This is the mail archive of the
glibc-cvs@sourceware.org
mailing list for the glibc project.
GNU C Library master sources branch master updated. glibc-2.22-274-g94ced92
- From: jsm28 at sourceware dot org
- To: glibc-cvs at sourceware dot org
- Date: 18 Sep 2015 20:35:32 -0000
- Subject: GNU C Library master sources branch master updated. glibc-2.22-274-g94ced92
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "GNU C Library master sources".
The branch, master has been updated
via 94ced920a951bcab51f2ef955ccd2cc51668e6f3 (commit)
from 89faa0340ad399ead8104f8fb6b7fed81f7d939c (commit)
Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.
- Log -----------------------------------------------------------------
http://sourceware.org/git/gitweb.cgi?p=glibc.git;a=commitdiff;h=94ced920a951bcab51f2ef955ccd2cc51668e6f3
commit 94ced920a951bcab51f2ef955ccd2cc51668e6f3
Author: Joseph Myers <joseph@codesourcery.com>
Date: Fri Sep 18 20:34:59 2015 +0000
Avoid excess range in results from i386 scalb functions (bug 18981).
i386 scalb / scalbn / scalbln (and thus ldexp) functions for float and
double can return results with excess range (and consequently excess
precision for subnormal results). As the results of these functions
are fully determined by reference to IEEE 754 operations, this is
unambiguously a bug, apart from the testsuite failures it causes.
This patch makes those functions store their results on the stack and
load them back to eliminate the excess range. Double rounding is not
a problem, as the only cases where it could occur are when the result
overflows or underflows for extended precision, and then the
double-rounded results are the same as the single-rounded results.
The new macros will be used for more functions, more such macros
added, and existing code refactored to use such macros, in subsequent
patches.
Tested for x86. Committed.
[BZ #18981]
* sysdeps/i386/fpu/i386-math-asm.h: New file.
* sysdeps/i386/fpu/e_scalb.S: Include <i386-math-asm.h>.
(__ieee754_scalb): Use DBL_NARROW_EVAL.
* sysdeps/i386/fpu/e_scalbf.S: Include <i386-math-asm.h>.
(__ieee754_scalbf): Use FLT_NARROW_EVAL.
* sysdeps/i386/fpu/s_scalbn.S: Include <i386-math-asm.h>.
(__scalbn): Use DBL_NARROW_EVAL.
* sysdeps/i386/fpu/s_scalbnf.S: Include <i386-math-asm.h>.
(__scalbnf): Use FLT_NARROW_EVAL.
diff --git a/ChangeLog b/ChangeLog
index 67c837d..cafe59e 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,16 @@
+2015-09-18 Joseph Myers <joseph@codesourcery.com>
+
+ [BZ #18981]
+ * sysdeps/i386/fpu/i386-math-asm.h: New file.
+ * sysdeps/i386/fpu/e_scalb.S: Include <i386-math-asm.h>.
+ (__ieee754_scalb): Use DBL_NARROW_EVAL.
+ * sysdeps/i386/fpu/e_scalbf.S: Include <i386-math-asm.h>.
+ (__ieee754_scalbf): Use FLT_NARROW_EVAL.
+ * sysdeps/i386/fpu/s_scalbn.S: Include <i386-math-asm.h>.
+ (__scalbn): Use DBL_NARROW_EVAL.
+ * sysdeps/i386/fpu/s_scalbnf.S: Include <i386-math-asm.h>.
+ (__scalbnf): Use FLT_NARROW_EVAL.
+
2015-09-18 Wilco Dijkstra <wdijkstr@arm.com>
* math/w_tgamma.c (__ieee754_gamma_r): Use explicit sign check.
diff --git a/NEWS b/NEWS
index 8f487fa..4d65df2 100644
--- a/NEWS
+++ b/NEWS
@@ -16,7 +16,7 @@ Version 2.23
18647, 18661, 18674, 18675, 18681, 18757, 18778, 18781, 18787, 18789,
18790, 18795, 18796, 18820, 18823, 18824, 18857, 18863, 18870, 18872,
18873, 18875, 18887, 18921, 18951, 18952, 18961, 18966, 18967, 18970,
- 18977.
+ 18977, 18981.
* The obsolete header <regexp.h> has been removed. Programs that require
this header must be updated to use <regex.h> instead.
diff --git a/sysdeps/i386/fpu/e_scalb.S b/sysdeps/i386/fpu/e_scalb.S
index 6a01aa6..370924c 100644
--- a/sysdeps/i386/fpu/e_scalb.S
+++ b/sysdeps/i386/fpu/e_scalb.S
@@ -6,6 +6,7 @@
*/
#include <machine/asm.h>
+#include <i386-math-asm.h>
.section .rodata
@@ -52,6 +53,7 @@ ENTRY(__ieee754_scalb)
jne 4f
fscale
fstp %st(1)
+ DBL_NARROW_EVAL
ret
/* y is -inf */
diff --git a/sysdeps/i386/fpu/e_scalbf.S b/sysdeps/i386/fpu/e_scalbf.S
index 12b2515..4f2dfa3 100644
--- a/sysdeps/i386/fpu/e_scalbf.S
+++ b/sysdeps/i386/fpu/e_scalbf.S
@@ -7,6 +7,7 @@
*/
#include <machine/asm.h>
+#include <i386-math-asm.h>
.section .rodata
@@ -54,6 +55,7 @@ ENTRY(__ieee754_scalbf)
jne 4f
fscale
fstp %st(1)
+ FLT_NARROW_EVAL
ret
/* y is -inf */
diff --git a/sysdeps/i386/fpu/i386-math-asm.h b/sysdeps/i386/fpu/i386-math-asm.h
new file mode 100644
index 0000000..fd43130
--- /dev/null
+++ b/sysdeps/i386/fpu/i386-math-asm.h
@@ -0,0 +1,39 @@
+/* Helper macros for x86 libm functions.
+ Copyright (C) 2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ 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
+ <http://www.gnu.org/licenses/>. */
+
+#ifndef _I386_MATH_ASM_H
+#define _I386_MATH_ASM_H 1
+
+/* Remove excess range and precision by storing a value on the stack
+ and loading it back. */
+#define FLT_NARROW_EVAL \
+ subl $4, %esp; \
+ cfi_adjust_cfa_offset (4); \
+ fstps (%esp); \
+ flds (%esp); \
+ addl $4, %esp; \
+ cfi_adjust_cfa_offset (-4);
+#define DBL_NARROW_EVAL \
+ subl $8, %esp; \
+ cfi_adjust_cfa_offset (8); \
+ fstpl (%esp); \
+ fldl (%esp); \
+ addl $8, %esp; \
+ cfi_adjust_cfa_offset (-8);
+
+#endif /* i386-math-asm.h. */
diff --git a/sysdeps/i386/fpu/s_scalbn.S b/sysdeps/i386/fpu/s_scalbn.S
index dab7580..4e90903 100644
--- a/sysdeps/i386/fpu/s_scalbn.S
+++ b/sysdeps/i386/fpu/s_scalbn.S
@@ -4,6 +4,7 @@
*/
#include <machine/asm.h>
+#include <i386-math-asm.h>
RCSID("$NetBSD: s_scalbn.S,v 1.4 1995/05/09 00:19:06 jtc Exp $")
@@ -12,6 +13,7 @@ ENTRY(__scalbn)
fldl 4(%esp)
fscale
fstp %st(1)
+ DBL_NARROW_EVAL
ret
END (__scalbn)
strong_alias (__scalbn, __scalbln)
diff --git a/sysdeps/i386/fpu/s_scalbnf.S b/sysdeps/i386/fpu/s_scalbnf.S
index ad97be8..f8353c4 100644
--- a/sysdeps/i386/fpu/s_scalbnf.S
+++ b/sysdeps/i386/fpu/s_scalbnf.S
@@ -4,6 +4,7 @@
*/
#include <machine/asm.h>
+#include <i386-math-asm.h>
RCSID("$NetBSD: s_scalbnf.S,v 1.3 1995/05/09 00:19:59 jtc Exp $")
@@ -12,6 +13,7 @@ ENTRY(__scalbnf)
flds 4(%esp)
fscale
fstp %st(1)
+ FLT_NARROW_EVAL
ret
END (__scalbnf)
strong_alias (__scalbnf, __scalblnf)
-----------------------------------------------------------------------
Summary of changes:
ChangeLog | 13 +++++++++
NEWS | 2 +-
sysdeps/i386/fpu/e_scalb.S | 2 +
sysdeps/i386/fpu/e_scalbf.S | 2 +
.../i386/fpu/i386-math-asm.h | 29 +++++++++++++------
sysdeps/i386/fpu/s_scalbn.S | 2 +
sysdeps/i386/fpu/s_scalbnf.S | 2 +
7 files changed, 42 insertions(+), 10 deletions(-)
copy benchtests/bench-util.h => sysdeps/i386/fpu/i386-math-asm.h (58%)
hooks/post-receive
--
GNU C Library master sources