Copyright (c) 1996, Sun Microsystems, Inc.
The contents of this file are subject to the terms of the Common Development and Distribution License (the "License"). You may not use this file except in compliance with the License.
You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE or http://www.opensolaris.org/os/licensing. See the License for the specific language governing permissions and limitations under the License.
When distributing Covered Code, include this CDDL HEADER in each file and include the License file at usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your own identifying information: Portions Copyright [yyyy] [name of copyright owner]
#include <floatingpoint.h> #include <siginfo.h> sigfpe_handler_type sigfpe(sigfpe_code_type code, sigfpe_handler_type hdl);
The sigfpe() function allows signal handling to be specified for particular SIGFPE codes. A call to sigfpe() defines a new handler hdl for a particular SIGFPE code and returns the old handler as the value of the function sigfpe(). Normally handlers are specified as pointers to functions; the special cases SIGFPE_IGNORE, SIGFPE_ABORT, and SIGFPE_DEFAULT allow ignoring, dumping core using abort(3C), or default handling respectively. Default handling is to dump core using abort(3C).
The code argument is usually one of the five IEEE\|754-related SIGFPE codes:
FPE_FLTRES fp_inexact - floating-point inexact result FPE_FLTDIV fp_division - floating-point division by zero FPE_FLTUND fp_underflow - floating-point underflow FPE_FLTOVF fp_overflow - floating-point overflow FPE_FLTINV fp_invalid - floating-point invalid operation
Three steps are required to intercept an IEEE\|754-related SIGFPE code with sigfpe():
1. Set up a handler with sigfpe().
2. Enable the relevant IEEE\|754 trapping capability in the hardware, perhaps by using assembly-language instructions.
3. Perform a floating-point operation that generates the intended IEEE\|754 exception.
The sigfpe() function never changes floating-point hardware mode bits affecting IEEE\|754 trapping. No IEEE\|754-related SIGFPE signals will be generated unless those hardware mode bits are enabled.
SIGFPE signals can be handled using sigfpe(), sigaction(2) or signal(3C). In a particular program, to avoid confusion, use only one of these interfaces to handle SIGFPE signals.
Example 1 Example Of A User-Specified Signal Handler
A user-specified signal handler might look like this:
#include <floatingpoint.h> #include <siginfo.h> #include <ucontext.h> /* * The sample_handler prints out a message then commits suicide. */ void sample_handler(int sig, siginfo_t *sip, ucontext_t *uap) { char *label; switch (sip->si_code) { case FPE_FLTINV: label = "invalid operand"; break; case FPE_FLTRES: label = "inexact"; break; case FPE_FLTDIV: label = "division-by-zero"; break; case FPE_FLTUND: label = "underflow"; break; case FPE_FLTOVF: label = "overflow"; break; default: label = "???"; break; } fprintf(stderr, "FP exception %s (0x%x) occurred at address %p.\en", label, sip->si_code, (void *) sip->si_addr); abort(); }
and it might be set up like this:
#include <floatingpoint.h> #include <siginfo.h> #include <ucontext.h> extern void sample_handler(int, siginfo_t *, ucontext_t *); main(void) { sigfpe_handler_type hdl, old_handler1, old_handler2; /* * save current fp_overflow and fp_invalid handlers; set the new * fp_overflow handler to sample_handler(\|) and set the new * fp_invalid handler to SIGFPE_ABORT (abort on invalid) */ hdl = (sigfpe_handler_type) sample_handler; old_handler1 = sigfpe(FPE_FLTOVF, hdl); old_handler2 = sigfpe(FPE_FLTINV, SIGFPE_ABORT); .\|.\|. /* * restore old fp_overflow and fp_invalid handlers */ sigfpe(FPE_FLTOVF, old_handler1); sigfpe(FPE_FLTINV, old_handler2); }
See attributes(5) for descriptions of the following attributes:
| ATTRIBUTE TYPE | ATTRIBUTE VALUE |
| MT-Level | Safe |
sigaction(2), abort(3C), signal(3C), attributes(5), floatingpoint.h(3HEAD)
The sigfpe() function returns (void(*)())-1 if code is not zero or a defined SIGFPE code.