Re: [PATCH tracing/kprobes v4] perf: Add perf probe subcommand for kprobe-event setup helper

["Frédéric Weisbecker" <fweisbec@gmail.com>]

2009/10/8 Masami Hiramatsu <mhiramat@redhat.com>:
> Add perf probe subcommand for kprobe-event setup helper to perf command.
> This allows user to define kprobe events by C expressions (C line numbers,
> C function names, and C local variables).
>
> Usage
> -----
>  perf probe [<options>] -P 'PROBEDEF' [-P 'PROBEDEF' ...]
>
>    -k, --vmlinux <file>  vmlinux/module pathname
>    -P, --probe <p|r:[GRP/]NAME FUNC[+OFFS][@SRC]|@SRC:LINE [ARG ...]>
>                          probe point definition, where
>                p:      kprobe probe
>                r:      kretprobe probe
>                GRP:    Group name (optional)
>                NAME:   Event name
>                FUNC:   Function name
>                OFFS:   Offset from function entry (in byte)
>                SRC:    Source code path
>                LINE:   Line number
>                ARG:    Probe argument (local variable name or
>                        kprobe-tracer argument format is supported.)
>
> Changes in v4:
>  - Add _GNU_SOURCE macro for strndup().
>
> Changes in v3:
>  - Remove -r option because perf always be used for online kernel.
>  - Check malloc/calloc results.
>
> Changes in v2:
>  - Check synthesized string length.
>  - Rename perf kprobe to perf probe.
>  - Use spaces for separator and update usage comment.
>  - Check error paths in parse_probepoint().
>  - Check optimized-out variables.
>
> Signed-off-by: Masami Hiramatsu <mhiramat@redhat.com>
> Cc: Frederic Weisbecker <fweisbec@gmail.com>
> Cc: Ingo Molnar <mingo@elte.hu>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
> Cc: Steven Rostedt <rostedt@goodmis.org>
> Cc: Mike Galbraith <efault@gmx.de>
> Cc: Paul Mackerras <paulus@samba.org>
> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
> Cc: Christoph Hellwig <hch@infradead.org>
> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
> Cc: Jim Keniston <jkenisto@us.ibm.com>
> Cc: Frank Ch. Eigler <fche@redhat.com>
> ---
>
[...]
> +/* Default vmlinux search paths */
> +#define NR_SEARCH_PATH 3
> +const char *default_search_path[NR_SEARCH_PATH] = {
> +"/lib/modules/%s/build/vmlinux",               /* Custom build kernel */
> +"/usr/lib/debug/lib/modules/%s/vmlinux",       /* Red Hat debuginfo */
> +"/boot/vmlinux-debug-%s",                      /* Ubuntu */
> +};
[...]
> +static int open_default_vmlinux(void)
> +{
> +       struct utsname uts;
> +       char fname[MAX_PATH_LEN];
> +       int fd, ret, i;
> +
> +       ret = uname(&uts);
> +       if (ret) {
> +               debug("uname() failed.\n");
> +               return -errno;
> +       }
> +       session.release = uts.release;
> +       for (i = 0; i < NR_SEARCH_PATH; i++) {
> +               ret = snprintf(fname, MAX_PATH_LEN,
> +                              default_search_path[i], session.release);
> +               if (ret >= MAX_PATH_LEN || ret < 0) {
> +                       debug("Filename(%d,%s) is too long.\n", i, uts.release);
> +                       errno = E2BIG;
> +                       return -E2BIG;
> +               }
> +               debug("try to open %s\n", fname);
> +               fd = open(fname, O_RDONLY);
> +               if (fd >= 0)
> +                       break;
> +       }
> +       return fd;
> +}


We have a kind of kernel path finder already inside perf. It might be
encapsulated
inside the load_kernel() helper, I don't remember exactly.

It would be better to make use of such centralized and already
existing facility.

The patchset looks good. I'll apply and push it soon.

Thanks a lot!