[PATCHv3] sysdeps/ieee754: prevent maybe-uninitialized errors with -O [BZ #19444]
Commit Message
With -O included in CFLAGS it fails to build with:
../sysdeps/ieee754/ldbl-96/e_jnl.c: In function '__ieee754_jnl':
../sysdeps/ieee754/ldbl-96/e_jnl.c:146:20: error: 'temp' may be used uninitialized in this function [-Werror=maybe-uninitialized]
b = invsqrtpi * temp / sqrtl (x);
~~~~~~~~~~^~~~~~
../sysdeps/ieee754/ldbl-96/e_jnl.c: In function '__ieee754_ynl':
../sysdeps/ieee754/ldbl-96/e_jnl.c:375:16: error: 'temp' may be used uninitialized in this function [-Werror=maybe-uninitialized]
b = invsqrtpi * temp / sqrtl (x);
~~~~~~~~~~^~~~~~
../sysdeps/ieee754/dbl-64/e_jn.c: In function '__ieee754_jn':
../sysdeps/ieee754/dbl-64/e_jn.c:113:20: error: 'temp' may be used uninitialized in this function [-Werror=maybe-uninitialized]
b = invsqrtpi * temp / sqrt (x);
~~~~~~~~~~^~~~~~
../sysdeps/ieee754/dbl-64/e_jn.c: In function '__ieee754_yn':
../sysdeps/ieee754/dbl-64/e_jn.c:320:16: error: 'temp' may be used uninitialized in this function [-Werror=maybe-uninitialized]
b = invsqrtpi * temp / sqrt (x);
~~~~~~~~~~^~~~~~
Build tested with Yocto for ARM, AARCH64, X86, X86_64, PPC, MIPS, MIPS64
with -O, -O1, -Os.
For soft-fp ARM it needs one more fix for -O1:
https://sourceware.org/ml/libc-alpha/2018-09/msg00300.html
For AARCH64 it needs one more fix in locale for -Os.
[BZ #23716]
* sysdeps/ieee754/dbl-96/e_jnl.c: Fix build with -O
* sysdeps/ieee754/ldbl-96/e_jnl.c: Likewise.
* sysdeps/ieee754/ldbl-128/e_jnl.c: Likewise.
* sysdeps/ieee754/ldbl-128ibm/e_jnl.c: Likewise.
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
---
ChangeLog | 7 +++++++
sysdeps/ieee754/dbl-64/e_jn.c | 21 +++++++++++++++++++++
sysdeps/ieee754/ldbl-128/e_jnl.c | 21 +++++++++++++++++++++
sysdeps/ieee754/ldbl-128ibm/e_jnl.c | 21 +++++++++++++++++++++
sysdeps/ieee754/ldbl-96/e_jnl.c | 21 +++++++++++++++++++++
5 files changed, 91 insertions(+)
Comments
On Sun, 30 Sep 2018, Martin Jansa wrote:
> + /* With GCC 8 (and older) when compiling with -O the compiler
> + warns that the variable 'temp', may be used uninitialized.
> + The switch above covers all possible values of n & 3
> + but GCC without VRP enabled isn't able to figure out the
> + range of possible values is [0,3] as explained in:
> + https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
> + so it's false possitive with -O1 and lower. */
"positive" (for all such comments). Also note there should be two spaces
at the end of a comment between "." and the trailing "*/".
Are you sure about the "and lower", or is this warning disabled for -O0?
(glibc disallows building with -O0 at present, but I think we *should* fix
things to be able to build with -O0, because being able to build with -O0
would be helpful for debugging - any particular pieces needing
optimization should then just have it force-enabled, or functions declared
as always_inline, or whatever is needed, locally.)
On Sun, Sep 30, 2018 at 04:52:56PM +0000, Joseph Myers wrote:
> On Sun, 30 Sep 2018, Martin Jansa wrote:
>
> > + /* With GCC 8 (and older) when compiling with -O the compiler
> > + warns that the variable 'temp', may be used uninitialized.
> > + The switch above covers all possible values of n & 3
> > + but GCC without VRP enabled isn't able to figure out the
> > + range of possible values is [0,3] as explained in:
> > + https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
> > + so it's false possitive with -O1 and lower. */
>
> "positive" (for all such comments). Also note there should be two spaces
> at the end of a comment between "." and the trailing "*/".
OK, will update.
> Are you sure about the "and lower", or is this warning disabled for -O0?
I'm pretty sure that -ftree-vrp isn't enabled with -O0, but will retry
if it also shows this warning with -O0.
> (glibc disallows building with -O0 at present, but I think we *should* fix
> things to be able to build with -O0, because being able to build with -O0
> would be helpful for debugging - any particular pieces needing
> optimization should then just have it force-enabled, or functions declared
> as always_inline, or whatever is needed, locally.)
I agree, but I'm not volunteering to fix all of -O0 issues reported in
[BZ #19444], my motivation was just to fix debug builds in Yocto which
is by default using -O (and issues with -Os which I'm using for some of
my builds):
http://lists.openembedded.org/pipermail/openembedded-core/2018-September/155693.html
Fixing:
https://sourceware.org/glibc/wiki/FAQ#Why_do_I_get:.60.23error_.22glibc_cannot_be_compiled_without_optimization.22.27.2C_when_trying_to_compile_GNU_libc_with_GNU_CC.3F
is out of my league and time I'm willing to spend, I'll leave it to
someone more knowledgable about glibc internals.
Cheers,
On Sun, Sep 30, 2018 at 07:57:11PM +0200, Martin Jansa wrote:
> On Sun, Sep 30, 2018 at 04:52:56PM +0000, Joseph Myers wrote:
> > On Sun, 30 Sep 2018, Martin Jansa wrote:
> >
> > > + /* With GCC 8 (and older) when compiling with -O the compiler
> > > + warns that the variable 'temp', may be used uninitialized.
> > > + The switch above covers all possible values of n & 3
> > > + but GCC without VRP enabled isn't able to figure out the
> > > + range of possible values is [0,3] as explained in:
> > > + https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
> > > + so it's false possitive with -O1 and lower. */
> >
> > "positive" (for all such comments). Also note there should be two spaces
> > at the end of a comment between "." and the trailing "*/".
>
> OK, will update.
>
> > Are you sure about the "and lower", or is this warning disabled for -O0?
>
> I'm pretty sure that -ftree-vrp isn't enabled with -O0, but will retry
> if it also shows this warning with -O0.
You were right, it warning isn't triggered with -O0, I'll drop the " and
lower" part in v4.
docker-shr @ ~/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/git/math $ x86_64-oe-linux-gcc -m64 -march=core2 -mtune=core2 -msse3 -mfpmath=sse --sysroot=/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot ../sysdeps/ieee754/ldbl-96/e_jnl.c -c -std=gnu11 -fgnu89-inline -O1 -pipe -g -feliminate-unused-debug-types -fdebug-prefix-map=/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0=/usr/src/debug/glibc/2.28-r0 -fdebug-prefix-map=/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot= -fdebug-prefix-map=/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot-native= -Wall -Wwrite-strings -Wundef -Werror -fmerge-all-constants -frounding-math -fno-stack-protector -Wstrict-prototypes -Wold-style-definition -fno-math-errno -D__NO_MATH_INLINES -D__LIBC_INTERNAL_MATH_INLINES -I../include -I/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/math -I/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux -I../sysdeps/unix/sysv/linux/x86_64/64 -I../sysdeps/unix/sysv/linux/x86_64 -I../sysdeps/unix/sysv/linux/x86/include -I../sysdeps/unix/sysv/linux/x86 -I../sysdeps/x86/nptl -I../sysdeps/unix/sysv/linux/wordsize-64 -I../sysdeps/x86_64/nptl -I../sysdeps/unix/sysv/linux/include -I../sysdeps/unix/sysv/linux -I../sysdeps/nptl -I../sysdeps/pthread -I../sysdeps/gnu -I../sysdeps/unix/inet -I../sysdeps/unix/sysv -I../sysdeps/unix/x86_64 -I../sysdeps/unix -I../sysdeps/posix -I../sysdeps/x86_64/64 -I../sysdeps/x86_64/fpu/multiarch -I../sysdeps/x86_64/fpu -I../sysdeps/x86/fpu/include -I../sysdeps/x86/fpu -I../sysdeps/x86_64/multiarch -I../sysdeps/x86_64 -I../sysdeps/x86 -I../sysdeps/ieee754/float128 -I../sysdeps/ieee754/ldbl-96/include -I../sysdeps/ieee754/ldbl-96 -I../sysdeps/ieee754/dbl-64/wordsize-64 -I../sysdeps/ieee754/dbl-64 -I../sysdeps/ieee754/flt-32 -I../sysdeps/wordsize-64 -I../sysdeps/ieee754 -I../sysdeps/generic -I.. -I../libio -I. -nostdinc -isystem /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot-native/usr/bin/x86_64-oe-linux.gcc-cross-initial-x86_64/../../lib/x86_64-oe-linux.gcc-cross-initial-x86_64/gcc/x86_64-oe-linux/8.2.0/include -isystem /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot-native/usr/bin/x86_64-oe-linux.gcc-cross-initial-x86_64/../../lib/x86_64-oe-linux.gcc-cross-initial-x86_64/gcc/x86_64-oe-linux/8.2.0/include-fixed -isystem /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot/usr/include -D_LIBC_REENTRANT -include /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/libc-modules.h -DMODULE_NAME=libm -include ../include/libc-symbols.h -DTOP_NAMESPACE=glibc -I../soft-fp -o /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/math/e_jnl.o -MD -MP -MF /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/math/e_jnl.o.dt -MT /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/math/e_jnl.o
../sysdeps/ieee754/ldbl-96/e_jnl.c: In function '__ieee754_jnl':
../sysdeps/ieee754/ldbl-96/e_jnl.c:147:20: error: 'temp' may be used uninitialized in this function [-Werror=maybe-uninitialized]
b = invsqrtpi * temp / sqrtl (x);
~~~~~~~~~~^~~~~~
../sysdeps/ieee754/ldbl-96/e_jnl.c: In function '__ieee754_ynl':
../sysdeps/ieee754/ldbl-96/e_jnl.c:376:16: error: 'temp' may be used uninitialized in this function [-Werror=maybe-uninitialized]
b = invsqrtpi * temp / sqrtl (x);
~~~~~~~~~~^~~~~~
cc1: all warnings being treated as errors
docker-shr @ ~/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/git/math $ x86_64-oe-linux-gcc -m64 -march=core2 -mtune=core2 -msse3 -mfpmath=sse --sysroot=/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot ../sysdeps/ieee754/ldbl-96/e_jnl.c -c -std=gnu11 -fgnu89-inline -O0 -pipe -g -feliminate-unused-debug-types -fdebug-prefix-map=/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0=/usr/src/debug/glibc/2.28-r0 -fdebug-prefix-map=/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot= -fdebug-prefix-map=/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot-native= -Wall -Wwrite-strings -Wundef -Werror -fmerge-all-constants -frounding-math -fno-stack-protector -Wstrict-prototypes -Wold-style-definition -fno-math-errno -D__NO_MATH_INLINES -D__LIBC_INTERNAL_MATH_INLINES -I../include -I/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/math -I/OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux -I../sysdeps/unix/sysv/linux/x86_64/64 -I../sysdeps/unix/sysv/linux/x86_64 -I../sysdeps/unix/sysv/linux/x86/include -I../sysdeps/unix/sysv/linux/x86 -I../sysdeps/x86/nptl -I../sysdeps/unix/sysv/linux/wordsize-64 -I../sysdeps/x86_64/nptl -I../sysdeps/unix/sysv/linux/include -I../sysdeps/unix/sysv/linux -I../sysdeps/nptl -I../sysdeps/pthread -I../sysdeps/gnu -I../sysdeps/unix/inet -I../sysdeps/unix/sysv -I../sysdeps/unix/x86_64 -I../sysdeps/unix -I../sysdeps/posix -I../sysdeps/x86_64/64 -I../sysdeps/x86_64/fpu/multiarch -I../sysdeps/x86_64/fpu -I../sysdeps/x86/fpu/include -I../sysdeps/x86/fpu -I../sysdeps/x86_64/multiarch -I../sysdeps/x86_64 -I../sysdeps/x86 -I../sysdeps/ieee754/float128 -I../sysdeps/ieee754/ldbl-96/include -I../sysdeps/ieee754/ldbl-96 -I../sysdeps/ieee754/dbl-64/wordsize-64 -I../sysdeps/ieee754/dbl-64 -I../sysdeps/ieee754/flt-32 -I../sysdeps/wordsize-64 -I../sysdeps/ieee754 -I../sysdeps/generic -I.. -I../libio -I. -nostdinc -isystem /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot-native/usr/bin/x86_64-oe-linux.gcc-cross-initial-x86_64/../../lib/x86_64-oe-linux.gcc-cross-initial-x86_64/gcc/x86_64-oe-linux/8.2.0/include -isystem /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot-native/usr/bin/x86_64-oe-linux.gcc-cross-initial-x86_64/../../lib/x86_64-oe-linux.gcc-cross-initial-x86_64/gcc/x86_64-oe-linux/8.2.0/include-fixed -isystem /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/recipe-sysroot/usr/include -D_LIBC_REENTRANT -include /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/libc-modules.h -DMODULE_NAME=libm -include ../include/libc-symbols.h -DTOP_NAMESPACE=glibc -I../soft-fp -o /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/math/e_jnl.o -MD -MP -MF /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/math/e_jnl.o.dt -MT /OE/build/oe-core/tmp-glibc/work/core2-64-oe-linux/glibc/2.28-r0/build-x86_64-oe-linux/math/e_jnl.o
> > (glibc disallows building with -O0 at present, but I think we *should* fix
> > things to be able to build with -O0, because being able to build with -O0
> > would be helpful for debugging - any particular pieces needing
> > optimization should then just have it force-enabled, or functions declared
> > as always_inline, or whatever is needed, locally.)
>
> I agree, but I'm not volunteering to fix all of -O0 issues reported in
> [BZ #19444], my motivation was just to fix debug builds in Yocto which
> is by default using -O (and issues with -Os which I'm using for some of
> my builds):
> http://lists.openembedded.org/pipermail/openembedded-core/2018-September/155693.html
>
> Fixing:
> https://sourceware.org/glibc/wiki/FAQ#Why_do_I_get:.60.23error_.22glibc_cannot_be_compiled_without_optimization.22.27.2C_when_trying_to_compile_GNU_libc_with_GNU_CC.3F
> is out of my league and time I'm willing to spend, I'll leave it to
> someone more knowledgable about glibc internals.
>
> Cheers,
> --
> Martin 'JaMa' Jansa jabber: Martin.Jansa@gmail.com
@@ -1,3 +1,10 @@
+2018-09-29 Martin Jansa <Martin.Jansa@gmail.com>
+ Partial fix for [BZ #23716]
+ * sysdeps/ieee754/dbl-96/e_jnl.c: Fix build with -O
+ * sysdeps/ieee754/ldbl-96/e_jnl.c: Likewise.
+ * sysdeps/ieee754/ldbl-128/e_jnl.c: Likewise.
+ * sysdeps/ieee754/ldbl-128ibm/e_jnl.c: Likewise.
+
2018-09-28 Joseph Myers <joseph@codesourcery.com>
* math/fromfp.h: Do not include <math_private.h>.
@@ -43,6 +43,7 @@
#include <math_private.h>
#include <fenv_private.h>
#include <math-underflow.h>
+#include <libc-diag.h>
static const double
invsqrtpi = 5.64189583547756279280e-01, /* 0x3FE20DD7, 0x50429B6D */
@@ -110,7 +111,17 @@ __ieee754_jn (int n, double x)
case 2: temp = -c - s; break;
case 3: temp = c - s; break;
}
+ /* With GCC 8 (and older) when compiling with -O the compiler
+ warns that the variable 'temp', may be used uninitialized.
+ The switch above covers all possible values of n & 3
+ but GCC without VRP enabled isn't able to figure out the
+ range of possible values is [0,3] as explained in:
+ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
+ so it's false possitive with -O1 and lower. */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
b = invsqrtpi * temp / sqrt (x);
+ DIAG_POP_NEEDS_COMMENT;
}
else
{
@@ -317,7 +328,17 @@ __ieee754_yn (int n, double x)
case 2: temp = -s + c; break;
case 3: temp = s + c; break;
}
+ /* With GCC 8 (and older) when compiling with -O the compiler
+ warns that the variable 'temp', may be used uninitialized.
+ The switch above covers all possible values of n & 3
+ but GCC without VRP enabled isn't able to figure out the
+ range of possible values is [0,3] as explained in:
+ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
+ so it's false possitive with -O1 and lower. */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
b = invsqrtpi * temp / sqrt (x);
+ DIAG_POP_NEEDS_COMMENT;
}
else
{
@@ -62,6 +62,7 @@
#include <math_private.h>
#include <fenv_private.h>
#include <math-underflow.h>
+#include <libc-diag.h>
static const _Float128
invsqrtpi = L(5.6418958354775628694807945156077258584405E-1),
@@ -151,7 +152,17 @@ __ieee754_jnl (int n, _Float128 x)
temp = c - s;
break;
}
+ /* With GCC 8 (and older) when compiling with -O the compiler
+ warns that the variable 'temp', may be used uninitialized.
+ The switch above covers all possible values of n & 3
+ but GCC without VRP enabled isn't able to figure out the
+ range of possible values is [0,3] as explained in:
+ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
+ so it's false possitive with -O1 and lower. */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
b = invsqrtpi * temp / sqrtl (x);
+ DIAG_POP_NEEDS_COMMENT;
}
else
{
@@ -387,7 +398,17 @@ __ieee754_ynl (int n, _Float128 x)
temp = s + c;
break;
}
+ /* With GCC 8 (and older) when compiling with -O the compiler
+ warns that the variable 'temp', may be used uninitialized.
+ The switch above covers all possible values of n & 3
+ but GCC without VRP enabled isn't able to figure out the
+ range of possible values is [0,3] as explained in:
+ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
+ so it's false possitive with -O1 and lower. */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
b = invsqrtpi * temp / sqrtl (x);
+ DIAG_POP_NEEDS_COMMENT;
}
else
{
@@ -62,6 +62,7 @@
#include <math_private.h>
#include <fenv_private.h>
#include <math-underflow.h>
+#include <libc-diag.h>
static const long double
invsqrtpi = 5.6418958354775628694807945156077258584405E-1L,
@@ -151,7 +152,17 @@ __ieee754_jnl (int n, long double x)
temp = c - s;
break;
}
+ /* With GCC 8 (and older) when compiling with -O the compiler
+ warns that the variable 'temp', may be used uninitialized.
+ The switch above covers all possible values of n & 3
+ but GCC without VRP enabled isn't able to figure out the
+ range of possible values is [0,3] as explained in:
+ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
+ so it's false possitive with -O1 and lower. */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
b = invsqrtpi * temp / sqrtl (x);
+ DIAG_POP_NEEDS_COMMENT;
}
else
{
@@ -387,7 +398,17 @@ __ieee754_ynl (int n, long double x)
temp = s + c;
break;
}
+ /* With GCC 8 (and older) when compiling with -O the compiler
+ warns that the variable 'temp', may be used uninitialized.
+ The switch above covers all possible values of n & 3
+ but GCC without VRP enabled isn't able to figure out the
+ range of possible values is [0,3] as explained in:
+ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
+ so it's false possitive with -O1 and lower. */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
b = invsqrtpi * temp / sqrtl (x);
+ DIAG_POP_NEEDS_COMMENT;
}
else
{
@@ -62,6 +62,7 @@
#include <math_private.h>
#include <fenv_private.h>
#include <math-underflow.h>
+#include <libc-diag.h>
static const long double
invsqrtpi = 5.64189583547756286948079e-1L, two = 2.0e0L, one = 1.0e0L;
@@ -144,7 +145,17 @@ __ieee754_jnl (int n, long double x)
temp = c - s;
break;
}
+ /* With GCC 8 (and older) when compiling with -O the compiler
+ warns that the variable 'temp', may be used uninitialized.
+ The switch above covers all possible values of n & 3
+ but GCC without VRP enabled isn't able to figure out the
+ range of possible values is [0,3] as explained in:
+ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
+ so it's false possitive with -O1 and lower. */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
b = invsqrtpi * temp / sqrtl (x);
+ DIAG_POP_NEEDS_COMMENT;
}
else
{
@@ -373,7 +384,17 @@ __ieee754_ynl (int n, long double x)
temp = s + c;
break;
}
+ /* With GCC 8 (and older) when compiling with -O the compiler
+ warns that the variable 'temp', may be used uninitialized.
+ The switch above covers all possible values of n & 3
+ but GCC without VRP enabled isn't able to figure out the
+ range of possible values is [0,3] as explained in:
+ https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69230
+ so it's false possitive with -O1 and lower. */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
b = invsqrtpi * temp / sqrtl (x);
+ DIAG_POP_NEEDS_COMMENT;
}
else
{