math: Remove const attribute from totalorder* functions
Commit Message
From: "Gabriel F. T. Gomes" <gabrielftg@linux.ibm.com>
NOTE:
If this is not a bug in glibc, then it might be a bug in gcc (I tested
with Debian's gcc and with the toolchain generated by build-many-glibcs
with top of the tree binutils, gcc, and glibc as of yesterday).
-- 8< --
Since the commit
commit 42760d764649ad82f5fe45a26cbdf2c2500409f7
Author: Joseph Myers <joseph@codesourcery.com>
Date: Thu Aug 15 15:18:34 2019 +0000
Make totalorder and totalordermag functions take pointer arguments.
the test case math/test-totalorderl-ldbl-128ibm fails on every input
pair, when compiled with -O2, which is the case for glibc test suite.
Debugging showed that the test case is passing arguments incorrectly to
totalorderl. This can also be inferred by the fact that compiling the
test case with -O0 hides the bug.
The documentation for the const attribute in GCC manual reads:
Note that a function that has pointer arguments and examines the data
pointed to must not be declared const if the pointed-to data might
change between successive invocations of the function. In general,
since a function cannot distinguish data that might change from data
that cannot, const functions should never take pointer or, in C++,
reference arguments. Likewise, a function that calls a non-const
function usually must not be const itself.
This commit removes the const attribute from the declarations of all
totalorder functions, since the pointed-to data is likely to be changed
by user code between invocations of totalorder*.
Tested for powerpc64le and x86_64.
* math/bits/mathcalls.h (totalorder, totalordermag): Remove
const attribute.
---
math/bits/mathcalls.h | 7 +++----
1 file changed, 3 insertions(+), 4 deletions(-)
Comments
On Wed, 4 Sep 2019, Gabriel F. T. Gomes wrote:
> This commit removes the const attribute from the declarations of all
> totalorder functions, since the pointed-to data is likely to be changed
> by user code between invocations of totalorder*.
This is OK, but it would be better to replace it with __attribute_pure__
(the patch with that replacement is also OK, if it works).
@@ -373,13 +373,12 @@ __MATHDECL_1 (int, canonicalize,, (_Mdouble_ *__cx, const _Mdouble_ *__x));
#if __GLIBC_USE (IEC_60559_BFP_EXT) || __MATH_DECLARING_FLOATN
/* Total order operation. */
-__MATHDECL_1 (int, totalorder,, (const _Mdouble_ *__x, const _Mdouble_ *__y))
- __attribute__ ((__const__));
+__MATHDECL_1 (int, totalorder,, (const _Mdouble_ *__x,
+ const _Mdouble_ *__y));
/* Total order operation on absolute values. */
__MATHDECL_1 (int, totalordermag,, (const _Mdouble_ *__x,
- const _Mdouble_ *__y))
- __attribute__ ((__const__));
+ const _Mdouble_ *__y));
/* Get NaN payload. */
__MATHCALL (getpayload,, (const _Mdouble_ *__x));