[PATCHv2] sysdeps/ieee754/soft-fp: ignore maybe-uninitialized with -O [BZ #19444]
Commit Message
* with -O, -O1, -Os it fails with:
In file included from ../soft-fp/soft-fp.h:318,
from ../sysdeps/ieee754/soft-fp/s_fdiv.c:28:
../sysdeps/ieee754/soft-fp/s_fdiv.c: In function '__fdiv':
../soft-fp/op-2.h:98:25: error: 'R_f1' may be used uninitialized in this function [-Werror=maybe-uninitialized]
X##_f0 = (X##_f1 << (_FP_W_TYPE_SIZE - (N)) | X##_f0 >> (N) \
^~
../sysdeps/ieee754/soft-fp/s_fdiv.c:38:14: note: 'R_f1' was declared here
FP_DECL_D (R);
^
../soft-fp/op-2.h:37:36: note: in definition of macro '_FP_FRAC_DECL_2'
_FP_W_TYPE X##_f0 _FP_ZERO_INIT, X##_f1 _FP_ZERO_INIT
^
../soft-fp/double.h:95:24: note: in expansion of macro '_FP_DECL'
# define FP_DECL_D(X) _FP_DECL (2, X)
^~~~~~~~
../sysdeps/ieee754/soft-fp/s_fdiv.c:38:3: note: in expansion of macro 'FP_DECL_D'
FP_DECL_D (R);
^~~~~~~~~
../soft-fp/op-2.h:101:17: error: 'R_f0' may be used uninitialized in this function [-Werror=maybe-uninitialized]
: (X##_f0 << (_FP_W_TYPE_SIZE - (N))) != 0)); \
^~
../sysdeps/ieee754/soft-fp/s_fdiv.c:38:14: note: 'R_f0' was declared here
FP_DECL_D (R);
^
../soft-fp/op-2.h:37:14: note: in definition of macro '_FP_FRAC_DECL_2'
_FP_W_TYPE X##_f0 _FP_ZERO_INIT, X##_f1 _FP_ZERO_INIT
^
../soft-fp/double.h:95:24: note: in expansion of macro '_FP_DECL'
# define FP_DECL_D(X) _FP_DECL (2, X)
^~~~~~~~
../sysdeps/ieee754/soft-fp/s_fdiv.c:38:3: note: in expansion of macro 'FP_DECL_D'
FP_DECL_D (R);
^~~~~~~~~
Build tested with Yocto for ARM, AARCH64, X86, X86_64, PPC, MIPS, MIPS64
with -O, -O1, -Os.
For AARCH64 it needs one more fix in locale for -Os.
Partial fix for [BZ #23716]
* sysdeps/ieee754/soft-fp/s_fdiv.c: Fix build with -O
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
---
ChangeLog | 4 ++++
sysdeps/ieee754/soft-fp/s_fdiv.c | 12 ++++++++++++
2 files changed, 16 insertions(+)
Comments
On Sun, 30 Sep 2018, Martin Jansa wrote:
> +#include <libc-diag.h>
> +/* R_f[01] are not set in cases where it is not used in packing, but the
> + compiler does not see that it is set in all cases where it is
> + used, resulting in warnings that it may be used uninitialized.
> + The location of the warning differs in different versions of GCC,
> + it may be where R is defined using a macro or it may be where the
> + macro is defined. */
This comment is missing the key information that it's specifically a
warning seen with -O1.
On Sun, Sep 30, 2018 at 04:53:36PM +0000, Joseph Myers wrote:
> On Sun, 30 Sep 2018, Martin Jansa wrote:
>
> > +#include <libc-diag.h>
> > +/* R_f[01] are not set in cases where it is not used in packing, but the
> > + compiler does not see that it is set in all cases where it is
> > + used, resulting in warnings that it may be used uninitialized.
> > + The location of the warning differs in different versions of GCC,
> > + it may be where R is defined using a macro or it may be where the
> > + macro is defined. */
>
> This comment is missing the key information that it's specifically a
> warning seen with -O1.
The other 3 cases of the same:
https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/ieee754/soft-fp/s_fma.c;h=41cf81a74bfc6a8c21c565fb04c90ced4bfd19e8;hb=refs/heads/master#l32
https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/ieee754/soft-fp/s_fmaf.c;h=afe64356b3a9c5f3467c35b80f33ba946fbeee46;hb=refs/heads/master#l32
https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/ieee754/soft-fp/s_fmal.c;h=4c2d4c5d9be09f0c8464fb20b81af3aff131e65d;hb=refs/heads/master#l32
are probably also triggered only with -O1, do you want me to update all
4 or just the new case?
On Sun, 30 Sep 2018, Martin Jansa wrote:
> On Sun, Sep 30, 2018 at 04:53:36PM +0000, Joseph Myers wrote:
> > On Sun, 30 Sep 2018, Martin Jansa wrote:
> >
> > > +#include <libc-diag.h>
> > > +/* R_f[01] are not set in cases where it is not used in packing, but the
> > > + compiler does not see that it is set in all cases where it is
> > > + used, resulting in warnings that it may be used uninitialized.
> > > + The location of the warning differs in different versions of GCC,
> > > + it may be where R is defined using a macro or it may be where the
> > > + macro is defined. */
> >
> > This comment is missing the key information that it's specifically a
> > warning seen with -O1.
>
> The other 3 cases of the same:
> https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/ieee754/soft-fp/s_fma.c;h=41cf81a74bfc6a8c21c565fb04c90ced4bfd19e8;hb=refs/heads/master#l32
> https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/ieee754/soft-fp/s_fmaf.c;h=afe64356b3a9c5f3467c35b80f33ba946fbeee46;hb=refs/heads/master#l32
> https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/ieee754/soft-fp/s_fmal.c;h=4c2d4c5d9be09f0c8464fb20b81af3aff131e65d;hb=refs/heads/master#l32
>
> are probably also triggered only with -O1, do you want me to update all
> 4 or just the new case?
They're definitely not limited to -O1 (or weren't when added in 2015,
commit dc6b5aed1b406a53c4512d355376b4e12c7da971). So it's only the new
case for which a comment about being limited to -O1 is appropriate.
@@ -1,3 +1,7 @@
+2018-09-30 Martin Jansa <Martin.Jansa@gmail.com>
+ Partial fix for [BZ #23716]
+ * sysdeps/ieee754/soft-fp/s_fdiv.c: Fix build with -O.
+
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
@@ -25,6 +25,16 @@
#undef fdivl
#include <math-narrow.h>
+
+#include <libc-diag.h>
+/* R_f[01] are not set in cases where it is not used in packing, but the
+ compiler does not see that it is set in all cases where it is
+ used, resulting in warnings that it may be used uninitialized.
+ The location of the warning differs in different versions of GCC,
+ it may be where R is defined using a macro or it may be where the
+ macro is defined. */
+DIAG_PUSH_NEEDS_COMMENT;
+DIAG_IGNORE_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
#include <soft-fp.h>
#include <single.h>
#include <double.h>
@@ -53,4 +63,6 @@ __fdiv (double x, double y)
CHECK_NARROW_DIV (ret, x, y);
return ret;
}
+DIAG_POP_NEEDS_COMMENT;
+
libm_alias_float_double (div)