about summary refs log tree commit diff
diff options
context:
space:
mode:
authorAdhemerval Zanella <adhemerval.zanella@linaro.org>2020-03-26 11:00:56 -0300
committerPranav Kant <prka@google.com>2023-12-28 23:45:34 +0000
commit66cff6cd8a8f477166939dd1df0ddd5c3e850adc (patch)
tree767c0732b7f986dad54976dbd0e0718ab08814e1
parent0778e25fe1f34789794689f99e25b0c5ff001795 (diff)
downloadglibc-66cff6cd8a8f477166939dd1df0ddd5c3e850adc.tar.gz
glibc-66cff6cd8a8f477166939dd1df0ddd5c3e850adc.tar.xz
glibc-66cff6cd8a8f477166939dd1df0ddd5c3e850adc.zip
x86_64: Add SSE sfp-exceptions
The exported x86_64 fenv.h functions operate on both i387 and SSE (since
they should work on both float, double, and long double) while the
internal libc_fe* set either SSE (float, double, and float128) or
i387 (long double).

The libgcc __sfp_handle_exceptions (used on float128 implementation),
however, will set either SEE or i387 exception depending of the
exception to raise.  This broke the internal assumption of float128
where only SSE operations will be used.

This patch reimplements the libgcc __sfp_handle_exceptions to use only
SSE operations and sets libgcc to use it instead of its own
implementation.

And I think we should fix libgcc in a similar manner, since checking on
config/i386/64/sfp-machine.h it already only supports SSE rounding mode
and x86_64 ABI also expectes float128 to use SSE registers [1]
(although it is not clear on how future implementation might implement
it).

Checked on x86_64-linux-gnu.

[1] https://github.com/hjl-tools/x86-psABI/wiki/X86-psABI
-rw-r--r--sysdeps/x86/fpu/sfp-exceptions.c57
-rw-r--r--sysdeps/x86_64/fpu/Makefile4
2 files changed, 60 insertions, 1 deletions
diff --git a/sysdeps/x86/fpu/sfp-exceptions.c b/sysdeps/x86/fpu/sfp-exceptions.c
new file mode 100644
index 0000000000..74ce0716e2
--- /dev/null
+++ b/sysdeps/x86/fpu/sfp-exceptions.c
@@ -0,0 +1,57 @@
+/* x86_64 soft-fp exception handling for _Float128.
+   Copyright (C) 2020 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
+   <https://www.gnu.org/licenses/>.  */
+
+#include <fenv.h>
+#include <float.h>
+#include <math-barriers.h>
+#include <soft-fp.h>
+
+void
+__sfp_handle_exceptions (int _fex)
+{
+  if (_fex & FP_EX_INVALID)
+    {
+      float f = 0.0f;
+      math_force_eval (f / f);
+    }
+  if (_fex & FP_EX_DENORM)
+    {
+      float f = FLT_MIN, g = 2.0f;
+      math_force_eval (f / g);
+    }
+  if (_fex & FP_EX_DIVZERO)
+    {
+      float f = 1.0f, g = 0.0f;
+      math_force_eval (f / g);
+    }
+  if (_fex & FP_EX_OVERFLOW)
+    {
+      float force_underflow = FLT_MAX * FLT_MAX;
+      math_force_eval (force_underflow);
+    }
+  if (_fex & FP_EX_UNDERFLOW)
+    {
+      float force_overflow = FLT_MIN * FLT_MIN;
+      math_force_eval (force_overflow);
+    }
+  if (_fex & FP_EX_INEXACT)
+    {
+      float f = 1.0f, g = 3.0f;
+      math_force_eval (f / g);
+    }
+}
diff --git a/sysdeps/x86_64/fpu/Makefile b/sysdeps/x86_64/fpu/Makefile
index 2b7d69bb50..ed26226623 100644
--- a/sysdeps/x86_64/fpu/Makefile
+++ b/sysdeps/x86_64/fpu/Makefile
@@ -25,8 +25,10 @@ libmvec-support += svml_d_cos2_core svml_d_cos4_core_avx \
 libmvec-static-only-routines = svml_finite_alias
 endif
 
-# Variables for libmvec tests.
 ifeq ($(subdir),math)
+libm-routines += sfp-exceptions
+
+# Variables for libmvec tests.
 ifeq ($(build-mathvec),yes)
 libmvec-tests += double-vlen2 double-vlen4 double-vlen4-avx2 \
 		 float-vlen4 float-vlen8 float-vlen8-avx2