EIC code displayed by LXR master/​include/​sigsegv.h	Source navigation Diff markup Identifier search General search
	Version: master test Revert   Change

File indexing completed on 2025-06-01 08:54:23

 /* Page fault handling library.
    Copyright (C) 1998-1999, 2002, 2004-2011, 2016-2018, 2021-2022  Bruno Haible <bruno@clisp.org>
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2 of the License, or
    (at your option) any later version.
 
    This program 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 General Public License for more details.
 
    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
 
 #ifndef _SIGSEGV_H
 #define _SIGSEGV_H
 
 /* Get size_t.  */
 #include <stddef.h>
 
 #include <ucontext.h>
 
 /* Correct the value of SIGSTKSZ on some systems.
    glibc >= 2.34: When _GNU_SOURCE is defined, SIGSTKSZ is no longer a
    compile-time constant.  But most programs need a simple constant.
    AIX 64-bit: original value 4096 is too small.
    HP-UX: original value 8192 is too small.
    Solaris 11/x86_64: original value 8192 is too small.  */
 #include <signal.h>
 #if __GLIBC__ >= 2
 # undef SIGSTKSZ
 # if defined __ia64__
 #  define SIGSTKSZ 262144
 # else
 #  define SIGSTKSZ 65536
 # endif
 #endif
 #if defined _AIX && defined _ARCH_PPC64
 # undef SIGSTKSZ
 # define SIGSTKSZ 8192
 #endif
 #if defined __hpux || (defined __sun && (defined __x86_64__ || defined __amd64__))
 # undef SIGSTKSZ
 # define SIGSTKSZ 16384
 #endif
 
 /* HAVE_SIGSEGV_RECOVERY
    is defined if the system supports catching SIGSEGV.  */
 #if 1
 # define HAVE_SIGSEGV_RECOVERY 1
 #endif
 
 /* HAVE_STACK_OVERFLOW_RECOVERY
    is defined if stack overflow can be caught.  */
 #if 1
 # define HAVE_STACK_OVERFLOW_RECOVERY 1
 #endif
 
 
 #ifdef __cplusplus
 extern "C" {
 #endif
 
 #define LIBSIGSEGV_VERSION 0x020E    /* version number: (major<<8) + minor */
 extern int libsigsegv_version;       /* Likewise */
 
 /* -------------------------------------------------------------------------- */
 
 /*
  * The mask of bits that are set to zero in a fault address that gets passed
  * to a global SIGSEGV handler.
  * On some platforms, the precise fault address is not known, only the memory
  * page into which the fault address falls. This is apparently allowed by POSIX:
  * <http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/signal.h.html>
  * says: "For some implementations, the value of si_addr may be inaccurate."
  * In this case, the returned fault address is rounded down to a multiple of
  * getpagesize() = sysconf(_SC_PAGESIZE).
  * On such platforms, we define SIGSEGV_FAULT_ADDRESS_ALIGNMENT to be an upper
  * bound for getpagesize() (and, like getpagesize(), also a power of 2).
  * On the platforms where the returned fault address is the precise one, we
  * define SIGSEGV_FAULT_ADDRESS_ALIGNMENT to 1.
  */
 #if defined __NetBSD__ && (defined __sparc__ || defined __sparc64__)
   /* getpagesize () is 0x1000 or 0x2000, depending on hardware.  */
 # define SIGSEGV_FAULT_ADDRESS_ALIGNMENT 0x2000UL
 #elif defined __linux__ && (defined __s390__ || defined __s390x__)
   /* getpagesize () is 0x1000.  */
 # define SIGSEGV_FAULT_ADDRESS_ALIGNMENT 0x1000UL
 #else
 # define SIGSEGV_FAULT_ADDRESS_ALIGNMENT 1UL
 #endif
 
 /*
  * The type of a global SIGSEGV handler.
  * The fault address, with the bits (SIGSEGV_FAULT_ADDRESS_ALIGNMENT - 1)
  * cleared, is passed as argument.
  * The access type (read access or write access) is not passed; your handler
  * has to know itself how to distinguish these two cases.
  * The second argument is 0, meaning it could also be a stack overflow, or 1,
  * meaning the handler should seriously try to fix the fault.
  * The return value should be nonzero if the handler has done its job
  * and no other handler should be called, or 0 if the handler declines
  * responsibility for the given address.
  *
  * The handler is run at a moment when nothing about the global state of the
  * program is known. Therefore it cannot use facilities that manipulate global
  * variables or locks. In particular, it cannot use malloc(); use mmap()
  * instead. It cannot use fopen(); use open() instead. Etc. All global
  * variables that are accessed by the handler should be marked 'volatile'.
  */
 typedef int (*sigsegv_handler_t) (void* fault_address, int serious);
 
 /*
  * Installs a global SIGSEGV handler.
  * This should be called once only, and it ignores any previously installed
  * SIGSEGV handler.
  * Returns 0 on success, or -1 if the system doesn't support catching SIGSEGV.
  */
 extern int sigsegv_install_handler (sigsegv_handler_t handler);
 
 /*
  * Deinstalls the global SIGSEGV handler.
  * This goes back to the state where no SIGSEGV handler is installed.
  */
 extern void sigsegv_deinstall_handler (void);
 
 #if LIBSIGSEGV_VERSION >= 0x0206
 /*
  * Prepares leaving a SIGSEGV handler (through longjmp or similar means).
  * Control is transferred by calling CONTINUATION with CONT_ARG1, CONT_ARG2,
  * CONT_ARG3 as arguments.
  * CONTINUATION must not return.
  * The sigsegv_leave_handler function may return if called from a SIGSEGV
  * handler; its return value should be used as the handler's return value.
  * The sigsegv_leave_handler function does not return if called from a
  * stack overflow handler.
  */
 extern int sigsegv_leave_handler (void (*continuation) (void*, void*, void*), void* cont_arg1, void* cont_arg2, void* cont_arg3);
 #else /* older versions of libsigsegv */
 /*
  * Prepares leaving a SIGSEGV handler (through longjmp or similar means).
  * Limitation: This function could only be called once on MacOS X.
  */
 extern void sigsegv_leave_handler (void);
 #endif
 
 /*
  * The type of a context passed to a stack overflow handler.
  * This type is system dependent; on some platforms it is an 'ucontext_t *',
  * on some platforms it is a 'struct sigcontext *', on others merely an
  * opaque 'void *'.
  */
 typedef ucontext_t *stackoverflow_context_t;
 
 /*
  * The type of a stack overflow handler.
  * Such a handler should perform a longjmp call in order to reduce the amount
  * of stack needed. It must not return.
  * The emergency argument is 0 when the stack could be repared, or 1 if the
  * application should better save its state and exit now.
  *
  * The handler is run at a moment when nothing about the global state of the
  * program is known. Therefore it cannot use facilities that manipulate global
  * variables or locks. In particular, it cannot use malloc(); use mmap()
  * instead. It cannot use fopen(); use open() instead. Etc. All global
  * variables that are accessed by the handler should be marked 'volatile'.
  */
 typedef void (*stackoverflow_handler_t) (int emergency, stackoverflow_context_t scp);
 
 /*
  * Installs a stack overflow handler.
  * The extra_stack argument is a pointer to a pre-allocated area used as a
  * stack for executing the handler. It typically comes from a static variable
  * or from heap-allocated memoty; placing it on the main stack may fail on
  * some operating systems.
  * Its size, passed in extra_stack_size, should be sufficiently large.  The
  * following code determines an appropriate size:
  *   #include <signal.h>
  *   #ifndef SIGSTKSZ         / * glibc defines SIGSTKSZ for this purpose * /
  *   # define SIGSTKSZ 16384  / * on most platforms, 16 KB are sufficient * /
  *   #endif
  * Returns 0 on success, or -1 if the system doesn't support catching stack
  * overflow.
  */
 extern int stackoverflow_install_handler (stackoverflow_handler_t handler,
                                           void* extra_stack, size_t extra_stack_size);
 
 /*
  * Deinstalls the stack overflow handler.
  */
 extern void stackoverflow_deinstall_handler (void);
 
 /* -------------------------------------------------------------------------- */
 
 /*
  * The following structure and functions permit to define different SIGSEGV
  * policies on different address ranges.
  */
 
 /*
  * The type of a local SIGSEGV handler.
  * The fault address is passed as argument.
  * The second argument is fixed arbitrary user data.
  * The return value should be nonzero if the handler has done its job
  * and no other handler should be called, or 0 if the handler declines
  * responsibility for the given address.
  */
 typedef int (*sigsegv_area_handler_t) (void* fault_address, void* user_arg);
 
 /*
  * This structure represents a table of memory areas (address range intervals),
  * with an local SIGSEGV handler for each.
  */
 typedef
 struct sigsegv_dispatcher {
   void* tree;
 }
 sigsegv_dispatcher;
 
 /*
  * Initializes a sigsegv_dispatcher structure.
  */
 extern void sigsegv_init (sigsegv_dispatcher* dispatcher);
 
 /*
  * Adds a local SIGSEGV handler to a sigsegv_dispatcher structure.
  * It will cover the interval [address..address+len-1].
  * The address and len arguments must be multiples of
  * SIGSEGV_FAULT_ADDRESS_ALIGNMENT.
  * Returns a "ticket" that can be used to remove the handler later.
  */
 extern void* sigsegv_register (sigsegv_dispatcher* dispatcher,
                                void* address, size_t len,
                                sigsegv_area_handler_t handler, void* handler_arg);
 
 /*
  * Removes a local SIGSEGV handler.
  */
 extern void sigsegv_unregister (sigsegv_dispatcher* dispatcher, void* ticket);
 
 /*
  * Call the local SIGSEGV handler responsible for the given fault address.
  * Return the handler's return value. 0 means that no handler has been found,
  * or that a handler was found but declined responsibility.
  */
 extern int sigsegv_dispatch (sigsegv_dispatcher* dispatcher, void* fault_address);
 
 /* -------------------------------------------------------------------------- */
 
 #ifdef __cplusplus
 }
 #endif
 
 #endif /* _SIGSEGV_H */

This page was automatically generated by the 2.3.7 LXR engine. The LXR team