avoid ifunc warnings in libatomic configure

Message ID	e0871a5b-8f69-7acd-a718-bd5bcd593289@gmail.com
State	New
Headers	show Return-Path: <gcc-patches-return-463371-incoming=patchwork.ozlabs.org@gcc.gnu.org> DomainKey-Signature: a=rsa-sha1; c=nofws; d=gcc.gnu.org; h=list-id :list-unsubscribe:list-archive:list-post:list-help:sender:to:cc :from:subject:message-id:date:mime-version:content-type; q=dns; s=default; b=ba3pmP2OkVBjv+JtkhbeGf5bb5zfd2lGGvn3UeRcVK61ZGgKyH X4RWkB0Cv5X50LeAPnSL+g5AVq0IusGP5aVWKrhO5Iv1nQ75aj9ffCPhnc1RqlC6 B6KaDSY8TCS/X+ONf4r5OqUMpPwbRPszKuBMFqVHjReJU8ZLMfl1hi6SI= Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk Sender: gcc-patches-owner@gcc.gnu.org To: Joseph Myers <joseph@codesourcery.com>, Steve Ellcey <sellcey@cavium.com> Cc: Gcc Patch List <gcc-patches@gcc.gnu.org> From: Martin Sebor <msebor@gmail.com> Subject: [PATCH] avoid ifunc warnings in libatomic configure Message-ID: <e0871a5b-8f69-7acd-a718-bd5bcd593289@gmail.com> Date: Mon, 2 Oct 2017 17:00:22 -0600 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.8.0 MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="------------94E9934D5A96411AC8FDF1E1"
Series	avoid ifunc warnings in libatomic configure \| expand avoid ifunc warnings in libatomic configure

Message ID

e0871a5b-8f69-7acd-a718-bd5bcd593289@gmail.com

State

New

Headers

DomainKey-Signature: a=rsa-sha1; c=nofws; d=gcc.gnu.org; h=list-id
	:list-unsubscribe:list-archive:list-post:list-help:sender:to:cc
	:from:subject:message-id:date:mime-version:content-type; q=dns;
	s=default; b=ba3pmP2OkVBjv+JtkhbeGf5bb5zfd2lGGvn3UeRcVK61ZGgKyH
	X4RWkB0Cv5X50LeAPnSL+g5AVq0IusGP5aVWKrhO5Iv1nQ75aj9ffCPhnc1RqlC6
	B6KaDSY8TCS/X+ONf4r5OqUMpPwbRPszKuBMFqVHjReJU8ZLMfl1hi6SI=
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
Sender: gcc-patches-owner@gcc.gnu.org
To: Joseph Myers <joseph@codesourcery.com>, Steve Ellcey <sellcey@cavium.com>
Cc: Gcc Patch List <gcc-patches@gcc.gnu.org>
From: Martin Sebor <msebor@gmail.com>
Subject: [PATCH] avoid ifunc warnings in libatomic configure
Message-ID: <e0871a5b-8f69-7acd-a718-bd5bcd593289@gmail.com>
Date: Mon, 2 Oct 2017 17:00:22 -0600
User-Agent: Mozilla/5.0 (X11; Linux x86_64;
	rv:45.0) Gecko/20100101 Thunderbird/45.8.0
MIME-Version: 1.0
Content-Type: multipart/mixed;
	boundary="------------94E9934D5A96411AC8FDF1E1"

Series

avoid ifunc warnings in libatomic configure | expand

Commit Message

Martin Sebor Oct. 2, 2017, 11 p.m. UTC

The attached patch tweaks the configuration check for attribute
ifunc support in libatomic to avoid the new type safety warnings.
It also changes the libatomic declarations of ifunc resolvers to
avoid the new warnings.  With it conftest.c compiles successfully
(see below) as does libatomic, and no there are regressions in
tests on x86_64.  Does it look good to check in?

I haven't been paying close attention so if Steve or someone else
already submitted a patch with the libatomic_i.h changes I can take
mine out.

Martin

configure:14689: checking whether the target supports the ifunc attribute
configure:14712: /ssd/build/gcc-git-maint/./gcc/xgcc 
-B/ssd/build/gcc-git-maint/./gcc/ -B/usr/local/x86_64-pc-linux-gnu/bin/ 
-B/usr/local/x86_64-pc-linux-gnu/lib/ -isystem 
/usr/local/x86_64-pc-linux-gnu/include -isystem 
/usr/local/x86_64-pc-linux-gnu/sys-include  -m32 -o conftest -O2 -O0 -g3 
  -pthread  -pthread -Werror   conftest.c  >&5
configure:14712: $? = 0
configure:14720: result: yes

Comments

Joseph Myers Oct. 2, 2017, 11:26 p.m. UTC | #1

On Mon, 2 Oct 2017, Martin Sebor wrote:

> The attached patch tweaks the configuration check for attribute
> ifunc support in libatomic to avoid the new type safety warnings.
> It also changes the libatomic declarations of ifunc resolvers to
> avoid the new warnings.  With it conftest.c compiles successfully
> (see below) as does libatomic, and no there are regressions in
> tests on x86_64.  Does it look good to check in?

OK.

Steve Ellcey Oct. 2, 2017, 11:29 p.m. UTC | #2

On Mon, 2017-10-02 at 17:00 -0600, Martin Sebor wrote:
> The attached patch tweaks the configuration check for attribute
> ifunc support in libatomic to avoid the new type safety warnings.
> It also changes the libatomic declarations of ifunc resolvers to
> avoid the new warnings.  With it conftest.c compiles successfully
> (see below) as does libatomic, and no there are regressions in
> tests on x86_64.  Does it look good to check in?
> 
> I haven't been paying close attention so if Steve or someone else
> already submitted a patch with the libatomic_i.h changes I can take
> mine out.
> 
> Martin

I haven't submitted anything for this problem.  This patch fixed things
for my aarch64 build.  I definitely needed the libatomic_i.h changes as
well as the configure change.

Steve Ellcey
sellcey@cavium.com

Martin Sebor Oct. 2, 2017, 11:59 p.m. UTC | #3

On 10/02/2017 05:29 PM, Steve Ellcey wrote:
> On Mon, 2017-10-02 at 17:00 -0600, Martin Sebor wrote:
>> The attached patch tweaks the configuration check for attribute
>> ifunc support in libatomic to avoid the new type safety warnings.
>> It also changes the libatomic declarations of ifunc resolvers to
>> avoid the new warnings.  With it conftest.c compiles successfully
>> (see below) as does libatomic, and no there are regressions in
>> tests on x86_64.  Does it look good to check in?
>>
>> I haven't been paying close attention so if Steve or someone else
>> already submitted a patch with the libatomic_i.h changes I can take
>> mine out.
>>
>> Martin
>
> I haven't submitted anything for this problem.  This patch fixed things
> for my aarch64 build.  I definitely needed the libatomic_i.h changes as
> well as the configure change.

Great, thank you for verifying it there!  I just committed it
in r253372.

Martin

PR c/81854 -  weak alias of an incompatible symbol accepted

libatomic/ChangeLog:

2017-10-02  Martin Sebor  <msebor@redhat.com>

	PR c/81854
	* acinclude.m4 (LIBAT_CHECK_IFUNC): Have ifunc resolver return
	a function pointer rather than void* to avoid GCC 8 warnings.
	* configure: Regenerate.
	* libatomic_i.h: Declare ifunc resolvers to return function
	pointers rather than void*.

diff --git a/libatomic/acinclude.m4 b/libatomic/acinclude.m4
index 485d731..383218f 100644
--- a/libatomic/acinclude.m4
+++ b/libatomic/acinclude.m4
@@ -195,7 +195,8 @@  AC_DEFUN([LIBAT_CHECK_IFUNC], [
   CFLAGS="$CFLAGS -Werror"
   AC_TRY_LINK([
     int foo_alt(void) { return 0; }
-    void *foo_sel(void) { return foo_alt; }
+    typedef int F (void);
+    F *foo_sel(void) { return foo_alt; }
     int foo(void) __attribute__((ifunc("foo_sel")));],
     [return foo();], libat_cv_have_ifunc=yes, libat_cv_have_ifunc=no)])
   LIBAT_DEFINE_YESNO([HAVE_IFUNC], [$libat_cv_have_ifunc],
diff --git a/libatomic/configure b/libatomic/configure
index c05fc9d..e88a7b8 100755
--- a/libatomic/configure
+++ b/libatomic/configure
@@ -12333,6 +12333,7 @@  _ACEOF
 
 
 
+
   { $as_echo "$as_me:${as_lineno-$LINENO}: checking for __atomic_load/store for size 2" >&5
 $as_echo_n "checking for __atomic_load/store for size 2... " >&6; }
 if test "${libat_cv_have_at_ldst_2+set}" = set; then :
@@ -12400,6 +12401,7 @@  _ACEOF
 
 
 
+
   { $as_echo "$as_me:${as_lineno-$LINENO}: checking for __atomic_load/store for size 4" >&5
 $as_echo_n "checking for __atomic_load/store for size 4... " >&6; }
 if test "${libat_cv_have_at_ldst_4+set}" = set; then :
@@ -12467,6 +12469,7 @@  _ACEOF
 
 
 
+
   { $as_echo "$as_me:${as_lineno-$LINENO}: checking for __atomic_load/store for size 8" >&5
 $as_echo_n "checking for __atomic_load/store for size 8... " >&6; }
 if test "${libat_cv_have_at_ldst_8+set}" = set; then :
@@ -12534,6 +12537,7 @@  _ACEOF
 
 
 
+
   { $as_echo "$as_me:${as_lineno-$LINENO}: checking for __atomic_load/store for size 16" >&5
 $as_echo_n "checking for __atomic_load/store for size 16... " >&6; }
 if test "${libat_cv_have_at_ldst_16+set}" = set; then :
@@ -12602,6 +12606,7 @@  _ACEOF
 
 
 
+
   { $as_echo "$as_me:${as_lineno-$LINENO}: checking for __atomic_test_and_set for size 1" >&5
 $as_echo_n "checking for __atomic_test_and_set for size 1... " >&6; }
 if test "${libat_cv_have_at_tas_1+set}" = set; then :
@@ -14693,7 +14698,8 @@  else
 /* end confdefs.h.  */
 
     int foo_alt(void) { return 0; }
-    void *foo_sel(void) { return foo_alt; }
+    typedef int F (void);
+    F *foo_sel(void) { return foo_alt; }
     int foo(void) __attribute__((ifunc("foo_sel")));
 int
 main ()
diff --git a/libatomic/libatomic_i.h b/libatomic/libatomic_i.h
index 4eb372a..2dad4a8 100644
--- a/libatomic/libatomic_i.h
+++ b/libatomic/libatomic_i.h
@@ -240,7 +240,7 @@  bool libat_is_lock_free (size_t, void *) MAN(is_lock_free);
 # if IFUNC_NCOND(N) == 1
 #  define GEN_SELECTOR(X)					\
 	extern typeof(C2(libat_,X)) C3(libat_,X,_i1) HIDDEN;	\
-	static void * C2(select_,X) (void)			\
+	static typeof(C2(libat_,X)) * C2(select_,X) (void)	\
 	{							\
 	  if (IFUNC_COND_1)					\
 	    return C3(libat_,X,_i1);				\
@@ -250,7 +250,7 @@  bool libat_is_lock_free (size_t, void *) MAN(is_lock_free);
 #  define GEN_SELECTOR(X)					\
 	extern typeof(C2(libat_,X)) C3(libat_,X,_i1) HIDDEN;	\
 	extern typeof(C2(libat_,X)) C3(libat_,X,_i2) HIDDEN;	\
-	static void * C2(select_,X) (void)			\
+	static typeof(C2(libat_,X)) * C2(select_,X) (void)	\
 	{							\
 	  if (IFUNC_COND_1)					\
 	    return C3(libat_,X,_i1);				\
@@ -263,7 +263,7 @@  bool libat_is_lock_free (size_t, void *) MAN(is_lock_free);
 	extern typeof(C2(libat_,X)) C3(libat_,X,_i1) HIDDEN;	\
 	extern typeof(C2(libat_,X)) C3(libat_,X,_i2) HIDDEN;	\
 	extern typeof(C2(libat_,X)) C3(libat_,X,_i3) HIDDEN;	\
-	static void * C2(select_,X) (void)			\
+	static typeof(C2(libat_,X)) * C2(select_,X) (void)	\
 	{							\
 	  if (IFUNC_COND_1)					\
 	    return C3(libat_,X,_i1);				\