unofficial mirror of emacs-devel@gnu.org 
 help / color / mirror / code / Atom feed
From: Eli Zaretskii <eliz@gnu.org>
To: Alex Gramiak <agrambot@gmail.com>
Cc: eggert@cs.ucla.edu, emacs-devel@gnu.org
Subject: Re: Using __builtin_expect (likely/unlikely macros)
Date: Tue, 16 Apr 2019 18:23:08 +0300	[thread overview]
Message-ID: <83bm16gejn.fsf@gnu.org> (raw)
In-Reply-To: <871s22jyyo.fsf@gmail.com> (message from Alex Gramiak on Mon, 15 Apr 2019 23:33:51 -0600)

> From: Alex Gramiak <agrambot@gmail.com>
> Cc: eggert@cs.ucla.edu,  emacs-devel@gnu.org
> Date: Mon, 15 Apr 2019 23:33:51 -0600
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > I'm not sure whether you put LIKELY and UNLIKELY somewhat randomly,
> > just for testing, or did you really think each place is
> > likely/unlikely as in the change, but at least some places in xdisp.c
> > are wrong: they use LIKELY where UNLIKELY is more appropriate, abd
> > vice versa.
> 
> Out of curiosity, which places would that be? I only put UNLIKELY calls
> around the checks for emacs_abort (which I presume to be unlikely), and
> the LIKELY calls are of the form:
> 
>   else if (LIKELY (<cond>))
>     {
>       ...
>     }
>   else
>     emacs_abort ();

OK, it's possible that I don't understand the exact semantics of
__builtin_expect in these situations.

The problem is that you used LIKELY like this:

  if (A)
    do_A;
  else if (B)
    do_B;
  else if (C)
    do_C;
  else if (LIKELY (D))
    do_D;
  else
    cant_happen ();

Essentially, the above is a moral equivalent of a 'switch' with the
'default' case aborting because it "cannot happen".  In such code, the
order of the clauses doesn't necessarily tell anything about their
likelihood; up front, they all are equally "likely".  So using LIKELY
only in the last one sends a wrong signal: that last condition is
neither more nor less likely than all the others.  Actually, in some
cases it might be _less_ likely than the preceding ones, because if I
knew that some of these conditions happens much more frequently, I'd
test it first.

Now, it's possible that the effect of __builtin_expect doesn't care
about this issue.  The GCC manual doesn't help to figure out whether
this is the case, because it only talks about a simple case of a
single 'if' clause, and doesn't tell any details about what GCC is
allowed to do when it sees __builtin_expect.  But just by looking at
how the code looks, I immediately raised a brow.



  reply	other threads:[~2019-04-16 15:23 UTC|newest]

Thread overview: 45+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-15  0:15 Using __builtin_expect (likely/unlikely macros) Alex Gramiak
2019-04-15  1:18 ` Paul Eggert
2019-04-15  3:11   ` Alex Gramiak
2019-04-15  4:41     ` Paul Eggert
2019-04-16  0:16       ` Alex Gramiak
2019-04-16  2:34         ` Eli Zaretskii
2019-04-16  5:33           ` Alex Gramiak
2019-04-16 15:23             ` Eli Zaretskii [this message]
2019-04-16 15:47               ` Alex Gramiak
2019-04-16  3:42         ` Paul Eggert
2019-04-16 13:05           ` Stefan Monnier
2019-04-16 15:22             ` Paul Eggert
2019-04-16 16:10               ` Alex Gramiak
2019-04-16 17:54                 ` Paul Eggert
2019-04-16 20:50                   ` Alex Gramiak
2019-04-16 21:11                     ` Alex Gramiak
2019-04-16 21:27                     ` Stefan Monnier
2019-04-16 21:27                     ` Konstantin Kharlamov
2019-04-18  8:25                       ` Paul Eggert
2019-04-18  8:43                         ` Konstantin Kharlamov
2019-04-18 13:47                         ` Andy Moreton
2019-04-18 17:27                           ` Paul Eggert
2019-04-18 17:56                             ` Andy Moreton
2019-04-18 19:32                               ` Paul Eggert
2019-04-19 13:45                         ` Alex Gramiak
2019-04-19 13:58                           ` Konstantin Kharlamov
2019-04-19 14:45                             ` Alex Gramiak
2019-04-19 17:33                           ` Paul Eggert
2019-04-19 20:53                             ` Alex Gramiak
2019-04-20  0:05                               ` Alan Mackenzie
2019-04-20  0:42                                 ` Paul Eggert
2019-04-20 19:46                                   ` Alan Mackenzie
2019-04-20 15:29                             ` Andy Moreton
2019-04-20 15:57                               ` Paul Eggert
2019-04-20 16:03                                 ` Eli Zaretskii
2019-04-20 16:11                                   ` Paul Eggert
2019-04-20 16:18                                     ` Eli Zaretskii
2019-04-20 16:57                                       ` Paul Eggert
2019-04-20 17:22                                         ` Eli Zaretskii
2019-04-20 19:13                                           ` Paul Eggert
2019-04-20 16:28                                 ` Óscar Fuentes
2019-04-20 18:58                                   ` Paul Eggert
2019-04-20 19:35                                     ` Óscar Fuentes
2019-04-20 22:54                                       ` Paul Eggert
2020-04-15  3:14   ` John Carter

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

  List information: https://www.gnu.org/software/emacs/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=83bm16gejn.fsf@gnu.org \
    --to=eliz@gnu.org \
    --cc=agrambot@gmail.com \
    --cc=eggert@cs.ucla.edu \
    --cc=emacs-devel@gnu.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.
Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/emacs.git

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 read-only IMAP folder(s) and NNTP newsgroup(s).