Skip to contentSkip to conversation

x86, nmi: Various fixes and cleanups

PATCH Superseded Don Zickus 3 participants 16 messages May 19, 2014 – Jun 12, 2014 perf events

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Don Zickus

author
#0 lore raw

v3 cover x86, nmi: Various fixes and cleanups

I started this patch by fixing a performance problem with the GHES NMI handler and then things evolved to more patches as I was poking around in the code.

The main focus was moving the GHES NMI driver to its own NMI subtype to avoid slowing down perf handling. Then I decided to move the default external NMI handler to its own routine. Then finally, I needed to see which NMI handlers were registered so I hacked up /proc/interrupts to show me.

Tested mostly on HP boxes that have GHES enabled and having the iLO send NMIs to panic the box (using hpwdt driver). Ran perf on other GHES enabled boxes to test performance results.

Note: I haven't been able to reproduce Jiri's perf hang yet, so I don't have a patch for that particular problem yet.

V3: added patch to get cacheline size at runtime
updated some changelogs
fix a test failure (after install libunwind-devel) fix header column length

V2: adding irq_work items to handled possible lost NMIs (new patch 1)
modified output of /proc/interrupts based on feedback (patch 6)

Don Zickus (5):
x86, nmi: Add new nmi type 'external'
x86, nmi: Add boot line option 'panic_on_unrecovered_nmi' and

    'panic_on_io_nmi'

x86, nmi: Remove 'reason' value from unknown nmi output x86, nmi: Move default external NMI handler to its own routine x86, nmi: Add better NMI stats to /proc/interrupts and show handlers

 Documentation/kernel-parameters.txt |    9 ++
 arch/x86/include/asm/nmi.h          |    5 +
 arch/x86/kernel/irq.c               |    3 +
 arch/x86/kernel/nmi.c               |  199 ++++++++++++++++++++++++++--------
 drivers/acpi/apei/ghes.c            |    4 +-
 drivers/watchdog/hpwdt.c            |   24 +++--
 6 files changed, 187 insertions(+), 57 deletions(-)

*** BLURB HERE ***

Don Zickus (6):
x86, nmi: Implement delayed irq_work mechanism to handle lost NMIs x86, nmi: Add new nmi type 'external'
x86, nmi: Add boot line option 'panic_on_unrecovered_nmi' and

    'panic_on_io_nmi'

x86, nmi: Remove 'reason' value from unknown nmi output x86, nmi: Move default external NMI handler to its own routine x86, nmi: Add better NMI stats to /proc/interrupts and show handlers

 Documentation/kernel-parameters.txt |    9 +
 arch/x86/include/asm/nmi.h          |    5 +
 arch/x86/kernel/irq.c               |    3 +
 arch/x86/kernel/nmi.c               |  302 +++++++++++++++++++++++++++++-----
 drivers/acpi/apei/ghes.c            |    4 +-
 drivers/watchdog/hpwdt.c            |   24 ++-
 6 files changed, 292 insertions(+), 55 deletions(-)

Don Zickus (7):
events, perf: Pass protection and flags bits through mmap2 interface Revert "perf: Disable PERF_RECORD_MMAP2 support"
perf: Update mmap2 interface with protection and flag bits perf report: Add mem-mode documentation to report command
perf: Add cpumode to struct hist_entry
perf: Add support to dynamically get cacheline size
perf: Add dcacheline sort

 include/uapi/linux/perf_event.h          |   1 +
 kernel/events/core.c                     |  37 ++++++++--
 tools/perf/Documentation/perf-report.txt |  23 +++++++
 tools/perf/builtin-report.c              |   2 +-
 tools/perf/tests/dwarf-unwind.c          |   2 +-
 tools/perf/util/cpumap.c                 |  31 +++++++++
 tools/perf/util/cpumap.h                 |  12 ++++
 tools/perf/util/event.c                  |  59 ++++++++++------
 tools/perf/util/event.h                  |   2 +
 tools/perf/util/evsel.c                  |   1 +
 tools/perf/util/hist.c                   |   9 ++-
 tools/perf/util/hist.h                   |   1 +
 tools/perf/util/machine.c                |   4 +-
 tools/perf/util/map.c                    |   4 +-
 tools/perf/util/map.h                    |   4 +-
 tools/perf/util/sort.c                   | 112 +++++++++++++++++++++++++++++++
 tools/perf/util/sort.h                   |   2 +
 17 files changed, 275 insertions(+), 31 deletions(-)
Signature · 1 line
1.7.11.7
author
#1 lore raw

PATCH v3 1/7 events, perf: Pass protection and flags bits through mmap2 interface

From: Peter Zijlstra <[email protected]>

The mmap2 interface was missing the protection and flags bits needed to accurately determine if a mmap memory area was shared or private and if it was readable or not.

[tweaked patch to compile and wrote changelog]
---
 include/uapi/linux/perf_event.h |  1 +
 kernel/events/core.c            | 33 +++++++++++++++++++++++++++++++++
 2 files changed, 34 insertions(+)
include/uapi/linux/perf_event.h +1 -0
index 853bc1c..2ed502f 100644
@@ -699,6 +699,7 @@ enum perf_event_type {
 	 *	u32				min;
 	 *	u64				ino;
 	 *	u64				ino_generation;
+	 *	u32				prot, flags;
 	 *	char				filename[];
 	 * 	struct sample_id		sample_id;
 	 * };
kernel/events/core.c +33 -0
index 5129b12..0711cc9 100644
@@ -40,6 +40,7 @@
 #include <linux/mm_types.h>
 #include <linux/cgroup.h>
 #include <linux/module.h>
+#include <linux/mman.h>
 
 #include "internal.h"
 
@@ -5089,6 +5090,7 @@ struct perf_mmap_event {
 	int			maj, min;
 	u64			ino;
 	u64			ino_generation;
+	u32			prot, flags;
 
 	struct {
 		struct perf_event_header	header;
@@ -5130,6 +5132,8 @@ static void perf_event_mmap_output(struct perf_event *event,
 		mmap_event->event_id.header.size += sizeof(mmap_event->min);
 		mmap_event->event_id.header.size += sizeof(mmap_event->ino);
 		mmap_event->event_id.header.size += sizeof(mmap_event->ino_generation);
+		mmap_event->event_id.header.size += sizeof(mmap_event->prot);
+		mmap_event->event_id.header.size += sizeof(mmap_event->flags);
 	}
 
 	perf_event_header__init_id(&mmap_event->event_id.header, &sample, event);
@@ -5148,6 +5152,8 @@ static void perf_event_mmap_output(struct perf_event *event,
 		perf_output_put(&handle, mmap_event->min);
 		perf_output_put(&handle, mmap_event->ino);
 		perf_output_put(&handle, mmap_event->ino_generation);
+		perf_output_put(&handle, mmap_event->prot);
+		perf_output_put(&handle, mmap_event->flags);
 	}
 
 	__output_copy(&handle, mmap_event->file_name,
@@ -5166,6 +5172,7 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
 	struct file *file = vma->vm_file;
 	int maj = 0, min = 0;
 	u64 ino = 0, gen = 0;
+	u32 prot = 0, flags = 0;
 	unsigned int size;
 	char tmp[16];
 	char *buf = NULL;
@@ -5196,6 +5203,28 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
 		gen = inode->i_generation;
 		maj = MAJOR(dev);
 		min = MINOR(dev);
+
+		if (vma->vm_flags & VM_READ)
+			prot |= PROT_READ;
+		if (vma->vm_flags & VM_WRITE)
+			prot |= PROT_WRITE;
+		if (vma->vm_flags & VM_EXEC)
+			prot |= PROT_EXEC;
+
+		if (vma->vm_flags & VM_MAYSHARE)
+			flags = MAP_SHARED;
+		else
+			flags = MAP_PRIVATE;
+
+		if (vma->vm_flags & VM_DENYWRITE)
+			flags |= MAP_DENYWRITE;
+		if (vma->vm_flags & VM_MAYEXEC)
+			flags |= MAP_EXECUTABLE;
+		if (vma->vm_flags & VM_LOCKED)
+			flags |= MAP_LOCKED;
+		if (vma->vm_flags & VM_HUGETLB)
+			flags |= MAP_HUGETLB;
+
 		goto got_name;
 	} else {
 		name = (char *)arch_vma_name(vma);
@@ -5236,6 +5265,8 @@ got_name:
 	mmap_event->min = min;
 	mmap_event->ino = ino;
 	mmap_event->ino_generation = gen;
+	mmap_event->prot = prot;
+	mmap_event->flags = flags;
 
 	if (!(vma->vm_flags & VM_EXEC))
 		mmap_event->event_id.header.misc |= PERF_RECORD_MISC_MMAP_DATA;
@@ -5276,6 +5307,8 @@ void perf_event_mmap(struct vm_area_struct *vma)
 		/* .min (attr_mmap2 only) */
 		/* .ino (attr_mmap2 only) */
 		/* .ino_generation (attr_mmap2 only) */
+		/* .prot (attr_mmap2 only) */
+		/* .flags (attr_mmap2 only) */
 	};
 
 	perf_event_mmap_event(&mmap_event);
Signature · 1 line
1.7.11.7
#2 lore raw

[tip:perf/core] perf: Pass protection and flags bits through mmap2 interface

perf: Pass protection and flags bits through mmap2 interface
Commit-ID:  f972eb63b1003fae68d7b7e9b674d4ba5db681c2
Gitweb:     http://git.kernel.org/tip/f972eb63b1003fae68d7b7e9b674d4ba5db681c2
Author:     Peter Zijlstra <[email protected]>
AuthorDate: Mon, 19 May 2014 15:13:47 -0400
Committer:  Jiri Olsa <[email protected]>
CommitDate: Mon, 9 Jun 2014 12:21:04 +0200

perf: Pass protection and flags bits through mmap2 interface

The mmap2 interface was missing the protection and flags bits needed to accurately determine if a mmap memory area was shared or private and if it was readable or not.

---
 include/uapi/linux/perf_event.h |  1 +
 kernel/events/core.c            | 33 +++++++++++++++++++++++++++++++++
 2 files changed, 34 insertions(+)
include/uapi/linux/perf_event.h +1 -0
index 5312fae..9269de2 100644
@@ -705,6 +705,7 @@ enum perf_event_type {
 	 *	u32				min;
 	 *	u64				ino;
 	 *	u64				ino_generation;
+	 *	u32				prot, flags;
 	 *	char				filename[];
 	 * 	struct sample_id		sample_id;
 	 * };
kernel/events/core.c +33 -0
index 7da5e56..eea1955 100644
@@ -40,6 +40,7 @@
 #include <linux/mm_types.h>
 #include <linux/cgroup.h>
 #include <linux/module.h>
+#include <linux/mman.h>
 
 #include "internal.h"
 
@@ -5127,6 +5128,7 @@ struct perf_mmap_event {
 	int			maj, min;
 	u64			ino;
 	u64			ino_generation;
+	u32			prot, flags;
 
 	struct {
 		struct perf_event_header	header;
@@ -5168,6 +5170,8 @@ static void perf_event_mmap_output(struct perf_event *event,
 		mmap_event->event_id.header.size += sizeof(mmap_event->min);
 		mmap_event->event_id.header.size += sizeof(mmap_event->ino);
 		mmap_event->event_id.header.size += sizeof(mmap_event->ino_generation);
+		mmap_event->event_id.header.size += sizeof(mmap_event->prot);
+		mmap_event->event_id.header.size += sizeof(mmap_event->flags);
 	}
 
 	perf_event_header__init_id(&mmap_event->event_id.header, &sample, event);
@@ -5186,6 +5190,8 @@ static void perf_event_mmap_output(struct perf_event *event,
 		perf_output_put(&handle, mmap_event->min);
 		perf_output_put(&handle, mmap_event->ino);
 		perf_output_put(&handle, mmap_event->ino_generation);
+		perf_output_put(&handle, mmap_event->prot);
+		perf_output_put(&handle, mmap_event->flags);
 	}
 
 	__output_copy(&handle, mmap_event->file_name,
@@ -5204,6 +5210,7 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
 	struct file *file = vma->vm_file;
 	int maj = 0, min = 0;
 	u64 ino = 0, gen = 0;
+	u32 prot = 0, flags = 0;
 	unsigned int size;
 	char tmp[16];
 	char *buf = NULL;
@@ -5234,6 +5241,28 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
 		gen = inode->i_generation;
 		maj = MAJOR(dev);
 		min = MINOR(dev);
+
+		if (vma->vm_flags & VM_READ)
+			prot |= PROT_READ;
+		if (vma->vm_flags & VM_WRITE)
+			prot |= PROT_WRITE;
+		if (vma->vm_flags & VM_EXEC)
+			prot |= PROT_EXEC;
+
+		if (vma->vm_flags & VM_MAYSHARE)
+			flags = MAP_SHARED;
+		else
+			flags = MAP_PRIVATE;
+
+		if (vma->vm_flags & VM_DENYWRITE)
+			flags |= MAP_DENYWRITE;
+		if (vma->vm_flags & VM_MAYEXEC)
+			flags |= MAP_EXECUTABLE;
+		if (vma->vm_flags & VM_LOCKED)
+			flags |= MAP_LOCKED;
+		if (vma->vm_flags & VM_HUGETLB)
+			flags |= MAP_HUGETLB;
+
 		goto got_name;
 	} else {
 		name = (char *)arch_vma_name(vma);
@@ -5274,6 +5303,8 @@ got_name:
 	mmap_event->min = min;
 	mmap_event->ino = ino;
 	mmap_event->ino_generation = gen;
+	mmap_event->prot = prot;
+	mmap_event->flags = flags;
 
 	if (!(vma->vm_flags & VM_EXEC))
 		mmap_event->event_id.header.misc |= PERF_RECORD_MISC_MMAP_DATA;
@@ -5314,6 +5345,8 @@ void perf_event_mmap(struct vm_area_struct *vma)
 		/* .min (attr_mmap2 only) */
 		/* .ino (attr_mmap2 only) */
 		/* .ino_generation (attr_mmap2 only) */
+		/* .prot (attr_mmap2 only) */
+		/* .flags (attr_mmap2 only) */
 	};
 
 	perf_event_mmap_event(&mmap_event);
author
#3 lore raw

PATCH v3 2/7 Revert "perf: Disable PERF_RECORD_MMAP2 support"

In reply to Don Zickus

This reverts commit 3090ffb5a2515990182f3f55b0688a7817325488.

Re-enable the mmap2 interface as we will have a user soon.

Since things have changed since perf disabled mmap2, small tweaks to the revert had to be done:

o commit 9d4ecc88 forced (n!=8) to become (n<7)
o a new libunwind test needed updating to use mmap2 interface

---
v2: added a better changelog
fix a unwind test (thanks Jiri)

---
 kernel/events/core.c            |  4 ----
 tools/perf/tests/dwarf-unwind.c |  2 +-
 tools/perf/util/event.c         | 36 +++++++++++++++++++-----------------
 tools/perf/util/evsel.c         |  1 +
 4 files changed, 21 insertions(+), 22 deletions(-)
kernel/events/core.c +0 -4
index 0711cc9..c793cb0 100644
@@ -6882,10 +6882,6 @@ static int perf_copy_attr(struct perf_event_attr __user *uattr,
 	if (ret)
 		return -EFAULT;
 
-	/* disabled for now */
-	if (attr->mmap2)
-		return -EINVAL;
-
 	if (attr->__reserved_1)
 		return -EINVAL;
 
tools/perf/tests/dwarf-unwind.c +1 -1
index 108f0cd..96adb73 100644
@@ -15,7 +15,7 @@ static int mmap_handler(struct perf_tool *tool __maybe_unused,
 			struct perf_sample *sample __maybe_unused,
 			struct machine *machine)
 {
-	return machine__process_mmap_event(machine, event, NULL);
+	return machine__process_mmap2_event(machine, event, NULL);
 }
 
 static int init_live_machine(struct machine *machine)
tools/perf/util/event.c +19 -17
index 65795b8..5e7c722 100644
@@ -178,13 +178,14 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
 		return -1;
 	}
 
-	event->header.type = PERF_RECORD_MMAP;
+	event->header.type = PERF_RECORD_MMAP2;
 
 	while (1) {
 		char bf[BUFSIZ];
 		char prot[5];
 		char execname[PATH_MAX];
 		char anonstr[] = "//anon";
+		unsigned int ino;
 		size_t size;
 		ssize_t n;
 
@@ -195,14 +196,15 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
 		strcpy(execname, "");
 
 		/* 00400000-0040c000 r-xp 00000000 fd:01 41038  /bin/cat */
-		n = sscanf(bf, "%"PRIx64"-%"PRIx64" %s %"PRIx64" %*x:%*x %*u %s\n",
-		       &event->mmap.start, &event->mmap.len, prot,
-		       &event->mmap.pgoff,
-		       execname);
-		/*
- 		 * Anon maps don't have the execname.
- 		 */
-		if (n < 4)
+		n = sscanf(bf, "%"PRIx64"-%"PRIx64" %s %"PRIx64" %x:%x %u %s\n",
+		       &event->mmap2.start, &event->mmap2.len, prot,
+		       &event->mmap2.pgoff, &event->mmap2.maj,
+		       &event->mmap2.min,
+		       &ino, execname);
+
+		event->mmap2.ino = (u64)ino;
+
+		if (n < 7)
 			continue;
 		/*
 		 * Just like the kernel, see __perf_event_mmap in kernel/perf_event.c
@@ -223,15 +225,15 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
 			strcpy(execname, anonstr);
 
 		size = strlen(execname) + 1;
-		memcpy(event->mmap.filename, execname, size);
+		memcpy(event->mmap2.filename, execname, size);
 		size = PERF_ALIGN(size, sizeof(u64));
-		event->mmap.len -= event->mmap.start;
-		event->mmap.header.size = (sizeof(event->mmap) -
-					(sizeof(event->mmap.filename) - size));
-		memset(event->mmap.filename + size, 0, machine->id_hdr_size);
-		event->mmap.header.size += machine->id_hdr_size;
-		event->mmap.pid = tgid;
-		event->mmap.tid = pid;
+		event->mmap2.len -= event->mmap.start;
+		event->mmap2.header.size = (sizeof(event->mmap2) -
+					(sizeof(event->mmap2.filename) - size));
+		memset(event->mmap2.filename + size, 0, machine->id_hdr_size);
+		event->mmap2.header.size += machine->id_hdr_size;
+		event->mmap2.pid = tgid;
+		event->mmap2.tid = pid;
 
 		if (process(tool, event, &synth_sample, machine) != 0) {
 			rc = -1;
tools/perf/util/evsel.c +1 -0
index 5c28d82..21154da 100644
@@ -659,6 +659,7 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts)
 		perf_evsel__set_sample_bit(evsel, WEIGHT);
 
 	attr->mmap  = track;
+	attr->mmap2 = track && !perf_missing_features.mmap2;
 	attr->comm  = track;
 
 	if (opts->sample_transaction)
Signature · 1 line
1.7.11.7
author
#4 lore raw

PATCH v3 3/7 perf: Update mmap2 interface with protection and flag bits

In reply to Don Zickus

The kernel piece passes more info now. Update the perf tool to reflect that and adjust the synthesized maps to play along.

---
 tools/perf/util/event.c   | 23 +++++++++++++++++++++--
 tools/perf/util/event.h   |  2 ++
 tools/perf/util/machine.c |  4 +++-
 tools/perf/util/map.c     |  4 +++-
 tools/perf/util/map.h     |  4 +++-
 5 files changed, 32 insertions(+), 5 deletions(-)
tools/perf/util/event.c +21 -2
index 5e7c722..e91d28e 100644
@@ -1,4 +1,5 @@
 #include <linux/types.h>
+#include <sys/mman.h>
 #include "event.h"
 #include "debug.h"
 #include "hist.h"
@@ -214,6 +215,21 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
 		else
 			event->header.misc = PERF_RECORD_MISC_GUEST_USER;
 
+		/* map protection and flags bits */
+		event->mmap2.prot = 0;
+		event->mmap2.flags = 0;
+		if (prot[0] == 'r')
+			event->mmap2.prot |= PROT_READ;
+		if (prot[1] == 'w')
+			event->mmap2.prot |= PROT_WRITE;
+		if (prot[2] == 'x')
+			event->mmap2.prot |= PROT_EXEC;
+
+		if (prot[3] == 's')
+			event->mmap2.flags |= MAP_SHARED;
+		else
+			event->mmap2.flags |= MAP_PRIVATE;
+
 		if (prot[2] != 'x') {
 			if (!mmap_data || prot[0] != 'r')
 				continue;
@@ -614,12 +630,15 @@ size_t perf_event__fprintf_mmap(union perf_event *event, FILE *fp)
 size_t perf_event__fprintf_mmap2(union perf_event *event, FILE *fp)
 {
 	return fprintf(fp, " %d/%d: [%#" PRIx64 "(%#" PRIx64 ") @ %#" PRIx64
-			   " %02x:%02x %"PRIu64" %"PRIu64"]: %c %s\n",
+			   " %02x:%02x %"PRIu64" %"PRIu64"]: %c%c%c%c %s\n",
 		       event->mmap2.pid, event->mmap2.tid, event->mmap2.start,
 		       event->mmap2.len, event->mmap2.pgoff, event->mmap2.maj,
 		       event->mmap2.min, event->mmap2.ino,
 		       event->mmap2.ino_generation,
-		       (event->header.misc & PERF_RECORD_MISC_MMAP_DATA) ? 'r' : 'x',
+		       (event->mmap2.prot & PROT_READ) ? 'r' : '-',
+		       (event->mmap2.prot & PROT_WRITE) ? 'w' : '-',
+		       (event->mmap2.prot & PROT_EXEC) ? 'x' : '-',
+		       (event->mmap2.flags & MAP_SHARED) ? 's' : 'p',
 		       event->mmap2.filename);
 }
 
tools/perf/util/event.h +2 -0
index d970232..9ba2eb3 100644
@@ -27,6 +27,8 @@ struct mmap2_event {
 	u32 min;
 	u64 ino;
 	u64 ino_generation;
+	u32 prot;
+	u32 flags;
 	char filename[PATH_MAX];
 };
 
tools/perf/util/machine.c +3 -1
index 7409ac8..0e5fea9 100644
@@ -1060,6 +1060,8 @@ int machine__process_mmap2_event(struct machine *machine,
 			event->mmap2.pid, event->mmap2.maj,
 			event->mmap2.min, event->mmap2.ino,
 			event->mmap2.ino_generation,
+			event->mmap2.prot,
+			event->mmap2.flags,
 			event->mmap2.filename, type);
 
 	if (map == NULL)
@@ -1105,7 +1107,7 @@ int machine__process_mmap_event(struct machine *machine, union perf_event *event
 
 	map = map__new(&machine->user_dsos, event->mmap.start,
 			event->mmap.len, event->mmap.pgoff,
-			event->mmap.pid, 0, 0, 0, 0,
+			event->mmap.pid, 0, 0, 0, 0, 0, 0,
 			event->mmap.filename,
 			type);
 
tools/perf/util/map.c +3 -1
index ba5f5c0c..7ca1d7e 100644
@@ -51,7 +51,7 @@ void map__init(struct map *map, enum map_type type,
 
 struct map *map__new(struct list_head *dsos__list, u64 start, u64 len,
 		     u64 pgoff, u32 pid, u32 d_maj, u32 d_min, u64 ino,
-		     u64 ino_gen, char *filename,
+		     u64 ino_gen, u32 prot, u32 flags, char *filename,
 		     enum map_type type)
 {
 	struct map *map = malloc(sizeof(*map));
@@ -69,6 +69,8 @@ struct map *map__new(struct list_head *dsos__list, u64 start, u64 len,
 		map->min = d_min;
 		map->ino = ino;
 		map->ino_generation = ino_gen;
+		map->prot = prot;
+		map->flags = flags;
 
 		if ((anon || no_dso) && type == MAP__FUNCTION) {
 			snprintf(newfilename, sizeof(newfilename), "/tmp/perf-%d.map", pid);
tools/perf/util/map.h +3 -1
index ae2d451..7758c72 100644
@@ -35,6 +35,8 @@ struct map {
 	bool			referenced;
 	bool			erange_warned;
 	u32			priv;
+	u32			prot;
+	u32			flags;
 	u64			pgoff;
 	u64			reloc;
 	u32			maj, min; /* only valid for MMAP2 record */
@@ -118,7 +120,7 @@ void map__init(struct map *map, enum map_type type,
 	       u64 start, u64 end, u64 pgoff, struct dso *dso);
 struct map *map__new(struct list_head *dsos__list, u64 start, u64 len,
 		     u64 pgoff, u32 pid, u32 d_maj, u32 d_min, u64 ino,
-		     u64 ino_gen,
+		     u64 ino_gen, u32 prot, u32 flags,
 		     char *filename, enum map_type type);
 struct map *map__new2(u64 start, struct dso *dso, enum map_type type);
 void map__delete(struct map *map);
Signature · 1 line
1.7.11.7
#5 lore raw

[tip:perf/core] perf tools: Update mmap2 interface with protection and flag bits

perf tools: Update mmap2 interface with protection and flag bits
Commit-ID:  7ef807034ef33f8afe33fa7957c73954e8e4f89c
Gitweb:     http://git.kernel.org/tip/7ef807034ef33f8afe33fa7957c73954e8e4f89c
Author:     Don Zickus <[email protected]>
AuthorDate: Mon, 19 May 2014 15:13:49 -0400
Committer:  Jiri Olsa <[email protected]>
CommitDate: Mon, 9 Jun 2014 13:34:45 +0200

perf tools: Update mmap2 interface with protection and flag bits

The kernel piece passes more info now. Update the perf tool to reflect that and adjust the synthesized maps to play along.

---
 tools/perf/util/event.c   | 23 +++++++++++++++++++++--
 tools/perf/util/event.h   |  2 ++
 tools/perf/util/machine.c |  4 +++-
 tools/perf/util/map.c     |  4 +++-
 tools/perf/util/map.h     |  4 +++-
 5 files changed, 32 insertions(+), 5 deletions(-)
tools/perf/util/event.c +21 -2
index 65795b8..ce43cba 100644
@@ -1,4 +1,5 @@
 #include <linux/types.h>
+#include <sys/mman.h>
 #include "event.h"
 #include "debug.h"
 #include "hist.h"
@@ -212,6 +213,21 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
 		else
 			event->header.misc = PERF_RECORD_MISC_GUEST_USER;
 
+		/* map protection and flags bits */
+		event->mmap2.prot = 0;
+		event->mmap2.flags = 0;
+		if (prot[0] == 'r')
+			event->mmap2.prot |= PROT_READ;
+		if (prot[1] == 'w')
+			event->mmap2.prot |= PROT_WRITE;
+		if (prot[2] == 'x')
+			event->mmap2.prot |= PROT_EXEC;
+
+		if (prot[3] == 's')
+			event->mmap2.flags |= MAP_SHARED;
+		else
+			event->mmap2.flags |= MAP_PRIVATE;
+
 		if (prot[2] != 'x') {
 			if (!mmap_data || prot[0] != 'r')
 				continue;
@@ -612,12 +628,15 @@ size_t perf_event__fprintf_mmap(union perf_event *event, FILE *fp)
 size_t perf_event__fprintf_mmap2(union perf_event *event, FILE *fp)
 {
 	return fprintf(fp, " %d/%d: [%#" PRIx64 "(%#" PRIx64 ") @ %#" PRIx64
-			   " %02x:%02x %"PRIu64" %"PRIu64"]: %c %s\n",
+			   " %02x:%02x %"PRIu64" %"PRIu64"]: %c%c%c%c %s\n",
 		       event->mmap2.pid, event->mmap2.tid, event->mmap2.start,
 		       event->mmap2.len, event->mmap2.pgoff, event->mmap2.maj,
 		       event->mmap2.min, event->mmap2.ino,
 		       event->mmap2.ino_generation,
-		       (event->header.misc & PERF_RECORD_MISC_MMAP_DATA) ? 'r' : 'x',
+		       (event->mmap2.prot & PROT_READ) ? 'r' : '-',
+		       (event->mmap2.prot & PROT_WRITE) ? 'w' : '-',
+		       (event->mmap2.prot & PROT_EXEC) ? 'x' : '-',
+		       (event->mmap2.flags & MAP_SHARED) ? 's' : 'p',
 		       event->mmap2.filename);
 }
 
tools/perf/util/event.h +2 -0
index d970232..9ba2eb3 100644
@@ -27,6 +27,8 @@ struct mmap2_event {
 	u32 min;
 	u64 ino;
 	u64 ino_generation;
+	u32 prot;
+	u32 flags;
 	char filename[PATH_MAX];
 };
 
tools/perf/util/machine.c +3 -1
index 7409ac8..0e5fea9 100644
@@ -1060,6 +1060,8 @@ int machine__process_mmap2_event(struct machine *machine,
 			event->mmap2.pid, event->mmap2.maj,
 			event->mmap2.min, event->mmap2.ino,
 			event->mmap2.ino_generation,
+			event->mmap2.prot,
+			event->mmap2.flags,
 			event->mmap2.filename, type);
 
 	if (map == NULL)
@@ -1105,7 +1107,7 @@ int machine__process_mmap_event(struct machine *machine, union perf_event *event
 
 	map = map__new(&machine->user_dsos, event->mmap.start,
 			event->mmap.len, event->mmap.pgoff,
-			event->mmap.pid, 0, 0, 0, 0,
+			event->mmap.pid, 0, 0, 0, 0, 0, 0,
 			event->mmap.filename,
 			type);
 
tools/perf/util/map.c +3 -1
index 8ccbb32..25c571f 100644
@@ -138,7 +138,7 @@ void map__init(struct map *map, enum map_type type,
 
 struct map *map__new(struct list_head *dsos__list, u64 start, u64 len,
 		     u64 pgoff, u32 pid, u32 d_maj, u32 d_min, u64 ino,
-		     u64 ino_gen, char *filename,
+		     u64 ino_gen, u32 prot, u32 flags, char *filename,
 		     enum map_type type)
 {
 	struct map *map = malloc(sizeof(*map));
@@ -157,6 +157,8 @@ struct map *map__new(struct list_head *dsos__list, u64 start, u64 len,
 		map->min = d_min;
 		map->ino = ino;
 		map->ino_generation = ino_gen;
+		map->prot = prot;
+		map->flags = flags;
 
 		if ((anon || no_dso) && type == MAP__FUNCTION) {
 			snprintf(newfilename, sizeof(newfilename), "/tmp/perf-%d.map", pid);
tools/perf/util/map.h +3 -1
index ae2d451..7758c72 100644
@@ -35,6 +35,8 @@ struct map {
 	bool			referenced;
 	bool			erange_warned;
 	u32			priv;
+	u32			prot;
+	u32			flags;
 	u64			pgoff;
 	u64			reloc;
 	u32			maj, min; /* only valid for MMAP2 record */
@@ -118,7 +120,7 @@ void map__init(struct map *map, enum map_type type,
 	       u64 start, u64 end, u64 pgoff, struct dso *dso);
 struct map *map__new(struct list_head *dsos__list, u64 start, u64 len,
 		     u64 pgoff, u32 pid, u32 d_maj, u32 d_min, u64 ino,
-		     u64 ino_gen,
+		     u64 ino_gen, u32 prot, u32 flags,
 		     char *filename, enum map_type type);
 struct map *map__new2(u64 start, struct dso *dso, enum map_type type);
 void map__delete(struct map *map);
author
#6 lore raw

PATCH v3 4/7 perf report: Add mem-mode documentation to report command

In reply to Don Zickus

Add mem-mode sorting types and mem-mode itself to perf-report documentation.

---
 tools/perf/Documentation/perf-report.txt | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
tools/perf/Documentation/perf-report.txt +22 -0
index 09af662..bea1c83 100644
@@ -98,6 +98,21 @@ OPTIONS
 	And default sort keys are changed to comm, dso_from, symbol_from, dso_to
 	and symbol_to, see '--branch-stack'.
 
+	If --mem-mode option is used, following sort keys are also available
+	(incompatible with --branch-stack):
+	symbol_daddr, dso_daddr, locked, tlb, mem, snoop.
+
+	- symbol_daddr: name of data symbol being executed on at the time of sample
+	- dso_daddr: name of library or module containing the data being executed
+	on at the time of sample
+	- locked: whether the bus was locked at the time of sample
+	- tlb: type of tlb access for the data at the time of sample
+	- mem: type of memory access for the data at the time of sample
+	- snoop: type of snoop (if any) for the data at the time of sample
+
+	And default sort keys are changed to local_weight, mem, sym, dso,
+	symbol_daddr, dso_daddr, snoop, tlb, locked, see '--mem-mode'.
+
 -p::
 --parent=<regex>::
         A regex filter to identify parent. The parent is a caller of this
@@ -236,6 +251,13 @@ OPTIONS
 	Demangle symbol names to human readable form. It's enabled by default,
 	disable with --no-demangle.
 
+--mem-mode::
+	Use the data addresses of samples in addition to instruction addresses
+	to build the histograms.  To generate meaningful output, the perf.data
+	file must have been obtained using perf record -d -W and using a 
+	special event -e cpu/mem-loads/ or -e cpu/mem-stores/. See
+	'perf mem' for simpler access.
+
 --percent-limit::
 	Do not show entries which have an overhead under that percent.
 	(Default: 0).
Signature · 1 line
1.7.11.7
#7 lore raw

[tip:perf/core] perf report: Add mem-mode documentation to report command

perf report: Add mem-mode documentation to report command
Commit-ID:  75e906c9601aee73b88d6e6dc02371f8c3ca24d7
Gitweb:     http://git.kernel.org/tip/75e906c9601aee73b88d6e6dc02371f8c3ca24d7
Author:     Don Zickus <[email protected]>
AuthorDate: Fri, 23 May 2014 18:41:23 +0200
Committer:  Jiri Olsa <[email protected]>
CommitDate: Mon, 9 Jun 2014 13:34:47 +0200

perf report: Add mem-mode documentation to report command

Add mem-mode sorting types and mem-mode itself to perf-report documentation.

---
 tools/perf/Documentation/perf-report.txt | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
tools/perf/Documentation/perf-report.txt +22 -0
index cefdf43..00fbfb6 100644
@@ -117,6 +117,21 @@ OPTIONS
 	By default, every sort keys not specified in -F will be appended
 	automatically.
 
+	If --mem-mode option is used, following sort keys are also available
+	(incompatible with --branch-stack):
+	symbol_daddr, dso_daddr, locked, tlb, mem, snoop.
+
+	- symbol_daddr: name of data symbol being executed on at the time of sample
+	- dso_daddr: name of library or module containing the data being executed
+	on at the time of sample
+	- locked: whether the bus was locked at the time of sample
+	- tlb: type of tlb access for the data at the time of sample
+	- mem: type of memory access for the data at the time of sample
+	- snoop: type of snoop (if any) for the data at the time of sample
+
+	And default sort keys are changed to local_weight, mem, sym, dso,
+	symbol_daddr, dso_daddr, snoop, tlb, locked, see '--mem-mode'.
+
 -p::
 --parent=<regex>::
         A regex filter to identify parent. The parent is a caller of this
@@ -260,6 +275,13 @@ OPTIONS
 	Demangle symbol names to human readable form. It's enabled by default,
 	disable with --no-demangle.
 
+--mem-mode::
+	Use the data addresses of samples in addition to instruction addresses
+	to build the histograms.  To generate meaningful output, the perf.data
+	file must have been obtained using perf record -d -W and using a
+	special event -e cpu/mem-loads/ or -e cpu/mem-stores/. See
+	'perf mem' for simpler access.
+
 --percent-limit::
 	Do not show entries which have an overhead under that percent.
 	(Default: 0).
author
#8 lore raw

PATCH v3 5/7 perf: Add cpumode to struct hist_entry

In reply to Don Zickus

The next patch needs to sort on cpumode, so add it to hist_entry to be tracked.

---
 tools/perf/util/hist.c | 7 ++++---
 tools/perf/util/sort.h | 1 +
 2 files changed, 5 insertions(+), 3 deletions(-)
tools/perf/util/hist.c +4 -3
index 7f0236c..b7160b1c 100644
@@ -410,9 +410,10 @@ struct hist_entry *__hists__add_entry(struct hists *hists,
 			.map	= al->map,
 			.sym	= al->sym,
 		},
-		.cpu	= al->cpu,
-		.ip	= al->addr,
-		.level	= al->level,
+		.cpu	 = al->cpu,
+		.cpumode = al->cpumode,
+		.ip	 = al->addr,
+		.level	 = al->level,
 		.stat = {
 			.nr_events = 1,
 			.period	= period,
tools/perf/util/sort.h +1 -0
index 43e5ff4..22cf912 100644
@@ -87,6 +87,7 @@ struct hist_entry {
 	u64			ip;
 	u64			transaction;
 	s32			cpu;
+	u8			cpumode;
 
 	struct hist_entry_diff	diff;
 
Signature · 1 line
1.7.11.7
author
#9 lore raw

PATCH v3 6/7 perf: Add support to dynamically get cacheline size

In reply to Don Zickus

Different arches may have different cacheline sizes. Look it up and set a global variable for reference.

---
 tools/perf/util/cpumap.c | 31 +++++++++++++++++++++++++++++++
 tools/perf/util/cpumap.h | 12 ++++++++++++
 tools/perf/util/sort.c   |  3 +++
 3 files changed, 46 insertions(+)
tools/perf/util/cpumap.c +31 -0
index c4e55b7..d833238 100644
@@ -477,3 +477,34 @@ int cpu__setup_cpunode_map(void)
 	closedir(dir1);
 	return 0;
 }
+
+int cpu__setup_cacheline_size(void)
+{
+	const char *mnt;
+	char path[PATH_MAX];
+	int n, ret, size;
+	FILE *fp;
+	
+
+	mnt = sysfs__mountpoint();
+	if (!mnt)
+		return -1;
+
+	n = snprintf(path, PATH_MAX, "%s/devices/system/cpu/cpu0/cache/index0/coherency_line_size", mnt);
+	if (n == PATH_MAX) {
+		pr_err("sysfs path crossed PATH_MAX(%d) size\n", PATH_MAX);
+		return -1;
+	}
+
+	fp = fopen(path, "r");
+	if (!fp)
+		return -1;
+	ret = fscanf(fp, "%d", &size);
+	fclose(fp);
+	if (ret != 1)
+		return -1;
+
+	cacheline_size = size;
+
+	return 0;
+}
tools/perf/util/cpumap.h +12 -0
index 61a6548..b3e7b22 100644
@@ -81,4 +81,16 @@ static inline int cpu__get_node(int cpu)
 	return cpunode_map[cpu];
 }
 
+int cacheline_size;
+
+int cpu__setup_cacheline_size(void);
+
+static inline int cpu__cacheline_size(void)
+{
+	if (unlikely(!cacheline_size))
+		pr_debug("cacheline size not initialized\n");
+
+	return cacheline_size;
+}
+
 #endif /* __PERF_CPUMAP_H */
tools/perf/util/sort.c +3 -0
index 635cd8f..50adbfb 100644
@@ -2,6 +2,7 @@
 #include "hist.h"
 #include "comm.h"
 #include "symbol.h"
+#include "cpumap.h"
 
 regex_t		parent_regex;
 const char	default_parent_pattern[] = "^sys_|^do_page_fault";
@@ -1117,6 +1118,8 @@ int setup_sorting(void)
 		return -ENOMEM;
 	}
 
+	cpu__setup_cacheline_size();
+
 	for (tok = strtok_r(str, ", ", &tmp);
 			tok; tok = strtok_r(NULL, ", ", &tmp)) {
 		ret = sort_dimension__add(tok);
Signature · 1 line
1.7.11.7
#10 lore raw

Re: [PATCH 6/7] perf: Add support to dynamically get cacheline size

Quoted 34 lines · Don Zickus

Different arches may have different cacheline sizes. Look it up and set a global variable for reference.

---
 tools/perf/util/cpumap.c | 31 +++++++++++++++++++++++++++++++
 tools/perf/util/cpumap.h | 12 ++++++++++++
 tools/perf/util/sort.c   |  3 +++
 3 files changed, 46 insertions(+)
tools/perf/util/cpumap.c +16 -0
index c4e55b7..d833238 100644
@@ -477,3 +477,34 @@ int cpu__setup_cpunode_map(void)
 	closedir(dir1);
 	return 0;
 }
+
+int cpu__setup_cacheline_size(void)
+{
+	const char *mnt;
+	char path[PATH_MAX];
+	int n, ret, size;
+	FILE *fp;
+	
+
+	mnt = sysfs__mountpoint();
+	if (!mnt)
+		return -1;
+
+	n = snprintf(path, PATH_MAX, "%s/devices/system/cpu/cpu0/cache/index0/coherency_line_size", mnt);
+	if (n == PATH_MAX) {
+		pr_err("sysfs path crossed PATH_MAX(%d) size\n", PATH_MAX);

tools/perf/util/cpumap.c int cpu__setup_cpunode_map(void) · line 505

Quoted 10 lines
tools/perf/util/cpumap.c +10 -0
+		return -1;
+	}
+
+	fp = fopen(path, "r");
+	if (!fp)
+		return -1;
+	ret = fscanf(fp, "%d", &size);
+	fclose(fp);
+	if (ret != 1)
+		return -1;

you can use filename__read_int

jirka

author
#11 lore raw

Re: [PATCH 6/7] perf: Add support to dynamically get cacheline size

Quoted 39 lines · Jiri Olsa
Quoted 38 lines · Don Zickus

Different arches may have different cacheline sizes. Look it up and set a global variable for reference.

---
 tools/perf/util/cpumap.c | 31 +++++++++++++++++++++++++++++++
 tools/perf/util/cpumap.h | 12 ++++++++++++
 tools/perf/util/sort.c   |  3 +++
 3 files changed, 46 insertions(+)
tools/perf/util/cpumap.c +20 -0
index c4e55b7..d833238 100644
@@ -477,3 +477,34 @@ int cpu__setup_cpunode_map(void)
 	closedir(dir1);
 	return 0;
 }
+
+int cpu__setup_cacheline_size(void)
+{
+	const char *mnt;
+	char path[PATH_MAX];
+	int n, ret, size;
+	FILE *fp;
+	
+
+	mnt = sysfs__mountpoint();
+	if (!mnt)
+		return -1;
+
+	n = snprintf(path, PATH_MAX, "%s/devices/system/cpu/cpu0/cache/index0/coherency_line_size", mnt);
+	if (n == PATH_MAX) {
+		pr_err("sysfs path crossed PATH_MAX(%d) size\n", PATH_MAX);
+		return -1;
+	}
+
+	fp = fopen(path, "r");
Quoted 8 lines

tools/perf/util/cpumap.c int cpu__setup_cpunode_map(void) · line 505

Quoted 6 lines
tools/perf/util/cpumap.c +6 -0
+	if (!fp)
+		return -1;
+	ret = fscanf(fp, "%d", &size);
+	fclose(fp);
+	if (ret != 1)
+		return -1;

you can use filename__read_int

Doh. I forgot about that. I can respin this patch or the whole series for that. Perhaps on Tuesday when we come back from our US Holiday.

Cheers,
Don

#12 lore raw

Re: [PATCH 6/7] perf: Add support to dynamically get cacheline size

In reply to Don Zickus

On Mon, May 19, 2014 at 03:13:52PM -0400, Don Zickus wrote:

SNIP

tools/perf/util/cpumap.h static inline int cpu__get_node(int cpu) · line 84

Quoted 9 lines
tools/perf/util/cpumap.h +1 -0
index 61a6548..b3e7b22 100644
@@ -81,4 +81,16 @@ static inline int cpu__get_node(int cpu)
 	return cpunode_map[cpu];
 }
 
+int cacheline_size;

hum, so cacheline_size does not need to be global, if you use cpu__cacheline_size in your next patch, right?

Quoted 25 lines
tools/perf/util/cpumap.h +11 -0
+
+int cpu__setup_cacheline_size(void);
+
+static inline int cpu__cacheline_size(void)
+{
+	if (unlikely(!cacheline_size))
+		pr_debug("cacheline size not initialized\n");
+
+	return cacheline_size;
+}
+
 #endif /* __PERF_CPUMAP_H */
tools/perf/util/sort.c +1 -0
index 635cd8f..50adbfb 100644
@@ -2,6 +2,7 @@
 #include "hist.h"
 #include "comm.h"
 #include "symbol.h"
+#include "cpumap.h"
 
 regex_t		parent_regex;
 const char	default_parent_pattern[] = "^sys_|^do_page_fault";

tools/perf/util/sort.c int setup_sorting(void) · line 1122

Quoted 5 lines
tools/perf/util/sort.c +2 -0
@@ -1117,6 +1118,8 @@ int setup_sorting(void)
 		return -ENOMEM;
 	}
 
+	cpu__setup_cacheline_size();
+

I dont think this is the best place to call the setup, on the other hand I can't think of any suitable ;-)

how about lazy initialization from cpu__cacheline_size?

We also have page_size variable which gets initialized in main and is globaly accesable. I wonder we should add the cacheline_size variable in the same way.

jirka

author
#13 lore raw

Re: [PATCH 6/7] perf: Add support to dynamically get cacheline size

Quoted 12 lines · Jiri Olsa

On Mon, May 19, 2014 at 03:13:52PM -0400, Don Zickus wrote:

SNIP

Quoted 8 lines
tools/perf/util/cpumap.h +0 -0
index 61a6548..b3e7b22 100644
@@ -81,4 +81,16 @@ static inline int cpu__get_node(int cpu)
 	return cpunode_map[cpu];
 }
 
Quoted 4 lines

tools/perf/util/cpumap.h static inline int cpu__get_node(int cpu) · line 84

Quoted 1 line
tools/perf/util/cpumap.h +1 -0
+int cacheline_size;

hum, so cacheline_size does not need to be global, if you use cpu__cacheline_size in your next patch, right?

I made cpu__cacheline_size an inline so I think it does. If I stick that function in cpumap.c then yeah I would agree it would not need to be global.

Being that is was going to be called repeatedly while processing each event I opted to make it an inline. I guess I could change that if it matters.

Quoted 35 lines
Quoted 26 lines
tools/perf/util/cpumap.h +11 -0
+
+int cpu__setup_cacheline_size(void);
+
+static inline int cpu__cacheline_size(void)
+{
+	if (unlikely(!cacheline_size))
+		pr_debug("cacheline size not initialized\n");
+
+	return cacheline_size;
+}
+
 #endif /* __PERF_CPUMAP_H */
tools/perf/util/sort.c +1 -0
index 635cd8f..50adbfb 100644
@@ -2,6 +2,7 @@
 #include "hist.h"
 #include "comm.h"
 #include "symbol.h"
+#include "cpumap.h"
 
 regex_t		parent_regex;
 const char	default_parent_pattern[] = "^sys_|^do_page_fault";

tools/perf/util/sort.c int setup_sorting(void) · line 1122

Quoted 5 lines
tools/perf/util/sort.c +2 -0
@@ -1117,6 +1118,8 @@ int setup_sorting(void)
 		return -ENOMEM;
 	}
 
+	cpu__setup_cacheline_size();
+

I dont think this is the best place to call the setup, on the other hand I can't think of any suitable ;-)

Quoted 5 lines

how about lazy initialization from cpu__cacheline_size?

We also have page_size variable which gets initialized in main and is globaly accesable. I wonder we should add the cacheline_size variable in the same way.

That seems to make sense. I can stick the variable in util.c|h then?

Cheers,
Don

author
#14 lore raw

PATCH v3 7/7 perf: Add dcacheline sort

In reply to Don Zickus

In perf's 'mem-mode', one can get access to a whole bunch of details specific to a particular sample instruction. A bunch of those details relate to the data address.

One interesting thing you can do with data addresses is to convert them into a unique cacheline they belong too. Organizing these data cachelines into similar groups and sorting them can reveal cache contention.

This patch creates an alogorithm based on various sample details that can help group entries together into data cachelines and allows 'perf report' to sort on it.

The algorithm relies on having proper mmap2 support in the kernel to help determine if the memory map the data address belongs to is private to a pid or globally shared.

The alogortithm is as follows:

o group cpumodes together
o group entries with discovered maps together
o sort on major, minor, inode and inode generation numbers
o if userspace anon, then sort on pid
o sort on cachelines based on data addresses

The 'dcacheline' sort option in 'perf report' only works in 'mem-mode'.

Sample output:

#
# Samples: 206 of event 'cpu/mem-loads/pp'
# Total weight : 2534
# Sort order : dcacheline,pid
#
# Overhead Samples Data Cacheline Command: Pid

 # ........  ............  ......................................................................  ..................
 #
    13.22%             1  [k] 0xffff88042f08ebc0                                                       swapper:    0
     9.27%             1  [k] 0xffff88082e8cea80                                                       swapper:    0
     3.59%             2  [k] 0xffffffff819ba180                                                       swapper:    0
     0.32%             1  [k] arch_trigger_all_cpu_backtrace_handler_na.23901+0xffffffffffffffe0       swapper:    0
     0.32%             1  [k] timekeeper_seq+0xfffffffffffffff8                                        swapper:    0

Note: Added a '+1' to symlen size in hists__calc_col_len to prevent the next column from prematurely tabbing over and mis-aligning. Not sure what the problem is.

---
v3: fix header column length
V2: update using cpu__cacheline_size()

---
 tools/perf/Documentation/perf-report.txt |   3 +-
 tools/perf/builtin-report.c              |   2 +-
 tools/perf/util/hist.c                   |   2 +
 tools/perf/util/hist.h                   |   1 +
 tools/perf/util/sort.c                   | 109 +++++++++++++++++++++++++++++++
 tools/perf/util/sort.h                   |   1 +
 6 files changed, 116 insertions(+), 2 deletions(-)
tools/perf/Documentation/perf-report.txt +2 -1
index bea1c83..9115d02 100644
@@ -100,7 +100,7 @@ OPTIONS
 
 	If --mem-mode option is used, following sort keys are also available
 	(incompatible with --branch-stack):
-	symbol_daddr, dso_daddr, locked, tlb, mem, snoop.
+	symbol_daddr, dso_daddr, locked, tlb, mem, snoop, dcacheline.
 
 	- symbol_daddr: name of data symbol being executed on at the time of sample
 	- dso_daddr: name of library or module containing the data being executed
@@ -109,6 +109,7 @@ OPTIONS
 	- tlb: type of tlb access for the data at the time of sample
 	- mem: type of memory access for the data at the time of sample
 	- snoop: type of snoop (if any) for the data at the time of sample
+	- dcacheline: the cacheline the data address is on at the time of sample
 
 	And default sort keys are changed to local_weight, mem, sym, dso,
 	symbol_daddr, dso_daddr, snoop, tlb, locked, see '--mem-mode'.
tools/perf/builtin-report.c +1 -1
index 89c9528..387459f 100644
@@ -702,7 +702,7 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 		   "sort by key(s): pid, comm, dso, symbol, parent, cpu, srcline,"
 		   " dso_to, dso_from, symbol_to, symbol_from, mispredict,"
 		   " weight, local_weight, mem, symbol_daddr, dso_daddr, tlb, "
-		   "snoop, locked, abort, in_tx, transaction"),
+		   "snoop, locked, abort, in_tx, transaction, dcacheline"),
 	OPT_BOOLEAN(0, "showcpuutilization", &symbol_conf.show_cpu_utilization,
 		    "Show sample percentage for different cpu modes"),
 	OPT_STRING('p', "parent", &parent_pattern, "regex",
tools/perf/util/hist.c +2 -0
index b7160b1c..5df514c 100644
@@ -127,6 +127,8 @@ void hists__calc_col_len(struct hists *hists, struct hist_entry *h)
 			       + unresolved_col_width + 2;
 			hists__new_col_len(hists, HISTC_MEM_DADDR_SYMBOL,
 					   symlen);
+			hists__new_col_len(hists, HISTC_MEM_DCACHELINE,
+					   symlen + 1);
 		} else {
 			symlen = unresolved_col_width + 4 + 2;
 			hists__new_col_len(hists, HISTC_MEM_DADDR_SYMBOL,
tools/perf/util/hist.h +1 -0
index 38c3e87..14d1dc1 100644
@@ -72,6 +72,7 @@ enum hist_column {
 	HISTC_MEM_TLB,
 	HISTC_MEM_LVL,
 	HISTC_MEM_SNOOP,
+	HISTC_MEM_DCACHELINE,
 	HISTC_TRANSACTION,
 	HISTC_NR_COLS, /* Last entry */
 };
tools/perf/util/sort.c +109 -0

130 lines not on this page · raw message

tools/perf/util/sort.h +1 -0
index 22cf912..460ec9c 100644
@@ -167,6 +167,7 @@ enum sort_type {
 	SORT_MEM_TLB,
 	SORT_MEM_LVL,
 	SORT_MEM_SNOOP,
+	SORT_MEM_DCACHELINE,
 };
 
 /*
Signature · 1 line
1.7.11.7
#15 lore raw

Re: [PATCH 7/7] perf: Add dcacheline sort

Quoted 40 lines · Don Zickus

In perf's 'mem-mode', one can get access to a whole bunch of details specific to a particular sample instruction. A bunch of those details relate to the data address.

One interesting thing you can do with data addresses is to convert them into a unique cacheline they belong too. Organizing these data cachelines into similar groups and sorting them can reveal cache contention.

This patch creates an alogorithm based on various sample details that can help group entries together into data cachelines and allows 'perf report' to sort on it.

The algorithm relies on having proper mmap2 support in the kernel to help determine if the memory map the data address belongs to is private to a pid or globally shared.

The alogortithm is as follows:

o group cpumodes together
o group entries with discovered maps together
o sort on major, minor, inode and inode generation numbers
o if userspace anon, then sort on pid
o sort on cachelines based on data addresses

The 'dcacheline' sort option in 'perf report' only works in 'mem-mode'.

Sample output:

#
# Samples: 206 of event 'cpu/mem-loads/pp'
# Total weight : 2534
# Sort order : dcacheline,pid
#
# Overhead Samples Data Cacheline Command: Pid

 # ........  ............  ......................................................................  ..................
 #
    13.22%             1  [k] 0xffff88042f08ebc0                                                       swapper:    0
     9.27%             1  [k] 0xffff88082e8cea80                                                       swapper:    0
     3.59%             2  [k] 0xffffffff819ba180                                                       swapper:    0
     0.32%             1  [k] arch_trigger_all_cpu_backtrace_handler_na.23901+0xffffffffffffffe0       swapper:    0
     0.32%             1  [k] timekeeper_seq+0xfffffffffffffff8                                        swapper:    0
Quoted 2 lines

Note: Added a '+1' to symlen size in hists__calc_col_len to prevent the next column from prematurely tabbing over and mis-aligning. Not sure what the problem is.

I think thats the extra '+' sign ;-) so +1 seems ok

jirka