Re: [PATCH bpf-next v1] selftests/bpf: Add test for indirect struct_ops trampoline

[bot+bpf-ci@kernel.org]

[-- Attachment #1: Type: text/plain, Size: 3214 bytes --]

> diff --git a/tools/testing/selftests/bpf/prog_tests/test_struct_ops_trampoline.c b/tools/testing/selftests/bpf/prog_tests/test_struct_ops_trampoline.c
> new file mode 100644
> index 0000000000000..5a21aedd02fc5
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/prog_tests/test_struct_ops_trampoline.c
> @@ -0,0 +1,41 @@
> +// SPDX-License-Identifier: GPL-2.0
> +#include <test_progs.h>
> +#include "struct_ops_trampoline.skel.h"

A subsystem pattern flags this as potentially concerning: the new
standalone prog_tests file plus a new progs file are added to cover a
9-argument struct_ops callback whose last argument is stack-passed.

tools/testing/selftests/bpf/progs/struct_ops_arena.c already exercises
this shape via bpf_testmod_ops3.test_arena_stack, whose declaration in
bpf_testmod.h carries the comment "enough leading args to force @ptr onto
the stack on x86 and arm64", and test_arena_stack_cb() already pins the
slot layout with argument checks before using ctx[8]. That test is driven
from prog_tests/test_struct_ops_arena.c in the same directory.

The plausible reason for a separate file is that the arena test is gated
to __x86_64__ || __aarch64__ and needs arena support, which LoongArch and
RISC-V lack.

Could a plain (non-arena) >8-argument member on bpf_testmod_ops3 plus a
new case in test_struct_ops_arena.c or test_struct_ops_multi_args.c have
avoided a new test binary?

[ ... ]

> diff --git a/tools/testing/selftests/bpf/progs/struct_ops_trampoline.c b/tools/testing/selftests/bpf/progs/struct_ops_trampoline.c
> new file mode 100644
> index 0000000000000..b0fc19a2a0c53
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/progs/struct_ops_trampoline.c
> @@ -0,0 +1,24 @@
> +// SPDX-License-Identifier: GPL-2.0
> +#include <vmlinux.h>
> +#include <bpf/bpf_tracing.h>
> +#include "../test_kmods/bpf_testmod.h"
> +#include "bpf_misc.h"

This isn't a bug, but could the bpf_misc.h include be dropped? The
program has no __failure/__success/__msg/__retval/__sink/__arch_*
annotation, unlike progs/struct_ops_multi_args.c which includes the same
header and does use __failure __msg(...).

[ ... ]

> diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
> index 850cf4f830c48..be96bbc1b7c04 100644
> --- a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
> +++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
> @@ -589,6 +589,9 @@ noinline int bpf_testmod_trampoline_count_test(void)
>  	return 0;
>  }
>
> +struct bpf_testmod_ops;
> +static struct bpf_testmod_ops *st_ops_trampoline;

This isn't a bug, but does the struct bpf_testmod_ops forward declaration
add anything here? bpf_testmod.c already includes "bpf_testmod.h" at the
top, which defines the struct completely, and the dereference below at
st_ops_trampoline->test_trampoline requires the complete type. The
neighbouring static struct bpf_testmod_ops3 *st_ops3 carries no such
forward declaration.

[ ... ]


---
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/33142391368