perf test: Extend branch stack sampling test for Arm64 BRBE

BRBE emits IRQ and ERET branches for branching and returning from
trapped instructions. Add a test that loops on a trapped instruction
(MRS - Read special register) for this.

Extend the expected 'any_call' branches to include FAULT_DATA and
FAULT_INST as these are emitted by BRBE.

Reviewed-by: Ian Rogers <[e-mail]>
Co-developed-by: German Gomez <[e-mail]>
Signed-off-by: German Gomez <[e-mail]>
Signed-off-by: James Clark <[e-mail]>
Cc: Adam Young <[e-mail]>
Cc: Adrian Hunter <[e-mail]>
Cc: Alexander Shishkin <[e-mail]>
Cc: Anshuman Khandual <[e-mail]>
Cc: Ingo Molnar <[e-mail]>
Cc: Jiri Olsa <[e-mail]>
Cc: Mark Rutland <[e-mail]>
Cc: Namhyung Kim <[e-mail]>
Cc: Peter Zijlstra <[e-mail]>
Cc: Rob Herring <[e-mail]>
Cc: Will Deacon <[e-mail]>
Signed-off-by: Arnaldo Carvalho de Melo <[e-mail]>

tools/perf/tests/builtin-test.c

···
 	&workload__brstack,
 	&workload__datasym,
 	&workload__landlock,
+	&workload__traploop,
 };
 
 #define workloads__for_each(workload) \

tools/perf/tests/shell/test_brstack.sh

···
 }
 trap trap_cleanup EXIT TERM INT
 
+is_arm64() {
+	[ "$(uname -m)" = "aarch64" ];
+}
+
 check_branches() {
 	if ! tr -s ' ' '\n' < "$TMPDIR/perf.script" | grep -E -m1 -q "$1"; then
 		echo "Branches missing $1"
···
 		err=1
 	fi
 	# some branch types are still not being tested:
-	# IND COND_CALL COND_RET SYSRET IRQ SERROR NO_TX
+	# IND COND_CALL COND_RET SYSRET SERROR NO_TX
 }
 
+test_trap_eret_branches() {
+	echo "Testing trap & eret branches"
+	if ! is_arm64; then
+		echo "skip: not arm64"
+	else
+		perf record -o $TMPDIR/perf.data --branch-filter any,save_type,u,k -- \
+			perf test -w traploop 1000
+		perf script -i $TMPDIR/perf.data --fields brstacksym | \
+			tr ' ' '\n' > $TMPDIR/perf.script
+
+		# BRBINF<n>.TYPE == TRAP are mapped to PERF_BR_IRQ by the BRBE driver
+		check_branches "^trap_bench\+[^ ]+/[^ ]/IRQ/"
+		check_branches "^[^ ]+/trap_bench\+[^ ]+/ERET/"
+	fi
+}
 
 test_kernel_branches() {
 	echo "Testing that k option only includes kernel source addresses"
···
 test_user_branches
 test_syscall
 test_kernel_branches
+test_trap_eret_branches
 
 any_call="CALL|IND_CALL|COND_CALL|SYSCALL|IRQ"
+
+if is_arm64; then
+	any_call="$any_call|FAULT_DATA|FAULT_INST"
+fi
 
 test_filter "any_call" "$any_call"
 test_filter "call"	"CALL|SYSCALL"

tools/perf/tests/tests.h

···
 DECLARE_WORKLOAD(brstack);
 DECLARE_WORKLOAD(datasym);
 DECLARE_WORKLOAD(landlock);
+DECLARE_WORKLOAD(traploop);
 
 extern const char *dso_to_test;
 extern const char *test_objdump_path;

tools/perf/tests/workloads/Build

···
 perf-test-y += brstack.o
 perf-test-y += datasym.o
 perf-test-y += landlock.o
+perf-test-y += traploop.o
 
 CFLAGS_sqrtloop.o         = -g -O0 -fno-inline -U_FORTIFY_SOURCE
 CFLAGS_leafloop.o         = -g -O0 -fno-inline -fno-omit-frame-pointer -U_FORTIFY_SOURCE
 CFLAGS_brstack.o          = -g -O0 -fno-inline -U_FORTIFY_SOURCE
 CFLAGS_datasym.o          = -g -O0 -fno-inline -U_FORTIFY_SOURCE
+CFLAGS_traploop.o         = -g -O0 -fno-inline -U_FORTIFY_SOURCE

tools/perf/tests/workloads/traploop.c

···
+// SPDX-License-Identifier: GPL-2.0
+#include <stdlib.h>
+#include "../tests.h"
+
+#define BENCH_RUNS 999999
+
+#ifdef __aarch64__
+static void trap_bench(void)
+{
+	unsigned long val;
+
+	asm("mrs %0, ID_AA64ISAR0_EL1" : "=r" (val));   /* TRAP + ERET */
+}
+#else
+static void trap_bench(void) { }
+#endif
+
+static int traploop(int argc, const char **argv)
+{
+	int num_loops = BENCH_RUNS;
+
+	if (argc > 0)
+		num_loops = atoi(argv[0]);
+
+	for (int i = 0; i < num_loops; i++)
+		trap_bench();
+
+	return 0;
+}
+
+DEFINE_WORKLOAD(traploop);