public inbox for libc-ports@sourceware.org
 help / color / mirror / Atom feed
From: Thomas Schwinge <thomas@codesourcery.com>
To: Carlos O'Donell <carlos@systemhalted.org>,
	Richard Henderson	<rth@twiddle.net>
Cc: <libc-alpha@sourceware.org>, <libc-ports@sourceware.org>,
	<joseph@codesourcery.com>
Subject: Re: Usage of __attribute__used__ in (system) headers?
Date: Wed, 20 Feb 2013 16:06:00 -0000	[thread overview]
Message-ID: <87vc9nc6ma.fsf@schwinge.name> (raw)
In-Reply-To: <CAE2sS1j8PQ=+LNCihphwr-RbxfNkRtB3xfOz9DmaQhp5x7kdOQ@mail.gmail.com>

[-- Attachment #1: Type: text/plain, Size: 5230 bytes --]

Hi!

On Mon, 18 Feb 2013 14:00:03 -0500, Carlos O'Donell <carlos@systemhalted.org> wrote:
> I think you should update the MIPS nan.h version to
> match the generic version [...]

Yes, I already had prepared such a patch and planned to commit it "as
obvious", without discussion.  Thusly pushed in commit
2636ffe65438af689e12b7977fe8609a6ca07c90:

ports/
	* sysdeps/mips/bits/nan.h: Align to generic IEEE 754 file.

diff --git ports/sysdeps/mips/bits/nan.h ports/sysdeps/mips/bits/nan.h
index ffbb3b5..af168ce 100644
--- ports/sysdeps/mips/bits/nan.h
+++ ports/sysdeps/mips/bits/nan.h
@@ -1,4 +1,4 @@
-/* `NAN' constant for IEEE 754 machines.
+/* `NAN' constant for IEEE 754 machines.  MIPS version.
    Copyright (C) 1992-2013 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
 
@@ -13,7 +13,7 @@
    Lesser General Public License for more details.
 
    You should have received a copy of the GNU Lesser General Public
-   License along with the GNU C Library.  If not, see
+   License along with the GNU C Library; if not, see
    <http://www.gnu.org/licenses/>.  */
 
 #ifndef _MATH_H
@@ -21,20 +21,21 @@
 #endif
 
 
-/* IEEE Not A Number (QNaN). Note that MIPS has the QNaN and SNaN patterns
-   reversed compared to most other architectures. The IEEE spec left
-   the definition of this open to implementations, and for MIPS the top
-   bit of the mantissa must be SET to indicate a SNaN.  */
+/* IEEE Not A Number.  */
+/* Note that MIPS has the QNaN and SNaN patterns reversed compared to most
+   other architectures.  The IEEE spec left the definition of this open to
+   implementations, and for MIPS the top bit of the mantissa must be SET to
+   indicate a SNaN.  */
 
 #if __GNUC_PREREQ(3,3)
 
-# define NAN	(__builtin_nanf(""))
+# define NAN	(__builtin_nanf (""))
 
 #elif defined __GNUC__
 
 # define NAN \
-  (__extension__                                                            \
-   ((union { unsigned __l __attribute__((__mode__(__SI__))); float __d; })  \
+  (__extension__							      \
+   ((union { unsigned __l __attribute__ ((__mode__ (__SI__))); float __d; })  \
     { __l: 0x7fbfffffUL }).__d)
 
 #else

I also added
<http://sourceware.org/glibc/wiki/Development_Todo/Master?action=diff&rev2=49&rev1=48>.

On Mon, 18 Feb 2013 14:41:02 -0500, Carlos O'Donell <carlos@systemhalted.org> wrote:
> On Mon, Feb 18, 2013 at 2:21 PM, Richard Henderson <rth@twiddle.net> wrote:
> > On 02/18/2013 11:00 AM, Carlos O'Donell wrote:
> >>> >  static union { unsigned char __c[4]; float __d; } __nan_union
> >>> > -    __attribute_used__ = { __nan_bytes };
> >>> > +  = { __nan_bytes };
> >>> >  # define NAN   (__nan_union.__d)
> >>> >
> >>> >  #endif /* GCC.  */

In commit 72f0ffdcbeb8135d04cf2dc73f8a5f5c7783a283, I first add the
missing __attribute_used__:

ports/
	* sysdeps/mips/bits/nan.h [!__GNUC__] (__nan_union): Add
	__attribute_used__.

diff --git ports/sysdeps/mips/bits/nan.h ports/sysdeps/mips/bits/nan.h
index af168ce..71f372d 100644
--- ports/sysdeps/mips/bits/nan.h
+++ ports/sysdeps/mips/bits/nan.h
@@ -49,7 +49,8 @@
 #  define __nan_bytes		{ 0xff, 0xff, 0xbf, 0x7f }
 # endif
 
-static union { unsigned char __c[4]; float __d; } __nan_union = { __nan_bytes };
+static union { unsigned char __c[4]; float __d; } __nan_union
+    __attribute_used__ = { __nan_bytes };
 # define NAN	(__nan_union.__d)
 
 #endif	/* GCC.  */

> >> I disagree, it's useful to mark the non-GCC version with
> >> an __attribute_used__ such that in the future we might
> >> use another compiler without problems.
> >
> > The annotation should not be attribute used at all, but rather unused.

Good catch, thanks!

> glibc doesn't have an __attribute_unused__.

It doesn't need one; per <sys/cdefs.h>'s __attribute_used__ definition,
we assume that any compiler that supports __attribute__ at all  also
supports __attribute__ ((unused)).  So, I see no reason to introduce
__attribute_unused__ bit instead directly use __attribute__ ((unused));
commit c7b275d6b3bceb6b400fa3044d13d1001bc605ca:

	* sysdeps/ieee754/bits/nan.h [!__GNUC__] (__nan_union): Change
	__attribute_used__ to __attribute__ ((unused)).

ports/
	* sysdeps/mips/bits/nan.h [!__GNUC__] (__nan_union): Change
	__attribute_used__ to __attribute__ ((unused)).

diff --git ports/sysdeps/mips/bits/nan.h ports/sysdeps/mips/bits/nan.h
index 71f372d..8f4666d 100644
--- ports/sysdeps/mips/bits/nan.h
+++ ports/sysdeps/mips/bits/nan.h
@@ -50,7 +50,7 @@
 # endif
 
 static union { unsigned char __c[4]; float __d; } __nan_union
-    __attribute_used__ = { __nan_bytes };
+  __attribute__ ((unused)) = { __nan_bytes };
 # define NAN	(__nan_union.__d)
 
 #endif	/* GCC.  */
diff --git sysdeps/ieee754/bits/nan.h sysdeps/ieee754/bits/nan.h
index d3ab38b..a1e6a51 100644
--- sysdeps/ieee754/bits/nan.h
+++ sysdeps/ieee754/bits/nan.h
@@ -46,7 +46,7 @@
 # endif
 
 static union { unsigned char __c[4]; float __d; } __nan_union
-    __attribute_used__ = { __nan_bytes };
+  __attribute__ ((unused)) = { __nan_bytes };
 # define NAN	(__nan_union.__d)
 
 #endif	/* GCC.  */


Grüße,
 Thomas

[-- Attachment #2: Type: application/pgp-signature, Size: 489 bytes --]

  reply	other threads:[~2013-02-20 16:06 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <87r4khnbqm.fsf@kepler.schwinge.homeip.net>
2013-02-18 19:00 ` Carlos O'Donell
2013-02-18 19:21   ` Richard Henderson
2013-02-18 19:41     ` Carlos O'Donell
2013-02-20 16:06       ` Thomas Schwinge [this message]
2013-02-20 16:25         ` Carlos O'Donell
2013-02-20 17:56           ` Thomas Schwinge
2013-02-20 18:02             ` Carlos O'Donell

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=87vc9nc6ma.fsf@schwinge.name \
    --to=thomas@codesourcery.com \
    --cc=carlos@systemhalted.org \
    --cc=joseph@codesourcery.com \
    --cc=libc-alpha@sourceware.org \
    --cc=libc-ports@sourceware.org \
    --cc=rth@twiddle.net \
    /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).