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 --]
next prev parent 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).