softfloat: Add float/double to 16 bit integer conversion functions
commitcbcef455a2d29cdd19f4f0a21990cd76c0d03102
authorPeter Maydell <peter.maydell@linaro.org>
Tue, 7 Dec 2010 15:37:34 +0000 (7 15:37 +0000)
committerPeter Maydell <peter.maydell@linaro.org>
Tue, 7 Dec 2010 15:37:34 +0000 (7 15:37 +0000)
tree8a0080abccbb4db822e5ce84085dd6355d372513
parent26a5e69aafd0434922316e4185c45c9bddd4b9ba
softfloat: Add float/double to 16 bit integer conversion functions

The ARM architecture needs float/double to 16 bit integer conversions.
(The 32 bit versions aren't sufficient because of the requirement
to saturate at 16 bit MAXINT/MININT and to get the exception bits right.)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Nathan Froyd <froydnj@codesourcery.com>
fpu/softfloat.c
fpu/softfloat.h