hooks
#3 lore raw

[tip:perf/core] perf/x86: Support CPU specific sysfs events

In reply to Stephane Eranian
perf/x86: Support CPU specific sysfs events
Commit-ID:  1a6461b12872e9622c231928e1620504d741cc79
Gitweb:     http://git.kernel.org/tip/1a6461b12872e9622c231928e1620504d741cc79
Author:     Andi Kleen <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:25 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Tue, 26 Mar 2013 16:50:23 -0300

perf/x86: Support CPU specific sysfs events

Add a way for the CPU initialization code to register additional events, and merge them into the events attribute directory. Used in the next patch.

---
 arch/x86/kernel/cpu/perf_event.c | 34 ++++++++++++++++++++++++++++++++++
 arch/x86/kernel/cpu/perf_event.h |  1 +
 2 files changed, 35 insertions(+)
arch/x86/kernel/cpu/perf_event.c +34 -0
index bf0f01a..c886dc8 100644
@@ -1330,6 +1330,32 @@ static void __init filter_events(struct attribute **attrs)
 	}
 }
 
+/* Merge two pointer arrays */
+static __init struct attribute **merge_attr(struct attribute **a, struct attribute **b)
+{
+	struct attribute **new;
+	int j, i;
+
+	for (j = 0; a[j]; j++)
+		;
+	for (i = 0; b[i]; i++)
+		j++;
+	j++;
+
+	new = kmalloc(sizeof(struct attribute *) * j, GFP_KERNEL);
+	if (!new)
+		return NULL;
+
+	j = 0;
+	for (i = 0; a[i]; i++)
+		new[j++] = a[i];
+	for (i = 0; b[i]; i++)
+		new[j++] = b[i];
+	new[j] = NULL;
+
+	return new;
+}
+
 static ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
 			  char *page)
 {
@@ -1469,6 +1495,14 @@ static int __init init_hw_perf_events(void)
 	else
 		filter_events(x86_pmu_events_group.attrs);
 
+	if (x86_pmu.cpu_events) {
+		struct attribute **tmp;
+
+		tmp = merge_attr(x86_pmu_events_group.attrs, x86_pmu.cpu_events);
+		if (!WARN_ON(!tmp))
+			x86_pmu_events_group.attrs = tmp;
+	}
+
 	pr_info("... version:                %d\n",     x86_pmu.version);
 	pr_info("... bit width:              %d\n",     x86_pmu.cntval_bits);
 	pr_info("... generic registers:      %d\n",     x86_pmu.num_counters);
arch/x86/kernel/cpu/perf_event.h +1 -0
index 7f5c75c..95152c1 100644
@@ -357,6 +357,7 @@ struct x86_pmu {
 	struct attribute **format_attrs;
 
 	ssize_t		(*events_sysfs_show)(char *page, u64 config);
+	struct attribute **cpu_events;
 
 	/*
 	 * CPU Hotplug hooks
#4 lore raw

PATCH v7 2/18 perf/x86: improve sysfs event mapping with event string

In reply to Stephane Eranian

This patch extends Jiri's changes to make generic events mapping visible via sysfs. The patch extends the mechanism to non-generic events by allowing the mappings to be hardcoded in strings.

This mechanism will be used by the PEBS-LL patch later on.

---
 arch/x86/kernel/cpu/perf_event.c |   27 ++++++++++++---------------
 arch/x86/kernel/cpu/perf_event.h |   23 +++++++++++++++++++++++
 2 files changed, 35 insertions(+), 15 deletions(-)
arch/x86/kernel/cpu/perf_event.c +12 -15
index 015668c..794f30e 100644
@@ -1310,20 +1310,22 @@ static struct attribute_group x86_pmu_format_group = {
 	.attrs = NULL,
 };
 
-struct perf_pmu_events_attr {
-	struct device_attribute attr;
-	u64 id;
-};
-
 /*
  * Remove all undefined events (x86_pmu.event_map(id) == 0)
  * out of events_attr attributes.
  */
 static void __init filter_events(struct attribute **attrs)
 {
+	struct device_attribute *d;
+	struct perf_pmu_events_attr *pmu_attr;
 	int i, j;
 
 	for (i = 0; attrs[i]; i++) {
+		d = (struct device_attribute *)attrs[i];
+		pmu_attr = container_of(d, struct perf_pmu_events_attr, attr);
+		/* str trumps id */
+		if (pmu_attr->event_str)
+			continue;
 		if (x86_pmu.event_map(i))
 			continue;
 
@@ -1364,19 +1366,14 @@ static ssize_t events_sysfs_show(struct device *dev, struct device_attribute *at
 {
 	struct perf_pmu_events_attr *pmu_attr = \
 		container_of(attr, struct perf_pmu_events_attr, attr);
-
 	u64 config = x86_pmu.event_map(pmu_attr->id);
-	return x86_pmu.events_sysfs_show(page, config);
-}
 
-#define EVENT_VAR(_id)  event_attr_##_id
-#define EVENT_PTR(_id) &event_attr_##_id.attr.attr
+	/* string trumps id */
+	if (pmu_attr->event_str)
+		return sprintf(page, "%s", pmu_attr->event_str);
 
-#define EVENT_ATTR(_name, _id)					\
-static struct perf_pmu_events_attr EVENT_VAR(_id) = {		\
-	.attr = __ATTR(_name, 0444, events_sysfs_show, NULL),	\
-	.id   =  PERF_COUNT_HW_##_id,				\
-};
+	return x86_pmu.events_sysfs_show(page, config);
+}
 
 EVENT_ATTR(cpu-cycles,			CPU_CYCLES		);
 EVENT_ATTR(instructions,		INSTRUCTIONS		);
arch/x86/kernel/cpu/perf_event.h +23 -0
index 4170043..3f4380c 100644
@@ -420,6 +420,29 @@ do {									\
 #define ERF_NO_HT_SHARING	1
 #define ERF_HAS_RSP_1		2
 
+#define EVENT_VAR(_id)  event_attr_##_id
+#define EVENT_PTR(_id) &event_attr_##_id.attr.attr
+
+#define EVENT_ATTR(_name, _id)					\
+static struct perf_pmu_events_attr EVENT_VAR(_id) = {		\
+	.attr = __ATTR(_name, 0444, events_sysfs_show, NULL),	\
+	.id   =  PERF_COUNT_HW_##_id,				\
+	.event_str = NULL,					\
+};
+
+#define EVENT_ATTR_STR(_name, v, str)				  \
+static struct perf_pmu_events_attr event_attr_##v = {		  \
+	.attr      = __ATTR(_name, 0444, events_sysfs_show, NULL),\
+	.id        =  0,					  \
+	.event_str =  str,					  \
+};
+
+struct perf_pmu_events_attr {
+	struct device_attribute attr;
+	u64 id;
+	const char *event_str;
+};
+
 extern struct x86_pmu x86_pmu __read_mostly;
 
 DECLARE_PER_CPU(struct cpu_hw_events, cpu_hw_events);
Signature · 1 line
1.7.9.5
#5 lore raw

[tip:perf/x86] perf/x86: Improve sysfs event mapping with event string

Commit-ID:  5256a7d63866c51815fb59d8b850d678b4e32357
Gitweb:     http://git.kernel.org/tip/5256a7d63866c51815fb59d8b850d678b4e32357
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:26 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 25 Jan 2013 10:19:01 +0100

perf/x86: Improve sysfs event mapping with event string

This patch extends Jiri's changes to make generic events mapping visible via sysfs. The patch extends the mechanism to non-generic events by allowing the mappings to be hardcoded in strings.

This mechanism will be used by the PEBS-LL patch later on.

---
 arch/x86/kernel/cpu/perf_event.c | 27 ++++++++++++---------------
 arch/x86/kernel/cpu/perf_event.h | 23 +++++++++++++++++++++++
 2 files changed, 35 insertions(+), 15 deletions(-)
arch/x86/kernel/cpu/perf_event.c +12 -15
index f133091..fd80a0c 100644
@@ -1310,20 +1310,22 @@ static struct attribute_group x86_pmu_format_group = {
 	.attrs = NULL,
 };
 
-struct perf_pmu_events_attr {
-	struct device_attribute attr;
-	u64 id;
-};
-
 /*
  * Remove all undefined events (x86_pmu.event_map(id) == 0)
  * out of events_attr attributes.
  */
 static void __init filter_events(struct attribute **attrs)
 {
+	struct device_attribute *d;
+	struct perf_pmu_events_attr *pmu_attr;
 	int i, j;
 
 	for (i = 0; attrs[i]; i++) {
+		d = (struct device_attribute *)attrs[i];
+		pmu_attr = container_of(d, struct perf_pmu_events_attr, attr);
+		/* str trumps id */
+		if (pmu_attr->event_str)
+			continue;
 		if (x86_pmu.event_map(i))
 			continue;
 
@@ -1366,19 +1368,14 @@ static ssize_t events_sysfs_show(struct device *dev, struct device_attribute *at
 {
 	struct perf_pmu_events_attr *pmu_attr = \
 		container_of(attr, struct perf_pmu_events_attr, attr);
-
 	u64 config = x86_pmu.event_map(pmu_attr->id);
-	return x86_pmu.events_sysfs_show(page, config);
-}
 
-#define EVENT_VAR(_id)  event_attr_##_id
-#define EVENT_PTR(_id) &event_attr_##_id.attr.attr
+	/* string trumps id */
+	if (pmu_attr->event_str)
+		return sprintf(page, "%s", pmu_attr->event_str);
 
-#define EVENT_ATTR(_name, _id)					\
-static struct perf_pmu_events_attr EVENT_VAR(_id) = {		\
-	.attr = __ATTR(_name, 0444, events_sysfs_show, NULL),	\
-	.id   =  PERF_COUNT_HW_##_id,				\
-};
+	return x86_pmu.events_sysfs_show(page, config);
+}
 
 EVENT_ATTR(cpu-cycles,			CPU_CYCLES		);
 EVENT_ATTR(instructions,		INSTRUCTIONS		);
arch/x86/kernel/cpu/perf_event.h +23 -0
index 4170043..dda2101 100644
@@ -420,6 +420,29 @@ do {									\
 #define ERF_NO_HT_SHARING	1
 #define ERF_HAS_RSP_1		2
 
+#define EVENT_VAR(_id)  event_attr_##_id
+#define EVENT_PTR(_id) &event_attr_##_id.attr.attr
+
+#define EVENT_ATTR(_name, _id)						\
+static struct perf_pmu_events_attr EVENT_VAR(_id) = {			\
+	.attr		= __ATTR(_name, 0444, events_sysfs_show, NULL),	\
+	.id		= PERF_COUNT_HW_##_id,				\
+	.event_str	= NULL,						\
+};
+
+#define EVENT_ATTR_STR(_name, v, str)					\
+static struct perf_pmu_events_attr event_attr_##v = {			\
+	.attr		= __ATTR(_name, 0444, events_sysfs_show, NULL),	\
+	.id		= 0,						\
+	.event_str	= str,						\
+};
+
+struct perf_pmu_events_attr {
+	struct device_attribute attr;
+	u64 id;
+	const char *event_str;
+};
+
 extern struct x86_pmu x86_pmu __read_mostly;
 
 DECLARE_PER_CPU(struct cpu_hw_events, cpu_hw_events);
#6 lore raw

[tip:perf/core] perf/x86: Improve sysfs event mapping with event string

In reply to Stephane Eranian
perf/x86: Improve sysfs event mapping with event string
Commit-ID:  3a54aaa0a3ddb2cf2ec1b94a94024e9a8a8af962
Gitweb:     http://git.kernel.org/tip/3a54aaa0a3ddb2cf2ec1b94a94024e9a8a8af962
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:26 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Tue, 26 Mar 2013 17:36:45 -0300

perf/x86: Improve sysfs event mapping with event string

This patch extends Jiri's changes to make generic events mapping visible via sysfs. The patch extends the mechanism to non-generic events by allowing the mappings to be hardcoded in strings.

This mechanism will be used by the PEBS-LL patch
later on.

---
 arch/x86/kernel/cpu/perf_event.c | 20 ++++++++++++--------
 arch/x86/kernel/cpu/perf_event.h | 17 +++++++++++++++++
 include/linux/perf_event.h       |  1 +
 3 files changed, 30 insertions(+), 8 deletions(-)
arch/x86/kernel/cpu/perf_event.c +12 -8
index c886dc8..6e8ab04 100644
@@ -1316,9 +1316,16 @@ static struct attribute_group x86_pmu_format_group = {
  */
 static void __init filter_events(struct attribute **attrs)
 {
+	struct device_attribute *d;
+	struct perf_pmu_events_attr *pmu_attr;
 	int i, j;
 
 	for (i = 0; attrs[i]; i++) {
+		d = (struct device_attribute *)attrs[i];
+		pmu_attr = container_of(d, struct perf_pmu_events_attr, attr);
+		/* str trumps id */
+		if (pmu_attr->event_str)
+			continue;
 		if (x86_pmu.event_map(i))
 			continue;
 
@@ -1361,17 +1368,14 @@ static ssize_t events_sysfs_show(struct device *dev, struct device_attribute *at
 {
 	struct perf_pmu_events_attr *pmu_attr = \
 		container_of(attr, struct perf_pmu_events_attr, attr);
-
 	u64 config = x86_pmu.event_map(pmu_attr->id);
-	return x86_pmu.events_sysfs_show(page, config);
-}
 
-#define EVENT_VAR(_id)  event_attr_##_id
-#define EVENT_PTR(_id) &event_attr_##_id.attr.attr
+	/* string trumps id */
+	if (pmu_attr->event_str)
+		return sprintf(page, "%s", pmu_attr->event_str);
 
-#define EVENT_ATTR(_name, _id)						\
-	PMU_EVENT_ATTR(_name, EVENT_VAR(_id), PERF_COUNT_HW_##_id,	\
-			events_sysfs_show)
+	return x86_pmu.events_sysfs_show(page, config);
+}
 
 EVENT_ATTR(cpu-cycles,			CPU_CYCLES		);
 EVENT_ATTR(instructions,		INSTRUCTIONS		);
arch/x86/kernel/cpu/perf_event.h +17 -0
index 95152c1..b1518ee 100644
@@ -422,6 +422,23 @@ do {									\
 #define ERF_NO_HT_SHARING	1
 #define ERF_HAS_RSP_1		2
 
+#define EVENT_VAR(_id)  event_attr_##_id
+#define EVENT_PTR(_id) &event_attr_##_id.attr.attr
+
+#define EVENT_ATTR(_name, _id)						\
+static struct perf_pmu_events_attr EVENT_VAR(_id) = {			\
+	.attr		= __ATTR(_name, 0444, events_sysfs_show, NULL),	\
+	.id		= PERF_COUNT_HW_##_id,				\
+	.event_str	= NULL,						\
+};
+
+#define EVENT_ATTR_STR(_name, v, str)					\
+static struct perf_pmu_events_attr event_attr_##v = {			\
+	.attr		= __ATTR(_name, 0444, events_sysfs_show, NULL),	\
+	.id		= 0,						\
+	.event_str	= str,						\
+};
+
 extern struct x86_pmu x86_pmu __read_mostly;
 
 DECLARE_PER_CPU(struct cpu_hw_events, cpu_hw_events);
include/linux/perf_event.h +1 -0
index 8737e1c..1c59211 100644
@@ -809,6 +809,7 @@ do {									\
 struct perf_pmu_events_attr {
 	struct device_attribute attr;
 	u64 id;
+	const char *event_str;
 };
 
 #define PMU_EVENT_ATTR(_name, _var, _id, _show)				\
#7 lore raw

PATCH v7 3/18 perf/x86: add flags to event constraints

In reply to Stephane Eranian

This patch adds a flags field to each event constraint. It can be used to store event specific features which can then later be used by scheduling code or low-level x86 code.

The flags are propagated into event->hw.flags during the get_event_constraint() call. They are cleared during the put_event_constraint() call.

This mechanism is going to be used by the PEBS-LL patches. It avoids defining yet another table to hold event specific information.

---
 arch/x86/kernel/cpu/perf_event.c              |    2 +-
 arch/x86/kernel/cpu/perf_event.h              |    8 +++++---
 arch/x86/kernel/cpu/perf_event_intel.c        |    6 +++++-
 arch/x86/kernel/cpu/perf_event_intel_ds.c     |    4 +++-
 arch/x86/kernel/cpu/perf_event_intel_uncore.c |    2 +-
 include/linux/perf_event.h                    |    1 +
 6 files changed, 16 insertions(+), 7 deletions(-)
arch/x86/kernel/cpu/perf_event.c +1 -1
index 794f30e..63f8dcf 100644
@@ -1487,7 +1487,7 @@ static int __init init_hw_perf_events(void)
 
 	unconstrained = (struct event_constraint)
 		__EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_counters) - 1,
-				   0, x86_pmu.num_counters, 0);
+				   0, x86_pmu.num_counters, 0, 0);
 
 	x86_pmu.attr_rdpmc = 1; /* enable userspace RDPMC usage by default */
 	x86_pmu_format_group.attrs = x86_pmu.format_attrs;
arch/x86/kernel/cpu/perf_event.h +5 -3
index 3f4380c..3f10cfe 100644
@@ -59,6 +59,7 @@ struct event_constraint {
 	u64	cmask;
 	int	weight;
 	int	overlap;
+	int	flags;
 };
 
 struct amd_nb {
@@ -170,16 +171,17 @@ struct cpu_hw_events {
 	void				*kfree_on_online;
 };
 
-#define __EVENT_CONSTRAINT(c, n, m, w, o) {\
+#define __EVENT_CONSTRAINT(c, n, m, w, o, f) {\
 	{ .idxmsk64 = (n) },		\
 	.code = (c),			\
 	.cmask = (m),			\
 	.weight = (w),			\
 	.overlap = (o),			\
+	.flags = f,			\
 }
 
 #define EVENT_CONSTRAINT(c, n, m)	\
-	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 0)
+	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 0, 0)
 
 /*
  * The overlap flag marks event constraints with overlapping counter
@@ -203,7 +205,7 @@ struct cpu_hw_events {
  * and its counter masks must be kept at a minimum.
  */
 #define EVENT_CONSTRAINT_OVERLAP(c, n, m)	\
-	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 1)
+	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 1, 0)
 
 /*
  * Constraint on the Event code.
arch/x86/kernel/cpu/perf_event_intel.c +5 -1
index 93b9e11..67a8dd6 100644
@@ -1367,8 +1367,11 @@ x86_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event)
 
 	if (x86_pmu.event_constraints) {
 		for_each_event_constraint(c, x86_pmu.event_constraints) {
-			if ((event->hw.config & c->cmask) == c->code)
+			if ((event->hw.config & c->cmask) == c->code) {
+				/* hw.flags zeroed at initialization */
+				event->hw.flags |= c->flags;
 				return c;
+			}
 		}
 	}
 
@@ -1413,6 +1416,7 @@ intel_put_shared_regs_event_constraints(struct cpu_hw_events *cpuc,
 static void intel_put_event_constraints(struct cpu_hw_events *cpuc,
 					struct perf_event *event)
 {
+	event->hw.flags = 0;
 	intel_put_shared_regs_event_constraints(cpuc, event);
 }
 
arch/x86/kernel/cpu/perf_event_intel_ds.c +3 -1
index 826054a..f30d85b 100644
@@ -430,8 +430,10 @@ struct event_constraint *intel_pebs_constraints(struct perf_event *event)
 
 	if (x86_pmu.pebs_constraints) {
 		for_each_event_constraint(c, x86_pmu.pebs_constraints) {
-			if ((event->hw.config & c->cmask) == c->code)
+			if ((event->hw.config & c->cmask) == c->code) {
+				event->hw.flags |= c->flags;
 				return c;
+			}
 		}
 	}
 
arch/x86/kernel/cpu/perf_event_intel_uncore.c +1 -1
index b43200d..75da9e1 100644
@@ -2438,7 +2438,7 @@ static int __init uncore_type_init(struct intel_uncore_type *type)
 
 	type->unconstrainted = (struct event_constraint)
 		__EVENT_CONSTRAINT(0, (1ULL << type->num_counters) - 1,
-				0, type->num_counters, 0);
+				0, type->num_counters, 0, 0);
 
 	for (i = 0; i < type->num_boxes; i++) {
 		pmus[i].func_id = -1;
include/linux/perf_event.h +1 -0
index 6bfb2faa..484cfbc 100644
@@ -128,6 +128,7 @@ struct hw_perf_event {
 			int		event_base_rdpmc;
 			int		idx;
 			int		last_cpu;
+			int		flags;
 
 			struct hw_perf_event_extra extra_reg;
 			struct hw_perf_event_extra branch_reg;
Signature · 1 line
1.7.9.5
#8 lore raw

[tip:perf/x86] perf/x86: Add flags to event constraints

Commit-ID:  1a4e0aca7416fd2b859e08a587a36a4bc00d4f46
Gitweb:     http://git.kernel.org/tip/1a4e0aca7416fd2b859e08a587a36a4bc00d4f46
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:27 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 25 Jan 2013 10:19:01 +0100

perf/x86: Add flags to event constraints

This patch adds a flags field to each event constraint. It can be used to store event specific features which can then later be used by scheduling code or low-level x86 code.

The flags are propagated into event->hw.flags during the get_event_constraint() call. They are cleared during the put_event_constraint() call.

This mechanism is going to be used by the PEBS-LL patches. It avoids defining yet another table to hold event specific information.

---
 arch/x86/kernel/cpu/perf_event.c              | 2 +-
 arch/x86/kernel/cpu/perf_event.h              | 8 +++++---
 arch/x86/kernel/cpu/perf_event_intel.c        | 6 +++++-
 arch/x86/kernel/cpu/perf_event_intel_ds.c     | 4 +++-
 arch/x86/kernel/cpu/perf_event_intel_uncore.c | 2 +-
 include/linux/perf_event.h                    | 1 +
 6 files changed, 16 insertions(+), 7 deletions(-)
arch/x86/kernel/cpu/perf_event.c +1 -1
index fd80a0c..49f760b 100644
@@ -1489,7 +1489,7 @@ static int __init init_hw_perf_events(void)
 
 	unconstrained = (struct event_constraint)
 		__EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_counters) - 1,
-				   0, x86_pmu.num_counters, 0);
+				   0, x86_pmu.num_counters, 0, 0);
 
 	x86_pmu.attr_rdpmc = 1; /* enable userspace RDPMC usage by default */
 	x86_pmu_format_group.attrs = x86_pmu.format_attrs;
arch/x86/kernel/cpu/perf_event.h +5 -3
index dda2101..74b59ec 100644
@@ -59,6 +59,7 @@ struct event_constraint {
 	u64	cmask;
 	int	weight;
 	int	overlap;
+	int	flags;
 };
 
 struct amd_nb {
@@ -170,16 +171,17 @@ struct cpu_hw_events {
 	void				*kfree_on_online;
 };
 
-#define __EVENT_CONSTRAINT(c, n, m, w, o) {\
+#define __EVENT_CONSTRAINT(c, n, m, w, o, f) {\
 	{ .idxmsk64 = (n) },		\
 	.code = (c),			\
 	.cmask = (m),			\
 	.weight = (w),			\
 	.overlap = (o),			\
+	.flags = f,			\
 }
 
 #define EVENT_CONSTRAINT(c, n, m)	\
-	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 0)
+	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 0, 0)
 
 /*
  * The overlap flag marks event constraints with overlapping counter
@@ -203,7 +205,7 @@ struct cpu_hw_events {
  * and its counter masks must be kept at a minimum.
  */
 #define EVENT_CONSTRAINT_OVERLAP(c, n, m)	\
-	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 1)
+	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 1, 0)
 
 /*
  * Constraint on the Event code.
arch/x86/kernel/cpu/perf_event_intel.c +5 -1
index 93b9e11..67a8dd6 100644
@@ -1367,8 +1367,11 @@ x86_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event)
 
 	if (x86_pmu.event_constraints) {
 		for_each_event_constraint(c, x86_pmu.event_constraints) {
-			if ((event->hw.config & c->cmask) == c->code)
+			if ((event->hw.config & c->cmask) == c->code) {
+				/* hw.flags zeroed at initialization */
+				event->hw.flags |= c->flags;
 				return c;
+			}
 		}
 	}
 
@@ -1413,6 +1416,7 @@ intel_put_shared_regs_event_constraints(struct cpu_hw_events *cpuc,
 static void intel_put_event_constraints(struct cpu_hw_events *cpuc,
 					struct perf_event *event)
 {
+	event->hw.flags = 0;
 	intel_put_shared_regs_event_constraints(cpuc, event);
 }
 
arch/x86/kernel/cpu/perf_event_intel_ds.c +3 -1
index 826054a..f30d85b 100644
@@ -430,8 +430,10 @@ struct event_constraint *intel_pebs_constraints(struct perf_event *event)
 
 	if (x86_pmu.pebs_constraints) {
 		for_each_event_constraint(c, x86_pmu.pebs_constraints) {
-			if ((event->hw.config & c->cmask) == c->code)
+			if ((event->hw.config & c->cmask) == c->code) {
+				event->hw.flags |= c->flags;
 				return c;
+			}
 		}
 	}
 
arch/x86/kernel/cpu/perf_event_intel_uncore.c +1 -1
index b43200d..75da9e1 100644
@@ -2438,7 +2438,7 @@ static int __init uncore_type_init(struct intel_uncore_type *type)
 
 	type->unconstrainted = (struct event_constraint)
 		__EVENT_CONSTRAINT(0, (1ULL << type->num_counters) - 1,
-				0, type->num_counters, 0);
+				0, type->num_counters, 0, 0);
 
 	for (i = 0; i < type->num_boxes; i++) {
 		pmus[i].func_id = -1;
include/linux/perf_event.h +1 -0
index 6bfb2faa..484cfbc 100644
@@ -128,6 +128,7 @@ struct hw_perf_event {
 			int		event_base_rdpmc;
 			int		idx;
 			int		last_cpu;
+			int		flags;
 
 			struct hw_perf_event_extra extra_reg;
 			struct hw_perf_event_extra branch_reg;
#9 lore raw

[tip:perf/core] perf/x86: Add flags to event constraints

In reply to Stephane Eranian
perf/x86: Add flags to event constraints
Commit-ID:  9fac2cf316b070ae43d2ae2525e381ff2d1d68aa
Gitweb:     http://git.kernel.org/tip/9fac2cf316b070ae43d2ae2525e381ff2d1d68aa
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:27 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:15:04 -0300

perf/x86: Add flags to event constraints

This patch adds a flags field to each event constraint. It can be used to store event specific features which can then later be used by scheduling code or low-level x86 code.

The flags are propagated into event->hw.flags during the get_event_constraint() call. They are cleared during the put_event_constraint() call.

This mechanism is going to be used by the PEBS-LL patches. It avoids defining yet another table to hold event specific information.

---
 arch/x86/kernel/cpu/perf_event.c              | 2 +-
 arch/x86/kernel/cpu/perf_event.h              | 8 +++++---
 arch/x86/kernel/cpu/perf_event_intel.c        | 6 +++++-
 arch/x86/kernel/cpu/perf_event_intel_ds.c     | 4 +++-
 arch/x86/kernel/cpu/perf_event_intel_uncore.c | 2 +-
 include/linux/perf_event.h                    | 1 +
 6 files changed, 16 insertions(+), 7 deletions(-)
arch/x86/kernel/cpu/perf_event.c +1 -1
index 6e8ab04..8ba5151 100644
@@ -1489,7 +1489,7 @@ static int __init init_hw_perf_events(void)
 
 	unconstrained = (struct event_constraint)
 		__EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_counters) - 1,
-				   0, x86_pmu.num_counters, 0);
+				   0, x86_pmu.num_counters, 0, 0);
 
 	x86_pmu.attr_rdpmc = 1; /* enable userspace RDPMC usage by default */
 	x86_pmu_format_group.attrs = x86_pmu.format_attrs;
arch/x86/kernel/cpu/perf_event.h +5 -3
index b1518ee..9686d38 100644
@@ -59,6 +59,7 @@ struct event_constraint {
 	u64	cmask;
 	int	weight;
 	int	overlap;
+	int	flags;
 };
 
 struct amd_nb {
@@ -170,16 +171,17 @@ struct cpu_hw_events {
 	void				*kfree_on_online;
 };
 
-#define __EVENT_CONSTRAINT(c, n, m, w, o) {\
+#define __EVENT_CONSTRAINT(c, n, m, w, o, f) {\
 	{ .idxmsk64 = (n) },		\
 	.code = (c),			\
 	.cmask = (m),			\
 	.weight = (w),			\
 	.overlap = (o),			\
+	.flags = f,			\
 }
 
 #define EVENT_CONSTRAINT(c, n, m)	\
-	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 0)
+	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 0, 0)
 
 /*
  * The overlap flag marks event constraints with overlapping counter
@@ -203,7 +205,7 @@ struct cpu_hw_events {
  * and its counter masks must be kept at a minimum.
  */
 #define EVENT_CONSTRAINT_OVERLAP(c, n, m)	\
-	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 1)
+	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 1, 0)
 
 /*
  * Constraint on the Event code.
arch/x86/kernel/cpu/perf_event_intel.c +5 -1
index dab7580..df3beaa 100644
@@ -1392,8 +1392,11 @@ x86_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event)
 
 	if (x86_pmu.event_constraints) {
 		for_each_event_constraint(c, x86_pmu.event_constraints) {
-			if ((event->hw.config & c->cmask) == c->code)
+			if ((event->hw.config & c->cmask) == c->code) {
+				/* hw.flags zeroed at initialization */
+				event->hw.flags |= c->flags;
 				return c;
+			}
 		}
 	}
 
@@ -1438,6 +1441,7 @@ intel_put_shared_regs_event_constraints(struct cpu_hw_events *cpuc,
 static void intel_put_event_constraints(struct cpu_hw_events *cpuc,
 					struct perf_event *event)
 {
+	event->hw.flags = 0;
 	intel_put_shared_regs_event_constraints(cpuc, event);
 }
 
arch/x86/kernel/cpu/perf_event_intel_ds.c +3 -1
index 826054a..f30d85b 100644
@@ -430,8 +430,10 @@ struct event_constraint *intel_pebs_constraints(struct perf_event *event)
 
 	if (x86_pmu.pebs_constraints) {
 		for_each_event_constraint(c, x86_pmu.pebs_constraints) {
-			if ((event->hw.config & c->cmask) == c->code)
+			if ((event->hw.config & c->cmask) == c->code) {
+				event->hw.flags |= c->flags;
 				return c;
+			}
 		}
 	}
 
arch/x86/kernel/cpu/perf_event_intel_uncore.c +1 -1
index b43200d..75da9e1 100644
@@ -2438,7 +2438,7 @@ static int __init uncore_type_init(struct intel_uncore_type *type)
 
 	type->unconstrainted = (struct event_constraint)
 		__EVENT_CONSTRAINT(0, (1ULL << type->num_counters) - 1,
-				0, type->num_counters, 0);
+				0, type->num_counters, 0, 0);
 
 	for (i = 0; i < type->num_boxes; i++) {
 		pmus[i].func_id = -1;
include/linux/perf_event.h +1 -0
index 1c59211..cd3bb2c 100644
@@ -127,6 +127,7 @@ struct hw_perf_event {
 			int		event_base_rdpmc;
 			int		idx;
 			int		last_cpu;
+			int		flags;
 
 			struct hw_perf_event_extra extra_reg;
 			struct hw_perf_event_extra branch_reg;
#10 lore raw

PATCH v7 4/18 perf, core: Add a concept of a weightened sample v2

In reply to Stephane Eranian

From: Andi Kleen <[email protected]>

For some events it's useful to weight sample with a hardware provided number. This expresses how expensive the action the sample represent was. This allows the profiler to scale the samples to be more informative to the programmer.

There is already the period which is used similarly, but it means something different, so I chose to not overload it. Instead a new sample type for WEIGHT is added.

Can be used for multiple things. Initially it is used for TSX abort costs and profiling by memory latencies (so to make expensive load appear higher up in the histograms) The concept is quite generic and can be extended to many other kinds of events or architectures, as long as the hardware provides suitable auxillary values. In principle it could be also used for software tracpoints.

This adds the generic glue. A new optional sample format for a 64bit weight value.

v2: Move weight format to the end. Remove *_FORMAT_WEIGHT

---
 include/linux/perf_event.h      |    2 ++
 include/uapi/linux/perf_event.h |    6 +++++-
 kernel/events/core.c            |    6 ++++++
 3 files changed, 13 insertions(+), 1 deletion(-)
include/linux/perf_event.h +2 -0
index 484cfbc..bb2429d 100644
@@ -584,6 +584,7 @@ struct perf_sample_data {
 	struct perf_branch_stack	*br_stack;
 	struct perf_regs_user		regs_user;
 	u64				stack_user_size;
+	u64				weight;
 };
 
 static inline void perf_sample_data_init(struct perf_sample_data *data,
@@ -597,6 +598,7 @@ static inline void perf_sample_data_init(struct perf_sample_data *data,
 	data->regs_user.abi = PERF_SAMPLE_REGS_ABI_NONE;
 	data->regs_user.regs = NULL;
 	data->stack_user_size = 0;
+	data->weight = 0;
 }
 
 extern void perf_output_sample(struct perf_output_handle *handle,
include/uapi/linux/perf_event.h +5 -1
index 4f63c05..3e6c394 100644
@@ -132,8 +132,10 @@ enum perf_event_sample_format {
 	PERF_SAMPLE_BRANCH_STACK		= 1U << 11,
 	PERF_SAMPLE_REGS_USER			= 1U << 12,
 	PERF_SAMPLE_STACK_USER			= 1U << 13,
+	PERF_SAMPLE_WEIGHT			= 1U << 14,
+
+	PERF_SAMPLE_MAX = 1U << 15,		/* non-ABI */
 
-	PERF_SAMPLE_MAX = 1U << 14,		/* non-ABI */
 };
 
 /*
@@ -587,6 +589,8 @@ enum perf_event_type {
 	 * 	{ u64			size;
 	 * 	  char			data[size];
 	 * 	  u64			dyn_size; } && PERF_SAMPLE_STACK_USER
+	 *
+	 *	{ u64			weight;   } && PERF_SAMPLE_WEIGHT
 	 * };
 	 */
 	PERF_RECORD_SAMPLE			= 9,
kernel/events/core.c +6 -0
index 301079d..749bdf4 100644
@@ -952,6 +952,9 @@ static void perf_event__header_size(struct perf_event *event)
 	if (sample_type & PERF_SAMPLE_PERIOD)
 		size += sizeof(data->period);
 
+	if (sample_type & PERF_SAMPLE_WEIGHT)
+		size += sizeof(data->weight);
+
 	if (sample_type & PERF_SAMPLE_READ)
 		size += event->read_size;
 
@@ -4169,6 +4172,9 @@ void perf_output_sample(struct perf_output_handle *handle,
 		perf_output_sample_ustack(handle,
 					  data->stack_user_size,
 					  data->regs_user.regs);
+
+	if (sample_type & PERF_SAMPLE_WEIGHT)
+		perf_output_put(handle, data->weight);
 }
 
 void perf_prepare_sample(struct perf_event_header *header,
Signature · 1 line
1.7.9.5
#11 lore raw

[tip:perf/x86] perf/core: Add weighted samples

Commit-ID:  648865900cd6cb5b2ed738206b41fb6ede1c8b17
Gitweb:     http://git.kernel.org/tip/648865900cd6cb5b2ed738206b41fb6ede1c8b17
Author:     Andi Kleen <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:28 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 25 Jan 2013 10:19:02 +0100

perf/core: Add weighted samples

For some events it's useful to weight sample with a hardware provided number. This expresses how expensive the action the sample represent was. This allows the profiler to scale the samples to be more informative to the programmer.

There is already the period which is used similarly, but it means something different, so I chose to not overload it. Instead a new sample type for WEIGHT is added.

Can be used for multiple things. Initially it is used for TSX abort costs and profiling by memory latencies (so to make expensive load appear higher up in the histograms). The concept is quite generic and can be extended to many other kinds of events or architectures, as long as the hardware provides suitable auxillary values. In principle it could be also used for software tracepoints.

This adds the generic glue. A new optional sample format for a 64-bit weight value.

---
 include/linux/perf_event.h      | 2 ++
 include/uapi/linux/perf_event.h | 6 +++++-
 kernel/events/core.c            | 6 ++++++
 3 files changed, 13 insertions(+), 1 deletion(-)
include/linux/perf_event.h +2 -0
index 484cfbc..bb2429d 100644
@@ -584,6 +584,7 @@ struct perf_sample_data {
 	struct perf_branch_stack	*br_stack;
 	struct perf_regs_user		regs_user;
 	u64				stack_user_size;
+	u64				weight;
 };
 
 static inline void perf_sample_data_init(struct perf_sample_data *data,
@@ -597,6 +598,7 @@ static inline void perf_sample_data_init(struct perf_sample_data *data,
 	data->regs_user.abi = PERF_SAMPLE_REGS_ABI_NONE;
 	data->regs_user.regs = NULL;
 	data->stack_user_size = 0;
+	data->weight = 0;
 }
 
 extern void perf_output_sample(struct perf_output_handle *handle,
include/uapi/linux/perf_event.h +5 -1
index 4f63c05..3e6c394 100644
@@ -132,8 +132,10 @@ enum perf_event_sample_format {
 	PERF_SAMPLE_BRANCH_STACK		= 1U << 11,
 	PERF_SAMPLE_REGS_USER			= 1U << 12,
 	PERF_SAMPLE_STACK_USER			= 1U << 13,
+	PERF_SAMPLE_WEIGHT			= 1U << 14,
+
+	PERF_SAMPLE_MAX = 1U << 15,		/* non-ABI */
 
-	PERF_SAMPLE_MAX = 1U << 14,		/* non-ABI */
 };
 
 /*
@@ -587,6 +589,8 @@ enum perf_event_type {
 	 * 	{ u64			size;
 	 * 	  char			data[size];
 	 * 	  u64			dyn_size; } && PERF_SAMPLE_STACK_USER
+	 *
+	 *	{ u64			weight;   } && PERF_SAMPLE_WEIGHT
 	 * };
 	 */
 	PERF_RECORD_SAMPLE			= 9,
kernel/events/core.c +6 -0
index 301079d..749bdf4 100644
@@ -952,6 +952,9 @@ static void perf_event__header_size(struct perf_event *event)
 	if (sample_type & PERF_SAMPLE_PERIOD)
 		size += sizeof(data->period);
 
+	if (sample_type & PERF_SAMPLE_WEIGHT)
+		size += sizeof(data->weight);
+
 	if (sample_type & PERF_SAMPLE_READ)
 		size += event->read_size;
 
@@ -4169,6 +4172,9 @@ void perf_output_sample(struct perf_output_handle *handle,
 		perf_output_sample_ustack(handle,
 					  data->stack_user_size,
 					  data->regs_user.regs);
+
+	if (sample_type & PERF_SAMPLE_WEIGHT)
+		perf_output_put(handle, data->weight);
 }
 
 void perf_prepare_sample(struct perf_event_header *header,
#12 lore raw

[tip:perf/core] perf/core: Add weighted samples

In reply to Stephane Eranian
perf/core: Add weighted samples
Commit-ID:  c3feedf2aaf9ac8bad6f19f5d21e4ee0b4b87e9c
Gitweb:     http://git.kernel.org/tip/c3feedf2aaf9ac8bad6f19f5d21e4ee0b4b87e9c
Author:     Andi Kleen <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:28 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:15:44 -0300

perf/core: Add weighted samples

For some events it's useful to weight sample with a hardware provided number. This expresses how expensive the action the sample represent was. This allows the profiler to scale the samples to be more informative to the programmer.

There is already the period which is used similarly, but it means something different, so I chose to not overload it. Instead a new sample type for WEIGHT is added.

Can be used for multiple things. Initially it is used for TSX abort costs and profiling by memory latencies (so to make expensive load appear higher up in the histograms). The concept is quite generic and can be extended to many other kinds of events or architectures, as long as the hardware provides suitable auxillary values. In principle it could be also used for software tracepoints.

This adds the generic glue. A new optional sample format for a 64-bit weight value.

---
 include/linux/perf_event.h      | 2 ++
 include/uapi/linux/perf_event.h | 6 +++++-
 kernel/events/core.c            | 6 ++++++
 3 files changed, 13 insertions(+), 1 deletion(-)
include/linux/perf_event.h +2 -0
index cd3bb2c..7ce0b37 100644
@@ -573,6 +573,7 @@ struct perf_sample_data {
 	struct perf_branch_stack	*br_stack;
 	struct perf_regs_user		regs_user;
 	u64				stack_user_size;
+	u64				weight;
 };
 
 static inline void perf_sample_data_init(struct perf_sample_data *data,
@@ -586,6 +587,7 @@ static inline void perf_sample_data_init(struct perf_sample_data *data,
 	data->regs_user.abi = PERF_SAMPLE_REGS_ABI_NONE;
 	data->regs_user.regs = NULL;
 	data->stack_user_size = 0;
+	data->weight = 0;
 }
 
 extern void perf_output_sample(struct perf_output_handle *handle,
include/uapi/linux/perf_event.h +5 -1
index 9fa9c62..cdc255d 100644
@@ -132,8 +132,10 @@ enum perf_event_sample_format {
 	PERF_SAMPLE_BRANCH_STACK		= 1U << 11,
 	PERF_SAMPLE_REGS_USER			= 1U << 12,
 	PERF_SAMPLE_STACK_USER			= 1U << 13,
+	PERF_SAMPLE_WEIGHT			= 1U << 14,
+
+	PERF_SAMPLE_MAX = 1U << 15,		/* non-ABI */
 
-	PERF_SAMPLE_MAX = 1U << 14,		/* non-ABI */
 };
 
 /*
@@ -588,6 +590,8 @@ enum perf_event_type {
 	 * 	{ u64			size;
 	 * 	  char			data[size];
 	 * 	  u64			dyn_size; } && PERF_SAMPLE_STACK_USER
+	 *
+	 *	{ u64			weight;   } && PERF_SAMPLE_WEIGHT
 	 * };
 	 */
 	PERF_RECORD_SAMPLE			= 9,
kernel/events/core.c +6 -0
index 7b4a55d..9e3edb2 100644
@@ -976,6 +976,9 @@ static void perf_event__header_size(struct perf_event *event)
 	if (sample_type & PERF_SAMPLE_PERIOD)
 		size += sizeof(data->period);
 
+	if (sample_type & PERF_SAMPLE_WEIGHT)
+		size += sizeof(data->weight);
+
 	if (sample_type & PERF_SAMPLE_READ)
 		size += event->read_size;
 
@@ -4193,6 +4196,9 @@ void perf_output_sample(struct perf_output_handle *handle,
 		perf_output_sample_ustack(handle,
 					  data->stack_user_size,
 					  data->regs_user.regs);
+
+	if (sample_type & PERF_SAMPLE_WEIGHT)
+		perf_output_put(handle, data->weight);
 }
 
 void perf_prepare_sample(struct perf_event_header *header,
#13 lore raw

PATCH v7 5/18 perf, tools: Add support for weight v7 (modified)

In reply to Stephane Eranian

From: Andi Kleen <[email protected]>

perf record has a new option -W that enables weightened sampling.

Add sorting support in top/report for the average weight per sample and the total weight sum. This allows to both compare relative cost per event and the total cost over the measurement period.

Add the necessary glue to perf report, record and the library.

v2: Merge with new hist refactoring.
v3: Fix manpage. Remove value check.
Rename global_weight to weight and weight to local_weight.
v4: Readd sort keys to manpage
v5: Move weight to end
v6: Move weight to template
v7: Rename weight key.

Original patch from Andi modified by Stephane Eranian <[email protected]> to include ONLY the weight supporting code and apply to pristine 3.8.0-rc4.

---
 tools/perf/Documentation/perf-record.txt |    6 ++++
 tools/perf/Documentation/perf-report.txt |    2 +-
 tools/perf/Documentation/perf-top.txt    |    2 +-
 tools/perf/builtin-annotate.c            |    2 +-
 tools/perf/builtin-diff.c                |    7 +++--
 tools/perf/builtin-record.c              |    2 ++
 tools/perf/builtin-report.c              |    7 +++--
 tools/perf/builtin-top.c                 |    5 +--
 tools/perf/perf.h                        |    1 +
 tools/perf/util/event.h                  |    1 +
 tools/perf/util/evsel.c                  |   10 ++++++
 tools/perf/util/hist.c                   |   23 +++++++++-----
 tools/perf/util/hist.h                   |    8 +++--
 tools/perf/util/session.c                |    3 ++
 tools/perf/util/sort.c                   |   49 ++++++++++++++++++++++++++++++
 tools/perf/util/sort.h                   |    3 ++
 16 files changed, 111 insertions(+), 20 deletions(-)
tools/perf/Documentation/perf-record.txt +6 -0
index 938e890..d4da111 100644
@@ -182,6 +182,12 @@ is enabled for all the sampling events. The sampled branch type is the same for
 The various filters must be specified as a comma separated list: --branch-filter any_ret,u,k
 Note that this feature may not be available on all processors.
 
+-W::
+--weight::
+Enable weightened sampling. An additional weight is recorded per sample and can be
+displayed with the weight and local_weight sort keys.  This currently works for TSX
+abort events and some memory events in precise mode on modern Intel CPUs.
+
 SEE ALSO
 --------
 linkperf:perf-stat[1], linkperf:perf-list[1]
tools/perf/Documentation/perf-report.txt +1 -1
index f4d91be..88ebd50 100644
@@ -57,7 +57,7 @@ OPTIONS
 
 -s::
 --sort=::
-	Sort by key(s): pid, comm, dso, symbol, parent, srcline.
+	Sort by key(s): pid, comm, dso, symbol, parent, srcline, weight, local_weight.
 
 -p::
 --parent=<regex>::
tools/perf/Documentation/perf-top.txt +1 -1
index 5b80d84..48a4089 100644
@@ -112,7 +112,7 @@ Default is to monitor all CPUS.
 
 -s::
 --sort::
-	Sort by key(s): pid, comm, dso, symbol, parent, srcline.
+	Sort by key(s): pid, comm, dso, symbol, parent, srcline, weight, local_weight.
 
 -n::
 --show-nr-samples::
tools/perf/builtin-annotate.c +1 -1
index dc870cf..1bacb7d 100644
@@ -62,7 +62,7 @@ static int perf_evsel__add_sample(struct perf_evsel *evsel,
 		return 0;
 	}
 
-	he = __hists__add_entry(&evsel->hists, al, NULL, 1);
+	he = __hists__add_entry(&evsel->hists, al, NULL, 1, 1);
 	if (he == NULL)
 		return -ENOMEM;
 
tools/perf/builtin-diff.c +4 -3
index 93b852f..03a322f 100644
@@ -248,9 +248,10 @@ int perf_diff__formula(char *buf, size_t size, struct hist_entry *he)
 }
 
 static int hists__add_entry(struct hists *self,
-			    struct addr_location *al, u64 period)
+			    struct addr_location *al, u64 period,
+			    u64 weight)
 {
-	if (__hists__add_entry(self, al, NULL, period) != NULL)
+	if (__hists__add_entry(self, al, NULL, period, weight) != NULL)
 		return 0;
 	return -ENOMEM;
 }
@@ -272,7 +273,7 @@ static int diff__process_sample_event(struct perf_tool *tool __maybe_unused,
 	if (al.filtered)
 		return 0;
 
-	if (hists__add_entry(&evsel->hists, &al, sample->period)) {
+	if (hists__add_entry(&evsel->hists, &al, sample->period, sample->weight)) {
 		pr_warning("problem incrementing symbol period, skipping event\n");
 		return -1;
 	}
tools/perf/builtin-record.c +2 -0
index f3151d3..8cdadd2 100644
@@ -1059,6 +1059,8 @@ const struct option record_options[] = {
 	OPT_CALLBACK('j', "branch-filter", &record.opts.branch_stack,
 		     "branch filter mask", "branch stack filter modes",
 		     parse_branch_stack),
+	OPT_BOOLEAN('W', "weight", &record.opts.sample_weight,
+		    "sample by weight (on special events only)"),
 	OPT_END()
 };
 
tools/perf/builtin-report.c +4 -3
index fc25100..3cfe259 100644
@@ -88,7 +88,7 @@ static int perf_report__add_branch_hist_entry(struct perf_tool *tool,
 		 * and not events sampled. Thus we use a pseudo period of 1.
 		 */
 		he = __hists__add_branch_entry(&evsel->hists, al, parent,
-				&bi[i], 1);
+				&bi[i], 1, 1);
 		if (he) {
 			struct annotation *notes;
 			err = -ENOMEM;
@@ -146,7 +146,8 @@ static int perf_evsel__add_hist_entry(struct perf_evsel *evsel,
 			return err;
 	}
 
-	he = __hists__add_entry(&evsel->hists, al, parent, sample->period);
+	he = __hists__add_entry(&evsel->hists, al, parent, sample->period,
+					sample->weight);
 	if (he == NULL)
 		return -ENOMEM;
 
@@ -596,7 +597,7 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 		    "Use the stdio interface"),
 	OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
 		   "sort by key(s): pid, comm, dso, symbol, parent, dso_to,"
-		   " dso_from, symbol_to, symbol_from, mispredict"),
+		   " dso_from, symbol_to, symbol_from, mispredict, weight, local_weight"),
 	OPT_BOOLEAN(0, "showcpuutilization", &symbol_conf.show_cpu_utilization,
 		    "Show sample percentage for different cpu modes"),
 	OPT_STRING('p', "parent", &parent_pattern, "regex",
tools/perf/builtin-top.c +3 -2
index c9ff395..df00fe2 100644
@@ -271,7 +271,8 @@ static struct hist_entry *perf_evsel__add_hist_entry(struct perf_evsel *evsel,
 {
 	struct hist_entry *he;
 
-	he = __hists__add_entry(&evsel->hists, al, NULL, sample->period);
+	he = __hists__add_entry(&evsel->hists, al, NULL, sample->period,
+				sample->weight);
 	if (he == NULL)
 		return NULL;
 
@@ -1230,7 +1231,7 @@ int cmd_top(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_INCR('v', "verbose", &verbose,
 		    "be more verbose (show counter open errors, etc)"),
 	OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
-		   "sort by key(s): pid, comm, dso, symbol, parent"),
+		   "sort by key(s): pid, comm, dso, symbol, parent, weight, local_weight"),
 	OPT_BOOLEAN('n', "show-nr-samples", &symbol_conf.show_nr_samples,
 		    "Show a column with the number of samples"),
 	OPT_CALLBACK_DEFAULT('G', "call-graph", &top, "output_type,min_percent, call_order",
tools/perf/perf.h +1 -0
index 2c340e7..827a9ea 100644
@@ -236,6 +236,7 @@ struct perf_record_opts {
 	bool	     pipe_output;
 	bool	     raw_samples;
 	bool	     sample_address;
+	bool	     sample_weight;
 	bool	     sample_time;
 	bool	     sample_id_all_missing;
 	bool	     exclude_guest_missing;
tools/perf/util/event.h +1 -0
index 0d573ff..a97fbbe 100644
@@ -88,6 +88,7 @@ struct perf_sample {
 	u64 id;
 	u64 stream_id;
 	u64 period;
+	u64 weight;
 	u32 cpu;
 	u32 raw_size;
 	void *raw_data;
tools/perf/util/evsel.c +10 -0
index 1b16dd1..805d33e 100644
@@ -510,6 +510,9 @@ void perf_evsel__config(struct perf_evsel *evsel,
 		attr->branch_sample_type = opts->branch_stack;
 	}
 
+	if (opts->sample_weight)
+		attr->sample_type	|= PERF_SAMPLE_WEIGHT;
+
 	attr->mmap = track;
 	attr->comm = track;
 
@@ -908,6 +911,7 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
 	data->cpu = data->pid = data->tid = -1;
 	data->stream_id = data->id = data->time = -1ULL;
 	data->period = 1;
+	data->weight = 0;
 
 	if (event->header.type != PERF_RECORD_SAMPLE) {
 		if (!evsel->attr.sample_id_all)
@@ -1058,6 +1062,12 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
 		}
 	}
 
+	data->weight = 0;
+	if (type & PERF_SAMPLE_WEIGHT) {
+		data->weight = *array;
+		array++;
+	}
+
 	return 0;
 }
 
tools/perf/util/hist.c +16 -7
index cb17e2a..a8d7647 100644
@@ -151,9 +151,11 @@ static void hist_entry__add_cpumode_period(struct hist_entry *he,
 	}
 }
 
-static void he_stat__add_period(struct he_stat *he_stat, u64 period)
+static void he_stat__add_period(struct he_stat *he_stat, u64 period,
+				u64 weight)
 {
 	he_stat->period		+= period;
+	he_stat->weight		+= weight;
 	he_stat->nr_events	+= 1;
 }
 
@@ -165,12 +167,14 @@ static void he_stat__add_stat(struct he_stat *dest, struct he_stat *src)
 	dest->period_guest_sys	+= src->period_guest_sys;
 	dest->period_guest_us	+= src->period_guest_us;
 	dest->nr_events		+= src->nr_events;
+	dest->weight		+= src->weight;
 }
 
 static void hist_entry__decay(struct hist_entry *he)
 {
 	he->stat.period = (he->stat.period * 7) / 8;
 	he->stat.nr_events = (he->stat.nr_events * 7) / 8;
+	/* XXX need decay for weight too? */
 }
 
 static bool hists__decay_entry(struct hists *hists, struct hist_entry *he)
@@ -270,7 +274,8 @@ static u8 symbol__parent_filter(const struct symbol *parent)
 static struct hist_entry *add_hist_entry(struct hists *hists,
 				      struct hist_entry *entry,
 				      struct addr_location *al,
-				      u64 period)
+				      u64 period,
+				      u64 weight)
 {
 	struct rb_node **p;
 	struct rb_node *parent = NULL;
@@ -288,7 +293,7 @@ static struct hist_entry *add_hist_entry(struct hists *hists,
 		cmp = hist_entry__cmp(entry, he);
 
 		if (!cmp) {
-			he_stat__add_period(&he->stat, period);
+			he_stat__add_period(&he->stat, period, weight);
 
 			/* If the map of an existing hist_entry has
 			 * become out-of-date due to an exec() or
@@ -327,7 +332,8 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 					     struct addr_location *al,
 					     struct symbol *sym_parent,
 					     struct branch_info *bi,
-					     u64 period)
+					     u64 period,
+					     u64 weight)
 {
 	struct hist_entry entry = {
 		.thread	= al->thread,
@@ -341,6 +347,7 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 		.stat = {
 			.period	= period,
 			.nr_events = 1,
+			.weight = weight,
 		},
 		.parent = sym_parent,
 		.filtered = symbol__parent_filter(sym_parent),
@@ -348,12 +355,13 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 		.hists	= self,
 	};
 
-	return add_hist_entry(self, &entry, al, period);
+	return add_hist_entry(self, &entry, al, period, weight);
 }
 
 struct hist_entry *__hists__add_entry(struct hists *self,
 				      struct addr_location *al,
-				      struct symbol *sym_parent, u64 period)
+				      struct symbol *sym_parent, u64 period,
+				      u64 weight)
 {
 	struct hist_entry entry = {
 		.thread	= al->thread,
@@ -367,13 +375,14 @@ struct hist_entry *__hists__add_entry(struct hists *self,
 		.stat = {
 			.period	= period,
 			.nr_events = 1,
+			.weight = weight,
 		},
 		.parent = sym_parent,
 		.filtered = symbol__parent_filter(sym_parent),
 		.hists	= self,
 	};
 
-	return add_hist_entry(self, &entry, al, period);
+	return add_hist_entry(self, &entry, al, period, weight);
 }
 
 int64_t
tools/perf/util/hist.h +6 -2
index 8b091a5..c769984a 100644
@@ -49,6 +49,8 @@ enum hist_column {
 	HISTC_DSO_FROM,
 	HISTC_DSO_TO,
 	HISTC_SRCLINE,
+	HISTC_LOCAL_WEIGHT,
+	HISTC_GLOBAL_WEIGHT,
 	HISTC_NR_COLS, /* Last entry */
 };
 
@@ -73,7 +75,8 @@ struct hists {
 
 struct hist_entry *__hists__add_entry(struct hists *self,
 				      struct addr_location *al,
-				      struct symbol *parent, u64 period);
+				      struct symbol *parent, u64 period,
+				      u64 weight);
 int64_t hist_entry__cmp(struct hist_entry *left, struct hist_entry *right);
 int64_t hist_entry__collapse(struct hist_entry *left, struct hist_entry *right);
 int hist_entry__sort_snprintf(struct hist_entry *self, char *bf, size_t size,
@@ -84,7 +87,8 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 					     struct addr_location *al,
 					     struct symbol *sym_parent,
 					     struct branch_info *bi,
-					     u64 period);
+					     u64 period,
+					     u64 weight);
 
 void hists__output_resort(struct hists *self);
 void hists__output_resort_threaded(struct hists *hists);
tools/perf/util/session.c +3 -0
index ce6f511..3de9097 100644
@@ -1006,6 +1006,9 @@ static void dump_sample(struct perf_evsel *evsel, union perf_event *event,
 
 	if (sample_type & PERF_SAMPLE_STACK_USER)
 		stack_user__printf(&sample->user_stack);
+
+	if (sample_type & PERF_SAMPLE_WEIGHT)
+		printf("... weight: %" PRIu64 "\n", sample->weight);
 }
 
 static struct machine *
tools/perf/util/sort.c +49 -0

67 lines not on this page · raw message

tools/perf/util/sort.h +3 -0
index b4e8c3b..9af5446 100644
@@ -49,6 +49,7 @@ struct he_stat {
 	u64			period_us;
 	u64			period_guest_sys;
 	u64			period_guest_us;
+	u64			weight;
 	u32			nr_events;
 };
 
@@ -137,6 +138,8 @@ enum sort_type {
 	SORT_SYM_TO,
 	SORT_MISPREDICT,
 	SORT_SRCLINE,
+	SORT_LOCAL_WEIGHT,
+	SORT_GLOBAL_WEIGHT,
 };
 
 /*
Signature · 1 line
1.7.9.5
#14 lore raw

[tip:perf/core] perf tools: Add support for weight v7 (modified)

perf tools: Add support for weight v7 (modified)
Commit-ID:  05484298cbfebbf8c8c55b000541a245bc286bec
Gitweb:     http://git.kernel.org/tip/05484298cbfebbf8c8c55b000541a245bc286bec
Author:     Andi Kleen <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:29 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:19:43 -0300

perf tools: Add support for weight v7 (modified)

perf record has a new option -W that enables weightened sampling.

Add sorting support in top/report for the average weight per sample and the total weight sum. This allows to both compare relative cost per event and the total cost over the measurement period.

Add the necessary glue to perf report, record and the library.

v2: Merge with new hist refactoring.
v3: Fix manpage. Remove value check.
Rename global_weight to weight and weight to local_weight.
v4: Readd sort keys to manpage
v5: Move weight to end
v6: Move weight to template
v7: Rename weight key.

Original patch from Andi modified by Stephane Eranian <[email protected]> to include ONLY the weight supporting code and apply to pristine 3.8.0-rc4.

Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Peter Zijlstra <[email protected]>
[ committer note: changed to cope with fc5871ed and the hists_link perf test entry ]
---
 tools/perf/Documentation/perf-record.txt |  6 +++++
 tools/perf/Documentation/perf-report.txt |  2 +-
 tools/perf/Documentation/perf-top.txt    |  2 +-
 tools/perf/builtin-annotate.c            |  2 +-
 tools/perf/builtin-diff.c                |  7 ++---
 tools/perf/builtin-record.c              |  2 ++
 tools/perf/builtin-report.c              |  8 +++---
 tools/perf/builtin-top.c                 |  5 ++--
 tools/perf/perf.h                        |  1 +
 tools/perf/tests/hists_link.c            |  4 +--
 tools/perf/util/event.h                  |  1 +
 tools/perf/util/evsel.c                  | 10 +++++++
 tools/perf/util/hist.c                   | 23 +++++++++++-----
 tools/perf/util/hist.h                   |  8 ++++--
 tools/perf/util/session.c                |  3 +++
 tools/perf/util/sort.c                   | 45 ++++++++++++++++++++++++++++++++
 tools/perf/util/sort.h                   |  3 +++
 17 files changed, 110 insertions(+), 22 deletions(-)
tools/perf/Documentation/perf-record.txt +6 -0
index 938e890..d4da111 100644
@@ -182,6 +182,12 @@ is enabled for all the sampling events. The sampled branch type is the same for
 The various filters must be specified as a comma separated list: --branch-filter any_ret,u,k
 Note that this feature may not be available on all processors.
 
+-W::
+--weight::
+Enable weightened sampling. An additional weight is recorded per sample and can be
+displayed with the weight and local_weight sort keys.  This currently works for TSX
+abort events and some memory events in precise mode on modern Intel CPUs.
+
 SEE ALSO
 --------
 linkperf:perf-stat[1], linkperf:perf-list[1]
tools/perf/Documentation/perf-report.txt +1 -1
index 71f1551..7d5f4f3 100644
@@ -59,7 +59,7 @@ OPTIONS
 --sort=::
 	Sort histogram entries by given key(s) - multiple keys can be specified
 	in CSV format.  Following sort keys are available:
-	pid, comm, dso, symbol, parent, cpu, srcline.
+	pid, comm, dso, symbol, parent, cpu, srcline, weight, local_weight.
 
 	Each key has following meaning:
 
tools/perf/Documentation/perf-top.txt +1 -1
index a414bc9..9f1a2fe 100644
@@ -112,7 +112,7 @@ Default is to monitor all CPUS.
 
 -s::
 --sort::
-	Sort by key(s): pid, comm, dso, symbol, parent, srcline.
+	Sort by key(s): pid, comm, dso, symbol, parent, srcline, weight, local_weight.
 
 -n::
 --show-nr-samples::
tools/perf/builtin-annotate.c +1 -1
index ae36f3c..db491e9 100644
@@ -63,7 +63,7 @@ static int perf_evsel__add_sample(struct perf_evsel *evsel,
 		return 0;
 	}
 
-	he = __hists__add_entry(&evsel->hists, al, NULL, 1);
+	he = __hists__add_entry(&evsel->hists, al, NULL, 1, 1);
 	if (he == NULL)
 		return -ENOMEM;
 
tools/perf/builtin-diff.c +4 -3
index d207a97..2d0462d 100644
@@ -231,9 +231,10 @@ int perf_diff__formula(struct hist_entry *he, struct hist_entry *pair,
 }
 
 static int hists__add_entry(struct hists *self,
-			    struct addr_location *al, u64 period)
+			    struct addr_location *al, u64 period,
+			    u64 weight)
 {
-	if (__hists__add_entry(self, al, NULL, period) != NULL)
+	if (__hists__add_entry(self, al, NULL, period, weight) != NULL)
 		return 0;
 	return -ENOMEM;
 }
@@ -255,7 +256,7 @@ static int diff__process_sample_event(struct perf_tool *tool __maybe_unused,
 	if (al.filtered)
 		return 0;
 
-	if (hists__add_entry(&evsel->hists, &al, sample->period)) {
+	if (hists__add_entry(&evsel->hists, &al, sample->period, sample->weight)) {
 		pr_warning("problem incrementing symbol period, skipping event\n");
 		return -1;
 	}
tools/perf/builtin-record.c +2 -0
index 78a41fd..cdf58ec 100644
@@ -953,6 +953,8 @@ const struct option record_options[] = {
 	OPT_CALLBACK('j', "branch-filter", &record.opts.branch_stack,
 		     "branch filter mask", "branch stack filter modes",
 		     parse_branch_stack),
+	OPT_BOOLEAN('W', "weight", &record.opts.sample_weight,
+		    "sample by weight (on special events only)"),
 	OPT_END()
 };
 
tools/perf/builtin-report.c +5 -3
index b5ea26c..e31f070 100644
@@ -98,7 +98,7 @@ static int perf_report__add_branch_hist_entry(struct perf_tool *tool,
 		 * and not events sampled. Thus we use a pseudo period of 1.
 		 */
 		he = __hists__add_branch_entry(&evsel->hists, al, parent,
-				&bi[i], 1);
+				&bi[i], 1, 1);
 		if (he) {
 			struct annotation *notes;
 			err = -ENOMEM;
@@ -156,7 +156,8 @@ static int perf_evsel__add_hist_entry(struct perf_evsel *evsel,
 			return err;
 	}
 
-	he = __hists__add_entry(&evsel->hists, al, parent, sample->period);
+	he = __hists__add_entry(&evsel->hists, al, parent, sample->period,
+					sample->weight);
 	if (he == NULL)
 		return -ENOMEM;
 
@@ -644,7 +645,8 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 		    "Use the stdio interface"),
 	OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
 		   "sort by key(s): pid, comm, dso, symbol, parent, cpu, srcline,"
-		   " dso_to, dso_from, symbol_to, symbol_from, mispredict"),
+		   " dso_to, dso_from, symbol_to, symbol_from, mispredict,"
+		   " weight, local_weight"),
 	OPT_BOOLEAN(0, "showcpuutilization", &symbol_conf.show_cpu_utilization,
 		    "Show sample percentage for different cpu modes"),
 	OPT_STRING('p', "parent", &parent_pattern, "regex",
tools/perf/builtin-top.c +3 -2
index b5520ad..67bdb9f 100644
@@ -251,7 +251,8 @@ static struct hist_entry *perf_evsel__add_hist_entry(struct perf_evsel *evsel,
 {
 	struct hist_entry *he;
 
-	he = __hists__add_entry(&evsel->hists, al, NULL, sample->period);
+	he = __hists__add_entry(&evsel->hists, al, NULL, sample->period,
+				sample->weight);
 	if (he == NULL)
 		return NULL;
 
@@ -1088,7 +1089,7 @@ int cmd_top(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_INCR('v', "verbose", &verbose,
 		    "be more verbose (show counter open errors, etc)"),
 	OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
-		   "sort by key(s): pid, comm, dso, symbol, parent"),
+		   "sort by key(s): pid, comm, dso, symbol, parent, weight, local_weight"),
 	OPT_BOOLEAN('n', "show-nr-samples", &symbol_conf.show_nr_samples,
 		    "Show a column with the number of samples"),
 	OPT_CALLBACK_DEFAULT('G', "call-graph", &top.record_opts,
tools/perf/perf.h +1 -0
index 74659ec..32bd102 100644
@@ -218,6 +218,7 @@ struct perf_record_opts {
 	bool	     pipe_output;
 	bool	     raw_samples;
 	bool	     sample_address;
+	bool	     sample_weight;
 	bool	     sample_time;
 	bool	     period;
 	unsigned int freq;
tools/perf/tests/hists_link.c +2 -2
index e0c0267..89085a9 100644
@@ -223,7 +223,7 @@ static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
 							  &sample, 0) < 0)
 				goto out;
 
-			he = __hists__add_entry(&evsel->hists, &al, NULL, 1);
+			he = __hists__add_entry(&evsel->hists, &al, NULL, 1, 1);
 			if (he == NULL)
 				goto out;
 
@@ -247,7 +247,7 @@ static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
 							  &sample, 0) < 0)
 				goto out;
 
-			he = __hists__add_entry(&evsel->hists, &al, NULL, 1);
+			he = __hists__add_entry(&evsel->hists, &al, NULL, 1, 1);
 			if (he == NULL)
 				goto out;
 
tools/perf/util/event.h +1 -0
index 0d573ff..a97fbbe 100644
@@ -88,6 +88,7 @@ struct perf_sample {
 	u64 id;
 	u64 stream_id;
 	u64 period;
+	u64 weight;
 	u32 cpu;
 	u32 raw_size;
 	void *raw_data;
tools/perf/util/evsel.c +10 -0
index 1adb824..23061a6 100644
@@ -563,6 +563,9 @@ void perf_evsel__config(struct perf_evsel *evsel,
 		attr->branch_sample_type = opts->branch_stack;
 	}
 
+	if (opts->sample_weight)
+		attr->sample_type	|= PERF_SAMPLE_WEIGHT;
+
 	attr->mmap = track;
 	attr->comm = track;
 
@@ -1017,6 +1020,7 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
 	data->cpu = data->pid = data->tid = -1;
 	data->stream_id = data->id = data->time = -1ULL;
 	data->period = 1;
+	data->weight = 0;
 
 	if (event->header.type != PERF_RECORD_SAMPLE) {
 		if (!evsel->attr.sample_id_all)
@@ -1167,6 +1171,12 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
 		}
 	}
 
+	data->weight = 0;
+	if (type & PERF_SAMPLE_WEIGHT) {
+		data->weight = *array;
+		array++;
+	}
+
 	return 0;
 }
 
tools/perf/util/hist.c +16 -7
index f855941..97ddd18 100644
@@ -155,9 +155,11 @@ static void hist_entry__add_cpumode_period(struct hist_entry *he,
 	}
 }
 
-static void he_stat__add_period(struct he_stat *he_stat, u64 period)
+static void he_stat__add_period(struct he_stat *he_stat, u64 period,
+				u64 weight)
 {
 	he_stat->period		+= period;
+	he_stat->weight		+= weight;
 	he_stat->nr_events	+= 1;
 }
 
@@ -169,12 +171,14 @@ static void he_stat__add_stat(struct he_stat *dest, struct he_stat *src)
 	dest->period_guest_sys	+= src->period_guest_sys;
 	dest->period_guest_us	+= src->period_guest_us;
 	dest->nr_events		+= src->nr_events;
+	dest->weight		+= src->weight;
 }
 
 static void hist_entry__decay(struct hist_entry *he)
 {
 	he->stat.period = (he->stat.period * 7) / 8;
 	he->stat.nr_events = (he->stat.nr_events * 7) / 8;
+	/* XXX need decay for weight too? */
 }
 
 static bool hists__decay_entry(struct hists *hists, struct hist_entry *he)
@@ -282,7 +286,8 @@ static u8 symbol__parent_filter(const struct symbol *parent)
 static struct hist_entry *add_hist_entry(struct hists *hists,
 				      struct hist_entry *entry,
 				      struct addr_location *al,
-				      u64 period)
+				      u64 period,
+				      u64 weight)
 {
 	struct rb_node **p;
 	struct rb_node *parent = NULL;
@@ -306,7 +311,7 @@ static struct hist_entry *add_hist_entry(struct hists *hists,
 		cmp = hist_entry__cmp(he, entry);
 
 		if (!cmp) {
-			he_stat__add_period(&he->stat, period);
+			he_stat__add_period(&he->stat, period, weight);
 
 			/* If the map of an existing hist_entry has
 			 * become out-of-date due to an exec() or
@@ -345,7 +350,8 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 					     struct addr_location *al,
 					     struct symbol *sym_parent,
 					     struct branch_info *bi,
-					     u64 period)
+					     u64 period,
+					     u64 weight)
 {
 	struct hist_entry entry = {
 		.thread	= al->thread,
@@ -359,6 +365,7 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 		.stat = {
 			.period	= period,
 			.nr_events = 1,
+			.weight = weight,
 		},
 		.parent = sym_parent,
 		.filtered = symbol__parent_filter(sym_parent),
@@ -366,12 +373,13 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 		.hists	= self,
 	};
 
-	return add_hist_entry(self, &entry, al, period);
+	return add_hist_entry(self, &entry, al, period, weight);
 }
 
 struct hist_entry *__hists__add_entry(struct hists *self,
 				      struct addr_location *al,
-				      struct symbol *sym_parent, u64 period)
+				      struct symbol *sym_parent, u64 period,
+				      u64 weight)
 {
 	struct hist_entry entry = {
 		.thread	= al->thread,
@@ -385,13 +393,14 @@ struct hist_entry *__hists__add_entry(struct hists *self,
 		.stat = {
 			.period	= period,
 			.nr_events = 1,
+			.weight = weight,
 		},
 		.parent = sym_parent,
 		.filtered = symbol__parent_filter(sym_parent),
 		.hists	= self,
 	};
 
-	return add_hist_entry(self, &entry, al, period);
+	return add_hist_entry(self, &entry, al, period, weight);
 }
 
 int64_t
tools/perf/util/hist.h +6 -2
index 8483313..121cc14 100644
@@ -49,6 +49,8 @@ enum hist_column {
 	HISTC_DSO_FROM,
 	HISTC_DSO_TO,
 	HISTC_SRCLINE,
+	HISTC_LOCAL_WEIGHT,
+	HISTC_GLOBAL_WEIGHT,
 	HISTC_NR_COLS, /* Last entry */
 };
 
@@ -73,7 +75,8 @@ struct hists {
 
 struct hist_entry *__hists__add_entry(struct hists *self,
 				      struct addr_location *al,
-				      struct symbol *parent, u64 period);
+				      struct symbol *parent, u64 period,
+				      u64 weight);
 int64_t hist_entry__cmp(struct hist_entry *left, struct hist_entry *right);
 int64_t hist_entry__collapse(struct hist_entry *left, struct hist_entry *right);
 int hist_entry__sort_snprintf(struct hist_entry *self, char *bf, size_t size,
@@ -84,7 +87,8 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 					     struct addr_location *al,
 					     struct symbol *sym_parent,
 					     struct branch_info *bi,
-					     u64 period);
+					     u64 period,
+					     u64 weight);
 
 void hists__output_resort(struct hists *self);
 void hists__output_resort_threaded(struct hists *hists);
tools/perf/util/session.c +3 -0
index c8ba120..627be09 100644
@@ -798,6 +798,9 @@ static void dump_sample(struct perf_evsel *evsel, union perf_event *event,
 
 	if (sample_type & PERF_SAMPLE_STACK_USER)
 		stack_user__printf(&sample->user_stack);
+
+	if (sample_type & PERF_SAMPLE_WEIGHT)
+		printf("... weight: %" PRIu64 "\n", sample->weight);
 }
 
 static struct machine *
tools/perf/util/sort.c +45 -0

57 lines not on this page · raw message

tools/perf/util/sort.h +3 -0
index b13e56f6..3939250 100644
@@ -49,6 +49,7 @@ struct he_stat {
 	u64			period_us;
 	u64			period_guest_sys;
 	u64			period_guest_us;
+	u64			weight;
 	u32			nr_events;
 };
 
@@ -130,6 +131,8 @@ enum sort_type {
 	SORT_PARENT,
 	SORT_CPU,
 	SORT_SRCLINE,
+	SORT_LOCAL_WEIGHT,
+	SORT_GLOBAL_WEIGHT,
 
 	/* branch stack specific sort keys */
 	__SORT_BRANCH_STACK,
#15 lore raw

PATCH v7 6/18 perf: add support for PERF_SAMPLE_ADDR in dump_sampple()

In reply to Stephane Eranian

Was missing from current code yet PERF_SAMPLE_ADDR has been present for a long time. Needed for PEBS-LL mode.

---
 tools/perf/util/session.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
tools/perf/util/session.c +4 -1
index 3de9097..8fe3688 100644
@@ -1008,7 +1008,10 @@ static void dump_sample(struct perf_evsel *evsel, union perf_event *event,
 		stack_user__printf(&sample->user_stack);
 
 	if (sample_type & PERF_SAMPLE_WEIGHT)
-		printf("... weight: %" PRIu64 "\n", sample->weight);
+		printf(" ... weight: %" PRIu64 "\n", sample->weight);
+
+	if (sample_type & PERF_SAMPLE_ADDR)
+		printf(" ..... data: 0x%"PRIx64"\n", sample->addr);
 }
 
 static struct machine *
Signature · 1 line
1.7.9.5
#16 lore raw

PATCH v7 7/18 perf: add generic memory sampling interface

In reply to Stephane Eranian

This patch adds PERF_SAMPLE_DSRC.

PERF_SAMPLE_DSRC collects the data source, i.e., where did the data associated with the sampled instruction come from. Information is stored in a perf_mem_dsrc structure. It contains opcode, mem level, tlb, snoop, lock information, subject to availability in hardware.

---
 include/linux/perf_event.h      |    2 ++
 include/uapi/linux/perf_event.h |   68 +++++++++++++++++++++++++++++++++++++--
 kernel/events/core.c            |    6 ++++
 3 files changed, 74 insertions(+), 2 deletions(-)
include/linux/perf_event.h +2 -0
index bb2429d..8fe4610 100644
@@ -579,6 +579,7 @@ struct perf_sample_data {
 		u32	reserved;
 	}				cpu_entry;
 	u64				period;
+	union  perf_mem_dsrc		dsrc;
 	struct perf_callchain_entry	*callchain;
 	struct perf_raw_record		*raw;
 	struct perf_branch_stack	*br_stack;
@@ -599,6 +600,7 @@ static inline void perf_sample_data_init(struct perf_sample_data *data,
 	data->regs_user.regs = NULL;
 	data->stack_user_size = 0;
 	data->weight = 0;
+	data->dsrc.val = 0;
 }
 
 extern void perf_output_sample(struct perf_output_handle *handle,
include/uapi/linux/perf_event.h +66 -2

85 lines not on this page · raw message

kernel/events/core.c +6 -0
index 749bdf4..56ca60b 100644
@@ -958,6 +958,9 @@ static void perf_event__header_size(struct perf_event *event)
 	if (sample_type & PERF_SAMPLE_READ)
 		size += event->read_size;
 
+	if (sample_type & PERF_SAMPLE_DSRC)
+		size += sizeof(data->dsrc.val);
+
 	event->header_size = size;
 }
 
@@ -4175,6 +4178,9 @@ void perf_output_sample(struct perf_output_handle *handle,
 
 	if (sample_type & PERF_SAMPLE_WEIGHT)
 		perf_output_put(handle, data->weight);
+
+	if (sample_type & PERF_SAMPLE_DSRC)
+		perf_output_put(handle, data->dsrc.val);
 }
 
 void perf_prepare_sample(struct perf_event_header *header,
Signature · 1 line
1.7.9.5
#17 lore raw

Re: [PATCH v7 07/18] perf: add generic memory sampling interface

Quoted 116 lines · Stephane Eranian

This patch adds PERF_SAMPLE_DSRC.

PERF_SAMPLE_DSRC collects the data source, i.e., where did the data associated with the sampled instruction come from. Information is stored in a perf_mem_dsrc structure. It contains opcode, mem level, tlb, snoop, lock information, subject to availability in hardware.

---
 include/linux/perf_event.h      |    2 ++
 include/uapi/linux/perf_event.h |   68 +++++++++++++++++++++++++++++++++++++--
 kernel/events/core.c            |    6 ++++
 3 files changed, 74 insertions(+), 2 deletions(-)
include/linux/perf_event.h +2 -0
index bb2429d..8fe4610 100644
@@ -579,6 +579,7 @@ struct perf_sample_data {
 		u32	reserved;
 	}				cpu_entry;
 	u64				period;
+	union  perf_mem_dsrc		dsrc;
 	struct perf_callchain_entry	*callchain;
 	struct perf_raw_record		*raw;
 	struct perf_branch_stack	*br_stack;
@@ -599,6 +600,7 @@ static inline void perf_sample_data_init(struct perf_sample_data *data,
 	data->regs_user.regs = NULL;
 	data->stack_user_size = 0;
 	data->weight = 0;
+	data->dsrc.val = 0;
 }
 
 extern void perf_output_sample(struct perf_output_handle *handle,
include/uapi/linux/perf_event.h +56 -2
index 3e6c394..3e4844c 100644
@@ -133,9 +133,9 @@ enum perf_event_sample_format {
 	PERF_SAMPLE_REGS_USER			= 1U << 12,
 	PERF_SAMPLE_STACK_USER			= 1U << 13,
 	PERF_SAMPLE_WEIGHT			= 1U << 14,
+	PERF_SAMPLE_DSRC			= 1U << 15,
 
-	PERF_SAMPLE_MAX = 1U << 15,		/* non-ABI */
-
+	PERF_SAMPLE_MAX = 1U << 16,		/* non-ABI */
 };
 
 /*
@@ -591,6 +591,7 @@ enum perf_event_type {
 	 * 	  u64			dyn_size; } && PERF_SAMPLE_STACK_USER
 	 *
 	 *	{ u64			weight;   } && PERF_SAMPLE_WEIGHT
+	 *	{ u64			dsrc;     } && PERF_SAMPLE_DSRC
 	 * };
 	 */
 	PERF_RECORD_SAMPLE			= 9,
@@ -616,4 +617,67 @@ enum perf_callchain_context {
 #define PERF_FLAG_FD_OUTPUT		(1U << 1)
 #define PERF_FLAG_PID_CGROUP		(1U << 2) /* pid=cgroup id, per-cpu mode only */
 
+union perf_mem_dsrc {
+	__u64 val;
+	struct {
+		__u64   mem_op:5,	/* type of opcode */
+			mem_lvl:14,	/* memory hierarchy level */
+			mem_snoop:5,	/* snoop mode */
+			mem_lock:2,	/* lock instr */
+			mem_dtlb:7,	/* tlb access */
+			mem_rsvd:31;
+	};
+};
+
+/* type of opcode (load/store/prefetch,code) */
+#define PERF_MEM_OP_NA		0x01 /* not available */
+#define PERF_MEM_OP_LOAD	0x02 /* load instruction */
+#define PERF_MEM_OP_STORE	0x04 /* store instruction */
+#define PERF_MEM_OP_PFETCH	0x08 /* prefetch */
+#define PERF_MEM_OP_EXEC	0x10 /* code (execution) */
+#define PERF_MEM_OP_SHIFT	0
+
+/* memory hierarchy (memory level, hit or miss) */
+#define PERF_MEM_LVL_NA		0x01  /* not available */
+#define PERF_MEM_LVL_HIT	0x02  /* hit level */
+#define PERF_MEM_LVL_MISS	0x04  /* miss level  */
+#define PERF_MEM_LVL_L1		0x08  /* L1 */
+#define PERF_MEM_LVL_LFB	0x10  /* Line Fill Buffer */
+#define PERF_MEM_LVL_L2		0x20  /* L2 hit */
+#define PERF_MEM_LVL_L3		0x40  /* L3 hit */
+#define PERF_MEM_LVL_LOC_RAM	0x80  /* Local DRAM */
+#define PERF_MEM_LVL_REM_RAM1	0x100 /* Remote DRAM (1 hop) */
+#define PERF_MEM_LVL_REM_RAM2	0x200 /* Remote DRAM (2 hops) */
+#define PERF_MEM_LVL_REM_CCE1	0x400 /* Remote Cache (1 hop) */
+#define PERF_MEM_LVL_REM_CCE2	0x800 /* Remote Cache (2 hops) */
+#define PERF_MEM_LVL_IO		0x1000 /* I/O memory */
+#define PERF_MEM_LVL_UNC	0x2000 /* Uncached memory */
+#define PERF_MEM_LVL_SHIFT	5
+
+/* snoop mode */
+#define PERF_MEM_SNOOP_NA	0x01 /* not available */
+#define PERF_MEM_SNOOP_NONE	0x02 /* no snoop */
+#define PERF_MEM_SNOOP_HIT	0x04 /* snoop hit */
+#define PERF_MEM_SNOOP_MISS	0x08 /* snoop miss */
+#define PERF_MEM_SNOOP_HITM	0x10 /* snoop hit modified */
+#define PERF_MEM_SNOOP_SHIFT	19
+
+/* locked instruction */
+#define PERF_MEM_LOCK_NA	0x01 /* not available */
+#define PERF_MEM_LOCK_LOCKED	0x02 /* locked transaction */
+#define PERF_MEM_LOCK_SHIFT	24
+
+/* TLB access */
+#define PERF_MEM_TLB_NA		0x01 /* not available */
+#define PERF_MEM_TLB_HIT	0x02 /* hit level */

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 682

Quoted 10 lines
include/uapi/linux/perf_event.h +10 -0
+#define PERF_MEM_TLB_MISS	0x04 /* miss level */
+#define PERF_MEM_TLB_L1		0x08 /* L1 */
+#define PERF_MEM_TLB_L2		0x10 /* L2 */
+#define PERF_MEM_TLB_WK		0x20 /* Hardware Walker*/
+#define PERF_MEM_TLB_OS		0x40 /* OS fault handler */
+#define PERF_MEM_TLB_SHIFT	26
+
+#define PERF_MEM_S(a, s) \
+	(((u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT)
+

Would be nice to get feedback from PowerPC folks to see how well this matches their memory profiling hw capabilities?

I suspect there's a lot of differences, but one can always hope
...

If there's some hope for unification we could at least shape it in a way that they could pick up and extend.

Thanks,

Ingo

#18 lore raw

Re: [PATCH v7 07/18] perf: add generic memory sampling interface

Quoted 124 lines · Ingo Molnar
Quoted 121 lines · Stephane Eranian

This patch adds PERF_SAMPLE_DSRC.

PERF_SAMPLE_DSRC collects the data source, i.e., where did the data associated with the sampled instruction come from. Information is stored in a perf_mem_dsrc structure. It contains opcode, mem level, tlb, snoop, lock information, subject to availability in hardware.

---
 include/linux/perf_event.h      |    2 ++
 include/uapi/linux/perf_event.h |   68 +++++++++++++++++++++++++++++++++++++--
 kernel/events/core.c            |    6 ++++
 3 files changed, 74 insertions(+), 2 deletions(-)
include/linux/perf_event.h +2 -0
index bb2429d..8fe4610 100644
@@ -579,6 +579,7 @@ struct perf_sample_data {
              u32     reserved;
      }                               cpu_entry;
      u64                             period;
+     union  perf_mem_dsrc            dsrc;
      struct perf_callchain_entry     *callchain;
      struct perf_raw_record          *raw;
      struct perf_branch_stack        *br_stack;
@@ -599,6 +600,7 @@ static inline void perf_sample_data_init(struct perf_sample_data *data,
      data->regs_user.regs = NULL;
      data->stack_user_size = 0;
      data->weight = 0;
+     data->dsrc.val = 0;
 }
 
 extern void perf_output_sample(struct perf_output_handle *handle,
include/uapi/linux/perf_event.h +61 -2
index 3e6c394..3e4844c 100644
@@ -133,9 +133,9 @@ enum perf_event_sample_format {
      PERF_SAMPLE_REGS_USER                   = 1U << 12,
      PERF_SAMPLE_STACK_USER                  = 1U << 13,
      PERF_SAMPLE_WEIGHT                      = 1U << 14,
+     PERF_SAMPLE_DSRC                        = 1U << 15,
 
-     PERF_SAMPLE_MAX = 1U << 15,             /* non-ABI */
-
+     PERF_SAMPLE_MAX = 1U << 16,             /* non-ABI */
 };
 
 /*
@@ -591,6 +591,7 @@ enum perf_event_type {
       *        u64                   dyn_size; } && PERF_SAMPLE_STACK_USER
       *
       *      { u64                   weight;   } && PERF_SAMPLE_WEIGHT
+      *      { u64                   dsrc;     } && PERF_SAMPLE_DSRC
       * };
       */
      PERF_RECORD_SAMPLE                      = 9,
@@ -616,4 +617,67 @@ enum perf_callchain_context {
 #define PERF_FLAG_FD_OUTPUT          (1U << 1)
 #define PERF_FLAG_PID_CGROUP         (1U << 2) /* pid=cgroup id, per-cpu mode only */
 
+union perf_mem_dsrc {
+     __u64 val;
+     struct {
+             __u64   mem_op:5,       /* type of opcode */
+                     mem_lvl:14,     /* memory hierarchy level */
+                     mem_snoop:5,    /* snoop mode */
+                     mem_lock:2,     /* lock instr */
+                     mem_dtlb:7,     /* tlb access */
+                     mem_rsvd:31;
+     };
+};
+
+/* type of opcode (load/store/prefetch,code) */
+#define PERF_MEM_OP_NA               0x01 /* not available */
+#define PERF_MEM_OP_LOAD     0x02 /* load instruction */
+#define PERF_MEM_OP_STORE    0x04 /* store instruction */
+#define PERF_MEM_OP_PFETCH   0x08 /* prefetch */
+#define PERF_MEM_OP_EXEC     0x10 /* code (execution) */
+#define PERF_MEM_OP_SHIFT    0
+
+/* memory hierarchy (memory level, hit or miss) */
+#define PERF_MEM_LVL_NA              0x01  /* not available */
+#define PERF_MEM_LVL_HIT     0x02  /* hit level */
+#define PERF_MEM_LVL_MISS    0x04  /* miss level  */
+#define PERF_MEM_LVL_L1              0x08  /* L1 */
+#define PERF_MEM_LVL_LFB     0x10  /* Line Fill Buffer */
+#define PERF_MEM_LVL_L2              0x20  /* L2 hit */
+#define PERF_MEM_LVL_L3              0x40  /* L3 hit */
+#define PERF_MEM_LVL_LOC_RAM 0x80  /* Local DRAM */
+#define PERF_MEM_LVL_REM_RAM1        0x100 /* Remote DRAM (1 hop) */
+#define PERF_MEM_LVL_REM_RAM2        0x200 /* Remote DRAM (2 hops) */
+#define PERF_MEM_LVL_REM_CCE1        0x400 /* Remote Cache (1 hop) */
+#define PERF_MEM_LVL_REM_CCE2        0x800 /* Remote Cache (2 hops) */
+#define PERF_MEM_LVL_IO              0x1000 /* I/O memory */
+#define PERF_MEM_LVL_UNC     0x2000 /* Uncached memory */
+#define PERF_MEM_LVL_SHIFT   5
+
+/* snoop mode */
+#define PERF_MEM_SNOOP_NA    0x01 /* not available */
+#define PERF_MEM_SNOOP_NONE  0x02 /* no snoop */
+#define PERF_MEM_SNOOP_HIT   0x04 /* snoop hit */
+#define PERF_MEM_SNOOP_MISS  0x08 /* snoop miss */
+#define PERF_MEM_SNOOP_HITM  0x10 /* snoop hit modified */
+#define PERF_MEM_SNOOP_SHIFT 19
+
+/* locked instruction */
+#define PERF_MEM_LOCK_NA     0x01 /* not available */
+#define PERF_MEM_LOCK_LOCKED 0x02 /* locked transaction */
+#define PERF_MEM_LOCK_SHIFT  24
+
+/* TLB access */
+#define PERF_MEM_TLB_NA              0x01 /* not available */
+#define PERF_MEM_TLB_HIT     0x02 /* hit level */
+#define PERF_MEM_TLB_MISS    0x04 /* miss level */
+#define PERF_MEM_TLB_L1              0x08 /* L1 */
+#define PERF_MEM_TLB_L2              0x10 /* L2 */
+#define PERF_MEM_TLB_WK              0x20 /* Hardware Walker*/
+#define PERF_MEM_TLB_OS              0x40 /* OS fault handler */
Quoted 9 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 682

Quoted 5 lines
include/uapi/linux/perf_event.h +5 -0
+#define PERF_MEM_TLB_SHIFT   26
+
+#define PERF_MEM_S(a, s) \
+     (((u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT)
+

Would be nice to get feedback from PowerPC folks to see how well this matches their memory profiling hw capabilities?

I agree, I tried to remain as generic as possible here but I probably don't have all the possibilities covered. I remember IBM asking me about the categories a long time ago. Haven't heard anything since then.

Quoted 6 lines

I suspect there's a lot of differences, but one can always hope
...

If there's some hope for unification we could at least shape it in a way that they could pick up and extend.

Agreed.

#19 lore raw

Re: [PATCH v7 07/18] perf: add generic memory sampling interface

Quoted 8 lines · Stephane Eranian
Quoted 4 lines · Ingo Molnar

Would be nice to get feedback from PowerPC folks to see how well this matches their memory profiling hw capabilities?

I agree, I tried to remain as generic as possible here but I probably don't have all the possibilities covered. I remember IBM asking me about the categories a long time ago. Haven't heard anything since then.

I'm at linux.conf.au this week so I probably won't have time to look at it properly 'til next week sorry.

cheers

#20 lore raw

Re: [PATCH v7 07/18] perf: add generic memory sampling interface

In reply to Ingo Molnar
Quoted 66 lines · Stephane Eranian
Quoted 64 lines

This patch adds PERF_SAMPLE_DSRC.

PERF_SAMPLE_DSRC collects the data source, i.e., where did the data associated with the sampled instruction come from. Information is stored in a perf_mem_dsrc structure. It contains opcode, mem level, tlb, snoop, lock information, subject to availability in hardware.

---
 include/linux/perf_event.h      |    2 ++
 include/uapi/linux/perf_event.h |   68 +++++++++++++++++++++++++++++++++++++--
 kernel/events/core.c            |    6 ++++
 3 files changed, 74 insertions(+), 2 deletions(-)
include/linux/perf_event.h +2 -0
index bb2429d..8fe4610 100644
@@ -579,6 +579,7 @@ struct perf_sample_data {
 		u32	reserved;
 	}				cpu_entry;
 	u64				period;
+	union  perf_mem_dsrc		dsrc;
 	struct perf_callchain_entry	*callchain;
 	struct perf_raw_record		*raw;
 	struct perf_branch_stack	*br_stack;
@@ -599,6 +600,7 @@ static inline void perf_sample_data_init(struct perf_sample_data *data,
 	data->regs_user.regs = NULL;
 	data->stack_user_size = 0;
 	data->weight = 0;
+	data->dsrc.val = 0;
 }
 
 extern void perf_output_sample(struct perf_output_handle *handle,
include/uapi/linux/perf_event.h +3 -2
index 3e6c394..3e4844c 100644
@@ -133,9 +133,9 @@ enum perf_event_sample_format {
 	PERF_SAMPLE_REGS_USER			= 1U << 12,
 	PERF_SAMPLE_STACK_USER			= 1U << 13,
 	PERF_SAMPLE_WEIGHT			= 1U << 14,
+	PERF_SAMPLE_DSRC			= 1U << 15,
 
-	PERF_SAMPLE_MAX = 1U << 15,		/* non-ABI */
-
+	PERF_SAMPLE_MAX = 1U << 16,		/* non-ABI */
 };
 
 /*
@@ -591,6 +591,7 @@ enum perf_event_type {
 	 * 	  u64			dyn_size; } && PERF_SAMPLE_STACK_USER
 	 *
 	 *	{ u64			weight;   } && PERF_SAMPLE_WEIGHT
+	 *	{ u64			dsrc;     } && PERF_SAMPLE_DSRC
 	 * };
 	 */
 	PERF_RECORD_SAMPLE			= 9,
@@ -616,4 +617,67 @@ enum perf_callchain_context {
 #define PERF_FLAG_FD_OUTPUT		(1U << 1)
 #define PERF_FLAG_PID_CGROUP		(1U << 2) /* pid=cgroup id, per-cpu mode only */
 
Quoted 10 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 629

Quoted 10 lines
include/uapi/linux/perf_event.h +10 -0
+union perf_mem_dsrc {
+	__u64 val;
+	struct {
+		__u64   mem_op:5,	/* type of opcode */
+			mem_lvl:14,	/* memory hierarchy level */
+			mem_snoop:5,	/* snoop mode */
+			mem_lock:2,	/* lock instr */
+			mem_dtlb:7,	/* tlb access */
+			mem_rsvd:31;
+	};

POWER could use an additional field:

mem_deratmiss:1

AFAICT, POWER does not currently save the mem_op, snoop or lock info for the sampled instruction. I guess we can leave them set to 0.

Quoted 18 lines
Quoted 18 lines
include/uapi/linux/perf_event.h +18 -0
+};
+
+/* type of opcode (load/store/prefetch,code) */
+#define PERF_MEM_OP_NA		0x01 /* not available */
+#define PERF_MEM_OP_LOAD	0x02 /* load instruction */
+#define PERF_MEM_OP_STORE	0x04 /* store instruction */
+#define PERF_MEM_OP_PFETCH	0x08 /* prefetch */
+#define PERF_MEM_OP_EXEC	0x10 /* code (execution) */
+#define PERF_MEM_OP_SHIFT	0
+
+/* memory hierarchy (memory level, hit or miss) */
+#define PERF_MEM_LVL_NA		0x01  /* not available */
+#define PERF_MEM_LVL_HIT	0x02  /* hit level */
+#define PERF_MEM_LVL_MISS	0x04  /* miss level  */
+#define PERF_MEM_LVL_L1		0x08  /* L1 */
+#define PERF_MEM_LVL_LFB	0x10  /* Line Fill Buffer */
+#define PERF_MEM_LVL_L2		0x20  /* L2 hit */
+#define PERF_MEM_LVL_L3		0x40  /* L3 hit */
Quoted 8 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 655

Quoted 8 lines
include/uapi/linux/perf_event.h +8 -0
+#define PERF_MEM_LVL_LOC_RAM	0x80  /* Local DRAM */
+#define PERF_MEM_LVL_REM_RAM1	0x100 /* Remote DRAM (1 hop) */
+#define PERF_MEM_LVL_REM_RAM2	0x200 /* Remote DRAM (2 hops) */
+#define PERF_MEM_LVL_REM_CCE1	0x400 /* Remote Cache (1 hop) */
+#define PERF_MEM_LVL_REM_CCE2	0x800 /* Remote Cache (2 hops) */
+#define PERF_MEM_LVL_IO		0x1000 /* I/O memory */
+#define PERF_MEM_LVL_UNC	0x2000 /* Uncached memory */
+#define PERF_MEM_LVL_SHIFT	5

POWER saves following information to describe where the data was loaded from after a Dcache or DTLB miss.

FROM_L2
FROM_L3

FROM_L2.1_SHR From another L2 or L3 on same chip, shared
FROM_L2.1_MOD From another L2 or L3 on same chip, modified

FROM_L3.1_SHR From remote L2 or L3, shared
FROM_L3.1_MOD From remote L2 or L3, modified

FROM_RL2L3_SHR From remote L2 or L3, shared
FROM_RL2L3_MOD From remote L2 or L3, modified

FROM_DL2L3_SHR From distant L2 or L3, shared
FROM_DL2L3_MOD From distant L2 or L3, modified

POWER uses 4 bits and a running count for its (currently) 13 possible values.

The macros in the patch use a separate bit for each level - is that to allow selecting more than one level at the same time ? If so, we will need to reserve a few more bits to allow for Power's memory levels that don't map to the above.

Quoted 15 lines
Quoted 15 lines
include/uapi/linux/perf_event.h +15 -0
+
+/* snoop mode */
+#define PERF_MEM_SNOOP_NA	0x01 /* not available */
+#define PERF_MEM_SNOOP_NONE	0x02 /* no snoop */
+#define PERF_MEM_SNOOP_HIT	0x04 /* snoop hit */
+#define PERF_MEM_SNOOP_MISS	0x08 /* snoop miss */
+#define PERF_MEM_SNOOP_HITM	0x10 /* snoop hit modified */
+#define PERF_MEM_SNOOP_SHIFT	19
+
+/* locked instruction */
+#define PERF_MEM_LOCK_NA	0x01 /* not available */
+#define PERF_MEM_LOCK_LOCKED	0x02 /* locked transaction */
+#define PERF_MEM_LOCK_SHIFT	24
+
+/* TLB access */
Quoted 8 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 678

Quoted 8 lines
include/uapi/linux/perf_event.h +8 -0
+#define PERF_MEM_TLB_NA		0x01 /* not available */
+#define PERF_MEM_TLB_HIT	0x02 /* hit level */
+#define PERF_MEM_TLB_MISS	0x04 /* miss level */
+#define PERF_MEM_TLB_L1		0x08 /* L1 */
+#define PERF_MEM_TLB_L2		0x10 /* L2 */
+#define PERF_MEM_TLB_WK		0x20 /* Hardware Walker*/
+#define PERF_MEM_TLB_OS		0x40 /* OS fault handler */
+#define PERF_MEM_TLB_SHIFT	26

On POWER, like with the Dcache source above, we have 4 bits to describe where the DTLB was loaded from after a dTLB miss.

We would probably need to allow more bits to for the memory level of the dTLB load source.

Quoted 8 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 682

Quoted 4 lines
include/uapi/linux/perf_event.h +4 -0
+
+#define PERF_MEM_S(a, s) \
+	(((u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT)
+

Would be nice to get feedback from PowerPC folks to see how well this matches their memory profiling hw capabilities?

Quoted 5 lines

I suspect there's a lot of differences, but one can always hope
...

If there's some hope for unification we could at least shape it in a way that they could pick up and extend.

Thanks for Ccing.

While on the topic of sampled instructions, POWER saves following information (in addition to the above memory info) for sampled instructions.

- whether the sampled instruction encountered a stall
- the reasons for the stall.
- whether the instruction was from hypervisor
- there was a branch mis-predict,
- thresholding information

These are clubbed into an "event vector" that is saved for sampled instructions. We have been meaning to find ways to present that to to user space. Are there plans to retreive and present these too.

Sukadev

#21 lore raw

Re: [PATCH v7 07/18] perf: add generic memory sampling interface

Quoted 4 lines · Sukadev Bhattiprolu

POWER could use an additional field:

mem_deratmiss:1

If you want to make that field more "generic" make it "lvl1_tlb_miss", ie, a miss in the internal "level 1" TLB which is the smallest/fastest TLB level in the load/store unit.

Quoted 2 lines

AFAICT, POWER does not currently save the mem_op, snoop or lock info for the sampled instruction. I guess we can leave them set to 0.

Well, we don't have lock instructions to begin with :-) If we can read the IP then we can deduce the memop tho.

Quoted 118 lines
Quoted 12 lines
Quoted 12 lines
include/uapi/linux/perf_event.h +12 -0
+};
+
+/* type of opcode (load/store/prefetch,code) */
+#define PERF_MEM_OP_NA             0x01 /* not available */
+#define PERF_MEM_OP_LOAD   0x02 /* load instruction */
+#define PERF_MEM_OP_STORE  0x04 /* store instruction */
+#define PERF_MEM_OP_PFETCH 0x08 /* prefetch */
+#define PERF_MEM_OP_EXEC   0x10 /* code (execution) */
+#define PERF_MEM_OP_SHIFT  0
+
+/* memory hierarchy (memory level, hit or miss) */
+#define PERF_MEM_LVL_NA            0x01  /* not available */
Quoted 8 lines

include/uapi/linux/perf_event.h enum perf_callchain_context {

Quoted 8 lines
include/uapi/linux/perf_event.h +8 -0
+#define PERF_MEM_LVL_HIT   0x02  /* hit level */
+#define PERF_MEM_LVL_MISS  0x04  /* miss level  */
+#define PERF_MEM_LVL_L1            0x08  /* L1 */
+#define PERF_MEM_LVL_LFB   0x10  /* Line Fill Buffer */
+#define PERF_MEM_LVL_L2            0x20  /* L2 hit */
+#define PERF_MEM_LVL_L3            0x40  /* L3 hit */
+#define PERF_MEM_LVL_LOC_RAM       0x80  /* Local DRAM */
+#define PERF_MEM_LVL_REM_RAM1      0x100 /* Remote DRAM (1 hop)

*/

Quoted 1 line
Quoted 1 line

+#define PERF_MEM_LVL_REM_RAM2 0x200 /* Remote DRAM (2 hops)

*/

Quoted 1 line
Quoted 1 line

+#define PERF_MEM_LVL_REM_CCE1 0x400 /* Remote Cache (1 hop)

*/

Quoted 1 line
Quoted 1 line

+#define PERF_MEM_LVL_REM_CCE2 0x800 /* Remote Cache (2 hops)

*/

Quoted 3 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 655

Quoted 3 lines
include/uapi/linux/perf_event.h +3 -0
+#define PERF_MEM_LVL_IO            0x1000 /* I/O memory */
+#define PERF_MEM_LVL_UNC   0x2000 /* Uncached memory */
+#define PERF_MEM_LVL_SHIFT 5

POWER saves following information to describe where the data was loaded from after a Dcache or DTLB miss.

        FROM_L2
        FROM_L3

        FROM_L2.1_SHR   From another L2 or L3 on same chip,

shared
FROM_L2.1_MOD From another L2 or L3 on same chip, modified

FROM_L3.1_SHR From remote L2 or L3, shared
FROM_L3.1_MOD From remote L2 or L3, modified

FROM_RL2L3_SHR From remote L2 or L3, shared FROM_RL2L3_MOD From remote L2 or L3, modified

FROM_DL2L3_SHR From distant L2 or L3, shared FROM_DL2L3_MOD From distant L2 or L3, modified

POWER uses 4 bits and a running count for its (currently) 13 possible values.

The macros in the patch use a separate bit for each level - is that to allow
selecting more than one level at the same time ? If so, we will need to reserve
a few more bits to allow for Power's memory levels that don't map to the above.

Quoted 15 lines
Quoted 15 lines
include/uapi/linux/perf_event.h +15 -0
+
+/* snoop mode */
+#define PERF_MEM_SNOOP_NA  0x01 /* not available */
+#define PERF_MEM_SNOOP_NONE        0x02 /* no snoop */
+#define PERF_MEM_SNOOP_HIT 0x04 /* snoop hit */
+#define PERF_MEM_SNOOP_MISS        0x08 /* snoop miss */
+#define PERF_MEM_SNOOP_HITM        0x10 /* snoop hit modified */
+#define PERF_MEM_SNOOP_SHIFT       19
+
+/* locked instruction */
+#define PERF_MEM_LOCK_NA   0x01 /* not available */
+#define PERF_MEM_LOCK_LOCKED       0x02 /* locked transaction */
+#define PERF_MEM_LOCK_SHIFT        24
+
+/* TLB access */
Quoted 8 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 678

Quoted 8 lines
include/uapi/linux/perf_event.h +8 -0
+#define PERF_MEM_TLB_NA            0x01 /* not available */
+#define PERF_MEM_TLB_HIT   0x02 /* hit level */
+#define PERF_MEM_TLB_MISS  0x04 /* miss level */
+#define PERF_MEM_TLB_L1            0x08 /* L1 */
+#define PERF_MEM_TLB_L2            0x10 /* L2 */
+#define PERF_MEM_TLB_WK            0x20 /* Hardware Walker*/
+#define PERF_MEM_TLB_OS            0x40 /* OS fault handler */
+#define PERF_MEM_TLB_SHIFT 26

On POWER, like with the Dcache source above, we have 4 bits to describe where
the DTLB was loaded from after a dTLB miss.

We would probably need to allow more bits to for the memory level of the dTLB
load source.

Quoted 8 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 682

Quoted 4 lines
include/uapi/linux/perf_event.h +4 -0
+
+#define PERF_MEM_S(a, s) \
+   (((u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT)
+

Would be nice to get feedback from PowerPC folks to see how well this matches their memory profiling hw capabilities?

Quoted 5 lines

I suspect there's a lot of differences, but one can always hope
...

If there's some hope for unification we could at least shape it in a way that they could pick up and extend.

Thanks for Ccing.

While on the topic of sampled instructions, POWER saves following information
(in addition to the above memory info) for sampled instructions.

- whether the sampled instruction encountered a stall
- the reasons for the stall.
- whether the instruction was from hypervisor
- there was a branch mis-predict,
- thresholding information

Quoted 3 lines

These are clubbed into an "event vector" that is saved for sampled instructions. We have been meaning to find ways to present that to to user space. Are there plans to retreive and present these too.

Ben.

#22 lore raw

Re: [PATCH v7 07/18] perf: add generic memory sampling interface

Quoted 10 lines · Benjamin Herrenschmidt
Quoted 4 lines · Sukadev Bhattiprolu

POWER could use an additional field:

mem_deratmiss:1

If you want to make that field more "generic" make it "lvl1_tlb_miss", ie, a miss in the internal "level 1" TLB which is the smallest/fastest TLB level in the load/store unit.

I'd also suggest adding spare bits generously, if this is a fixed width ABI component - or a size field.

Thanks,

Ingo

#23 lore raw

Re: [PATCH v7 07/18] perf: add generic memory sampling interface

In reply to Benjamin Herrenschmidt
Quoted 10 lines · Benjamin Herrenschmidt
Quoted 4 lines · Sukadev Bhattiprolu

POWER could use an additional field:

mem_deratmiss:1

If you want to make that field more "generic" make it "lvl1_tlb_miss", ie, a miss in the internal "level 1" TLB which is the smallest/fastest TLB level in the load/store unit.

If you want to express a L1 TLB miss, you already can do that:

PERF_MEM_S(TLB, MISS) | PERF_MEM_S(TLB, L1)

If this is a feature you do not support, then use the NA macro.

For instance:
PERF_MEM_S(LOCK, NA)

Need to be able to differentiate not supported from did not happen.

Quoted 131 lines
Quoted 2 lines

AFAICT, POWER does not currently save the mem_op, snoop or lock info for the sampled instruction. I guess we can leave them set to 0.

Well, we don't have lock instructions to begin with :-) If we can read the IP then we can deduce the memop tho.

Quoted 118 lines
Quoted 12 lines
Quoted 12 lines
include/uapi/linux/perf_event.h +12 -0
+};
+
+/* type of opcode (load/store/prefetch,code) */
+#define PERF_MEM_OP_NA             0x01 /* not available */
+#define PERF_MEM_OP_LOAD   0x02 /* load instruction */
+#define PERF_MEM_OP_STORE  0x04 /* store instruction */
+#define PERF_MEM_OP_PFETCH 0x08 /* prefetch */
+#define PERF_MEM_OP_EXEC   0x10 /* code (execution) */
+#define PERF_MEM_OP_SHIFT  0
+
+/* memory hierarchy (memory level, hit or miss) */
+#define PERF_MEM_LVL_NA            0x01  /* not available */
Quoted 8 lines

include/uapi/linux/perf_event.h enum perf_callchain_context {

Quoted 8 lines
include/uapi/linux/perf_event.h +8 -0
+#define PERF_MEM_LVL_HIT   0x02  /* hit level */
+#define PERF_MEM_LVL_MISS  0x04  /* miss level  */
+#define PERF_MEM_LVL_L1            0x08  /* L1 */
+#define PERF_MEM_LVL_LFB   0x10  /* Line Fill Buffer */
+#define PERF_MEM_LVL_L2            0x20  /* L2 hit */
+#define PERF_MEM_LVL_L3            0x40  /* L3 hit */
+#define PERF_MEM_LVL_LOC_RAM       0x80  /* Local DRAM */
+#define PERF_MEM_LVL_REM_RAM1      0x100 /* Remote DRAM (1 hop)

*/

Quoted 1 line
Quoted 1 line

+#define PERF_MEM_LVL_REM_RAM2 0x200 /* Remote DRAM (2 hops)

*/

Quoted 1 line
Quoted 1 line

+#define PERF_MEM_LVL_REM_CCE1 0x400 /* Remote Cache (1 hop)

*/

Quoted 1 line
Quoted 1 line

+#define PERF_MEM_LVL_REM_CCE2 0x800 /* Remote Cache (2 hops)

*/

Quoted 3 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 655

Quoted 3 lines
include/uapi/linux/perf_event.h +3 -0
+#define PERF_MEM_LVL_IO            0x1000 /* I/O memory */
+#define PERF_MEM_LVL_UNC   0x2000 /* Uncached memory */
+#define PERF_MEM_LVL_SHIFT 5

POWER saves following information to describe where the data was loaded from after a Dcache or DTLB miss.

        FROM_L2
        FROM_L3

        FROM_L2.1_SHR   From another L2 or L3 on same chip,

shared
FROM_L2.1_MOD From another L2 or L3 on same chip, modified

FROM_L3.1_SHR From remote L2 or L3, shared
FROM_L3.1_MOD From remote L2 or L3, modified

FROM_RL2L3_SHR From remote L2 or L3, shared FROM_RL2L3_MOD From remote L2 or L3, modified

FROM_DL2L3_SHR From distant L2 or L3, shared FROM_DL2L3_MOD From distant L2 or L3, modified

POWER uses 4 bits and a running count for its (currently) 13 possible values.

The macros in the patch use a separate bit for each level - is that to allow
selecting more than one level at the same time ? If so, we will need to reserve
a few more bits to allow for Power's memory levels that don't map to the above.

Quoted 15 lines
Quoted 15 lines
include/uapi/linux/perf_event.h +15 -0
+
+/* snoop mode */
+#define PERF_MEM_SNOOP_NA  0x01 /* not available */
+#define PERF_MEM_SNOOP_NONE        0x02 /* no snoop */
+#define PERF_MEM_SNOOP_HIT 0x04 /* snoop hit */
+#define PERF_MEM_SNOOP_MISS        0x08 /* snoop miss */
+#define PERF_MEM_SNOOP_HITM        0x10 /* snoop hit modified */
+#define PERF_MEM_SNOOP_SHIFT       19
+
+/* locked instruction */
+#define PERF_MEM_LOCK_NA   0x01 /* not available */
+#define PERF_MEM_LOCK_LOCKED       0x02 /* locked transaction */
+#define PERF_MEM_LOCK_SHIFT        24
+
+/* TLB access */
Quoted 8 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 678

Quoted 8 lines
include/uapi/linux/perf_event.h +8 -0
+#define PERF_MEM_TLB_NA            0x01 /* not available */
+#define PERF_MEM_TLB_HIT   0x02 /* hit level */
+#define PERF_MEM_TLB_MISS  0x04 /* miss level */
+#define PERF_MEM_TLB_L1            0x08 /* L1 */
+#define PERF_MEM_TLB_L2            0x10 /* L2 */
+#define PERF_MEM_TLB_WK            0x20 /* Hardware Walker*/
+#define PERF_MEM_TLB_OS            0x40 /* OS fault handler */
+#define PERF_MEM_TLB_SHIFT 26

On POWER, like with the Dcache source above, we have 4 bits to describe where
the DTLB was loaded from after a dTLB miss.

We would probably need to allow more bits to for the memory level of the dTLB
load source.

Quoted 8 lines

include/uapi/linux/perf_event.h enum perf_callchain_context { · line 682

Quoted 4 lines
include/uapi/linux/perf_event.h +4 -0
+
+#define PERF_MEM_S(a, s) \
+   (((u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT)
+

Would be nice to get feedback from PowerPC folks to see how well this matches their memory profiling hw capabilities?

Quoted 5 lines

I suspect there's a lot of differences, but one can always hope
...

If there's some hope for unification we could at least shape it in a way that they could pick up and extend.

Thanks for Ccing.

While on the topic of sampled instructions, POWER saves following information
(in addition to the above memory info) for sampled instructions.

- whether the sampled instruction encountered a stall
- the reasons for the stall.
- whether the instruction was from hypervisor
- there was a branch mis-predict,
- thresholding information

Quoted 3 lines

These are clubbed into an "event vector" that is saved for sampled instructions. We have been meaning to find ways to present that to to user space. Are there plans to retreive and present these too.

Ben.

#24 lore raw

[tip:perf/x86] perf: Add generic memory sampling interface

In reply to Stephane Eranian
Commit-ID:  4b18e27234adb05597e7ec1f0eacda048aed080f
Gitweb:     http://git.kernel.org/tip/4b18e27234adb05597e7ec1f0eacda048aed080f
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:31 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 25 Jan 2013 10:19:03 +0100

perf: Add generic memory sampling interface

This patch adds PERF_SAMPLE_DATA_SRC.

PERF_SAMPLE_DATA_SRC collects the data source, i.e., where did the data associated with the sampled instruction come from. Information is stored in a perf_mem_data_src structure. It contains opcode, mem level, tlb, snoop, lock information, subject to availability in hardware.

---
 include/linux/perf_event.h      |  2 ++
 include/uapi/linux/perf_event.h | 68 +++++++++++++++++++++++++++++++++++++++--
 kernel/events/core.c            |  6 ++++
 3 files changed, 74 insertions(+), 2 deletions(-)
include/linux/perf_event.h +2 -0
index bb2429d..dad68a3 100644
@@ -579,6 +579,7 @@ struct perf_sample_data {
 		u32	reserved;
 	}				cpu_entry;
 	u64				period;
+	union  perf_mem_data_src	data_src;
 	struct perf_callchain_entry	*callchain;
 	struct perf_raw_record		*raw;
 	struct perf_branch_stack	*br_stack;
@@ -599,6 +600,7 @@ static inline void perf_sample_data_init(struct perf_sample_data *data,
 	data->regs_user.regs = NULL;
 	data->stack_user_size = 0;
 	data->weight = 0;
+	data->data_src.val = 0;
 }
 
 extern void perf_output_sample(struct perf_output_handle *handle,
include/uapi/linux/perf_event.h +66 -2

85 lines not on this page · raw message

kernel/events/core.c +6 -0
index 749bdf4..4661009 100644
@@ -958,6 +958,9 @@ static void perf_event__header_size(struct perf_event *event)
 	if (sample_type & PERF_SAMPLE_READ)
 		size += event->read_size;
 
+	if (sample_type & PERF_SAMPLE_DATA_SRC)
+		size += sizeof(data->data_src.val);
+
 	event->header_size = size;
 }
 
@@ -4175,6 +4178,9 @@ void perf_output_sample(struct perf_output_handle *handle,
 
 	if (sample_type & PERF_SAMPLE_WEIGHT)
 		perf_output_put(handle, data->weight);
+
+	if (sample_type & PERF_SAMPLE_DATA_SRC)
+		perf_output_put(handle, data->data_src.val);
 }
 
 void perf_prepare_sample(struct perf_event_header *header,
#25 lore raw

[tip:perf/core] perf: Add generic memory sampling interface

In reply to Stephane Eranian
perf: Add generic memory sampling interface
Commit-ID:  d6be9ad6c960f43800a6f118932bc8a5a4eadcd1
Gitweb:     http://git.kernel.org/tip/d6be9ad6c960f43800a6f118932bc8a5a4eadcd1
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:31 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:15:59 -0300

perf: Add generic memory sampling interface

This patch adds PERF_SAMPLE_DATA_SRC.

PERF_SAMPLE_DATA_SRC collects the data source, i.e., where did the data associated with the sampled instruction come from. Information is stored in a perf_mem_data_src structure. It contains opcode, mem level, tlb, snoop, lock information, subject to availability in hardware.

---
 include/linux/perf_event.h      |  2 ++
 include/uapi/linux/perf_event.h | 68 +++++++++++++++++++++++++++++++++++++++--
 kernel/events/core.c            |  6 ++++
 3 files changed, 74 insertions(+), 2 deletions(-)
include/linux/perf_event.h +2 -0
index 7ce0b37..42a6daa 100644
@@ -568,6 +568,7 @@ struct perf_sample_data {
 		u32	reserved;
 	}				cpu_entry;
 	u64				period;
+	union  perf_mem_data_src	data_src;
 	struct perf_callchain_entry	*callchain;
 	struct perf_raw_record		*raw;
 	struct perf_branch_stack	*br_stack;
@@ -588,6 +589,7 @@ static inline void perf_sample_data_init(struct perf_sample_data *data,
 	data->regs_user.regs = NULL;
 	data->stack_user_size = 0;
 	data->weight = 0;
+	data->data_src.val = 0;
 }
 
 extern void perf_output_sample(struct perf_output_handle *handle,
include/uapi/linux/perf_event.h +66 -2

85 lines not on this page · raw message

kernel/events/core.c +6 -0
index 9e3edb2..77c96d1 100644
@@ -982,6 +982,9 @@ static void perf_event__header_size(struct perf_event *event)
 	if (sample_type & PERF_SAMPLE_READ)
 		size += event->read_size;
 
+	if (sample_type & PERF_SAMPLE_DATA_SRC)
+		size += sizeof(data->data_src.val);
+
 	event->header_size = size;
 }
 
@@ -4199,6 +4202,9 @@ void perf_output_sample(struct perf_output_handle *handle,
 
 	if (sample_type & PERF_SAMPLE_WEIGHT)
 		perf_output_put(handle, data->weight);
+
+	if (sample_type & PERF_SAMPLE_DATA_SRC)
+		perf_output_put(handle, data->data_src.val);
 }
 
 void perf_prepare_sample(struct perf_event_header *header,
#26 lore raw

PATCH v7 8/18 perf/x86: add memory profiling via PEBS Load Latency

In reply to Stephane Eranian

This patch adds support for memory profiling using the PEBS Load Latency facility.

Load accesses are sampled by HW and the instruction address, data address, load latency, data source, tlb, locked information can be saved in the sampling buffer if using the PERF_SAMPLE_COST (for latency), PERF_SAMPLE_ADDR, PERF_SAMPLE_DSRC types.

To enable PEBS Load Latency, users have to use the model specific event:
- on NHM/WSM: MEM_INST_RETIRED:LATENCY_ABOVE_THRESHOLD
- on SNB/IVB: MEM_TRANS_RETIRED:LATENCY_ABOVE_THRESHOLD

To make things easier, this patch also exports a generic alias via sysfs: mem-loads. It export the right event encoding based on the host CPU and can be used directly by the perf tool.

Loosely based on Intel's Lin Ming patch posted on LKML in July 2011.

---
 arch/x86/include/uapi/asm/msr-index.h     |    1 +
 arch/x86/kernel/cpu/perf_event.c          |    5 +-
 arch/x86/kernel/cpu/perf_event.h          |   25 +++++-
 arch/x86/kernel/cpu/perf_event_intel.c    |   24 ++++++
 arch/x86/kernel/cpu/perf_event_intel_ds.c |  133 +++++++++++++++++++++++++++--
 5 files changed, 178 insertions(+), 10 deletions(-)
arch/x86/include/uapi/asm/msr-index.h +1 -0
index 433a59f..1031604 100644
@@ -71,6 +71,7 @@
 #define MSR_IA32_PEBS_ENABLE		0x000003f1
 #define MSR_IA32_DS_AREA		0x00000600
 #define MSR_IA32_PERF_CAPABILITIES	0x00000345
+#define MSR_PEBS_LD_LAT_THRESHOLD	0x000003f6
 
 #define MSR_MTRRfix64K_00000		0x00000250
 #define MSR_MTRRfix16K_80000		0x00000258
arch/x86/kernel/cpu/perf_event.c +4 -1
index 63f8dcf..aa53f07 100644
@@ -1361,7 +1361,7 @@ static __init struct attribute **merge_attr(struct attribute **a,
 	return new;
 }
 
-static ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
+ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
 			  char *page)
 {
 	struct perf_pmu_events_attr *pmu_attr = \
@@ -1492,6 +1492,9 @@ static int __init init_hw_perf_events(void)
 	x86_pmu.attr_rdpmc = 1; /* enable userspace RDPMC usage by default */
 	x86_pmu_format_group.attrs = x86_pmu.format_attrs;
 
+	if (x86_pmu.event_attrs)
+		x86_pmu_events_group.attrs = x86_pmu.event_attrs;
+
 	if (!x86_pmu.events_sysfs_show)
 		x86_pmu_events_group.attrs = &empty_attrs;
 	else
arch/x86/kernel/cpu/perf_event.h +24 -1
index 3f10cfe..3f91411 100644
@@ -46,6 +46,7 @@ enum extra_reg_type {
 	EXTRA_REG_RSP_0 = 0,	/* offcore_response_0 */
 	EXTRA_REG_RSP_1 = 1,	/* offcore_response_1 */
 	EXTRA_REG_LBR   = 2,	/* lbr_select */
+	EXTRA_REG_LDLAT = 3,	/* ld_lat_threshold */
 
 	EXTRA_REG_MAX		/* number of entries needed */
 };
@@ -61,6 +62,10 @@ struct event_constraint {
 	int	overlap;
 	int	flags;
 };
+/*
+ * struct event_constraint flags
+ */
+#define PERF_X86_EVENT_PEBS_LDLAT	0x1 /* ld+ldlat data address sampling */
 
 struct amd_nb {
 	int nb_id;  /* NorthBridge id */
@@ -233,6 +238,10 @@ struct cpu_hw_events {
 #define INTEL_UEVENT_CONSTRAINT(c, n)	\
 	EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK)
 
+#define INTEL_PLD_CONSTRAINT(c, n)	\
+	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \
+			   HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LDLAT)
+
 #define EVENT_CONSTRAINT_END		\
 	EVENT_CONSTRAINT(0, 0, 0)
 
@@ -262,12 +271,22 @@ struct extra_reg {
 	.msr = (ms),		\
 	.config_mask = (m),	\
 	.valid_mask = (vm),	\
-	.idx = EXTRA_REG_##i	\
+	.idx = EXTRA_REG_##i,	\
 	}
 
 #define INTEL_EVENT_EXTRA_REG(event, msr, vm, idx)	\
 	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT, vm, idx)
 
+#define INTEL_UEVENT_EXTRA_REG(event, msr, vm, idx) \
+	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT | \
+			ARCH_PERFMON_EVENTSEL_UMASK, vm, idx)
+
+#define INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(c) \
+	INTEL_UEVENT_EXTRA_REG(c, \
+			       MSR_PEBS_LD_LAT_THRESHOLD, \
+			       0xffff, \
+			       LDLAT)
+
 #define EVENT_EXTRA_END EVENT_EXTRA_REG(0, 0, 0, 0, RSP_0)
 
 union perf_capabilities {
@@ -355,6 +374,7 @@ struct x86_pmu {
 	 */
 	int		attr_rdpmc;
 	struct attribute **format_attrs;
+	struct attribute **event_attrs;
 
 	ssize_t		(*events_sysfs_show)(char *page, u64 config);
 	struct attribute **cpu_events;
@@ -659,6 +679,9 @@ int p6_pmu_init(void);
 
 int knc_pmu_init(void);
 
+ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
+			  char *page);
+
 #else /* CONFIG_CPU_SUP_INTEL */
 
 static inline void reserve_ds_buffers(void)
arch/x86/kernel/cpu/perf_event_intel.c +24 -0
index 67a8dd6..f30027a 100644
@@ -81,6 +81,7 @@ static struct event_constraint intel_nehalem_event_constraints[] __read_mostly =
 static struct extra_reg intel_nehalem_extra_regs[] __read_mostly =
 {
 	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0),
+	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x100b),
 	EVENT_EXTRA_END
 };
 
@@ -111,6 +112,7 @@ static struct extra_reg intel_westmere_extra_regs[] __read_mostly =
 {
 	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0),
 	INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0xffff, RSP_1),
+	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x100b),
 	EVENT_EXTRA_END
 };
 
@@ -130,9 +132,23 @@ static struct event_constraint intel_gen_event_constraints[] __read_mostly =
 static struct extra_reg intel_snb_extra_regs[] __read_mostly = {
 	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0x3fffffffffull, RSP_0),
 	INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0x3fffffffffull, RSP_1),
+	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x01cd),
 	EVENT_EXTRA_END
 };
 
+EVENT_ATTR_STR(mem-loads, mem_ld_nhm, "event=0x0b,umask=0x10,ldlat=3");
+EVENT_ATTR_STR(mem-loads, mem_ld_snb, "event=0xcd,umask=0x1,ldlat=3");
+
+struct attribute *nhm_events_attrs[] = {
+	EVENT_PTR(mem_ld_nhm),
+	NULL,
+};
+
+struct attribute *snb_events_attrs[] = {
+	EVENT_PTR(mem_ld_snb),
+	NULL,
+};
+
 static u64 intel_pmu_event_map(int hw_event)
 {
 	return intel_perfmon_event_map[hw_event];
@@ -2010,6 +2026,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.enable_all = intel_pmu_nhm_enable_all;
 		x86_pmu.extra_regs = intel_nehalem_extra_regs;
 
+		x86_pmu.cpu_events = nhm_events_attrs;
+
 		/* UOPS_ISSUED.STALLED_CYCLES */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
@@ -2050,6 +2068,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.extra_regs = intel_westmere_extra_regs;
 		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 
+		x86_pmu.cpu_events = nhm_events_attrs;
+
 		/* UOPS_ISSUED.STALLED_CYCLES */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
@@ -2078,6 +2098,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 		x86_pmu.er_flags |= ERF_NO_HT_SHARING;
 
+		x86_pmu.cpu_events = snb_events_attrs;
+
 		/* UOPS_ISSUED.ANY,c=1,i=1 to count stall cycles */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
@@ -2103,6 +2125,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 		x86_pmu.er_flags |= ERF_NO_HT_SHARING;
 
+		x86_pmu.cpu_events = snb_events_attrs;
+
 		/* UOPS_ISSUED.ANY,c=1,i=1 to count stall cycles */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
arch/x86/kernel/cpu/perf_event_intel_ds.c +125 -8

185 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#27 lore raw

[tip:perf/x86] perf/x86: Add memory profiling via PEBS Load Latency

Commit-ID:  66172a094e5286fefa70de78f23220e9ff264f77
Gitweb:     http://git.kernel.org/tip/66172a094e5286fefa70de78f23220e9ff264f77
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:32 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 25 Jan 2013 10:19:03 +0100

perf/x86: Add memory profiling via PEBS Load Latency

This patch adds support for memory profiling using the PEBS Load Latency facility.

Load accesses are sampled by HW and the instruction address, data address, load latency, data source, tlb, locked information can be saved in the sampling buffer if using the PERF_SAMPLE_COST (for latency), PERF_SAMPLE_ADDR, PERF_SAMPLE_DATA_SRC types.

To enable PEBS Load Latency, users have to use the model specific event:

- on NHM/WSM: MEM_INST_RETIRED:LATENCY_ABOVE_THRESHOLD
- on SNB/IVB: MEM_TRANS_RETIRED:LATENCY_ABOVE_THRESHOLD

To make things easier, this patch also exports a generic alias via sysfs: mem-loads. It export the right event encoding based on the host CPU and can be used directly by the perf tool.

Loosely based on Intel's Lin Ming patch posted on LKML in July 2011.

---
 arch/x86/include/uapi/asm/msr-index.h     |   1 +
 arch/x86/kernel/cpu/perf_event.c          |   5 +-
 arch/x86/kernel/cpu/perf_event.h          |  25 +++++-
 arch/x86/kernel/cpu/perf_event_intel.c    |  24 ++++++
 arch/x86/kernel/cpu/perf_event_intel_ds.c | 133 ++++++++++++++++++++++++++++--
 5 files changed, 178 insertions(+), 10 deletions(-)
arch/x86/include/uapi/asm/msr-index.h +1 -0
index 433a59f..1031604 100644
@@ -71,6 +71,7 @@
 #define MSR_IA32_PEBS_ENABLE		0x000003f1
 #define MSR_IA32_DS_AREA		0x00000600
 #define MSR_IA32_PERF_CAPABILITIES	0x00000345
+#define MSR_PEBS_LD_LAT_THRESHOLD	0x000003f6
 
 #define MSR_MTRRfix64K_00000		0x00000250
 #define MSR_MTRRfix16K_80000		0x00000258
arch/x86/kernel/cpu/perf_event.c +4 -1
index 49f760b..1e2519d 100644
@@ -1363,7 +1363,7 @@ static __init struct attribute **merge_attr(struct attribute **a, struct attribu
 	return new;
 }
 
-static ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
+ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
 			  char *page)
 {
 	struct perf_pmu_events_attr *pmu_attr = \
@@ -1494,6 +1494,9 @@ static int __init init_hw_perf_events(void)
 	x86_pmu.attr_rdpmc = 1; /* enable userspace RDPMC usage by default */
 	x86_pmu_format_group.attrs = x86_pmu.format_attrs;
 
+	if (x86_pmu.event_attrs)
+		x86_pmu_events_group.attrs = x86_pmu.event_attrs;
+
 	if (!x86_pmu.events_sysfs_show)
 		x86_pmu_events_group.attrs = &empty_attrs;
 	else
arch/x86/kernel/cpu/perf_event.h +24 -1
index 74b59ec..8f49e45 100644
@@ -46,6 +46,7 @@ enum extra_reg_type {
 	EXTRA_REG_RSP_0 = 0,	/* offcore_response_0 */
 	EXTRA_REG_RSP_1 = 1,	/* offcore_response_1 */
 	EXTRA_REG_LBR   = 2,	/* lbr_select */
+	EXTRA_REG_LDLAT = 3,	/* ld_lat_threshold */
 
 	EXTRA_REG_MAX		/* number of entries needed */
 };
@@ -61,6 +62,10 @@ struct event_constraint {
 	int	overlap;
 	int	flags;
 };
+/*
+ * struct event_constraint flags
+ */
+#define PERF_X86_EVENT_PEBS_LDLAT	0x1 /* ld+ldlat data address sampling */
 
 struct amd_nb {
 	int nb_id;  /* NorthBridge id */
@@ -233,6 +238,10 @@ struct cpu_hw_events {
 #define INTEL_UEVENT_CONSTRAINT(c, n)	\
 	EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK)
 
+#define INTEL_PLD_CONSTRAINT(c, n)	\
+	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \
+			   HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LDLAT)
+
 #define EVENT_CONSTRAINT_END		\
 	EVENT_CONSTRAINT(0, 0, 0)
 
@@ -262,12 +271,22 @@ struct extra_reg {
 	.msr = (ms),		\
 	.config_mask = (m),	\
 	.valid_mask = (vm),	\
-	.idx = EXTRA_REG_##i	\
+	.idx = EXTRA_REG_##i,	\
 	}
 
 #define INTEL_EVENT_EXTRA_REG(event, msr, vm, idx)	\
 	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT, vm, idx)
 
+#define INTEL_UEVENT_EXTRA_REG(event, msr, vm, idx) \
+	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT | \
+			ARCH_PERFMON_EVENTSEL_UMASK, vm, idx)
+
+#define INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(c) \
+	INTEL_UEVENT_EXTRA_REG(c, \
+			       MSR_PEBS_LD_LAT_THRESHOLD, \
+			       0xffff, \
+			       LDLAT)
+
 #define EVENT_EXTRA_END EVENT_EXTRA_REG(0, 0, 0, 0, RSP_0)
 
 union perf_capabilities {
@@ -355,6 +374,7 @@ struct x86_pmu {
 	 */
 	int		attr_rdpmc;
 	struct attribute **format_attrs;
+	struct attribute **event_attrs;
 
 	ssize_t		(*events_sysfs_show)(char *page, u64 config);
 	struct attribute **cpu_events;
@@ -659,6 +679,9 @@ int p6_pmu_init(void);
 
 int knc_pmu_init(void);
 
+ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
+			  char *page);
+
 #else /* CONFIG_CPU_SUP_INTEL */
 
 static inline void reserve_ds_buffers(void)
arch/x86/kernel/cpu/perf_event_intel.c +24 -0
index 67a8dd6..f30027a 100644
@@ -81,6 +81,7 @@ static struct event_constraint intel_nehalem_event_constraints[] __read_mostly =
 static struct extra_reg intel_nehalem_extra_regs[] __read_mostly =
 {
 	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0),
+	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x100b),
 	EVENT_EXTRA_END
 };
 
@@ -111,6 +112,7 @@ static struct extra_reg intel_westmere_extra_regs[] __read_mostly =
 {
 	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0),
 	INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0xffff, RSP_1),
+	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x100b),
 	EVENT_EXTRA_END
 };
 
@@ -130,9 +132,23 @@ static struct event_constraint intel_gen_event_constraints[] __read_mostly =
 static struct extra_reg intel_snb_extra_regs[] __read_mostly = {
 	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0x3fffffffffull, RSP_0),
 	INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0x3fffffffffull, RSP_1),
+	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x01cd),
 	EVENT_EXTRA_END
 };
 
+EVENT_ATTR_STR(mem-loads, mem_ld_nhm, "event=0x0b,umask=0x10,ldlat=3");
+EVENT_ATTR_STR(mem-loads, mem_ld_snb, "event=0xcd,umask=0x1,ldlat=3");
+
+struct attribute *nhm_events_attrs[] = {
+	EVENT_PTR(mem_ld_nhm),
+	NULL,
+};
+
+struct attribute *snb_events_attrs[] = {
+	EVENT_PTR(mem_ld_snb),
+	NULL,
+};
+
 static u64 intel_pmu_event_map(int hw_event)
 {
 	return intel_perfmon_event_map[hw_event];
@@ -2010,6 +2026,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.enable_all = intel_pmu_nhm_enable_all;
 		x86_pmu.extra_regs = intel_nehalem_extra_regs;
 
+		x86_pmu.cpu_events = nhm_events_attrs;
+
 		/* UOPS_ISSUED.STALLED_CYCLES */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
@@ -2050,6 +2068,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.extra_regs = intel_westmere_extra_regs;
 		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 
+		x86_pmu.cpu_events = nhm_events_attrs;
+
 		/* UOPS_ISSUED.STALLED_CYCLES */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
@@ -2078,6 +2098,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 		x86_pmu.er_flags |= ERF_NO_HT_SHARING;
 
+		x86_pmu.cpu_events = snb_events_attrs;
+
 		/* UOPS_ISSUED.ANY,c=1,i=1 to count stall cycles */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
@@ -2103,6 +2125,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 		x86_pmu.er_flags |= ERF_NO_HT_SHARING;
 
+		x86_pmu.cpu_events = snb_events_attrs;
+
 		/* UOPS_ISSUED.ANY,c=1,i=1 to count stall cycles */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
arch/x86/kernel/cpu/perf_event_intel_ds.c +125 -8

185 lines not on this page · raw message

#28 lore raw

[tip:perf/core] perf/x86: Add memory profiling via PEBS Load Latency

In reply to Stephane Eranian
perf/x86: Add memory profiling via PEBS Load Latency
Commit-ID:  f20093eef5f7843a25adfc0512617d4b1ff1aa6e
Gitweb:     http://git.kernel.org/tip/f20093eef5f7843a25adfc0512617d4b1ff1aa6e
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:32 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:16:31 -0300

perf/x86: Add memory profiling via PEBS Load Latency

This patch adds support for memory profiling using the PEBS Load Latency facility.

Load accesses are sampled by HW and the instruction address, data address, load latency, data source, tlb, locked information can be saved in the sampling buffer if using the PERF_SAMPLE_COST (for latency), PERF_SAMPLE_ADDR, PERF_SAMPLE_DATA_SRC types.

To enable PEBS Load Latency, users have to use the model specific event:

- on NHM/WSM: MEM_INST_RETIRED:LATENCY_ABOVE_THRESHOLD
- on SNB/IVB: MEM_TRANS_RETIRED:LATENCY_ABOVE_THRESHOLD

To make things easier, this patch also exports a generic alias via sysfs: mem-loads. It export the right event encoding based on the host CPU and can be used directly by the perf tool.

Loosely based on Intel's Lin Ming patch posted on LKML in July 2011.

---
 arch/x86/include/uapi/asm/msr-index.h     |   1 +
 arch/x86/kernel/cpu/perf_event.c          |   5 +-
 arch/x86/kernel/cpu/perf_event.h          |  25 +++++-
 arch/x86/kernel/cpu/perf_event_intel.c    |  24 ++++++
 arch/x86/kernel/cpu/perf_event_intel_ds.c | 133 ++++++++++++++++++++++++++++--
 5 files changed, 178 insertions(+), 10 deletions(-)
arch/x86/include/uapi/asm/msr-index.h +1 -0
index 892ce40..b31798d 100644
@@ -71,6 +71,7 @@
 #define MSR_IA32_PEBS_ENABLE		0x000003f1
 #define MSR_IA32_DS_AREA		0x00000600
 #define MSR_IA32_PERF_CAPABILITIES	0x00000345
+#define MSR_PEBS_LD_LAT_THRESHOLD	0x000003f6
 
 #define MSR_MTRRfix64K_00000		0x00000250
 #define MSR_MTRRfix16K_80000		0x00000258
arch/x86/kernel/cpu/perf_event.c +4 -1
index 8ba5151..5ed7a4c 100644
@@ -1363,7 +1363,7 @@ static __init struct attribute **merge_attr(struct attribute **a, struct attribu
 	return new;
 }
 
-static ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
+ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
 			  char *page)
 {
 	struct perf_pmu_events_attr *pmu_attr = \
@@ -1494,6 +1494,9 @@ static int __init init_hw_perf_events(void)
 	x86_pmu.attr_rdpmc = 1; /* enable userspace RDPMC usage by default */
 	x86_pmu_format_group.attrs = x86_pmu.format_attrs;
 
+	if (x86_pmu.event_attrs)
+		x86_pmu_events_group.attrs = x86_pmu.event_attrs;
+
 	if (!x86_pmu.events_sysfs_show)
 		x86_pmu_events_group.attrs = &empty_attrs;
 	else
arch/x86/kernel/cpu/perf_event.h +24 -1
index 9686d38..f3a9a94 100644
@@ -46,6 +46,7 @@ enum extra_reg_type {
 	EXTRA_REG_RSP_0 = 0,	/* offcore_response_0 */
 	EXTRA_REG_RSP_1 = 1,	/* offcore_response_1 */
 	EXTRA_REG_LBR   = 2,	/* lbr_select */
+	EXTRA_REG_LDLAT = 3,	/* ld_lat_threshold */
 
 	EXTRA_REG_MAX		/* number of entries needed */
 };
@@ -61,6 +62,10 @@ struct event_constraint {
 	int	overlap;
 	int	flags;
 };
+/*
+ * struct event_constraint flags
+ */
+#define PERF_X86_EVENT_PEBS_LDLAT	0x1 /* ld+ldlat data address sampling */
 
 struct amd_nb {
 	int nb_id;  /* NorthBridge id */
@@ -233,6 +238,10 @@ struct cpu_hw_events {
 #define INTEL_UEVENT_CONSTRAINT(c, n)	\
 	EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK)
 
+#define INTEL_PLD_CONSTRAINT(c, n)	\
+	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \
+			   HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LDLAT)
+
 #define EVENT_CONSTRAINT_END		\
 	EVENT_CONSTRAINT(0, 0, 0)
 
@@ -262,12 +271,22 @@ struct extra_reg {
 	.msr = (ms),		\
 	.config_mask = (m),	\
 	.valid_mask = (vm),	\
-	.idx = EXTRA_REG_##i	\
+	.idx = EXTRA_REG_##i,	\
 	}
 
 #define INTEL_EVENT_EXTRA_REG(event, msr, vm, idx)	\
 	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT, vm, idx)
 
+#define INTEL_UEVENT_EXTRA_REG(event, msr, vm, idx) \
+	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT | \
+			ARCH_PERFMON_EVENTSEL_UMASK, vm, idx)
+
+#define INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(c) \
+	INTEL_UEVENT_EXTRA_REG(c, \
+			       MSR_PEBS_LD_LAT_THRESHOLD, \
+			       0xffff, \
+			       LDLAT)
+
 #define EVENT_EXTRA_END EVENT_EXTRA_REG(0, 0, 0, 0, RSP_0)
 
 union perf_capabilities {
@@ -357,6 +376,7 @@ struct x86_pmu {
 	 */
 	int		attr_rdpmc;
 	struct attribute **format_attrs;
+	struct attribute **event_attrs;
 
 	ssize_t		(*events_sysfs_show)(char *page, u64 config);
 	struct attribute **cpu_events;
@@ -648,6 +668,9 @@ int p6_pmu_init(void);
 
 int knc_pmu_init(void);
 
+ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
+			  char *page);
+
 #else /* CONFIG_CPU_SUP_INTEL */
 
 static inline void reserve_ds_buffers(void)
arch/x86/kernel/cpu/perf_event_intel.c +24 -0
index df3beaa..d5ea5a0 100644
@@ -81,6 +81,7 @@ static struct event_constraint intel_nehalem_event_constraints[] __read_mostly =
 static struct extra_reg intel_nehalem_extra_regs[] __read_mostly =
 {
 	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0),
+	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x100b),
 	EVENT_EXTRA_END
 };
 
@@ -136,6 +137,7 @@ static struct extra_reg intel_westmere_extra_regs[] __read_mostly =
 {
 	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0),
 	INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0xffff, RSP_1),
+	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x100b),
 	EVENT_EXTRA_END
 };
 
@@ -155,9 +157,23 @@ static struct event_constraint intel_gen_event_constraints[] __read_mostly =
 static struct extra_reg intel_snb_extra_regs[] __read_mostly = {
 	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0x3fffffffffull, RSP_0),
 	INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0x3fffffffffull, RSP_1),
+	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x01cd),
 	EVENT_EXTRA_END
 };
 
+EVENT_ATTR_STR(mem-loads, mem_ld_nhm, "event=0x0b,umask=0x10,ldlat=3");
+EVENT_ATTR_STR(mem-loads, mem_ld_snb, "event=0xcd,umask=0x1,ldlat=3");
+
+struct attribute *nhm_events_attrs[] = {
+	EVENT_PTR(mem_ld_nhm),
+	NULL,
+};
+
+struct attribute *snb_events_attrs[] = {
+	EVENT_PTR(mem_ld_snb),
+	NULL,
+};
+
 static u64 intel_pmu_event_map(int hw_event)
 {
 	return intel_perfmon_event_map[hw_event];
@@ -2035,6 +2051,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.enable_all = intel_pmu_nhm_enable_all;
 		x86_pmu.extra_regs = intel_nehalem_extra_regs;
 
+		x86_pmu.cpu_events = nhm_events_attrs;
+
 		/* UOPS_ISSUED.STALLED_CYCLES */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
@@ -2078,6 +2096,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.extra_regs = intel_westmere_extra_regs;
 		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 
+		x86_pmu.cpu_events = nhm_events_attrs;
+
 		/* UOPS_ISSUED.STALLED_CYCLES */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
@@ -2106,6 +2126,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 		x86_pmu.er_flags |= ERF_NO_HT_SHARING;
 
+		x86_pmu.cpu_events = snb_events_attrs;
+
 		/* UOPS_ISSUED.ANY,c=1,i=1 to count stall cycles */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
@@ -2132,6 +2154,8 @@ __init int intel_pmu_init(void)
 		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 		x86_pmu.er_flags |= ERF_NO_HT_SHARING;
 
+		x86_pmu.cpu_events = snb_events_attrs;
+
 		/* UOPS_ISSUED.ANY,c=1,i=1 to count stall cycles */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] =
 			X86_CONFIG(.event=0x0e, .umask=0x01, .inv=1, .cmask=1);
arch/x86/kernel/cpu/perf_event_intel_ds.c +125 -8

185 lines not on this page · raw message

#29 lore raw

PATCH v7 9/18 perf/x86: export PEBS load latency threshold register to sysfs

In reply to Stephane Eranian

Make the PEBS Load Latency threshold register layout and encoding visible to user level tools.

---
 arch/x86/kernel/cpu/perf_event_intel.c |    3 +++
 1 file changed, 3 insertions(+)
arch/x86/kernel/cpu/perf_event_intel.c +3 -0
index f30027a..4ee1211 100644
@@ -1756,6 +1756,8 @@ static void intel_pmu_flush_branch_stack(void)
 
 PMU_FORMAT_ATTR(offcore_rsp, "config1:0-63");
 
+PMU_FORMAT_ATTR(ldlat, "config1:0-15");
+
 static struct attribute *intel_arch3_formats_attr[] = {
 	&format_attr_event.attr,
 	&format_attr_umask.attr,
@@ -1766,6 +1768,7 @@ static struct attribute *intel_arch3_formats_attr[] = {
 	&format_attr_cmask.attr,
 
 	&format_attr_offcore_rsp.attr, /* XXX do NHM/WSM + SNB breakout */
+	&format_attr_ldlat.attr, /* PEBS load latency */
 	NULL,
 };
 
Signature · 1 line
1.7.9.5
#30 lore raw

[tip:perf/x86] perf/x86: Export PEBS load latency threshold register to sysfs

Commit-ID:  5a54b18a9e9c77ebd1d30d4544e4f936fcafa6b9
Gitweb:     http://git.kernel.org/tip/5a54b18a9e9c77ebd1d30d4544e4f936fcafa6b9
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:33 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 25 Jan 2013 10:19:04 +0100

perf/x86: Export PEBS load latency threshold register to sysfs

Make the PEBS Load Latency threshold register layout and encoding visible to user level tools.

---
 arch/x86/kernel/cpu/perf_event_intel.c | 3 +++
 1 file changed, 3 insertions(+)
arch/x86/kernel/cpu/perf_event_intel.c +3 -0
index f30027a..4ee1211 100644
@@ -1756,6 +1756,8 @@ static void intel_pmu_flush_branch_stack(void)
 
 PMU_FORMAT_ATTR(offcore_rsp, "config1:0-63");
 
+PMU_FORMAT_ATTR(ldlat, "config1:0-15");
+
 static struct attribute *intel_arch3_formats_attr[] = {
 	&format_attr_event.attr,
 	&format_attr_umask.attr,
@@ -1766,6 +1768,7 @@ static struct attribute *intel_arch3_formats_attr[] = {
 	&format_attr_cmask.attr,
 
 	&format_attr_offcore_rsp.attr, /* XXX do NHM/WSM + SNB breakout */
+	&format_attr_ldlat.attr, /* PEBS load latency */
 	NULL,
 };
 
#31 lore raw

[tip:perf/core] perf/x86: Export PEBS load latency threshold register to sysfs

In reply to Stephane Eranian
perf/x86: Export PEBS load latency threshold register to sysfs
Commit-ID:  a63fcab45273174e665e6a8c9fa1a79a9046d0d5
Gitweb:     http://git.kernel.org/tip/a63fcab45273174e665e6a8c9fa1a79a9046d0d5
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:33 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:16:49 -0300

perf/x86: Export PEBS load latency threshold register to sysfs

Make the PEBS Load Latency threshold register layout and encoding visible to user level tools.

---
 arch/x86/kernel/cpu/perf_event_intel.c | 3 +++
 1 file changed, 3 insertions(+)
arch/x86/kernel/cpu/perf_event_intel.c +3 -0
index d5ea5a0..ae6096b 100644
@@ -1781,6 +1781,8 @@ static void intel_pmu_flush_branch_stack(void)
 
 PMU_FORMAT_ATTR(offcore_rsp, "config1:0-63");
 
+PMU_FORMAT_ATTR(ldlat, "config1:0-15");
+
 static struct attribute *intel_arch3_formats_attr[] = {
 	&format_attr_event.attr,
 	&format_attr_umask.attr,
@@ -1791,6 +1793,7 @@ static struct attribute *intel_arch3_formats_attr[] = {
 	&format_attr_cmask.attr,
 
 	&format_attr_offcore_rsp.attr, /* XXX do NHM/WSM + SNB breakout */
+	&format_attr_ldlat.attr, /* PEBS load latency */
 	NULL,
 };
 
#32 lore raw

PATCH v7 10/18 perf/x86: add support for PEBS Precise Store

In reply to Stephane Eranian

This patch adds support for PEBS Precise Store which is available on Intel Sandy Bridge and Ivy Bridge processors.

To use Precise store, the proper PEBS event must be used: mem_trans_retired:precise_stores. For the perf tool, the generic mem-stores event exported via sysfs can be used directly.

---
 arch/x86/kernel/cpu/perf_event.h          |    5 +++
 arch/x86/kernel/cpu/perf_event_intel.c    |    2 ++
 arch/x86/kernel/cpu/perf_event_intel_ds.c |   49 +++++++++++++++++++++++++++--
 3 files changed, 54 insertions(+), 2 deletions(-)
arch/x86/kernel/cpu/perf_event.h +5 -0
index 3f91411..645b864 100644
@@ -66,6 +66,7 @@ struct event_constraint {
  * struct event_constraint flags
  */
 #define PERF_X86_EVENT_PEBS_LDLAT	0x1 /* ld+ldlat data address sampling */
+#define PERF_X86_EVENT_PEBS_ST		0x2 /* st data address sampling */
 
 struct amd_nb {
 	int nb_id;  /* NorthBridge id */
@@ -242,6 +243,10 @@ struct cpu_hw_events {
 	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \
 			   HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LDLAT)
 
+#define INTEL_PST_CONSTRAINT(c, n)	\
+	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \
+			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_ST)
+
 #define EVENT_CONSTRAINT_END		\
 	EVENT_CONSTRAINT(0, 0, 0)
 
arch/x86/kernel/cpu/perf_event_intel.c +2 -0
index 4ee1211..9d0d036 100644
@@ -138,6 +138,7 @@ static struct extra_reg intel_snb_extra_regs[] __read_mostly = {
 
 EVENT_ATTR_STR(mem-loads, mem_ld_nhm, "event=0x0b,umask=0x10,ldlat=3");
 EVENT_ATTR_STR(mem-loads, mem_ld_snb, "event=0xcd,umask=0x1,ldlat=3");
+EVENT_ATTR_STR(mem-stores, mem_st_snb, "event=0xcd,umask=0x2");
 
 struct attribute *nhm_events_attrs[] = {
 	EVENT_PTR(mem_ld_nhm),
@@ -146,6 +147,7 @@ struct attribute *nhm_events_attrs[] = {
 
 struct attribute *snb_events_attrs[] = {
 	EVENT_PTR(mem_ld_snb),
+	EVENT_PTR(mem_st_snb),
 	NULL,
 };
 
arch/x86/kernel/cpu/perf_event_intel_ds.c +47 -2

96 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#33 lore raw

[tip:perf/x86] perf/x86: Add support for PEBS Precise Store

Commit-ID:  6c538c1cc9f0049803227c6ca1d1528edf397589
Gitweb:     http://git.kernel.org/tip/6c538c1cc9f0049803227c6ca1d1528edf397589
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:34 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 25 Jan 2013 10:19:05 +0100

perf/x86: Add support for PEBS Precise Store

This patch adds support for PEBS Precise Store which is available on Intel Sandy Bridge and Ivy Bridge processors.

To use Precise store, the proper PEBS event must be used: mem_trans_retired:precise_stores. For the perf tool, the generic mem-stores event exported via sysfs can be used directly.

---
 arch/x86/kernel/cpu/perf_event.h          |  5 ++++
 arch/x86/kernel/cpu/perf_event_intel.c    |  2 ++
 arch/x86/kernel/cpu/perf_event_intel_ds.c | 49 +++++++++++++++++++++++++++++--
 3 files changed, 54 insertions(+), 2 deletions(-)
arch/x86/kernel/cpu/perf_event.h +5 -0
index 8f49e45..158f46b 100644
@@ -66,6 +66,7 @@ struct event_constraint {
  * struct event_constraint flags
  */
 #define PERF_X86_EVENT_PEBS_LDLAT	0x1 /* ld+ldlat data address sampling */
+#define PERF_X86_EVENT_PEBS_ST		0x2 /* st data address sampling */
 
 struct amd_nb {
 	int nb_id;  /* NorthBridge id */
@@ -242,6 +243,10 @@ struct cpu_hw_events {
 	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \
 			   HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LDLAT)
 
+#define INTEL_PST_CONSTRAINT(c, n)	\
+	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \
+			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_ST)
+
 #define EVENT_CONSTRAINT_END		\
 	EVENT_CONSTRAINT(0, 0, 0)
 
arch/x86/kernel/cpu/perf_event_intel.c +2 -0
index 4ee1211..9d0d036 100644
@@ -138,6 +138,7 @@ static struct extra_reg intel_snb_extra_regs[] __read_mostly = {
 
 EVENT_ATTR_STR(mem-loads, mem_ld_nhm, "event=0x0b,umask=0x10,ldlat=3");
 EVENT_ATTR_STR(mem-loads, mem_ld_snb, "event=0xcd,umask=0x1,ldlat=3");
+EVENT_ATTR_STR(mem-stores, mem_st_snb, "event=0xcd,umask=0x2");
 
 struct attribute *nhm_events_attrs[] = {
 	EVENT_PTR(mem_ld_nhm),
@@ -146,6 +147,7 @@ struct attribute *nhm_events_attrs[] = {
 
 struct attribute *snb_events_attrs[] = {
 	EVENT_PTR(mem_ld_snb),
+	EVENT_PTR(mem_st_snb),
 	NULL,
 };
 
arch/x86/kernel/cpu/perf_event_intel_ds.c +47 -2

96 lines not on this page · raw message

#34 lore raw

[tip:perf/core] perf/x86: Add support for PEBS Precise Store

In reply to Stephane Eranian
perf/x86: Add support for PEBS Precise Store
Commit-ID:  9ad64c0f481c37a63dd39842a0fd264bee44a097
Gitweb:     http://git.kernel.org/tip/9ad64c0f481c37a63dd39842a0fd264bee44a097
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:34 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:17:06 -0300

perf/x86: Add support for PEBS Precise Store

This patch adds support for PEBS Precise Store which is available on Intel Sandy Bridge and Ivy Bridge processors.

To use Precise store, the proper PEBS event must be used: mem_trans_retired:precise_stores. For the perf tool, the generic mem-stores event exported via sysfs can be used directly.

---
 arch/x86/kernel/cpu/perf_event.h          |  5 ++++
 arch/x86/kernel/cpu/perf_event_intel.c    |  2 ++
 arch/x86/kernel/cpu/perf_event_intel_ds.c | 49 +++++++++++++++++++++++++++++--
 3 files changed, 54 insertions(+), 2 deletions(-)
arch/x86/kernel/cpu/perf_event.h +5 -0
index f3a9a94..ba9aadf 100644
@@ -66,6 +66,7 @@ struct event_constraint {
  * struct event_constraint flags
  */
 #define PERF_X86_EVENT_PEBS_LDLAT	0x1 /* ld+ldlat data address sampling */
+#define PERF_X86_EVENT_PEBS_ST		0x2 /* st data address sampling */
 
 struct amd_nb {
 	int nb_id;  /* NorthBridge id */
@@ -242,6 +243,10 @@ struct cpu_hw_events {
 	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \
 			   HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LDLAT)
 
+#define INTEL_PST_CONSTRAINT(c, n)	\
+	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK, \
+			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_ST)
+
 #define EVENT_CONSTRAINT_END		\
 	EVENT_CONSTRAINT(0, 0, 0)
 
arch/x86/kernel/cpu/perf_event_intel.c +2 -0
index ae6096b..e84c4ba 100644
@@ -163,6 +163,7 @@ static struct extra_reg intel_snb_extra_regs[] __read_mostly = {
 
 EVENT_ATTR_STR(mem-loads, mem_ld_nhm, "event=0x0b,umask=0x10,ldlat=3");
 EVENT_ATTR_STR(mem-loads, mem_ld_snb, "event=0xcd,umask=0x1,ldlat=3");
+EVENT_ATTR_STR(mem-stores, mem_st_snb, "event=0xcd,umask=0x2");
 
 struct attribute *nhm_events_attrs[] = {
 	EVENT_PTR(mem_ld_nhm),
@@ -171,6 +172,7 @@ struct attribute *nhm_events_attrs[] = {
 
 struct attribute *snb_events_attrs[] = {
 	EVENT_PTR(mem_ld_snb),
+	EVENT_PTR(mem_st_snb),
 	NULL,
 };
 
arch/x86/kernel/cpu/perf_event_intel_ds.c +47 -2

96 lines not on this page · raw message

#35 lore raw

PATCH v7 11/18 perf tools: add mem access sampling core support

In reply to Stephane Eranian

This patch adds the sorting and histogram support
functions to enable profiling of memory accesses.

The following sorting orders are added:
- symbol_daddr: data address symbol (or raw address)
- dso_daddr: data address shared object
- locked: access uses locked transaction
- tlb : TLB access
- mem : memory level of the access (L1, L2, L3, RAM, ...)
- snoop: access snoop mode

---
 tools/perf/util/event.h   |    1 +
 tools/perf/util/evsel.c   |    6 +
 tools/perf/util/hist.c    |   77 +++++++++++-
 tools/perf/util/hist.h    |   13 ++
 tools/perf/util/session.c |   38 ++++++
 tools/perf/util/session.h |    4 +
 tools/perf/util/sort.c    |  297 ++++++++++++++++++++++++++++++++++++++++++++-
 tools/perf/util/sort.h    |    9 +-
 tools/perf/util/symbol.h  |    6 +
 9 files changed, 441 insertions(+), 10 deletions(-)
tools/perf/util/event.h +1 -0
index a97fbbe..4008b7f 100644
@@ -91,6 +91,7 @@ struct perf_sample {
 	u64 weight;
 	u32 cpu;
 	u32 raw_size;
+	u64 dsrc;
 	void *raw_data;
 	struct ip_callchain *callchain;
 	struct branch_stack *branch_stack;
tools/perf/util/evsel.c +6 -0
index 805d33e..49daa7b 100644
@@ -1068,6 +1068,12 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
 		array++;
 	}
 
+	data->dsrc = 0;
+	if (type & PERF_SAMPLE_DSRC) {
+		data->dsrc = *array;
+		array++;
+	}
+
 	return 0;
 }
 
tools/perf/util/hist.c +74 -3

137 lines not on this page · raw message

tools/perf/util/hist.h +13 -0
index c769984a..012fb99 100644
@@ -51,6 +51,12 @@ enum hist_column {
 	HISTC_SRCLINE,
 	HISTC_LOCAL_WEIGHT,
 	HISTC_GLOBAL_WEIGHT,
+	HISTC_MEM_DADDR_SYMBOL,
+	HISTC_MEM_DADDR_DSO,
+	HISTC_MEM_LOCKED,
+	HISTC_MEM_TLB,
+	HISTC_MEM_LVL,
+	HISTC_MEM_SNOOP,
 	HISTC_NR_COLS, /* Last entry */
 };
 
@@ -90,6 +96,13 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 					     u64 period,
 					     u64 weight);
 
+struct hist_entry *__hists__add_mem_entry(struct hists *self,
+					  struct addr_location *al,
+					  struct symbol *sym_parent,
+					  struct mem_info *mi,
+					  u64 period,
+					  u64 weight);
+
 void hists__output_resort(struct hists *self);
 void hists__output_resort_threaded(struct hists *hists);
 void hists__collapse_resort(struct hists *self);
tools/perf/util/session.c +38 -0
index 8fe3688..698be69 100644
@@ -273,6 +273,41 @@ static void ip__resolve_ams(struct machine *self, struct thread *thread,
 	ams->map = al.map;
 }
 
+static void ip__resolve_data(struct machine *self, struct thread *thread,
+			     u8 m,
+			    struct addr_map_symbol *ams,
+			    u64 addr)
+{
+	struct addr_location al;
+
+	memset(&al, 0, sizeof(al));
+
+	thread__find_addr_location(thread, self, m, MAP__VARIABLE, addr, &al,
+				   NULL);
+	ams->addr = addr;
+	ams->al_addr = al.addr;
+	ams->sym = al.sym;
+	ams->map = al.map;
+}
+
+struct mem_info *machine__resolve_mem(struct machine *self,
+				      struct thread *thr,
+				      struct perf_sample *sample,
+				      u8 cpumode)
+{
+	struct mem_info *mi;
+
+	mi = calloc(1, sizeof(struct mem_info));
+	if (!mi)
+		return NULL;
+
+	ip__resolve_ams(self, thr, &mi->iaddr, sample->ip);
+	ip__resolve_data(self, thr, cpumode, &mi->daddr, sample->addr);
+	mi->dsrc.val = sample->dsrc;
+
+	return mi;
+}
+
 struct branch_info *machine__resolve_bstack(struct machine *self,
 					    struct thread *thr,
 					    struct branch_stack *bs)
@@ -1012,6 +1047,9 @@ static void dump_sample(struct perf_evsel *evsel, union perf_event *event,
 
 	if (sample_type & PERF_SAMPLE_ADDR)
 		printf(" ..... data: 0x%"PRIx64"\n", sample->addr);
+
+	if (sample_type & PERF_SAMPLE_DSRC)
+		printf(" . data_src: 0x%"PRIx64"\n", sample->dsrc);
 }
 
 static struct machine *
tools/perf/util/session.h +4 -0
index cea133a..f3ea026 100644
@@ -69,6 +69,10 @@ int perf_session__resolve_callchain(struct perf_session *self, struct perf_evsel
 				    struct ip_callchain *chain,
 				    struct symbol **parent);
 
+struct mem_info *machine__resolve_mem(struct machine *self,
+				      struct thread *thread,
+				      struct perf_sample *sample, u8 cpumode);
+
 bool perf_session__has_traces(struct perf_session *self, const char *msg);
 
 void mem_bswap_64(void *src, int byte_size);
tools/perf/util/sort.c +291 -6

333 lines not on this page · raw message

tools/perf/util/sort.h +8 -1
index 9af5446..0184f1c 100644
@@ -104,7 +104,8 @@ struct hist_entry {
 	struct rb_root		sorted_chain;
 	struct branch_info	*branch_info;
 	struct hists		*hists;
-	struct callchain_root	callchain[0];
+	struct mem_info		*mem_info;
+	struct callchain_root	callchain[0]; /* must be last member */
 };
 
 static inline bool hist_entry__has_pairs(struct hist_entry *he)
@@ -140,6 +141,12 @@ enum sort_type {
 	SORT_SRCLINE,
 	SORT_LOCAL_WEIGHT,
 	SORT_GLOBAL_WEIGHT,
+	SORT_MEM_DADDR_SYMBOL,
+	SORT_MEM_DADDR_DSO,
+	SORT_MEM_LOCKED,
+	SORT_MEM_TLB,
+	SORT_MEM_LVL,
+	SORT_MEM_SNOOP,
 };
 
 /*
tools/perf/util/symbol.h +6 -0
index de68f98..d0dab17 100644
@@ -152,6 +152,12 @@ struct branch_info {
 	struct branch_flags flags;
 };
 
+struct mem_info {
+	struct addr_map_symbol iaddr;
+	struct addr_map_symbol daddr;
+	union perf_mem_dsrc dsrc;
+};
+
 struct addr_location {
 	struct thread *thread;
 	struct map    *map;
Signature · 1 line
1.7.9.5
#36 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

On Thu, Jan 24, 2013 at 04:10:35PM +0100, Stephane Eranian wrote:

SNIP

Quoted 21 lines
tools/perf/util/session.c +20 -0
+static void ip__resolve_data(struct machine *self, struct thread *thread,
+			     u8 m,
+			    struct addr_map_symbol *ams,
+			    u64 addr)
+{
+	struct addr_location al;
+
+	memset(&al, 0, sizeof(al));
+
+	thread__find_addr_location(thread, self, m, MAP__VARIABLE, addr, &al,
+				   NULL);
+	ams->addr = addr;
+	ams->al_addr = al.addr;
+	ams->sym = al.sym;
+	ams->map = al.map;
+}
+
+struct mem_info *machine__resolve_mem(struct machine *self,
+				      struct thread *thr,
+				      struct perf_sample *sample,

tools/perf/util/session.c static void ip__resolve_ams(struct machine *self, struct thread *thread, · line 305

Quoted 10 lines
tools/perf/util/session.c +10 -0
+				      u8 cpumode)
+{
+	struct mem_info *mi;
+
+	mi = calloc(1, sizeof(struct mem_info));
+	if (!mi)
+		return NULL;
+
+	ip__resolve_ams(self, thr, &mi->iaddr, sample->ip);
+	ip__resolve_data(self, thr, cpumode, &mi->daddr, sample->addr);

question, should this be the other way around? like:

	ip__resolve_ams(machine, thr, &mi->daddr, sample->addr);
	ip__resolve_data(machine, thr, cpumode, &mi->iaddr, sample->ip);

we have correct cpumode for sample->ip, but I think it's the PEBS->dla (sample->addr) where we need to guess.. right?

that makes me think that we could probably use ip__resolve_data for both.. hummm.. but we could access data cross the user/kernel boundary, so cpumode would be different for ip and accessed data

thanks,
jirka

maintainer
#37 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

Quoted 2 lines · Jiri Olsa

we have correct cpumode for sample->ip, but I think it's the PEBS->dla (sample->addr) where we need to guess.. right?

kernel mode very much fakes the cpumode/segment stuff for PEBS. PEBS assumes you're running in a linear/flat mode.

#38 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

Quoted 7 lines · Peter Zijlstra
Quoted 2 lines · Jiri Olsa

we have correct cpumode for sample->ip, but I think it's the PEBS->dla (sample->addr) where we need to guess.. right?

kernel mode very much fakes the cpumode/segment stuff for PEBS. PEBS assumes you're running in a linear/flat mode.

say we hit the sample when kernel accesses the user data, we will endup with IP in kernel space and DATA ptr in user space.. in theory ;-)

and that would need the cpumode guessing for DATA ptr, because cpumode value is deduced from cs register

jirka

#39 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

Quoted 6 lines · Jiri Olsa
Quoted 4 lines · Peter Zijlstra
Quoted 2 lines · Jiri Olsa

we have correct cpumode for sample->ip, but I think it's the PEBS->dla (sample->addr) where we need to guess.. right?

Quoted 7 lines
Quoted 3 lines

kernel mode very much fakes the cpumode/segment stuff for PEBS. PEBS assumes you're running in a linear/flat mode.

say we hit the sample when kernel accesses the user data, we will endup with IP in kernel space and DATA ptr in user space.. in theory ;-)

Yes, this is possible. So I think we could probably leaverage ip__resolve_ams() and pass the extra parameter for MAP_VARIABLE vs. MAP_FUNCTION.

Quoted 5 lines

and that would need the cpumode guessing for DATA ptr, because cpumode value is deduced from cs register

jirka

#40 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

Em Wed, Mar 27, 2013 at 03:48:15PM +0100, Stephane Eranian escreveu:

Quoted 6 lines · Jiri Olsa
Quoted 4 lines · Peter Zijlstra
Quoted 3 lines · Jiri Olsa
Quoted 2 lines

we have correct cpumode for sample->ip, but I think it's the PEBS->dla (sample->addr) where we need to guess.. right?

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

kernel mode very much fakes the cpumode/segment stuff for PEBS. PEBS assumes you're running in a linear/flat mode.

say we hit the sample when kernel accesses the user data, we will endup with IP in kernel space and DATA ptr in user space.. in theory ;-)

Yes, this is possible. So I think we could probably leaverage ip__resolve_ams() and pass the extra parameter for MAP_VARIABLE vs. MAP_FUNCTION.

BTW, I fixed up the patches (kernel and user parts) and have them in perf/mem at:

git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux

Hope to push it to Ingo today/tomorrow, after some more testing, thanks Jiri for reviewing it.

Stephane, if you could give it a try again to see that the fixups I did (documented in the commit logs, just before my Signed-off-by) are ok, that would be good.

- Arnaldo

Quoted 6 lines
Quoted 6 lines

and that would need the cpumode guessing for DATA ptr, because cpumode value is deduced from cs register

jirka

#41 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

Arnaldo,

Quoted 19 lines · Arnaldo Carvalho de Melo

Em Wed, Mar 27, 2013 at 03:48:15PM +0100, Stephane Eranian escreveu:

Quoted 6 lines · Jiri Olsa
Quoted 4 lines · Peter Zijlstra
Quoted 3 lines · Jiri Olsa
Quoted 2 lines

we have correct cpumode for sample->ip, but I think it's the PEBS->dla (sample->addr) where we need to guess.. right?

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

kernel mode very much fakes the cpumode/segment stuff for PEBS. PEBS assumes you're running in a linear/flat mode.

say we hit the sample when kernel accesses the user data, we will endup with IP in kernel space and DATA ptr in user space.. in theory ;-)

Yes, this is possible. So I think we could probably leaverage ip__resolve_ams() and pass the extra parameter for MAP_VARIABLE vs. MAP_FUNCTION.

BTW, I fixed up the patches (kernel and user parts) and have them in perf/mem at:

Quoted 9 lines

git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux

Hope to push it to Ingo today/tomorrow, after some more testing, thanks Jiri for reviewing it.

Stephane, if you could give it a try again to see that the fixups I did (documented in the commit logs, just before my Signed-off-by) are ok, that would be good.

I tried on a few examples on both NHM (only loads, no TLB) and SNB and got the right answers for my tests, including data symbol resolution.

What we discussed with Jiri yesterday can be added later on.

Thanks for the integration work. Looks good to me.

Quoted 8 lines

- Arnaldo

Quoted 6 lines
Quoted 6 lines

and that would need the cpumode guessing for DATA ptr, because cpumode value is deduced from cs register

jirka

#42 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

Em Thu, Mar 28, 2013 at 03:24:30PM +0100, Stephane Eranian escreveu:

Quoted 11 lines

On Wed, Mar 27, 2013 at 5:56 PM, Arnaldo Carvalho de Melo

Quoted 3 lines

Stephane, if you could give it a try again to see that the fixups I did (documented in the commit logs, just before my Signed-off-by) are ok, that would be good.

I tried on a few examples on both NHM (only loads, no TLB) and SNB and got the right answers for my tests, including data symbol resolution.

What we discussed with Jiri yesterday can be added later on.

Thanks for the integration work. Looks good to me.

Humm, I just tried it with a simple:

perf mem -t load rec

And got an OOPS, trying again, and this machine was suspended, perhaps perf/core doesn't have that PEBS fix, will check.

- Arnaldo

#43 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

Quoted 9 lines · Arnaldo Carvalho de Melo

Em Thu, Mar 28, 2013 at 03:24:30PM +0100, Stephane Eranian escreveu:

Quoted 7 lines

On Wed, Mar 27, 2013 at 5:56 PM, Arnaldo Carvalho de Melo

Quoted 3 lines

Stephane, if you could give it a try again to see that the fixups I did (documented in the commit logs, just before my Signed-off-by) are ok, that would be good.

I tried on a few examples on both NHM (only loads, no TLB) and SNB and got the right answers for my tests, including data symbol resolution.

Quoted 8 lines
Quoted 3 lines

What we discussed with Jiri yesterday can be added later on.

Thanks for the integration work. Looks good to me.

Humm, I just tried it with a simple:

perf mem -t load rec

That cmdline should not product anything. It is missing a command to run?

Quoted 3 lines

And got an OOPS, trying again, and this machine was suspended, perhaps perf/core doesn't have that PEBS fix, will check.

What HW is this running on?
I was running with your kernel.

#44 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

In reply to Arnaldo Carvalho de Melo

Em Thu, Mar 28, 2013 at 12:00:18PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 11 lines

Em Thu, Mar 28, 2013 at 03:24:30PM +0100, Stephane Eranian escreveu:

Quoted 9 lines

On Wed, Mar 27, 2013 at 5:56 PM, Arnaldo Carvalho de Melo

Quoted 3 lines

Stephane, if you could give it a try again to see that the fixups I did (documented in the commit logs, just before my Signed-off-by) are ok, that would be good.

I tried on a few examples on both NHM (only loads, no TLB) and SNB and got the right answers for my tests, including data symbol resolution.

What we discussed with Jiri yesterday can be added later on.

Quoted 8 lines
Quoted 1 line

Thanks for the integration work. Looks good to me.

Humm, I just tried it with a simple:

perf mem -t load rec

And got an OOPS, trying again, and this machine was suspended, perhaps perf/core doesn't have that PEBS fix, will check.

Yeah, after a fresh reboot it doesn't OOPses, the fix:

commit 1d9d8639c063caf6efc2447f5f26aa637f844ff6

Author: Stephane Eranian <[email protected]>
Date:   Fri Mar 15 14:26:07 2013 +0100

perf,x86: fix kernel crash with PEBS/BTS after suspend/resume

--------

Isn't in perf/core, cool, before your test results I thougt I had messed up something :-)

- Arnaldo

#45 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

Quoted 21 lines · Arnaldo Carvalho de Melo

Em Thu, Mar 28, 2013 at 12:00:18PM -0300, Arnaldo Carvalho de Melo escreveu:

Quoted 11 lines

Em Thu, Mar 28, 2013 at 03:24:30PM +0100, Stephane Eranian escreveu:

Quoted 9 lines

On Wed, Mar 27, 2013 at 5:56 PM, Arnaldo Carvalho de Melo

Quoted 3 lines

Stephane, if you could give it a try again to see that the fixups I did (documented in the commit logs, just before my Signed-off-by) are ok, that would be good.

I tried on a few examples on both NHM (only loads, no TLB) and SNB and got the right answers for my tests, including data symbol resolution.

What we discussed with Jiri yesterday can be added later on.

Quoted 8 lines
Quoted 1 line

Thanks for the integration work. Looks good to me.

Humm, I just tried it with a simple:

perf mem -t load rec

And got an OOPS, trying again, and this machine was suspended, perhaps perf/core doesn't have that PEBS fix, will check.

Quoted 8 lines

Yeah, after a fresh reboot it doesn't OOPses, the fix:

commit 1d9d8639c063caf6efc2447f5f26aa637f844ff6

Author: Stephane Eranian <[email protected]>
Date:   Fri Mar 15 14:26:07 2013 +0100

perf,x86: fix kernel crash with PEBS/BTS after suspend/resume

Okay, so you too fell into that trap! I mean laptop guys, trying to save battery, yet running perf, c'mon... ;->

Quoted 5 lines

--------

Isn't in perf/core, cool, before your test results I thougt I had messed up something :-)

yeah, closed your laptop lid.....
Should work better with the suspend/resume fix now, hopefully. Of course, on SNB systems, you also need that firmware patch to enable PEBS, IIRC.

#46 lore raw

Re: [PATCH v7 11/18] perf tools: add mem access sampling core support

In reply to Jiri Olsa
Quoted 38 lines · Jiri Olsa

On Thu, Jan 24, 2013 at 04:10:35PM +0100, Stephane Eranian wrote:

SNIP

Quoted 21 lines
tools/perf/util/session.c +20 -0
+static void ip__resolve_data(struct machine *self, struct thread *thread,
+			     u8 m,
+			    struct addr_map_symbol *ams,
+			    u64 addr)
+{
+	struct addr_location al;
+
+	memset(&al, 0, sizeof(al));
+
+	thread__find_addr_location(thread, self, m, MAP__VARIABLE, addr, &al,
+				   NULL);
+	ams->addr = addr;
+	ams->al_addr = al.addr;
+	ams->sym = al.sym;
+	ams->map = al.map;
+}
+
+struct mem_info *machine__resolve_mem(struct machine *self,
+				      struct thread *thr,
+				      struct perf_sample *sample,

tools/perf/util/session.c static void ip__resolve_ams(struct machine *self, struct thread *thread, · line 305

Quoted 10 lines
tools/perf/util/session.c +10 -0
+				      u8 cpumode)
+{
+	struct mem_info *mi;
+
+	mi = calloc(1, sizeof(struct mem_info));
+	if (!mi)
+		return NULL;
+
+	ip__resolve_ams(self, thr, &mi->iaddr, sample->ip);
+	ip__resolve_data(self, thr, cpumode, &mi->daddr, sample->addr);

question, should this be the other way around? like:

Quoted 2 lines
	ip__resolve_ams(machine, thr, &mi->daddr, sample->addr);
	ip__resolve_data(machine, thr, cpumode, &mi->iaddr, sample->ip);

ugh, I missed the MAP__VARIABLE/MAP__FUNCTION difference there, thanks Arnaldo! ;-)

still, no need to guess the cpumode for ip and guess it for data?

jirka

#47 lore raw

[tip:perf/core] perf tools: Add mem access sampling core support

In reply to Stephane Eranian
perf tools: Add mem access sampling core support
Commit-ID:  98a3b32c99ada4bca8aaf4f91efd96fc906dd5c4
Gitweb:     http://git.kernel.org/tip/98a3b32c99ada4bca8aaf4f91efd96fc906dd5c4
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:35 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:20:13 -0300

perf tools: Add mem access sampling core support

This patch adds the sorting and histogram support
functions to enable profiling of memory accesses.

The following sorting orders are added:
- symbol_daddr: data address symbol (or raw address)
- dso_daddr: data address shared object
- locked: access uses locked transaction
- tlb : TLB access
- mem : memory level of the access (L1, L2, L3, RAM, ...)
- snoop: access snoop mode

[ committer note: changed to cope with fc5871ed, the move of methods to
machine.[ch], and the rename of dsrc to data_src, to match the change made in the PERF_SAMPLE_DSRC in a previous patch. ]

---
 tools/perf/util/event.h   |   8 +
 tools/perf/util/evsel.c   |   6 +
 tools/perf/util/hist.c    |  86 ++++++++++-
 tools/perf/util/hist.h    |  13 ++
 tools/perf/util/machine.c |  32 ++++
 tools/perf/util/machine.h |   3 +
 tools/perf/util/session.c |   3 +
 tools/perf/util/sort.c    | 369 +++++++++++++++++++++++++++++++++++++++++++++-
 tools/perf/util/sort.h    |   9 +-
 tools/perf/util/symbol.h  |   6 +
 10 files changed, 525 insertions(+), 10 deletions(-)
tools/perf/util/event.h +8 -0
index a97fbbe..1813895 100644
@@ -91,6 +91,7 @@ struct perf_sample {
 	u64 weight;
 	u32 cpu;
 	u32 raw_size;
+	u64 data_src;
 	void *raw_data;
 	struct ip_callchain *callchain;
 	struct branch_stack *branch_stack;
@@ -98,6 +99,13 @@ struct perf_sample {
 	struct stack_dump user_stack;
 };
 
+#define PERF_MEM_DATA_SRC_NONE \
+	(PERF_MEM_S(OP, NA) |\
+	 PERF_MEM_S(LVL, NA) |\
+	 PERF_MEM_S(SNOOP, NA) |\
+	 PERF_MEM_S(LOCK, NA) |\
+	 PERF_MEM_S(TLB, NA))
+
 struct build_id_event {
 	struct perf_event_header header;
 	pid_t			 pid;
tools/perf/util/evsel.c +6 -0
index 23061a6..5c4ca51 100644
@@ -1177,6 +1177,12 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
 		array++;
 	}
 
+	data->data_src = PERF_MEM_DATA_SRC_NONE;
+	if (type & PERF_SAMPLE_DATA_SRC) {
+		data->data_src = *array;
+		array++;
+	}
+
 	return 0;
 }
 
tools/perf/util/hist.c +83 -3

145 lines not on this page · raw message

tools/perf/util/hist.h +13 -0
index 121cc14..fd63134 100644
@@ -51,6 +51,12 @@ enum hist_column {
 	HISTC_SRCLINE,
 	HISTC_LOCAL_WEIGHT,
 	HISTC_GLOBAL_WEIGHT,
+	HISTC_MEM_DADDR_SYMBOL,
+	HISTC_MEM_DADDR_DSO,
+	HISTC_MEM_LOCKED,
+	HISTC_MEM_TLB,
+	HISTC_MEM_LVL,
+	HISTC_MEM_SNOOP,
 	HISTC_NR_COLS, /* Last entry */
 };
 
@@ -90,6 +96,13 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 					     u64 period,
 					     u64 weight);
 
+struct hist_entry *__hists__add_mem_entry(struct hists *self,
+					  struct addr_location *al,
+					  struct symbol *sym_parent,
+					  struct mem_info *mi,
+					  u64 period,
+					  u64 weight);
+
 void hists__output_resort(struct hists *self);
 void hists__output_resort_threaded(struct hists *hists);
 void hists__collapse_resort(struct hists *self);
tools/perf/util/machine.c +32 -0
index c5e3b12..d77ba86 100644
@@ -1097,6 +1097,38 @@ found:
 	ams->map = al.map;
 }
 
+static void ip__resolve_data(struct machine *machine, struct thread *thread,
+			     u8 m, struct addr_map_symbol *ams, u64 addr)
+{
+	struct addr_location al;
+
+	memset(&al, 0, sizeof(al));
+
+	thread__find_addr_location(thread, machine, m, MAP__VARIABLE, addr, &al,
+				   NULL);
+	ams->addr = addr;
+	ams->al_addr = al.addr;
+	ams->sym = al.sym;
+	ams->map = al.map;
+}
+
+struct mem_info *machine__resolve_mem(struct machine *machine,
+				      struct thread *thr,
+				      struct perf_sample *sample,
+				      u8 cpumode)
+{
+	struct mem_info *mi = zalloc(sizeof(*mi));
+
+	if (!mi)
+		return NULL;
+
+	ip__resolve_ams(machine, thr, &mi->iaddr, sample->ip);
+	ip__resolve_data(machine, thr, cpumode, &mi->daddr, sample->addr);
+	mi->data_src.val = sample->data_src;
+
+	return mi;
+}
+
 struct branch_info *machine__resolve_bstack(struct machine *machine,
 					    struct thread *thr,
 					    struct branch_stack *bs)
tools/perf/util/machine.h +3 -0
index e0b2c00..7794068 100644
@@ -76,6 +76,9 @@ void machine__delete(struct machine *machine);
 struct branch_info *machine__resolve_bstack(struct machine *machine,
 					    struct thread *thread,
 					    struct branch_stack *bs);
+struct mem_info *machine__resolve_mem(struct machine *machine,
+				      struct thread *thread,
+				      struct perf_sample *sample, u8 cpumode);
 int machine__resolve_callchain(struct machine *machine,
 			       struct perf_evsel *evsel,
 			       struct thread *thread,
tools/perf/util/session.c +3 -0
index 627be09..cf1fe01 100644
@@ -801,6 +801,9 @@ static void dump_sample(struct perf_evsel *evsel, union perf_event *event,
 
 	if (sample_type & PERF_SAMPLE_WEIGHT)
 		printf("... weight: %" PRIu64 "\n", sample->weight);
+
+	if (sample_type & PERF_SAMPLE_DATA_SRC)
+		printf(" . data_src: 0x%"PRIx64"\n", sample->data_src);
 }
 
 static struct machine *
tools/perf/util/sort.c +363 -6

399 lines not on this page · raw message

tools/perf/util/sort.h +8 -1
index 3939250..f24bdf6 100644
@@ -101,7 +101,8 @@ struct hist_entry {
 	struct rb_root		sorted_chain;
 	struct branch_info	*branch_info;
 	struct hists		*hists;
-	struct callchain_root	callchain[0];
+	struct mem_info		*mem_info;
+	struct callchain_root	callchain[0]; /* must be last member */
 };
 
 static inline bool hist_entry__has_pairs(struct hist_entry *he)
@@ -133,6 +134,12 @@ enum sort_type {
 	SORT_SRCLINE,
 	SORT_LOCAL_WEIGHT,
 	SORT_GLOBAL_WEIGHT,
+	SORT_MEM_DADDR_SYMBOL,
+	SORT_MEM_DADDR_DSO,
+	SORT_MEM_LOCKED,
+	SORT_MEM_TLB,
+	SORT_MEM_LVL,
+	SORT_MEM_SNOOP,
 
 	/* branch stack specific sort keys */
 	__SORT_BRANCH_STACK,
tools/perf/util/symbol.h +6 -0
index d7654c2..5f720dc 100644
@@ -156,6 +156,12 @@ struct branch_info {
 	struct branch_flags flags;
 };
 
+struct mem_info {
+	struct addr_map_symbol iaddr;
+	struct addr_map_symbol daddr;
+	union perf_mem_data_src data_src;
+};
+
 struct addr_location {
 	struct thread *thread;
 	struct map    *map;
#48 lore raw

PATCH v7 12/18 perf report: add support for mem access profiling

In reply to Stephane Eranian

This patch adds the --mem-mode option to perf report.

This mode requires a perf.data file created with memory access samples.

---
 tools/perf/builtin-report.c |  140 +++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 135 insertions(+), 5 deletions(-)
tools/perf/builtin-report.c +135 -5
index 3cfe259..f433011 100644
@@ -46,6 +46,7 @@ struct perf_report {
 	bool			show_full_info;
 	bool			show_threads;
 	bool			inverted_callchain;
+	bool			mem_mode;
 	struct perf_read_values	show_threads_values;
 	const char		*pretty_printing_style;
 	symbol_filter_t		annotate_init;
@@ -54,6 +55,100 @@ struct perf_report {
 	DECLARE_BITMAP(cpu_bitmap, MAX_NR_CPUS);
 };
 
+static int perf_report__add_mem_hist_entry(struct perf_tool *tool,
+					   struct addr_location *al,
+					   struct perf_sample *sample,
+					   struct perf_evsel *evsel,
+					   struct machine *machine,
+					   union perf_event *event)
+{
+	struct perf_report *rep = container_of(tool, struct perf_report, tool);
+	struct symbol *parent = NULL;
+	u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
+	int err = 0;
+	struct hist_entry *he;
+	struct mem_info *mi, *mx;
+	uint64_t cost;
+
+	if ((sort__has_parent || symbol_conf.use_callchain)
+	    && sample->callchain) {
+		err = machine__resolve_callchain(machine, evsel, al->thread,
+						 sample, &parent);
+		if (err)
+			return err;
+	}
+
+	mi = machine__resolve_mem(machine, al->thread, sample, cpumode);
+	if (!mi)
+		return -ENOMEM;
+
+	if (rep->hide_unresolved && !al->sym)
+		return 0;
+
+	cost = sample->weight;
+	if (!cost)
+		cost = 1;
+
+	/*
+	 * must pass period=weight in order to get the correct
+	 * sorting from hists__collapse_resort() which is solely
+	 * based on periods. We want sorting be done on nr_events * weight
+	 * and this is indirectly achieved by passing period=weight here
+	 * and the he_stat__add_period() function.
+	 */
+	he = __hists__add_mem_entry(&evsel->hists, al, parent, mi, cost, cost);
+	if (!he)
+		return -ENOMEM;
+
+	/*
+	 * In the newt browser, we are doing integrated annotation,
+	 * so we don't allocate the extra space needed because the stdio
+	 * code will not use it.
+	 */
+	if (sort__has_sym && he->ms.sym && use_browser > 0) {
+		struct annotation *notes = symbol__annotation(he->ms.sym);
+
+		assert(evsel != NULL);
+
+		if (notes->src == NULL && symbol__alloc_hist(he->ms.sym) < 0)
+			goto out;
+
+		err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
+		if (err)
+			goto out;
+	}
+
+	if (sort__has_sym && he->mem_info->daddr.sym && use_browser > 0) {
+		struct annotation *notes;
+
+		mx = he->mem_info;
+
+		notes = symbol__annotation(mx->daddr.sym);
+		if (!notes->src
+		    && symbol__alloc_hist(mx->daddr.sym) < 0)
+			goto out;
+
+		err = symbol__inc_addr_samples(mx->daddr.sym,
+					       mx->daddr.map,
+					       evsel->idx,
+					       mx->daddr.al_addr);
+		if (err)
+			goto out;
+	}
+
+	evsel->hists.stats.total_period += cost;
+	hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
+	err = 0;
+
+	if (symbol_conf.use_callchain) {
+		err = callchain_append(he->callchain,
+				       &callchain_cursor,
+				       sample->period);
+	}
+out:
+	return err;
+}
+
 static int perf_report__add_branch_hist_entry(struct perf_tool *tool,
 					struct addr_location *al,
 					struct perf_sample *sample,
@@ -210,6 +305,12 @@ static int process_sample_event(struct perf_tool *tool,
 			pr_debug("problem adding lbr entry, skipping event\n");
 			return -1;
 		}
+	} else if (rep->mem_mode == 1) {
+		if (perf_report__add_mem_hist_entry(tool, &al, sample,
+						    evsel, machine, event)) {
+			pr_debug("problem adding mem entry, skipping event\n");
+			return -1;
+		}
 	} else {
 		if (al.map != NULL)
 			al.map->dso->hit = 1;
@@ -293,7 +394,8 @@ static void sig_handler(int sig __maybe_unused)
 	session_done = 1;
 }
 
-static size_t hists__fprintf_nr_sample_events(struct hists *self,
+static size_t hists__fprintf_nr_sample_events(struct perf_report *rep,
+					      struct hists *self,
 					      const char *evname, FILE *fp)
 {
 	size_t ret;
@@ -306,7 +408,11 @@ static size_t hists__fprintf_nr_sample_events(struct hists *self,
 	if (evname != NULL)
 		ret += fprintf(fp, " of event '%s'", evname);
 
-	ret += fprintf(fp, "\n# Event count (approx.): %" PRIu64, nr_events);
+	if (rep->mem_mode) {
+		ret += fprintf(fp, "\n# Total weight : %" PRIu64, nr_events);
+		ret += fprintf(fp, "\n# Sort order   : %s", sort_order);
+	} else
+		ret += fprintf(fp, "\n# Event count (approx.): %" PRIu64, nr_events);
 	return ret + fprintf(fp, "\n#\n");
 }
 
@@ -320,7 +426,7 @@ static int perf_evlist__tty_browse_hists(struct perf_evlist *evlist,
 		struct hists *hists = &pos->hists;
 		const char *evname = perf_evsel__name(pos);
 
-		hists__fprintf_nr_sample_events(hists, evname, stdout);
+		hists__fprintf_nr_sample_events(rep, hists, evname, stdout);
 		hists__fprintf(hists, true, 0, 0, stdout);
 		fprintf(stdout, "\n\n");
 	}
@@ -596,8 +702,11 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_BOOLEAN(0, "stdio", &report.use_stdio,
 		    "Use the stdio interface"),
 	OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
-		   "sort by key(s): pid, comm, dso, symbol, parent, dso_to,"
-		   " dso_from, symbol_to, symbol_from, mispredict, weight, local_weight"),
+		   "sort by key(s): pid, comm, dso, symbol, parent, dso_to, "
+		   "dso_from, symbol_to, symbol_from, mispredict, weight, "
+		   "local_weight, dso_from, symbol_to, symbol_from, "
+		   "mispredict, mem, symbol_daddr, dso_daddr, tlb, snoop"
+		   ", locked"),
 	OPT_BOOLEAN(0, "showcpuutilization", &symbol_conf.show_cpu_utilization,
 		    "Show sample percentage for different cpu modes"),
 	OPT_STRING('p', "parent", &parent_pattern, "regex",
@@ -643,6 +752,7 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 		    "use branch records for histogram filling", parse_branch_mode),
 	OPT_STRING(0, "objdump", &objdump_path, "path",
 		   "objdump binary to use for disassembly and annotations"),
+	OPT_BOOLEAN(0, "mem-mode", &report.mem_mode, "mem access profile"),
 	OPT_END()
 	};
 
@@ -688,6 +798,18 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 				     "dso_to,symbol_to";
 
 	}
+	if (report.mem_mode) {
+		if (sort__branch_mode == 1) {
+			fprintf(stderr, "branch and mem mode incompatible\n");
+			goto error;
+		}
+		/*
+		 * if no sort_order is provided, then specify
+		 * branch-mode specific order
+		 */
+		if (sort_order == default_sort_order)
+			sort_order = "local_weight,mem,sym,dso,symbol_daddr,dso_daddr,snoop,tlb,locked";
+	}
 
 	if (strcmp(input_name, "-") != 0)
 		setup_browser(true);
@@ -759,6 +881,14 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 		sort_entry__setup_elide(&sort_sym_from, symbol_conf.sym_from_list, "sym_from", stdout);
 		sort_entry__setup_elide(&sort_sym_to, symbol_conf.sym_to_list, "sym_to", stdout);
 	} else {
+		if (report.mem_mode) {
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "symbol_daddr", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "dso_daddr", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "mem", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "local_weight", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "tlb", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "snoop", stdout);
+		}
 		sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "dso", stdout);
 		sort_entry__setup_elide(&sort_sym, symbol_conf.sym_list, "symbol", stdout);
 	}
Signature · 1 line
1.7.9.5
#49 lore raw

[tip:perf/core] perf report: Add support for mem access profiling

perf report: Add support for mem access profiling
Commit-ID:  f4f7e28d0e813ddb997f49ae718ddf98db972292
Gitweb:     http://git.kernel.org/tip/f4f7e28d0e813ddb997f49ae718ddf98db972292
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:36 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:21:28 -0300

perf report: Add support for mem access profiling

This patch adds the --mem-mode option to perf report.

This mode requires a perf.data file created with memory access samples.

[ Removed duplicates in the --sort help, man page needs updating,
Fixed minor conflict with 328ccda "perf report: Add --no-demangle option" ]

---
 tools/perf/builtin-report.c | 135 ++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 131 insertions(+), 4 deletions(-)
tools/perf/builtin-report.c +131 -4
index e31f070..a20550c 100644
@@ -46,6 +46,7 @@ struct perf_report {
 	bool			show_full_info;
 	bool			show_threads;
 	bool			inverted_callchain;
+	bool			mem_mode;
 	struct perf_read_values	show_threads_values;
 	const char		*pretty_printing_style;
 	symbol_filter_t		annotate_init;
@@ -64,6 +65,99 @@ static int perf_report_config(const char *var, const char *value, void *cb)
 	return perf_default_config(var, value, cb);
 }
 
+static int perf_report__add_mem_hist_entry(struct perf_tool *tool,
+					   struct addr_location *al,
+					   struct perf_sample *sample,
+					   struct perf_evsel *evsel,
+					   struct machine *machine,
+					   union perf_event *event)
+{
+	struct perf_report *rep = container_of(tool, struct perf_report, tool);
+	struct symbol *parent = NULL;
+	u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
+	int err = 0;
+	struct hist_entry *he;
+	struct mem_info *mi, *mx;
+	uint64_t cost;
+
+	if ((sort__has_parent || symbol_conf.use_callchain) &&
+	    sample->callchain) {
+		err = machine__resolve_callchain(machine, evsel, al->thread,
+						 sample, &parent);
+		if (err)
+			return err;
+	}
+
+	mi = machine__resolve_mem(machine, al->thread, sample, cpumode);
+	if (!mi)
+		return -ENOMEM;
+
+	if (rep->hide_unresolved && !al->sym)
+		return 0;
+
+	cost = sample->weight;
+	if (!cost)
+		cost = 1;
+
+	/*
+	 * must pass period=weight in order to get the correct
+	 * sorting from hists__collapse_resort() which is solely
+	 * based on periods. We want sorting be done on nr_events * weight
+	 * and this is indirectly achieved by passing period=weight here
+	 * and the he_stat__add_period() function.
+	 */
+	he = __hists__add_mem_entry(&evsel->hists, al, parent, mi, cost, cost);
+	if (!he)
+		return -ENOMEM;
+
+	/*
+	 * In the newt browser, we are doing integrated annotation,
+	 * so we don't allocate the extra space needed because the stdio
+	 * code will not use it.
+	 */
+	if (sort__has_sym && he->ms.sym && use_browser > 0) {
+		struct annotation *notes = symbol__annotation(he->ms.sym);
+
+		assert(evsel != NULL);
+
+		if (notes->src == NULL && symbol__alloc_hist(he->ms.sym) < 0)
+			goto out;
+
+		err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
+		if (err)
+			goto out;
+	}
+
+	if (sort__has_sym && he->mem_info->daddr.sym && use_browser > 0) {
+		struct annotation *notes;
+
+		mx = he->mem_info;
+
+		notes = symbol__annotation(mx->daddr.sym);
+		if (notes->src == NULL && symbol__alloc_hist(mx->daddr.sym) < 0)
+			goto out;
+
+		err = symbol__inc_addr_samples(mx->daddr.sym,
+					       mx->daddr.map,
+					       evsel->idx,
+					       mx->daddr.al_addr);
+		if (err)
+			goto out;
+	}
+
+	evsel->hists.stats.total_period += cost;
+	hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
+	err = 0;
+
+	if (symbol_conf.use_callchain) {
+		err = callchain_append(he->callchain,
+				       &callchain_cursor,
+				       sample->period);
+	}
+out:
+	return err;
+}
+
 static int perf_report__add_branch_hist_entry(struct perf_tool *tool,
 					struct addr_location *al,
 					struct perf_sample *sample,
@@ -220,6 +314,12 @@ static int process_sample_event(struct perf_tool *tool,
 			pr_debug("problem adding lbr entry, skipping event\n");
 			return -1;
 		}
+	} else if (rep->mem_mode == 1) {
+		if (perf_report__add_mem_hist_entry(tool, &al, sample,
+						    evsel, machine, event)) {
+			pr_debug("problem adding mem entry, skipping event\n");
+			return -1;
+		}
 	} else {
 		if (al.map != NULL)
 			al.map->dso->hit = 1;
@@ -303,7 +403,8 @@ static void sig_handler(int sig __maybe_unused)
 	session_done = 1;
 }
 
-static size_t hists__fprintf_nr_sample_events(struct hists *self,
+static size_t hists__fprintf_nr_sample_events(struct perf_report *rep,
+					      struct hists *self,
 					      const char *evname, FILE *fp)
 {
 	size_t ret;
@@ -331,7 +432,11 @@ static size_t hists__fprintf_nr_sample_events(struct hists *self,
 	if (evname != NULL)
 		ret += fprintf(fp, " of event '%s'", evname);
 
-	ret += fprintf(fp, "\n# Event count (approx.): %" PRIu64, nr_events);
+	if (rep->mem_mode) {
+		ret += fprintf(fp, "\n# Total weight : %" PRIu64, nr_events);
+		ret += fprintf(fp, "\n# Sort order   : %s", sort_order);
+	} else
+		ret += fprintf(fp, "\n# Event count (approx.): %" PRIu64, nr_events);
 	return ret + fprintf(fp, "\n#\n");
 }
 
@@ -349,7 +454,7 @@ static int perf_evlist__tty_browse_hists(struct perf_evlist *evlist,
 		    !perf_evsel__is_group_leader(pos))
 			continue;
 
-		hists__fprintf_nr_sample_events(hists, evname, stdout);
+		hists__fprintf_nr_sample_events(rep, hists, evname, stdout);
 		hists__fprintf(hists, true, 0, 0, stdout);
 		fprintf(stdout, "\n\n");
 	}
@@ -646,7 +751,8 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
 		   "sort by key(s): pid, comm, dso, symbol, parent, cpu, srcline,"
 		   " dso_to, dso_from, symbol_to, symbol_from, mispredict,"
-		   " weight, local_weight"),
+		   " weight, local_weight, mem, symbol_daddr, dso_daddr, tlb, "
+		   "snoop, locked"),
 	OPT_BOOLEAN(0, "showcpuutilization", &symbol_conf.show_cpu_utilization,
 		    "Show sample percentage for different cpu modes"),
 	OPT_STRING('p', "parent", &parent_pattern, "regex",
@@ -696,6 +802,7 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 		   "objdump binary to use for disassembly and annotations"),
 	OPT_BOOLEAN(0, "demangle", &symbol_conf.demangle,
 		    "Disable symbol demangling"),
+	OPT_BOOLEAN(0, "mem-mode", &report.mem_mode, "mem access profile"),
 	OPT_END()
 	};
 
@@ -753,6 +860,18 @@ repeat:
 				     "dso_to,symbol_to";
 
 	}
+	if (report.mem_mode) {
+		if (sort__branch_mode == 1) {
+			fprintf(stderr, "branch and mem mode incompatible\n");
+			goto error;
+		}
+		/*
+		 * if no sort_order is provided, then specify
+		 * branch-mode specific order
+		 */
+		if (sort_order == default_sort_order)
+			sort_order = "local_weight,mem,sym,dso,symbol_daddr,dso_daddr,snoop,tlb,locked";
+	}
 
 	if (setup_sorting() < 0)
 		usage_with_options(report_usage, options);
@@ -818,6 +937,14 @@ repeat:
 		sort_entry__setup_elide(&sort_sym_from, symbol_conf.sym_from_list, "sym_from", stdout);
 		sort_entry__setup_elide(&sort_sym_to, symbol_conf.sym_to_list, "sym_to", stdout);
 	} else {
+		if (report.mem_mode) {
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "symbol_daddr", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "dso_daddr", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "mem", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "local_weight", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "tlb", stdout);
+			sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "snoop", stdout);
+		}
 		sort_entry__setup_elide(&sort_dso, symbol_conf.dso_list, "dso", stdout);
 		sort_entry__setup_elide(&sort_sym, symbol_conf.sym_list, "symbol", stdout);
 	}
#50 lore raw

PATCH v7 13/18 perf record: add support for mem access profiling

In reply to Stephane Eranian

We use the -W option to obtain the cost of the
memory accesses.

Data address sampling is obtained via the -d option.

---
 tools/perf/util/evsel.c |    3 +++
 1 file changed, 3 insertions(+)
tools/perf/util/evsel.c +3 -0
index 49daa7b..277b98d 100644
@@ -501,6 +501,9 @@ void perf_evsel__config(struct perf_evsel *evsel,
 		attr->sample_type	|= PERF_SAMPLE_CPU;
 	}
 
+	if (opts->sample_address)
+		attr->sample_type	|= PERF_SAMPLE_DSRC;
+
 	if (opts->no_delay) {
 		attr->watermark = 0;
 		attr->wakeup_events = 1;
Signature · 1 line
1.7.9.5
#51 lore raw

[tip:perf/core] perf record: Add support for mem access profiling

perf record: Add support for mem access profiling
Commit-ID:  ccf49bfc6bb1025788637417780e9f1eeae9fc37
Gitweb:     http://git.kernel.org/tip/ccf49bfc6bb1025788637417780e9f1eeae9fc37
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:37 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:20:28 -0300

perf record: Add support for mem access profiling

We use the -W option to obtain the cost of the memory accesses.

Data address sampling is obtained via the -d option.

---
 tools/perf/util/evsel.c | 3 +++
 1 file changed, 3 insertions(+)
tools/perf/util/evsel.c +3 -0
index 5c4ca51..07b1a3a 100644
@@ -554,6 +554,9 @@ void perf_evsel__config(struct perf_evsel *evsel,
 		perf_evsel__set_sample_bit(evsel, CPU);
 	}
 
+	if (opts->sample_address)
+		attr->sample_type	|= PERF_SAMPLE_DATA_SRC;
+
 	if (opts->no_delay) {
 		attr->watermark = 0;
 		attr->wakeup_events = 1;
#52 lore raw

PATCH v7 14/18 perf tools: add new mem command for memory access profiling

In reply to Stephane Eranian

This new command is a wrapper on top of perf record and perf report to make it easier to configure for memory access profiling.

To record loads:

$ perf mem -t load rec .....

To record stores:

$ perf mem -t store rec .....

To get the report:

$ perf mem -t load rep
---
 tools/perf/Documentation/perf-mem.txt |   48 +++++++
 tools/perf/Makefile                   |    1 +
 tools/perf/builtin-mem.c              |  242 +++++++++++++++++++++++++++++++++
 tools/perf/builtin.h                  |    1 +
 tools/perf/command-list.txt           |    1 +
 tools/perf/perf.c                     |    1 +
 tools/perf/util/hist.c                |    1 +
 7 files changed, 295 insertions(+)
 create mode 100644 tools/perf/Documentation/perf-mem.txt
 create mode 100644 tools/perf/builtin-mem.c
tools/perf/Documentation/perf-mem.txt new file +48 -0
new file mode 100644
index 0000000..888d511
@@ -0,0 +1,48 @@
+perf-mem(1)
+===========
+
+NAME
+----
+perf-mem - Profile memory accesses
+
+SYNOPSIS
+--------
+[verse]
+'perf mem' [<options>] (record [<command>] | report)
+
+DESCRIPTION
+-----------
+"perf mem -t <TYPE> record" runs a command and gathers memory operation data
+from it, into perf.data. Perf record options are accepted and are passed through.
+
+"perf mem -t <TYPE> report" displays the result. It invokes perf report with the
+right set of options to display a memory access profile.
+
+OPTIONS
+-------
+<command>...::
+	Any command you can specify in a shell.
+
+-t::
+--type=::
+	Select the memory operation type: load or store (default: load)
+
+-D::
+--dump-raw-samples=::
+	Dump the raw decoded samples on the screen in a format that is easy to parse with
+	one sample per line.
+
+-x::
+--field-separator::
+	Specify the field separator used when dump raw samples (-D option). By default,
+	The separator is the space character.
+
+-C::
+--cpu-list::
+	Restrict dump of raw samples to those provided via this option. Note that the same
+	option can be passed in record mode. It will be interpreted the same way as perf
+	record.
+
+SEE ALSO
+--------
+linkperf:perf-record[1], linkperf:perf-report[1]
tools/perf/Makefile +1 -0
index 994f0f6..be8f4697 100644
@@ -505,6 +505,7 @@ BUILTIN_OBJS += $(OUTPUT)builtin-lock.o
 BUILTIN_OBJS += $(OUTPUT)builtin-kvm.o
 BUILTIN_OBJS += $(OUTPUT)builtin-inject.o
 BUILTIN_OBJS += $(OUTPUT)tests/builtin-test.o
+BUILTIN_OBJS += $(OUTPUT)builtin-mem.o
 
 PERFLIBS = $(LIB_FILE) $(LIBTRACEEVENT)
 
tools/perf/builtin-mem.c new file +242 -0

242 lines not on this page · raw message

tools/perf/builtin.h +1 -0
index 08143bd..b210d62 100644
@@ -36,6 +36,7 @@ extern int cmd_kvm(int argc, const char **argv, const char *prefix);
 extern int cmd_test(int argc, const char **argv, const char *prefix);
 extern int cmd_trace(int argc, const char **argv, const char *prefix);
 extern int cmd_inject(int argc, const char **argv, const char *prefix);
+extern int cmd_mem(int argc, const char **argv, const char *prefix);
 
 extern int find_scripts(char **scripts_array, char **scripts_path_array);
 #endif
tools/perf/command-list.txt +1 -0
index 3e86bbd..2c5b621 100644
@@ -24,3 +24,4 @@ perf-kmem			mainporcelain common
 perf-lock			mainporcelain common
 perf-kvm			mainporcelain common
 perf-test			mainporcelain common
+perf-mem			mainporcelain common
tools/perf/perf.c +1 -0
index 0f661fb..682340e 100644
@@ -60,6 +60,7 @@ static struct cmd_struct commands[] = {
 	{ "trace",	cmd_trace,	0 },
 #endif
 	{ "inject",	cmd_inject,	0 },
+	{ "mem",	cmd_mem,	0 },
 };
 
 struct pager_config {
tools/perf/util/hist.c +1 -0
index 9203683..8482a0a1 100644
@@ -493,6 +493,7 @@ hist_entry__collapse(struct hist_entry *left, struct hist_entry *right)
 void hist_entry__free(struct hist_entry *he)
 {
 	free(he->branch_info);
+	free(he->mem_info);
 	free(he);
 }
 
Signature · 1 line
1.7.9.5
#53 lore raw

[tip:perf/core] perf tools: Add new mem command for memory access profiling

perf tools: Add new mem command for memory access profiling
Commit-ID:  028f12ee6beff0961781c5ed3f740e5f3b56f781
Gitweb:     http://git.kernel.org/tip/028f12ee6beff0961781c5ed3f740e5f3b56f781
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:38 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:21:44 -0300

perf tools: Add new mem command for memory access profiling

This new command is a wrapper on top of perf record and perf report to make it easier to configure for memory access profiling.

To record loads:

$ perf mem -t load rec .....

To record stores:

$ perf mem -t store rec .....

To get the report:

$ perf mem -t load rep
Cc Andi Kleen <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Peter Zijlstra <[email protected]>
[ Fixed minor conflict with 66857b5 "Sort command-list.txt alphabetically" ]
---
 tools/perf/Documentation/perf-mem.txt |  48 +++++++
 tools/perf/Makefile                   |   1 +
 tools/perf/builtin-mem.c              | 242 ++++++++++++++++++++++++++++++++++
 tools/perf/builtin.h                  |   1 +
 tools/perf/command-list.txt           |   1 +
 tools/perf/perf.c                     |   1 +
 tools/perf/util/hist.c                |   1 +
 7 files changed, 295 insertions(+)
tools/perf/Documentation/perf-mem.txt new file +48 -0
new file mode 100644
index 0000000..888d511
@@ -0,0 +1,48 @@
+perf-mem(1)
+===========
+
+NAME
+----
+perf-mem - Profile memory accesses
+
+SYNOPSIS
+--------
+[verse]
+'perf mem' [<options>] (record [<command>] | report)
+
+DESCRIPTION
+-----------
+"perf mem -t <TYPE> record" runs a command and gathers memory operation data
+from it, into perf.data. Perf record options are accepted and are passed through.
+
+"perf mem -t <TYPE> report" displays the result. It invokes perf report with the
+right set of options to display a memory access profile.
+
+OPTIONS
+-------
+<command>...::
+	Any command you can specify in a shell.
+
+-t::
+--type=::
+	Select the memory operation type: load or store (default: load)
+
+-D::
+--dump-raw-samples=::
+	Dump the raw decoded samples on the screen in a format that is easy to parse with
+	one sample per line.
+
+-x::
+--field-separator::
+	Specify the field separator used when dump raw samples (-D option). By default,
+	The separator is the space character.
+
+-C::
+--cpu-list::
+	Restrict dump of raw samples to those provided via this option. Note that the same
+	option can be passed in record mode. It will be interpreted the same way as perf
+	record.
+
+SEE ALSO
+--------
+linkperf:perf-record[1], linkperf:perf-report[1]
tools/perf/Makefile +1 -0
index 0230b75..07feae7 100644
@@ -547,6 +547,7 @@ BUILTIN_OBJS += $(OUTPUT)builtin-lock.o
 BUILTIN_OBJS += $(OUTPUT)builtin-kvm.o
 BUILTIN_OBJS += $(OUTPUT)builtin-inject.o
 BUILTIN_OBJS += $(OUTPUT)tests/builtin-test.o
+BUILTIN_OBJS += $(OUTPUT)builtin-mem.o
 
 PERFLIBS = $(LIB_FILE) $(LIBLK) $(LIBTRACEEVENT)
 
tools/perf/builtin-mem.c new file +242 -0

242 lines not on this page · raw message

tools/perf/builtin.h +1 -0
index 08143bd..b210d62 100644
@@ -36,6 +36,7 @@ extern int cmd_kvm(int argc, const char **argv, const char *prefix);
 extern int cmd_test(int argc, const char **argv, const char *prefix);
 extern int cmd_trace(int argc, const char **argv, const char *prefix);
 extern int cmd_inject(int argc, const char **argv, const char *prefix);
+extern int cmd_mem(int argc, const char **argv, const char *prefix);
 
 extern int find_scripts(char **scripts_array, char **scripts_path_array);
 #endif
tools/perf/command-list.txt +1 -0
index a28e31b..0906fc4 100644
@@ -14,6 +14,7 @@ perf-kmem			mainporcelain common
 perf-kvm			mainporcelain common
 perf-list			mainporcelain common
 perf-lock			mainporcelain common
+perf-mem			mainporcelain common
 perf-probe			mainporcelain full
 perf-record			mainporcelain common
 perf-report			mainporcelain common
tools/perf/perf.c +1 -0
index f6ba7b7..31c9380 100644
@@ -60,6 +60,7 @@ static struct cmd_struct commands[] = {
 	{ "trace",	cmd_trace,	0 },
 #endif
 	{ "inject",	cmd_inject,	0 },
+	{ "mem",	cmd_mem,	0 },
 };
 
 struct pager_config {
tools/perf/util/hist.c +1 -0
index 99cc719..6b32721 100644
@@ -520,6 +520,7 @@ hist_entry__collapse(struct hist_entry *left, struct hist_entry *right)
 void hist_entry__free(struct hist_entry *he)
 {
 	free(he->branch_info);
+	free(he->mem_info);
 	free(he);
 }
 
#54 lore raw

PATCH v7 15/18 perf: add PERF_RECORD_MISC_MMAP_DATA to RECORD_MMAP

In reply to Stephane Eranian

Type of mapping was lost and made it hard for a tool to distinguish code vs. data mmaps. Perf has the ability to distinguish the two.

Use a bit in the header->misc bitmask to keep track of the mmap type. If PERF_RECORD_MISC_MMAP_DATA is set then the mapping is not executable (!VM_EXEC). If not set, then the mapping is executable.

---
 include/uapi/linux/perf_event.h |    1 +
 kernel/events/core.c            |    3 +++
 2 files changed, 4 insertions(+)
include/uapi/linux/perf_event.h +1 -0
index 3e4844c..3907b29 100644
@@ -445,6 +445,7 @@ struct perf_event_mmap_page {
 #define PERF_RECORD_MISC_GUEST_KERNEL		(4 << 0)
 #define PERF_RECORD_MISC_GUEST_USER		(5 << 0)
 
+#define PERF_RECORD_MISC_MMAP_DATA		(1 << 13)
 /*
  * Indicates that the content of PERF_SAMPLE_IP points to
  * the actual instruction that triggered the event. See also
kernel/events/core.c +3 -0
index 56ca60b..966069d 100644
@@ -4764,6 +4764,9 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
 	mmap_event->file_name = name;
 	mmap_event->file_size = size;
 
+	if (!(vma->vm_flags & VM_EXEC))
+		mmap_event->event_id.header.misc |= PERF_RECORD_MISC_MMAP_DATA;
+
 	mmap_event->event_id.header.size = sizeof(mmap_event->event_id) + size;
 
 	rcu_read_lock();
Signature · 1 line
1.7.9.5
#55 lore raw

[tip:perf/x86] perf: Add PERF_RECORD_MISC_MMAP_DATA to RECORD_MMAP

Commit-ID:  59e1a00ac0e4d33ffc46814a3509b5cc2c94c56f
Gitweb:     http://git.kernel.org/tip/59e1a00ac0e4d33ffc46814a3509b5cc2c94c56f
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:39 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 25 Jan 2013 10:19:05 +0100

perf: Add PERF_RECORD_MISC_MMAP_DATA to RECORD_MMAP

Type of mapping was lost and made it hard for a tool to distinguish code vs. data mmaps. Perf has the ability to distinguish the two.

Use a bit in the header->misc bitmask to keep track of the mmap type. If PERF_RECORD_MISC_MMAP_DATA is set then the mapping is not executable (!VM_EXEC). If not set, then the mapping is executable.

---
 include/uapi/linux/perf_event.h | 1 +
 kernel/events/core.c            | 3 +++
 2 files changed, 4 insertions(+)
include/uapi/linux/perf_event.h +1 -0
index 0c46659..62e9f25 100644
@@ -445,6 +445,7 @@ struct perf_event_mmap_page {
 #define PERF_RECORD_MISC_GUEST_KERNEL		(4 << 0)
 #define PERF_RECORD_MISC_GUEST_USER		(5 << 0)
 
+#define PERF_RECORD_MISC_MMAP_DATA		(1 << 13)
 /*
  * Indicates that the content of PERF_SAMPLE_IP points to
  * the actual instruction that triggered the event. See also
kernel/events/core.c +3 -0
index 4661009..dd1c130 100644
@@ -4764,6 +4764,9 @@ got_name:
 	mmap_event->file_name = name;
 	mmap_event->file_size = size;
 
+	if (!(vma->vm_flags & VM_EXEC))
+		mmap_event->event_id.header.misc |= PERF_RECORD_MISC_MMAP_DATA;
+
 	mmap_event->event_id.header.size = sizeof(mmap_event->event_id) + size;
 
 	rcu_read_lock();
#56 lore raw

[tip:perf/core] perf: Add PERF_RECORD_MISC_MMAP_DATA to RECORD_MMAP

In reply to Stephane Eranian
perf: Add PERF_RECORD_MISC_MMAP_DATA to RECORD_MMAP
Commit-ID:  2fe85427e3bf65d791700d065132772fc26e4d75
Gitweb:     http://git.kernel.org/tip/2fe85427e3bf65d791700d065132772fc26e4d75
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:39 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:19:02 -0300

perf: Add PERF_RECORD_MISC_MMAP_DATA to RECORD_MMAP

Type of mapping was lost and made it hard for a tool to distinguish code vs. data mmaps. Perf has the ability to distinguish the two.

Use a bit in the header->misc bitmask to keep track of the mmap type. If PERF_RECORD_MISC_MMAP_DATA is set then the mapping is not executable (!VM_EXEC). If not set, then the mapping is executable.

---
 include/uapi/linux/perf_event.h | 1 +
 kernel/events/core.c            | 3 +++
 2 files changed, 4 insertions(+)
include/uapi/linux/perf_event.h +1 -0
index 5b57620..964a450 100644
@@ -445,6 +445,7 @@ struct perf_event_mmap_page {
 #define PERF_RECORD_MISC_GUEST_KERNEL		(4 << 0)
 #define PERF_RECORD_MISC_GUEST_USER		(5 << 0)
 
+#define PERF_RECORD_MISC_MMAP_DATA		(1 << 13)
 /*
  * Indicates that the content of PERF_SAMPLE_IP points to
  * the actual instruction that triggered the event. See also
kernel/events/core.c +3 -0
index 77c96d1..98c0845 100644
@@ -4791,6 +4791,9 @@ got_name:
 	mmap_event->file_name = name;
 	mmap_event->file_size = size;
 
+	if (!(vma->vm_flags & VM_EXEC))
+		mmap_event->event_id.header.misc |= PERF_RECORD_MISC_MMAP_DATA;
+
 	mmap_event->event_id.header.size = sizeof(mmap_event->event_id) + size;
 
 	rcu_read_lock();
#57 lore raw

PATCH v7 16/18 perf tools: detect data vs. text mappings

In reply to Stephane Eranian

Leverages the PERF_RECORD_MISC_MMAP_DATA bit in the RECORD_MMAP record header. When the bit is set then the mapping type is set to MAP__VARIABLE.

---
 tools/perf/util/machine.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
tools/perf/util/machine.c +9 -1
index 1f09d05..d1c3e48 100644
@@ -379,6 +379,7 @@ int machine__process_mmap_event(struct machine *machine, union perf_event *event
 	u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
 	struct thread *thread;
 	struct map *map;
+	enum map_type type;
 	int ret = 0;
 
 	if (dump_trace)
@@ -395,10 +396,17 @@ int machine__process_mmap_event(struct machine *machine, union perf_event *event
 	thread = machine__findnew_thread(machine, event->mmap.pid);
 	if (thread == NULL)
 		goto out_problem;
+
+	if (event->header.misc & PERF_RECORD_MISC_MMAP_DATA)
+		type = MAP__VARIABLE;
+	else
+		type = MAP__FUNCTION;
+
 	map = map__new(&machine->user_dsos, event->mmap.start,
 			event->mmap.len, event->mmap.pgoff,
 			event->mmap.pid, event->mmap.filename,
-			MAP__FUNCTION);
+			type);
+
 	if (map == NULL)
 		goto out_problem;
 
Signature · 1 line
1.7.9.5
#58 lore raw

[tip:perf/core] perf machine: Detect data vs. text mappings

perf machine: Detect data vs. text mappings
Commit-ID:  bad4091791b0bb8c2d7919ddefe2f0d109299b5a
Gitweb:     http://git.kernel.org/tip/bad4091791b0bb8c2d7919ddefe2f0d109299b5a
Author:     Stephane Eranian <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:40 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:22:00 -0300

perf machine: Detect data vs. text mappings

Leverages the PERF_RECORD_MISC_MMAP_DATA bit in the RECORD_MMAP record header. When the bit is set then the mapping type is set to MAP__VARIABLE.

---
 tools/perf/util/machine.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
tools/perf/util/machine.c +9 -1
index d77ba86..b2ecad6 100644
@@ -955,6 +955,7 @@ int machine__process_mmap_event(struct machine *machine, union perf_event *event
 	u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
 	struct thread *thread;
 	struct map *map;
+	enum map_type type;
 	int ret = 0;
 
 	if (dump_trace)
@@ -971,10 +972,17 @@ int machine__process_mmap_event(struct machine *machine, union perf_event *event
 	thread = machine__findnew_thread(machine, event->mmap.pid);
 	if (thread == NULL)
 		goto out_problem;
+
+	if (event->header.misc & PERF_RECORD_MISC_MMAP_DATA)
+		type = MAP__VARIABLE;
+	else
+		type = MAP__FUNCTION;
+
 	map = map__new(&machine->user_dsos, event->mmap.start,
 			event->mmap.len, event->mmap.pgoff,
 			event->mmap.pid, event->mmap.filename,
-			MAP__FUNCTION);
+			type);
+
 	if (map == NULL)
 		goto out_problem;
 
#59 lore raw

PATCH v7 17/18 perf tools: Ignore ABS symbols when loading data maps

In reply to Stephane Eranian

From: Namhyung Kim <[email protected]>

When loading symbols in a data mapping, ABS symbols (which has a value of SHN_ABS in its st_shndx) failed at elf_getscn(). And it marks the loading as a failure so already loaded symbols cannot be fixed up.

I'm not sure what should be done. Just ignore them for now. :)

Cc Stephane Eranian <[email protected]>
---
 tools/perf/util/symbol-elf.c |    3 +++
 1 file changed, 3 insertions(+)
tools/perf/util/symbol-elf.c +3 -0
index db0cc92..00cf128 100644
@@ -719,6 +719,9 @@ int dso__load_sym(struct dso *dso, struct map *map,
 			used_opd = true;
 		}
 
+		if (sym.st_shndx == SHN_ABS)
+			continue;
+
 		sec = elf_getscn(runtime_ss->elf, sym.st_shndx);
 		if (!sec)
 			goto out_elf_end;
Signature · 1 line
1.7.9.5
#60 lore raw

PATCH v7 18/18 perf tools: Fix output of symbol_daddr offset

In reply to Stephane Eranian

From: Namhyung Kim <[email protected]>

The symbol addresses in a dso have relative offsets from the start of a mapping. So in order to ouput correct offset value from @ip, one of them should be converted.

Cc Stephane Eranian <[email protected]>
---
 tools/perf/util/sort.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/util/sort.c +1 -1
index 0625ea7..ce2f18b 100644
@@ -186,7 +186,7 @@ static int _hist_entry__sym_snprintf(struct map *map, struct symbol *sym,
 		if (map->type == MAP__VARIABLE) {
 			ret += repsep_snprintf(bf + ret, size - ret, "%s", sym->name);
 			ret += repsep_snprintf(bf + ret, size - ret, "+0x%llx",
-					ip - sym->start);
+					ip - map->unmap_ip(map, sym->start));
 			ret += repsep_snprintf(bf + ret, size - ret, "%-*s",
 				       width - ret, "");
 		} else {
Signature · 1 line
1.7.9.5
#61 lore raw

[tip:perf/core] perf tools: Fix output of symbol_daddr offset

perf tools: Fix output of symbol_daddr offset
Commit-ID:  62667746a6ded2a1fc8dac2e6258f46150b5e46c
Gitweb:     http://git.kernel.org/tip/62667746a6ded2a1fc8dac2e6258f46150b5e46c
Author:     Namhyung Kim <[email protected]>
AuthorDate: Thu, 24 Jan 2013 16:10:42 +0100
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Mon, 1 Apr 2013 12:22:15 -0300

perf tools: Fix output of symbol_daddr offset

The symbol addresses in a dso have relative offsets from the start of a mapping. So in order to ouput correct offset value from @ip, one of them should be converted.

---
 tools/perf/util/sort.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/util/sort.c +1 -1
index 32a1ef1..5f52d49 100644
@@ -202,7 +202,7 @@ static int _hist_entry__sym_snprintf(struct map *map, struct symbol *sym,
 		if (map->type == MAP__VARIABLE) {
 			ret += repsep_snprintf(bf + ret, size - ret, "%s", sym->name);
 			ret += repsep_snprintf(bf + ret, size - ret, "+0x%llx",
-					ip - sym->start);
+					ip - map->unmap_ip(map, sym->start));
 			ret += repsep_snprintf(bf + ret, size - ret, "%-*s",
 				       width - ret, "");
 		} else {
#62 lore raw
In reply to Stephane Eranian
Quoted 2 lines · Stephane Eranian

This patch series had a new feature to the kernel perf_events interface and corresponding user level tool, perf.

Can I add your Signed-off-by tag to the patches you picked up from Andi?

Thanks,

Ingo

#63 lore raw
Quoted 9 lines · Ingo Molnar
Quoted 2 lines · Stephane Eranian

This patch series had a new feature to the kernel perf_events interface and corresponding user level tool, perf.

Can I add your Signed-off-by tag to the patches you picked up from Andi?

Yes. But note that I have to simplify one of them because it included TSX changes. It's the following patch:

[PATCH v7 05/18] perf, tools: Add support for weight v7 (modified)

I was going to suggest to him that he breaks his original into two or move it earlier in his stack so has to make it independent of HSW specific extensions.

Quoted 3 lines

Thanks,

Ingo

#64 lore raw
In reply to Stephane Eranian
Quoted 2 lines · Stephane Eranian

This patch series had a new feature to the kernel perf_events interface and corresponding user level tool, perf.

Ok, so I have created a topic tree for this, tip:perf/x86.

I have applied the kernel bits (with some minor renaming changes). Arnaldo, if you agree with the tooling bits you can merge that branch into your tree and apply the tooling bits from Stephane.

Thanks,

Ingo

#65 lore raw
Quoted 6 lines · Ingo Molnar
Quoted 2 lines · Stephane Eranian

This patch series had a new feature to the kernel perf_events interface and corresponding user level tool, perf.

Quoted 7 lines

Ok, so I have created a topic tree for this, tip:perf/x86.

I have applied the kernel bits (with some minor renaming changes). Arnaldo, if you agree with the tooling bits you can merge that branch into your tree and apply the tooling bits from Stephane.

Arnaldo, did you incorporate the perf changes for this somewhere in your tree yet? The kernel bits are in.

#66 lore raw

Em Tue, Feb 05, 2013 at 02:03:55PM +0100, Stephane Eranian escreveu:

Quoted 7 lines · Ingo Molnar
Quoted 5 lines · Stephane Eranian
Quoted 2 lines

This patch series had a new feature to the kernel perf_events interface and corresponding user level tool, perf.

Ok, so I have created a topic tree for this, tip:perf/x86.

Quoted 7 lines
Quoted 4 lines

I have applied the kernel bits (with some minor renaming changes). Arnaldo, if you agree with the tooling bits you can merge that branch into your tree and apply the tooling bits from Stephane.

Arnaldo, did you incorporate the perf changes for this somewhere in your tree yet? The kernel bits are in.

I'm doing it, updating some csets on the go, will post here after I go thru some tests.

- Arnaldo

#67 lore raw
Quoted 13 lines · Arnaldo Carvalho de Melo

Em Tue, Feb 05, 2013 at 02:03:55PM +0100, Stephane Eranian escreveu:

Quoted 11 lines · Ingo Molnar
Quoted 10 lines · Stephane Eranian
Quoted 2 lines

This patch series had a new feature to the kernel perf_events interface and corresponding user level tool, perf.

Ok, so I have created a topic tree for this, tip:perf/x86.

I have applied the kernel bits (with some minor renaming changes). Arnaldo, if you agree with the tooling bits you can merge that branch into your tree and apply the tooling bits from Stephane.

Quoted 5 lines
Quoted 2 lines

Arnaldo, did you incorporate the perf changes for this somewhere in your tree yet? The kernel bits are in.

I'm doing it, updating some csets on the go, will post here after I go thru some tests.

Great. Please pull/merge tip:perf/x86 directly before you apply and test the tooling bits - right now it's not in tip:master nor in perf/core. That way it will be in the tree as a coherent unit.

Thanks,

Ingo