Re: [PATCH v1 bpf-next 17/18] selftests/bpf: Enable bpf_scan_ir in the config

From: bot+bpf-ci

Date: Fri Nov 06 2026 - 16:35:36 EST


> selftests/bpf: Enable bpf_scan_ir in the config

TL;DR: Commit message only: the list of cases where Kconfig ignores the plugin
options looks incomplete (other arches, missing matching llvm-config), so "CI
builds and tests it" may not hold and fmodret_retval may silently skip.

> selftests/bpf: Enable bpf_scan_ir in the config
>
> The bpf_scan_ir plugin is off by default. Without it bpf_testmod has no
> fixtures for the plugin and fmodret_retval skips. Enable the plugin in
> the selftests config, so CI builds and tests it. Kconfig ignores the
> options with GCC and with clang older than 19.
> Signed-off-by: Anton Protopopov <a.s.protopopov@xxxxxxxxx>

Is the list of conditions in the last sentence complete? The commit message
says Kconfig ignores the options "with GCC and with clang older than 19", and
that CI "builds and tests it".

scripts/llvm-plugins/Kconfig has more conditions. LLVM_PLUGINS needs
CC_IS_CLANG, a working $(LLVM_CONFIG), an llvm-config with the same major
version as $(CC), and a $(HOSTCXX) that accepts the LLVM headers.
CLANG_PLUGIN_BPF_SCAN_IR needs X86_64 || ARM64, DEBUG_INFO_BTF, !LTO_CLANG and
CLANG_VERSION >= 190000.

So config.s390x, config.ppc64el and config.riscv64 never get the plugin, and a
clang >= 19 host without a matching llvm-dev drops both options silently. In
those cases the build succeeds, bpf_testmod is built without the
bpf_testmod_retval*.o objects, and fmodret_retval reports SKIP.

Could the message list the full set of conditions, or say that "CI builds and
tests it" only holds when a matching llvm-config is available?


---
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

CI run summary: https://github.com/kernel-patches/bpf/actions/runs/42294987431