Messages in this thread |  | | From | Andrii Nakryiko <> | | Date | Tue, 31 Mar 2026 16:05:14 -0700 | | Subject | Re: [PATCH bpf-next v4 1/2] libbpf: clarify raw-address single kprobe attach behavior |
| |
On Tue, Mar 31, 2026 at 8:25 AM Hoyeon Lee <hoyeon.lee@suse.com> wrote: > > bpf_program__attach_kprobe_opts() documents single-kprobe attach > through func_name, with an optional offset. For the PMU-based > non-legacy path, func_name = NULL with an absolute address in offset > already works as well, but that is not described in the API. > > This commit clarifies this existing non-legacy behavior. For PMU-based > attach, callers can use func_name = NULL with an absolute address in > offset as the raw-address form. For legacy tracefs/debugfs kprobes, > reject this form explicitly. > > Signed-off-by: Hoyeon Lee <hoyeon.lee@suse.com> > --- > tools/lib/bpf/libbpf.c | 21 ++++++++++++--------- > tools/lib/bpf/libbpf.h | 3 ++- > 2 files changed, 14 insertions(+), 10 deletions(-) > > diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c > index 9ea41f40dc82..9083f542a3b0 100644 > --- a/tools/lib/bpf/libbpf.c > +++ b/tools/lib/bpf/libbpf.c > @@ -11523,7 +11523,8 @@ static int determine_uprobe_retprobe_bit(void) > #define PERF_UPROBE_REF_CTR_OFFSET_SHIFT 32 > > static int perf_event_open_probe(bool uprobe, bool retprobe, const char *name, > - uint64_t offset, int pid, size_t ref_ctr_off) > + uint64_t offset_or_addr, int pid,
don't rename it, keep it as "offset"
> + size_t ref_ctr_off) > { > const size_t attr_sz = sizeof(struct perf_event_attr); > struct perf_event_attr attr; > @@ -11558,7 +11559,7 @@ static int perf_event_open_probe(bool uprobe, bool retprobe, const char *name, > attr.type = type; > attr.config |= (__u64)ref_ctr_off << PERF_UPROBE_REF_CTR_OFFSET_SHIFT; > attr.config1 = ptr_to_u64(name); /* kprobe_func or uprobe_path */ > - attr.config2 = offset; /* kprobe_addr or probe_offset */ > + attr.config2 = offset_or_addr; /* kprobe_addr or probe_offset */ > > /* pid filter is meaningful only for uprobes */ > pfd = syscall(__NR_perf_event_open, &attr, > @@ -11816,6 +11817,8 @@ bpf_program__attach_kprobe_opts(const struct bpf_program *prog, > default: > return libbpf_err_ptr(-EINVAL); > } > + if (!func_name && legacy) > + return libbpf_err_ptr(-ENOTSUP);
-EOPNOTSUPP
pw-bot: cr
> > if (!legacy) { > pfd = perf_event_open_probe(false /* uprobe */, retprobe, > @@ -11835,21 +11838,21 @@ bpf_program__attach_kprobe_opts(const struct bpf_program *prog, > offset, -1 /* pid */); > } > if (pfd < 0) { > - err = -errno; > - pr_warn("prog '%s': failed to create %s '%s+0x%zx' perf event: %s\n", > + err = pfd; > + pr_warn("prog '%s': failed to create %s '%s%s0x%zx' perf event: %s\n", > prog->name, retprobe ? "kretprobe" : "kprobe", > - func_name, offset, > - errstr(err)); > + func_name ?: "", func_name ? "+" : "", > + offset, errstr(err)); > goto err_out; > } > link = bpf_program__attach_perf_event_opts(prog, pfd, &pe_opts); > err = libbpf_get_error(link); > if (err) { > close(pfd); > - pr_warn("prog '%s': failed to attach to %s '%s+0x%zx': %s\n", > + pr_warn("prog '%s': failed to attach to %s '%s%s0x%zx': %s\n", > prog->name, retprobe ? "kretprobe" : "kprobe", > - func_name, offset, > - errstr(err)); > + func_name ?: "", func_name ? "+" : "", > + offset, errstr(err)); > goto err_clean_legacy; > } > if (legacy) { > diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h > index 0be34852350f..f75f3ab0f20c 100644 > --- a/tools/lib/bpf/libbpf.h > +++ b/tools/lib/bpf/libbpf.h > @@ -557,7 +557,7 @@ struct bpf_kprobe_opts { > size_t sz; > /* custom user-provided value fetchable through bpf_get_attach_cookie() */ > __u64 bpf_cookie; > - /* function's offset to install kprobe to */ > + /* function offset, or raw address if func_name == NULL (non-legacy) */
legacy mode will be useful on very-very old kernels only, let's not distract with constant mentions for "non-legacy", please remove "(non-legacy)" here and other legacy mentions above
but also, while you are at it, let's add a proper doc comment to bpf_program__attach_kprobe and bpf_program__attach_kprobe_opts APIs? That seems like a more obvious place to explain name+offset and offset as absolute address behavior, IMO
> size_t offset; > /* kprobe is return probe */ > bool retprobe; > @@ -565,6 +565,7 @@ struct bpf_kprobe_opts { > enum probe_attach_mode attach_mode; > size_t :0; > }; > + > #define bpf_kprobe_opts__last_field attach_mode > > LIBBPF_API struct bpf_link * > -- > 2.52.0 >
|  |