Document C99 floating-point environment.

4 files changed, 465 insertions, 0 deletions

diff --git a/lib/libm/man/feclearexcept.3 b/lib/libm/man/feclearexcept.3
new file mode 100644
index 00000000000..9a6eaf6a500
--- /dev/null
+++ b/lib/libm/man/feclearexcept.3
@@ -0,0 +1,170 @@
+.\"	$OpenBSD: feclearexcept.3,v 1.1 2011/05/25 21:28:43 martynas Exp $
+.\"
+.\" Copyright (c) 2011 Martynas Venckus <martynas@openbsd.org>
+.\"
+.\" Permission to use, copy, modify, and distribute this software for any
+.\" purpose with or without fee is hereby granted, provided that the above
+.\" copyright notice and this permission notice appear in all copies.
+.\"
+.\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+.\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+.\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+.\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+.\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+.\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+.\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+.\"
+.Dd $Mdocdate: May 25 2011 $
+.Dt FECLEAREXCEPT 3
+.Os
+.Sh NAME
+.Nm feclearexcept ,
+.Nm fegetexceptflag ,
+.Nm feraiseexcept ,
+.Nm fesetexceptflag ,
+.Nm fetestexcept
+.Nd access floating-point status flags
+.Sh SYNOPSIS
+.Fd #include <fenv.h>
+.Ft int
+.Fn feclearexcept "int excepts"
+.Ft int
+.Fn fegetexceptflag "fexcept_t *flagp" "int excepts"
+.Ft int
+.Fn feraiseexcept "int excepts"
+.Ft int
+.Fn fesetexceptflag "const fexcept_t *flagp" "int excepts"
+.Ft int
+.Fn fetestexcept "int excepts"
+.Sh DESCRIPTION
+These functions provide access to the floating-point status flags.
+The
+.Fa excepts
+input argument is a bitmask specifying an exception type and
+containing any of the values listed below.
+.Bl -tag -width ".Dv FE_DIVBYZERO"
+.It Dv FE_DIVBYZERO
+A divide-by-zero exception occurs when the program attempts to
+divide a finite non-zero number by zero.
+.It Dv FE_INEXACT
+An inexact exception is raised whenever there is a loss of precision
+due to rounding.
+.It Dv FE_INVALID
+Invalid operation exceptions occur when a program attempts to
+perform calculations for which there is no reasonable representable
+answer.
+
+For instance, subtraction of infinities, division of zero by zero,
+ordered comparison involving \*(Nas, and taking the square root of a
+negative number are all invalid operations.
+.It Dv FE_OVERFLOW
+An overflow exception occurs when the magnitude of the result of a
+computation is too large to fit in the destination type.
+.It Dv FE_UNDERFLOW
+Underflow occurs when the result of a computation is too close to zero
+to be represented as a non-zero value in the destination type.
+.It Dv FE_DENORMAL
+Denormalization exception occurs when the result of a floating-point
+expression is a denormalized number.
+
+This is available only on the floating-point implementations of
+amd64 and i386 processors.
+.El
+.Pp
+Additionally, the macro
+.Dv FE_ALL_EXCEPT
+is simply the bitwise OR of all floating-point exception macros
+listed above.
+.Pp
+The
+.Fn feclearexcept
+function clears the floating-point exceptions represented by
+.Fa excepts .
+.Pp
+The
+.Fn fegetexceptflag
+function stores representation of the states of the floating-point
+flags indicated by
+.Pa excepts
+in the object pointed to by
+.Pa flagp .
+.Pp
+The
+.Fn feraiseexcept
+funtion raises floating-point exceptions represented by
+.Pa excepts .
+.Pp
+The
+.Fn fesetexceptflag
+function sets the floating-point status flags indicated by
+.Pa excepts
+to the states stored in the object pointed to by
+.Pa flagp .
+The value of
+.Pa flagp
+shall have been set by a previous call to
+.Fn fegetexceptflag
+whose second argument represented at least those floating-point
+exceptions represented by
+.Pa excepts .
+This function does not raise floating-point exceptions, but only
+sets the state of the flags.
+.Pp
+The
+.Fn fetesteexcept
+function determines which of a specified subset of the floating-point
+exception flags are currently set.
+The
+.Pa excepts
+specifies the floating-point status flags to be queried.
+.Sh RETURN VALUES
+The
+.Fn feclearexcept ,
+.Fn fegetexceptflag ,
+.Fn feraiseexcept ,
+and
+.Fn fesetexceptflag
+functions return zero on success, and non-zero if an error occurred.
+The
+.Fn fetestexcept
+function returns bitmask of a specified subset of the floating-point
+exception flags are currently set.
+.Sh SEE ALSO
+.Xr fegetround 3 ,
+.Xr fegetenv 3 ,
+.Xr feenableexcept 3
+.Sh STANDARDS
+The
+.Fn feclearexcept ,
+.Fn fegetexceptflag ,
+.Fn feraiseexcept ,
+.Fn fesetexceptflag ,
+and
+.Fn fetestexcept
+functions conform to
+.St -isoC-99 .
+.Pp
+The return types for
+.Fn feclearexcept ,
+.Fn fegetexceptflag ,
+.Fn feraiseexcept ,
+and
+.Fn fesetexceptflag
+are
+.Vt int
+for alignment with
+.St -isoC-99
+Defect Report #202.
+.Sh HISTORY
+These functions first appeared in
+.Ox 5.0 .
+.Sh CAVEATS
+On some architectures,
+.Fn feraiseexcept
+function additionally raises the 
+.Dq inexact
+floating-point exception whenever it raises the
+.Dq overflow
+or
+.Dq underflow
+floating-point exception.
diff --git a/lib/libm/man/feenableexcept.3 b/lib/libm/man/feenableexcept.3
new file mode 100644
index 00000000000..d550c6575eb
--- /dev/null
+++ b/lib/libm/man/feenableexcept.3
@@ -0,0 +1,87 @@
+.\"	$OpenBSD: feenableexcept.3,v 1.1 2011/05/25 21:28:43 martynas Exp $
+.\"
+.\" Copyright (c) 2011 Martynas Venckus <martynas@openbsd.org>
+.\"
+.\" Permission to use, copy, modify, and distribute this software for any
+.\" purpose with or without fee is hereby granted, provided that the above
+.\" copyright notice and this permission notice appear in all copies.
+.\"
+.\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+.\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+.\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+.\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+.\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+.\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+.\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+.\"
+.Dd $Mdocdate: May 25 2011 $
+.Dt FEENABLEEXCEPT 3
+.Os
+.Sh NAME
+.Nm feenableexcept ,
+.Nm fedisableexcept ,
+.Nm fegetexcept
+.Nd control floating-point exception masks
+.Sh SYNOPSIS
+.Fd #include <fenv.h>
+.Ft int
+.Fn feenableexcept "int excepts"
+.Ft int
+.Fn fedisableexcept "int excepts"
+.Ft int
+.Fn fegetexcept void
+.Sh DESCRIPTION
+These functions provide control of the floating-point exception
+masks.
+The
+.Fa excepts
+input argument is a bitmask specifying an exception type and
+containing any of the values listed in
+.Xr feclearexcept 3 .
+.El
+.Pp
+The
+.Fn feenableexcept
+function unmasks the floating-point exceptions represented by
+.Fa excepts .
+The future floating-point operations that produce
+.Fa excepts
+will trap, and a
+.Dv SIGFPE
+will be delivered to the process.
+.Pp
+The
+.Fn fedisableexcept
+function masks the floating-point exceptions represented by
+.Fa excepts .
+All exceptions are masked by default.
+.Pp
+The
+.Fn fegetexcept
+function returns the current exception mask.
+.Sh RETURN VALUES
+The
+.Fn feenableexcept ,
+and
+.Fn fedisableexcept
+functions return the previous exception mask.
+The
+.Fn fegetexcept
+function returns the current exception mask.
+.Sh SEE ALSO
+.Xr sigaction 2 ,
+.Xr feclearexcept 3 ,
+.Xr fegetround 3 ,
+.Xr fegetenv 3
+.Sh STANDARDS
+The
+.Fn feenableexcept ,
+.Fn fedisableexcept ,
+and
+.Fn fegetexcept
+functions are
+.Ox
+extensions.
+.Sh HISTORY
+These functions first appeared in
+.Ox 5.0 .
diff --git a/lib/libm/man/fegetenv.3 b/lib/libm/man/fegetenv.3
new file mode 100644
index 00000000000..cb6553e5c63
--- /dev/null
+++ b/lib/libm/man/fegetenv.3
@@ -0,0 +1,127 @@
+.\"	$OpenBSD: fegetenv.3,v 1.1 2011/05/25 21:28:43 martynas Exp $
+.\"
+.\" Copyright (c) 2011 Martynas Venckus <martynas@openbsd.org>
+.\"
+.\" Permission to use, copy, modify, and distribute this software for any
+.\" purpose with or without fee is hereby granted, provided that the above
+.\" copyright notice and this permission notice appear in all copies.
+.\"
+.\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+.\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+.\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+.\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+.\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+.\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+.\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+.\"
+.Dd $Mdocdate: May 25 2011 $
+.Dt FEGETENV 3
+.Os
+.Sh NAME
+.Nm fegetenv ,
+.Nm feholdexcept ,
+.Nm fesetenv ,
+.Nm feupdateenv
+.Nd manage floating-point environment
+.Sh SYNOPSIS
+.Fd #include <fenv.h>
+.Ft int
+.Fn fegetenv "fenv_t *envp"
+.Ft int
+.Fn feholdexcept "fenv_t *envp"
+.Ft int
+.Fn fesetenv "const fenv_t *envp"
+.Ft int
+.Fn feupdateenv "const fenv_t *envp"
+.Sh DESCRIPTION
+These functions manage the floating-point environment -- status
+flags, rounding direction modes and exception masks -- as one entity.
+The
+.Fa envp
+input argument is an object representing floating-point environment.
+The macro
+.Dv FE_DFL_ENV
+represents the default floating-point environment -- the one installed
+at program startup.
+.Pp
+The
+.Fn fegetenv
+function stores the current floating-point environment in the object
+pointed to by
+.Pa envp .
+.Pp
+The
+.Fn feholdexcept
+function saves the current floating-point environment in the object
+pointed to by
+.Pa envp ,
+clears the floating-point status flags, and then installs a non-stop
+(continue on floating-point exceptions) mode for all floating-point
+exceptions.
+.Pp
+The
+.Fn fesetenv
+function establishes the floating-point environment represented by
+the object pointed to by
+.Pa envp .
+The argument
+.Pa envp
+shall point to an object set by a call to
+.Fn fegetenv
+or
+.Fn feholdexcept ,
+or equal the macro
+.Dv FE_DFL_ENV .
+Note that
+.Fn fesetenv
+merely installs the state of the floating-point status flags
+represented through its argument, and does not raise these
+floating-point exceptions.
+.Pp
+The
+.Fn feupdateenv
+function saves the currently raised floating-point exceptions in
+its automatic storage, installs the floating-point environment
+represented by the object pointed to by
+.Pa envp ,
+and then raises the saved floating-point exceptions.
+The argument
+.Pa envp
+shall point to an object set by a call to
+.Fn feholdexcept
+or
+.Fn fegetenv ,
+or equal the macro
+.Dv FE_DFL_ENV .
+.Sh RETURN VALUES
+The
+.Fn fegetenv ,
+.Fn feholdexcept ,
+.Fn fesetenv ,
+and
+.Fn feupdateenv
+functions return zero on success, and non-zero if an error occurred.
+.Sh SEE ALSO
+.Xr feclearexcept 3 ,
+.Xr fegetround 3 ,
+.Xr feenableexcept 3
+.Sh STANDARDS
+The
+.Fn fegetenv ,
+.Fn feholdexcept ,
+.Fn fesetenv ,
+and
+.Fn feupdateenv
+functions conform to
+.St -isoC-99 .
+.Pp
+The return types for
+.Fn fegetenv ,
+.Fn fesetenv ,
+and
+.Fn feupdateenv
+are
+.Vt int
+for alignment with
+.St -isoC-99
+Defect Report #202.
diff --git a/lib/libm/man/fegetround.3 b/lib/libm/man/fegetround.3
new file mode 100644
index 00000000000..1f93010feef
--- /dev/null
+++ b/lib/libm/man/fegetround.3
@@ -0,0 +1,81 @@
+.\"	$OpenBSD: fegetround.3,v 1.1 2011/05/25 21:28:43 martynas Exp $
+.\"
+.\" Copyright (c) 2011 Martynas Venckus <martynas@openbsd.org>
+.\"
+.\" Permission to use, copy, modify, and distribute this software for any
+.\" purpose with or without fee is hereby granted, provided that the above
+.\" copyright notice and this permission notice appear in all copies.
+.\"
+.\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+.\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+.\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+.\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+.\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+.\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+.\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+.\"
+.Dd $Mdocdate: May 25 2011 $
+.Dt FEGETROUND 3
+.Os
+.Sh NAME
+.Nm fegetround ,
+.Nm fesetround
+.Nd control floating-point rounding direction modes
+.Sh SYNOPSIS
+.Fd #include <fenv.h>
+.Ft int
+.Fn fegetround void
+.Ft int
+.Fn fesetround "int round"
+.Sh DESCRIPTION
+These functions provide control of floating-point rounding direction
+modes.
+The
+.Fa round
+input argument is a value specifying a rounding direction mode and
+containing any of the values listed below.
+.Bl -tag -width ".Dv FE_TOWARDZERO"
+.It Dv FE_TONEAREST
+Results are rounded to the closest representable value.
+If the exact result is exactly half way between two representable
+values, the value whose last binary digit is even (zero) is chosen.
+This is the default mode.
+.It Dv FE_DOWNWARD
+Results are rounded towards negative \*[If].
+.It Dv FE_UPWARD
+Results are rounded towards positive \*[If].
+.It Dv FE_TOWARDZERO
+Results are rounded towards zero.
+.El
+.Pp
+The
+.Fn fegetround
+function gets the current rounding direction.
+.Pp
+The
+.Fn fesetround
+function establishes the rounding direction represented by
+.Pa round .
+If the argument is not equal to the value of a rounding direction
+macro, the rounding direction is not changed.
+.Sh RETURN VALUES
+The
+.Fn fegetround
+function returns the current rounding direction.
+The
+.Fn fesetround
+function return zero on success, and non-zero if an error occurred.
+.Sh SEE ALSO
+.Xr feclearexcept 3 ,
+.Xr fegetenv 3 ,
+.Xr feenableexcept 3
+.Sh STANDARDS
+The
+.Fn fegetround ,
+and
+.Fn fesetround
+functions conform to
+.St -isoC-99 .
+.Sh HISTORY
+These functions first appeared in
+.Ox 5.0 .