Re: [PATCHv3 bpf-next 6/7] libbpf: allow specification of "kprobe/function+offset"

On Wed, Jul 07, 2021 at 05:14:20PM -0700, Andrii Nakryiko wrote:

On Wed, Jul 7, 2021 at 2:54 PM Jiri Olsa [off-list ref] wrote:

quoted

From: Alan Maguire <redacted>

kprobes can be placed on most instructions in a function, not
just entry, and ftrace and bpftrace support the function+offset
notification for probe placement.  Adding parsing of func_name
into func+offset to bpf_program__attach_kprobe() allows the
user to specify

SEC("kprobe/bpf_fentry_test5+0x6")

...for example, and the offset can be passed to perf_event_open_probe()
to support kprobe attachment.

Signed-off-by: Alan Maguire <redacted>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
 tools/lib/bpf/libbpf.c | 20 +++++++++++++++++---
 1 file changed, 17 insertions(+), 3 deletions(-)

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 1e04ce724240..60c9e3e77684 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c

@@ -10309,11 +10309,25 @@ struct bpf_link *bpf_program__attach_kprobe(struct bpf_program *prog,
                                            const char *func_name)

I think we should add bpf_program__attach_kprobe_opts instead for the
programmatic API instead of parsing it here from func_name. It's a
cumbersome API.

Parsing SEC() is fine, of course, but then it has to call into
bpf_program__attach_kprobe_opts() internally.

ok, Alan, will you make the change, or should I do that?

thanks,
jirka

quoted

 {
        char errmsg[STRERR_BUFSIZE];
+       char func[BPF_OBJ_NAME_LEN];
+       unsigned long offset = 0;
        struct bpf_link *link;
-       int pfd, err;
+       int pfd, err, n;
+
+       n = sscanf(func_name, "%[a-zA-Z0-9_.]+%lx", func, &offset);
+       if (n < 1) {
+               err = -EINVAL;
+               pr_warn("kprobe name is invalid: %s\n", func_name);
+               return libbpf_err_ptr(err);
+       }
+       if (retprobe && offset != 0) {
+               err = -EINVAL;
+               pr_warn("kretprobes do not support offset specification\n");
+               return libbpf_err_ptr(err);
+       }

-       pfd = perf_event_open_probe(false /* uprobe */, retprobe, func_name,
-                                   0 /* offset */, -1 /* pid */);
+       pfd = perf_event_open_probe(false /* uprobe */, retprobe, func,
+                                   offset, -1 /* pid */);
        if (pfd < 0) {
                pr_warn("prog '%s': failed to create %s '%s' perf event: %s\n",
                        prog->name, retprobe ? "kretprobe" : "kprobe", func_name,
--
2.31.1