Re: [PATCH bpf-next v6 6/9] selftests/bpf: derive the bench object list from the sources

> The bench binary links every benchs/bench_*.c object, but the link
> rule names them one by one in a hand-maintained list - adding a
> benchmark means writing its name three times (source file, skeleton
> dependency line, link list). Derive the list with a wildcard, as the
> per-object skeleton dependency lines already imply: any bench_*.c
> placed in benchs/ is compiled and linked in.
>
> The derived list is sorted, which changes the link order of the bench
> objects (previously roughly chronological) and with it the symbol
> layout of the binary; no benchmark behaves differently. The trailing
> '#' terminator goes away: the list now ends in a variable, so the
> append-friendliness it provided no longer applies.
>
> Signed-off-by: Mykola Lysenko <[email protected]>

This isn't a bug, but could this last paragraph be trimmed or corrected?
The list still ends in $(OUTPUT)/usdt_2.o rather than a variable, and
the '#' could have stayed.

Similarly, is it the pattern rule rather than the skeleton dependency
lines that already makes any benchs/bench_*.c buildable? The statement
that "adding a benchmark means writing its name three times" also
doesn't hold for benchmarks without a skeleton like benchs/bench_count.c,
which has no dependency line at all.


---
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/32610372822