perf updates/fixes

GIT PULL New Ingo Molnar 3 participants 6 messages Oct 12, 2012

Subsystems (3)

x86 Library code perf events

lore

Share & tools

lore

Ingo Molnar

author

2012-10-12 08:20 UTC #0 lore raw

Linus,

Please pull the latest perf-urgent-for-linus git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf-urgent-for-linus

HEAD: 95cf59ea72331d0093010543b8951bb43f262cac perf: Fix perf_cgroup_switch for sw-events

This tree includes some late late perf items that missed the first round:

tools:

* Bash auto completion improvements, now we can auto complete the tools long options, tracepoint event names, etc, from Namhyung Kim.

* Look up thread using tid instead of pid in 'perf sched'.

* Move global variables into a perf_kvm struct, from David Ahern.

* Hists refactorings, preparatory for improved 'diff' command, from Jiri Olsa.

* Hists refactorings, preparatory for event group viewieng work, from Namhyung Kim.

* Remove double negation on optional feature macro definitions, from Namhyung Kim.

* Remove several cases of needless global variables, on most builtins.

* misc fixes

kernel:

* sysfs support for IBS on AMD CPUs, from Robert Richter.

* Support for an upcoming Intel CPU, the Xeon-Phi / Knights Corner HPC blade PMU, from Vince Weaver.

* misc fixes

Thanks,

Ingo

------------------>
Arnaldo Carvalho de Melo (19):
perf trace: Use evsel->handler.func
perf inject: Remove unused 'input_name' static var perf inject: Remove static variables
perf sched: Look up thread using tid instead of pid perf stat: Don't use globals where not needed to perf script: Don't use globals where not needed to perf help: Don't use globals where not needed to perf kmem: Don't use globals where not needed to perf lock: Don't use globals where not needed to perf timechart: Don't use globals where not needed to perf buildid-cache: Don't use globals where not needed to perf buildid-list: Don't use globals where not needed to perf probe: Don't use globals where not needed to perf top: Don't use globals where not needed to perf evlist: Don't use globals where not needed to perf record: Don't use globals where not needed to perf inject: Don't use globals where not needed to perf evlist: Introduce add_newtp method
perf evlist: Remove some unused methods

David Ahern (1):
perf kvm: Move global variables into a perf_kvm struct

Jiri Olsa (6):
perf hists: Add struct hists pointer to struct hist_entry perf diff: Refactor diff displacement possition info perf hists: Separate overhead and baseline columns perf tools: Removing hists pair argument from output path perf tool: Add hpp interface to enable/disable hpp column perf diff: Removing the total_period argument from output code

Namhyung Kim (16):
perf tools: Move libdw availability check before arch Makefile perf tools: Remove unused PYRF_OBJS variable on Makefile
perf tools: Convert to LIBELF_SUPPORT
perf tools: Convert to LIBUNWIND_SUPPORT
perf tools: Convert to LIBAUDIT_SUPPORT
perf tools: Convert to NEWT_SUPPORT
perf tools: Convert to GTK2_SUPPORT
perf tools: Convert to HAVE_STRLCPY
perf tools: Check existence of _get_comp_words_by_ref when bash completing perf tools: Complete long option names of perf command
perf tools: Long option completion support for each subcommands perf tools: Convert to BACKTRACE_SUPPORT
perf tools: Complete tracepoint event names
perf hists: Introduce struct he_stat
perf hists: Move he->stat.nr_events initialization to a template perf hists: Add more helpers for hist entry stat

Peter Zijlstra (2):
perf: Clarify perf_cpu_context::active_pmu usage by renaming it to ::unique_pmu
perf: Fix perf_cgroup_switch for sw-events

Robert Richter (1):
perf/AMD/IBS: Add sysfs support

Vince Weaver (1):
perf/x86: Add support for Intel Xeon-Phi Knights Corner PMU

 arch/x86/include/asm/msr-index.h         |   5 +
 arch/x86/kernel/cpu/Makefile             |   2 +-
 arch/x86/kernel/cpu/perf_event.h         |   2 +
 arch/x86/kernel/cpu/perf_event_amd_ibs.c |  61 +++-
 arch/x86/kernel/cpu/perf_event_intel.c   |   2 +
 arch/x86/kernel/cpu/perf_event_knc.c     | 248 +++++++++++++++++
 arch/x86/kernel/cpu/perfctr-watchdog.c   |   4 +
 include/linux/perf_event.h               |   2 +-
 kernel/events/core.c                     |  21 +-
 tools/perf/Makefile                      |  83 ++----
 tools/perf/bash_completion               |  50 +++-
 tools/perf/builtin-buildid-cache.c       |  58 ++--
 tools/perf/builtin-buildid-list.c        |  55 ++--
 tools/perf/builtin-diff.c                |  68 +++--
 tools/perf/builtin-evlist.c              |  21 +-
 tools/perf/builtin-help.c                |  40 +--
 tools/perf/builtin-inject.c              |  88 +++---
 tools/perf/builtin-kmem.c                |  66 ++---
 tools/perf/builtin-kvm.c                 | 460 +++++++++++++++++--------------
 tools/perf/builtin-lock.c                |  90 +++---
 tools/perf/builtin-probe.c               |  26 +-
 tools/perf/builtin-record.c              |  27 +-
 tools/perf/builtin-report.c              |   4 +-
 tools/perf/builtin-sched.c               |   2 +-
 tools/perf/builtin-script.c              |  90 +++---
 tools/perf/builtin-stat.c                | 328 +++++++++++-----------
 tools/perf/builtin-timechart.c           | 100 +++----
 tools/perf/builtin-top.c                 |  11 +-
 tools/perf/builtin-trace.c               | 134 +++++----
 tools/perf/perf.c                        |   4 +-
 tools/perf/ui/browsers/hists.c           |  12 +-
 tools/perf/ui/gtk/browser.c              |   6 +-
 tools/perf/ui/gtk/util.c                 |   2 +-
 tools/perf/ui/helpline.h                 |  18 +-
 tools/perf/ui/hist.c                     | 145 ++++++----
 tools/perf/ui/setup.c                    |   2 +-
 tools/perf/ui/stdio/hist.c               |  45 ++-
 tools/perf/util/annotate.h               |   8 +-
 tools/perf/util/cache.h                  |  38 +--
 tools/perf/util/debug.c                  |   2 +-
 tools/perf/util/debug.h                  |  17 +-
 tools/perf/util/evlist.c                 |  88 +-----
 tools/perf/util/evlist.h                 |  18 +-
 tools/perf/util/generate-cmdlist.sh      |   4 +-
 tools/perf/util/hist.c                   |  66 +++--
 tools/perf/util/hist.h                   |  38 ++-
 tools/perf/util/map.c                    |   2 +-
 tools/perf/util/parse-options.c          |   8 +
 tools/perf/util/parse-options.h          |   1 +
 tools/perf/util/path.c                   |   2 +-
 tools/perf/util/perf_regs.h              |   4 +-
 tools/perf/util/sort.h                   |  19 +-
 tools/perf/util/symbol.h                 |  10 +-
 tools/perf/util/unwind.h                 |   4 +-
 tools/perf/util/util.c                   |   4 +-
 55 files changed, 1517 insertions(+), 1198 deletions(-)
 create mode 100644 arch/x86/kernel/cpu/perf_event_knc.c

arch/x86/include/asm/msr-index.h +5 -0

index 957ec87..07f96cb 100644
@@ -121,6 +121,11 @@
 #define MSR_P6_EVNTSEL0			0x00000186
 #define MSR_P6_EVNTSEL1			0x00000187
 
+#define MSR_KNC_PERFCTR0               0x00000020
+#define MSR_KNC_PERFCTR1               0x00000021
+#define MSR_KNC_EVNTSEL0               0x00000028
+#define MSR_KNC_EVNTSEL1               0x00000029
+
 /* AMD64 MSRs. Not complete. See the architecture manual for a more
    complete list. */

arch/x86/kernel/cpu/Makefile +1 -1

index d30a6a9..a0e067d 100644
@@ -32,7 +32,7 @@ obj-$(CONFIG_PERF_EVENTS)		+= perf_event.o
 
 ifdef CONFIG_PERF_EVENTS
 obj-$(CONFIG_CPU_SUP_AMD)		+= perf_event_amd.o
-obj-$(CONFIG_CPU_SUP_INTEL)		+= perf_event_p6.o perf_event_p4.o
+obj-$(CONFIG_CPU_SUP_INTEL)		+= perf_event_p6.o perf_event_knc.o perf_event_p4.o
 obj-$(CONFIG_CPU_SUP_INTEL)		+= perf_event_intel_lbr.o perf_event_intel_ds.o perf_event_intel.o
 obj-$(CONFIG_CPU_SUP_INTEL)		+= perf_event_intel_uncore.o
 endif

arch/x86/kernel/cpu/perf_event.h +2 -0

index 8b6defe..271d257 100644
@@ -626,6 +626,8 @@ int p4_pmu_init(void);
 
 int p6_pmu_init(void);
 
+int knc_pmu_init(void);
+
 #else /* CONFIG_CPU_SUP_INTEL */
 
 static inline void reserve_ds_buffers(void)

arch/x86/kernel/cpu/perf_event_amd_ibs.c +49 ~~-12~~

103 lines not on this page · raw message

arch/x86/kernel/cpu/perf_event_intel.c +2 -0

index 6bca492..324bb52 100644
@@ -1906,6 +1906,8 @@ __init int intel_pmu_init(void)
 		switch (boot_cpu_data.x86) {
 		case 0x6:
 			return p6_pmu_init();
+		case 0xb:
+			return knc_pmu_init();
 		case 0xf:
 			return p4_pmu_init();
 		}

arch/x86/kernel/cpu/perf_event_knc.c new file +248 -0

248 lines not on this page · raw message

arch/x86/kernel/cpu/perfctr-watchdog.c +4 -0

index 966512b..2e8caf0 100644
@@ -56,6 +56,8 @@ static inline unsigned int nmi_perfctr_msr_to_bit(unsigned int msr)
 		switch (boot_cpu_data.x86) {
 		case 6:
 			return msr - MSR_P6_PERFCTR0;
+		case 11:
+			return msr - MSR_KNC_PERFCTR0;
 		case 15:
 			return msr - MSR_P4_BPU_PERFCTR0;
 		}

@@ -82,6 +84,8 @@ static inline unsigned int nmi_evntsel_msr_to_bit(unsigned int msr)
 		switch (boot_cpu_data.x86) {
 		case 6:
 			return msr - MSR_P6_EVNTSEL0;
+		case 11:
+			return msr - MSR_KNC_EVNTSEL0;
 		case 15:
 			return msr - MSR_P4_BSU_ESCR0;
 		}

include/linux/perf_event.h +1 -1

index 599afc4..b4166cd 100644
@@ -1110,7 +1110,7 @@ struct perf_cpu_context {
 	int				exclusive;
 	struct list_head		rotation_list;
 	int				jiffies_interval;
-	struct pmu			*active_pmu;
+	struct pmu			*unique_pmu;
 	struct perf_cgroup		*cgrp;
 };

kernel/events/core.c +12 -9

index 7b9df35..fd15593 100644
@@ -372,6 +372,8 @@ void perf_cgroup_switch(struct task_struct *task, int mode)
 
 	list_for_each_entry_rcu(pmu, &pmus, entry) {
 		cpuctx = this_cpu_ptr(pmu->pmu_cpu_context);
+		if (cpuctx->unique_pmu != pmu)
+			continue; /* ensure we process each cpuctx once */
 
 		/*
 		 * perf_cgroup_events says at least one

@@ -395,9 +397,10 @@ void perf_cgroup_switch(struct task_struct *task, int mode)
 
 			if (mode & PERF_CGROUP_SWIN) {
 				WARN_ON_ONCE(cpuctx->cgrp);
-				/* set cgrp before ctxsw in to
-				 * allow event_filter_match() to not
-				 * have to pass task around
+				/*
+				 * set cgrp before ctxsw in to allow
+				 * event_filter_match() to not have to pass
+				 * task around
 				 */
 				cpuctx->cgrp = perf_cgroup_from_task(task);
 				cpu_ctx_sched_in(cpuctx, EVENT_ALL, task);

@@ -4419,7 +4422,7 @@ static void perf_event_task_event(struct perf_task_event *task_event)
 	rcu_read_lock();
 	list_for_each_entry_rcu(pmu, &pmus, entry) {
 		cpuctx = get_cpu_ptr(pmu->pmu_cpu_context);
-		if (cpuctx->active_pmu != pmu)
+		if (cpuctx->unique_pmu != pmu)
 			goto next;
 		perf_event_task_ctx(&cpuctx->ctx, task_event);

@@ -4565,7 +4568,7 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event)
 	rcu_read_lock();
 	list_for_each_entry_rcu(pmu, &pmus, entry) {
 		cpuctx = get_cpu_ptr(pmu->pmu_cpu_context);
-		if (cpuctx->active_pmu != pmu)
+		if (cpuctx->unique_pmu != pmu)
 			goto next;
 		perf_event_comm_ctx(&cpuctx->ctx, comm_event);

@@ -4761,7 +4764,7 @@ got_name:
 	rcu_read_lock();
 	list_for_each_entry_rcu(pmu, &pmus, entry) {
 		cpuctx = get_cpu_ptr(pmu->pmu_cpu_context);
-		if (cpuctx->active_pmu != pmu)
+		if (cpuctx->unique_pmu != pmu)
 			goto next;
 		perf_event_mmap_ctx(&cpuctx->ctx, mmap_event,
 					vma->vm_flags & VM_EXEC);

@@ -5862,8 +5865,8 @@ static void update_pmu_context(struct pmu *pmu, struct pmu *old_pmu)
 
 		cpuctx = per_cpu_ptr(pmu->pmu_cpu_context, cpu);
 
-		if (cpuctx->active_pmu == old_pmu)
-			cpuctx->active_pmu = pmu;
+		if (cpuctx->unique_pmu == old_pmu)
+			cpuctx->unique_pmu = pmu;
 	}
 }

@@ -5998,7 +6001,7 @@ skip_type:
 		cpuctx->ctx.pmu = pmu;
 		cpuctx->jiffies_interval = 1;
 		INIT_LIST_HEAD(&cpuctx->rotation_list);
-		cpuctx->active_pmu = pmu;
+		cpuctx->unique_pmu = pmu;
 	}
 
 got_cpu_context:

tools/perf/Makefile +29 ~~-54~~

179 lines not on this page · raw message

tools/perf/bash_completion +43 -7

66 lines not on this page · raw message

tools/perf/builtin-buildid-cache.c +26 ~~-32~~

83 lines not on this page · raw message

tools/perf/builtin-buildid-list.c +24 ~~-31~~

76 lines not on this page · raw message

tools/perf/builtin-diff.c +49 ~~-19~~

125 lines not on this page · raw message

tools/perf/builtin-evlist.c +9 ~~-12~~

index 1fb1641..997afb8 100644
@@ -108,23 +108,20 @@ static int __cmd_evlist(const char *input_name, struct perf_attr_details *detail
 	return 0;
 }
 
-static const char * const evlist_usage[] = {
-	"perf evlist [<options>]",
-	NULL
-};
-
 int cmd_evlist(int argc, const char **argv, const char *prefix __maybe_unused)
 {
 	struct perf_attr_details details = { .verbose = false, };
 	const char *input_name = NULL;
 	const struct option options[] = {
-		OPT_STRING('i', "input", &input_name, "file",
-			    "Input file name"),
-		OPT_BOOLEAN('F', "freq", &details.freq,
-			    "Show the sample frequency"),
-		OPT_BOOLEAN('v', "verbose", &details.verbose,
-			    "Show all event attr details"),
-		OPT_END()
+	OPT_STRING('i', "input", &input_name, "file", "Input file name"),
+	OPT_BOOLEAN('F', "freq", &details.freq, "Show the sample frequency"),
+	OPT_BOOLEAN('v', "verbose", &details.verbose,
+		    "Show all event attr details"),
+	OPT_END()
+	};
+	const char * const evlist_usage[] = {
+		"perf evlist [<options>]",
+		NULL
 	};
 
 	argc = parse_options(argc, argv, options, evlist_usage, 0);

tools/perf/builtin-help.c +20 ~~-20~~

index 25c8b94..411ee56 100644
@@ -30,23 +30,6 @@ enum help_format {
 	HELP_FORMAT_WEB,
 };
 
-static bool show_all = false;
-static enum help_format help_format = HELP_FORMAT_NONE;
-static struct option builtin_help_options[] = {
-	OPT_BOOLEAN('a', "all", &show_all, "print all available commands"),
-	OPT_SET_UINT('m', "man", &help_format, "show man page", HELP_FORMAT_MAN),
-	OPT_SET_UINT('w', "web", &help_format, "show manual in web browser",
-			HELP_FORMAT_WEB),
-	OPT_SET_UINT('i', "info", &help_format, "show info page",
-			HELP_FORMAT_INFO),
-	OPT_END(),
-};
-
-static const char * const builtin_help_usage[] = {
-	"perf help [--all] [--man|--web|--info] [command]",
-	NULL
-};
-
 static enum help_format parse_help_format(const char *format)
 {
 	if (!strcmp(format, "man"))

@@ -258,11 +241,13 @@ static int add_man_viewer_info(const char *var, const char *value)
 
 static int perf_help_config(const char *var, const char *value, void *cb)
 {
+	enum help_format *help_formatp = cb;
+
 	if (!strcmp(var, "help.format")) {
 		if (!value)
 			return config_error_nonbool(var);
-		help_format = parse_help_format(value);
-		if (help_format == HELP_FORMAT_NONE)
+		*help_formatp = parse_help_format(value);
+		if (*help_formatp == HELP_FORMAT_NONE)
 			return -1;
 		return 0;
 	}

@@ -428,12 +413,27 @@ static int show_html_page(const char *perf_cmd)
 
 int cmd_help(int argc, const char **argv, const char *prefix __maybe_unused)
 {
+	bool show_all = false;
+	enum help_format help_format = HELP_FORMAT_NONE;
+	struct option builtin_help_options[] = {
+	OPT_BOOLEAN('a', "all", &show_all, "print all available commands"),
+	OPT_SET_UINT('m', "man", &help_format, "show man page", HELP_FORMAT_MAN),
+	OPT_SET_UINT('w', "web", &help_format, "show manual in web browser",
+			HELP_FORMAT_WEB),
+	OPT_SET_UINT('i', "info", &help_format, "show info page",
+			HELP_FORMAT_INFO),
+	OPT_END(),
+	};
+	const char * const builtin_help_usage[] = {
+		"perf help [--all] [--man|--web|--info] [command]",
+		NULL
+	};
 	const char *alias;
 	int rc = 0;
 
 	load_command_list("perf-", &main_cmds, &other_cmds);
 
-	perf_config(perf_help_config, NULL);
+	perf_config(perf_help_config, &help_format);
 
 	argc = parse_options(argc, argv, builtin_help_options,
 			builtin_help_usage, 0);

tools/perf/builtin-inject.c +45 ~~-43~~

138 lines not on this page · raw message

tools/perf/builtin-kmem.c +26 ~~-40~~

114 lines not on this page · raw message

tools/perf/builtin-kvm.c +253 ~~-207~~

836 lines not on this page · raw message

tools/perf/builtin-lock.c +39 ~~-51~~

116 lines not on this page · raw message

tools/perf/builtin-probe.c +12 ~~-14~~

index 118aa89..de38a03 100644
@@ -250,19 +250,20 @@ static int opt_set_filter(const struct option *opt __maybe_unused,
 	return 0;
 }
 
-static const char * const probe_usage[] = {
-	"perf probe [<options>] 'PROBEDEF' ['PROBEDEF' ...]",
-	"perf probe [<options>] --add 'PROBEDEF' [--add 'PROBEDEF' ...]",
-	"perf probe [<options>] --del '[GROUP:]EVENT' ...",
-	"perf probe --list",
+int cmd_probe(int argc, const char **argv, const char *prefix __maybe_unused)
+{
+	const char * const probe_usage[] = {
+		"perf probe [<options>] 'PROBEDEF' ['PROBEDEF' ...]",
+		"perf probe [<options>] --add 'PROBEDEF' [--add 'PROBEDEF' ...]",
+		"perf probe [<options>] --del '[GROUP:]EVENT' ...",
+		"perf probe --list",
 #ifdef DWARF_SUPPORT
-	"perf probe [<options>] --line 'LINEDESC'",
-	"perf probe [<options>] --vars 'PROBEPOINT'",
+		"perf probe [<options>] --line 'LINEDESC'",
+		"perf probe [<options>] --vars 'PROBEPOINT'",
 #endif
-	NULL
+		NULL
 };
-
-static const struct option options[] = {
+	const struct option options[] = {
 	OPT_INCR('v', "verbose", &verbose,
 		    "be more verbose (show parsed arguments, etc)"),
 	OPT_BOOLEAN('l', "list", &params.list_events,

@@ -325,10 +326,7 @@ static const struct option options[] = {
 	OPT_CALLBACK('x', "exec", NULL, "executable|path",
 			"target executable name or path", opt_set_target),
 	OPT_END()
-};
-
-int cmd_probe(int argc, const char **argv, const char *prefix __maybe_unused)
-{
+	};
 	int ret;
 
 	argc = parse_options(argc, argv, options, probe_usage,

tools/perf/builtin-record.c +14 ~~-13~~

index f14cb5f..e9231659 100644
@@ -31,15 +31,6 @@
 #include <sched.h>
 #include <sys/mman.h>
 
-#define CALLCHAIN_HELP "do call-graph (stack chain/backtrace) recording: "
-
-#ifdef NO_LIBUNWIND_SUPPORT
-static char callchain_help[] = CALLCHAIN_HELP "[fp]";
-#else
-static unsigned long default_stack_dump_size = 8192;
-static char callchain_help[] = CALLCHAIN_HELP "[fp] dwarf";
-#endif
-
 enum write_mode_t {
 	WRITE_FORCE,
 	WRITE_APPEND

@@ -800,7 +791,7 @@ error:
 	return ret;
 }
 
-#ifndef NO_LIBUNWIND_SUPPORT
+#ifdef LIBUNWIND_SUPPORT
 static int get_stack_size(char *str, unsigned long *_size)
 {
 	char *endptr;

@@ -826,7 +817,7 @@ static int get_stack_size(char *str, unsigned long *_size)
 	       max_size, str);
 	return -1;
 }
-#endif /* !NO_LIBUNWIND_SUPPORT */
+#endif /* LIBUNWIND_SUPPORT */
 
 static int
 parse_callchain_opt(const struct option *opt __maybe_unused, const char *arg,

@@ -865,9 +856,11 @@ parse_callchain_opt(const struct option *opt __maybe_unused, const char *arg,
 				       "needed for -g fp\n");
 			break;
 
-#ifndef NO_LIBUNWIND_SUPPORT
+#ifdef LIBUNWIND_SUPPORT
 		/* Dwarf style */
 		} else if (!strncmp(name, "dwarf", sizeof("dwarf"))) {
+			const unsigned long default_stack_dump_size = 8192;
+
 			ret = 0;
 			rec->opts.call_graph = CALLCHAIN_DWARF;
 			rec->opts.stack_dump_size = default_stack_dump_size;

@@ -883,7 +876,7 @@ parse_callchain_opt(const struct option *opt __maybe_unused, const char *arg,
 			if (!ret)
 				pr_debug("callchain: stack dump size %d\n",
 					 rec->opts.stack_dump_size);
-#endif /* !NO_LIBUNWIND_SUPPORT */
+#endif /* LIBUNWIND_SUPPORT */
 		} else {
 			pr_err("callchain: Unknown -g option "
 			       "value: %s\n", arg);

@@ -930,6 +923,14 @@ static struct perf_record record = {
 	.file_new   = true,
 };
 
+#define CALLCHAIN_HELP "do call-graph (stack chain/backtrace) recording: "
+
+#ifdef LIBUNWIND_SUPPORT
+static const char callchain_help[] = CALLCHAIN_HELP "[fp] dwarf";
+#else
+static const char callchain_help[] = CALLCHAIN_HELP "[fp]";
+#endif
+
 /*
  * XXX Will stay a global variable till we fix builtin-script.c to stop messing
  * with it and switch to use the library functions in perf_evlist that came

tools/perf/builtin-report.c +2 -2

index 1da243d..a61725d 100644
@@ -320,7 +320,7 @@ static int perf_evlist__tty_browse_hists(struct perf_evlist *evlist,
 		const char *evname = perf_evsel__name(pos);
 
 		hists__fprintf_nr_sample_events(hists, evname, stdout);
-		hists__fprintf(hists, NULL, false, true, 0, 0, stdout);
+		hists__fprintf(hists, true, 0, 0, stdout);
 		fprintf(stdout, "\n\n");
 	}

@@ -691,7 +691,7 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 		setup_browser(true);
 	else {
 		use_browser = 0;
-		perf_hpp__init(false, false);
+		perf_hpp__init();
 	}
 
 	setup_sorting(report_usage, options);

tools/perf/builtin-sched.c +1 -1

index 9b9e32e..3488ead 100644
@@ -1426,7 +1426,7 @@ static int perf_sched__process_tracepoint_sample(struct perf_tool *tool __maybe_
 						 struct perf_evsel *evsel,
 						 struct machine *machine)
 {
-	struct thread *thread = machine__findnew_thread(machine, sample->pid);
+	struct thread *thread = machine__findnew_thread(machine, sample->tid);
 	int err = 0;
 
 	if (thread == NULL) {

tools/perf/builtin-script.c +42 ~~-48~~

132 lines not on this page · raw message

tools/perf/builtin-stat.c +159 ~~-169~~

377 lines not on this page · raw message

tools/perf/builtin-timechart.c +41 ~~-59~~

57 lines not on this page · raw message

tools/perf/builtin-top.c +5 -6

index e434a16..ff6db80 100644

Lines of this file are not shown. View the raw message.

tools/perf/builtin-trace.c +82 ~~-52~~

134 lines not on this page · raw message

tools/perf/perf.c +2 -2

index fc2f770..6d50eb0 100644

Lines of this file are not shown. View the raw message.

tools/perf/ui/browsers/hists.c +6 -6

index a21f40b..0568536 100644

Lines of this file are not shown. View the raw message.

tools/perf/ui/gtk/browser.c +3 -3

index 7ff99ec..4125c62 100644

Lines of this file are not shown. View the raw message.

tools/perf/ui/gtk/util.c +1 -1

index 8aada5b..ccb046a 100644

Lines of this file are not shown. View the raw message.

tools/perf/ui/helpline.h +9 -9

index 2b667ee..baa28a4 100644

Lines of this file are not shown. View the raw message.

tools/perf/ui/hist.c +91 ~~-54~~

145 lines not on this page · raw message

tools/perf/ui/setup.c +1 -1

index bd7d460..ebb4cc1 100644

Lines of this file are not shown. View the raw message.

tools/perf/ui/stdio/hist.c +16 ~~-29~~

45 lines not on this page · raw message

tools/perf/util/annotate.h +4 -4

index 9b5b21e..39242dc 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/cache.h +20 ~~-18~~

index ab17694..2bd5137 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/debug.c +1 -1

index 66eb382..03f830b 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/debug.h +9 -8

index bb2e7d1..dec9875 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/evlist.c +10 ~~-78~~

88 lines not on this page · raw message

tools/perf/util/evlist.h +3 ~~-15~~

index 3f1fb66..56003f7 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/generate-cmdlist.sh +2 -2

index 389590c..3ac3803 100755

Lines of this file are not shown. View the raw message.

tools/perf/util/hist.c +42 ~~-24~~

66 lines not on this page · raw message

tools/perf/util/hist.h +18 ~~-20~~

index f011ad4..66cb31f 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/map.c +1 -1

index ead5316..6109fa4 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/parse-options.c +8 -0

index 443fc11..2bc9e70 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/parse-options.h +1 -0

index abc31a1..7bb5999 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/path.c +1 -1

index bd74977..a8c4954 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/perf_regs.h +2 -2

index 316dbe7..5a4f2b6f 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/sort.h +12 -7

index 12d6347..5786f32 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/symbol.h +5 -5

index b441b07..8b6ef7f 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/unwind.h +2 -2

index a78c8b3..cb6bc50 100644

Lines of this file are not shown. View the raw message.

tools/perf/util/util.c +2 -2

index 2055cf3..9966459 100644

Lines of this file are not shown. View the raw message.

The rest of this diff is not shown. View the raw message.

Part of the diff is not shown here. View the raw message.

Ingo Molnar

author

2012-10-12 08:39 UTC #1 lore raw

Quoted 7 lines · Ingo Molnar

Linus,

Please pull the latest perf-urgent-for-linus git tree from:

   git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf-urgent-for-linus

   HEAD: 95cf59ea72331d0093010543b8951bb43f262cac perf: Fix perf_cgroup_switch for sw-events

Note that if you merge it then there's a new semantic conflict with recent rbtree.c changes in your tree, causing a tools/perf/ build failure:

../../lib/rbtree.c:24:36: fatal error: linux/rbtree_augmented.h: No such file or directory compilation terminated.

See the fix below.

( If the __maybe_unused annotations are too ugly for
lib/rbtree.c then we'll fix that in tools/perf in a cleaner way, weakening the compiler checks for the rbtree build. We are using stronger compiler checks in tools/perf/, which has served us very well so far and is a big net win - the price is the occasional extra annotation of dummy inline function parameters. )

Thanks,

Ingo

Signed-off-by Ingo Molnar <[email protected]>

---
 lib/rbtree.c                                     | 7 ++++---
 tools/perf/util/include/linux/rbtree_augmented.h | 2 ++
 2 files changed, 6 insertions(+), 3 deletions(-)

lib/rbtree.c +4 -3

index 4f56a11..993ffcb 100644
@@ -22,6 +22,7 @@
 */
 
 #include <linux/rbtree_augmented.h>
+#include <linux/compiler.h>
 #include <linux/export.h>
 
 /*

@@ -364,9 +365,9 @@ EXPORT_SYMBOL(__rb_erase_color);
  * out of the rb_insert_color() and rb_erase() function definitions.
  */
 
-static inline void dummy_propagate(struct rb_node *node, struct rb_node *stop) {}
-static inline void dummy_copy(struct rb_node *old, struct rb_node *new) {}
-static inline void dummy_rotate(struct rb_node *old, struct rb_node *new) {}
+static inline void dummy_propagate(struct rb_node *node __maybe_unused, struct rb_node *stop __maybe_unused) {}
+static inline void dummy_copy(struct rb_node *old __maybe_unused, struct rb_node *new __maybe_unused) {}
+static inline void dummy_rotate(struct rb_node *old __maybe_unused, struct rb_node *new __maybe_unused) {}
 
 static const struct rb_augment_callbacks dummy_callbacks = {
 	dummy_propagate, dummy_copy, dummy_rotate

tools/perf/util/include/linux/rbtree_augmented.h new file +2 -0

new file mode 100644
index 0000000..9d6fcdf
@@ -0,0 +1,2 @@
+#include <stdbool.h>
+#include "../../../../include/linux/rbtree_augmented.h"

Borislav Petkov

maintainer

2012-10-12 09:07 UTC #2 lore raw

Quoted 19 lines · Ingo Molnar

Quoted 7 lines · Ingo Molnar
Linus,
Please pull the latest perf-urgent-for-linus git tree from:
   git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf-urgent-for-linus

   HEAD: 95cf59ea72331d0093010543b8951bb43f262cac perf: Fix perf_cgroup_switch for sw-events
Note that if you merge it then there's a new semantic conflict with recent rbtree.c changes in your tree, causing a tools/perf/ build failure:
../../lib/rbtree.c:24:36: fatal error: linux/rbtree_augmented.h: No such file or directory compilation terminated.
See the fix below.

Quoted 7 lines

( If the __maybe_unused annotations are too ugly for
lib/rbtree.c then we'll fix that in tools/perf in a cleaner way, weakening the compiler checks for the rbtree build. We are using stronger compiler checks in tools/perf/, which has served us very well so far and is a big net win - the price is the occasional extra annotation of dummy inline function parameters. )

Btw,

Markus fixed it that way recently:

http://marc.info/?l=linux-kernel&m=134980573527738

by adding the -Wno-unused-parameter switch only to lib/rbtree.o when built from within perf.

I see now though that his patch has some unrelated changes to trace-event-perl.c's permissions which shouldn't be there.

Signature · 2 lines

Regards/Gruss,
Boris.

Ingo Molnar

author Acked-by

2012-10-12 09:11 UTC #3 lore raw

Quoted 32 lines · Borislav Petkov

Quoted 19 lines · Ingo Molnar
Quoted 7 lines · Ingo Molnar
Linus,
Please pull the latest perf-urgent-for-linus git tree from:
   git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf-urgent-for-linus

   HEAD: 95cf59ea72331d0093010543b8951bb43f262cac perf: Fix perf_cgroup_switch for sw-events
Note that if you merge it then there's a new semantic conflict with recent rbtree.c changes in your tree, causing a tools/perf/ build failure:
../../lib/rbtree.c:24:36: fatal error: linux/rbtree_augmented.h: No such file or directory compilation terminated.
See the fix below.
Quoted 7 lines
( If the __maybe_unused annotations are too ugly for
lib/rbtree.c then we'll fix that in tools/perf in a cleaner way, weakening the compiler checks for the rbtree build. We are using stronger compiler checks in tools/perf/, which has served us very well so far and is a big net win - the price is the occasional extra annotation of dummy inline function parameters. )
Btw,
Markus fixed it that way recently:

Quoted 7 lines

http://marc.info/?l=linux-kernel&m=134980573527738
by adding the -Wno-unused-parameter switch only to lib/rbtree.o when built from within perf.
I see now though that his patch has some unrelated changes to trace-event-perl.c's permissions which shouldn't be there.

Yes, that final hunk should be removed. Other than that, Markus's solution is fine to me too:

Acked-by Ingo Molnar <[email protected]>

Thanks,

Ingo

Ingo Molnar

author

2012-10-12 09:23 UTC #4 lore raw

PATCH 1/1 perf: Handle new rbtree implementation

Quoted 1 line

From 6c4d077c467d689b5dd2f89cc69713f1bf4e31d7 Mon Sep 17 00:00:00 2001

From: Markus Trippelsdorf <[email protected]>
Date: Tue, 9 Oct 2012 20:01:56 +0200
Subject: [PATCH] perf: Handle new rbtree implementation

Perf build fails with the new rbtree implementation:

../../lib/rbtree.c:24:36: fatal error: linux/rbtree_augmented.h: No such file or directory compilation terminated.

Fix by updating the Makefile and adding a btree_augmented.h wrapper.

Reported-and-tested-by Borislav Petkov <[email protected]>

Signed-off-by Markus Trippelsdorf <[email protected]>

Cc Arnaldo Carvalho de Melo <[email protected]>

Cc Borislav Petkov <[email protected]>

Link http://lkml.kernel.org/r/20121009180156.GA245@x4 open here

Signed-off-by Ingo Molnar <[email protected]>

---
 tools/perf/Makefile                              | 3 ++-
 tools/perf/util/include/linux/rbtree_augmented.h | 2 ++
 2 files changed, 4 insertions(+), 1 deletion(-)

tools/perf/Makefile +2 -1

index d7b8b54..2472645 100644
@@ -270,6 +270,7 @@ LIB_H += util/include/linux/magic.h
 LIB_H += util/include/linux/poison.h
 LIB_H += util/include/linux/prefetch.h
 LIB_H += util/include/linux/rbtree.h
+LIB_H += util/include/linux/rbtree_augmented.h
 LIB_H += util/include/linux/string.h
 LIB_H += util/include/linux/types.h
 LIB_H += util/include/linux/linkage.h

@@ -881,7 +882,7 @@ $(OUTPUT)ui/browsers/map.o: ui/browsers/map.c $(OUTPUT)PERF-CFLAGS
 	$(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) -DENABLE_SLFUTURE_CONST $<
 
 $(OUTPUT)util/rbtree.o: ../../lib/rbtree.c $(OUTPUT)PERF-CFLAGS
-	$(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) -DETC_PERFCONFIG='"$(ETC_PERFCONFIG_SQ)"' $<
+	$(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) -Wno-unused-parameter -DETC_PERFCONFIG='"$(ETC_PERFCONFIG_SQ)"' $<
 
 $(OUTPUT)util/parse-events.o: util/parse-events.c $(OUTPUT)PERF-CFLAGS
 	$(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) -Wno-redundant-decls $<

tools/perf/util/include/linux/rbtree_augmented.h new file +2 -0

new file mode 100644
index 0000000..9d6fcdf
@@ -0,0 +1,2 @@
+#include <stdbool.h>
+#include "../../../../include/linux/rbtree_augmented.h"

Markus Trippelsdorf

2012-10-12 09:19 UTC #5 lore raw

In reply to Borislav Petkov

Quoted 32 lines · Borislav Petkov

Quoted 19 lines · Ingo Molnar
Quoted 7 lines · Ingo Molnar
Linus,
Please pull the latest perf-urgent-for-linus git tree from:
   git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf-urgent-for-linus

   HEAD: 95cf59ea72331d0093010543b8951bb43f262cac perf: Fix perf_cgroup_switch for sw-events
Note that if you merge it then there's a new semantic conflict with recent rbtree.c changes in your tree, causing a tools/perf/ build failure:
../../lib/rbtree.c:24:36: fatal error: linux/rbtree_augmented.h: No such file or directory compilation terminated.
See the fix below.
Quoted 7 lines
( If the __maybe_unused annotations are too ugly for
lib/rbtree.c then we'll fix that in tools/perf in a cleaner way, weakening the compiler checks for the rbtree build. We are using stronger compiler checks in tools/perf/, which has served us very well so far and is a big net win - the price is the occasional extra annotation of dummy inline function parameters. )
Btw,
Markus fixed it that way recently:

Quoted 7 lines

http://marc.info/?l=linux-kernel&m=134980573527738
by adding the -Wno-unused-parameter switch only to lib/rbtree.o when built from within perf.
I see now though that his patch has some unrelated changes to trace-event-perl.c's permissions which shouldn't be there.

Yes that was part of a different issue:

CC util/scripting-engines/trace-event-perl.o
util/scripting-engines/trace-event-perl.c: In function ‘perl_process_tracepoint’:
util/scripting-engines/trace-event-perl.c:285:3: error: format ‘%lu’ expects argument of type ‘long unsigned int’, but argument 2 has type ‘__u64’ [-Werror=format]
die("ug! no event found for type %" PRIu64, evsel->attr.config);
^
cc1: all warnings being treated as errors

Which I fixed locally like this:

tools/perf/util/scripting-engines/trace-event-perl.c +1 -1

index f80605e..f53889d 100644
@@ -282,7 +282,7 @@ static void perl_process_tracepoint(union perf_event *perf_event __maybe_unused,
 
 	event = find_cache_event(evsel);
 	if (!event)
-		die("ug! no event found for type %" PRIu64, evsel->attr.config);
+		die("ug! no event found for type %" PRIu64, (unsigned long)evsel->attr.config);
 
 	pid = raw_field_value(event, "common_pid", data);

Signature · 1 line

Markus