[PATCH bpf-next v2 10/21] bpf: Shrink CONFIG_HAVE_SINGLE_FTRACE_DIRECT_OPS guard scope for bpf_tracing_multi_attach()
flat view
COOLING4d
From: Leon Hwang <hidden>
Date: 2026-10-12 15:43:42
Also in:
bpf, linux-kselftest, lkml
Subsystem:
bpf [core], bpf [general] (safe dynamic programs and tools), bpf [security & lsm] (security audit and enforcement using bpf), bpf [tracing], the rest, tracing · Maintainers:
Alexei Starovoitov, Daniel Borkmann, Andrii Nakryiko, Eduard Zingerman, Kumar Kartikeya Dwivedi, KP Singh, Matt Bobrowski, Song Liu, Linus Torvalds, Steven Rostedt, Masami Hiramatsu
Revision v2 of 2 in this series.
Revisions (2)
- v1 [diff vs current]
- v2 current
When adding tracing_multi link support for bpf progs, it will use the same UAPI of the tracing_multi link for kernel functions in bpf_tracing_multi_attach(). However, the tracing_multi link support for bpf progs does not require CONFIG_HAVE_SINGLE_FTRACE_DIRECT_OPS. So, shrink its guard scope to add new code in bpf_tracing_multi_attach() for the tracing_multi link support for bpf progs. To simplify the change, validate CONFIG_HAVE_SINGLE_FTRACE_DIRECT_OPS only here, because the execution path in trampoline of the tracing_multi link support for kernel functions is already guarded with the CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS and CONFIG_HAVE_SINGLE_FTRACE_DIRECT_OPS. Meanwhile, provide a stub of bpf_trampoline_multi_detach() out of the CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS and CONFIG_HAVE_SINGLE_FTRACE_DIRECT_OPS guard. Signed-off-by: Leon Hwang <redacted> --- kernel/bpf/trampoline.c | 5 +++++ kernel/trace/bpf_trace.c | 15 +++------------ 2 files changed, 8 insertions(+), 12 deletions(-)
diff --git a/kernel/bpf/trampoline.c b/kernel/bpf/trampoline.c
index e57c687955fd..6e18f69a63ad 100644
--- a/kernel/bpf/trampoline.c
+++ b/kernel/bpf/trampoline.c@@ -1867,6 +1867,11 @@ void bpf_trampoline_multi_detach(struct bpf_prog *prog, #undef for_each_mnode_cnt #undef for_each_mnode +#else + +void bpf_trampoline_multi_detach(struct bpf_prog *prog, + struct bpf_tracing_multi_link *link) {} + #endif /* CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS && CONFIG_HAVE_SINGLE_FTRACE_DIRECT_OPS && CONFIG_BPF_SYSCALL */
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index 195f78db9bda..69559158ab0b 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c@@ -3679,9 +3679,6 @@ __bpf_kfunc int bpf_copy_from_user_task_str_dynptr(const struct bpf_dynptr *dptr __bpf_kfunc_end_defs(); -#if defined(CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS) && \ - defined(CONFIG_HAVE_SINGLE_FTRACE_DIRECT_OPS) - static void bpf_tracing_multi_link_release(struct bpf_link *link) { struct bpf_tracing_multi_link *tr_link =
@@ -3858,6 +3855,9 @@ int bpf_tracing_multi_attach(struct bpf_prog *prog, const union bpf_attr *attr) if (attr->link_create.flags || attr->link_create.target_fd) return -EINVAL; + if (!IS_ENABLED(CONFIG_HAVE_SINGLE_FTRACE_DIRECT_OPS)) + return -EOPNOTSUPP; + ids = kvmalloc_objs(*ids, cnt); if (!ids) return -ENOMEM;
@@ -3924,12 +3924,3 @@ int bpf_tracing_multi_attach(struct bpf_prog *prog, const union bpf_attr *attr) kvfree(link); return err; } - -#else - -int bpf_tracing_multi_attach(struct bpf_prog *prog, const union bpf_attr *attr) -{ - return -EOPNOTSUPP; -} - -#endif /* CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS && CONFIG_HAVE_SINGLE_FTRACE_DIRECT_OPS */
--
2.55.0