------------------------------------------------------------------------------ -- -- -- GNU ADA RUNTIME LIBRARY (GNARL) COMPONENTS -- -- -- -- S Y S T E M . I N T E R R U P T _ M A N A G E M E N T -- -- -- -- B o d y -- -- (Version for new GNARL) -- -- -- -- $Revision: 1.1 $ -- -- -- -- Copyright (C) 1991,1992,1993,1994,1995,1996 Florida State University -- -- -- -- GNARL is free software; you can redistribute it and/or modify it under -- -- terms of the GNU General Public License as published by the Free Soft- -- -- ware Foundation; either version 2, or (at your option) any later ver- -- -- sion. GNARL is distributed in the hope that it will be useful, but WITH- -- -- OUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY -- -- or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License -- -- for more details. You should have received a copy of the GNU General -- -- Public License distributed with GNARL; see file COPYING. If not, write -- -- to the Free Software Foundation, 59 Temple Place - Suite 330, Boston, -- -- MA 02111-1307, USA. -- -- -- -- As a special exception, if other files instantiate generics from this -- -- unit, or you link this unit with other files to produce an executable, -- -- this unit does not by itself cause the resulting executable to be -- -- covered by the GNU General Public License. This exception does not -- -- however invalidate any other reasons why the executable file might be -- -- covered by the GNU Public License. -- -- -- -- GNARL was developed by the GNARL team at Florida State University. It is -- -- now maintained by Ada Core Technologies Inc. in cooperation with Florida -- -- State University (http://www.gnat.com). -- -- -- ------------------------------------------------------------------------------ -- This is the FreeBSD PTHREADS version of this package -- This is only a first approximation. -- It should be autogenerated by the m4 macro processor. -- Contributed by Peter Burwood (gnat@arcangel.dircon.co.uk). -- This file performs the system-dependent translation between machine -- exceptions and the Ada exceptions, if any, that should be raised when -- they occur. This version works for FreeBSD. Contributed by -- Daniel M. Eischen (deischen@iworks.InterWorks.org). -- PLEASE DO NOT add any dependences on other packages. -- This package is designed to work with or without tasking support. -- See the other warnings in the package specification before making -- any modifications to this file. -- Make a careful study of all signals available under the OS, -- to see which need to be reserved, kept always unmasked, -- or kept always unmasked. -- Be on the lookout for special signals that -- may be used by the thread library. with Interfaces.C; -- used for int and other types with System.Error_Reporting; -- used for Shutdown with System.OS_Interface; -- used for various Constants, Signal and types package body System.Interrupt_Management is use Interfaces.C; use System.Error_Reporting; use System.OS_Interface; type Interrupt_List is array (Interrupt_ID range <>) of Interrupt_ID; Exception_Interrupts : constant Interrupt_List := (SIGFPE, SIGILL, SIGSEGV, SIGBUS); ---------------------- -- Notify_Exception -- ---------------------- -- This function identifies the Ada exception to be raised using -- the information when the system received a synchronous signal. -- Since this function is machine and OS dependent, different code -- has to be provided for different target. -- Language specs say signal handlers take exactly one arg, even -- though FreeBSD actually supplies three. Ugh! procedure Notify_Exception (signo : Signal; code : Interfaces.C.int; context : access struct_sigcontext); procedure Notify_Exception (signo : Signal; code : Interfaces.C.int; context : access struct_sigcontext) is begin -- As long as we are using a longjmp to return control to the -- exception handler on the runtime stack, we are safe. The original -- signal mask (the one we had before coming into this signal catching -- function) will be restored by the longjmp. Therefore, raising -- an exception in this handler should be a safe operation. -- Check that treatment of exception propagation here -- is consistent with treatment of the abort signal in -- System.Task_Primitives.Operations. -- ????? -- The code below is first approximation. -- It would be nice to figure out more -- precisely what exception has occurred. -- One also should arrange to use an alternate stack for -- recovery from stack overflow. -- I don't understand the Linux kernel code well -- enough to figure out how to do this yet. -- I hope someone will look at this. --Ted Baker -- How can SIGSEGV be split into constraint and storage errors ? -- What should SIGILL really raise ? Some implemenations have -- codes for different types of SIGILL and some raise Storage_Error. -- What causes SIGBUS and should it be caught ? -- Peter Burwood case signo is when SIGFPE => raise Constraint_Error; when SIGILL => raise Constraint_Error; when SIGSEGV => raise Storage_Error; when SIGBUS => raise Storage_Error; when others => pragma Assert (Shutdown ("Unexpected signal")); null; end case; end Notify_Exception; ---------------- -- Initialize -- ---------------- procedure Initialize is act : aliased struct_sigaction; old_act : aliased struct_sigaction; mask : aliased sigset_t; Result : Interfaces.C.int; begin Abort_Task_Interrupt := SIGABRT; -- Change this if you want to use another signal for task abort. -- SIGTERM might be a good one. act.sa_handler := Notify_Exception'Address; act.sa_flags := 16#010#; -- Set sa_flags to SA_NODEFER so that during the handler execution -- we do not change the Signal_Mask to be masked for the Signal. -- This is a temporary fix to the problem that the Signal_Mask is -- not restored after the exception (longjmp) from the handler. -- The right fix should be made in sigsetjmp so that we save -- the Signal_Set and restore it after a longjmp. -- In that case, this field should be changed back to 0. ??? (Dong-Ik) Result := sigemptyset (mask'Access); pragma Assert (Result = 0 or else Shutdown ("GNULLI failure---sigemptyset")); for I in Exception_Interrupts'Range loop Result := sigaddset (mask'Access, Signal (Exception_Interrupts (I))); pragma Assert (Result = 0 or else Shutdown ("GNULLI failure---sigaddset")); end loop; act.sa_mask := mask; for I in Exception_Interrupts'Range loop Keep_Unmasked (Exception_Interrupts (I)) := True; Result := sigaction (Signal (Exception_Interrupts (I)), act'Access, old_act'Unchecked_Access); pragma Assert (Result = 0 or else Shutdown ("GNULLI failure---sigaction")); end loop; Keep_Unmasked (Abort_Task_Interrupt) := true; -- Keep_Unmasked (SIGBUS) := true; Keep_Unmasked (SIGSTOP) := true; Keep_Unmasked (SIGKILL) := true; Keep_Unmasked (SIGINT) := true; -- Keep_Unmasked (SIGEMT) := true; -- Keep_Unmasked (SIGCHLD) := true; -- Keep_Unmasked (SIGALRM) := true; -- ???? The above signals have been found to need to be -- kept unmasked on some systems, per Dong-Ik Oh. -- I don't know whether the MIT/Provenzano threads -- need these or any other signals unmasked at the thread level. -- I hope somebody will take -- the time to look it up. -- Ted Baker -- FreeBSD uses SIGINFO to dump thread status to stdout. If -- the user really wants to attach his own handler, let him. -- FreeBSD pthreads uses setitimer/getitimer for thread scheduling. -- It's not clear, but it looks as if it only needs SIGVTALRM -- in order to handle the setitimer/getitimer operations. We -- could probably allow SIGALARM, but we'll leave it as unmasked -- for now. FreeBSD pthreads also needs SIGCHLD. Keep_Unmasked (SIGCHLD) := true; Keep_Unmasked (SIGALRM) := true; Keep_Unmasked (SIGVTALRM) := true; Reserve := Reserve or Keep_Unmasked or Keep_Masked; Reserve (0) := true; -- We do not have Signal 0 in reality. We just use this value -- to identify non-existent signals (see s-intnam.ads). Therefore, -- Signal 0 should not be used in all signal related operations hence -- mark it as reserved. end Initialize; begin Initialize; end System.Interrupt_Management;