[PATCH v4 00/18] perf: add memory access sampling support

[PATCH v4 00/18] perf: add memory access sampling support

[Stephane Eranian]

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 cost : 1013994
# Sort order : cost,mem,sym,dso,symbol_daddr,dso_daddr,snoop,tlb,locked
#
# Overhead      Samples     Cost  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.

One powerful feature of perf is that users can toy with sort order to display
the information in different format 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 cost : 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,cost
# Samples: 19K of event 'cpu/mem-loads/pp'
# Total cost : 1013994
# Sort order : symbol_daddr,cost
#
# Overhead      Samples             Data Symbol     Cost
# ........  ...........  ......................  .......
#
     0.10%            1  [.] 0x00007f67dffe8038      986
     0.09%            1  [.] 0x00007f67df91a750      890
     0.08%            1  [.] 0x00007f67e288fba8      826


One note on the 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 cost 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.

Andi Kleen (2):
  perf, x86: Support CPU specific sysfs events
  perf, core: Add a concept of a weightened sample

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

Stephane Eranian (14):
  perf/x86: improve sysfs event mapping with event string
  perf/x86: add flags to event constraints
  perf: add minimal support for PERF_SAMPLE_WEIGHT
  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              |   63 +++--
 arch/x86/kernel/cpu/perf_event.h              |   62 ++++-
 arch/x86/kernel/cpu/perf_event_intel.c        |   34 ++-
 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               |   74 +++++-
 kernel/events/core.c                          |   15 ++
 tools/perf/Makefile                           |    1 +
 tools/perf/builtin-mem.c                      |  238 ++++++++++++++++++
 tools/perf/builtin-record.c                   |    2 +
 tools/perf/builtin-report.c                   |  131 +++++++++-
 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                       |   15 ++
 tools/perf/util/hist.c                        |   67 ++++-
 tools/perf/util/hist.h                        |   13 +
 tools/perf/util/machine.c                     |   10 +-
 tools/perf/util/session.c                     |   44 ++++
 tools/perf/util/session.h                     |    4 +
 tools/perf/util/sort.c                        |  324 ++++++++++++++++++++++++-
 tools/perf/util/sort.h                        |   10 +-
 tools/perf/util/symbol-elf.c                  |    3 +
 tools/perf/util/symbol.h                      |    7 +
 28 files changed, 1264 insertions(+), 49 deletions(-)
 create mode 100644 tools/perf/builtin-mem.c

-- 
1.7.9.5

[PATCH v4 01/18] perf, x86: Support CPU specific sysfs events

[Stephane Eranian]

From: Andi Kleen <ak@linux.intel.com>

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.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 arch/x86/kernel/cpu/perf_event.c |   29 +++++++++++++++++++++++++++++
 arch/x86/kernel/cpu/perf_event.h |    1 +
 2 files changed, 30 insertions(+)

diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 4428fd1..83f4c1a 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1341,6 +1341,30 @@ 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 a;
+	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)
 {
@@ -1482,6 +1506,11 @@ static int __init init_hw_perf_events(void)
 	else
 		filter_events(x86_pmu_events_group.attrs);
 
+	if (x86_pmu.cpu_events)
+		x86_pmu_events_group.attrs =
+			merge_attr(x86_pmu_events_group.attrs,
+				   x86_pmu.cpu_events);
+
 	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);
diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h
index 115c1ea..4170043 100644
--- a/arch/x86/kernel/cpu/perf_event.h
+++ b/arch/x86/kernel/cpu/perf_event.h
@@ -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
-- 
1.7.9.5

Re: [PATCH v4 01/18] perf, x86: Support CPU specific sysfs events

[Jiri Olsa]

On Thu, Dec 20, 2012 at 04:41:31PM +0100, Stephane Eranian wrote:
> From: Andi Kleen <ak@linux.intel.com>
> 
> 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.
> 
> Signed-off-by: Andi Kleen <ak@linux.intel.com>
> ---
>  arch/x86/kernel/cpu/perf_event.c |   29 +++++++++++++++++++++++++++++
>  arch/x86/kernel/cpu/perf_event.h |    1 +
>  2 files changed, 30 insertions(+)
> 
> diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
> index 4428fd1..83f4c1a 100644
> --- a/arch/x86/kernel/cpu/perf_event.c
> +++ b/arch/x86/kernel/cpu/perf_event.c
> @@ -1341,6 +1341,30 @@ 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 a;

not likely to happen, but still.. do we want proper error handling?

other than that the patch looks ok to me

jirka

Re: [PATCH v4 01/18] perf, x86: Support CPU specific sysfs events

[Stephane Eranian]

On Wed, Jan 2, 2013 at 3:04 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> On Thu, Dec 20, 2012 at 04:41:31PM +0100, Stephane Eranian wrote:
>> From: Andi Kleen <ak@linux.intel.com>
>>
>> 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.
>>
>> Signed-off-by: Andi Kleen <ak@linux.intel.com>
>> ---
>>  arch/x86/kernel/cpu/perf_event.c |   29 +++++++++++++++++++++++++++++
>>  arch/x86/kernel/cpu/perf_event.h |    1 +
>>  2 files changed, 30 insertions(+)
>>
>> diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
>> index 4428fd1..83f4c1a 100644
>> --- a/arch/x86/kernel/cpu/perf_event.c
>> +++ b/arch/x86/kernel/cpu/perf_event.c
>> @@ -1341,6 +1341,30 @@ 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 a;
>
> not likely to happen, but still.. do we want proper error handling?
>
Ok, will fix the error path from Andi's patch.
Thanks.

> other than that the patch looks ok to me
>
> jirka

[PATCH v4 02/18] perf/x86: improve sysfs event mapping with event string

[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.

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 arch/x86/kernel/cpu/perf_event.c |   27 ++++++++++++---------------
 arch/x86/kernel/cpu/perf_event.h |   23 +++++++++++++++++++++++
 2 files changed, 35 insertions(+), 15 deletions(-)

diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 83f4c1a..f4057bb 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1316,20 +1316,22 @@ 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;
 
@@ -1370,19 +1372,14 @@ 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		);
diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h
index 4170043..3f4380c 100644
--- a/arch/x86/kernel/cpu/perf_event.h
+++ b/arch/x86/kernel/cpu/perf_event.h
@@ -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);
-- 
1.7.9.5

Re: [PATCH v4 02/18] perf/x86: improve sysfs event mapping with event string

[Jiri Olsa]

On Thu, Dec 20, 2012 at 04:41:32PM +0100, Stephane Eranian wrote:
> 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.
> 
> Signed-off-by: Stephane Eranian <eranian@google.com>

looks ok,

Acked-by: Jiri Olsa <jolsa@redhat.com>

[PATCH v4 03/18] perf/x86: add flags to event constraints

[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.

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 arch/x86/kernel/cpu/perf_event.c              |    2 +-
 arch/x86/kernel/cpu/perf_event.h              |    8 +++++---
 arch/x86/kernel/cpu/perf_event_intel.c        |    5 ++++-
 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, 15 insertions(+), 7 deletions(-)

diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index f4057bb..1a3f693 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1493,7 +1493,7 @@ 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;
diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h
index 3f4380c..3f10cfe 100644
--- a/arch/x86/kernel/cpu/perf_event.h
+++ b/arch/x86/kernel/cpu/perf_event.h
@@ -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.
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 93b9e11..57d6527 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -1367,8 +1367,10 @@ 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) {
+				event->hw.flags |= c->flags;
 				return c;
+			}
 		}
 	}
 
@@ -1413,6 +1415,7 @@ 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);
 }
 
diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c
index 826054a..f30d85b 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_ds.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c
@@ -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;
+			}
 		}
 	}
 
diff --git a/arch/x86/kernel/cpu/perf_event_intel_uncore.c b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
index 3cf3d97..cc39e64 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_uncore.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
@@ -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;
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 6bfb2faa..484cfbc 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -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;
-- 
1.7.9.5

[PATCH v4 04/18] perf, core: Add a concept of a weightened sample

[Stephane Eranian]

From: Andi Kleen <ak@linux.intel.com>

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.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 include/linux/perf_event.h      |    2 ++
 include/uapi/linux/perf_event.h |    8 ++++++--
 kernel/events/core.c            |    6 ++++++
 3 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 484cfbc..bb2429d 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -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,
diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h
index 4f63c05..c52caab 100644
--- a/include/uapi/linux/perf_event.h
+++ b/include/uapi/linux/perf_event.h
@@ -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 */
 };
 
 /*
@@ -198,8 +200,9 @@ enum perf_event_read_format {
 	PERF_FORMAT_TOTAL_TIME_RUNNING		= 1U << 1,
 	PERF_FORMAT_ID				= 1U << 2,
 	PERF_FORMAT_GROUP			= 1U << 3,
+	PERF_FORMAT_WEIGHT			= 1U << 4,
 
-	PERF_FORMAT_MAX = 1U << 4,		/* non-ABI */
+	PERF_FORMAT_MAX = 1U << 5,		/* non-ABI */
 };
 
 #define PERF_ATTR_SIZE_VER0	64	/* sizeof first published struct */
@@ -559,6 +562,7 @@ enum perf_event_type {
 	 *	{ u64			stream_id;} && PERF_SAMPLE_STREAM_ID
 	 *	{ u32			cpu, res; } && PERF_SAMPLE_CPU
 	 *	{ u64			period;   } && PERF_SAMPLE_PERIOD
+	 *	{ u64			weight;   } && PERF_SAMPLE_WEIGHT
 	 *
 	 *	{ struct read_format	values;	  } && PERF_SAMPLE_READ
 	 *
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 301079d..236c58e 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -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;
 
@@ -4080,6 +4083,9 @@ void perf_output_sample(struct perf_output_handle *handle,
 	if (sample_type & PERF_SAMPLE_PERIOD)
 		perf_output_put(handle, data->period);
 
+	if (sample_type & PERF_SAMPLE_WEIGHT)
+		perf_output_put(handle, data->weight);
+
 	if (sample_type & PERF_SAMPLE_READ)
 		perf_output_read(handle, event);
 
-- 
1.7.9.5

[PATCH v4 05/18] perf: add minimal support for PERF_SAMPLE_WEIGHT

[Stephane Eranian]

Ensure we grab the weight from  raw sample struct
and that we can dump it via perf report -D.

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 tools/perf/util/event.h   |    1 +
 tools/perf/util/evsel.c   |    5 +++++
 tools/perf/util/session.c |    3 +++
 3 files changed, 9 insertions(+)

diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h
index 0d573ff..cf52977 100644
--- a/tools/perf/util/event.h
+++ b/tools/perf/util/event.h
@@ -90,6 +90,7 @@ struct perf_sample {
 	u64 period;
 	u32 cpu;
 	u32 raw_size;
+	u64 weight;
 	void *raw_data;
 	struct ip_callchain *callchain;
 	struct branch_stack *branch_stack;
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 1b16dd1..4374c07 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -1058,6 +1058,11 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
 		}
 	}
 
+	if (type & PERF_SAMPLE_WEIGHT) {
+		data->weight= *array;
+		array++;
+	}
+
 	return 0;
 }
 
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index ce6f511..117983e 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -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 *
-- 
1.7.9.5

Re: [PATCH v4 05/18] perf: add minimal support for PERF_SAMPLE_WEIGHT

[Jiri Olsa]

On Thu, Dec 20, 2012 at 04:41:35PM +0100, Stephane Eranian wrote:
> Ensure we grab the weight from  raw sample struct
> and that we can dump it via perf report -D.
> 
> Signed-off-by: Stephane Eranian <eranian@google.com>
> ---
>  tools/perf/util/event.h   |    1 +
>  tools/perf/util/evsel.c   |    5 +++++
>  tools/perf/util/session.c |    3 +++
>  3 files changed, 9 insertions(+)
> 
> diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h
> index 0d573ff..cf52977 100644
> --- a/tools/perf/util/event.h
> +++ b/tools/perf/util/event.h
> @@ -90,6 +90,7 @@ struct perf_sample {
>  	u64 period;
>  	u32 cpu;
>  	u32 raw_size;
> +	u64 weight;
>  	void *raw_data;
>  	struct ip_callchain *callchain;
>  	struct branch_stack *branch_stack;
> diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
> index 1b16dd1..4374c07 100644
> --- a/tools/perf/util/evsel.c
> +++ b/tools/perf/util/evsel.c
> @@ -1058,6 +1058,11 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
>  		}
>  	}
>  
> +	if (type & PERF_SAMPLE_WEIGHT) {
> +		data->weight= *array;
> +		array++;
> +	}
> +

this bit should be placed right after PERF_SAMPLE_PERIOD (where
you placed it in kernel path), otherwise you get wrong data

jirka

Re: [PATCH v4 05/18] perf: add minimal support for PERF_SAMPLE_WEIGHT

[Stephane Eranian]

On Wed, Jan 2, 2013 at 3:07 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> On Thu, Dec 20, 2012 at 04:41:35PM +0100, Stephane Eranian wrote:
>> Ensure we grab the weight from  raw sample struct
>> and that we can dump it via perf report -D.
>>
>> Signed-off-by: Stephane Eranian <eranian@google.com>
>> ---
>>  tools/perf/util/event.h   |    1 +
>>  tools/perf/util/evsel.c   |    5 +++++
>>  tools/perf/util/session.c |    3 +++
>>  3 files changed, 9 insertions(+)
>>
>> diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h
>> index 0d573ff..cf52977 100644
>> --- a/tools/perf/util/event.h
>> +++ b/tools/perf/util/event.h
>> @@ -90,6 +90,7 @@ struct perf_sample {
>>       u64 period;
>>       u32 cpu;
>>       u32 raw_size;
>> +     u64 weight;
>>       void *raw_data;
>>       struct ip_callchain *callchain;
>>       struct branch_stack *branch_stack;
>> diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
>> index 1b16dd1..4374c07 100644
>> --- a/tools/perf/util/evsel.c
>> +++ b/tools/perf/util/evsel.c
>> @@ -1058,6 +1058,11 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
>>               }
>>       }
>>
>> +     if (type & PERF_SAMPLE_WEIGHT) {
>> +             data->weight= *array;
>> +             array++;
>> +     }
>> +
>
> this bit should be placed right after PERF_SAMPLE_PERIOD (where
> you placed it in kernel path), otherwise you get wrong data
>

Thought some more about this. I think this new WEIGHT sample type is
in the wrong place in Andi's patch.
It's new, it should be put at the end of perf_output_sample(). That's
necessary to ensure older versions
of perf (without the knowledge about WEIGHT) can still decode
PERF_SAMPLE_READ in case
PERF_SAMPLE_WEIGHT is present in the perf.data file.

I will update his patch to reflect this change.

Re: [PATCH v4 05/18] perf: add minimal support for PERF_SAMPLE_WEIGHT

[Jiri Olsa]

On Mon, Jan 07, 2013 at 02:10:20PM +0100, Stephane Eranian wrote:
> On Wed, Jan 2, 2013 at 3:07 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> > On Thu, Dec 20, 2012 at 04:41:35PM +0100, Stephane Eranian wrote:
> >> Ensure we grab the weight from  raw sample struct
> >> and that we can dump it via perf report -D.
> >>
> >> Signed-off-by: Stephane Eranian <eranian@google.com>
> >> ---
> >>  tools/perf/util/event.h   |    1 +
> >>  tools/perf/util/evsel.c   |    5 +++++
> >>  tools/perf/util/session.c |    3 +++
> >>  3 files changed, 9 insertions(+)
> >>
> >> diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h
> >> index 0d573ff..cf52977 100644
> >> --- a/tools/perf/util/event.h
> >> +++ b/tools/perf/util/event.h
> >> @@ -90,6 +90,7 @@ struct perf_sample {
> >>       u64 period;
> >>       u32 cpu;
> >>       u32 raw_size;
> >> +     u64 weight;
> >>       void *raw_data;
> >>       struct ip_callchain *callchain;
> >>       struct branch_stack *branch_stack;
> >> diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
> >> index 1b16dd1..4374c07 100644
> >> --- a/tools/perf/util/evsel.c
> >> +++ b/tools/perf/util/evsel.c
> >> @@ -1058,6 +1058,11 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
> >>               }
> >>       }
> >>
> >> +     if (type & PERF_SAMPLE_WEIGHT) {
> >> +             data->weight= *array;
> >> +             array++;
> >> +     }
> >> +
> >
> > this bit should be placed right after PERF_SAMPLE_PERIOD (where
> > you placed it in kernel path), otherwise you get wrong data
> >
> 
> Thought some more about this. I think this new WEIGHT sample type is
> in the wrong place in Andi's patch.
> It's new, it should be put at the end of perf_output_sample(). That's
> necessary to ensure older versions
> of perf (without the knowledge about WEIGHT) can still decode
> PERF_SAMPLE_READ in case
> PERF_SAMPLE_WEIGHT is present in the perf.data file.

agreed

Re: [PATCH v4 05/18] perf: add minimal support for PERF_SAMPLE_WEIGHT

[Stephane Eranian]

Andi,

I also think we should drop PERF_FORMAT_WEIGHT. You added it in your patch
but there is no code to actually perform the read. Will remove it from
your patch
in my series. We can always add it later on.


On Mon, Jan 7, 2013 at 2:27 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> On Mon, Jan 07, 2013 at 02:10:20PM +0100, Stephane Eranian wrote:
>> On Wed, Jan 2, 2013 at 3:07 PM, Jiri Olsa <jolsa@redhat.com> wrote:
>> > On Thu, Dec 20, 2012 at 04:41:35PM +0100, Stephane Eranian wrote:
>> >> Ensure we grab the weight from  raw sample struct
>> >> and that we can dump it via perf report -D.
>> >>
>> >> Signed-off-by: Stephane Eranian <eranian@google.com>
>> >> ---
>> >>  tools/perf/util/event.h   |    1 +
>> >>  tools/perf/util/evsel.c   |    5 +++++
>> >>  tools/perf/util/session.c |    3 +++
>> >>  3 files changed, 9 insertions(+)
>> >>
>> >> diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h
>> >> index 0d573ff..cf52977 100644
>> >> --- a/tools/perf/util/event.h
>> >> +++ b/tools/perf/util/event.h
>> >> @@ -90,6 +90,7 @@ struct perf_sample {
>> >>       u64 period;
>> >>       u32 cpu;
>> >>       u32 raw_size;
>> >> +     u64 weight;
>> >>       void *raw_data;
>> >>       struct ip_callchain *callchain;
>> >>       struct branch_stack *branch_stack;
>> >> diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
>> >> index 1b16dd1..4374c07 100644
>> >> --- a/tools/perf/util/evsel.c
>> >> +++ b/tools/perf/util/evsel.c
>> >> @@ -1058,6 +1058,11 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
>> >>               }
>> >>       }
>> >>
>> >> +     if (type & PERF_SAMPLE_WEIGHT) {
>> >> +             data->weight= *array;
>> >> +             array++;
>> >> +     }
>> >> +
>> >
>> > this bit should be placed right after PERF_SAMPLE_PERIOD (where
>> > you placed it in kernel path), otherwise you get wrong data
>> >
>>
>> Thought some more about this. I think this new WEIGHT sample type is
>> in the wrong place in Andi's patch.
>> It's new, it should be put at the end of perf_output_sample(). That's
>> necessary to ensure older versions
>> of perf (without the knowledge about WEIGHT) can still decode
>> PERF_SAMPLE_READ in case
>> PERF_SAMPLE_WEIGHT is present in the perf.data file.
>
> agreed

[PATCH v4 06/18] perf: add support for PERF_SAMPLE_ADDR in dump_sampple()

[Stephane Eranian]

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

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 tools/perf/util/session.c |    4 ++++
 1 file changed, 4 insertions(+)

diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 117983e..3982fc3 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -1009,6 +1009,10 @@ 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_ADDR)
+                printf(" ..... data: 0x%"PRIx64"\n", sample->addr);
+
 }
 
 static struct machine *
-- 
1.7.9.5

[PATCH v4 07/18] perf: add generic memory sampling interface

[Stephane Eranian]

This patch adds PERF_SAMPLE_COST and PERF_SAMPLE_DSRC.
The first collects a cost associated with the sampled
event. In case of memory access, the cost would be
the latency of the load, otherwise it defaults to
the sampling period.

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.

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 include/linux/perf_event.h      |    2 ++
 include/uapi/linux/perf_event.h |   69 +++++++++++++++++++++++++++++++++++++--
 kernel/events/core.c            |    6 ++++
 3 files changed, 75 insertions(+), 2 deletions(-)

diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index bb2429d..8fe4610 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -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,
diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h
index c52caab..b2a78bb 100644
--- a/include/uapi/linux/perf_event.h
+++ b/include/uapi/linux/perf_event.h
@@ -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 +591,8 @@ enum perf_event_type {
 	 * 	{ u64			size;
 	 * 	  char			data[size];
 	 * 	  u64			dyn_size; } && PERF_SAMPLE_STACK_USER
+	 *
+	 *	{ u64			dsrc;  } && PERF_SAMPLE_DSRC
 	 * };
 	 */
 	PERF_RECORD_SAMPLE			= 9,
@@ -616,4 +618,67 @@ 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 */
+#define PERF_MEM_TLB_SHIFT	26
+
+#define PERF_MEM_S(a, s) \
+	(((u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT)
+
 #endif /* _UAPI_LINUX_PERF_EVENT_H */
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 236c58e..3366f89 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -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,
 		perf_output_sample_ustack(handle,
 					  data->stack_user_size,
 					  data->regs_user.regs);
+
+	if (sample_type & PERF_SAMPLE_DSRC)
+		perf_output_put(handle, data->dsrc.val);
 }
 
 void perf_prepare_sample(struct perf_event_header *header,
-- 
1.7.9.5

Re: [PATCH v4 07/18] perf: add generic memory sampling interface

[Jiri Olsa]

On Thu, Dec 20, 2012 at 04:41:37PM +0100, Stephane Eranian wrote:
> This patch adds PERF_SAMPLE_COST and PERF_SAMPLE_DSRC.

I guess PERF_SAMPLE_COST was replaced by PERF_SAMPLE_WEIGHT added by Andi

jirka

> The first collects a cost associated with the sampled
> event. In case of memory access, the cost would be
> the latency of the load, otherwise it defaults to
> the sampling period.
> 
> 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.
> 
> Signed-off-by: Stephane Eranian <eranian@google.com>

Re: [PATCH v4 07/18] perf: add generic memory sampling interface

[Jiri Olsa]

On Thu, Dec 20, 2012 at 04:41:37PM +0100, Stephane Eranian wrote:
> This patch adds PERF_SAMPLE_COST and PERF_SAMPLE_DSRC.
> The first collects a cost associated with the sampled
> event. In case of memory access, the cost would be
> the latency of the load, otherwise it defaults to
> the sampling period.
> 
> 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.
> 
> Signed-off-by: Stephane Eranian <eranian@google.com>

SNIP

> +
> +/* 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 */
> +#define PERF_MEM_TLB_SHIFT	26
> +
> +#define PERF_MEM_S(a, s) \
> +	(((u64)PERF_MEM_##a##_##s) << PERF_MEM_##a##_SHIFT)

Why dont use enums for this?

jirka

Re: [PATCH v4 07/18] perf: add generic memory sampling interface

[Andi Kleen]

> Why dont use enums for this?

enums can have unpredictable signed/unsignedness issues. #defines for
hardware constants is usually far safer.

-Andi

-- 
ak@linux.intel.com -- Speaking for myself only

[PATCH v4 08/18] perf/x86: add memory profiling via PEBS Load Latency

[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.

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 arch/x86/include/uapi/asm/msr-index.h     |    1 +
 arch/x86/kernel/cpu/perf_event.c          |    7 +-
 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, 179 insertions(+), 11 deletions(-)

diff --git a/arch/x86/include/uapi/asm/msr-index.h b/arch/x86/include/uapi/asm/msr-index.h
index 433a59f..1031604 100644
--- a/arch/x86/include/uapi/asm/msr-index.h
+++ b/arch/x86/include/uapi/asm/msr-index.h
@@ -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
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 1a3f693..c69317a 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1366,8 +1366,8 @@ static __init struct attribute **merge_attr(struct attribute **a,
 	new[j] = NULL;
 	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 = \
@@ -1498,6 +1498,9 @@ 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
diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h
index 3f10cfe..3f91411 100644
--- a/arch/x86/kernel/cpu/perf_event.h
+++ b/arch/x86/kernel/cpu/perf_event.h
@@ -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)
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 57d6527..1d2a7ca 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -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];
@@ -2009,6 +2025,8 @@ __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);
@@ -2049,6 +2067,8 @@ __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);
@@ -2077,6 +2097,8 @@ __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);
@@ -2102,6 +2124,8 @@ __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);
diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c
index f30d85b..759b0c0 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_ds.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c
@@ -24,6 +24,92 @@ struct pebs_record_32 {
 
  */
 
+union intel_x86_pebs_dse {
+	u64 val;
+	struct {
+		unsigned int ld_dse:4;
+		unsigned int ld_stlb_miss:1;
+		unsigned int ld_locked:1;
+		unsigned int ld_reserved:26;
+	};
+	struct {
+		unsigned int st_l1d_hit:1;
+		unsigned int st_reserved1:3;
+		unsigned int st_stlb_miss:1;
+		unsigned int st_locked:1;
+		unsigned int st_reserved2:26;
+	};
+};
+
+
+/*
+ * Map PEBS Load Latency Data Source encodings to generic
+ * memory data source information
+ */
+#define P(a, b) PERF_MEM_S(a, b)
+#define OP_LH (P(OP, LOAD) | P(LVL, HIT))
+#define SNOOP_NONE_MISS (P(SNOOP, NONE) | P(SNOOP, MISS))
+
+static const u64 pebs_data_source[] = {
+	P(OP, LOAD) | P(LVL, MISS) | P(LVL, L3) | P(SNOOP, NA),/* 0x00:ukn L3 */
+	OP_LH | P(LVL, L1) | P(SNOOP, NONE),	/* 0x01: L1 local */
+	OP_LH | P(LVL, LFB)| P(SNOOP, NONE),	/* 0x02: LFB hit */
+	OP_LH | P(LVL, L2) | P(SNOOP, NONE),	/* 0x03: L2 hit */
+	OP_LH | P(LVL, L3) | P(SNOOP, NONE),	/* 0x04: L3 hit */
+	OP_LH | P(LVL, L3) | P(SNOOP, MISS),	/* 0x05: L3 hit, snoop miss */
+	OP_LH | P(LVL, L3) | P(SNOOP, HIT),	/* 0x06: L3 hit, snoop hit */
+	OP_LH | P(LVL, L3) | P(SNOOP, HITM),	/* 0x07: L3 hit, snoop hitm */
+	OP_LH | P(LVL, REM_CCE1) | P(SNOOP, HIT),  /* 0x08: L3 miss snoop hit */
+	OP_LH | P(LVL, REM_CCE1) | P(SNOOP, HITM), /* 0x09: L3 miss snoop hitm*/
+	OP_LH | P(LVL, LOC_RAM)  | P(SNOOP, HIT),  /* 0x0a: L3 miss, shared */
+	OP_LH | P(LVL, REM_RAM1) | P(SNOOP, HIT),  /* 0x0b: L3 miss, shared */
+	OP_LH | P(LVL, LOC_RAM)  | SNOOP_NONE_MISS,/* 0x0c: L3 miss, excl */
+	OP_LH | P(LVL, REM_RAM1) | SNOOP_NONE_MISS,/* 0x0d: L3 miss, excl */
+	OP_LH | P(LVL, IO) | P(SNOOP, NONE), /* 0x0e: I/O */
+	OP_LH | P(LVL,UNC) | P(SNOOP, NONE), /* 0x0f: uncached */
+};
+
+static u64 load_latency_data(u64 status)
+{
+	union intel_x86_pebs_dse dse;
+	u64 val;
+	int model = boot_cpu_data.x86_model;
+	int fam = boot_cpu_data.x86;
+
+	dse.val = status;
+
+	/*
+	 * use the mapping table for bit 0-3
+	 */
+	val = pebs_data_source[dse.ld_dse];
+
+	/*
+	 * Nehalem models do not support TLB, Lock infos
+	 */
+	if (fam == 0x6 && (model == 26 || model == 30
+	    || model == 31 || model == 46)) {
+		val |= P(TLB, NA) | P(LOCK, NA);
+		return val;
+	}
+	/*
+	 * bit 4: TLB access
+	 * 0 = did not miss 2nd level TLB
+	 * 1 = missed 2nd level TLB
+	 */
+	if (dse.ld_stlb_miss)
+		val |= P(TLB, MISS) | P(TLB, L2);
+	else
+		val |= P(TLB, HIT) | P(TLB,L1) | P(TLB, L2);
+
+	/*
+	 * bit 5: locked prefix
+	 */
+	if (dse.ld_locked)
+		val |= P(LOCK, LOCKED);
+
+	return val;
+}
+
 struct pebs_record_core {
 	u64 flags, ip;
 	u64 ax, bx, cx, dx;
@@ -364,7 +450,7 @@ struct event_constraint intel_atom_pebs_event_constraints[] = {
 };
 
 struct event_constraint intel_nehalem_pebs_event_constraints[] = {
-	INTEL_EVENT_CONSTRAINT(0x0b, 0xf),    /* MEM_INST_RETIRED.* */
+	INTEL_PLD_CONSTRAINT(0x100b, 0xf),      /* MEM_INST_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0x0f, 0xf),    /* MEM_UNCORE_RETIRED.* */
 	INTEL_UEVENT_CONSTRAINT(0x010c, 0xf), /* MEM_STORE_RETIRED.DTLB_MISS */
 	INTEL_EVENT_CONSTRAINT(0xc0, 0xf),    /* INST_RETIRED.ANY */
@@ -379,7 +465,7 @@ struct event_constraint intel_nehalem_pebs_event_constraints[] = {
 };
 
 struct event_constraint intel_westmere_pebs_event_constraints[] = {
-	INTEL_EVENT_CONSTRAINT(0x0b, 0xf),    /* MEM_INST_RETIRED.* */
+	INTEL_PLD_CONSTRAINT(0x100b, 0xf),      /* MEM_INST_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0x0f, 0xf),    /* MEM_UNCORE_RETIRED.* */
 	INTEL_UEVENT_CONSTRAINT(0x010c, 0xf), /* MEM_STORE_RETIRED.DTLB_MISS */
 	INTEL_EVENT_CONSTRAINT(0xc0, 0xf),    /* INSTR_RETIRED.* */
@@ -399,7 +485,7 @@ struct event_constraint intel_snb_pebs_event_constraints[] = {
 	INTEL_UEVENT_CONSTRAINT(0x02c2, 0xf), /* UOPS_RETIRED.RETIRE_SLOTS */
 	INTEL_EVENT_CONSTRAINT(0xc4, 0xf),    /* BR_INST_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0xc5, 0xf),    /* BR_MISP_RETIRED.* */
-	INTEL_EVENT_CONSTRAINT(0xcd, 0x8),    /* MEM_TRANS_RETIRED.* */
+	INTEL_PLD_CONSTRAINT(0x01cd, 0x8),    /* MEM_TRANS_RETIRED.LAT_ABOVE_THR */
 	INTEL_EVENT_CONSTRAINT(0xd0, 0xf),    /* MEM_UOP_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0xd1, 0xf),    /* MEM_LOAD_UOPS_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0xd2, 0xf),    /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */
@@ -413,7 +499,7 @@ struct event_constraint intel_ivb_pebs_event_constraints[] = {
         INTEL_UEVENT_CONSTRAINT(0x02c2, 0xf), /* UOPS_RETIRED.RETIRE_SLOTS */
         INTEL_EVENT_CONSTRAINT(0xc4, 0xf),    /* BR_INST_RETIRED.* */
         INTEL_EVENT_CONSTRAINT(0xc5, 0xf),    /* BR_MISP_RETIRED.* */
-        INTEL_EVENT_CONSTRAINT(0xcd, 0x8),    /* MEM_TRANS_RETIRED.* */
+        INTEL_PLD_CONSTRAINT(0x01cd, 0x8),    /* MEM_TRANS_RETIRED.LAT_ABOVE_THR */
         INTEL_EVENT_CONSTRAINT(0xd0, 0xf),    /* MEM_UOP_RETIRED.* */
         INTEL_EVENT_CONSTRAINT(0xd1, 0xf),    /* MEM_LOAD_UOPS_RETIRED.* */
         INTEL_EVENT_CONSTRAINT(0xd2, 0xf),    /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */
@@ -448,6 +534,9 @@ void intel_pmu_pebs_enable(struct perf_event *event)
 	hwc->config &= ~ARCH_PERFMON_EVENTSEL_INT;
 
 	cpuc->pebs_enabled |= 1ULL << hwc->idx;
+
+	if (event->hw.flags & PERF_X86_EVENT_PEBS_LDLAT)
+		cpuc->pebs_enabled |= 1ULL << (hwc->idx + 32);
 }
 
 void intel_pmu_pebs_disable(struct perf_event *event)
@@ -560,20 +649,48 @@ static void __intel_pmu_pebs_event(struct perf_event *event,
 				   struct pt_regs *iregs, void *__pebs)
 {
 	/*
-	 * We cast to pebs_record_core since that is a subset of
-	 * both formats and we don't use the other fields in this
-	 * routine.
+	 * We cast to pebs_record_nhm to get the load latency data
+	 * if extra_reg MSR_PEBS_LD_LAT_THRESHOLD used
 	 */
 	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
-	struct pebs_record_core *pebs = __pebs;
+	struct pebs_record_nhm *pebs = __pebs;
 	struct perf_sample_data data;
 	struct pt_regs regs;
+	u64 sample_type;
+	int fll;
 
 	if (!intel_pmu_save_and_restart(event))
 		return;
 
+	fll = event->hw.flags & PERF_X86_EVENT_PEBS_LDLAT;
+
 	perf_sample_data_init(&data, 0, event->hw.last_period);
 
+	data.period = event->hw.last_period;
+	sample_type = event->attr.sample_type;
+
+	/*
+	 * if PEBS-LL or PreciseStore
+	 */
+	if (fll) {
+		if (sample_type & PERF_SAMPLE_ADDR)
+			data.addr = pebs->dla;
+
+		/*
+		 * Use latency for weight (only avail with PEBS-LL)
+		 */
+		if (fll && (sample_type & PERF_SAMPLE_WEIGHT))
+			data.weight = pebs->lat;
+
+		/*
+		 * data.dsrc encodes the data source
+		 */
+		if (sample_type & PERF_SAMPLE_DSRC) {
+			if (fll)
+				data.dsrc.val = load_latency_data(pebs->dse);
+		}
+	}
+
 	/*
 	 * We use the interrupt regs as a base because the PEBS record
 	 * does not contain a full regs set, specifically it seems to
-- 
1.7.9.5

Re: [PATCH v4 08/18] perf/x86: add memory profiling via PEBS Load Latency

[Jiri Olsa]

On Thu, Dec 20, 2012 at 04:41:38PM +0100, Stephane Eranian wrote:
> 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_WEIGHT ?

> 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.
> 
> Signed-off-by: Stephane Eranian <eranian@google.com>

SNIP

> +/*
> + * Map PEBS Load Latency Data Source encodings to generic
> + * memory data source information
> + */
> +#define P(a, b) PERF_MEM_S(a, b)
> +#define OP_LH (P(OP, LOAD) | P(LVL, HIT))
> +#define SNOOP_NONE_MISS (P(SNOOP, NONE) | P(SNOOP, MISS))
> +

I checked Intel SDM 'Table 18-13. Data Source Encoding for Load Latency Record'
and it seems to be different (below) at some points.. did you use another source?

> +static const u64 pebs_data_source[] = {
> +	P(OP, LOAD) | P(LVL, MISS) | P(LVL, L3) | P(SNOOP, NA),/* 0x00:ukn L3 */
> +	OP_LH | P(LVL, L1) | P(SNOOP, NONE),	/* 0x01: L1 local */
> +	OP_LH | P(LVL, LFB)| P(SNOOP, NONE),	/* 0x02: LFB hit */
> +	OP_LH | P(LVL, L2) | P(SNOOP, NONE),	/* 0x03: L2 hit */
> +	OP_LH | P(LVL, L3) | P(SNOOP, NONE),	/* 0x04: L3 hit */
> +	OP_LH | P(LVL, L3) | P(SNOOP, MISS),	/* 0x05: L3 hit, snoop miss */
> +	OP_LH | P(LVL, L3) | P(SNOOP, HIT),	/* 0x06: L3 hit, snoop hit */

0x6:
L3 HIT. Local or Remote home requests that hit the L3 cache and was serviced by
another processor core with a cross core snoop where modified copies were found.
(HITM).


> +	OP_LH | P(LVL, L3) | P(SNOOP, HITM),	/* 0x07: L3 hit, snoop hitm */

0x7:
Reserved

> +	OP_LH | P(LVL, REM_CCE1) | P(SNOOP, HIT),  /* 0x08: L3 miss snoop hit */
> +	OP_LH | P(LVL, REM_CCE1) | P(SNOOP, HITM), /* 0x09: L3 miss snoop hitm*/

0x9:
Reserved

> +	OP_LH | P(LVL, LOC_RAM)  | P(SNOOP, HIT),  /* 0x0a: L3 miss, shared */
> +	OP_LH | P(LVL, REM_RAM1) | P(SNOOP, HIT),  /* 0x0b: L3 miss, shared */
> +	OP_LH | P(LVL, LOC_RAM)  | SNOOP_NONE_MISS,/* 0x0c: L3 miss, excl */
> +	OP_LH | P(LVL, REM_RAM1) | SNOOP_NONE_MISS,/* 0x0d: L3 miss, excl */
> +	OP_LH | P(LVL, IO) | P(SNOOP, NONE), /* 0x0e: I/O */
> +	OP_LH | P(LVL,UNC) | P(SNOOP, NONE), /* 0x0f: uncached */
> +};

thanks,
jirka

Re: [PATCH v4 08/18] perf/x86: add memory profiling via PEBS Load Latency

[Stephane Eranian]

On Sat, Jan 5, 2013 at 7:43 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> On Thu, Dec 20, 2012 at 04:41:38PM +0100, Stephane Eranian wrote:
>> 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_WEIGHT ?
>
No I switched to using Andi's PERF_SAMPLE_WEIGHT patch for this.
So it's PERF_SAMPLE_COST now.

>> 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.
>>
>> Signed-off-by: Stephane Eranian <eranian@google.com>
>
> SNIP
>
>> +/*
>> + * Map PEBS Load Latency Data Source encodings to generic
>> + * memory data source information
>> + */
>> +#define P(a, b) PERF_MEM_S(a, b)
>> +#define OP_LH (P(OP, LOAD) | P(LVL, HIT))
>> +#define SNOOP_NONE_MISS (P(SNOOP, NONE) | P(SNOOP, MISS))
>> +
>
> I checked Intel SDM 'Table 18-13. Data Source Encoding for Load Latency Record'
> and it seems to be different (below) at some points.. did you use another source?
>
Yeah, and the table is wrong in the SDM. What I have is correct and approved by
Intel.

>> +static const u64 pebs_data_source[] = {
>> +     P(OP, LOAD) | P(LVL, MISS) | P(LVL, L3) | P(SNOOP, NA),/* 0x00:ukn L3 */
>> +     OP_LH | P(LVL, L1) | P(SNOOP, NONE),    /* 0x01: L1 local */
>> +     OP_LH | P(LVL, LFB)| P(SNOOP, NONE),    /* 0x02: LFB hit */
>> +     OP_LH | P(LVL, L2) | P(SNOOP, NONE),    /* 0x03: L2 hit */
>> +     OP_LH | P(LVL, L3) | P(SNOOP, NONE),    /* 0x04: L3 hit */
>> +     OP_LH | P(LVL, L3) | P(SNOOP, MISS),    /* 0x05: L3 hit, snoop miss */
>> +     OP_LH | P(LVL, L3) | P(SNOOP, HIT),     /* 0x06: L3 hit, snoop hit */
>
> 0x6:
> L3 HIT. Local or Remote home requests that hit the L3 cache and was serviced by
> another processor core with a cross core snoop where modified copies were found.
> (HITM).
>
>
>> +     OP_LH | P(LVL, L3) | P(SNOOP, HITM),    /* 0x07: L3 hit, snoop hitm */
>
> 0x7:
> Reserved
>
>> +     OP_LH | P(LVL, REM_CCE1) | P(SNOOP, HIT),  /* 0x08: L3 miss snoop hit */
>> +     OP_LH | P(LVL, REM_CCE1) | P(SNOOP, HITM), /* 0x09: L3 miss snoop hitm*/
>
> 0x9:
> Reserved
>
>> +     OP_LH | P(LVL, LOC_RAM)  | P(SNOOP, HIT),  /* 0x0a: L3 miss, shared */
>> +     OP_LH | P(LVL, REM_RAM1) | P(SNOOP, HIT),  /* 0x0b: L3 miss, shared */
>> +     OP_LH | P(LVL, LOC_RAM)  | SNOOP_NONE_MISS,/* 0x0c: L3 miss, excl */
>> +     OP_LH | P(LVL, REM_RAM1) | SNOOP_NONE_MISS,/* 0x0d: L3 miss, excl */
>> +     OP_LH | P(LVL, IO) | P(SNOOP, NONE), /* 0x0e: I/O */
>> +     OP_LH | P(LVL,UNC) | P(SNOOP, NONE), /* 0x0f: uncached */
>> +};
>
> thanks,
> jirka

Re: [PATCH v4 08/18] perf/x86: add memory profiling via PEBS Load Latency

[Jiri Olsa]

On Sun, Jan 06, 2013 at 09:37:25PM +0100, Stephane Eranian wrote:
> On Sat, Jan 5, 2013 at 7:43 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> > On Thu, Dec 20, 2012 at 04:41:38PM +0100, Stephane Eranian wrote:
> >> 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_WEIGHT ?
> >
> No I switched to using Andi's PERF_SAMPLE_WEIGHT patch for this.
> So it's PERF_SAMPLE_COST now.

Is it? I dont see PERF_SAMPLE_COST being used in the code..

> 
> >> 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.
> >>
> >> Signed-off-by: Stephane Eranian <eranian@google.com>
> >
> > SNIP
> >
> >> +/*
> >> + * Map PEBS Load Latency Data Source encodings to generic
> >> + * memory data source information
> >> + */
> >> +#define P(a, b) PERF_MEM_S(a, b)
> >> +#define OP_LH (P(OP, LOAD) | P(LVL, HIT))
> >> +#define SNOOP_NONE_MISS (P(SNOOP, NONE) | P(SNOOP, MISS))
> >> +
> >
> > I checked Intel SDM 'Table 18-13. Data Source Encoding for Load Latency Record'
> > and it seems to be different (below) at some points.. did you use another source?
> >
> Yeah, and the table is wrong in the SDM. What I have is correct and approved by
> Intel.

Is there any public link?

thanks,
jirka

Re: [PATCH v4 08/18] perf/x86: add memory profiling via PEBS Load Latency

[Stephane Eranian]

On Mon, Jan 7, 2013 at 11:11 AM, Jiri Olsa <jolsa@redhat.com> wrote:
> On Sun, Jan 06, 2013 at 09:37:25PM +0100, Stephane Eranian wrote:
>> On Sat, Jan 5, 2013 at 7:43 PM, Jiri Olsa <jolsa@redhat.com> wrote:
>> > On Thu, Dec 20, 2012 at 04:41:38PM +0100, Stephane Eranian wrote:
>> >> 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_WEIGHT ?
>> >
>> No I switched to using Andi's PERF_SAMPLE_WEIGHT patch for this.
>> So it's PERF_SAMPLE_COST now.
>
Sorry, got it backwards. I was using PERF_SAMPLE_COST originally, but
then Andi added PERF_SAMPLE_WEIGHT. So I am using that one instead
now.


> Is it? I dont see PERF_SAMPLE_COST being used in the code..
>
>>
>> >> 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.
>> >>
>> >> Signed-off-by: Stephane Eranian <eranian@google.com>
>> >
>> > SNIP
>> >
>> >> +/*
>> >> + * Map PEBS Load Latency Data Source encodings to generic
>> >> + * memory data source information
>> >> + */
>> >> +#define P(a, b) PERF_MEM_S(a, b)
>> >> +#define OP_LH (P(OP, LOAD) | P(LVL, HIT))
>> >> +#define SNOOP_NONE_MISS (P(SNOOP, NONE) | P(SNOOP, MISS))
>> >> +
>> >
>> > I checked Intel SDM 'Table 18-13. Data Source Encoding for Load Latency Record'
>> > and it seems to be different (below) at some points.. did you use another source?
>> >
>> Yeah, and the table is wrong in the SDM. What I have is correct and approved by
>> Intel.
>
> Is there any public link?
>
> thanks,
> jirka

[PATCH v4 09/18] perf/x86: export PEBS load latency threshold register to sysfs

[Stephane Eranian]

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

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 arch/x86/kernel/cpu/perf_event_intel.c |    3 +++
 1 file changed, 3 insertions(+)

diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 1d2a7ca..8eced4c 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -1755,6 +1755,8 @@ 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,
@@ -1765,6 +1767,7 @@ 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,
 };
 
-- 
1.7.9.5

[PATCH v4 10/18] perf/x86: add support for PEBS Precise Store

[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.

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 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(-)

diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h
index 3f91411..645b864 100644
--- a/arch/x86/kernel/cpu/perf_event.h
+++ b/arch/x86/kernel/cpu/perf_event.h
@@ -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)
 
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 8eced4c..665e26c 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -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,
 };
 
diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c
index 759b0c0..ed0ca3e 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_ds.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c
@@ -69,6 +69,44 @@ static const u64 pebs_data_source[] = {
 	OP_LH | P(LVL,UNC) | P(SNOOP, NONE), /* 0x0f: uncached */
 };
 
+static u64 precise_store_data(u64 status)
+{
+	union intel_x86_pebs_dse dse;
+	u64 val = P(OP, STORE) | P(SNOOP, NA) | P(LVL, L1) | P(TLB, L2);
+
+	dse.val = status;
+
+	/*
+	 * bit 4: TLB access
+	 * 1 = stored missed 2nd level TLB
+	 *
+	 * so it either hit the walker or the OS
+	 * otherwise hit 2nd level TLB
+	 */
+	if (dse.st_stlb_miss)
+		val |= P(TLB, MISS);
+	else
+		val |= P(TLB, HIT);
+
+	/*
+	 * bit 0: hit L1 data cache
+	 * if not set, then all we know is that
+	 * it missed L1D
+	 */
+	if (dse.st_l1d_hit)
+		val |= P(LVL, HIT);
+	else
+		val |= P(LVL, MISS);
+
+	/*
+	 * bit 5: Locked prefix
+	 */
+	if (dse.st_locked)
+		val |= P(LOCK, LOCKED);
+
+	return val;
+}
+
 static u64 load_latency_data(u64 status)
 {
 	union intel_x86_pebs_dse dse;
@@ -486,6 +524,7 @@ struct event_constraint intel_snb_pebs_event_constraints[] = {
 	INTEL_EVENT_CONSTRAINT(0xc4, 0xf),    /* BR_INST_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0xc5, 0xf),    /* BR_MISP_RETIRED.* */
 	INTEL_PLD_CONSTRAINT(0x01cd, 0x8),    /* MEM_TRANS_RETIRED.LAT_ABOVE_THR */
+	INTEL_PST_CONSTRAINT(0x02cd, 0x8),    /* MEM_TRANS_RETIRED.PRECISE_STORES */
 	INTEL_EVENT_CONSTRAINT(0xd0, 0xf),    /* MEM_UOP_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0xd1, 0xf),    /* MEM_LOAD_UOPS_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0xd2, 0xf),    /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */
@@ -500,6 +539,7 @@ struct event_constraint intel_ivb_pebs_event_constraints[] = {
         INTEL_EVENT_CONSTRAINT(0xc4, 0xf),    /* BR_INST_RETIRED.* */
         INTEL_EVENT_CONSTRAINT(0xc5, 0xf),    /* BR_MISP_RETIRED.* */
         INTEL_PLD_CONSTRAINT(0x01cd, 0x8),    /* MEM_TRANS_RETIRED.LAT_ABOVE_THR */
+	INTEL_PST_CONSTRAINT(0x02cd, 0x8),    /* MEM_TRANS_RETIRED.PRECISE_STORES */
         INTEL_EVENT_CONSTRAINT(0xd0, 0xf),    /* MEM_UOP_RETIRED.* */
         INTEL_EVENT_CONSTRAINT(0xd1, 0xf),    /* MEM_LOAD_UOPS_RETIRED.* */
         INTEL_EVENT_CONSTRAINT(0xd2, 0xf),    /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */
@@ -537,6 +577,8 @@ void intel_pmu_pebs_enable(struct perf_event *event)
 
 	if (event->hw.flags & PERF_X86_EVENT_PEBS_LDLAT)
 		cpuc->pebs_enabled |= 1ULL << (hwc->idx + 32);
+	else if (event->hw.flags & PERF_X86_EVENT_PEBS_ST)
+		cpuc->pebs_enabled |= 1ULL << 63;
 }
 
 void intel_pmu_pebs_disable(struct perf_event *event)
@@ -657,12 +699,13 @@ static void __intel_pmu_pebs_event(struct perf_event *event,
 	struct perf_sample_data data;
 	struct pt_regs regs;
 	u64 sample_type;
-	int fll;
+	int fll, fst;
 
 	if (!intel_pmu_save_and_restart(event))
 		return;
 
 	fll = event->hw.flags & PERF_X86_EVENT_PEBS_LDLAT;
+	fst = event->hw.flags & PERF_X86_EVENT_PEBS_ST;
 
 	perf_sample_data_init(&data, 0, event->hw.last_period);
 
@@ -672,7 +715,7 @@ static void __intel_pmu_pebs_event(struct perf_event *event,
 	/*
 	 * if PEBS-LL or PreciseStore
 	 */
-	if (fll) {
+	if (fll || fst) {
 		if (sample_type & PERF_SAMPLE_ADDR)
 			data.addr = pebs->dla;
 
@@ -688,6 +731,8 @@ static void __intel_pmu_pebs_event(struct perf_event *event,
 		if (sample_type & PERF_SAMPLE_DSRC) {
 			if (fll)
 				data.dsrc.val = load_latency_data(pebs->dse);
+			else
+				data.dsrc.val = precise_store_data(pebs->dse);
 		}
 	}
 
-- 
1.7.9.5

[PATCH v4 11/18] perf tools: add mem access sampling core support

[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
 - cost: access cost
 - locked: access uses locked transaction
 - tlb : TLB access
 - mem : memory level of the access (L1, L2, L3, RAM, ...)
 - snoop: access snoop mode

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 tools/perf/util/event.h   |    1 +
 tools/perf/util/evsel.c   |    4 +
 tools/perf/util/hist.c    |   66 ++++++++-
 tools/perf/util/hist.h    |   13 ++
 tools/perf/util/session.c |   37 ++++++
 tools/perf/util/session.h |    4 +
 tools/perf/util/sort.c    |  324 ++++++++++++++++++++++++++++++++++++++++++++-
 tools/perf/util/sort.h    |   10 +-
 tools/perf/util/symbol.h  |    7 +
 9 files changed, 456 insertions(+), 10 deletions(-)

diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h
index cf52977..ad66b44 100644
--- a/tools/perf/util/event.h
+++ b/tools/perf/util/event.h
@@ -91,6 +91,7 @@ struct perf_sample {
 	u32 cpu;
 	u32 raw_size;
 	u64 weight;
+	u64 dsrc;
 	void *raw_data;
 	struct ip_callchain *callchain;
 	struct branch_stack *branch_stack;
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 4374c07..0049d8e 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -1063,6 +1063,10 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
 		array++;
 	}
 
+	if (type & PERF_SAMPLE_DSRC) {
+		data->dsrc = *array;
+		array++;
+	}
 	return 0;
 }
 
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index cb17e2a..4b95510 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -66,12 +66,16 @@ static void hists__set_unres_dso_col_len(struct hists *hists, int dso)
 void hists__calc_col_len(struct hists *hists, struct hist_entry *h)
 {
 	const unsigned int unresolved_col_width = BITS_PER_LONG / 4;
+	int symlen;
 	u16 len;
 
 	if (h->ms.sym)
 		hists__new_col_len(hists, HISTC_SYMBOL, h->ms.sym->namelen + 4);
-	else
+	else {
+		symlen = unresolved_col_width + 4 + 2;
+		hists__new_col_len(hists, HISTC_SYMBOL, symlen);
 		hists__set_unres_dso_col_len(hists, HISTC_DSO);
+	}
 
 	len = thread__comm_len(h->thread);
 	if (hists__new_col_len(hists, HISTC_COMM, len))
@@ -83,7 +87,6 @@ void hists__calc_col_len(struct hists *hists, struct hist_entry *h)
 	}
 
 	if (h->branch_info) {
-		int symlen;
 		/*
 		 * +4 accounts for '[x] ' priv level info
 		 * +2 account of 0x prefix on raw addresses
@@ -111,7 +114,32 @@ void hists__calc_col_len(struct hists *hists, struct hist_entry *h)
 			hists__new_col_len(hists, HISTC_SYMBOL_TO, symlen);
 			hists__set_unres_dso_col_len(hists, HISTC_DSO_TO);
 		}
+	} else if (h->mem_info) {
+		/*
+		 * +4 accounts for '[x] ' priv level info
+		 * +2 account of 0x prefix on raw addresses
+		 */
+		if (h->mem_info->daddr.sym) {
+			symlen = (int)h->mem_info->daddr.sym->namelen + 4 + unresolved_col_width + 2;
+			hists__new_col_len(hists, HISTC_MEM_DADDR_SYMBOL, symlen);
+		} else {
+			symlen = unresolved_col_width + 4 + 2;
+			hists__new_col_len(hists, HISTC_MEM_DADDR_SYMBOL, symlen);
+		}
+		if (h->mem_info->daddr.map) {
+			symlen = dso__name_len(h->mem_info->daddr.map->dso);
+			hists__new_col_len(hists, HISTC_MEM_DADDR_DSO, symlen);
+		} else {
+			symlen = unresolved_col_width + 4 + 2;
+			hists__set_unres_dso_col_len(hists, HISTC_MEM_DADDR_DSO);
+		}
+		hists__new_col_len(hists, HISTC_MEM_COST, 7);
+		hists__new_col_len(hists, HISTC_MEM_LOCKED, 6);
+		hists__new_col_len(hists, HISTC_MEM_TLB, 22);
+		hists__new_col_len(hists, HISTC_MEM_SNOOP, 12);
+		hists__new_col_len(hists, HISTC_MEM_LVL, 21+3);
 	}
+
 }
 
 void hists__output_recalc_col_len(struct hists *hists, int max_rows)
@@ -235,7 +263,7 @@ void hists__decay_entries_threaded(struct hists *hists,
 static struct hist_entry *hist_entry__new(struct hist_entry *template)
 {
 	size_t callchain_size = symbol_conf.use_callchain ? sizeof(struct callchain_root) : 0;
-	struct hist_entry *he = malloc(sizeof(*he) + callchain_size);
+	struct hist_entry *he = calloc(1, sizeof(*he) + callchain_size);
 
 	if (he != NULL) {
 		*he = *template;
@@ -323,6 +351,35 @@ static struct hist_entry *add_hist_entry(struct hists *hists,
 	return he;
 }
 
+struct hist_entry *__hists__add_mem_entry(struct hists *self,
+					  struct addr_location *al,
+					  struct symbol *sym_parent,
+					  struct mem_info *mi,
+					  u64 weight)
+{
+	struct hist_entry entry = {
+		.thread	= al->thread,
+		.ms = {
+			.map	= al->map,
+			.sym	= al->sym,
+		},
+		.cpu	= al->cpu,
+		.ip	= al->addr,
+		.level	= al->level,
+		.stat = {
+			.period = weight,
+			.nr_events = 1,
+		},
+		.parent = sym_parent,
+		.filtered = symbol__parent_filter(sym_parent),
+		.hists = self,
+		.mem_info = mi,
+		.branch_info = NULL,
+	};
+
+	return add_hist_entry(self, &entry, al, weight);
+}
+
 struct hist_entry *__hists__add_branch_entry(struct hists *self,
 					     struct addr_location *al,
 					     struct symbol *sym_parent,
@@ -346,6 +403,7 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 		.filtered = symbol__parent_filter(sym_parent),
 		.branch_info = bi,
 		.hists	= self,
+		.mem_info = NULL,
 	};
 
 	return add_hist_entry(self, &entry, al, period);
@@ -371,6 +429,8 @@ struct hist_entry *__hists__add_entry(struct hists *self,
 		.parent = sym_parent,
 		.filtered = symbol__parent_filter(sym_parent),
 		.hists	= self,
+		.branch_info = NULL,
+		.mem_info = NULL,
 	};
 
 	return add_hist_entry(self, &entry, al, period);
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index 8b091a5..464d9d2 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -49,6 +49,13 @@ enum hist_column {
 	HISTC_DSO_FROM,
 	HISTC_DSO_TO,
 	HISTC_SRCLINE,
+	HISTC_MEM_DADDR_SYMBOL,
+	HISTC_MEM_DADDR_DSO,
+	HISTC_MEM_COST,
+	HISTC_MEM_LOCKED,
+	HISTC_MEM_TLB,
+	HISTC_MEM_LVL,
+	HISTC_MEM_SNOOP,
 	HISTC_NR_COLS, /* Last entry */
 };
 
@@ -86,6 +93,12 @@ struct hist_entry *__hists__add_branch_entry(struct hists *self,
 					     struct branch_info *bi,
 					     u64 period);
 
+struct hist_entry *__hists__add_mem_entry(struct hists *self,
+					  struct addr_location *al,
+					  struct symbol *sym_parent,
+					  struct mem_info *mi,
+					  u64 period);
+
 void hists__output_resort(struct hists *self);
 void hists__output_resort_threaded(struct hists *hists);
 void hists__collapse_resort(struct hists *self);
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 3982fc3..b27ee86 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -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->cost = sample->weight;
+	mi->dsrc.val = sample->dsrc;
+
+	return mi;
+}
+
 struct branch_info *machine__resolve_bstack(struct machine *self,
 					    struct thread *thr,
 					    struct branch_stack *bs)
@@ -1013,6 +1048,8 @@ 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 *
diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h
index cea133a..f3ea026 100644
--- a/tools/perf/util/session.h
+++ b/tools/perf/util/session.h
@@ -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);
diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
index cfd1c0f..e2e466d 100644
--- a/tools/perf/util/sort.c
+++ b/tools/perf/util/sort.c
@@ -182,11 +182,19 @@ static int _hist_entry__sym_snprintf(struct map *map, struct symbol *sym,
 	}
 
 	ret += repsep_snprintf(bf + ret, size - ret, "[%c] ", level);
-	if (sym)
-		ret += repsep_snprintf(bf + ret, size - ret, "%-*s",
-				       width - ret,
-				       sym->name);
-	else {
+	if (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);
+			ret += repsep_snprintf(bf + ret, size - ret, "%-*s",
+				       width - ret, "");
+		} else {
+			ret += repsep_snprintf(bf + ret, size - ret, "%-*s",
+					       width - ret,
+					       sym->name);
+		}
+	} else {
 		size_t len = BITS_PER_LONG / 4;
 		ret += repsep_snprintf(bf + ret, size - ret, "%-#.*llx",
 				       len, ip);
@@ -469,6 +477,238 @@ static int hist_entry__mispredict_snprintf(struct hist_entry *self, char *bf,
 	return repsep_snprintf(bf, size, "%-*s", width, out);
 }
 
+/* --sort daddr_sym */
+static int64_t
+sort__daddr_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+	struct addr_map_symbol *l = &left->mem_info->daddr;
+	struct addr_map_symbol *r = &right->mem_info->daddr;
+
+	return (int64_t)(r->addr - l->addr);
+}
+
+static int hist_entry__daddr_snprintf(struct hist_entry *self, char *bf,
+				    size_t size, unsigned int width)
+{
+	return _hist_entry__sym_snprintf(self->mem_info->daddr.map,
+					 self->mem_info->daddr.sym,
+					 self->mem_info->daddr.addr,
+					 self->level, bf, size, width);
+}
+
+static int64_t
+sort__dso_daddr_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+	return _sort__dso_cmp(left->mem_info->daddr.map, right->mem_info->daddr.map);
+}
+
+static int hist_entry__dso_daddr_snprintf(struct hist_entry *self, char *bf,
+				    size_t size, unsigned int width)
+{
+	return _hist_entry__dso_snprintf(self->mem_info->daddr.map, bf, size, width);
+}
+
+static int64_t
+sort__cost_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+	u64 cost_l = left->mem_info->cost;
+	u64 cost_r = right->mem_info->cost;
+
+	return (int64_t)(cost_r - cost_l);
+}
+
+static int hist_entry__cost_snprintf(struct hist_entry *self, char *bf,
+				    size_t size, unsigned int width)
+{
+	if (self->mem_info->cost == 0)
+		return repsep_snprintf(bf, size, "%*s", width, "N/A");
+	return repsep_snprintf(bf, size, "%*"PRIu64, width, self->mem_info->cost);
+}
+
+static int64_t
+sort__locked_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+	union perf_mem_dsrc dsrc_l = left->mem_info->dsrc;
+	union perf_mem_dsrc dsrc_r = right->mem_info->dsrc;
+
+	return (int64_t)(dsrc_r.mem_lock - dsrc_l.mem_lock);
+}
+
+static int hist_entry__locked_snprintf(struct hist_entry *self, char *bf,
+				    size_t size, unsigned int width)
+{
+	const char *out = "??";
+	u64 mask = self->mem_info->dsrc.mem_lock;
+
+	if (mask & PERF_MEM_LOCK_NA)
+		out = "N/A";
+	else if (mask & PERF_MEM_LOCK_LOCKED)
+		out = "Yes";
+	else
+		out = "No";
+
+	return repsep_snprintf(bf, size, "%-*s", width, out);
+}
+
+static int64_t
+sort__tlb_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+	union perf_mem_dsrc dsrc_l = left->mem_info->dsrc;
+	union perf_mem_dsrc dsrc_r = right->mem_info->dsrc;
+
+	return (int64_t)(dsrc_r.mem_dtlb - dsrc_l.mem_dtlb);
+}
+
+static const char *tlb_access[] = {
+	"N/A",
+	"HIT",
+	"MISS",
+	"L1",
+	"L2",
+	"Walker",
+	"Fault",
+};
+#define NUM_TLB_ACCESS (sizeof(tlb_access)/sizeof(const char *))
+
+static int hist_entry__tlb_snprintf(struct hist_entry *self, char *bf,
+				    size_t size, unsigned int width)
+{
+	char out[64];
+	size_t sz = sizeof(out) - 1; /* -1 for null termination */
+	size_t l = 0, i;
+	u64 m = self->mem_info->dsrc.mem_dtlb;
+	u64 hit, miss;
+
+	out[0] = '\0';
+
+	hit = m & PERF_MEM_TLB_HIT;
+	miss = m & PERF_MEM_TLB_MISS;
+
+	/* already taken care of */
+	m &= ~(PERF_MEM_TLB_HIT|PERF_MEM_TLB_MISS);
+
+	for (i = 0; m && i < NUM_TLB_ACCESS; i++, m >>= 1) {
+		if (!(m & 0x1))
+			continue;
+		if (l) {
+			strcat(out, " or ");
+			l += 4;
+		}
+		strncat(out, tlb_access[i], sz - l);
+		l += strlen(tlb_access[i]);
+	}
+	if (hit)
+		strncat(out, " hit", sz - l);
+	if (miss)
+		strncat(out, " miss", sz - l);
+
+	return repsep_snprintf(bf, size, "%-*s", width, out);
+}
+
+static int64_t
+sort__lvl_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+	union perf_mem_dsrc dsrc_l = left->mem_info->dsrc;
+	union perf_mem_dsrc dsrc_r = right->mem_info->dsrc;
+
+	return (int64_t)(dsrc_r.mem_lvl - dsrc_l.mem_lvl);
+}
+
+static const char *mem_lvl[] = {
+	"N/A",
+	"HIT",
+	"MISS",
+	"L1",
+	"LFB",
+	"L2",
+	"L3",
+	"Local RAM",
+	"Remote RAM (1 hop)",
+	"Remote RAM (2 hops)",
+	"Remote Cache (1 hop)",
+	"Remote Cache (2 hops)",
+	"I/O",
+	"Uncached",
+};
+#define NUM_MEM_LVL (sizeof(mem_lvl)/sizeof(const char *))
+
+static int hist_entry__lvl_snprintf(struct hist_entry *self, char *bf,
+				    size_t size, unsigned int width)
+{
+	char out[64];
+	size_t sz = sizeof(out) - 1; /* -1 for null termination */
+	size_t i, l = 0;
+	u64 m = self->mem_info->dsrc.mem_lvl;
+	u64 hit, miss;
+
+	out[0] = '\0';
+
+	hit = m & PERF_MEM_LVL_HIT;
+	miss = m & PERF_MEM_LVL_MISS;
+
+	/* already taken care of */
+	m &= ~(PERF_MEM_LVL_HIT|PERF_MEM_LVL_MISS);
+
+	for (i = 0; m && i < NUM_MEM_LVL; i++, m >>= 1) {
+		if (!(m & 0x1))
+			continue;
+		if (l) {
+			strcat(out, " or ");
+			l += 4;
+		}
+		strncat(out, mem_lvl[i], sz - l);
+		l += strlen(mem_lvl[i]);
+	}
+	if (hit)
+		strncat(out, " hit", sz - l);
+	if (miss)
+		strncat(out, " miss", sz - l);
+
+	return repsep_snprintf(bf, size, "%-*s", width, out);
+}
+
+static int64_t
+sort__snoop_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+	union perf_mem_dsrc dsrc_l = left->mem_info->dsrc;
+	union perf_mem_dsrc dsrc_r = right->mem_info->dsrc;
+
+	return (int64_t)(dsrc_r.mem_snoop - dsrc_l.mem_snoop);
+}
+
+static const char *snoop_access[] = {
+	"N/A",
+	"None",
+	"Miss",
+	"Hit",
+	"HitM",
+};
+#define NUM_SNOOP_ACCESS (sizeof(snoop_access)/sizeof(const char *))
+
+static int hist_entry__snoop_snprintf(struct hist_entry *self, char *bf,
+				    size_t size, unsigned int width)
+{
+	char out[64];
+	size_t sz = sizeof(out) - 1; /* -1 for null termination */
+	size_t i, l = 0;
+	u64 m = self->mem_info->dsrc.mem_snoop;
+
+	out[0] = '\0';
+
+	for (i = 0; m && i < NUM_SNOOP_ACCESS; i++, m >>= 1) {
+		if (!(m & 0x1))
+			continue;
+		if (l) {
+			strcat(out, " or ");
+			l += 4;
+		}
+		strncat(out, snoop_access[i], sz - l);
+		l += strlen(snoop_access[i]);
+	}
+
+	return repsep_snprintf(bf, size, "%-*s", width, out);
+}
+
 struct sort_entry sort_mispredict = {
 	.se_header	= "Branch Mispredicted",
 	.se_cmp		= sort__mispredict_cmp,
@@ -476,6 +716,56 @@ struct sort_entry sort_mispredict = {
 	.se_width_idx	= HISTC_MISPREDICT,
 };
 
+struct sort_entry sort_mem_daddr_sym = {
+	.se_header	= "Data Symbol",
+	.se_cmp		= sort__daddr_cmp,
+	.se_snprintf	= hist_entry__daddr_snprintf,
+	.se_width_idx	= HISTC_MEM_DADDR_SYMBOL,
+};
+
+struct sort_entry sort_mem_daddr_dso = {
+	.se_header	= "Data Object",
+	.se_cmp		= sort__dso_daddr_cmp,
+	.se_snprintf	= hist_entry__dso_daddr_snprintf,
+	.se_width_idx	= HISTC_MEM_DADDR_SYMBOL,
+};
+
+struct sort_entry sort_mem_cost = {
+	.se_header	= "Cost",
+	.se_cmp		= sort__cost_cmp,
+	.se_snprintf	= hist_entry__cost_snprintf,
+	.se_width_idx	= HISTC_MEM_COST,
+};
+
+struct sort_entry sort_mem_locked = {
+	.se_header	= "Locked",
+	.se_cmp		= sort__locked_cmp,
+	.se_snprintf	= hist_entry__locked_snprintf,
+	.se_width_idx	= HISTC_MEM_LOCKED,
+};
+
+struct sort_entry sort_mem_tlb = {
+	.se_header	= "TLB access",
+	.se_cmp		= sort__tlb_cmp,
+	.se_snprintf	= hist_entry__tlb_snprintf,
+	.se_width_idx	= HISTC_MEM_TLB,
+};
+
+struct sort_entry sort_mem_lvl = {
+	.se_header	= "Memory access",
+	.se_cmp		= sort__lvl_cmp,
+	.se_snprintf	= hist_entry__lvl_snprintf,
+	.se_width_idx	= HISTC_MEM_LVL,
+};
+
+struct sort_entry sort_mem_snoop = {
+	.se_header	= "Snoop",
+	.se_cmp		= sort__snoop_cmp,
+	.se_snprintf	= hist_entry__snoop_snprintf,
+	.se_width_idx	= HISTC_MEM_SNOOP,
+};
+
+
 struct sort_dimension {
 	const char		*name;
 	struct sort_entry	*entry;
@@ -497,6 +787,13 @@ static struct sort_dimension sort_dimensions[] = {
 	DIM(SORT_CPU, "cpu", sort_cpu),
 	DIM(SORT_MISPREDICT, "mispredict", sort_mispredict),
 	DIM(SORT_SRCLINE, "srcline", sort_srcline),
+	DIM(SORT_MEM_DADDR_SYMBOL, "symbol_daddr", sort_mem_daddr_sym),
+	DIM(SORT_MEM_DADDR_DSO, "dso_daddr", sort_mem_daddr_dso),
+	DIM(SORT_MEM_COST, "cost", sort_mem_cost),
+	DIM(SORT_MEM_LOCKED, "locked", sort_mem_locked),
+	DIM(SORT_MEM_TLB, "tlb", sort_mem_tlb),
+	DIM(SORT_MEM_LVL, "mem", sort_mem_lvl),
+	DIM(SORT_MEM_SNOOP, "snoop", sort_mem_snoop),
 };
 
 int sort_dimension__add(const char *tok)
@@ -520,7 +817,8 @@ int sort_dimension__add(const char *tok)
 			sort__has_parent = 1;
 		} else if (sd->entry == &sort_sym ||
 			   sd->entry == &sort_sym_from ||
-			   sd->entry == &sort_sym_to) {
+			   sd->entry == &sort_sym_to ||
+			   sd->entry == &sort_mem_daddr_sym) {
 			sort__has_sym = 1;
 		}
 
@@ -553,6 +851,20 @@ int sort_dimension__add(const char *tok)
 				sort__first_dimension = SORT_DSO_TO;
 			else if (!strcmp(sd->name, "mispredict"))
 				sort__first_dimension = SORT_MISPREDICT;
+			else if (!strcmp(sd->name, "symbol_daddr"))
+				sort__first_dimension = SORT_MEM_DADDR_SYMBOL;
+			else if (!strcmp(sd->name, "dso_daddr"))
+				sort__first_dimension = SORT_MEM_DADDR_DSO;
+			else if (!strcmp(sd->name, "cost"))
+				sort__first_dimension = SORT_MEM_COST;
+			else if (!strcmp(sd->name, "locked"))
+				sort__first_dimension = SORT_MEM_LOCKED;
+			else if (!strcmp(sd->name, "tlb"))
+				sort__first_dimension = SORT_MEM_TLB;
+			else if (!strcmp(sd->name, "mem_lvl"))
+				sort__first_dimension = SORT_MEM_LVL;
+			else if (!strcmp(sd->name, "snoop"))
+				sort__first_dimension = SORT_MEM_SNOOP;
 		}
 
 		list_add_tail(&sd->entry->list, &hist_entry__sort_list);
diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h
index b4e8c3b..adee5eb 100644
--- a/tools/perf/util/sort.h
+++ b/tools/perf/util/sort.h
@@ -103,7 +103,8 @@ 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)
@@ -137,6 +138,13 @@ enum sort_type {
 	SORT_SYM_TO,
 	SORT_MISPREDICT,
 	SORT_SRCLINE,
+	SORT_MEM_DADDR_SYMBOL,
+	SORT_MEM_DADDR_DSO,
+	SORT_MEM_COST,
+	SORT_MEM_LOCKED,
+	SORT_MEM_TLB,
+	SORT_MEM_LVL,
+	SORT_MEM_SNOOP,
 };
 
 /*
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index de68f98..fdf6371 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -152,6 +152,13 @@ struct branch_info {
 	struct branch_flags flags;
 };
 
+struct mem_info {
+	struct addr_map_symbol iaddr;
+	struct addr_map_symbol daddr;
+	u64	cost;
+	union perf_mem_dsrc dsrc;
+};
+
 struct addr_location {
 	struct thread *thread;
 	struct map    *map;
-- 
1.7.9.5

[PATCH v4 12/18] perf report: add support for mem access profiling

[Stephane Eranian]

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

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

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 tools/perf/builtin-report.c |  131 +++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 127 insertions(+), 4 deletions(-)

diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index fc25100..231165a 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -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 +55,96 @@ 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 = mi->cost;
+	if (!cost)
+		cost = 1;
+
+	/*
+	 * The report shows the percentage of total branches captured
+	 * and not events sampled. Thus we use a pseudo period of 1.
+	 * Only in the newt browser we are doing integrated annotation,
+	 * so we don't allocated the extra space needed because the stdio
+	 * code will not use it.
+	 */
+	he = __hists__add_mem_entry(&evsel->hists, al, parent, mi,
+				    cost);
+	if (!he)
+		return -ENOMEM;
+
+	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 += sample->period;
+	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,
@@ -209,6 +300,12 @@ 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;
@@ -292,7 +389,8 @@ 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;
@@ -305,7 +403,11 @@ 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 cost : %" 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");
 }
 
@@ -319,7 +421,7 @@ 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");
 	}
@@ -596,7 +698,7 @@ 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, mem, cost, 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",
@@ -642,6 +744,7 @@ 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()
 	};
 
@@ -687,6 +790,18 @@ 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 = "cost,mem,sym,dso,symbol_daddr,dso_daddr,snoop,tlb,locked";
+	}
 
 	if (strcmp(input_name, "-") != 0)
 		setup_browser(true);
@@ -758,6 +873,14 @@ 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, "cost", 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);
 	}
-- 
1.7.9.5

[PATCH v4 13/18] perf record: add support for mem access profiling

[Stephane Eranian]

Add the -l option to perf record to enable sampling
access cost sampling.

Data address sampling is obtained via the -d option.

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 tools/perf/builtin-record.c |    2 ++
 tools/perf/perf.h           |    1 +
 tools/perf/util/evsel.c     |    6 ++++++
 3 files changed, 9 insertions(+)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index f3151d3..341f9c1 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -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('l', "cost", &record.opts.weight,
+		    "event cost"),
 	OPT_END()
 };
 
diff --git a/tools/perf/perf.h b/tools/perf/perf.h
index 2c340e7..c976553 100644
--- a/tools/perf/perf.h
+++ b/tools/perf/perf.h
@@ -240,6 +240,7 @@ struct perf_record_opts {
 	bool	     sample_id_all_missing;
 	bool	     exclude_guest_missing;
 	bool	     period;
+	bool         weight;
 	unsigned int freq;
 	unsigned int mmap_pages;
 	unsigned int user_freq;
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 0049d8e..603410c 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -501,6 +501,12 @@ void perf_evsel__config(struct perf_evsel *evsel,
 		attr->sample_type	|= PERF_SAMPLE_CPU;
 	}
 
+	if (opts->weight)
+		attr->sample_type	|= PERF_SAMPLE_WEIGHT;
+
+	if (opts->sample_address)
+		attr->sample_type	|= PERF_SAMPLE_DSRC;
+
 	if (opts->no_delay) {
 		attr->watermark = 0;
 		attr->wakeup_events = 1;
-- 
1.7.9.5

[PATCH v4 14/18] perf tools: add new mem command for memory access profiling

[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

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 tools/perf/Makefile         |    1 +
 tools/perf/builtin-mem.c    |  238 +++++++++++++++++++++++++++++++++++++++++++
 tools/perf/builtin.h        |    1 +
 tools/perf/command-list.txt |    1 +
 tools/perf/perf.c           |    1 +
 tools/perf/util/hist.c      |    1 +
 6 files changed, 243 insertions(+)
 create mode 100644 tools/perf/builtin-mem.c

diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 994f0f6..be8f4697 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -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)
 
diff --git a/tools/perf/builtin-mem.c b/tools/perf/builtin-mem.c
new file mode 100644
index 0000000..c2980b8
--- /dev/null
+++ b/tools/perf/builtin-mem.c
@@ -0,0 +1,238 @@
+#include "builtin.h"
+#include "perf.h"
+
+#include "util/parse-options.h"
+#include "util/trace-event.h"
+#include "util/tool.h"
+#include "util/session.h"
+
+#define MEM_OPERATION_LOAD	"load"
+#define MEM_OPERATION_STORE	"store"
+
+static const char	*mem_operation		= MEM_OPERATION_LOAD;
+
+struct perf_mem {
+	struct perf_tool	tool;
+	char const		*input_name;
+	symbol_filter_t		annotate_init;
+	bool			hide_unresolved;
+	bool			dump_raw;
+	const char		*cpu_list;
+	DECLARE_BITMAP(cpu_bitmap, MAX_NR_CPUS);
+};
+
+static const char * const mem_usage[] = {
+	"perf mem [<options>] {record <command> |report}",
+	NULL
+};
+
+static int __cmd_record(int argc, const char **argv)
+{
+	int rec_argc, i = 0, j;
+	const char **rec_argv;
+	char event[64];
+	int ret;
+
+	rec_argc = argc + 4;
+	rec_argv = calloc(rec_argc + 1, sizeof(char *));
+	if (!rec_argv)
+		return -1;
+
+	rec_argv[i++] = strdup("record");
+	if (!strcmp(mem_operation, MEM_OPERATION_LOAD))
+		rec_argv[i++] = strdup("-l");
+	rec_argv[i++] = strdup("-d");
+	rec_argv[i++] = strdup("-e");
+
+	if (strcmp(mem_operation, MEM_OPERATION_LOAD))
+		sprintf(event, "cpu/mem-stores/pp");
+	else
+		sprintf(event, "cpu/mem-loads/pp");
+
+	rec_argv[i++] = strdup(event);
+	for (j = 1; j < argc; j++, i++)
+		rec_argv[i] = argv[j];
+
+	ret = cmd_record(i, rec_argv, NULL);
+	free(rec_argv);
+	return ret;
+}
+
+static int
+dump_raw_samples(struct perf_tool *tool,
+		 union perf_event *event,
+		 struct perf_sample *sample,
+		 struct perf_evsel *evsel __maybe_unused,
+		 struct machine *machine)
+{
+	struct perf_mem *mem = container_of(tool, struct perf_mem, tool);
+	struct addr_location al;
+	const char *fmt;
+
+	if (perf_event__preprocess_sample(event, machine, &al, sample,
+				mem->annotate_init) < 0) {
+		fprintf(stderr, "problem processing %d event, skipping it.\n",
+				event->header.type);
+		return -1;
+	}
+
+	if (al.filtered || (mem->hide_unresolved && al.sym == NULL))
+		return 0;
+
+	if (al.map != NULL)
+		al.map->dso->hit = 1;
+
+	if (symbol_conf.field_sep) {
+		fmt = "%d%s%d%s0x%"PRIx64"%s0x%"PRIx64"%s%"PRIu64
+		      "%s0x%"PRIx64"%s%s:%s\n";
+	} else {
+		fmt = "%5d%s%5d%s0x%016"PRIx64"%s0x016%"PRIx64
+		      "%s%5"PRIu64"%s0x%06"PRIx64"%s%s:%s\n";
+		symbol_conf.field_sep = " ";
+	}
+
+	printf( fmt,
+		sample->pid,
+		symbol_conf.field_sep,
+		sample->tid,
+		symbol_conf.field_sep,
+		event->ip.ip,
+		symbol_conf.field_sep,
+		sample->addr,
+		symbol_conf.field_sep,
+		sample->weight,
+		symbol_conf.field_sep,
+		sample->dsrc,
+		symbol_conf.field_sep,
+		al.map ? (al.map->dso ? al.map->dso->long_name : "???") : "???",
+		al.sym ? al.sym->name : "???");
+
+	return 0;
+}
+
+static int process_sample_event(struct perf_tool *tool,
+				union perf_event *event,
+				struct perf_sample *sample,
+                                struct perf_evsel *evsel,
+				struct machine *machine)
+{
+	return dump_raw_samples(tool, event, sample, evsel, machine);
+}
+
+static int report_raw_events(struct perf_mem *mem)
+{
+	int err = -EINVAL;
+	int ret;
+	struct perf_session *session = perf_session__new(input_name, O_RDONLY,
+							 0, false, &mem->tool);
+
+	if (session == NULL)
+		return -ENOMEM;
+
+	if (mem->cpu_list) {
+		ret = perf_session__cpu_bitmap(session, mem->cpu_list,
+					       mem->cpu_bitmap);
+		if (ret)
+			goto out_delete;
+	}
+
+	if (symbol__init() < 0)
+		return -1;
+
+	printf("# PID, TID, IP, ADDR, COST, DSRC, SYMBOL\n");
+
+	err = perf_session__process_events(session, &mem->tool);
+	if (err)
+		return err;
+
+	return 0;
+
+out_delete:
+	perf_session__delete(session);
+	return err;
+}
+
+static int report_events(int argc, const char **argv, struct perf_mem *mem)
+{
+	const char **rep_argv;
+	int ret, i = 0, j, rep_argc;
+
+	if (mem->dump_raw)
+		return report_raw_events(mem);
+
+	rep_argc = argc + 3;
+	rep_argv = calloc(rep_argc + 1, sizeof(char *));
+	if (!rep_argv)
+		return -1;
+
+	rep_argv[i++] = strdup("report");
+	rep_argv[i++] = strdup("--mem-mode");
+	rep_argv[i++] = strdup("-n"); /* display number of samples */
+
+	/*
+	 * there is no cost associated with stores, so don't print
+	 * the column
+	 */
+	if (strcmp(mem_operation, MEM_OPERATION_LOAD))
+		rep_argv[i++] = strdup("--sort=mem,sym,dso,symbol_daddr,dso_daddr,tlb,locked");
+
+	for (j = 1; j < argc; j++, i++)
+		rep_argv[i] = argv[j];
+
+	ret = cmd_report(i, rep_argv, NULL);
+	free(rep_argv);
+	return ret;
+}
+
+int cmd_mem(int argc, const char **argv, const char *prefix __maybe_unused)
+{
+	struct stat st;
+	struct perf_mem mem = {
+		.tool = {
+			.sample		= process_sample_event,
+			.mmap		= perf_event__process_mmap,
+			.comm		= perf_event__process_comm,
+			.lost		= perf_event__process_lost,
+			.fork		= perf_event__process_fork,
+			.build_id	= perf_event__process_build_id,
+			.ordered_samples= true,
+		},
+		.input_name		 = "perf.data",
+	};
+	const struct option mem_options[] = {
+		OPT_STRING('t', "type", &mem_operation,
+			   "type", "memory operations(load/store)"),
+		OPT_BOOLEAN('D', "dump-raw-samples", &mem.dump_raw,
+			    "dump raw samples in ASCII"),
+		OPT_BOOLEAN('U', "hide-unresolved", &mem.hide_unresolved,
+			    "Only display entries resolved to a symbol"),
+		OPT_STRING('i', "input", &input_name, "file", "input file name"),
+		OPT_STRING('C', "cpu", &mem.cpu_list, "cpu", "list of cpus to profile"),
+		OPT_STRING('x', "field-separator", &symbol_conf.field_sep, "separator",
+			   "separator for columns, no spaces will be added between "
+			   "columns '.' is reserved."),
+		OPT_END()
+	};
+
+	argc = parse_options(argc, argv, mem_options, mem_usage,
+                        PARSE_OPT_STOP_AT_NON_OPTION);
+
+	if (!argc || !(strncmp(argv[0], "rec", 3) || mem_operation))
+		usage_with_options(mem_usage, mem_options);
+
+	if (!mem.input_name || !strlen(mem.input_name)) {
+		if (!fstat(STDIN_FILENO, &st) && S_ISFIFO(st.st_mode))
+			mem.input_name = "-";
+		else
+			mem.input_name = "perf.data";
+	}
+
+        if (!strncmp(argv[0], "rec", 3))
+		return __cmd_record(argc, argv);
+	else if (!strncmp(argv[0], "rep", 3))
+		return report_events(argc, argv, &mem);
+	else
+		usage_with_options(mem_usage, mem_options);
+
+	return 0;
+}
diff --git a/tools/perf/builtin.h b/tools/perf/builtin.h
index 08143bd..b210d62 100644
--- a/tools/perf/builtin.h
+++ b/tools/perf/builtin.h
@@ -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
diff --git a/tools/perf/command-list.txt b/tools/perf/command-list.txt
index 3e86bbd..2c5b621 100644
--- a/tools/perf/command-list.txt
+++ b/tools/perf/command-list.txt
@@ -24,3 +24,4 @@ perf-kmem			mainporcelain common
 perf-lock			mainporcelain common
 perf-kvm			mainporcelain common
 perf-test			mainporcelain common
+perf-mem			mainporcelain common
diff --git a/tools/perf/perf.c b/tools/perf/perf.c
index 0f661fb..682340e 100644
--- a/tools/perf/perf.c
+++ b/tools/perf/perf.c
@@ -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 {
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 4b95510..efd9929 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -473,6 +473,7 @@ 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);
 }
 
-- 
1.7.9.5

Re: [PATCH v4 14/18] perf tools: add new mem command for memory access profiling

[Jiri Olsa]

On Thu, Dec 20, 2012 at 04:41:44PM +0100, Stephane Eranian wrote:
> 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

I get following segfault occasionally:

  $ perf mem -t load rec ls
  $ perf mem -t load rep --stdio
    Failed to open /etc/ld.so.cache, continuing without symbols
    Segmentation fault
  $ gdb ./perf
  GNU gdb (GDB) Fedora (7.4.50.20120120-52.fc17)
  Copyright (C) 2012 Free Software Foundation, Inc.
  License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
  This is free software: you are free to change and redistribute it.
  There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
  and "show warranty" for details.
  This GDB was configured as "x86_64-redhat-linux-gnu".
  For bug reporting instructions, please see:
  <http://www.gnu.org/software/gdb/bugs/>...
  Reading symbols from /home/jolsa/kernel/tip-local/tools/perf/perf...done.
  (gdb) r  mem -t load rep --stdio
  Starting program: /home/jolsa/kernel/tip-local/tools/perf/perf mem -t load rep --stdio
  [Thread debugging using libthread_db enabled]
  Using host libthread_db library "/lib64/libthread_db.so.1".
  Detaching after fork from child process 3130.
  Failed to open /etc/ld.so.cache, continuing without symbols

  Program received signal SIGSEGV, Segmentation fault.
  dso__name_len (dso=0x68100142) at util/dso.c:389
  389             return dso->short_name_len;
  Missing separate debuginfos, use: debuginfo-install atk-2.4.0-1.fc17.x86_64 audit-libs-2.2.1-1.fc17.x86_64 bzip2-libs-1.0.6-4.fc17.x86_64 cairo-1.10.2-7.fc17.x86_64 elfutils-libelf-0.154-2.fc17.x86_64 elfutils-libs-0.154-2.fc17.x86_64 expat-2.1.0-1.fc17.x86_64 fontconfig-2.8.0-8.fc17.x86_64 freetype-2.4.8-3.fc17.x86_64 gdk-pixbuf2-2.26.1-1.fc17.x86_64 glib2-2.32.4-2.fc17.x86_64 glibc-2.15-58.fc17.x86_64 gtk2-2.24.13-1.fc17.x86_64 libX11-1.5.0-2.fc17.x86_64 libXau-1.0.6-3.fc17.x86_64 libXcomposite-0.4.3-3.fc17.x86_64 libXcursor-1.1.13-1.fc17.x86_64 libXdamage-1.1.3-3.fc17.x86_64 libXext-1.3.1-1.fc17.x86_64 libXfixes-5.0-2.fc17.x86_64 libXi-1.6.1-1.fc17.x86_64 libXinerama-1.1.2-1.fc17.x86_64 libXrandr-1.3.1-3.fc17.x86_64 libXrender-0.9.7-1.fc17.x86_64 libffi-3.0.10-2.fc17.x86_64 libgcc-4.7.2-2.fc17.x86_64 libpng-1.5.10-1.fc17.x86_64 libselinux-2.1.10-3.fc17.x86_64 libunwind-1.0.1-3.fc17.x86_64 libxcb-1.9-1.fc17.x86_64 newt-0.52.14-2.fc17.x86_64 numactl-libs-2.0.7-6.fc17.x86_64 pango-1.30.0-1.fc17.x86_64 pixman-0.24.4-2.fc17.x86_64 python-libs-2.7.3-7.2.fc17.x86_64 slang-2.2.4-3.fc17.x86_64 xz-libs-5.1.2-1alpha.fc17.x86_64 zlib-1.2.5-7.fc17.x86_64
  (gdb) bt
  #0  dso__name_len (dso=0x68100142) at util/dso.c:389
  #1  0x000000000048d8cd in hists__calc_col_len (hists=hists@entry=0x8a1c10, h=h@entry=0xd7de50) at util/hist.c:130
  #2  0x000000000048dc47 in hists__inc_nr_entries (h=0xd7de50, hists=0x8a1c10) at util/hist.c:285
  #3  hists__inc_nr_entries (h=0xd7de50, hists=0x8a1c10) at util/hist.c:613
  #4  __hists__output_resort (hists=hists@entry=0x8a1c10, threaded=threaded@entry=false) at util/hist.c:639
  #5  0x000000000048e2de in hists__output_resort (hists=hists@entry=0x8a1c10) at util/hist.c:645
  #6  0x0000000000436240 in __cmd_report (rep=0x7fffffffd330) at builtin-report.c:521
  #7  cmd_report (argc=<optimized out>, argv=argv@entry=0x89fd10, prefix=prefix@entry=0x0) at builtin-report.c:888
  #8  0x0000000000449e30 in report_events (mem=0x7fffffffde90, argv=0x7fffffffe4a0, argc=<optimized out>) at builtin-mem.c:182
  #9  cmd_mem (argc=<optimized out>, argv=0x7fffffffe4a0, prefix=<optimized out>) at builtin-mem.c:233
  #10 0x00000000004191d3 in run_builtin (p=p@entry=0x7bf768, argc=argc@entry=5, argv=argv@entry=0x7fffffffe4a0) at perf.c:319
  #11 0x0000000000418976 in handle_internal_command (argv=0x7fffffffe4a0, argc=5) at perf.c:367
  #12 run_argv (argv=0x7fffffffe290, argcp=0x7fffffffe29c) at perf.c:411
  #13 main (argc=5, argv=0x7fffffffe4a0) at perf.c:513


jirka

Re: [PATCH v4 14/18] perf tools: add new mem command for memory access profiling

[Stephane Eranian]

On Wed, Jan 2, 2013 at 3:16 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> On Thu, Dec 20, 2012 at 04:41:44PM +0100, Stephane Eranian wrote:
>> 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
>
> I get following segfault occasionally:
>
>   $ perf mem -t load rec ls
>   $ perf mem -t load rep --stdio
>     Failed to open /etc/ld.so.cache, continuing without symbols
>     Segmentation fault
>   $ gdb ./perf
>   GNU gdb (GDB) Fedora (7.4.50.20120120-52.fc17)
>   Copyright (C) 2012 Free Software Foundation, Inc.
>   License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
>   This is free software: you are free to change and redistribute it.
>   There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
>   and "show warranty" for details.
>   This GDB was configured as "x86_64-redhat-linux-gnu".
>   For bug reporting instructions, please see:
>   <http://www.gnu.org/software/gdb/bugs/>...
>   Reading symbols from /home/jolsa/kernel/tip-local/tools/perf/perf...done.
>   (gdb) r  mem -t load rep --stdio
>   Starting program: /home/jolsa/kernel/tip-local/tools/perf/perf mem -t load rep --stdio
>   [Thread debugging using libthread_db enabled]
>   Using host libthread_db library "/lib64/libthread_db.so.1".
>   Detaching after fork from child process 3130.
>   Failed to open /etc/ld.so.cache, continuing without symbols
>
Can't reproduce this problem on my 3.8.0-rc1 patch series kernel, but
then I am not
using the same distro as you. Is your bin/ls statically linked and stripped?


>   Program received signal SIGSEGV, Segmentation fault.
>   dso__name_len (dso=0x68100142) at util/dso.c:389
>   389             return dso->short_name_len;
>   Missing separate debuginfos, use: debuginfo-install atk-2.4.0-1.fc17.x86_64 audit-libs-2.2.1-1.fc17.x86_64 bzip2-libs-1.0.6-4.fc17.x86_64 cairo-1.10.2-7.fc17.x86_64 elfutils-libelf-0.154-2.fc17.x86_64 elfutils-libs-0.154-2.fc17.x86_64 expat-2.1.0-1.fc17.x86_64 fontconfig-2.8.0-8.fc17.x86_64 freetype-2.4.8-3.fc17.x86_64 gdk-pixbuf2-2.26.1-1.fc17.x86_64 glib2-2.32.4-2.fc17.x86_64 glibc-2.15-58.fc17.x86_64 gtk2-2.24.13-1.fc17.x86_64 libX11-1.5.0-2.fc17.x86_64 libXau-1.0.6-3.fc17.x86_64 libXcomposite-0.4.3-3.fc17.x86_64 libXcursor-1.1.13-1.fc17.x86_64 libXdamage-1.1.3-3.fc17.x86_64 libXext-1.3.1-1.fc17.x86_64 libXfixes-5.0-2.fc17.x86_64 libXi-1.6.1-1.fc17.x86_64 libXinerama-1.1.2-1.fc17.x86_64 libXrandr-1.3.1-3.fc17.x86_64 libXrender-0.9.7-1.fc17.x86_64 libffi-3.0.10-2.fc17.x86_64 libgcc-4.7.2-2.fc17.x86_64 libpng-1.5.10-1.fc17.x86_64 libselinux-2.1.10-3.fc17.x86_64 libunwind-1.0.1-3.fc17.x86_64 libxcb-1.9-1.fc17.x86_64 newt-0.52.14-2.fc17.x86_64 numactl-libs-2.0.7-6.fc17.x86_64 pango-1.30.0-1.fc17.x86_64 pixman-0.24.4-2.fc17.x86_64 python-libs-2.7.3-7.2.fc17.x86_64 slang-2.2.4-3.fc17.x86_64 xz-libs-5.1.2-1alpha.fc17.x86_64 zlib-1.2.5-7.fc17.x86_64
>   (gdb) bt
>   #0  dso__name_len (dso=0x68100142) at util/dso.c:389
>   #1  0x000000000048d8cd in hists__calc_col_len (hists=hists@entry=0x8a1c10, h=h@entry=0xd7de50) at util/hist.c:130
>   #2  0x000000000048dc47 in hists__inc_nr_entries (h=0xd7de50, hists=0x8a1c10) at util/hist.c:285
>   #3  hists__inc_nr_entries (h=0xd7de50, hists=0x8a1c10) at util/hist.c:613
>   #4  __hists__output_resort (hists=hists@entry=0x8a1c10, threaded=threaded@entry=false) at util/hist.c:639
>   #5  0x000000000048e2de in hists__output_resort (hists=hists@entry=0x8a1c10) at util/hist.c:645
>   #6  0x0000000000436240 in __cmd_report (rep=0x7fffffffd330) at builtin-report.c:521
>   #7  cmd_report (argc=<optimized out>, argv=argv@entry=0x89fd10, prefix=prefix@entry=0x0) at builtin-report.c:888
>   #8  0x0000000000449e30 in report_events (mem=0x7fffffffde90, argv=0x7fffffffe4a0, argc=<optimized out>) at builtin-mem.c:182
>   #9  cmd_mem (argc=<optimized out>, argv=0x7fffffffe4a0, prefix=<optimized out>) at builtin-mem.c:233
>   #10 0x00000000004191d3 in run_builtin (p=p@entry=0x7bf768, argc=argc@entry=5, argv=argv@entry=0x7fffffffe4a0) at perf.c:319
>   #11 0x0000000000418976 in handle_internal_command (argv=0x7fffffffe4a0, argc=5) at perf.c:367
>   #12 run_argv (argv=0x7fffffffe290, argcp=0x7fffffffe29c) at perf.c:411
>   #13 main (argc=5, argv=0x7fffffffe4a0) at perf.c:513
>
>
> jirka

Re: [PATCH v4 14/18] perf tools: add new mem command for memory access profiling

[Jiri Olsa]

On Mon, Jan 07, 2013 at 02:04:58PM +0100, Stephane Eranian wrote:
> On Wed, Jan 2, 2013 at 3:16 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> > On Thu, Dec 20, 2012 at 04:41:44PM +0100, Stephane Eranian wrote:
> >> 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
> >
> > I get following segfault occasionally:
> >
> >   $ perf mem -t load rec ls
> >   $ perf mem -t load rep --stdio
> >     Failed to open /etc/ld.so.cache, continuing without symbols
> >     Segmentation fault
> >   $ gdb ./perf
> >   GNU gdb (GDB) Fedora (7.4.50.20120120-52.fc17)
> >   Copyright (C) 2012 Free Software Foundation, Inc.
> >   License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
> >   This is free software: you are free to change and redistribute it.
> >   There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
> >   and "show warranty" for details.
> >   This GDB was configured as "x86_64-redhat-linux-gnu".
> >   For bug reporting instructions, please see:
> >   <http://www.gnu.org/software/gdb/bugs/>...
> >   Reading symbols from /home/jolsa/kernel/tip-local/tools/perf/perf...done.
> >   (gdb) r  mem -t load rep --stdio
> >   Starting program: /home/jolsa/kernel/tip-local/tools/perf/perf mem -t load rep --stdio
> >   [Thread debugging using libthread_db enabled]
> >   Using host libthread_db library "/lib64/libthread_db.so.1".
> >   Detaching after fork from child process 3130.
> >   Failed to open /etc/ld.so.cache, continuing without symbols
> >
> Can't reproduce this problem on my 3.8.0-rc1 patch series kernel, but
> then I am not
> using the same distro as you. Is your bin/ls statically linked and stripped?

[jolsa@dhcp-26-214 ~]$ file /bin/ls
/bin/ls: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.32, stripped
[jolsa@dhcp-26-214 ~]$ ldd /bin/ls
/usr/bin/ldd: line 163: /lib/ld-linux.so.2: cannot execute binary file
        linux-vdso.so.1 =>  (0x00007fff499ff000)
        libselinux.so.1 => /lib64/libselinux.so.1 (0x00007f0eebdc1000)
        librt.so.1 => /lib64/librt.so.1 (0x00007f0eebbb9000)
        libcap.so.2 => /lib64/libcap.so.2 (0x00007f0eeb9b4000)
        libacl.so.1 => /lib64/libacl.so.1 (0x00007f0eeb7ac000)
        libc.so.6 => /lib64/libc.so.6 (0x00007f0eeb3f4000)
        libdl.so.2 => /lib64/libdl.so.2 (0x00007f0eeb1ef000)
        /lib64/ld-linux-x86-64.so.2 (0x00007f0eebff5000)
        libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f0eeafd3000)
        libattr.so.1 => /lib64/libattr.so.1 (0x00007f0eeadcf000)
[jolsa@dhcp-26-214 ~]$ cat /etc/fedora-release 
Fedora release 16 (Verne)

I planed to look on it if I dont hear from you ;-) I'll check and let you know

jirka

[PATCH v4 15/18] perf: add PERF_RECORD_MISC_MMAP_DATA to RECORD_MMAP

[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.

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 include/uapi/linux/perf_event.h |    1 +
 kernel/events/core.c            |    3 +++
 2 files changed, 4 insertions(+)

diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h
index b2a78bb..0fa3feb 100644
--- a/include/uapi/linux/perf_event.h
+++ b/include/uapi/linux/perf_event.h
@@ -446,6 +446,7 @@ 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
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 3366f89..3a19a45 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -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();
-- 
1.7.9.5

[PATCH v4 16/18] perf tools: detect data vs. text mappings

[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.

Signed-off-by: Stephane Eranian <eranian@google.com>
---
 tools/perf/util/machine.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c
index 1f09d05..d1c3e48 100644
--- a/tools/perf/util/machine.c
+++ b/tools/perf/util/machine.c
@@ -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;
 
-- 
1.7.9.5

[PATCH v4 17/18] perf tools: Ignore ABS symbols when loading data maps

[Stephane Eranian]

From: Namhyung Kim <namhyung.kim@lge.com>

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 <eranian@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/symbol-elf.c |    3 +++
 1 file changed, 3 insertions(+)

diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
index db0cc92..00cf128 100644
--- a/tools/perf/util/symbol-elf.c
+++ b/tools/perf/util/symbol-elf.c
@@ -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;
-- 
1.7.9.5

[PATCH v4 18/18] perf tools: Fix output of symbol_daddr offset

[Stephane Eranian]

From: Namhyung Kim <namhyung.kim@lge.com>

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 <eranian@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/sort.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
index e2e466d..166480c 100644
--- a/tools/perf/util/sort.c
+++ b/tools/perf/util/sort.c
@@ -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 {
-- 
1.7.9.5

Re: [PATCH v4 00/18] perf: add memory access sampling support

[Jiri Olsa]

On Thu, Dec 20, 2012 at 04:41:30PM +0100, Stephane Eranian wrote:
> This patch series had a new feature to the kernel perf_events
> interface and corresponding user level tool, perf.

hi,
would you have git tree branch?

thanks,
jirka

Re: [PATCH v4 00/18] perf: add memory access sampling support

[Stephane Eranian]

Hi,

It's relative to tip.git.
Forgot to mention it in the Changelog.


On Thu, Dec 20, 2012 at 5:05 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> On Thu, Dec 20, 2012 at 04:41:30PM +0100, Stephane Eranian wrote:
>> This patch series had a new feature to the kernel perf_events
>> interface and corresponding user level tool, perf.
>
> hi,
> would you have git tree branch?
>
> thanks,
> jirka

Re: [PATCH v4 00/18] perf: add memory access sampling support

[Jiri Olsa]

On Thu, Dec 20, 2012 at 05:07:21PM +0100, Stephane Eranian wrote:
> Hi,
> 
> It's relative to tip.git.
> Forgot to mention it in the Changelog.

well I meant git branch I could pull from, so I dont need to do the
mutt patch store/apply stuff.. no big deal though ;-)

jirka

Re: [PATCH v4 00/18] perf: add memory access sampling support

[Stephane Eranian]

On Thu, Dec 20, 2012 at 5:12 PM, Jiri Olsa <jolsa@redhat.com> wrote:
> On Thu, Dec 20, 2012 at 05:07:21PM +0100, Stephane Eranian wrote:
>> Hi,
>>
>> It's relative to tip.git.
>> Forgot to mention it in the Changelog.
>
> well I meant git branch I could pull from, so I dont need to do the
> mutt patch store/apply stuff.. no big deal though ;-)
>
Don't have a git tree on kernel.org unfortunately.
So you have to save each one and then use git am < patchN
> jirka