perf report: Display columns Predicted/Abort/Cycles in --branch-history

The original commit message:

"
Use current sort mechanism but the real .se_cmp() just returns 0 so
that new columns "Predicted", "Abort" and "Cycles" are created in display
but actually these keys are not the sort keys.

For example:

Overhead Source:Line Symbol Shared Object Predicted Abort Cycles
........ ............ ........ ............. ......... ..... ......

38.25% div.c:45 [.] main div 97.6% 0 3
"

Update missed commit from series "perf report: Show branch flags/cycles
in --branch-history callgraph view" to apply to current repository so that
new columns described above are visible.

Link to original series:
https://lore.kernel.org/lkml/[e-mail]/

Reported-by: Dr. David Alan Gilbert <[e-mail]>
Suggested-by: Kan Liang <[e-mail]>
Co-developed-by: Jin Yao <[e-mail]>
Signed-off-by: Jin Yao <[e-mail]>
Tested-by: Thomas Falcon <[e-mail]>
Signed-off-by: Thomas Falcon <[e-mail]>
Link: https://lore.kernel.org/r/[e-mail]
Signed-off-by: Namhyung Kim <[e-mail]>

tools/perf/Documentation/perf-report.txt

···
 	This allows to examine the path the program took to each sample.
 	The data collection must have used -b (or -j) and -g.
 
+	Also show with some branch flags that can be:
+	- Predicted: display the average percentage of predicated branches.
+		     (predicated number / total number)
+	- Abort: display the number of tsx aborted branches.
+	- Cycles: cycles in basic block.
+
+	- iterations: display the average number of iterations in callchain list.
+
 --addr2line=<path>::
         Path to addr2line binary.
 

tools/perf/builtin-report.c

···
 	return 0;
 }
 
+#define CALLCHAIN_BRANCH_SORT_ORDER	\
+	"srcline,symbol,dso,callchain_branch_predicted," \
+	"callchain_branch_abort,callchain_branch_cycles"
+
 int cmd_report(int argc, const char **argv)
 {
 	struct perf_session *session;
···
 		symbol_conf.use_callchain = true;
 		callchain_register_param(&callchain_param);
 		if (sort_order == NULL)
-			sort_order = "srcline,symbol,dso";
+			sort_order = CALLCHAIN_BRANCH_SORT_ORDER;
 	}
 
 	if (report.mem_mode) {

tools/perf/util/hist.c

···
 	hists__new_col_len(hists, HISTC_LOCAL_P_STAGE_CYC, 13);
 	hists__new_col_len(hists, HISTC_GLOBAL_P_STAGE_CYC, 13);
 	hists__new_col_len(hists, HISTC_ADDR, BITS_PER_LONG / 4 + 2);
+	hists__new_col_len(hists, HISTC_CALLCHAIN_BRANCH_PREDICTED, 9);
+	hists__new_col_len(hists, HISTC_CALLCHAIN_BRANCH_ABORT, 5);
+	hists__new_col_len(hists, HISTC_CALLCHAIN_BRANCH_CYCLES, 6);
 
 	if (symbol_conf.nanosecs)
 		hists__new_col_len(hists, HISTC_TIME, 16);

tools/perf/util/hist.h

···
 	HISTC_TYPE_OFFSET,
 	HISTC_SYMBOL_OFFSET,
 	HISTC_TYPE_CACHELINE,
+	HISTC_CALLCHAIN_BRANCH_PREDICTED,
+	HISTC_CALLCHAIN_BRANCH_ABORT,
+	HISTC_CALLCHAIN_BRANCH_CYCLES,
 	HISTC_NR_COLS, /* Last entry */
 };
 

tools/perf/util/sort.c

···
 	.se_width_idx	= HISTC_SYMBOL_IPC,
 };
 
+/* --sort callchain_branch_predicted */
+
+static int64_t
+sort__callchain_branch_predicted_cmp(struct hist_entry *left __maybe_unused,
+				     struct hist_entry *right __maybe_unused)
+{
+	return 0;
+}
+
+static int hist_entry__callchain_branch_predicted_snprintf(
+	struct hist_entry *he, char *bf, size_t size, unsigned int width)
+{
+	u64 branch_count, predicted_count;
+	double percent = 0.0;
+	char str[32];
+
+	callchain_branch_counts(he->callchain, &branch_count,
+				&predicted_count, NULL, NULL);
+
+	if (branch_count)
+		percent = predicted_count * 100.0 / branch_count;
+
+	snprintf(str, sizeof(str), "%.1f%%", percent);
+	return repsep_snprintf(bf, size, "%-*.*s", width, width, str);
+}
+
+struct sort_entry sort_callchain_branch_predicted = {
+	.se_header	= "Predicted",
+	.se_cmp		= sort__callchain_branch_predicted_cmp,
+	.se_snprintf	= hist_entry__callchain_branch_predicted_snprintf,
+	.se_width_idx	= HISTC_CALLCHAIN_BRANCH_PREDICTED,
+};
+
+/* --sort callchain_branch_abort */
+
+static int64_t
+sort__callchain_branch_abort_cmp(struct hist_entry *left __maybe_unused,
+				 struct hist_entry *right __maybe_unused)
+{
+	return 0;
+}
+
+static int hist_entry__callchain_branch_abort_snprintf(struct hist_entry *he,
+						       char *bf, size_t size,
+						       unsigned int width)
+{
+	u64 branch_count, abort_count;
+	char str[32];
+
+	callchain_branch_counts(he->callchain, &branch_count,
+				NULL, &abort_count, NULL);
+
+	snprintf(str, sizeof(str), "%" PRId64, abort_count);
+	return repsep_snprintf(bf, size, "%-*.*s", width, width, str);
+}
+
+struct sort_entry sort_callchain_branch_abort = {
+	.se_header	= "Abort",
+	.se_cmp		= sort__callchain_branch_abort_cmp,
+	.se_snprintf	= hist_entry__callchain_branch_abort_snprintf,
+	.se_width_idx	= HISTC_CALLCHAIN_BRANCH_ABORT,
+};
+
+/* --sort callchain_branch_cycles */
+
+static int64_t
+sort__callchain_branch_cycles_cmp(struct hist_entry *left __maybe_unused,
+				  struct hist_entry *right __maybe_unused)
+{
+	return 0;
+}
+
+static int hist_entry__callchain_branch_cycles_snprintf(struct hist_entry *he,
+							char *bf, size_t size,
+							unsigned int width)
+{
+	u64 branch_count, cycles_count, cycles = 0;
+	char str[32];
+
+	callchain_branch_counts(he->callchain, &branch_count,
+				NULL, NULL, &cycles_count);
+
+	if (branch_count)
+		cycles = cycles_count / branch_count;
+
+	snprintf(str, sizeof(str), "%" PRId64 "", cycles);
+	return repsep_snprintf(bf, size, "%-*.*s", width, width, str);
+}
+
+struct sort_entry sort_callchain_branch_cycles = {
+	.se_header	= "Cycles",
+	.se_cmp		= sort__callchain_branch_cycles_cmp,
+	.se_snprintf	= hist_entry__callchain_branch_cycles_snprintf,
+	.se_width_idx	= HISTC_CALLCHAIN_BRANCH_CYCLES,
+};
+
 /* --sort srcfile */
 
 static char no_srcfile[1];
···
 	DIM(SORT_SYM_IPC, "ipc_lbr", sort_sym_ipc),
 	DIM(SORT_ADDR_FROM, "addr_from", sort_addr_from),
 	DIM(SORT_ADDR_TO, "addr_to", sort_addr_to),
+	DIM(SORT_CALLCHAIN_BRANCH_PREDICTED,
+		"callchain_branch_predicted",
+		sort_callchain_branch_predicted),
+	DIM(SORT_CALLCHAIN_BRANCH_ABORT,
+		"callchain_branch_abort",
+		sort_callchain_branch_abort),
+	DIM(SORT_CALLCHAIN_BRANCH_CYCLES,
+		"callchain_branch_cycles",
+		sort_callchain_branch_cycles)
 };
 
 #undef DIM
···
 		if (!sd->name || strncasecmp(tok, sd->name, strlen(tok)))
 			continue;
 
-		if (sort__mode != SORT_MODE__BRANCH)
+		if ((sort__mode != SORT_MODE__BRANCH) &&
+			strncasecmp(tok, "callchain_branch_predicted",
+				    strlen(tok)) &&
+			strncasecmp(tok, "callchain_branch_abort",
+				    strlen(tok)) &&
+			strncasecmp(tok, "callchain_branch_cycles",
+				    strlen(tok)))
 			return -EINVAL;
 
 		if (sd->entry == &sort_sym_from || sd->entry == &sort_sym_to)

tools/perf/util/sort.h

···
 	SORT_SYM_IPC,
 	SORT_ADDR_FROM,
 	SORT_ADDR_TO,
+	SORT_CALLCHAIN_BRANCH_PREDICTED,
+	SORT_CALLCHAIN_BRANCH_ABORT,
+	SORT_CALLCHAIN_BRANCH_CYCLES,
 
 	/* memory mode specific sort keys */
 	__SORT_MEMORY_MODE,