libbpf: Make optimized uprobes backward compatible

[kernel-patches-daemon-bpf]

Pull request for series with
subject: libbpf: Make optimized uprobes backward compatible
version: 1
url: https://patchwork.kernel.org/project/netdevbpf/list/?series=1052727

[kernel-patches-daemon-bpf]

Upstream branch: db975de
series: https://patchwork.kernel.org/project/netdevbpf/list/?series=1052727
version: 1

[kernel-patches-review-bot]

commit 70e5527050f3499bd935d6d5bd5dedebf29326cb
Author: Jiri Olsa <jolsa@kernel.org>

selftests/bpf: Emit nop,no5 instructions combo for x86_64 arch

This commit syncs the selftests usdt.h with the upstream libbpf/usdt
project to emit a nop,nop5 instruction combo for x86_64 USDT probes,
allowing libbpf to use the desirable nop instruction for optimization.

> selftests/bpf: Emit nop,no5 instructions combo for x86_64 arch

This isn't a bug, but there's a typo in the subject line. "no5" should
be "nop5".

> Now that we have no5 optimization support in kernel, let's emit

The same typo appears here in the commit body.

---

AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

In-Reply-To-Subject: selftests/bpf: Emit nop,no5 instructions combo for x86_64 arch
CI run summary: https://github.com/kernel-patches/bpf/actions/runs/21867569087

AI-authorship-score: low
AI-authorship-explanation: The commit shows typical human development patterns including a typo in the commit message, iterative design changes between v1 and v2 based on real-world testing, and normal kernel mailing list review interactions.
issues-found: 1
issue-severity-score: low
issue-severity-explanation: Single typo in commit message subject and body where 'no5' should be 'nop5'; does not affect code functionality.

[kernel-patches-review-bot]

commit cd9efa94104e93bf2a6238dc581ef561b1f69aed
Author: Jiri Olsa <jolsa@kernel.org>

selftests/bpf: Add test for checking correct nop of optimized usdt

This commit adds a test that attaches a BPF program to USDT probes in two
scenarios: one using a single nop instruction (usdt_1), and one using the
nop,nop5 combo (usdt_2) to verify probe optimization behavior.

> diff --git a/tools/testing/selftests/bpf/prog_tests/usdt.c b/tools/testing/selftests/bpf/prog_tests/usdt.c
> --- a/tools/testing/selftests/bpf/prog_tests/usdt.c
> +++ b/tools/testing/selftests/bpf/prog_tests/usdt.c

[ ... ]

> +#ifdef __x86_64
          ^^^^^^^^
Should this be __x86_64__ (with trailing underscores) to match the standard
GCC predefined macro? The call site in test_usdt() uses __x86_64__:

> @@ -516,6 +599,8 @@ void test_usdt(void)
> #ifdef __x86_64__
> 	if (test__start_subtest("basic_optimized"))
> 		subtest_basic_usdt(true);
> +	if (test__start_subtest("optimized_attach"))
> +		subtest_optimized_attach();
> #endif

Without the trailing underscores, the function definition will not be
compiled on x86_64 while the call will be, leading to a linker error.

---

AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

In-Reply-To-Subject: selftests/bpf: Add test for checking correct nop of optimized usdt
CI run summary: https://github.com/kernel-patches/bpf/actions/runs/21867569087

AI-authorship-score: low
AI-authorship-explanation: The commit follows standard kernel selftest patterns with a natural coding style and contains a typo characteristic of human error rather than AI generation.
issues-found: 1
issue-severity-score: medium
issue-severity-explanation: The mismatched preprocessor macro will cause a linker error on x86_64 builds, preventing the test from compiling.

[kernel-patches-review-bot]

commit dabc30135d3b4a26c0348604c610154df2b3c078
Author: Jiri Olsa <jolsa@kernel.org>

selftests/bpf: Add usdt trigger bench

This commit adds USDT trigger benchmarks for measuring USDT probe performance
with single nop and nop/nop5 combo instruction patterns.

> diff --git a/tools/testing/selftests/bpf/progs/trigger_bench.c b/tools/testing/selftests/bpf/progs/trigger_bench.c
> index 4ea0422d1042..f631706bbdc9 100644

[ ... ]

> @@ -180,3 +181,10 @@ int bench_trigger_rawtp(void *ctx)
>  	handle(ctx);
>  	return 0;
>  }
> +
> +SEC("usdt")
> +int bench_trigger_usdt(void *ctx)

Should this use SEC("?usdt") instead of SEC("usdt")?

All other programs in this file use the '?' prefix to disable autoload by
default:

    SEC("?uprobe")
    SEC("?uprobe.multi")
    SEC("?raw_tp")
    SEC("?kprobe/bpf_get_numa_node_id")
    ...

Without the '?' prefix, bench_trigger_usdt will be auto-loaded whenever the
trigger_bench skeleton is loaded by any benchmark using setup_ctx()/load_ctx(),
adding unnecessary overhead to unrelated benchmarks.

> +{
> +	inc_counter();
> +	return 0;
> +}

---

AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

In-Reply-To-Subject: selftests/bpf: Add usdt trigger bench
CI run summary: https://github.com/kernel-patches/bpf/actions/runs/21867569087

AI-authorship-score: low
AI-authorship-explanation: The commit exhibits natural kernel developer style with real benchmark data, informal comments, and follows existing codebase patterns with minor inconsistencies typical of human development.
issues-found: 1
issue-severity-score: low
issue-severity-explanation: Minor pattern inconsistency in test code where SEC("usdt") should use SEC("?usdt") to match the established convention for disabling autoload by default.

[kernel-patches-daemon-bpf]

Forwarding comment 3877838387 via email
In-Reply-To: 20260210133649.524292-4-jolsa@kernel.org
Patch: https://patchwork.kernel.org/project/netdevbpf/patch/20260210133649.524292-4-jolsa@kernel.org/

[kernel-patches-daemon-bpf]

Forwarding comment 3877846166 via email
In-Reply-To: 20260210133649.524292-5-jolsa@kernel.org
Patch: https://patchwork.kernel.org/project/netdevbpf/patch/20260210133649.524292-5-jolsa@kernel.org/

[kernel-patches-daemon-bpf]

Forwarding comment 3877853344 via email
In-Reply-To: 20260210133649.524292-6-jolsa@kernel.org
Patch: https://patchwork.kernel.org/project/netdevbpf/patch/20260210133649.524292-6-jolsa@kernel.org/

[kernel-patches-daemon-bpf]

Upstream branch: a9aabb3
series: https://patchwork.kernel.org/project/netdevbpf/list/?series=1052727
version: 1

[kernel-patches-daemon-bpf]

Upstream branch: dc855b7
series: https://patchwork.kernel.org/project/netdevbpf/list/?series=1052727
version: 1

[kernel-patches-daemon-bpf]

Upstream branch: dc855b7
series: https://patchwork.kernel.org/project/netdevbpf/list/?series=1052727
version: 1