Skip to contentSkip to conversation

selftests/bpf: skip test_perf_branches_hw() on unsupported platforms

PATCH Applied Matt Bobrowski 3 participants 3 messages Nov 20, 2025 – Nov 22, 2025

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 selftests/bpf: skip test_perf_branches_hw() on unsupported platforms

Gracefully skip the test_perf_branches_hw subtest on platforms that do not support LBR or require specialized perf event attributes to enable branch sampling.

For example, AMD's Milan (Zen 3) supports BRS rather than traditional LBR. This requires specific configurations (attr.type = PERF_TYPE_RAW, attr.config = RETIRED_TAKEN_BRANCH_INSTRUCTIONS) that differ from the generic setup used within this test. Notably, it also probably doesn't hold much value to special case perf event configurations for selected micro architectures.

Fixes 67306f84ca78c ("selftests/bpf: Add bpf_read_branch_records() selftest")
---
 tools/testing/selftests/bpf/prog_tests/perf_branches.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
tools/testing/selftests/bpf/prog_tests/perf_branches.c +3 -3
index 1d51ec5f171a..0a7ef770c487 100644
@@ -128,11 +128,11 @@ static void test_perf_branches_hw(void)
 	pfd = syscall(__NR_perf_event_open, &attr, -1, 0, -1, PERF_FLAG_FD_CLOEXEC);
 
 	/*
-	 * Some setups don't support branch records (virtual machines, !x86),
-	 * so skip test in this case.
+	 * Some setups don't support LBR (virtual machines, !x86, AMD Milan Zen
+	 * 3 which only supports BRS), so skip test in this case.
 	 */
 	if (pfd < 0) {
-		if (errno == ENOENT || errno == EOPNOTSUPP) {
+		if (errno == ENOENT || errno == EOPNOTSUPP || errno == EINVAL) {
 			printf("%s:SKIP:no PERF_SAMPLE_BRANCH_STACK\n",
 			       __func__);
 			test__skip();
Signature · 1 line
2.52.0.rc1.455.g30608eb744-goog
reviewer Acked-by on v1 1/1
#1 lore raw
Quoted 12 lines · Matt Bobrowski

Gracefully skip the test_perf_branches_hw subtest on platforms that do not support LBR or require specialized perf event attributes to enable branch sampling.

For example, AMD's Milan (Zen 3) supports BRS rather than traditional LBR. This requires specific configurations (attr.type = PERF_TYPE_RAW, attr.config = RETIRED_TAKEN_BRANCH_INSTRUCTIONS) that differ from the generic setup used within this test. Notably, it also probably doesn't hold much value to special case perf event configurations for selected micro architectures.

Quoted 2 lines
Fixes 67306f84ca78c ("selftests/bpf: Add bpf_read_branch_records() selftest")

Thanks for the fix!

#2 lore raw
In reply to Matt Bobrowski
This patch was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:

Quoted 4 lines · On Thu, 20 Nov 2025 14:20:59 +0000 you wrote:

Gracefully skip the test_perf_branches_hw subtest on platforms that do not support LBR or require specialized perf event attributes to enable branch sampling.

Quoted 8 lines

For example, AMD's Milan (Zen 3) supports BRS rather than traditional LBR. This requires specific configurations (attr.type = PERF_TYPE_RAW, attr.config = RETIRED_TAKEN_BRANCH_INSTRUCTIONS) that differ from the generic setup used within this test. Notably, it also probably doesn't hold much value to special case perf event configurations for selected micro architectures.

[...]

Here is the summary with links:
- [bpf-next] selftests/bpf: skip test_perf_branches_hw() on unsupported platforms

    https://git.kernel.org/bpf/bpf-next/c/27746aaf1b20

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html