public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "bugdal at aerifal dot cx" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug middle-end/32667] block copy with exact overlap is expanded as memcpy
Date: Wed, 22 Nov 2023 19:07:14 +0000	[thread overview]
Message-ID: <bug-32667-4-i4ZwKuV1HW@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-32667-4@http.gcc.gnu.org/bugzilla/>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=32667

--- Comment #28 from Rich Felker <bugdal at aerifal dot cx> ---
> No, that is not a reasonable fix, because it severely pessimizes common code for a theoretical only problem.

Far less than a call to memmove (which necessarily has something comparable to
that and other unnecessary branches) pessimizes it.

I also disagree that it's severe. On basically any machine with branch
prediction, the branch will be predicted correctly all the time and has
basically zero cost. On the other hand, the branches in memmove could go
different ways depending on the caller, so it's much more
machine-capability-dependent whether they can be predicted.

In some sense the optimal thing to do is "nothing", just assuming it would be
hard to write a memcpy that fails on src==dest. However, at the very least this
precludes hardened memcpy trapping on src==dest, which might be a useful
hardening feature (or rather on a range test for overlapping, which would
happen to also catch exact overlap). So it would be nice if it were fixed.

FWIW, I don't think single branches are relevant to overall performance in
cases where the compiler is doing something reasonable by emitting a call to
memcpy to implement assignment. If the object is small enough that the branch
is relevant, the call overhead is even more of a big deal, and it should be
inlining loads/stores to perform the assignment.

  parent reply	other threads:[~2023-11-22 19:07 UTC|newest]

Thread overview: 55+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <bug-32667-4@http.gcc.gnu.org/bugzilla/>
2011-12-05 17:00 ` [Bug middle-end/32667] builtin operator= generates memcpy with overlapping memory regions lu_zero at gentoo dot org
2011-12-05 20:38 ` mans at mansr dot com
2011-12-06  8:46 ` [Bug middle-end/32667] block copy with exact overlap is expanded as memcpy rguenth at gcc dot gnu.org
2015-02-12  5:38 ` pinskia at gcc dot gnu.org
2015-02-12  9:00 ` rguenth at gcc dot gnu.org
2015-02-12  9:42 ` jakub at gcc dot gnu.org
2015-09-28 18:18 ` hugh at mimosa dot com
2021-06-09 18:50 ` public at timruffing dot de
2021-06-10  6:36 ` rguenther at suse dot de
2023-01-05  8:53 ` pinskia at gcc dot gnu.org
2023-01-05  9:19 ` nyh at math dot technion.ac.il
2023-01-06 18:48 ` lopresti at gmail dot com
2023-10-28 10:04 ` post+gcc at ralfj dot de
2023-11-21 15:57 ` bugdal at aerifal dot cx
2023-11-22  7:33 ` rguenther at suse dot de
2023-11-22 14:57 ` bugdal at aerifal dot cx
2023-11-22 15:06 ` jakub at gcc dot gnu.org
2023-11-22 19:07 ` bugdal at aerifal dot cx [this message]
2023-11-23  0:00 ` lopresti at gmail dot com
2023-11-23  7:44 ` post+gcc at ralfj dot de
2023-11-23  7:58 ` rguenth at gcc dot gnu.org
2023-11-23  8:03 ` fw at gcc dot gnu.org
2023-11-23  8:08 ` rguenther at suse dot de
2023-11-23  8:32 ` cvs-commit at gcc dot gnu.org
2023-11-23  8:39 ` jakub at gcc dot gnu.org
2023-11-23 13:54 ` bugdal at aerifal dot cx
2023-11-23 13:57 ` bugdal at aerifal dot cx
2023-11-23 14:41 ` public at timruffing dot de
2023-11-23 15:00 ` public at timruffing dot de
2023-11-23 15:01 ` sjames at gcc dot gnu.org
2023-11-23 15:29 ` post+gcc at ralfj dot de
2023-11-23 18:55 ` bugdal at aerifal dot cx
2023-11-23 19:04 ` post+gcc at ralfj dot de
2023-11-23 20:27 ` bugdal at aerifal dot cx
2023-11-24  8:19 ` public at timruffing dot de
2023-11-24  9:29 ` rguenth at gcc dot gnu.org
2023-11-24 10:09 ` jakub at gcc dot gnu.org
2023-11-24 19:08 ` post+gcc at ralfj dot de
2023-11-27 14:12 ` post+gcc at ralfj dot de
2023-11-27 23:32 ` joseph at codesourcery dot com
2023-11-28  7:17 ` post+gcc at ralfj dot de
2023-11-28  7:20 ` post+gcc at ralfj dot de
2023-11-28 10:58 ` rguenther at suse dot de
2023-11-28 11:03 ` rguenth at gcc dot gnu.org
2023-11-28 11:07 ` fw at gcc dot gnu.org
2024-01-04 12:52 ` rearnsha at gcc dot gnu.org
2024-01-04 13:22 ` bugdal at aerifal dot cx
2024-01-04 13:37 ` jakub at gcc dot gnu.org
2024-01-04 14:28 ` rearnsha at gcc dot gnu.org
2024-01-04 14:29 ` bugdal at aerifal dot cx
2024-01-04 14:41 ` rearnsha at gcc dot gnu.org
2024-01-04 14:44 ` bugdal at aerifal dot cx
2024-06-18 16:28 ` pinskia at gcc dot gnu.org
2024-06-18 16:54 ` lopresti at gmail dot com
2024-06-18 17:00 ` pinskia at gcc dot gnu.org

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=bug-32667-4-i4ZwKuV1HW@http.gcc.gnu.org/bugzilla/ \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@gcc.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.
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).