diff mbox

[v4,09/10] RISC-V: Fix llrint and llround missing exceptions on RV32

Message ID cdc854fe543e20da1095059d2c7fbe471fff1635.1543572707.git.zongbox@gmail.com
State New, archived
Headers show

Commit Message

Zong Li Dec. 1, 2018, 3:10 a.m. UTC
Similar to the recent fix for MIPS, ARM and S/390, RV32 is missing
correct exception on overflow from llrint and llround functions because
cast from floating-point types to long long do not result in correct
exceptions on overflow.

2018-11-29  Zong Li  <zong@andestech.com>

	* sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h: New file.
---
 ChangeLog                                        |  1 +
 sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h | 38 ++++++++++++++++++++++++
 2 files changed, 39 insertions(+)
 create mode 100644 sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h

Comments

Joseph Myers Dec. 3, 2018, 6:05 p.m. UTC | #1
On Sat, 1 Dec 2018, Zong Li wrote:

> Similar to the recent fix for MIPS, ARM and S/390, RV32 is missing

That's not "recent" now.
Zong Li Dec. 4, 2018, 10:01 a.m. UTC | #2
Joseph Myers <joseph@codesourcery.com> 於 2018年12月4日 週二 上午2:05寫道:
>
> On Sat, 1 Dec 2018, Zong Li wrote:
>
> > Similar to the recent fix for MIPS, ARM and S/390, RV32 is missing
>
> That's not "recent" now.
>
Ok, I would modify the commit message.
Zong Li Dec. 10, 2018, 2:46 a.m. UTC | #3
DJ Delorie <dj@redhat.com> 於 2018年12月8日 週六 上午8:08寫道:
>
>
> Zong Li <zongbox@gmail.com> writes:
> > +     * sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h: New file.
> >
> > +/* Fix for conversion of floating point to integer overflow.  ARM version.
>
> RISC-V version, you mean.

Thanks for pointing out the mistake.
diff mbox

Patch

diff --git a/ChangeLog b/ChangeLog
index 9c9d2f9..ee74249 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -42,6 +42,7 @@ 
 	* sysdeps/unix/sysv/linux/riscv/configure.ac: Likewise.
 	* sysdeps/unix/sysv/linux/riscv/shlib-versions: Likewise.
 	* sysdeps/riscv/preconfigure: Likewise.
+	* sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h: New file.
 
 2018-11-28  Florian Weimer  <fweimer@redhat.com>
 
diff --git a/sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h b/sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h
new file mode 100644
index 0000000..6b34a84
--- /dev/null
+++ b/sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h
@@ -0,0 +1,38 @@ 
+/* Fix for conversion of floating point to integer overflow.  ARM version.
+   Copyright (C) 2015-2018 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 FIX_FP_INT_CONVERT_OVERFLOW_H
+#define FIX_FP_INT_CONVERT_OVERFLOW_H	1
+
+/* As of GCC 5, the generic libgcc2.c conversions from floating point
+   to long long may not raise the correct exceptions on overflow (and
+   may raise spurious "inexact" exceptions even in non-overflow cases,
+   see <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=59412>).  */
+#define FIX_FLT_LONG_CONVERT_OVERFLOW 0
+#define FIX_FLT_LLONG_CONVERT_OVERFLOW 1
+
+#define FIX_DBL_LONG_CONVERT_OVERFLOW 0
+#define FIX_DBL_LLONG_CONVERT_OVERFLOW 1
+
+#define FIX_LDBL_LONG_CONVERT_OVERFLOW 0
+#define FIX_LDBL_LLONG_CONVERT_OVERFLOW 0
+
+#define FIX_FLT128_LONG_CONVERT_OVERFLOW 0
+#define FIX_FLT128_LLONG_CONVERT_OVERFLOW 0
+
+#endif /* fix-fp-int-convert-overflow.h */