bpf: Expose symbol's respective address
authorYafang Shao <laoar.shao@gmail.com>
Sun, 9 Jul 2023 02:56:26 +0000 (02:56 +0000)
committerAlexei Starovoitov <ast@kernel.org>
Wed, 12 Jul 2023 03:07:51 +0000 (20:07 -0700)
Since different symbols can share the same name, it is insufficient to only
expose the symbol name. It is essential to also expose the symbol address
so that users can accurately identify which one is being probed.

Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20230709025630.3735-7-laoar.shao@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
kernel/trace/trace_kprobe.c

index 249eb14d0f7aba79884a94c6a8d669aff6b6cc80..bf2872ca5aaf6184825366e0b9fc9a91f08fbf6b 100644 (file)
@@ -1544,18 +1544,10 @@ int bpf_get_kprobe_info(const struct perf_event *event, u32 *fd_type,
 
        *fd_type = trace_kprobe_is_return(tk) ? BPF_FD_TYPE_KRETPROBE
                                              : BPF_FD_TYPE_KPROBE;
-       if (tk->symbol) {
-               *symbol = tk->symbol;
-               *probe_offset = tk->rp.kp.offset;
-               *probe_addr = 0;
-       } else {
-               *symbol = NULL;
-               *probe_offset = 0;
-               if (kallsyms_show_value(current_cred()))
-                       *probe_addr = (unsigned long)tk->rp.kp.addr;
-               else
-                       *probe_addr = 0;
-       }
+       *probe_offset = tk->rp.kp.offset;
+       *probe_addr = kallsyms_show_value(current_cred()) ?
+                     (unsigned long)tk->rp.kp.addr : 0;
+       *symbol = tk->symbol;
        return 0;
 }
 #endif /* CONFIG_PERF_EVENTS */