Re: [PATCH 01/25] ldv/unwind: improve stacktrace feature(v3)

[Luca C. <luc...@gm...>]

On Sat, Nov 9, 2013 at 10:41 PM, Masatake YAMATO <ya...@re...> wrote:
> Based on the discussion on strace-devel, I improved following points:
>

This patch does not apply cleanly on the ldv/unwind branch of Denys,
he amended a commit on the top of the old  ldv/unwind, so you need to
fetch and git reset --hard remotes/origin/ldv/unwind.

The conflicts are trivial.

And the patch series works well.
Just two minor comments.

> +
> +static void
> +stacktrace_capture(struct tcb *tcp)
> +{
> +       stacktrace_walk(tcp, queue_put_call, queue_put_error,
> +                       tcp->queue);
> +}

stacktrace_capture is called only once, is there any reason why it
can't be removed?

> +
> +
> +static void
> +print_call_cb(void *dummy,
> +             char *binary_filename,
> +             char *symbol_name,
> +             unw_word_t function_off_set,
> +             unsigned long true_offset)
> +{
> +       print_call(binary_filename,
> +                  symbol_name,
> +                  function_off_set,
> +                  true_offset);
> +}
> +
> +static void
> +print_error_cb(void *dummy,
> +              const char *error,
> +              unsigned long true_offset)
> +{
> +       print_error(error, true_offset);
> +}
> +
> +static void
> +stacktrace_print(struct tcb *tcp)
> +{
> +       DPRINTF("tcp=%p, queue=%p", "stackprint", tcp, tcp->queue->head);
> +       stacktrace_walk(tcp, print_call_cb, print_error_cb, NULL);
> +}
> +

stacktrace_print ditto...

I'm just looking if the patch can be simplified a little.

Luca