From: Tzvetomir Stoyanov <tz.stoyanov@gmail.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: linux-trace-devel@vger.kernel.org
Subject: Re: [PATCH] trace-cmd: Optimize pretty_print() function
Date: Fri, 15 May 2020 07:53:22 +0300 [thread overview]
Message-ID: <CAPpZLN6g3CmBngdH==3yjTdoyT5th2M-Cd=xU4FBMQ534xYM+g@mail.gmail.com> (raw)
In-Reply-To: <20200514211410.0362161b@oasis.local.home>
On Fri, May 15, 2020 at 4:14 AM Steven Rostedt <rostedt@goodmis.org> wrote:
>
[...]
> > +static int parse_arg_format(struct tep_print_parse **parse,
> > + struct tep_event *event,
> > + const char *format, struct tep_print_arg
> > **arg) +{
> > + struct tep_print_arg *len_arg = NULL;
> > + char print_format[32];
> > + const char *start = format;
> > + int ret = 0;
> > + int ls = 0;
> > + int res;
> > + int len;
> > +
> > + if (*format != '%')
> > + return ret;
> > + format++;
> > +
> > + /* %% */
> > + if (*format == '%')
> > + return ret;
>
> The above two checks for '%' isn't needed as it is checked by the
> caller. If for some reason that the caller didn't catch it, and we
> return here, it would return zero and the calling loop would never
> proceed. It's not protecting anything, I would just remove these two
> checks.
>
> > + ret++;
> > +
> > + for (; *format; format++) {
> > + switch (*format) {
> > + case '#':
> > + /* FIXME: need to handle properly */
> > + break;
> > + case 'h':
> > + ls--;
> > + break;
> > + case 'l':
> > + ls++;
> > + break;
> > + case 'L':
> > + ls = 2;
> > + break;
> > + case '.':
> > + case 'z':
> > + case 'Z':
> > + case '0' ... '9':
> > + case '-':
> > + break;
> > + case '*':
> > + /* The argument is the length. */
> > + if (!arg) {
> > + do_warning_event(event, "no argument match");
> > + event->flags |= TEP_EVENT_FL_FAILED;
> > + goto out_failed;
> > + }
>
> Since all cases to exit this loop do the same check, I think we can
> move this check before the loop. The old code didn't do this check
> properly for %p, but that was a bug in the old code.
>
We still have to check for a valid argument inside the switch for almost all
exit cases, as the '*' case can consume the argument that would be verified
before the loop. And there is one exit that does not require a valid argument -
the default case. The old code just prints the unknown character, without
consuming an argument.
>
> > + len_arg = *arg;
>
[...]
> -- Steve
Thanks, Steve!
--
Tzvetomir (Ceco) Stoyanov
VMware Open Source Technology Center
prev parent reply other threads:[~2020-05-15 4:53 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-14 7:12 [PATCH] trace-cmd: Optimize pretty_print() function Tzvetomir Stoyanov (VMware)
2020-05-15 1:14 ` Steven Rostedt
2020-05-15 4:53 ` Tzvetomir Stoyanov [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAPpZLN6g3CmBngdH==3yjTdoyT5th2M-Cd=xU4FBMQ534xYM+g@mail.gmail.com' \
--to=tz.stoyanov@gmail.com \
--cc=linux-trace-devel@vger.kernel.org \
--cc=rostedt@goodmis.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).