[GIT PULL 00/21] perf/core improvements and fixes

[GIT PULL 00/21] perf/core improvements and fixes

[Arnaldo Carvalho de Melo]

Hi Ingo,

	Besides these 21 patches there are 65 other patches, all present in the
perf-core-for-mingo tag, that I sent a pull request for but had some issues
building on older distros (got reports and fixes for OL6, CentOS6, tested it
all on RHEL6), minor stuff, all noted on the comments just before my
Signed-off-by lines.

	Please consider pulling,

- Arnaldo

The following changes since commit b64aa553d8430aabd24f303899cfa4de678e2c3a:

  perf bench numa: Show more stats of particular threads in verbose mode (2015-05-04 12:43:41 -0300)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git tags/perf-core-for-mingo-2

for you to fetch changes up to 0c160d495b5616e071bb4f873812e8f473128149:

  perf kmem: Add kmem.default config option (2015-05-04 13:34:48 -0300)

----------------------------------------------------------------
perf/core improvements and fixes:

User visible:

- perf kmem improvements: (Namhyung Kim)

  - Support sort keys on page analysis
  - New --live option
  - Humand readable gfp flags
  - Allow setting the default in perfconfig files

- perf probe --filter improvements (Masami Hiramatsu)

- Improve detection of file/function name in the 'perf probe' pattern (Naveen Rao)

Infrastructure:

- Some more Intel PT prep patches (Adrian Hunter)

- Fix ppc64 ABIv2 symbol decoding (Ananth N Mavinakayanahalli)

Build fixes:

- bison-related build failure on CentOS 6 (Namhyung Kim)

- perf probe fixes for better support powerpc (Naveen Rao)

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

----------------------------------------------------------------
Adrian Hunter (3):
      perf evlist: Amend mmap ref counting for the AUX area mmap
      perf script: Always allow fields 'addr' and 'cpu' for auxtrace
      perf report: Add Instruction Tracing support

Ananth N Mavinakayanahalli (1):
      perf probe ppc64le: Fix ppc64 ABIv2 symbol decoding

Masami Hiramatsu (4):
      perf tools: Improve strfilter to append additional rules
      perf tools: Add strfilter__string to recover rules string
      perf probe: Accept multiple filter options
      perf probe: Accept filter argument for --list

Namhyung Kim (6):
      perf tools: Fix bison-related build failure on CentOS 6
      perf kmem: Implement stat --page --caller
      perf kmem: Support sort keys on page analysis
      perf kmem: Add --live option for current allocation stat
      perf kmem: Print gfp flags in human readable string
      perf kmem: Add kmem.default config option

Naveen N. Rao (7):
      perf probe ppc: Fix symbol fixup issues due to ELF type
      perf probe ppc: Use the right prefix when ignoring SyS symbols on ppc
      perf probe ppc: Enable matching against dot symbols automatically
      perf probe ppc64le: Prefer symbol table lookup over DWARF
      perf probe ppc64le: Fixup function entry if using kallsyms lookup
      perf symbols: Warn on build id mismatch
      perf probe: Improve detection of file/function name in the probe pattern

 tools/perf/Documentation/perf-kmem.txt      |  11 +-
 tools/perf/Documentation/perf-probe.txt     |   6 +-
 tools/perf/Documentation/perf-report.txt    |  27 +
 tools/perf/arch/powerpc/util/Build          |   1 +
 tools/perf/arch/powerpc/util/sym-handling.c |  82 +++
 tools/perf/builtin-kmem.c                   | 964 +++++++++++++++++++++++++---
 tools/perf/builtin-probe.c                  |  64 +-
 tools/perf/builtin-report.c                 |  11 +
 tools/perf/builtin-script.c                 |  29 +-
 tools/perf/util/Build                       |   2 +-
 tools/perf/util/evlist.c                    |   2 +-
 tools/perf/util/map.c                       |   5 +
 tools/perf/util/map.h                       |   3 +-
 tools/perf/util/probe-event.c               |  69 +-
 tools/perf/util/probe-event.h               |   5 +-
 tools/perf/util/strfilter.c                 | 107 +++
 tools/perf/util/strfilter.h                 |  35 +
 tools/perf/util/symbol-elf.c                |  13 +-
 tools/perf/util/symbol.c                    |  25 +-
 tools/perf/util/symbol.h                    |  10 +
 20 files changed, 1313 insertions(+), 158 deletions(-)
 create mode 100644 tools/perf/arch/powerpc/util/sym-handling.c

[PATCH 01/21] perf probe ppc: Fix symbol fixup issues due to ELF type

[Arnaldo Carvalho de Melo]

From: "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>

If using the symbol table, symbol addresses are not being fixed up
properly, resulting in probes being placed at wrong addresses:

  # perf probe do_fork
  Added new event:
    probe:do_fork        (on do_fork)

  You can now use it in all perf tools, such as:

	  perf record -e probe:do_fork -aR sleep 1

  # cat /sys/kernel/debug/tracing/kprobe_events
  p:probe/do_fork _text+635952
  # printf "%x" 635952
  9b430
  # grep do_fork /boot/System.map
  c0000000000ab430 T .do_fork

Fix by checking for ELF type ET_DYN used by ppc64 kernels.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: http://lkml.kernel.org/r/41392bb856ef62d929995e0b61967689b7915207.1430217967.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/arch/powerpc/util/Build          |  1 +
 tools/perf/arch/powerpc/util/sym-handling.c | 19 +++++++++++++++++++
 tools/perf/util/symbol-elf.c                |  8 ++++++--
 tools/perf/util/symbol.h                    |  4 ++++
 4 files changed, 30 insertions(+), 2 deletions(-)
 create mode 100644 tools/perf/arch/powerpc/util/sym-handling.c

diff --git a/tools/perf/arch/powerpc/util/Build b/tools/perf/arch/powerpc/util/Build
index 0af6e9b..7b8b0d1 100644
--- a/tools/perf/arch/powerpc/util/Build
+++ b/tools/perf/arch/powerpc/util/Build
@@ -1,4 +1,5 @@
 libperf-y += header.o
+libperf-y += sym-handling.o
 
 libperf-$(CONFIG_DWARF) += dwarf-regs.o
 libperf-$(CONFIG_DWARF) += skip-callchain-idx.o
diff --git a/tools/perf/arch/powerpc/util/sym-handling.c b/tools/perf/arch/powerpc/util/sym-handling.c
new file mode 100644
index 0000000..c9de001
--- /dev/null
+++ b/tools/perf/arch/powerpc/util/sym-handling.c
@@ -0,0 +1,19 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License, version 2, as
+ * published by the Free Software Foundation.
+ *
+ * Copyright (C) 2015 Naveen N. Rao, IBM Corporation
+ */
+
+#include "debug.h"
+#include "symbol.h"
+
+#ifdef HAVE_LIBELF_SUPPORT
+bool elf__needs_adjust_symbols(GElf_Ehdr ehdr)
+{
+	return ehdr.e_type == ET_EXEC ||
+	       ehdr.e_type == ET_REL ||
+	       ehdr.e_type == ET_DYN;
+}
+#endif
diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
index a7ab606..54347ba 100644
--- a/tools/perf/util/symbol-elf.c
+++ b/tools/perf/util/symbol-elf.c
@@ -630,6 +630,11 @@ void symsrc__destroy(struct symsrc *ss)
 	close(ss->fd);
 }
 
+bool __weak elf__needs_adjust_symbols(GElf_Ehdr ehdr)
+{
+	return ehdr.e_type == ET_EXEC || ehdr.e_type == ET_REL;
+}
+
 int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
 		 enum dso_binary_type type)
 {
@@ -711,8 +716,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
 						     ".gnu.prelink_undo",
 						     NULL) != NULL);
 	} else {
-		ss->adjust_symbols = ehdr.e_type == ET_EXEC ||
-				     ehdr.e_type == ET_REL;
+		ss->adjust_symbols = elf__needs_adjust_symbols(ehdr);
 	}
 
 	ss->name   = strdup(name);
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index 0956150..8cb0af4 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -303,4 +303,8 @@ int setup_list(struct strlist **list, const char *list_str,
 int setup_intlist(struct intlist **list, const char *list_str,
 		  const char *list_name);
 
+#ifdef HAVE_LIBELF_SUPPORT
+bool elf__needs_adjust_symbols(GElf_Ehdr ehdr);
+#endif
+
 #endif /* __PERF_SYMBOL */
-- 
2.1.0

[PATCH 02/21] perf probe ppc: Use the right prefix when ignoring SyS symbols on ppc

[Arnaldo Carvalho de Melo]

From: "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>

Use the proper prefix when ignoring SyS symbols on ppc ABIv1. While at
it, generalize symbol selection so architectures can implement their own
logic.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: http://lkml.kernel.org/r/adf1f98b121ecaf292777fe5cc69fe1038feabce.1430217967.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/arch/powerpc/util/sym-handling.c | 20 ++++++++++++++++++++
 tools/perf/util/symbol.c                    | 21 ++++++++++++---------
 tools/perf/util/symbol.h                    |  5 +++++
 3 files changed, 37 insertions(+), 9 deletions(-)

diff --git a/tools/perf/arch/powerpc/util/sym-handling.c b/tools/perf/arch/powerpc/util/sym-handling.c
index c9de001..5522a40 100644
--- a/tools/perf/arch/powerpc/util/sym-handling.c
+++ b/tools/perf/arch/powerpc/util/sym-handling.c
@@ -17,3 +17,23 @@ bool elf__needs_adjust_symbols(GElf_Ehdr ehdr)
 	       ehdr.e_type == ET_DYN;
 }
 #endif
+
+#if !defined(_CALL_ELF) || _CALL_ELF != 2
+int arch__choose_best_symbol(struct symbol *syma,
+			     struct symbol *symb __maybe_unused)
+{
+	char *sym = syma->name;
+
+	/* Skip over any initial dot */
+	if (*sym == '.')
+		sym++;
+
+	/* Avoid "SyS" kernel syscall aliases */
+	if (strlen(sym) >= 3 && !strncmp(sym, "SyS", 3))
+		return SYMBOL_B;
+	if (strlen(sym) >= 10 && !strncmp(sym, "compat_SyS", 10))
+		return SYMBOL_B;
+
+	return SYMBOL_A;
+}
+#endif
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index 201f6c4c..f805757 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -85,8 +85,17 @@ static int prefix_underscores_count(const char *str)
 	return tail - str;
 }
 
-#define SYMBOL_A 0
-#define SYMBOL_B 1
+int __weak arch__choose_best_symbol(struct symbol *syma,
+				    struct symbol *symb __maybe_unused)
+{
+	/* Avoid "SyS" kernel syscall aliases */
+	if (strlen(syma->name) >= 3 && !strncmp(syma->name, "SyS", 3))
+		return SYMBOL_B;
+	if (strlen(syma->name) >= 10 && !strncmp(syma->name, "compat_SyS", 10))
+		return SYMBOL_B;
+
+	return SYMBOL_A;
+}
 
 static int choose_best_symbol(struct symbol *syma, struct symbol *symb)
 {
@@ -134,13 +143,7 @@ static int choose_best_symbol(struct symbol *syma, struct symbol *symb)
 	else if (na < nb)
 		return SYMBOL_B;
 
-	/* Avoid "SyS" kernel syscall aliases */
-	if (na >= 3 && !strncmp(syma->name, "SyS", 3))
-		return SYMBOL_B;
-	if (na >= 10 && !strncmp(syma->name, "compat_SyS", 10))
-		return SYMBOL_B;
-
-	return SYMBOL_A;
+	return arch__choose_best_symbol(syma, symb);
 }
 
 void symbols__fixup_duplicate(struct rb_root *symbols)
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index 8cb0af4..bd50ba0 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -307,4 +307,9 @@ int setup_intlist(struct intlist **list, const char *list_str,
 bool elf__needs_adjust_symbols(GElf_Ehdr ehdr);
 #endif
 
+#define SYMBOL_A 0
+#define SYMBOL_B 1
+
+int arch__choose_best_symbol(struct symbol *syma, struct symbol *symb);
+
 #endif /* __PERF_SYMBOL */
-- 
2.1.0

[PATCH 03/21] perf probe ppc: Enable matching against dot symbols automatically

[Arnaldo Carvalho de Melo]

From: "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>

Allow perf probe to work on ppc ABIv1 without the need to specify the
leading dot '.' for functions. 'perf probe do_fork' works with this
patch.

We do this by changing how symbol name comparison works on ppc ABIv1 -
we simply ignore and skip over the initial dot, if one exists, during
symbol name comparison.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: http://lkml.kernel.org/r/652a8f3bfa919bd02a1836a128370eaed59b4a34.1430217967.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/arch/powerpc/util/sym-handling.c | 13 +++++++++++++
 tools/perf/util/map.c                       |  5 +++++
 tools/perf/util/map.h                       |  3 ++-
 tools/perf/util/symbol.c                    |  4 ++--
 4 files changed, 22 insertions(+), 3 deletions(-)

diff --git a/tools/perf/arch/powerpc/util/sym-handling.c b/tools/perf/arch/powerpc/util/sym-handling.c
index 5522a40..2de2cc4 100644
--- a/tools/perf/arch/powerpc/util/sym-handling.c
+++ b/tools/perf/arch/powerpc/util/sym-handling.c
@@ -8,6 +8,7 @@
 
 #include "debug.h"
 #include "symbol.h"
+#include "map.h"
 
 #ifdef HAVE_LIBELF_SUPPORT
 bool elf__needs_adjust_symbols(GElf_Ehdr ehdr)
@@ -36,4 +37,16 @@ int arch__choose_best_symbol(struct symbol *syma,
 
 	return SYMBOL_A;
 }
+
+/* Allow matching against dot variants */
+int arch__compare_symbol_names(const char *namea, const char *nameb)
+{
+	/* Skip over initial dot */
+	if (*namea == '.')
+		namea++;
+	if (*nameb == '.')
+		nameb++;
+
+	return strcmp(namea, nameb);
+}
 #endif
diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c
index a14f08f..cd0e335 100644
--- a/tools/perf/util/map.c
+++ b/tools/perf/util/map.c
@@ -292,6 +292,11 @@ int map__load(struct map *map, symbol_filter_t filter)
 	return 0;
 }
 
+int __weak arch__compare_symbol_names(const char *namea, const char *nameb)
+{
+	return strcmp(namea, nameb);
+}
+
 struct symbol *map__find_symbol(struct map *map, u64 addr,
 				symbol_filter_t filter)
 {
diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h
index ec19c59..4e0c729 100644
--- a/tools/perf/util/map.h
+++ b/tools/perf/util/map.h
@@ -124,7 +124,7 @@ struct thread;
  */
 #define __map__for_each_symbol_by_name(map, sym_name, pos, filter)	\
 	for (pos = map__find_symbol_by_name(map, sym_name, filter);	\
-	     pos && strcmp(pos->name, sym_name) == 0;		\
+	     pos && arch__compare_symbol_names(pos->name, sym_name) == 0;	\
 	     pos = symbol__next_by_name(pos))
 
 #define map__for_each_symbol_by_name(map, sym_name, pos)		\
@@ -132,6 +132,7 @@ struct thread;
 
 typedef int (*symbol_filter_t)(struct map *map, struct symbol *sym);
 
+int arch__compare_symbol_names(const char *namea, const char *nameb);
 void map__init(struct map *map, enum map_type type,
 	       u64 start, u64 end, u64 pgoff, struct dso *dso);
 struct map *map__new(struct machine *machine, u64 start, u64 len,
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index f805757..45ba48a 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -411,7 +411,7 @@ static struct symbol *symbols__find_by_name(struct rb_root *symbols,
 		int cmp;
 
 		s = rb_entry(n, struct symbol_name_rb_node, rb_node);
-		cmp = strcmp(name, s->sym.name);
+		cmp = arch__compare_symbol_names(name, s->sym.name);
 
 		if (cmp < 0)
 			n = n->rb_left;
@@ -429,7 +429,7 @@ static struct symbol *symbols__find_by_name(struct rb_root *symbols,
 		struct symbol_name_rb_node *tmp;
 
 		tmp = rb_entry(n, struct symbol_name_rb_node, rb_node);
-		if (strcmp(tmp->sym.name, s->sym.name))
+		if (arch__compare_symbol_names(tmp->sym.name, s->sym.name))
 			break;
 
 		s = tmp;
-- 
2.1.0

[PATCH 04/21] perf probe ppc64le: Fix ppc64 ABIv2 symbol decoding

[Arnaldo Carvalho de Melo]

From: Ananth N Mavinakayanahalli <ananth@in.ibm.com>

ppc64 ELF ABIv2 has a Global Entry Point (GEP) and a Local Entry Point
(LEP). For purposes of probing, we need the LEP - the offset to which is
encoded in st_other.

Signed-off-by: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: http://lkml.kernel.org/r/ab9cc5e2b9de4cbaaf50f6ef2346a6a81100bad1.1430217967.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/arch/powerpc/util/sym-handling.c | 7 +++++++
 tools/perf/util/symbol-elf.c                | 4 ++++
 tools/perf/util/symbol.h                    | 1 +
 3 files changed, 12 insertions(+)

diff --git a/tools/perf/arch/powerpc/util/sym-handling.c b/tools/perf/arch/powerpc/util/sym-handling.c
index 2de2cc4..012a0f8 100644
--- a/tools/perf/arch/powerpc/util/sym-handling.c
+++ b/tools/perf/arch/powerpc/util/sym-handling.c
@@ -17,6 +17,13 @@ bool elf__needs_adjust_symbols(GElf_Ehdr ehdr)
 	       ehdr.e_type == ET_REL ||
 	       ehdr.e_type == ET_DYN;
 }
+
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+void arch__elf_sym_adjust(GElf_Sym *sym)
+{
+	sym->st_value += PPC64_LOCAL_ENTRY_OFFSET(sym->st_other);
+}
+#endif
 #endif
 
 #if !defined(_CALL_ELF) || _CALL_ELF != 2
diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
index 54347ba..d99b442 100644
--- a/tools/perf/util/symbol-elf.c
+++ b/tools/perf/util/symbol-elf.c
@@ -775,6 +775,8 @@ static bool want_demangle(bool is_kernel_sym)
 	return is_kernel_sym ? symbol_conf.demangle_kernel : symbol_conf.demangle;
 }
 
+void __weak arch__elf_sym_adjust(GElf_Sym *sym __maybe_unused) { }
+
 int dso__load_sym(struct dso *dso, struct map *map,
 		  struct symsrc *syms_ss, struct symsrc *runtime_ss,
 		  symbol_filter_t filter, int kmodule)
@@ -939,6 +941,8 @@ int dso__load_sym(struct dso *dso, struct map *map,
 		    (sym.st_value & 1))
 			--sym.st_value;
 
+		arch__elf_sym_adjust(&sym);
+
 		if (dso->kernel || kmodule) {
 			char dso_name[PATH_MAX];
 
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index bd50ba0..9096529 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -305,6 +305,7 @@ int setup_intlist(struct intlist **list, const char *list_str,
 
 #ifdef HAVE_LIBELF_SUPPORT
 bool elf__needs_adjust_symbols(GElf_Ehdr ehdr);
+void arch__elf_sym_adjust(GElf_Sym *sym);
 #endif
 
 #define SYMBOL_A 0
-- 
2.1.0

[PATCH 05/21] perf probe ppc64le: Prefer symbol table lookup over DWARF

[Arnaldo Carvalho de Melo]

From: "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>

Use symbol table lookups by default if DWARF is not necessary, since
powerpc ABIv2 encodes local entry points in the symbol table and the
function entry address in DWARF may not be appropriate for kprobes, as
described here:

https://sourceware.org/bugzilla/show_bug.cgi?id=17638

"The DWARF address ranges deliberately include the *whole* function,
both global and local entry points."
...
"If you want to set probes on a local entry point, you should look up
the symbol in the main symbol table (not DWARF), and check the st_other
bits; they will indicate whether the function has a local entry point,
and what its offset from the global entry point is.  Note that GDB does
the same when setting a breakpoint on a function entry."

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: http://lkml.kernel.org/r/88a10e22f4aaba2aef812824ca4b10d7beeea012.1430217967.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/arch/powerpc/util/sym-handling.c | 8 ++++++++
 tools/perf/util/probe-event.c               | 8 ++++++++
 tools/perf/util/probe-event.h               | 1 +
 3 files changed, 17 insertions(+)

diff --git a/tools/perf/arch/powerpc/util/sym-handling.c b/tools/perf/arch/powerpc/util/sym-handling.c
index 012a0f8..a170060 100644
--- a/tools/perf/arch/powerpc/util/sym-handling.c
+++ b/tools/perf/arch/powerpc/util/sym-handling.c
@@ -9,6 +9,7 @@
 #include "debug.h"
 #include "symbol.h"
 #include "map.h"
+#include "probe-event.h"
 
 #ifdef HAVE_LIBELF_SUPPORT
 bool elf__needs_adjust_symbols(GElf_Ehdr ehdr)
@@ -57,3 +58,10 @@ int arch__compare_symbol_names(const char *namea, const char *nameb)
 	return strcmp(namea, nameb);
 }
 #endif
+
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+bool arch__prefers_symtab(void)
+{
+	return true;
+}
+#endif
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index 291bf23..4dfb412 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -2567,6 +2567,8 @@ err_out:
 	goto out;
 }
 
+bool __weak arch__prefers_symtab(void) { return false; }
+
 static int convert_to_probe_trace_events(struct perf_probe_event *pev,
 					  struct probe_trace_event **tevs,
 					  int max_tevs, const char *target)
@@ -2582,6 +2584,12 @@ static int convert_to_probe_trace_events(struct perf_probe_event *pev,
 		}
 	}
 
+	if (arch__prefers_symtab() && !perf_probe_event_need_dwarf(pev)) {
+		ret = find_probe_trace_events_from_map(pev, tevs, max_tevs, target);
+		if (ret > 0)
+			return ret; /* Found in symbol table */
+	}
+
 	/* Convert perf_probe_event with debuginfo */
 	ret = try_to_find_probe_trace_events(pev, tevs, max_tevs, target);
 	if (ret != 0)
diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h
index d6b7834..52bca4b 100644
--- a/tools/perf/util/probe-event.h
+++ b/tools/perf/util/probe-event.h
@@ -135,6 +135,7 @@ extern int show_available_vars(struct perf_probe_event *pevs, int npevs,
 			       struct strfilter *filter, bool externs);
 extern int show_available_funcs(const char *module, struct strfilter *filter,
 				bool user);
+bool arch__prefers_symtab(void);
 
 /* Maximum index number of event-name postfix */
 #define MAX_EVENT_INDEX	1024
-- 
2.1.0

[PATCH 06/21] perf probe ppc64le: Fixup function entry if using kallsyms lookup

[Arnaldo Carvalho de Melo]

From: "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>

On powerpc ABIv2, if no debug-info is found and we use kallsyms, we need
to fixup the function entry to point to the local entry point. Use
offset of 8 since current toolchains always generate 2 instructions (8
bytes).

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: http://lkml.kernel.org/r/92253021e77a104b23b615c8c23bf9501dfe60bf.1430217967.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/arch/powerpc/util/sym-handling.c | 15 +++++++++++++++
 tools/perf/util/probe-event.c               |  5 +++++
 tools/perf/util/probe-event.h               |  2 ++
 3 files changed, 22 insertions(+)

diff --git a/tools/perf/arch/powerpc/util/sym-handling.c b/tools/perf/arch/powerpc/util/sym-handling.c
index a170060..bbc1a50 100644
--- a/tools/perf/arch/powerpc/util/sym-handling.c
+++ b/tools/perf/arch/powerpc/util/sym-handling.c
@@ -64,4 +64,19 @@ bool arch__prefers_symtab(void)
 {
 	return true;
 }
+
+#define PPC64LE_LEP_OFFSET	8
+
+void arch__fix_tev_from_maps(struct perf_probe_event *pev,
+			     struct probe_trace_event *tev, struct map *map)
+{
+	/*
+	 * ppc64 ABIv2 local entry point is currently always 2 instructions
+	 * (8 bytes) after the global entry point.
+	 */
+	if (!pev->uprobes && map->dso->symtab_type == DSO_BINARY_TYPE__KALLSYMS) {
+		tev->point.address += PPC64LE_LEP_OFFSET;
+		tev->point.offset += PPC64LE_LEP_OFFSET;
+	}
+}
 #endif
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index 4dfb412..eb75a5e 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -2447,6 +2447,10 @@ static int find_probe_functions(struct map *map, char *name)
 #define strdup_or_goto(str, label)	\
 	({ char *__p = strdup(str); if (!__p) goto label; __p; })
 
+void __weak arch__fix_tev_from_maps(struct perf_probe_event *pev __maybe_unused,
+				struct probe_trace_event *tev __maybe_unused,
+				struct map *map __maybe_unused) { }
+
 /*
  * Find probe function addresses from map.
  * Return an error or the number of found probe_trace_event
@@ -2553,6 +2557,7 @@ static int find_probe_trace_events_from_map(struct perf_probe_event *pev,
 					strdup_or_goto(pev->args[i].type,
 							nomem_out);
 		}
+		arch__fix_tev_from_maps(pev, tev, map);
 	}
 
 out:
diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h
index 52bca4b..180f142 100644
--- a/tools/perf/util/probe-event.h
+++ b/tools/perf/util/probe-event.h
@@ -136,6 +136,8 @@ extern int show_available_vars(struct perf_probe_event *pevs, int npevs,
 extern int show_available_funcs(const char *module, struct strfilter *filter,
 				bool user);
 bool arch__prefers_symtab(void);
+void arch__fix_tev_from_maps(struct perf_probe_event *pev,
+			     struct probe_trace_event *tev, struct map *map);
 
 /* Maximum index number of event-name postfix */
 #define MAX_EVENT_INDEX	1024
-- 
2.1.0

[PATCH 07/21] perf evlist: Amend mmap ref counting for the AUX area mmap

[Arnaldo Carvalho de Melo]

From: Adrian Hunter <adrian.hunter@intel.com>

Reference counting of the mmap buffer does not work correctly when there
is an AUX area mmap also.

In snapshot mode it is not easy to know if the AUX area mmap buffer
contains usefull information. Equally the evlist does not know if the
recording is in sanpshot mode anyway.

Consequently, for now just assume the AUX area mmap always has data,
which will just cause the mmap buffer to remain mmapped for the duration
of the recording.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1429903807-20559-2-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/evlist.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 0ec8c3b..7ec1bf9 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -695,7 +695,7 @@ union perf_event *perf_evlist__mmap_read(struct perf_evlist *evlist, int idx)
 
 static bool perf_mmap__empty(struct perf_mmap *md)
 {
-	return perf_mmap__read_head(md) == md->prev;
+	return perf_mmap__read_head(md) == md->prev && !md->auxtrace_mmap.base;
 }
 
 static void perf_evlist__mmap_get(struct perf_evlist *evlist, int idx)
-- 
2.1.0

[PATCH 08/21] perf script: Always allow fields 'addr' and 'cpu' for auxtrace

[Arnaldo Carvalho de Melo]

From: Adrian Hunter <adrian.hunter@intel.com>

If a file contains AUX area tracing data then always allow fields 'addr'
and 'cpu' to be selected as options for perf script.  This is necessary
because AUX area decoding may synthesize events with that information.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1429903807-20559-3-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-script.c | 29 ++++++++++++++++++++++-------
 1 file changed, 22 insertions(+), 7 deletions(-)

diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index cd2f38b..6805098 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -148,9 +148,10 @@ static const char *output_field2str(enum perf_output_field field)
 
 #define PRINT_FIELD(x)  (output[attr->type].fields & PERF_OUTPUT_##x)
 
-static int perf_evsel__check_stype(struct perf_evsel *evsel,
-				   u64 sample_type, const char *sample_msg,
-				   enum perf_output_field field)
+static int perf_evsel__do_check_stype(struct perf_evsel *evsel,
+				      u64 sample_type, const char *sample_msg,
+				      enum perf_output_field field,
+				      bool allow_user_set)
 {
 	struct perf_event_attr *attr = &evsel->attr;
 	int type = attr->type;
@@ -160,6 +161,8 @@ static int perf_evsel__check_stype(struct perf_evsel *evsel,
 		return 0;
 
 	if (output[type].user_set) {
+		if (allow_user_set)
+			return 0;
 		evname = perf_evsel__name(evsel);
 		pr_err("Samples for '%s' event do not have %s attribute set. "
 		       "Cannot print '%s' field.\n",
@@ -177,10 +180,22 @@ static int perf_evsel__check_stype(struct perf_evsel *evsel,
 	return 0;
 }
 
+static int perf_evsel__check_stype(struct perf_evsel *evsel,
+				   u64 sample_type, const char *sample_msg,
+				   enum perf_output_field field)
+{
+	return perf_evsel__do_check_stype(evsel, sample_type, sample_msg, field,
+					  false);
+}
+
 static int perf_evsel__check_attr(struct perf_evsel *evsel,
 				  struct perf_session *session)
 {
 	struct perf_event_attr *attr = &evsel->attr;
+	bool allow_user_set;
+
+	allow_user_set = perf_header__has_feat(&session->header,
+					       HEADER_AUXTRACE);
 
 	if (PRINT_FIELD(TRACE) &&
 		!perf_session__has_traces(session, "record -R"))
@@ -193,8 +208,8 @@ static int perf_evsel__check_attr(struct perf_evsel *evsel,
 	}
 
 	if (PRINT_FIELD(ADDR) &&
-		perf_evsel__check_stype(evsel, PERF_SAMPLE_ADDR, "ADDR",
-					PERF_OUTPUT_ADDR))
+		perf_evsel__do_check_stype(evsel, PERF_SAMPLE_ADDR, "ADDR",
+					   PERF_OUTPUT_ADDR, allow_user_set))
 		return -EINVAL;
 
 	if (PRINT_FIELD(SYM) && !PRINT_FIELD(IP) && !PRINT_FIELD(ADDR)) {
@@ -231,8 +246,8 @@ static int perf_evsel__check_attr(struct perf_evsel *evsel,
 		return -EINVAL;
 
 	if (PRINT_FIELD(CPU) &&
-		perf_evsel__check_stype(evsel, PERF_SAMPLE_CPU, "CPU",
-					PERF_OUTPUT_CPU))
+		perf_evsel__do_check_stype(evsel, PERF_SAMPLE_CPU, "CPU",
+					   PERF_OUTPUT_CPU, allow_user_set))
 		return -EINVAL;
 
 	if (PRINT_FIELD(PERIOD) &&
-- 
2.1.0

[PATCH 09/21] perf report: Add Instruction Tracing support

[Arnaldo Carvalho de Melo]

From: Adrian Hunter <adrian.hunter@intel.com>

Add support for decoding an AUX area assuming it contains instruction
tracing data.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1429903807-20559-4-git-send-email-adrian.hunter@intel.com
[ Do not use -Z as an alternative to --itrace ]
[ Fixed initialization of itrace_synth_opts struct fields on older gcc versions ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/Documentation/perf-report.txt | 27 +++++++++++++++++++++++++++
 tools/perf/builtin-report.c              | 11 +++++++++++
 2 files changed, 38 insertions(+)

diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt
index 896672b..b66f6b4 100644
--- a/tools/perf/Documentation/perf-report.txt
+++ b/tools/perf/Documentation/perf-report.txt
@@ -327,6 +327,33 @@ OPTIONS
 
 include::callchain-overhead-calculation.txt[]
 
+--itrace::
+	Options for decoding instruction tracing data. The options are:
+
+		i	synthesize instructions events
+		b	synthesize branches events
+		c	synthesize branches events (calls only)
+		r	synthesize branches events (returns only)
+		e	synthesize error events
+		d	create a debug log
+		g	synthesize a call chain for instructions events
+
+	The default is all events i.e. the same as --itrace=ibe
+
+	In addition, the period (default 100000) for instructions events
+	can be specified in units of:
+
+		i	instructions
+		t	ticks
+		ms	milliseconds
+		us	microseconds
+		ns	nanoseconds (default)
+
+	Also the call chain size (default 16, max. 1024) for instructions
+	events can be specified.
+
+	To disable decoding entirely, use --no-itrace.
+
 SEE ALSO
 --------
 linkperf:perf-stat[1], linkperf:perf-annotate[1]
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 476cdf7..18cb0ff 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -36,6 +36,8 @@
 #include "util/data.h"
 #include "arch/common.h"
 
+#include "util/auxtrace.h"
+
 #include <dlfcn.h>
 #include <linux/bitmap.h>
 
@@ -585,6 +587,7 @@ parse_percent_limit(const struct option *opt, const char *str,
 int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 {
 	struct perf_session *session;
+	struct itrace_synth_opts itrace_synth_opts = { .set = 0, };
 	struct stat st;
 	bool has_br_stack = false;
 	int branch_mode = -1;
@@ -607,6 +610,9 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 			.attr		 = perf_event__process_attr,
 			.tracing_data	 = perf_event__process_tracing_data,
 			.build_id	 = perf_event__process_build_id,
+			.id_index	 = perf_event__process_id_index,
+			.auxtrace_info	 = perf_event__process_auxtrace_info,
+			.auxtrace	 = perf_event__process_auxtrace,
 			.ordered_events	 = true,
 			.ordering_requires_timestamps = true,
 		},
@@ -717,6 +723,9 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 		     "Don't show entries under that percent", parse_percent_limit),
 	OPT_CALLBACK(0, "percentage", NULL, "relative|absolute",
 		     "how to display percentage of filtered entries", parse_filter_percentage),
+	OPT_CALLBACK_OPTARG(0, "itrace", &itrace_synth_opts, NULL, "opts",
+			    "Instruction Tracing options",
+			    itrace_parse_synth_opts),
 	OPT_END()
 	};
 	struct perf_data_file file = {
@@ -761,6 +770,8 @@ repeat:
 					       report.queue_size);
 	}
 
+	session->itrace_synth_opts = &itrace_synth_opts;
+
 	report.session = session;
 
 	has_br_stack = perf_header__has_feat(&session->header,
-- 
2.1.0

[PATCH 10/21] perf symbols: Warn on build id mismatch

[Arnaldo Carvalho de Melo]

From: "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>

Add a debug message to indicate that the build id didn't match.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1429904686-16516-1-git-send-email-naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/symbol-elf.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
index d99b442..9d526a5 100644
--- a/tools/perf/util/symbol-elf.c
+++ b/tools/perf/util/symbol-elf.c
@@ -683,6 +683,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
 		}
 
 		if (!dso__build_id_equal(dso, build_id)) {
+			pr_debug("%s: build id mismatch for %s.\n", __func__, name);
 			dso->load_errno = DSO_LOAD_ERRNO__MISMATCHING_BUILDID;
 			goto out_elf_end;
 		}
-- 
2.1.0

[PATCH 11/21] perf tools: Fix bison-related build failure on CentOS 6

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

The YYLTYPE_IS_TRIVIAL is defined in the Build file, but unlike
pmu-bison.c, gcc complained about it for parse-events-bison.c:

    CC       util/parse-events-bison.o
  In file included from util/parse-events.y:16:
  util/parse-events-bison.h:101:1: error: "YYLTYPE_IS_TRIVIAL" redefined
  <command-line>: error: this is the location of the previous definition
  make[3]: *** [util/parse-events-bison.o] Error 1

Comments from Jiri Olsa:

"Reason is the parse error handling that was added just recently: it
adds YYLTYPE type (which is not present in pmu-bison.h), so
YYLTYPE_IS_TRIVIAL gets redefined, which is ok in F20 that handle the
error via '-w' option, but it's not ok for RHEL6 where the '-w' does not
work for this kind of error."

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1430322871-18107-1-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/Build | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/util/Build b/tools/perf/util/Build
index b6c3f39..28af8e2 100644
--- a/tools/perf/util/Build
+++ b/tools/perf/util/Build
@@ -118,7 +118,7 @@ $(OUTPUT)util/pmu-bison.c: util/pmu.y
 
 CFLAGS_parse-events-flex.o  += -w
 CFLAGS_pmu-flex.o           += -w
-CFLAGS_parse-events-bison.o += -DYYENABLE_NLS=0 -DYYLTYPE_IS_TRIVIAL=0 -w
+CFLAGS_parse-events-bison.o += -DYYENABLE_NLS=0 -w
 CFLAGS_pmu-bison.o          += -DYYENABLE_NLS=0 -DYYLTYPE_IS_TRIVIAL=0 -w
 
 $(OUTPUT)util/parse-events.o: $(OUTPUT)util/parse-events-flex.c $(OUTPUT)util/parse-events-bison.c
-- 
2.1.0

[PATCH 12/21] perf probe: Improve detection of file/function name in the probe pattern

[Arnaldo Carvalho de Melo]

From: "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>

Currently, perf probe considers patterns including a '.' to be a file.
However, this causes problems on powerpc ABIv1 where all functions have
a leading '.':

  $ perf probe -F | grep schedule_timeout_interruptible
  .schedule_timeout_interruptible
  $ perf probe .schedule_timeout_interruptible
  Semantic error :File always requires line number or lazy pattern.
    Error: Command Parse Error.

Fix this:
- by checking the probe pattern in more detail, and
- skipping leading dot if one exists when creating/deleting events.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: http://lkml.kernel.org/r/db680f7cb11c4452b632f908e67151f3aa0f4602.1430217967.git.naveen.n.rao@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/probe-event.c | 29 ++++++++++++++++++++++++++---
 1 file changed, 26 insertions(+), 3 deletions(-)

diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index eb75a5e..416c10f 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -1077,6 +1077,7 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
 	struct perf_probe_point *pp = &pev->point;
 	char *ptr, *tmp;
 	char c, nc = 0;
+	bool file_spec = false;
 	/*
 	 * <Syntax>
 	 * perf probe [EVENT=]SRC[:LN|;PTN]
@@ -1105,6 +1106,23 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
 		arg = tmp;
 	}
 
+	/*
+	 * Check arg is function or file name and copy it.
+	 *
+	 * We consider arg to be a file spec if and only if it satisfies
+	 * all of the below criteria::
+	 * - it does not include any of "+@%",
+	 * - it includes one of ":;", and
+	 * - it has a period '.' in the name.
+	 *
+	 * Otherwise, we consider arg to be a function specification.
+	 */
+	if (!strpbrk(arg, "+@%") && (ptr = strpbrk(arg, ";:")) != NULL) {
+		/* This is a file spec if it includes a '.' before ; or : */
+		if (memchr(arg, '.', ptr - arg))
+			file_spec = true;
+	}
+
 	ptr = strpbrk(arg, ";:+@%");
 	if (ptr) {
 		nc = *ptr;
@@ -1115,10 +1133,9 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
 	if (tmp == NULL)
 		return -ENOMEM;
 
-	/* Check arg is function or file and copy it */
-	if (strchr(tmp, '.'))	/* File */
+	if (file_spec)
 		pp->file = tmp;
-	else			/* Function */
+	else
 		pp->function = tmp;
 
 	/* Parse other options */
@@ -2265,6 +2282,9 @@ static int get_new_event_name(char *buf, size_t len, const char *base,
 {
 	int i, ret;
 
+	if (*base == '.')
+		base++;
+
 	/* Try no suffix */
 	ret = e_snprintf(buf, len, "%s", base);
 	if (ret < 0) {
@@ -2766,6 +2786,9 @@ int del_perf_probe_events(struct strlist *dellist)
 			event = str;
 		}
 
+		if (event && *event == '.')
+			event++;
+
 		ret = e_snprintf(buf, 128, "%s:%s", group, event);
 		if (ret < 0) {
 			pr_err("Failed to copy event.");
-- 
2.1.0

[PATCH 13/21] perf tools: Improve strfilter to append additional rules

[Arnaldo Carvalho de Melo]

From: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>

Add strfilter__or/and to append additional rules to existing strfilter.

Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20150424094744.23967.97614.stgit@localhost.localdomain
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/strfilter.c | 40 ++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/strfilter.h | 26 ++++++++++++++++++++++++++
 2 files changed, 66 insertions(+)

diff --git a/tools/perf/util/strfilter.c b/tools/perf/util/strfilter.c
index 79a757a..f3429cd 100644
--- a/tools/perf/util/strfilter.c
+++ b/tools/perf/util/strfilter.c
@@ -170,6 +170,46 @@ struct strfilter *strfilter__new(const char *rules, const char **err)
 	return filter;
 }
 
+static int strfilter__append(struct strfilter *filter, bool _or,
+			     const char *rules, const char **err)
+{
+	struct strfilter_node *right, *root;
+	const char *ep = NULL;
+
+	if (!filter || !rules)
+		return -EINVAL;
+
+	right = strfilter_node__new(rules, &ep);
+	if (!right || *ep != '\0') {
+		if (err)
+			*err = ep;
+		goto error;
+	}
+	root = strfilter_node__alloc(_or ? OP_or : OP_and, filter->root, right);
+	if (!root) {
+		ep = NULL;
+		goto error;
+	}
+
+	filter->root = root;
+	return 0;
+
+error:
+	strfilter_node__delete(right);
+	return ep ? -EINVAL : -ENOMEM;
+}
+
+int strfilter__or(struct strfilter *filter, const char *rules, const char **err)
+{
+	return strfilter__append(filter, true, rules, err);
+}
+
+int strfilter__and(struct strfilter *filter, const char *rules,
+		   const char **err)
+{
+	return strfilter__append(filter, false, rules, err);
+}
+
 static bool strfilter_node__compare(struct strfilter_node *node,
 				    const char *str)
 {
diff --git a/tools/perf/util/strfilter.h b/tools/perf/util/strfilter.h
index fe611f3..d007cdc 100644
--- a/tools/perf/util/strfilter.h
+++ b/tools/perf/util/strfilter.h
@@ -29,6 +29,32 @@ struct strfilter {
 struct strfilter *strfilter__new(const char *rules, const char **err);
 
 /**
+ * strfilter__or - Append an additional rule by logical-or
+ * @filter: Original string filter
+ * @rules: Filter rule to be appended at left of the root of
+ *         @filter by using logical-or.
+ * @err: Pointer which points an error detected on @rules
+ *
+ * Parse @rules and join it to the @filter by using logical-or.
+ * Return 0 if success, or return the error code.
+ */
+int strfilter__or(struct strfilter *filter,
+		  const char *rules, const char **err);
+
+/**
+ * strfilter__add - Append an additional rule by logical-and
+ * @filter: Original string filter
+ * @rules: Filter rule to be appended at left of the root of
+ *         @filter by using logical-and.
+ * @err: Pointer which points an error detected on @rules
+ *
+ * Parse @rules and join it to the @filter by using logical-and.
+ * Return 0 if success, or return the error code.
+ */
+int strfilter__and(struct strfilter *filter,
+		   const char *rules, const char **err);
+
+/**
  * strfilter__compare - compare given string and a string filter
  * @filter: String filter
  * @str: target string
-- 
2.1.0

[PATCH 14/21] perf tools: Add strfilter__string to recover rules string

[Arnaldo Carvalho de Melo]

From: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>

Add strfilter__string to recover rules string from strfilter.  This will
be good for debugging.

Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20150424094746.23967.52434.stgit@localhost.localdomain
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/strfilter.c | 67 +++++++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/strfilter.h |  9 ++++++
 2 files changed, 76 insertions(+)

diff --git a/tools/perf/util/strfilter.c b/tools/perf/util/strfilter.c
index f3429cd..bcae659 100644
--- a/tools/perf/util/strfilter.c
+++ b/tools/perf/util/strfilter.c
@@ -237,3 +237,70 @@ bool strfilter__compare(struct strfilter *filter, const char *str)
 		return false;
 	return strfilter_node__compare(filter->root, str);
 }
+
+static int strfilter_node__sprint(struct strfilter_node *node, char *buf);
+
+/* sprint node in parenthesis if needed */
+static int strfilter_node__sprint_pt(struct strfilter_node *node, char *buf)
+{
+	int len;
+	int pt = node->r ? 2 : 0;	/* don't need to check node->l */
+
+	if (buf && pt)
+		*buf++ = '(';
+	len = strfilter_node__sprint(node, buf);
+	if (len < 0)
+		return len;
+	if (buf && pt)
+		*(buf + len) = ')';
+	return len + pt;
+}
+
+static int strfilter_node__sprint(struct strfilter_node *node, char *buf)
+{
+	int len = 0, rlen;
+
+	if (!node || !node->p)
+		return -EINVAL;
+
+	switch (*node->p) {
+	case '|':
+	case '&':
+		len = strfilter_node__sprint_pt(node->l, buf);
+		if (len < 0)
+			return len;
+	case '!':
+		if (buf) {
+			*(buf + len++) = *node->p;
+			buf += len;
+		} else
+			len++;
+		rlen = strfilter_node__sprint_pt(node->r, buf);
+		if (rlen < 0)
+			return rlen;
+		len += rlen;
+		break;
+	default:
+		len = strlen(node->p);
+		if (buf)
+			strcpy(buf, node->p);
+	}
+
+	return len;
+}
+
+char *strfilter__string(struct strfilter *filter)
+{
+	int len;
+	char *ret = NULL;
+
+	len = strfilter_node__sprint(filter->root, NULL);
+	if (len < 0)
+		return NULL;
+
+	ret = malloc(len + 1);
+	if (ret)
+		strfilter_node__sprint(filter->root, ret);
+
+	return ret;
+}
diff --git a/tools/perf/util/strfilter.h b/tools/perf/util/strfilter.h
index d007cdc..cff5eda 100644
--- a/tools/perf/util/strfilter.h
+++ b/tools/perf/util/strfilter.h
@@ -71,4 +71,13 @@ bool strfilter__compare(struct strfilter *filter, const char *str);
  */
 void strfilter__delete(struct strfilter *filter);
 
+/**
+ * strfilter__string - Reconstruct a rule string from filter
+ * @filter: String filter to reconstruct
+ *
+ * Reconstruct a rule string from @filter. This will be good for
+ * debug messages. Note that returning string must be freed afterward.
+ */
+char *strfilter__string(struct strfilter *filter);
+
 #endif
-- 
2.1.0

[PATCH 15/21] perf probe: Accept multiple filter options

[Arnaldo Carvalho de Melo]

From: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>

Accept multiple filter options. Each filters are combined by logical-or.
E.g. --filter abc* --filter *def is same as --filter abc*|*def

Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20150424094748.23967.63355.stgit@localhost.localdomain
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-probe.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c
index 92dcce0..be17075 100644
--- a/tools/perf/builtin-probe.c
+++ b/tools/perf/builtin-probe.c
@@ -262,21 +262,25 @@ static int opt_set_filter(const struct option *opt __maybe_unused,
 			  const char *str, int unset __maybe_unused)
 {
 	const char *err;
+	int ret = 0;
 
 	if (str) {
 		pr_debug2("Set filter: %s\n", str);
-		if (params.filter)
-			strfilter__delete(params.filter);
-		params.filter = strfilter__new(str, &err);
 		if (!params.filter) {
+			params.filter = strfilter__new(str, &err);
+			if (!params.filter)
+				ret = err ? -EINVAL : -ENOMEM;
+		} else
+			ret = strfilter__or(params.filter, str, &err);
+
+		if (ret == -EINVAL) {
 			pr_err("Filter parse error at %td.\n", err - str + 1);
 			pr_err("Source: \"%s\"\n", str);
 			pr_err("         %*c\n", (int)(err - str + 1), '^');
-			return -EINVAL;
 		}
 	}
 
-	return 0;
+	return ret;
 }
 
 static int init_params(void)
-- 
2.1.0

[PATCH 16/21] perf probe: Accept filter argument for --list

[Arnaldo Carvalho de Melo]

From: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>

Currently, perf-probe --list option ignores given event filter.
  ----
  # ./perf probe -l vfs\*
    probe:vfs_read       (on vfs_read@ksrc/linux-3/fs/read_write.c)
    probe_libc:malloc    (on __libc_malloc@malloc/malloc.c in /usr/lib64/libc-2.17.so)
  ----

This changes --list option to accept the event filter argument as below.
  ----
  # ./perf probe -l vfs\*
    probe:vfs_read       (on vfs_read@ksrc/linux-3/fs/read_write.c)
  # ./perf probe -l \*libc:\*
    probe_libc:malloc    (on __libc_malloc@malloc/malloc.c in /usr/lib64/libc-2.17.so)
  ----

Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20150424094750.23967.53868.stgit@localhost.localdomain
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/Documentation/perf-probe.txt |  6 +--
 tools/perf/builtin-probe.c              | 70 ++++++++++++++++++++++-----------
 tools/perf/util/probe-event.c           | 27 +++++++++++--
 tools/perf/util/probe-event.h           |  2 +-
 4 files changed, 73 insertions(+), 32 deletions(-)

diff --git a/tools/perf/Documentation/perf-probe.txt b/tools/perf/Documentation/perf-probe.txt
index a4a3cc7..d0feb8e 100644
--- a/tools/perf/Documentation/perf-probe.txt
+++ b/tools/perf/Documentation/perf-probe.txt
@@ -14,7 +14,7 @@ or
 or
 'perf probe' [options] --del='[GROUP:]EVENT' [...]
 or
-'perf probe' --list
+'perf probe' --list[=[GROUP:]EVENT]
 or
 'perf probe' [options] --line='LINE'
 or
@@ -66,8 +66,8 @@ OPTIONS
 	classes(e.g. [a-z], [!A-Z]).
 
 -l::
---list::
-	List up current probe events.
+--list[=[GROUP:]EVENT]::
+	List up current probe events. This can also accept filtering patterns of event names.
 
 -L::
 --line=::
diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c
index be17075..feca316 100644
--- a/tools/perf/builtin-probe.c
+++ b/tools/perf/builtin-probe.c
@@ -44,6 +44,7 @@
 
 #define DEFAULT_VAR_FILTER "!__k???tab_* & !__crc_*"
 #define DEFAULT_FUNC_FILTER "!_*"
+#define DEFAULT_LIST_FILTER "*:*"
 
 /* Session management structure */
 static struct {
@@ -93,6 +94,28 @@ static int parse_probe_event(const char *str)
 	return ret;
 }
 
+static int params_add_filter(const char *str)
+{
+	const char *err = NULL;
+	int ret = 0;
+
+	pr_debug2("Add filter: %s\n", str);
+	if (!params.filter) {
+		params.filter = strfilter__new(str, &err);
+		if (!params.filter)
+			ret = err ? -EINVAL : -ENOMEM;
+	} else
+		ret = strfilter__or(params.filter, str, &err);
+
+	if (ret == -EINVAL) {
+		pr_err("Filter parse error at %td.\n", err - str + 1);
+		pr_err("Source: \"%s\"\n", str);
+		pr_err("         %*c\n", (int)(err - str + 1), '^');
+	}
+
+	return ret;
+}
+
 static int set_target(const char *ptr)
 {
 	int found = 0;
@@ -180,6 +203,18 @@ static int opt_del_probe_event(const struct option *opt __maybe_unused,
 	return 0;
 }
 
+static int opt_list_probe_event(const struct option *opt __maybe_unused,
+				const char *str, int unset)
+{
+	if (!unset)
+		params.list_events = true;
+
+	if (str)
+		return params_add_filter(str);
+
+	return 0;
+}
+
 static int opt_set_target(const struct option *opt, const char *str,
 			int unset __maybe_unused)
 {
@@ -261,26 +296,10 @@ static int opt_show_vars(const struct option *opt __maybe_unused,
 static int opt_set_filter(const struct option *opt __maybe_unused,
 			  const char *str, int unset __maybe_unused)
 {
-	const char *err;
-	int ret = 0;
+	if (str)
+		return params_add_filter(str);
 
-	if (str) {
-		pr_debug2("Set filter: %s\n", str);
-		if (!params.filter) {
-			params.filter = strfilter__new(str, &err);
-			if (!params.filter)
-				ret = err ? -EINVAL : -ENOMEM;
-		} else
-			ret = strfilter__or(params.filter, str, &err);
-
-		if (ret == -EINVAL) {
-			pr_err("Filter parse error at %td.\n", err - str + 1);
-			pr_err("Source: \"%s\"\n", str);
-			pr_err("         %*c\n", (int)(err - str + 1), '^');
-		}
-	}
-
-	return ret;
+	return 0;
 }
 
 static int init_params(void)
@@ -320,21 +339,22 @@ __cmd_probe(int argc, const char **argv, const char *prefix __maybe_unused)
 		"perf probe [<options>] 'PROBEDEF' ['PROBEDEF' ...]",
 		"perf probe [<options>] --add 'PROBEDEF' [--add 'PROBEDEF' ...]",
 		"perf probe [<options>] --del '[GROUP:]EVENT' ...",
-		"perf probe --list",
+		"perf probe --list [GROUP:]EVENT ...",
 #ifdef HAVE_DWARF_SUPPORT
 		"perf probe [<options>] --line 'LINEDESC'",
 		"perf probe [<options>] --vars 'PROBEPOINT'",
 #endif
 		"perf probe [<options>] --funcs",
 		NULL
-};
+	};
 	struct option options[] = {
 	OPT_INCR('v', "verbose", &verbose,
 		    "be more verbose (show parsed arguments, etc)"),
 	OPT_BOOLEAN('q', "quiet", &params.quiet,
 		    "be quiet (do not show any mesages)"),
-	OPT_BOOLEAN('l', "list", &params.list_events,
-		    "list up current probe events"),
+	OPT_CALLBACK_DEFAULT('l', "list", NULL, "[GROUP:]EVENT",
+			     "list up probe events", opt_list_probe_event,
+			     DEFAULT_LIST_FILTER),
 	OPT_CALLBACK('d', "del", NULL, "[GROUP:]EVENT", "delete a probe event.",
 		opt_del_probe_event),
 	OPT_CALLBACK('a', "add", NULL,
@@ -448,7 +468,9 @@ __cmd_probe(int argc, const char **argv, const char *prefix __maybe_unused)
 			pr_warning("  Error: Don't use --list with --exec.\n");
 			usage_with_options(probe_usage, options);
 		}
-		ret = show_perf_probe_events();
+		ret = show_perf_probe_events(params.filter);
+		strfilter__delete(params.filter);
+		params.filter = NULL;
 		if (ret < 0)
 			pr_err_with_code("  Error: Failed to show event list.", ret);
 		return ret;
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index 416c10f..5995d81 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -2146,7 +2146,23 @@ static int show_perf_probe_event(struct perf_probe_event *pev,
 	return ret;
 }
 
-static int __show_perf_probe_events(int fd, bool is_kprobe)
+static bool filter_probe_trace_event(struct probe_trace_event *tev,
+				     struct strfilter *filter)
+{
+	char tmp[128];
+
+	/* At first, check the event name itself */
+	if (strfilter__compare(filter, tev->event))
+		return true;
+
+	/* Next, check the combination of name and group */
+	if (e_snprintf(tmp, 128, "%s:%s", tev->group, tev->event) < 0)
+		return false;
+	return strfilter__compare(filter, tmp);
+}
+
+static int __show_perf_probe_events(int fd, bool is_kprobe,
+				    struct strfilter *filter)
 {
 	int ret = 0;
 	struct probe_trace_event tev;
@@ -2164,12 +2180,15 @@ static int __show_perf_probe_events(int fd, bool is_kprobe)
 	strlist__for_each(ent, rawlist) {
 		ret = parse_probe_trace_command(ent->s, &tev);
 		if (ret >= 0) {
+			if (!filter_probe_trace_event(&tev, filter))
+				goto next;
 			ret = convert_to_perf_probe_event(&tev, &pev,
 								is_kprobe);
 			if (ret >= 0)
 				ret = show_perf_probe_event(&pev,
 							    tev.point.module);
 		}
+next:
 		clear_perf_probe_event(&pev);
 		clear_probe_trace_event(&tev);
 		if (ret < 0)
@@ -2181,7 +2200,7 @@ static int __show_perf_probe_events(int fd, bool is_kprobe)
 }
 
 /* List up current perf-probe events */
-int show_perf_probe_events(void)
+int show_perf_probe_events(struct strfilter *filter)
 {
 	int kp_fd, up_fd, ret;
 
@@ -2193,7 +2212,7 @@ int show_perf_probe_events(void)
 
 	kp_fd = open_kprobe_events(false);
 	if (kp_fd >= 0) {
-		ret = __show_perf_probe_events(kp_fd, true);
+		ret = __show_perf_probe_events(kp_fd, true, filter);
 		close(kp_fd);
 		if (ret < 0)
 			goto out;
@@ -2207,7 +2226,7 @@ int show_perf_probe_events(void)
 	}
 
 	if (up_fd >= 0) {
-		ret = __show_perf_probe_events(up_fd, false);
+		ret = __show_perf_probe_events(up_fd, false, filter);
 		close(up_fd);
 	}
 out:
diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h
index 180f142..ec13362 100644
--- a/tools/perf/util/probe-event.h
+++ b/tools/perf/util/probe-event.h
@@ -127,7 +127,7 @@ extern const char *kernel_get_module_path(const char *module);
 extern int add_perf_probe_events(struct perf_probe_event *pevs, int npevs,
 				 int max_probe_points, bool force_add);
 extern int del_perf_probe_events(struct strlist *dellist);
-extern int show_perf_probe_events(void);
+extern int show_perf_probe_events(struct strfilter *filter);
 extern int show_line_range(struct line_range *lr, const char *module,
 			   bool user);
 extern int show_available_vars(struct perf_probe_event *pevs, int npevs,
-- 
2.1.0

[PATCH 17/21] perf kmem: Implement stat --page --caller

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

It is 'perf kmem' support caller statistics for page.  Unlike slab case,
the tracepoints in page allocator don't provide callsite info.  So it
records with callchain and extracts callsite info.

Note that the callchain contains several memory allocation functions
which has no meaning for users.  So skip those functions to get proper
callsites.  I used following regex pattern to skip the allocator
functions:

  ^_?_?(alloc|get_free|get_zeroed)_pages?

This gave me a following list of functions:

  # perf kmem record --page sleep 3
  # perf kmem stat --page -v
  ...
  alloc func: __get_free_pages
  alloc func: get_zeroed_page
  alloc func: alloc_pages_exact
  alloc func: __alloc_pages_direct_compact
  alloc func: __alloc_pages_nodemask
  alloc func: alloc_page_interleave
  alloc func: alloc_pages_current
  alloc func: alloc_pages_vma
  alloc func: alloc_page_buffers
  alloc func: alloc_pages_exact_nid
  ...

The output looks mostly same as --alloc (I also added callsite column
to that) but groups entries by callsite.  Currently, the order,
migrate type and GFP flag info is for the last allocation and not
guaranteed to be same for all allocations from the callsite.

  ---------------------------------------------------------------------------------------------
   Total_alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite
  ---------------------------------------------------------------------------------------------
              1,064 |       266 |     0 | UNMOVABL |  000000d0 | __pollwait
                 52 |        13 |     0 | UNMOVABL |  002084d0 | pte_alloc_one
                 44 |        11 |     0 |  MOVABLE |  000280da | handle_mm_fault
                 20 |         5 |     0 |  MOVABLE |  000200da | do_cow_fault
                 20 |         5 |     0 |  MOVABLE |  000200da | do_wp_page
                 16 |         4 |     0 | UNMOVABL |  000084d0 | __pmd_alloc
                 16 |         4 |     0 | UNMOVABL |  00000200 | __tlb_remove_page
                 12 |         3 |     0 | UNMOVABL |  000084d0 | __pud_alloc
                  8 |         2 |     0 | UNMOVABL |  00000010 | bio_copy_user_iov
                  4 |         1 |     0 | UNMOVABL |  000200d2 | pipe_write
                  4 |         1 |     0 |  MOVABLE |  000280da | do_wp_page
                  4 |         1 |     0 | UNMOVABL |  002084d0 | pgd_alloc
  ---------------------------------------------------------------------------------------------

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Joonsoo Kim <js1304@gmail.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/1429592107-1807-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-kmem.c | 327 +++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 306 insertions(+), 21 deletions(-)

diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 4f0f384..3649eec 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -10,6 +10,7 @@
 #include "util/header.h"
 #include "util/session.h"
 #include "util/tool.h"
+#include "util/callchain.h"
 
 #include "util/parse-options.h"
 #include "util/trace-event.h"
@@ -21,6 +22,7 @@
 #include <linux/rbtree.h>
 #include <linux/string.h>
 #include <locale.h>
+#include <regex.h>
 
 static int	kmem_slab;
 static int	kmem_page;
@@ -241,6 +243,7 @@ static unsigned long nr_page_fails;
 static unsigned long nr_page_nomatch;
 
 static bool use_pfn;
+static struct perf_session *kmem_session;
 
 #define MAX_MIGRATE_TYPES  6
 #define MAX_PAGE_ORDER     11
@@ -250,6 +253,7 @@ static int order_stats[MAX_PAGE_ORDER][MAX_MIGRATE_TYPES];
 struct page_stat {
 	struct rb_node 	node;
 	u64 		page;
+	u64 		callsite;
 	int 		order;
 	unsigned 	gfp_flags;
 	unsigned 	migrate_type;
@@ -262,8 +266,144 @@ struct page_stat {
 static struct rb_root page_tree;
 static struct rb_root page_alloc_tree;
 static struct rb_root page_alloc_sorted;
+static struct rb_root page_caller_tree;
+static struct rb_root page_caller_sorted;
 
-static struct page_stat *search_page(unsigned long page, bool create)
+struct alloc_func {
+	u64 start;
+	u64 end;
+	char *name;
+};
+
+static int nr_alloc_funcs;
+static struct alloc_func *alloc_func_list;
+
+static int funcmp(const void *a, const void *b)
+{
+	const struct alloc_func *fa = a;
+	const struct alloc_func *fb = b;
+
+	if (fa->start > fb->start)
+		return 1;
+	else
+		return -1;
+}
+
+static int callcmp(const void *a, const void *b)
+{
+	const struct alloc_func *fa = a;
+	const struct alloc_func *fb = b;
+
+	if (fb->start <= fa->start && fa->end < fb->end)
+		return 0;
+
+	if (fa->start > fb->start)
+		return 1;
+	else
+		return -1;
+}
+
+static int build_alloc_func_list(void)
+{
+	int ret;
+	struct map *kernel_map;
+	struct symbol *sym;
+	struct rb_node *node;
+	struct alloc_func *func;
+	struct machine *machine = &kmem_session->machines.host;
+	regex_t alloc_func_regex;
+	const char pattern[] = "^_?_?(alloc|get_free|get_zeroed)_pages?";
+
+	ret = regcomp(&alloc_func_regex, pattern, REG_EXTENDED);
+	if (ret) {
+		char err[BUFSIZ];
+
+		regerror(ret, &alloc_func_regex, err, sizeof(err));
+		pr_err("Invalid regex: %s\n%s", pattern, err);
+		return -EINVAL;
+	}
+
+	kernel_map = machine->vmlinux_maps[MAP__FUNCTION];
+	if (map__load(kernel_map, NULL) < 0) {
+		pr_err("cannot load kernel map\n");
+		return -ENOENT;
+	}
+
+	map__for_each_symbol(kernel_map, sym, node) {
+		if (regexec(&alloc_func_regex, sym->name, 0, NULL, 0))
+			continue;
+
+		func = realloc(alloc_func_list,
+			       (nr_alloc_funcs + 1) * sizeof(*func));
+		if (func == NULL)
+			return -ENOMEM;
+
+		pr_debug("alloc func: %s\n", sym->name);
+		func[nr_alloc_funcs].start = sym->start;
+		func[nr_alloc_funcs].end   = sym->end;
+		func[nr_alloc_funcs].name  = sym->name;
+
+		alloc_func_list = func;
+		nr_alloc_funcs++;
+	}
+
+	qsort(alloc_func_list, nr_alloc_funcs, sizeof(*func), funcmp);
+
+	regfree(&alloc_func_regex);
+	return 0;
+}
+
+/*
+ * Find first non-memory allocation function from callchain.
+ * The allocation functions are in the 'alloc_func_list'.
+ */
+static u64 find_callsite(struct perf_evsel *evsel, struct perf_sample *sample)
+{
+	struct addr_location al;
+	struct machine *machine = &kmem_session->machines.host;
+	struct callchain_cursor_node *node;
+
+	if (alloc_func_list == NULL) {
+		if (build_alloc_func_list() < 0)
+			goto out;
+	}
+
+	al.thread = machine__findnew_thread(machine, sample->pid, sample->tid);
+	sample__resolve_callchain(sample, NULL, evsel, &al, 16);
+
+	callchain_cursor_commit(&callchain_cursor);
+	while (true) {
+		struct alloc_func key, *caller;
+		u64 addr;
+
+		node = callchain_cursor_current(&callchain_cursor);
+		if (node == NULL)
+			break;
+
+		key.start = key.end = node->ip;
+		caller = bsearch(&key, alloc_func_list, nr_alloc_funcs,
+				 sizeof(key), callcmp);
+		if (!caller) {
+			/* found */
+			if (node->map)
+				addr = map__unmap_ip(node->map, node->ip);
+			else
+				addr = node->ip;
+
+			return addr;
+		} else
+			pr_debug3("skipping alloc function: %s\n", caller->name);
+
+		callchain_cursor_advance(&callchain_cursor);
+	}
+
+out:
+	pr_debug2("unknown callsite: %"PRIx64 "\n", sample->ip);
+	return sample->ip;
+}
+
+static struct page_stat *
+__page_stat__findnew_page(u64 page, bool create)
 {
 	struct rb_node **node = &page_tree.rb_node;
 	struct rb_node *parent = NULL;
@@ -298,6 +438,16 @@ static struct page_stat *search_page(unsigned long page, bool create)
 	return data;
 }
 
+static struct page_stat *page_stat__find_page(u64 page)
+{
+	return __page_stat__findnew_page(page, false);
+}
+
+static struct page_stat *page_stat__findnew_page(u64 page)
+{
+	return __page_stat__findnew_page(page, true);
+}
+
 static int page_stat_cmp(struct page_stat *a, struct page_stat *b)
 {
 	if (a->page > b->page)
@@ -319,7 +469,8 @@ static int page_stat_cmp(struct page_stat *a, struct page_stat *b)
 	return 0;
 }
 
-static struct page_stat *search_page_alloc_stat(struct page_stat *pstat, bool create)
+static struct page_stat *
+__page_stat__findnew_alloc(struct page_stat *pstat, bool create)
 {
 	struct rb_node **node = &page_alloc_tree.rb_node;
 	struct rb_node *parent = NULL;
@@ -357,6 +508,62 @@ static struct page_stat *search_page_alloc_stat(struct page_stat *pstat, bool cr
 	return data;
 }
 
+static struct page_stat *page_stat__find_alloc(struct page_stat *pstat)
+{
+	return __page_stat__findnew_alloc(pstat, false);
+}
+
+static struct page_stat *page_stat__findnew_alloc(struct page_stat *pstat)
+{
+	return __page_stat__findnew_alloc(pstat, true);
+}
+
+static struct page_stat *
+__page_stat__findnew_caller(u64 callsite, bool create)
+{
+	struct rb_node **node = &page_caller_tree.rb_node;
+	struct rb_node *parent = NULL;
+	struct page_stat *data;
+
+	while (*node) {
+		s64 cmp;
+
+		parent = *node;
+		data = rb_entry(*node, struct page_stat, node);
+
+		cmp = data->callsite - callsite;
+		if (cmp < 0)
+			node = &parent->rb_left;
+		else if (cmp > 0)
+			node = &parent->rb_right;
+		else
+			return data;
+	}
+
+	if (!create)
+		return NULL;
+
+	data = zalloc(sizeof(*data));
+	if (data != NULL) {
+		data->callsite = callsite;
+
+		rb_link_node(&data->node, parent, node);
+		rb_insert_color(&data->node, &page_caller_tree);
+	}
+
+	return data;
+}
+
+static struct page_stat *page_stat__find_caller(u64 callsite)
+{
+	return __page_stat__findnew_caller(callsite, false);
+}
+
+static struct page_stat *page_stat__findnew_caller(u64 callsite)
+{
+	return __page_stat__findnew_caller(callsite, true);
+}
+
 static bool valid_page(u64 pfn_or_page)
 {
 	if (use_pfn && pfn_or_page == -1UL)
@@ -375,6 +582,7 @@ static int perf_evsel__process_page_alloc_event(struct perf_evsel *evsel,
 	unsigned int migrate_type = perf_evsel__intval(evsel, sample,
 						       "migratetype");
 	u64 bytes = kmem_page_size << order;
+	u64 callsite;
 	struct page_stat *pstat;
 	struct page_stat this = {
 		.order = order,
@@ -397,25 +605,40 @@ static int perf_evsel__process_page_alloc_event(struct perf_evsel *evsel,
 		return 0;
 	}
 
+	callsite = find_callsite(evsel, sample);
+
 	/*
 	 * This is to find the current page (with correct gfp flags and
 	 * migrate type) at free event.
 	 */
-	pstat = search_page(page, true);
+	pstat = page_stat__findnew_page(page);
 	if (pstat == NULL)
 		return -ENOMEM;
 
 	pstat->order = order;
 	pstat->gfp_flags = gfp_flags;
 	pstat->migrate_type = migrate_type;
+	pstat->callsite = callsite;
 
 	this.page = page;
-	pstat = search_page_alloc_stat(&this, true);
+	pstat = page_stat__findnew_alloc(&this);
 	if (pstat == NULL)
 		return -ENOMEM;
 
 	pstat->nr_alloc++;
 	pstat->alloc_bytes += bytes;
+	pstat->callsite = callsite;
+
+	pstat = page_stat__findnew_caller(callsite);
+	if (pstat == NULL)
+		return -ENOMEM;
+
+	pstat->order = order;
+	pstat->gfp_flags = gfp_flags;
+	pstat->migrate_type = migrate_type;
+
+	pstat->nr_alloc++;
+	pstat->alloc_bytes += bytes;
 
 	order_stats[order][migrate_type]++;
 
@@ -441,7 +664,7 @@ static int perf_evsel__process_page_free_event(struct perf_evsel *evsel,
 	nr_page_frees++;
 	total_page_free_bytes += bytes;
 
-	pstat = search_page(page, false);
+	pstat = page_stat__find_page(page);
 	if (pstat == NULL) {
 		pr_debug2("missing free at page %"PRIx64" (order: %d)\n",
 			  page, order);
@@ -455,11 +678,19 @@ static int perf_evsel__process_page_free_event(struct perf_evsel *evsel,
 	this.page = page;
 	this.gfp_flags = pstat->gfp_flags;
 	this.migrate_type = pstat->migrate_type;
+	this.callsite = pstat->callsite;
 
 	rb_erase(&pstat->node, &page_tree);
 	free(pstat);
 
-	pstat = search_page_alloc_stat(&this, false);
+	pstat = page_stat__find_alloc(&this);
+	if (pstat == NULL)
+		return -ENOENT;
+
+	pstat->nr_free++;
+	pstat->free_bytes += bytes;
+
+	pstat = page_stat__find_caller(this.callsite);
 	if (pstat == NULL)
 		return -ENOENT;
 
@@ -576,41 +807,89 @@ static const char * const migrate_type_str[] = {
 	"UNKNOWN",
 };
 
-static void __print_page_result(struct rb_root *root,
-				struct perf_session *session __maybe_unused,
-				int n_lines)
+static void __print_page_alloc_result(struct perf_session *session, int n_lines)
 {
-	struct rb_node *next = rb_first(root);
+	struct rb_node *next = rb_first(&page_alloc_sorted);
+	struct machine *machine = &session->machines.host;
 	const char *format;
 
-	printf("\n%.80s\n", graph_dotted_line);
-	printf(" %-16s | Total alloc (KB) | Hits      | Order | Mig.type | GFP flags\n",
+	printf("\n%.105s\n", graph_dotted_line);
+	printf(" %-16s | Total alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite\n",
 	       use_pfn ? "PFN" : "Page");
-	printf("%.80s\n", graph_dotted_line);
+	printf("%.105s\n", graph_dotted_line);
 
 	if (use_pfn)
-		format = " %16llu | %'16llu | %'9d | %5d | %8s |  %08lx\n";
+		format = " %16llu | %'16llu | %'9d | %5d | %8s |  %08lx | %s\n";
 	else
-		format = " %016llx | %'16llu | %'9d | %5d | %8s |  %08lx\n";
+		format = " %016llx | %'16llu | %'9d | %5d | %8s |  %08lx | %s\n";
 
 	while (next && n_lines--) {
 		struct page_stat *data;
+		struct symbol *sym;
+		struct map *map;
+		char buf[32];
+		char *caller = buf;
 
 		data = rb_entry(next, struct page_stat, node);
+		sym = machine__find_kernel_function(machine, data->callsite,
+						    &map, NULL);
+		if (sym && sym->name)
+			caller = sym->name;
+		else
+			scnprintf(buf, sizeof(buf), "%"PRIx64, data->callsite);
 
 		printf(format, (unsigned long long)data->page,
 		       (unsigned long long)data->alloc_bytes / 1024,
 		       data->nr_alloc, data->order,
 		       migrate_type_str[data->migrate_type],
-		       (unsigned long)data->gfp_flags);
+		       (unsigned long)data->gfp_flags, caller);
 
 		next = rb_next(next);
 	}
 
 	if (n_lines == -1)
-		printf(" ...              | ...              | ...       | ...   | ...      | ...     \n");
+		printf(" ...              | ...              | ...       | ...   | ...      | ...       | ...\n");
 
-	printf("%.80s\n", graph_dotted_line);
+	printf("%.105s\n", graph_dotted_line);
+}
+
+static void __print_page_caller_result(struct perf_session *session, int n_lines)
+{
+	struct rb_node *next = rb_first(&page_caller_sorted);
+	struct machine *machine = &session->machines.host;
+
+	printf("\n%.105s\n", graph_dotted_line);
+	printf(" Total alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite\n");
+	printf("%.105s\n", graph_dotted_line);
+
+	while (next && n_lines--) {
+		struct page_stat *data;
+		struct symbol *sym;
+		struct map *map;
+		char buf[32];
+		char *caller = buf;
+
+		data = rb_entry(next, struct page_stat, node);
+		sym = machine__find_kernel_function(machine, data->callsite,
+						    &map, NULL);
+		if (sym && sym->name)
+			caller = sym->name;
+		else
+			scnprintf(buf, sizeof(buf), "%"PRIx64, data->callsite);
+
+		printf(" %'16llu | %'9d | %5d | %8s |  %08lx | %s\n",
+		       (unsigned long long)data->alloc_bytes / 1024,
+		       data->nr_alloc, data->order,
+		       migrate_type_str[data->migrate_type],
+		       (unsigned long)data->gfp_flags, caller);
+
+		next = rb_next(next);
+	}
+
+	if (n_lines == -1)
+		printf(" ...              | ...       | ...   | ...      | ...       | ...\n");
+
+	printf("%.105s\n", graph_dotted_line);
 }
 
 static void print_slab_summary(void)
@@ -682,8 +961,10 @@ static void print_slab_result(struct perf_session *session)
 
 static void print_page_result(struct perf_session *session)
 {
+	if (caller_flag)
+		__print_page_caller_result(session, caller_lines);
 	if (alloc_flag)
-		__print_page_result(&page_alloc_sorted, session, alloc_lines);
+		__print_page_alloc_result(session, alloc_lines);
 	print_page_summary();
 }
 
@@ -802,6 +1083,7 @@ static void sort_result(void)
 	}
 	if (kmem_page) {
 		__sort_page_result(&page_alloc_tree, &page_alloc_sorted);
+		__sort_page_result(&page_caller_tree, &page_caller_sorted);
 	}
 }
 
@@ -1084,7 +1366,7 @@ static int __cmd_record(int argc, const char **argv)
 	if (kmem_slab)
 		rec_argc += ARRAY_SIZE(slab_events);
 	if (kmem_page)
-		rec_argc += ARRAY_SIZE(page_events);
+		rec_argc += ARRAY_SIZE(page_events) + 1; /* for -g */
 
 	rec_argv = calloc(rec_argc + 1, sizeof(char *));
 
@@ -1099,6 +1381,8 @@ static int __cmd_record(int argc, const char **argv)
 			rec_argv[i] = strdup(slab_events[j]);
 	}
 	if (kmem_page) {
+		rec_argv[i++] = strdup("-g");
+
 		for (j = 0; j < ARRAY_SIZE(page_events); j++, i++)
 			rec_argv[i] = strdup(page_events[j]);
 	}
@@ -1159,7 +1443,7 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	file.path = input_name;
 
-	session = perf_session__new(&file, false, &perf_kmem);
+	kmem_session = session = perf_session__new(&file, false, &perf_kmem);
 	if (session == NULL)
 		return -1;
 
@@ -1172,6 +1456,7 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
 		}
 
 		kmem_page_size = pevent_get_page_size(evsel->tp_format->pevent);
+		symbol_conf.use_callchain = true;
 	}
 
 	symbol__init(&session->header.env);
-- 
2.1.0

[PATCH 18/21] perf kmem: Support sort keys on page analysis

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

Add new sort keys for page: page, order, migtype, gfp - existing
'bytes', 'hit' and 'callsite' sort keys also work for page.  Note that
-s/--sort option should be preceded by either of --slab or --page option
to determine where the sort keys applies.

Now it properly groups and sorts allocation stats - so same
page/caller with different order/migtype/gfp will be printed on a
different line.

 # perf kmem stat --page --caller -l 10 -s order,hit

 -----------------------------------------------------------------------------
 Total alloc (KB) | Hits   | Order | Mig.type | GFP flags | Callsite
 -----------------------------------------------------------------------------
               64 |      4 |     2 |  RECLAIM |  00285250 | new_slab
           50,144 | 12,536 |     0 |  MOVABLE |  0102005a | __page_cache_alloc
               52 |     13 |     0 | UNMOVABL |  002084d0 | pte_alloc_one
               40 |     10 |     0 |  MOVABLE |  000280da | handle_mm_fault
               28 |      7 |     0 | UNMOVABL |  000000d0 | __pollwait
               20 |      5 |     0 |  MOVABLE |  000200da | do_wp_page
               20 |      5 |     0 |  MOVABLE |  000200da | do_cow_fault
               16 |      4 |     0 | UNMOVABL |  00000200 | __tlb_remove_page
               16 |      4 |     0 | UNMOVABL |  000084d0 | __pmd_alloc
                8 |      2 |     0 | UNMOVABL |  000084d0 | __pud_alloc
 ...              | ...    | ...   | ...      | ...       | ...
 -----------------------------------------------------------------------------

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Joonsoo Kim <js1304@gmail.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/1429592107-1807-3-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/Documentation/perf-kmem.txt |   6 +-
 tools/perf/builtin-kmem.c              | 393 ++++++++++++++++++++++++++-------
 2 files changed, 313 insertions(+), 86 deletions(-)

diff --git a/tools/perf/Documentation/perf-kmem.txt b/tools/perf/Documentation/perf-kmem.txt
index 23219c6..69e1812 100644
--- a/tools/perf/Documentation/perf-kmem.txt
+++ b/tools/perf/Documentation/perf-kmem.txt
@@ -37,7 +37,11 @@ OPTIONS
 
 -s <key[,key2...]>::
 --sort=<key[,key2...]>::
-	Sort the output (default: frag,hit,bytes)
+	Sort the output (default: 'frag,hit,bytes' for slab and 'bytes,hit'
+	for page).  Available sort keys are 'ptr, callsite, bytes, hit,
+	pingpong, frag' for slab and 'page, callsite, bytes, hit, order,
+	migtype, gfp' for page.  This option should be preceded by one of the
+	mode selection options - i.e. --slab, --page, --alloc and/or --caller.
 
 -l <num>::
 --line=<num>::
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 3649eec..0393a7f 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -30,7 +30,7 @@ static int	kmem_page;
 static long	kmem_page_size;
 
 struct alloc_stat;
-typedef int (*sort_fn_t)(struct alloc_stat *, struct alloc_stat *);
+typedef int (*sort_fn_t)(void *, void *);
 
 static int			alloc_flag;
 static int			caller_flag;
@@ -181,8 +181,8 @@ static int perf_evsel__process_alloc_node_event(struct perf_evsel *evsel,
 	return ret;
 }
 
-static int ptr_cmp(struct alloc_stat *, struct alloc_stat *);
-static int callsite_cmp(struct alloc_stat *, struct alloc_stat *);
+static int ptr_cmp(void *, void *);
+static int slab_callsite_cmp(void *, void *);
 
 static struct alloc_stat *search_alloc_stat(unsigned long ptr,
 					    unsigned long call_site,
@@ -223,7 +223,8 @@ static int perf_evsel__process_free_event(struct perf_evsel *evsel,
 		s_alloc->pingpong++;
 
 		s_caller = search_alloc_stat(0, s_alloc->call_site,
-					     &root_caller_stat, callsite_cmp);
+					     &root_caller_stat,
+					     slab_callsite_cmp);
 		if (!s_caller)
 			return -1;
 		s_caller->pingpong++;
@@ -448,26 +449,14 @@ static struct page_stat *page_stat__findnew_page(u64 page)
 	return __page_stat__findnew_page(page, true);
 }
 
-static int page_stat_cmp(struct page_stat *a, struct page_stat *b)
-{
-	if (a->page > b->page)
-		return -1;
-	if (a->page < b->page)
-		return 1;
-	if (a->order > b->order)
-		return -1;
-	if (a->order < b->order)
-		return 1;
-	if (a->migrate_type > b->migrate_type)
-		return -1;
-	if (a->migrate_type < b->migrate_type)
-		return 1;
-	if (a->gfp_flags > b->gfp_flags)
-		return -1;
-	if (a->gfp_flags < b->gfp_flags)
-		return 1;
-	return 0;
-}
+struct sort_dimension {
+	const char		name[20];
+	sort_fn_t		cmp;
+	struct list_head	list;
+};
+
+static LIST_HEAD(page_alloc_sort_input);
+static LIST_HEAD(page_caller_sort_input);
 
 static struct page_stat *
 __page_stat__findnew_alloc(struct page_stat *pstat, bool create)
@@ -475,14 +464,20 @@ __page_stat__findnew_alloc(struct page_stat *pstat, bool create)
 	struct rb_node **node = &page_alloc_tree.rb_node;
 	struct rb_node *parent = NULL;
 	struct page_stat *data;
+	struct sort_dimension *sort;
 
 	while (*node) {
-		s64 cmp;
+		int cmp = 0;
 
 		parent = *node;
 		data = rb_entry(*node, struct page_stat, node);
 
-		cmp = page_stat_cmp(data, pstat);
+		list_for_each_entry(sort, &page_alloc_sort_input, list) {
+			cmp = sort->cmp(pstat, data);
+			if (cmp)
+				break;
+		}
+
 		if (cmp < 0)
 			node = &parent->rb_left;
 		else if (cmp > 0)
@@ -519,19 +514,25 @@ static struct page_stat *page_stat__findnew_alloc(struct page_stat *pstat)
 }
 
 static struct page_stat *
-__page_stat__findnew_caller(u64 callsite, bool create)
+__page_stat__findnew_caller(struct page_stat *pstat, bool create)
 {
 	struct rb_node **node = &page_caller_tree.rb_node;
 	struct rb_node *parent = NULL;
 	struct page_stat *data;
+	struct sort_dimension *sort;
 
 	while (*node) {
-		s64 cmp;
+		int cmp = 0;
 
 		parent = *node;
 		data = rb_entry(*node, struct page_stat, node);
 
-		cmp = data->callsite - callsite;
+		list_for_each_entry(sort, &page_caller_sort_input, list) {
+			cmp = sort->cmp(pstat, data);
+			if (cmp)
+				break;
+		}
+
 		if (cmp < 0)
 			node = &parent->rb_left;
 		else if (cmp > 0)
@@ -545,7 +546,10 @@ __page_stat__findnew_caller(u64 callsite, bool create)
 
 	data = zalloc(sizeof(*data));
 	if (data != NULL) {
-		data->callsite = callsite;
+		data->callsite = pstat->callsite;
+		data->order = pstat->order;
+		data->gfp_flags = pstat->gfp_flags;
+		data->migrate_type = pstat->migrate_type;
 
 		rb_link_node(&data->node, parent, node);
 		rb_insert_color(&data->node, &page_caller_tree);
@@ -554,14 +558,14 @@ __page_stat__findnew_caller(u64 callsite, bool create)
 	return data;
 }
 
-static struct page_stat *page_stat__find_caller(u64 callsite)
+static struct page_stat *page_stat__find_caller(struct page_stat *pstat)
 {
-	return __page_stat__findnew_caller(callsite, false);
+	return __page_stat__findnew_caller(pstat, false);
 }
 
-static struct page_stat *page_stat__findnew_caller(u64 callsite)
+static struct page_stat *page_stat__findnew_caller(struct page_stat *pstat)
 {
-	return __page_stat__findnew_caller(callsite, true);
+	return __page_stat__findnew_caller(pstat, true);
 }
 
 static bool valid_page(u64 pfn_or_page)
@@ -629,14 +633,11 @@ static int perf_evsel__process_page_alloc_event(struct perf_evsel *evsel,
 	pstat->alloc_bytes += bytes;
 	pstat->callsite = callsite;
 
-	pstat = page_stat__findnew_caller(callsite);
+	this.callsite = callsite;
+	pstat = page_stat__findnew_caller(&this);
 	if (pstat == NULL)
 		return -ENOMEM;
 
-	pstat->order = order;
-	pstat->gfp_flags = gfp_flags;
-	pstat->migrate_type = migrate_type;
-
 	pstat->nr_alloc++;
 	pstat->alloc_bytes += bytes;
 
@@ -690,7 +691,7 @@ static int perf_evsel__process_page_free_event(struct perf_evsel *evsel,
 	pstat->nr_free++;
 	pstat->free_bytes += bytes;
 
-	pstat = page_stat__find_caller(this.callsite);
+	pstat = page_stat__find_caller(&this);
 	if (pstat == NULL)
 		return -ENOENT;
 
@@ -976,14 +977,10 @@ static void print_result(struct perf_session *session)
 		print_page_result(session);
 }
 
-struct sort_dimension {
-	const char		name[20];
-	sort_fn_t		cmp;
-	struct list_head	list;
-};
-
-static LIST_HEAD(caller_sort);
-static LIST_HEAD(alloc_sort);
+static LIST_HEAD(slab_caller_sort);
+static LIST_HEAD(slab_alloc_sort);
+static LIST_HEAD(page_caller_sort);
+static LIST_HEAD(page_alloc_sort);
 
 static void sort_slab_insert(struct rb_root *root, struct alloc_stat *data,
 			     struct list_head *sort_list)
@@ -1032,10 +1029,12 @@ static void __sort_slab_result(struct rb_root *root, struct rb_root *root_sorted
 	}
 }
 
-static void sort_page_insert(struct rb_root *root, struct page_stat *data)
+static void sort_page_insert(struct rb_root *root, struct page_stat *data,
+			     struct list_head *sort_list)
 {
 	struct rb_node **new = &root->rb_node;
 	struct rb_node *parent = NULL;
+	struct sort_dimension *sort;
 
 	while (*new) {
 		struct page_stat *this;
@@ -1044,8 +1043,11 @@ static void sort_page_insert(struct rb_root *root, struct page_stat *data)
 		this = rb_entry(*new, struct page_stat, node);
 		parent = *new;
 
-		/* TODO: support more sort key */
-		cmp = data->alloc_bytes - this->alloc_bytes;
+		list_for_each_entry(sort, sort_list, list) {
+			cmp = sort->cmp(data, this);
+			if (cmp)
+				break;
+		}
 
 		if (cmp > 0)
 			new = &parent->rb_left;
@@ -1057,7 +1059,8 @@ static void sort_page_insert(struct rb_root *root, struct page_stat *data)
 	rb_insert_color(&data->node, root);
 }
 
-static void __sort_page_result(struct rb_root *root, struct rb_root *root_sorted)
+static void __sort_page_result(struct rb_root *root, struct rb_root *root_sorted,
+			       struct list_head *sort_list)
 {
 	struct rb_node *node;
 	struct page_stat *data;
@@ -1069,7 +1072,7 @@ static void __sort_page_result(struct rb_root *root, struct rb_root *root_sorted
 
 		rb_erase(node, root);
 		data = rb_entry(node, struct page_stat, node);
-		sort_page_insert(root_sorted, data);
+		sort_page_insert(root_sorted, data, sort_list);
 	}
 }
 
@@ -1077,13 +1080,15 @@ static void sort_result(void)
 {
 	if (kmem_slab) {
 		__sort_slab_result(&root_alloc_stat, &root_alloc_sorted,
-				   &alloc_sort);
+				   &slab_alloc_sort);
 		__sort_slab_result(&root_caller_stat, &root_caller_sorted,
-				   &caller_sort);
+				   &slab_caller_sort);
 	}
 	if (kmem_page) {
-		__sort_page_result(&page_alloc_tree, &page_alloc_sorted);
-		__sort_page_result(&page_caller_tree, &page_caller_sorted);
+		__sort_page_result(&page_alloc_tree, &page_alloc_sorted,
+				   &page_alloc_sort);
+		__sort_page_result(&page_caller_tree, &page_caller_sorted,
+				   &page_caller_sort);
 	}
 }
 
@@ -1132,8 +1137,12 @@ out:
 	return err;
 }
 
-static int ptr_cmp(struct alloc_stat *l, struct alloc_stat *r)
+/* slab sort keys */
+static int ptr_cmp(void *a, void *b)
 {
+	struct alloc_stat *l = a;
+	struct alloc_stat *r = b;
+
 	if (l->ptr < r->ptr)
 		return -1;
 	else if (l->ptr > r->ptr)
@@ -1146,8 +1155,11 @@ static struct sort_dimension ptr_sort_dimension = {
 	.cmp	= ptr_cmp,
 };
 
-static int callsite_cmp(struct alloc_stat *l, struct alloc_stat *r)
+static int slab_callsite_cmp(void *a, void *b)
 {
+	struct alloc_stat *l = a;
+	struct alloc_stat *r = b;
+
 	if (l->call_site < r->call_site)
 		return -1;
 	else if (l->call_site > r->call_site)
@@ -1157,11 +1169,14 @@ static int callsite_cmp(struct alloc_stat *l, struct alloc_stat *r)
 
 static struct sort_dimension callsite_sort_dimension = {
 	.name	= "callsite",
-	.cmp	= callsite_cmp,
+	.cmp	= slab_callsite_cmp,
 };
 
-static int hit_cmp(struct alloc_stat *l, struct alloc_stat *r)
+static int hit_cmp(void *a, void *b)
 {
+	struct alloc_stat *l = a;
+	struct alloc_stat *r = b;
+
 	if (l->hit < r->hit)
 		return -1;
 	else if (l->hit > r->hit)
@@ -1174,8 +1189,11 @@ static struct sort_dimension hit_sort_dimension = {
 	.cmp	= hit_cmp,
 };
 
-static int bytes_cmp(struct alloc_stat *l, struct alloc_stat *r)
+static int bytes_cmp(void *a, void *b)
 {
+	struct alloc_stat *l = a;
+	struct alloc_stat *r = b;
+
 	if (l->bytes_alloc < r->bytes_alloc)
 		return -1;
 	else if (l->bytes_alloc > r->bytes_alloc)
@@ -1188,9 +1206,11 @@ static struct sort_dimension bytes_sort_dimension = {
 	.cmp	= bytes_cmp,
 };
 
-static int frag_cmp(struct alloc_stat *l, struct alloc_stat *r)
+static int frag_cmp(void *a, void *b)
 {
 	double x, y;
+	struct alloc_stat *l = a;
+	struct alloc_stat *r = b;
 
 	x = fragmentation(l->bytes_req, l->bytes_alloc);
 	y = fragmentation(r->bytes_req, r->bytes_alloc);
@@ -1207,8 +1227,11 @@ static struct sort_dimension frag_sort_dimension = {
 	.cmp	= frag_cmp,
 };
 
-static int pingpong_cmp(struct alloc_stat *l, struct alloc_stat *r)
+static int pingpong_cmp(void *a, void *b)
 {
+	struct alloc_stat *l = a;
+	struct alloc_stat *r = b;
+
 	if (l->pingpong < r->pingpong)
 		return -1;
 	else if (l->pingpong > r->pingpong)
@@ -1221,7 +1244,135 @@ static struct sort_dimension pingpong_sort_dimension = {
 	.cmp	= pingpong_cmp,
 };
 
-static struct sort_dimension *avail_sorts[] = {
+/* page sort keys */
+static int page_cmp(void *a, void *b)
+{
+	struct page_stat *l = a;
+	struct page_stat *r = b;
+
+	if (l->page < r->page)
+		return -1;
+	else if (l->page > r->page)
+		return 1;
+	return 0;
+}
+
+static struct sort_dimension page_sort_dimension = {
+	.name	= "page",
+	.cmp	= page_cmp,
+};
+
+static int page_callsite_cmp(void *a, void *b)
+{
+	struct page_stat *l = a;
+	struct page_stat *r = b;
+
+	if (l->callsite < r->callsite)
+		return -1;
+	else if (l->callsite > r->callsite)
+		return 1;
+	return 0;
+}
+
+static struct sort_dimension page_callsite_sort_dimension = {
+	.name	= "callsite",
+	.cmp	= page_callsite_cmp,
+};
+
+static int page_hit_cmp(void *a, void *b)
+{
+	struct page_stat *l = a;
+	struct page_stat *r = b;
+
+	if (l->nr_alloc < r->nr_alloc)
+		return -1;
+	else if (l->nr_alloc > r->nr_alloc)
+		return 1;
+	return 0;
+}
+
+static struct sort_dimension page_hit_sort_dimension = {
+	.name	= "hit",
+	.cmp	= page_hit_cmp,
+};
+
+static int page_bytes_cmp(void *a, void *b)
+{
+	struct page_stat *l = a;
+	struct page_stat *r = b;
+
+	if (l->alloc_bytes < r->alloc_bytes)
+		return -1;
+	else if (l->alloc_bytes > r->alloc_bytes)
+		return 1;
+	return 0;
+}
+
+static struct sort_dimension page_bytes_sort_dimension = {
+	.name	= "bytes",
+	.cmp	= page_bytes_cmp,
+};
+
+static int page_order_cmp(void *a, void *b)
+{
+	struct page_stat *l = a;
+	struct page_stat *r = b;
+
+	if (l->order < r->order)
+		return -1;
+	else if (l->order > r->order)
+		return 1;
+	return 0;
+}
+
+static struct sort_dimension page_order_sort_dimension = {
+	.name	= "order",
+	.cmp	= page_order_cmp,
+};
+
+static int migrate_type_cmp(void *a, void *b)
+{
+	struct page_stat *l = a;
+	struct page_stat *r = b;
+
+	/* for internal use to find free'd page */
+	if (l->migrate_type == -1U)
+		return 0;
+
+	if (l->migrate_type < r->migrate_type)
+		return -1;
+	else if (l->migrate_type > r->migrate_type)
+		return 1;
+	return 0;
+}
+
+static struct sort_dimension migrate_type_sort_dimension = {
+	.name	= "migtype",
+	.cmp	= migrate_type_cmp,
+};
+
+static int gfp_flags_cmp(void *a, void *b)
+{
+	struct page_stat *l = a;
+	struct page_stat *r = b;
+
+	/* for internal use to find free'd page */
+	if (l->gfp_flags == -1U)
+		return 0;
+
+	if (l->gfp_flags < r->gfp_flags)
+		return -1;
+	else if (l->gfp_flags > r->gfp_flags)
+		return 1;
+	return 0;
+}
+
+static struct sort_dimension gfp_flags_sort_dimension = {
+	.name	= "gfp",
+	.cmp	= gfp_flags_cmp,
+};
+
+static struct sort_dimension *slab_sorts[] = {
 	&ptr_sort_dimension,
 	&callsite_sort_dimension,
 	&hit_sort_dimension,
@@ -1230,16 +1381,24 @@ static struct sort_dimension *avail_sorts[] = {
 	&pingpong_sort_dimension,
 };
 
-#define NUM_AVAIL_SORTS	((int)ARRAY_SIZE(avail_sorts))
+static struct sort_dimension *page_sorts[] = {
+	&page_sort_dimension,
+	&page_callsite_sort_dimension,
+	&page_hit_sort_dimension,
+	&page_bytes_sort_dimension,
+	&page_order_sort_dimension,
+	&migrate_type_sort_dimension,
+	&gfp_flags_sort_dimension,
+};
 
-static int sort_dimension__add(const char *tok, struct list_head *list)
+static int slab_sort_dimension__add(const char *tok, struct list_head *list)
 {
 	struct sort_dimension *sort;
 	int i;
 
-	for (i = 0; i < NUM_AVAIL_SORTS; i++) {
-		if (!strcmp(avail_sorts[i]->name, tok)) {
-			sort = memdup(avail_sorts[i], sizeof(*avail_sorts[i]));
+	for (i = 0; i < (int)ARRAY_SIZE(slab_sorts); i++) {
+		if (!strcmp(slab_sorts[i]->name, tok)) {
+			sort = memdup(slab_sorts[i], sizeof(*slab_sorts[i]));
 			if (!sort) {
 				pr_err("%s: memdup failed\n", __func__);
 				return -1;
@@ -1252,7 +1411,27 @@ static int sort_dimension__add(const char *tok, struct list_head *list)
 	return -1;
 }
 
-static int setup_sorting(struct list_head *sort_list, const char *arg)
+static int page_sort_dimension__add(const char *tok, struct list_head *list)
+{
+	struct sort_dimension *sort;
+	int i;
+
+	for (i = 0; i < (int)ARRAY_SIZE(page_sorts); i++) {
+		if (!strcmp(page_sorts[i]->name, tok)) {
+			sort = memdup(page_sorts[i], sizeof(*page_sorts[i]));
+			if (!sort) {
+				pr_err("%s: memdup failed\n", __func__);
+				return -1;
+			}
+			list_add_tail(&sort->list, list);
+			return 0;
+		}
+	}
+
+	return -1;
+}
+
+static int setup_slab_sorting(struct list_head *sort_list, const char *arg)
 {
 	char *tok;
 	char *str = strdup(arg);
@@ -1267,8 +1446,34 @@ static int setup_sorting(struct list_head *sort_list, const char *arg)
 		tok = strsep(&pos, ",");
 		if (!tok)
 			break;
-		if (sort_dimension__add(tok, sort_list) < 0) {
-			error("Unknown --sort key: '%s'", tok);
+		if (slab_sort_dimension__add(tok, sort_list) < 0) {
+			error("Unknown slab --sort key: '%s'", tok);
+			free(str);
+			return -1;
+		}
+	}
+
+	free(str);
+	return 0;
+}
+
+static int setup_page_sorting(struct list_head *sort_list, const char *arg)
+{
+	char *tok;
+	char *str = strdup(arg);
+	char *pos = str;
+
+	if (!str) {
+		pr_err("%s: strdup failed\n", __func__);
+		return -1;
+	}
+
+	while (true) {
+		tok = strsep(&pos, ",");
+		if (!tok)
+			break;
+		if (page_sort_dimension__add(tok, sort_list) < 0) {
+			error("Unknown page --sort key: '%s'", tok);
 			free(str);
 			return -1;
 		}
@@ -1284,10 +1489,17 @@ static int parse_sort_opt(const struct option *opt __maybe_unused,
 	if (!arg)
 		return -1;
 
-	if (caller_flag > alloc_flag)
-		return setup_sorting(&caller_sort, arg);
-	else
-		return setup_sorting(&alloc_sort, arg);
+	if (kmem_page > kmem_slab) {
+		if (caller_flag > alloc_flag)
+			return setup_page_sorting(&page_caller_sort, arg);
+		else
+			return setup_page_sorting(&page_alloc_sort, arg);
+	} else {
+		if (caller_flag > alloc_flag)
+			return setup_slab_sorting(&slab_caller_sort, arg);
+		else
+			return setup_slab_sorting(&slab_alloc_sort, arg);
+	}
 
 	return 0;
 }
@@ -1395,7 +1607,8 @@ static int __cmd_record(int argc, const char **argv)
 
 int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
 {
-	const char * const default_sort_order = "frag,hit,bytes";
+	const char * const default_slab_sort = "frag,hit,bytes";
+	const char * const default_page_sort = "bytes,hit";
 	struct perf_data_file file = {
 		.mode = PERF_DATA_MODE_READ,
 	};
@@ -1408,8 +1621,8 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_CALLBACK_NOOPT(0, "alloc", NULL, NULL,
 			   "show per-allocation statistics", parse_alloc_opt),
 	OPT_CALLBACK('s', "sort", NULL, "key[,key2...]",
-		     "sort by keys: ptr, call_site, bytes, hit, pingpong, frag",
-		     parse_sort_opt),
+		     "sort by keys: ptr, callsite, bytes, hit, pingpong, frag, "
+		     "page, order, migtype, gfp", parse_sort_opt),
 	OPT_CALLBACK('l', "line", NULL, "num", "show n lines", parse_line_opt),
 	OPT_BOOLEAN(0, "raw-ip", &raw_ip, "show raw ip instead of symbol"),
 	OPT_BOOLEAN('f', "force", &file.force, "don't complain, do it"),
@@ -1467,11 +1680,21 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
 		if (cpu__setup_cpunode_map())
 			goto out_delete;
 
-		if (list_empty(&caller_sort))
-			setup_sorting(&caller_sort, default_sort_order);
-		if (list_empty(&alloc_sort))
-			setup_sorting(&alloc_sort, default_sort_order);
-
+		if (list_empty(&slab_caller_sort))
+			setup_slab_sorting(&slab_caller_sort, default_slab_sort);
+		if (list_empty(&slab_alloc_sort))
+			setup_slab_sorting(&slab_alloc_sort, default_slab_sort);
+		if (list_empty(&page_caller_sort))
+			setup_page_sorting(&page_caller_sort, default_page_sort);
+		if (list_empty(&page_alloc_sort))
+			setup_page_sorting(&page_alloc_sort, default_page_sort);
+
+		if (kmem_page) {
+			setup_page_sorting(&page_alloc_sort_input,
+					   "page,order,migtype,gfp");
+			setup_page_sorting(&page_caller_sort_input,
+					   "callsite,order,migtype,gfp");
+		}
 		ret = __cmd_kmem(session);
 	} else
 		usage_with_options(kmem_usage, kmem_options);
-- 
2.1.0

[PATCH 19/21] perf kmem: Add --live option for current allocation stat

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

Currently 'perf kmem stat --page' shows total (page) allocation stat by
default, but sometimes one might want to see live (total alloc-only)
requests/pages only.  The new --live option does this by subtracting freed
allocation from the stat.

E.g.:

 # perf kmem stat --page

 SUMMARY (page allocator)
 ========================
 Total allocation requests     :          988,858   [        4,045,368 KB ]
 Total free requests           :          886,484   [        3,624,996 KB ]

 Total alloc+freed requests    :          885,969   [        3,622,628 KB ]
 Total alloc-only requests     :          102,889   [          422,740 KB ]
 Total free-only requests      :              515   [            2,368 KB ]

 Total allocation failures     :                0   [                0 KB ]

 Order     Unmovable   Reclaimable       Movable      Reserved  CMA/Isolated
 -----  ------------  ------------  ------------  ------------  ------------
     0       172,173         3,083       806,686             .             .
     1           284             .             .             .             .
     2         6,124            58             .             .             .
     3           114           335             .             .             .
     4             .             .             .             .             .
     5             .             .             .             .             .
     6             .             .             .             .             .
     7             .             .             .             .             .
     8             .             .             .             .             .
     9             .             .             1             .             .
    10             .             .             .             .             .
 # perf kmem stat --page --live

 SUMMARY (page allocator)
 ========================
 Total allocation requests     :          988,858   [        4,045,368 KB ]
 Total free requests           :          886,484   [        3,624,996 KB ]

 Total alloc+freed requests    :          885,969   [        3,622,628 KB ]
 Total alloc-only requests     :          102,889   [          422,740 KB ]
 Total free-only requests      :              515   [            2,368 KB ]

 Total allocation failures     :                0   [                0 KB ]

 Order     Unmovable   Reclaimable       Movable      Reserved  CMA/Isolated
 -----  ------------  ------------  ------------  ------------  ------------
     0         2,214         3,025        97,156             .             .
     1            59             .             .             .             .
     2            19            58             .             .             .
     3            23           335             .             .             .
     4             .             .             .             .             .
     5             .             .             .             .             .
     6             .             .             .             .             .
     7             .             .             .             .             .
     8             .             .             .             .             .
     9             .             .             .             .             .
    10             .             .             .             .             .
 #

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Joonsoo Kim <js1304@gmail.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/1429592107-1807-4-git-send-email-namhyung@kernel.org
[ Added examples to the changeset log ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/Documentation/perf-kmem.txt |   5 ++
 tools/perf/builtin-kmem.c              | 110 ++++++++++++++++++++-------------
 2 files changed, 73 insertions(+), 42 deletions(-)

diff --git a/tools/perf/Documentation/perf-kmem.txt b/tools/perf/Documentation/perf-kmem.txt
index 69e1812..ff0f433 100644
--- a/tools/perf/Documentation/perf-kmem.txt
+++ b/tools/perf/Documentation/perf-kmem.txt
@@ -56,6 +56,11 @@ OPTIONS
 --page::
 	Analyze page allocator events
 
+--live::
+	Show live page stat.  The perf kmem shows total allocation stat by
+	default, but this option shows live (currently allocated) pages
+	instead.  (This option works with --page option only)
+
 SEE ALSO
 --------
 linkperf:perf-record[1]
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 0393a7f..7ead942 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -244,6 +244,7 @@ static unsigned long nr_page_fails;
 static unsigned long nr_page_nomatch;
 
 static bool use_pfn;
+static bool live_page;
 static struct perf_session *kmem_session;
 
 #define MAX_MIGRATE_TYPES  6
@@ -264,7 +265,7 @@ struct page_stat {
 	int 		nr_free;
 };
 
-static struct rb_root page_tree;
+static struct rb_root page_live_tree;
 static struct rb_root page_alloc_tree;
 static struct rb_root page_alloc_sorted;
 static struct rb_root page_caller_tree;
@@ -403,10 +404,19 @@ out:
 	return sample->ip;
 }
 
+struct sort_dimension {
+	const char		name[20];
+	sort_fn_t		cmp;
+	struct list_head	list;
+};
+
+static LIST_HEAD(page_alloc_sort_input);
+static LIST_HEAD(page_caller_sort_input);
+
 static struct page_stat *
-__page_stat__findnew_page(u64 page, bool create)
+__page_stat__findnew_page(struct page_stat *pstat, bool create)
 {
-	struct rb_node **node = &page_tree.rb_node;
+	struct rb_node **node = &page_live_tree.rb_node;
 	struct rb_node *parent = NULL;
 	struct page_stat *data;
 
@@ -416,7 +426,7 @@ __page_stat__findnew_page(u64 page, bool create)
 		parent = *node;
 		data = rb_entry(*node, struct page_stat, node);
 
-		cmp = data->page - page;
+		cmp = data->page - pstat->page;
 		if (cmp < 0)
 			node = &parent->rb_left;
 		else if (cmp > 0)
@@ -430,34 +440,28 @@ __page_stat__findnew_page(u64 page, bool create)
 
 	data = zalloc(sizeof(*data));
 	if (data != NULL) {
-		data->page = page;
+		data->page = pstat->page;
+		data->order = pstat->order;
+		data->gfp_flags = pstat->gfp_flags;
+		data->migrate_type = pstat->migrate_type;
 
 		rb_link_node(&data->node, parent, node);
-		rb_insert_color(&data->node, &page_tree);
+		rb_insert_color(&data->node, &page_live_tree);
 	}
 
 	return data;
 }
 
-static struct page_stat *page_stat__find_page(u64 page)
+static struct page_stat *page_stat__find_page(struct page_stat *pstat)
 {
-	return __page_stat__findnew_page(page, false);
+	return __page_stat__findnew_page(pstat, false);
 }
 
-static struct page_stat *page_stat__findnew_page(u64 page)
+static struct page_stat *page_stat__findnew_page(struct page_stat *pstat)
 {
-	return __page_stat__findnew_page(page, true);
+	return __page_stat__findnew_page(pstat, true);
 }
 
-struct sort_dimension {
-	const char		name[20];
-	sort_fn_t		cmp;
-	struct list_head	list;
-};
-
-static LIST_HEAD(page_alloc_sort_input);
-static LIST_HEAD(page_caller_sort_input);
-
 static struct page_stat *
 __page_stat__findnew_alloc(struct page_stat *pstat, bool create)
 {
@@ -615,17 +619,8 @@ static int perf_evsel__process_page_alloc_event(struct perf_evsel *evsel,
 	 * This is to find the current page (with correct gfp flags and
 	 * migrate type) at free event.
 	 */
-	pstat = page_stat__findnew_page(page);
-	if (pstat == NULL)
-		return -ENOMEM;
-
-	pstat->order = order;
-	pstat->gfp_flags = gfp_flags;
-	pstat->migrate_type = migrate_type;
-	pstat->callsite = callsite;
-
 	this.page = page;
-	pstat = page_stat__findnew_alloc(&this);
+	pstat = page_stat__findnew_page(&this);
 	if (pstat == NULL)
 		return -ENOMEM;
 
@@ -633,6 +628,16 @@ static int perf_evsel__process_page_alloc_event(struct perf_evsel *evsel,
 	pstat->alloc_bytes += bytes;
 	pstat->callsite = callsite;
 
+	if (!live_page) {
+		pstat = page_stat__findnew_alloc(&this);
+		if (pstat == NULL)
+			return -ENOMEM;
+
+		pstat->nr_alloc++;
+		pstat->alloc_bytes += bytes;
+		pstat->callsite = callsite;
+	}
+
 	this.callsite = callsite;
 	pstat = page_stat__findnew_caller(&this);
 	if (pstat == NULL)
@@ -665,7 +670,8 @@ static int perf_evsel__process_page_free_event(struct perf_evsel *evsel,
 	nr_page_frees++;
 	total_page_free_bytes += bytes;
 
-	pstat = page_stat__find_page(page);
+	this.page = page;
+	pstat = page_stat__find_page(&this);
 	if (pstat == NULL) {
 		pr_debug2("missing free at page %"PRIx64" (order: %d)\n",
 			  page, order);
@@ -676,20 +682,23 @@ static int perf_evsel__process_page_free_event(struct perf_evsel *evsel,
 		return 0;
 	}
 
-	this.page = page;
 	this.gfp_flags = pstat->gfp_flags;
 	this.migrate_type = pstat->migrate_type;
 	this.callsite = pstat->callsite;
 
-	rb_erase(&pstat->node, &page_tree);
+	rb_erase(&pstat->node, &page_live_tree);
 	free(pstat);
 
-	pstat = page_stat__find_alloc(&this);
-	if (pstat == NULL)
-		return -ENOENT;
+	if (live_page) {
+		order_stats[this.order][this.migrate_type]--;
+	} else {
+		pstat = page_stat__find_alloc(&this);
+		if (pstat == NULL)
+			return -ENOMEM;
 
-	pstat->nr_free++;
-	pstat->free_bytes += bytes;
+		pstat->nr_free++;
+		pstat->free_bytes += bytes;
+	}
 
 	pstat = page_stat__find_caller(&this);
 	if (pstat == NULL)
@@ -698,6 +707,16 @@ static int perf_evsel__process_page_free_event(struct perf_evsel *evsel,
 	pstat->nr_free++;
 	pstat->free_bytes += bytes;
 
+	if (live_page) {
+		pstat->nr_alloc--;
+		pstat->alloc_bytes -= bytes;
+
+		if (pstat->nr_alloc == 0) {
+			rb_erase(&pstat->node, &page_caller_tree);
+			free(pstat);
+		}
+	}
+
 	return 0;
 }
 
@@ -815,8 +834,8 @@ static void __print_page_alloc_result(struct perf_session *session, int n_lines)
 	const char *format;
 
 	printf("\n%.105s\n", graph_dotted_line);
-	printf(" %-16s | Total alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite\n",
-	       use_pfn ? "PFN" : "Page");
+	printf(" %-16s | %5s alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite\n",
+	       use_pfn ? "PFN" : "Page", live_page ? "Live" : "Total");
 	printf("%.105s\n", graph_dotted_line);
 
 	if (use_pfn)
@@ -860,7 +879,8 @@ static void __print_page_caller_result(struct perf_session *session, int n_lines
 	struct machine *machine = &session->machines.host;
 
 	printf("\n%.105s\n", graph_dotted_line);
-	printf(" Total alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite\n");
+	printf(" %5s alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite\n",
+	       live_page ? "Live" : "Total");
 	printf("%.105s\n", graph_dotted_line);
 
 	while (next && n_lines--) {
@@ -1085,8 +1105,13 @@ static void sort_result(void)
 				   &slab_caller_sort);
 	}
 	if (kmem_page) {
-		__sort_page_result(&page_alloc_tree, &page_alloc_sorted,
-				   &page_alloc_sort);
+		if (live_page)
+			__sort_page_result(&page_live_tree, &page_alloc_sorted,
+					   &page_alloc_sort);
+		else
+			__sort_page_result(&page_alloc_tree, &page_alloc_sorted,
+					   &page_alloc_sort);
+
 		__sort_page_result(&page_caller_tree, &page_caller_sorted,
 				   &page_caller_sort);
 	}
@@ -1630,6 +1655,7 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
 			   parse_slab_opt),
 	OPT_CALLBACK_NOOPT(0, "page", NULL, NULL, "Analyze page allocator",
 			   parse_page_opt),
+	OPT_BOOLEAN(0, "live", &live_page, "Show live page stat"),
 	OPT_END()
 	};
 	const char *const kmem_subcommands[] = { "record", "stat", NULL };
-- 
2.1.0

[PATCH 20/21] perf kmem: Print gfp flags in human readable string

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

Save libtraceevent output and print it in the header.

  # perf kmem stat --page --caller
  #
  # GFP flags
  # ---------
  # 00000010:       NI: GFP_NOIO
  # 000000d0:        K: GFP_KERNEL
  # 00000200:      NWR: GFP_NOWARN
  # 000084d0:    K|R|Z: GFP_KERNEL|GFP_REPEAT|GFP_ZERO
  # 000200d2:       HU: GFP_HIGHUSER
  # 000200da:      HUM: GFP_HIGHUSER_MOVABLE
  # 000280da:    HUM|Z: GFP_HIGHUSER_MOVABLE|GFP_ZERO
  # 002084d0: K|R|Z|NT: GFP_KERNEL|GFP_REPEAT|GFP_ZERO|GFP_NOTRACK
  # 0102005a:  NF|HW|M: GFP_NOFS|GFP_HARDWALL|GFP_MOVABLE

  ---------------------------------------------------------------------------------------------------------
   Total alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite
  ---------------------------------------------------------------------------------------------------------
                 60 |        15 |     0 | UNMOVABL | K|R|Z|NT  | pte_alloc_one
                 40 |        10 |     0 |  MOVABLE | HUM|Z     | handle_mm_fault
                 24 |         6 |     0 |  MOVABLE | HUM       | do_wp_page
                 24 |         6 |     0 | UNMOVABL | K         | __pollwait
   ...

Requested-by: Joonsoo Kim <js1304@gmail.com>
Suggested-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Joonsoo Kim <js1304@gmail.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/1429592107-1807-5-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-kmem.c | 222 +++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 209 insertions(+), 13 deletions(-)

diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 7ead942..1c66895 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -581,6 +581,176 @@ static bool valid_page(u64 pfn_or_page)
 	return true;
 }
 
+struct gfp_flag {
+	unsigned int flags;
+	char *compact_str;
+	char *human_readable;
+};
+
+static struct gfp_flag *gfps;
+static int nr_gfps;
+
+static int gfpcmp(const void *a, const void *b)
+{
+	const struct gfp_flag *fa = a;
+	const struct gfp_flag *fb = b;
+
+	return fa->flags - fb->flags;
+}
+
+/* see include/trace/events/gfpflags.h */
+static const struct {
+	const char *original;
+	const char *compact;
+} gfp_compact_table[] = {
+	{ "GFP_TRANSHUGE",		"THP" },
+	{ "GFP_HIGHUSER_MOVABLE",	"HUM" },
+	{ "GFP_HIGHUSER",		"HU" },
+	{ "GFP_USER",			"U" },
+	{ "GFP_TEMPORARY",		"TMP" },
+	{ "GFP_KERNEL",			"K" },
+	{ "GFP_NOFS",			"NF" },
+	{ "GFP_ATOMIC",			"A" },
+	{ "GFP_NOIO",			"NI" },
+	{ "GFP_HIGH",			"H" },
+	{ "GFP_WAIT",			"W" },
+	{ "GFP_IO",			"I" },
+	{ "GFP_COLD",			"CO" },
+	{ "GFP_NOWARN",			"NWR" },
+	{ "GFP_REPEAT",			"R" },
+	{ "GFP_NOFAIL",			"NF" },
+	{ "GFP_NORETRY",		"NR" },
+	{ "GFP_COMP",			"C" },
+	{ "GFP_ZERO",			"Z" },
+	{ "GFP_NOMEMALLOC",		"NMA" },
+	{ "GFP_MEMALLOC",		"MA" },
+	{ "GFP_HARDWALL",		"HW" },
+	{ "GFP_THISNODE",		"TN" },
+	{ "GFP_RECLAIMABLE",		"RC" },
+	{ "GFP_MOVABLE",		"M" },
+	{ "GFP_NOTRACK",		"NT" },
+	{ "GFP_NO_KSWAPD",		"NK" },
+	{ "GFP_OTHER_NODE",		"ON" },
+	{ "GFP_NOWAIT",			"NW" },
+};
+
+static size_t max_gfp_len;
+
+static char *compact_gfp_flags(char *gfp_flags)
+{
+	char *orig_flags = strdup(gfp_flags);
+	char *new_flags = NULL;
+	char *str, *pos;
+	size_t len = 0;
+
+	if (orig_flags == NULL)
+		return NULL;
+
+	str = strtok_r(orig_flags, "|", &pos);
+	while (str) {
+		size_t i;
+		char *new;
+		const char *cpt;
+
+		for (i = 0; i < ARRAY_SIZE(gfp_compact_table); i++) {
+			if (strcmp(gfp_compact_table[i].original, str))
+				continue;
+
+			cpt = gfp_compact_table[i].compact;
+			new = realloc(new_flags, len + strlen(cpt) + 2);
+			if (new == NULL) {
+				free(new_flags);
+				return NULL;
+			}
+
+			new_flags = new;
+
+			if (!len) {
+				strcpy(new_flags, cpt);
+			} else {
+				strcat(new_flags, "|");
+				strcat(new_flags, cpt);
+				len++;
+			}
+
+			len += strlen(cpt);
+		}
+
+		str = strtok_r(NULL, "|", &pos);
+	}
+
+	if (max_gfp_len < len)
+		max_gfp_len = len;
+
+	free(orig_flags);
+	return new_flags;
+}
+
+static char *compact_gfp_string(unsigned long gfp_flags)
+{
+	struct gfp_flag key = {
+		.flags = gfp_flags,
+	};
+	struct gfp_flag *gfp;
+
+	gfp = bsearch(&key, gfps, nr_gfps, sizeof(*gfps), gfpcmp);
+	if (gfp)
+		return gfp->compact_str;
+
+	return NULL;
+}
+
+static int parse_gfp_flags(struct perf_evsel *evsel, struct perf_sample *sample,
+			   unsigned int gfp_flags)
+{
+	struct pevent_record record = {
+		.cpu = sample->cpu,
+		.data = sample->raw_data,
+		.size = sample->raw_size,
+	};
+	struct trace_seq seq;
+	char *str, *pos;
+
+	if (nr_gfps) {
+		struct gfp_flag key = {
+			.flags = gfp_flags,
+		};
+
+		if (bsearch(&key, gfps, nr_gfps, sizeof(*gfps), gfpcmp))
+			return 0;
+	}
+
+	trace_seq_init(&seq);
+	pevent_event_info(&seq, evsel->tp_format, &record);
+
+	str = strtok_r(seq.buffer, " ", &pos);
+	while (str) {
+		if (!strncmp(str, "gfp_flags=", 10)) {
+			struct gfp_flag *new;
+
+			new = realloc(gfps, (nr_gfps + 1) * sizeof(*gfps));
+			if (new == NULL)
+				return -ENOMEM;
+
+			gfps = new;
+			new += nr_gfps++;
+
+			new->flags = gfp_flags;
+			new->human_readable = strdup(str + 10);
+			new->compact_str = compact_gfp_flags(str + 10);
+			if (!new->human_readable || !new->compact_str)
+				return -ENOMEM;
+
+			qsort(gfps, nr_gfps, sizeof(*gfps), gfpcmp);
+		}
+
+		str = strtok_r(NULL, " ", &pos);
+	}
+
+	trace_seq_destroy(&seq);
+	return 0;
+}
+
 static int perf_evsel__process_page_alloc_event(struct perf_evsel *evsel,
 						struct perf_sample *sample)
 {
@@ -613,6 +783,9 @@ static int perf_evsel__process_page_alloc_event(struct perf_evsel *evsel,
 		return 0;
 	}
 
+	if (parse_gfp_flags(evsel, sample, gfp_flags) < 0)
+		return -1;
+
 	callsite = find_callsite(evsel, sample);
 
 	/*
@@ -832,16 +1005,18 @@ static void __print_page_alloc_result(struct perf_session *session, int n_lines)
 	struct rb_node *next = rb_first(&page_alloc_sorted);
 	struct machine *machine = &session->machines.host;
 	const char *format;
+	int gfp_len = max(strlen("GFP flags"), max_gfp_len);
 
 	printf("\n%.105s\n", graph_dotted_line);
-	printf(" %-16s | %5s alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite\n",
-	       use_pfn ? "PFN" : "Page", live_page ? "Live" : "Total");
+	printf(" %-16s | %5s alloc (KB) | Hits      | Order | Mig.type | %-*s | Callsite\n",
+	       use_pfn ? "PFN" : "Page", live_page ? "Live" : "Total",
+	       gfp_len, "GFP flags");
 	printf("%.105s\n", graph_dotted_line);
 
 	if (use_pfn)
-		format = " %16llu | %'16llu | %'9d | %5d | %8s |  %08lx | %s\n";
+		format = " %16llu | %'16llu | %'9d | %5d | %8s | %-*s | %s\n";
 	else
-		format = " %016llx | %'16llu | %'9d | %5d | %8s |  %08lx | %s\n";
+		format = " %016llx | %'16llu | %'9d | %5d | %8s | %-*s | %s\n";
 
 	while (next && n_lines--) {
 		struct page_stat *data;
@@ -862,13 +1037,15 @@ static void __print_page_alloc_result(struct perf_session *session, int n_lines)
 		       (unsigned long long)data->alloc_bytes / 1024,
 		       data->nr_alloc, data->order,
 		       migrate_type_str[data->migrate_type],
-		       (unsigned long)data->gfp_flags, caller);
+		       gfp_len, compact_gfp_string(data->gfp_flags), caller);
 
 		next = rb_next(next);
 	}
 
-	if (n_lines == -1)
-		printf(" ...              | ...              | ...       | ...   | ...      | ...       | ...\n");
+	if (n_lines == -1) {
+		printf(" ...              | ...              | ...       | ...   | ...      | %-*s | ...\n",
+		       gfp_len, "...");
+	}
 
 	printf("%.105s\n", graph_dotted_line);
 }
@@ -877,10 +1054,11 @@ static void __print_page_caller_result(struct perf_session *session, int n_lines
 {
 	struct rb_node *next = rb_first(&page_caller_sorted);
 	struct machine *machine = &session->machines.host;
+	int gfp_len = max(strlen("GFP flags"), max_gfp_len);
 
 	printf("\n%.105s\n", graph_dotted_line);
-	printf(" %5s alloc (KB) | Hits      | Order | Mig.type | GFP flags | Callsite\n",
-	       live_page ? "Live" : "Total");
+	printf(" %5s alloc (KB) | Hits      | Order | Mig.type | %-*s | Callsite\n",
+	       live_page ? "Live" : "Total", gfp_len, "GFP flags");
 	printf("%.105s\n", graph_dotted_line);
 
 	while (next && n_lines--) {
@@ -898,21 +1076,37 @@ static void __print_page_caller_result(struct perf_session *session, int n_lines
 		else
 			scnprintf(buf, sizeof(buf), "%"PRIx64, data->callsite);
 
-		printf(" %'16llu | %'9d | %5d | %8s |  %08lx | %s\n",
+		printf(" %'16llu | %'9d | %5d | %8s | %-*s | %s\n",
 		       (unsigned long long)data->alloc_bytes / 1024,
 		       data->nr_alloc, data->order,
 		       migrate_type_str[data->migrate_type],
-		       (unsigned long)data->gfp_flags, caller);
+		       gfp_len, compact_gfp_string(data->gfp_flags), caller);
 
 		next = rb_next(next);
 	}
 
-	if (n_lines == -1)
-		printf(" ...              | ...       | ...   | ...      | ...       | ...\n");
+	if (n_lines == -1) {
+		printf(" ...              | ...       | ...   | ...      | %-*s | ...\n",
+		       gfp_len, "...");
+	}
 
 	printf("%.105s\n", graph_dotted_line);
 }
 
+static void print_gfp_flags(void)
+{
+	int i;
+
+	printf("#\n");
+	printf("# GFP flags\n");
+	printf("# ---------\n");
+	for (i = 0; i < nr_gfps; i++) {
+		printf("# %08x: %*s: %s\n", gfps[i].flags,
+		       (int) max_gfp_len, gfps[i].compact_str,
+		       gfps[i].human_readable);
+	}
+}
+
 static void print_slab_summary(void)
 {
 	printf("\nSUMMARY (SLAB allocator)");
@@ -982,6 +1176,8 @@ static void print_slab_result(struct perf_session *session)
 
 static void print_page_result(struct perf_session *session)
 {
+	if (caller_flag || alloc_flag)
+		print_gfp_flags();
 	if (caller_flag)
 		__print_page_caller_result(session, caller_lines);
 	if (alloc_flag)
-- 
2.1.0

[PATCH 21/21] perf kmem: Add kmem.default config option

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

Currently perf kmem command will select --slab if neither --slab nor
--page is given for backward compatibility.  Add kmem.default config
option to select the default value ('page' or 'slab').

  # cat ~/.perfconfig
  [kmem]
  	default = page

  # perf kmem stat

  SUMMARY (page allocator)
  ========================
  Total allocation requests     :            1,518   [            6,096 KB ]
  Total free requests           :            1,431   [            5,748 KB ]

  Total alloc+freed requests    :            1,330   [            5,344 KB ]
  Total alloc-only requests     :              188   [              752 KB ]
  Total free-only requests      :              101   [              404 KB ]

  Total allocation failures     :                0   [                0 KB ]
  ...

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Joonsoo Kim <js1304@gmail.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Taeung Song <treeze.taeung@gmail.com>
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/1429592107-1807-6-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-kmem.c | 32 +++++++++++++++++++++++++++++---
 1 file changed, 29 insertions(+), 3 deletions(-)

diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 1c66895..828b728 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -28,6 +28,10 @@ static int	kmem_slab;
 static int	kmem_page;
 
 static long	kmem_page_size;
+static enum {
+	KMEM_SLAB,
+	KMEM_PAGE,
+} kmem_default = KMEM_SLAB;  /* for backward compatibility */
 
 struct alloc_stat;
 typedef int (*sort_fn_t)(void *, void *);
@@ -1710,7 +1714,8 @@ static int parse_sort_opt(const struct option *opt __maybe_unused,
 	if (!arg)
 		return -1;
 
-	if (kmem_page > kmem_slab) {
+	if (kmem_page > kmem_slab ||
+	    (kmem_page == 0 && kmem_slab == 0 && kmem_default == KMEM_PAGE)) {
 		if (caller_flag > alloc_flag)
 			return setup_page_sorting(&page_caller_sort, arg);
 		else
@@ -1826,6 +1831,22 @@ static int __cmd_record(int argc, const char **argv)
 	return cmd_record(i, rec_argv, NULL);
 }
 
+static int kmem_config(const char *var, const char *value, void *cb)
+{
+	if (!strcmp(var, "kmem.default")) {
+		if (!strcmp(value, "slab"))
+			kmem_default = KMEM_SLAB;
+		else if (!strcmp(value, "page"))
+			kmem_default = KMEM_PAGE;
+		else
+			pr_err("invalid default value ('slab' or 'page' required): %s\n",
+			       value);
+		return 0;
+	}
+
+	return perf_default_config(var, value, cb);
+}
+
 int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
 {
 	const char * const default_slab_sort = "frag,hit,bytes";
@@ -1862,14 +1883,19 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
 	struct perf_session *session;
 	int ret = -1;
 
+	perf_config(kmem_config, NULL);
 	argc = parse_options_subcommand(argc, argv, kmem_options,
 					kmem_subcommands, kmem_usage, 0);
 
 	if (!argc)
 		usage_with_options(kmem_usage, kmem_options);
 
-	if (kmem_slab == 0 && kmem_page == 0)
-		kmem_slab = 1;  /* for backward compatibility */
+	if (kmem_slab == 0 && kmem_page == 0) {
+		if (kmem_default == KMEM_SLAB)
+			kmem_slab = 1;
+		else
+			kmem_page = 1;
+	}
 
 	if (!strncmp(argv[0], "rec", 3)) {
 		symbol__init(NULL);
-- 
2.1.0