public inbox for gcc-prs@sourceware.org
help / color / mirror / Atom feed
From: Hallvard B Furuseth <h.b.furuseth@usit.uio.no>
To: gcc-gnats@gcc.gnu.org
Cc: h.b.furuseth@usit.uio.no
Subject: c/3414: Inaccurate documentation of __attribute__((malloc))
Date: Mon, 25 Jun 2001 13:36:00 -0000	[thread overview]
Message-ID: <E15Ed3R-00015d-00@bombur.uio.no> (raw)

>Number:         3414
>Category:       c
>Synopsis:       Inaccurate documentation of __attribute__((malloc))
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    unassigned
>State:          open
>Class:          doc-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Jun 25 13:36:00 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator:     Hallvard B Furuseth
>Release:        3.0
>Organization:
>Environment:
System: SunOS bombur.uio.no 5.8 Generic_108528-03 sun4u sparc SUNW,Ultra-5_10
Architecture: sun4

	
host: sparc-sun-solaris2.8
build: sparc-sun-solaris2.8
target: sparc-sun-solaris2.8
configured with: ./configure --quiet --enable-version-specific-runtime-libs --prefix=/usit/bombur/hbf --program-suffix=-3.0
>Description:
        From @item malloc, @node Function Attributes:

        "The `malloc' attribute is used to tell the compiler that a
        function may be treated as if it were the malloc function."

	The compiler could assume several things about that; e.g.
	that the return value will be aligned for any data type.
	However, the rest of the paragraph is:

	"The compiler assumes that calls to malloc result in a
	pointers that cannot alias anything.  This will often
	improve optimization."

	Is this the *only* assumption __attribute__((malloc)) causes?

	Incidentally, it also assumes that the program uses malloc
	correctly - here is a malloc which _can_ alias something:
	  free(x = malloc(10));
	  if (x == (y = malloc(10))) { ... }

	(BTW, typo: "a pointers" should be "a pointer".

	
>How-To-Repeat:
	
>Fix:
	I'm guessing that you mean something like this, but I
	have not checked.  Sorry if my English is a bit clumsy:

        The @code{malloc} attribute is used to tell the compiler
        that a function returns a pointer which cannot alias
        anything (or a null pointer), i.e. i behaves like the
        @code{malloc} function.  This will often improve
        optimization.

	Note that this may break programs that incorrectly uses
	the value of such a pointer after it has been
	invalidated (in the case of malloced pointers by calling
	@code{free()}, even if the pointer is not dereferenced.
>Release-Note:
>Audit-Trail:
>Unformatted:


             reply	other threads:[~2001-06-25 13:36 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2001-06-25 13:36 Hallvard B Furuseth [this message]
2003-02-18  2:54 bje
2003-03-27 21:04 neroden

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=E15Ed3R-00015d-00@bombur.uio.no \
    --to=h.b.furuseth@usit.uio.no \
    --cc=gcc-gnats@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).