Merge branch 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/l…

…inux/kernel/git/tip/linux-2.6-tip

* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (30 commits)
  perf: Add back list_head data types
  perf ui hist browser: Fixup key bindings
  perf ui browser: Add ui_browser__show counterpart: __hide
  perf annotate: Cycle thru sorted lines with samples
  perf ui: Make SPACE work as PGDN in all browsers
  perf annotate: Sort by hottest lines in the TUI
  perf ui: Complete the breakdown of util/newt.c
  perf ui: Move hists browser to util/ui/browsers/
  perf symbols: Ignore mapping symbols on ARM
  perf ui: Move map browser to util/ui/browsers/
  perf ui: Move annotate browser to util/ui/browsers/
  perf ui: Move ui_progress routines to separate file in util/ui/
  perf ui: Move ui_helpline routines to separate file in util/ui/
  perf ui: Shorten ui_browser member names
  perf, x86: P4 PMU -- update nmi irq statistics and unmask lvt entry properly
  perf ui: Start breaking down newt.c into multiple files
  perf tui: Introduce list_head based generic ui_browser refresh routine
  perf probe: Fix memory leaks in add_perf_probe_events
  perf probe: Fix to copy the type for raw parameters
  perf report: Speed up exit path
  ...

arch/x86/kernel/cpu/perf_event_p4.c

@@ -656,6 +656,7 @@ static int p4_pmu_handle_irq(struct pt_regs *regs)
 	cpuc = &__get_cpu_var(cpu_hw_events);
 
 	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
+		int overflow;
 
 		if (!test_bit(idx, cpuc->active_mask))
 			continue;
@@ -666,12 +667,14 @@ static int p4_pmu_handle_irq(struct pt_regs *regs)
 		WARN_ON_ONCE(hwc->idx != idx);
 
 		/* it might be unflagged overflow */
-		handled = p4_pmu_clear_cccr_ovf(hwc);
+		overflow = p4_pmu_clear_cccr_ovf(hwc);
 
 		val = x86_perf_event_update(event);
-		if (!handled && (val & (1ULL << (x86_pmu.cntval_bits - 1))))
+		if (!overflow && (val & (1ULL << (x86_pmu.cntval_bits - 1))))
 			continue;
 
+		handled += overflow;
+
 		/* event overflow for sure */
 		data.period = event->hw.last_period;
 
@@ -687,7 +690,7 @@ static int p4_pmu_handle_irq(struct pt_regs *regs)
 		inc_irq_stat(apic_perf_irqs);
 	}
 
-	return handled;
+	return handled > 0;
 }
 
 /*

arch/x86/oprofile/nmi_int.c

@@ -668,6 +668,7 @@ static int __init ppro_init(char **cpu_type)
 		*cpu_type = "i386/core_2";
 		break;
 	case 0x1a:
+	case 0x1e:
 	case 0x2e:
 		spec = &op_arch_perfmon_spec;
 		*cpu_type = "i386/core_i7";

tools/perf/Makefile

@@ -157,9 +157,8 @@ all::
 #
 # Define NO_DWARF if you do not want debug-info analysis feature at all.
 
-$(shell sh -c 'mkdir -p $(OUTPUT)scripts/python/Perf-Trace-Util/' 2> /dev/null)
-$(shell sh -c 'mkdir -p $(OUTPUT)scripts/perl/Perf-Trace-Util/' 2> /dev/null)
-$(shell sh -c 'mkdir -p $(OUTPUT)util/scripting-engines/' 2> /dev/null)
+$(shell sh -c 'mkdir -p $(OUTPUT)scripts/{perl,python}/Perf-Trace-Util/' 2> /dev/null)
+$(shell sh -c 'mkdir -p $(OUTPUT)util/{ui/browsers,scripting-engines}/' 2> /dev/null)
 $(shell sh -c 'mkdir $(OUTPUT)bench' 2> /dev/null)
 
 $(OUTPUT)PERF-VERSION-FILE: .FORCE-PERF-VERSION-FILE
@@ -568,7 +567,20 @@ else
 		# Fedora has /usr/include/slang/slang.h, but ubuntu /usr/include/slang.h
 		BASIC_CFLAGS += -I/usr/include/slang
 		EXTLIBS += -lnewt -lslang
-		LIB_OBJS += $(OUTPUT)util/newt.o
+		LIB_OBJS += $(OUTPUT)util/ui/setup.o
+		LIB_OBJS += $(OUTPUT)util/ui/browser.o
+		LIB_OBJS += $(OUTPUT)util/ui/browsers/annotate.o
+		LIB_OBJS += $(OUTPUT)util/ui/browsers/hists.o
+		LIB_OBJS += $(OUTPUT)util/ui/browsers/map.o
+		LIB_OBJS += $(OUTPUT)util/ui/helpline.o
+		LIB_OBJS += $(OUTPUT)util/ui/progress.o
+		LIB_OBJS += $(OUTPUT)util/ui/util.o
+		LIB_H += util/ui/browser.h
+		LIB_H += util/ui/browsers/map.h
+		LIB_H += util/ui/helpline.h
+		LIB_H += util/ui/libslang.h
+		LIB_H += util/ui/progress.h
+		LIB_H += util/ui/util.h
 	endif
 endif
 
@@ -966,7 +978,16 @@ $(OUTPUT)builtin-init-db.o: builtin-init-db.c $(OUTPUT)PERF-CFLAGS
 $(OUTPUT)util/config.o: util/config.c $(OUTPUT)PERF-CFLAGS
 	$(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) -DETC_PERFCONFIG='"$(ETC_PERFCONFIG_SQ)"' $<
 
-$(OUTPUT)util/newt.o: util/newt.c $(OUTPUT)PERF-CFLAGS
+$(OUTPUT)util/ui/browser.o: util/ui/browser.c $(OUTPUT)PERF-CFLAGS
+	$(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) -DENABLE_SLFUTURE_CONST $<
+
+$(OUTPUT)util/ui/browsers/annotate.o: util/ui/browsers/annotate.c $(OUTPUT)PERF-CFLAGS
+	$(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) -DENABLE_SLFUTURE_CONST $<
+
+$(OUTPUT)util/ui/browsers/hists.o: util/ui/browsers/hists.c $(OUTPUT)PERF-CFLAGS
+	$(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) -DENABLE_SLFUTURE_CONST $<
+
+$(OUTPUT)util/ui/browsers/map.o: util/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

tools/perf/builtin-annotate.c

@@ -285,7 +285,7 @@ static int hist_entry__tty_annotate(struct hist_entry *he)
 	LIST_HEAD(head);
 	struct objdump_line *pos, *n;
 
-	if (hist_entry__annotate(he, &head) < 0)
+	if (hist_entry__annotate(he, &head, 0) < 0)
 		return -1;
 
 	if (full_paths)

tools/perf/builtin-report.c

@@ -348,7 +348,18 @@ static int __cmd_report(void)
 		hists__tty_browse_tree(&session->hists_tree, help);
 
 out_delete:
-	perf_session__delete(session);
+	/*
+	 * Speed up the exit process, for large files this can
+	 * take quite a while.
+	 *
+	 * XXX Enable this when using valgrind or if we ever
+	 * librarize this command.
+	 *
+	 * Also experiment with obstacks to see how much speed
+	 * up we'll get here.
+	 *
+ 	 * perf_session__delete(session);
+ 	 */
 	return ret;
 }
 
@@ -478,8 +489,24 @@ int cmd_report(int argc, const char **argv, const char *prefix __used)
 	 * so don't allocate extra space that won't be used in the stdio
 	 * implementation.
 	 */
-	if (use_browser > 0)
+	if (use_browser > 0) {
 		symbol_conf.priv_size = sizeof(struct sym_priv);
+		/*
+ 		 * For searching by name on the "Browse map details".
+ 		 * providing it only in verbose mode not to bloat too
+ 		 * much struct symbol.
+ 		 */
+		if (verbose) {
+			/*
+			 * XXX: Need to provide a less kludgy way to ask for
+			 * more space per symbol, the u32 is for the index on
+			 * the ui browser.
+			 * See symbol__browser_index.
+			 */
+			symbol_conf.priv_size += sizeof(u32);
+			symbol_conf.sort_by_name = true;
+		}
+	}
 
 	if (symbol__init() < 0)
 		return -1;

tools/perf/builtin-timechart.c

@@ -455,8 +455,8 @@ static void sched_switch(int cpu, u64 timestamp, struct trace_entry *te)
 		if (p->current->state != TYPE_NONE)
 			pid_put_sample(sw->next_pid, p->current->state, cpu, p->current->state_since, timestamp);
 
-			p->current->state_since = timestamp;
-			p->current->state = TYPE_RUNNING;
+		p->current->state_since = timestamp;
+		p->current->state = TYPE_RUNNING;
 	}
 
 	if (prev_p->current) {

tools/perf/builtin-trace.c

@@ -1,13 +1,16 @@
 #include "builtin.h"
 
-#include "util/util.h"
+#include "perf.h"
 #include "util/cache.h"
+#include "util/debug.h"
+#include "util/exec_cmd.h"
+#include "util/header.h"
+#include "util/parse-options.h"
+#include "util/session.h"
 #include "util/symbol.h"
 #include "util/thread.h"
-#include "util/header.h"
-#include "util/exec_cmd.h"
 #include "util/trace-event.h"
-#include "util/session.h"
+#include "util/util.h"
 
 static char const		*script_name;
 static char const		*generate_script_lang;
@@ -59,14 +62,6 @@ static int cleanup_scripting(void)
 	return scripting_ops->stop_script();
 }
 
-#include "util/parse-options.h"
-
-#include "perf.h"
-#include "util/debug.h"
-
-#include "util/trace-event.h"
-#include "util/exec_cmd.h"
-
 static char const		*input_name = "perf.data";
 
 static int process_sample_event(event_t *event, struct perf_session *session)

tools/perf/util/debug.c

@@ -23,7 +23,7 @@ int eprintf(int level, const char *fmt, ...)
 	if (verbose >= level) {
 		va_start(args, fmt);
 		if (use_browser > 0)
-			ret = browser__show_help(fmt, args);
+			ret = ui_helpline__show_help(fmt, args);
 		else
 			ret = vfprintf(stderr, fmt, args);
 		va_end(args);

tools/perf/util/debug.h

@@ -14,7 +14,7 @@ void trace_event(event_t *event);
 struct ui_progress;
 
 #ifdef NO_NEWT_SUPPORT
-static inline int browser__show_help(const char *format __used, va_list ap __used)
+static inline int ui_helpline__show_help(const char *format __used, va_list ap __used)
 {
 	return 0;
 }
@@ -30,10 +30,9 @@ static inline void ui_progress__update(struct ui_progress *self __used,
 
 static inline void ui_progress__delete(struct ui_progress *self __used) {}
 #else
-int browser__show_help(const char *format, va_list ap);
-struct ui_progress *ui_progress__new(const char *title, u64 total);
-void ui_progress__update(struct ui_progress *self, u64 curr);
-void ui_progress__delete(struct ui_progress *self);
+extern char ui_helpline__last_msg[];
+int ui_helpline__show_help(const char *format, va_list ap);
+#include "ui/progress.h"
 #endif
 
 #endif	/* __PERF_DEBUG_H */

tools/perf/util/hist.c

@@ -876,6 +876,9 @@ unsigned int hists__sort_list_width(struct hists *self)
 		if (!se->elide)
 			ret += 2 + hists__col_len(self, se->se_width_idx);
 
+	if (verbose) /* Addr + origin */
+		ret += 3 + BITS_PER_LONG / 4;
+
 	return ret;
 }
 
@@ -980,9 +983,9 @@ int hist_entry__inc_addr_samples(struct hist_entry *self, u64 ip)
 	return 0;
 }
 
-static struct objdump_line *objdump_line__new(s64 offset, char *line)
+static struct objdump_line *objdump_line__new(s64 offset, char *line, size_t privsize)
 {
-	struct objdump_line *self = malloc(sizeof(*self));
+	struct objdump_line *self = malloc(sizeof(*self) + privsize);
 
 	if (self != NULL) {
 		self->offset = offset;
@@ -1014,7 +1017,7 @@ struct objdump_line *objdump__get_next_ip_line(struct list_head *head,
 }
 
 static int hist_entry__parse_objdump_line(struct hist_entry *self, FILE *file,
-					  struct list_head *head)
+					  struct list_head *head, size_t privsize)
 {
 	struct symbol *sym = self->ms.sym;
 	struct objdump_line *objdump_line;
@@ -1065,7 +1068,7 @@ static int hist_entry__parse_objdump_line(struct hist_entry *self, FILE *file,
 			offset = -1;
 	}
 
-	objdump_line = objdump_line__new(offset, line);
+	objdump_line = objdump_line__new(offset, line, privsize);
 	if (objdump_line == NULL) {
 		free(line);
 		return -1;
@@ -1075,7 +1078,8 @@ static int hist_entry__parse_objdump_line(struct hist_entry *self, FILE *file,
 	return 0;
 }
 
-int hist_entry__annotate(struct hist_entry *self, struct list_head *head)
+int hist_entry__annotate(struct hist_entry *self, struct list_head *head,
+			 size_t privsize)
 {
 	struct symbol *sym = self->ms.sym;
 	struct map *map = self->ms.map;
@@ -1140,7 +1144,7 @@ int hist_entry__annotate(struct hist_entry *self, struct list_head *head)
 		goto out_free_filename;
 
 	while (!feof(file))
-		if (hist_entry__parse_objdump_line(self, file, head) < 0)
+		if (hist_entry__parse_objdump_line(self, file, head, privsize) < 0)
 			break;
 
 	pclose(file);

tools/perf/util/hist.h

@@ -101,7 +101,8 @@ size_t hists__fprintf(struct hists *self, struct hists *pair,
 		      bool show_displacement, FILE *fp);
 
 int hist_entry__inc_addr_samples(struct hist_entry *self, u64 ip);
-int hist_entry__annotate(struct hist_entry *self, struct list_head *head);
+int hist_entry__annotate(struct hist_entry *self, struct list_head *head,
+			 size_t privsize);
 
 void hists__filter_by_dso(struct hists *self, const struct dso *dso);
 void hists__filter_by_thread(struct hists *self, const struct thread *thread);

tools/perf/util/include/linux/list.h

@@ -15,4 +15,12 @@ static inline void list_del_range(struct list_head *begin,
 	begin->prev->next = end->next;
 	end->next->prev = begin->prev;
 }
+
+/**
+ * list_for_each_from	-	iterate over a list from one of its nodes
+ * @pos:  the &struct list_head to use as a loop cursor, from where to start
+ * @head: the head for your list.
+ */
+#define list_for_each_from(pos, head) \
+	for (; prefetch(pos->next), pos != (head); pos = pos->next)
 #endif

tools/perf/util/include/linux/types.h

@@ -6,4 +6,16 @@
 #define DECLARE_BITMAP(name,bits) \
 	unsigned long name[BITS_TO_LONGS(bits)]
 
+struct list_head {
+	struct list_head *next, *prev;
+};
+
+struct hlist_head {
+	struct hlist_node *first;
+};
+
+struct hlist_node {
+	struct hlist_node *next, **pprev;
+};
+
 #endif

tools/perf/util/probe-event.c

@@ -1606,8 +1606,10 @@ int add_perf_probe_events(struct perf_probe_event *pevs, int npevs,
 
 	/* Init vmlinux path */
 	ret = init_vmlinux();
-	if (ret < 0)
+	if (ret < 0) {
+		free(pkgs);
 		return ret;
+	}
 
 	/* Loop 1: convert all events */
 	for (i = 0; i < npevs; i++) {
@@ -1625,10 +1627,13 @@ int add_perf_probe_events(struct perf_probe_event *pevs, int npevs,
 		ret = __add_probe_trace_events(pkgs[i].pev, pkgs[i].tevs,
 						pkgs[i].ntevs, force_add);
 end:
-	/* Loop 3: cleanup trace events  */
-	for (i = 0; i < npevs; i++)
+	/* Loop 3: cleanup and free trace events  */
+	for (i = 0; i < npevs; i++) {
 		for (j = 0; j < pkgs[i].ntevs; j++)
 			clear_probe_trace_event(&pkgs[i].tevs[j]);
+		free(pkgs[i].tevs);
+	}
+	free(pkgs);
 
 	return ret;
 }

tools/perf/util/probe-finder.c

@@ -33,7 +33,6 @@
 #include <ctype.h>
 #include <dwarf-regs.h>
 
-#include "string.h"
 #include "event.h"
 #include "debug.h"
 #include "util.h"
@@ -706,8 +705,12 @@ static int find_variable(Dwarf_Die *sp_die, struct probe_finder *pf)
 		pf->tvar->value = strdup(pf->pvar->var);
 		if (pf->tvar->value == NULL)
 			return -ENOMEM;
-		else
-			return 0;
+		if (pf->pvar->type) {
+			pf->tvar->type = strdup(pf->pvar->type);
+			if (pf->tvar->type == NULL)
+				return -ENOMEM;
+		}
+		return 0;
 	}
 
 	pr_debug("Searching '%s' variable in context.\n",

tools/perf/util/pstack.h

@@ -1,6 +1,8 @@
 #ifndef _PERF_PSTACK_
 #define _PERF_PSTACK_
 
+#include <stdbool.h>
+
 struct pstack;
 struct pstack *pstack__new(unsigned short max_nr_entries);
 void pstack__delete(struct pstack *self);