/* IBM RS/6000 "XCOFF" back-end for BFD.
- Copyright (C) 1990, 1991, 1995 Free Software Foundation, Inc.
- FIXME: Can someone provide a transliteration of this name into ASCII?
- Using the following chars caused a compiler warning on HIUX (so I replaced
- them with octal escapes), and isn't useful without an understanding of what
- character set it is.
- Written by Metin G. Ozisik, Mimi Ph\373\364ng-Th\345o V\365,
- and John Gilmore.
+ Copyright (C) 1990-2020 Free Software Foundation, Inc.
+ Written by Metin G. Ozisik, Mimi Phuong-Thao Vo, and John Gilmore.
Archive support from Damon A. Permezel.
Contributed by IBM Corporation and Cygnus Support.
-This file is part of BFD, the Binary File Descriptor library.
+ This file is part of BFD, the Binary File Descriptor library.
-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 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 3 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.
+ 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, write to the Free Software
+ Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
+ MA 02110-1301, USA. */
-You should have received a copy of the GNU General Public License
-along with this program; if not, write to the Free Software
-Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
/* This port currently only handles reading object files, except when
compiled on an RS/6000 host. -- no archive support, no core files.
In all cases, it does not support writing.
- FIXMEmgo comments are left from Metin Ozisik's original port.
-
This is in a separate file from coff-rs6000.c, because it includes
- system include files that conflict with coff/rs6000.h.
- */
+ system include files that conflict with coff/rs6000.h. */
/* Internalcoff.h and coffcode.h modify themselves based on this flag. */
#define RS6000COFF_C 1
-#include "bfd.h"
+/* The AIX 4.1 kernel is obviously compiled with -D_LONG_LONG, so
+ we have to define _LONG_LONG for older versions of gcc to get the
+ proper alignments in the user structure. */
+#if defined(_AIX41) && !defined(_LONG_LONG)
+#define _LONG_LONG
+#endif
+
#include "sysdep.h"
+#include "bfd.h"
#include "libbfd.h"
#ifdef AIX_CORE
#undef SCNHDR
-
-/* ------------------------------------------------------------------------ */
-/* Support for core file stuff.. */
-/* ------------------------------------------------------------------------ */
+/* Support for core file stuff. */
#include <sys/user.h>
+#define __LDINFO_PTRACE32__ /* for __ld_info32 */
+#define __LDINFO_PTRACE64__ /* for __ld_info64 */
#include <sys/ldr.h>
#include <sys/core.h>
+#include <sys/systemcfg.h>
+
+/* Borrowed from <sys/inttypes.h> on recent AIX versions. */
+typedef unsigned long ptr_to_uint;
+
+#define core_hdr(bfd) ((CoreHdr *) bfd->tdata.any)
+
+/* AIX 4.1 changed the names and locations of a few items in the core file.
+ AIX 4.3 defined an entirely new structure, core_dumpx, but kept support for
+ the previous 4.1 structure, core_dump.
+
+ AIX_CORE_DUMPX_CORE is defined (by configure) on AIX 4.3+, and
+ CORE_VERSION_1 is defined (by AIX core.h) as 2 on AIX 4.3+ and as 1 on AIX
+ 4.1 and 4.2. AIX pre-4.1 (aka 3.x) either doesn't define CORE_VERSION_1
+ or else defines it as 0. */
+
+#if defined(CORE_VERSION_1) && !CORE_VERSION_1
+# undef CORE_VERSION_1
+#endif
+/* The following union and macros allow this module to compile on all AIX
+ versions and to handle both core_dumpx and core_dump on 4.3+. CNEW_*()
+ and COLD_*() macros respectively retrieve core_dumpx and core_dump
+ values. */
-/* Number of special purpose registers supported by gdb. This value
- should match `tm.h' in gdb directory. Clean this mess up and use
- the macros in sys/reg.h. FIXMEmgo. */
+/* Union of 32-bit and 64-bit versions of ld_info. */
-#define NUM_OF_SPEC_REGS 7
+typedef union
+{
+#if defined (__ld_info32) || defined (__ld_info64)
+ struct __ld_info32 l32;
+ struct __ld_info64 l64;
+#else
+ struct ld_info l32;
+ struct ld_info l64;
+#endif
+} LdInfo;
+
+/* Union of old and new core dump structures. */
-#define core_hdr(bfd) (((Rs6kCorData*)(bfd->tdata.any))->hdr)
-#define core_datasec(bfd) (((Rs6kCorData*)(bfd->tdata.any))->data_section)
-#define core_stacksec(bfd) (((Rs6kCorData*)(bfd->tdata.any))->stack_section)
-#define core_regsec(bfd) (((Rs6kCorData*)(bfd->tdata.any))->reg_section)
-#define core_reg2sec(bfd) (((Rs6kCorData*)(bfd->tdata.any))->reg2_section)
+typedef union
+{
+#ifdef AIX_CORE_DUMPX_CORE
+ struct core_dumpx new_dump; /* New AIX 4.3+ core dump. */
+#else
+ struct core_dump new_dump; /* For simpler coding. */
+#endif
+#ifndef BFD64 /* Use old only if gdb is 32-bit. */
+ struct core_dump old; /* Old AIX 4.2- core dump, still used on
+ 4.3+ with appropriate SMIT config. */
+#endif
+} CoreHdr;
+
+/* Union of old and new vm_info structures. */
+
+#ifdef CORE_VERSION_1
+typedef union
+{
+#ifdef AIX_CORE_DUMPX_CORE
+ struct vm_infox new_dump;
+#else
+ struct vm_info new_dump;
+#endif
+#ifndef BFD64
+ struct vm_info old;
+#endif
+} VmInfo;
+#endif
+
+/* Return whether CoreHdr C is in new or old format. */
+
+#ifdef AIX_CORE_DUMPX_CORE
+# ifndef BFD64
+# define CORE_NEW(c) (!(c).old.c_entries)
+# else
+# define CORE_NEW(c) 1
+# endif
+#else
+# define CORE_NEW(c) 0
+#endif
+
+/* Return whether CoreHdr C usese core_dumpxx structure.
+
+ FIXME: the core file format version number used here definitely indicates
+ that struct core_dumpxx should be used to represent the core file header,
+ but that may not be the only such format version number. */
+
+#ifdef AIX_5_CORE
+# define CORE_DUMPXX_VERSION 267312562
+# define CNEW_IS_CORE_DUMPXX(c) ((c).new_dump.c_version == CORE_DUMPXX_VERSION)
+#else
+# define CNEW_IS_CORE_DUMPXX(c) 0
+#endif
+
+/* Return the c_stackorg field from struct core_dumpx C. */
+
+#ifdef AIX_CORE_DUMPX_CORE
+# define CNEW_STACKORG(c) (c).c_stackorg
+#else
+# define CNEW_STACKORG(c) 0
+#endif
+
+/* Return the offset to the loader region from struct core_dump C. */
+
+#ifdef AIX_CORE_DUMPX_CORE
+# define CNEW_LOADER(c) (c).c_loader
+#else
+# define CNEW_LOADER(c) 0
+#endif
+
+/* Return the offset to the loader region from struct core_dump C. */
+
+#define COLD_LOADER(c) (c).c_tab
+
+/* Return the c_lsize field from struct core_dumpx C. */
+
+#ifdef AIX_CORE_DUMPX_CORE
+# define CNEW_LSIZE(c) (c).c_lsize
+#else
+# define CNEW_LSIZE(c) 0
+#endif
-/* AIX 4.1 Changed the names and locations of a few items in the core file,
- this seems to be the quickest easiet way to deal with it.
+/* Return the c_dataorg field from struct core_dumpx C. */
- Note however that encoding magic addresses (STACK_END_ADDR) is going
- to be _very_ fragile. But I don't see any easy way to get that info
- right now. */
-#ifdef ALTERNATE_AIX_CORE_FORMAT
-#define CORE_DATA_SIZE_FIELD c_u.U_dsize
-#define CORE_COMM_FIELD c_u.U_comm
-#define SAVE_FIELD c_mst
-#define STACK_END_ADDR 0x2ff23000
+#ifdef AIX_CORE_DUMPX_CORE
+# define CNEW_DATAORG(c) (c).c_dataorg
#else
-#define CORE_DATA_SIZE_FIELD c_u.u_dsize
-#define CORE_COMM_FIELD c_u.u_comm
-#define SAVE_FIELD c_u.u_save
-#define STACK_END_ADDR 0x2ff80000
+# define CNEW_DATAORG(c) 0
#endif
-/* These are stored in the bfd's tdata */
-typedef struct {
- struct core_dump hdr; /* core file header */
- asection *data_section,
- *stack_section,
- *reg_section, /* section for GPRs and special registers. */
- *reg2_section; /* section for FPRs. */
+/* Return the c_datasize field from struct core_dumpx C. */
+
+#ifdef AIX_CORE_DUMPX_CORE
+# define CNEW_DATASIZE(c) (c).c_datasize
+#else
+# define CNEW_DATASIZE(c) 0
+#endif
+
+/* Return the c_impl field from struct core_dumpx C. */
+
+#if defined (HAVE_ST_C_IMPL) || defined (AIX_5_CORE)
+# define CNEW_IMPL(c) (c).c_impl
+#else
+# define CNEW_IMPL(c) 0
+#endif
+
+/* Return the command string from struct core_dumpx C. */
+
+#ifdef AIX_CORE_DUMPX_CORE
+# define CNEW_COMM(c) (c).c_u.U_proc.pi_comm
+#else
+# define CNEW_COMM(c) 0
+#endif
+
+/* Return the command string from struct core_dump C. */
+
+#ifdef CORE_VERSION_1
+# define COLD_COMM(c) (c).c_u.U_comm
+#else
+# define COLD_COMM(c) (c).c_u.u_comm
+#endif
+
+/* Return the struct __context64 pointer from struct core_dumpx C. */
+
+#ifdef AIX_CORE_DUMPX_CORE
+# define CNEW_CONTEXT64(c) (c).c_flt.hctx.r64
+#else
+# define CNEW_CONTEXT64(c) c
+#endif
+
+/* Return the struct mstsave pointer from struct core_dumpx C. */
+
+#ifdef AIX_CORE_DUMPX_CORE
+# define CNEW_MSTSAVE(c) (c).c_flt.hctx.r32
+#else
+# define CNEW_MSTSAVE(c) c
+#endif
+
+/* Return the struct mstsave pointer from struct core_dump C. */
+
+#ifdef CORE_VERSION_1
+# define COLD_MSTSAVE(c) (c).c_mst
+#else
+# define COLD_MSTSAVE(c) (c).c_u.u_save
+#endif
+
+/* Return whether struct core_dumpx is from a 64-bit process. */
+
+#ifdef AIX_CORE_DUMPX_CORE
+# define CNEW_PROC64(c) IS_PROC64(&(c).c_u.U_proc)
+#else
+# define CNEW_PROC64(c) 0
+#endif
+
+/* Magic end-of-stack addresses for old core dumps. This is _very_ fragile,
+ but I don't see any easy way to get that info right now. */
+
+#ifdef CORE_VERSION_1
+# define COLD_STACKEND 0x2ff23000
+#else
+# define COLD_STACKEND 0x2ff80000
+#endif
+
+/* Size of the leading portion that old and new core dump structures have in
+ common. */
+#ifdef AIX_CORE_DUMPX_CORE
+#define CORE_COMMONSZ ((long) &((struct core_dumpx *) 0)->c_entries \
+ + sizeof (((struct core_dumpx *) 0)->c_entries))
+#else
+#define CORE_COMMONSZ ((int) &((struct core_dump *) 0)->c_entries \
+ + sizeof (((struct core_dump *) 0)->c_entries))
+#endif
+/* Define prototypes for certain functions, to avoid a compiler warning
+ saying that they are missing. */
+
+const bfd_cleanup rs6000coff_core_p (bfd *abfd);
+bfd_boolean rs6000coff_core_file_matches_executable_p (bfd *core_bfd,
+ bfd *exec_bfd);
+char * rs6000coff_core_file_failing_command (bfd *abfd);
+int rs6000coff_core_file_failing_signal (bfd *abfd);
+
+/* Try to read into CORE the header from the core file associated with ABFD.
+ Return success. */
+
+static bfd_boolean
+read_hdr (bfd *abfd, CoreHdr *core)
+{
+ bfd_size_type size;
+
+ if (bfd_seek (abfd, (file_ptr) 0, SEEK_SET) != 0)
+ return FALSE;
+
+ /* Read the leading portion that old and new core dump structures have in
+ common. */
+ size = CORE_COMMONSZ;
+ if (bfd_bread (core, size, abfd) != size)
+ return FALSE;
+
+ /* Read the trailing portion of the structure. */
+ if (CORE_NEW (*core))
+ size = sizeof (core->new_dump);
+#ifndef BFD64
+ else
+ size = sizeof (core->old);
+#endif
+ size -= CORE_COMMONSZ;
+ return bfd_bread ((char *) core + CORE_COMMONSZ, size, abfd) == size;
+}
+
+static asection *
+make_bfd_asection (bfd *abfd, const char *name, flagword flags,
+ bfd_size_type size, bfd_vma vma, file_ptr filepos)
+{
+ asection *asect;
+
+ asect = bfd_make_section_anyway_with_flags (abfd, name, flags);
+ if (!asect)
+ return NULL;
- /* This tells us where everything is mapped (shared libraries and so on).
- GDB needs it. */
- asection *ldinfo_section;
-#define core_ldinfosec(bfd) (((Rs6kCorData *)(bfd->tdata.any))->ldinfo_section)
-} Rs6kCorData;
+ asect->size = size;
+ asect->vma = vma;
+ asect->filepos = filepos;
+ asect->alignment_power = 8;
+ return asect;
+}
/* Decide if a given bfd represents a `core' file or not. There really is no
- magic number or anything like, in rs6000coff. */
+ magic number or anything like, in rs6000coff. */
-const bfd_target *
-rs6000coff_core_p (abfd)
- bfd *abfd;
+bfd_cleanup
+rs6000coff_core_p (bfd *abfd)
{
- int fd;
- struct core_dump coredata;
+ CoreHdr core;
struct stat statbuf;
+ bfd_size_type size;
char *tmpptr;
- /* Use bfd_xxx routines, rather than O/S primitives to read coredata. FIXMEmgo */
- fd = open (abfd->filename, O_RDONLY);
- if (fd < 0)
+ /* Values from new and old core structures. */
+ int c_flag;
+ file_ptr c_stack, c_regoff, c_loader;
+ bfd_size_type c_size, c_regsize, c_lsize;
+ bfd_vma c_stackend;
+ void *c_regptr;
+ int proc64;
+
+ if (!read_hdr (abfd, &core))
{
- bfd_set_error (bfd_error_system_call);
+ if (bfd_get_error () != bfd_error_system_call)
+ bfd_set_error (bfd_error_wrong_format);
return NULL;
}
- if (fstat (fd, &statbuf) < 0)
+ /* This isn't the right handler for 64-bit core files on AIX 5.x. */
+ if (CORE_NEW (core) && CNEW_IS_CORE_DUMPXX (core))
{
- bfd_set_error (bfd_error_system_call);
- close (fd);
+ bfd_set_error (bfd_error_wrong_format);
return NULL;
}
- if (read (fd, &coredata, sizeof (struct core_dump))
- != sizeof (struct core_dump))
+
+ /* Copy fields from new or old core structure. */
+ if (CORE_NEW (core))
{
- bfd_set_error (bfd_error_wrong_format);
- close (fd);
- return NULL;
+ c_flag = core.new_dump.c_flag;
+ c_stack = (file_ptr) core.new_dump.c_stack;
+ c_size = core.new_dump.c_size;
+ c_stackend = CNEW_STACKORG (core.new_dump) + c_size;
+ c_lsize = CNEW_LSIZE (core.new_dump);
+ c_loader = CNEW_LOADER (core.new_dump);
+#ifndef BFD64
+ proc64 = CNEW_PROC64 (core.new_dump);
+ }
+ else
+ {
+ c_flag = core.old.c_flag;
+ c_stack = (file_ptr) (ptr_to_uint) core.old.c_stack;
+ c_size = core.old.c_size;
+ c_stackend = COLD_STACKEND;
+ c_lsize = 0x7ffffff;
+ c_loader = (file_ptr) (ptr_to_uint) COLD_LOADER (core.old);
+#endif
+ proc64 = 0;
}
- if (close (fd) < 0)
+ if (proc64)
+ {
+ c_regsize = sizeof (CNEW_CONTEXT64 (core.new_dump));
+ c_regptr = &CNEW_CONTEXT64 (core.new_dump);
+ }
+ else if (CORE_NEW (core))
+ {
+ c_regsize = sizeof (CNEW_MSTSAVE (core.new_dump));
+ c_regptr = &CNEW_MSTSAVE (core.new_dump);
+ }
+#ifndef BFD64
+ else
+ {
+ c_regsize = sizeof (COLD_MSTSAVE (core.old));
+ c_regptr = &COLD_MSTSAVE (core.old);
+ }
+#endif
+ c_regoff = (char *) c_regptr - (char *) &core;
+
+ if (bfd_stat (abfd, &statbuf) < 0)
{
bfd_set_error (bfd_error_system_call);
return NULL;
are always set) (this is based on experimentation on AIX 3.2).
Now, the thing is that GDB users will be surprised
if segments just silently don't appear (well, maybe they would
- think to check "info files", I don't know), but we have no way of
- returning warnings (as opposed to errors).
+ think to check "info files", I don't know).
For the data segment, we have no choice but to keep going if it's
not there, since the default behavior is not to dump it (regardless
if it's not there, we refuse to have anything to do with this core
file. The usefulness of a core dump without a stack segment is pretty
limited anyway. */
-
- if (!(coredata.c_flag & UBLOCK_VALID)
- || !(coredata.c_flag & LE_VALID))
+
+ if (!(c_flag & UBLOCK_VALID)
+ || !(c_flag & LE_VALID))
{
bfd_set_error (bfd_error_wrong_format);
return NULL;
}
- if ((coredata.c_flag & CORE_TRUNC)
- || !(coredata.c_flag & USTACK_VALID))
+ if (!(c_flag & USTACK_VALID))
{
bfd_set_error (bfd_error_file_truncated);
return NULL;
/* Don't check the core file size for a full core, AIX 4.1 includes
additional shared library sections in a full core. */
- if (!(coredata.c_flag & FULL_CORE)
- && ((bfd_vma)coredata.c_stack + coredata.c_size) != statbuf.st_size)
+ if (!(c_flag & (FULL_CORE | CORE_TRUNC)))
{
/* If the size is wrong, it means we're misinterpreting something. */
- bfd_set_error (bfd_error_wrong_format);
- return NULL;
+ if (c_stack + (file_ptr) c_size != statbuf.st_size)
+ {
+ bfd_set_error (bfd_error_wrong_format);
+ return NULL;
+ }
}
/* Sanity check on the c_tab field. */
- if ((u_long) coredata.c_tab < sizeof coredata ||
- (u_long) coredata.c_tab >= statbuf.st_size ||
- (long) coredata.c_tab >= (long)coredata.c_stack)
+ if (!CORE_NEW (core)
+ && (
+#ifndef BFD64
+ c_loader < (file_ptr) sizeof core.old
+#else
+ c_loader < (file_ptr) sizeof core.new_dump
+#endif
+ || c_loader >= statbuf.st_size
+ || c_loader >= c_stack))
{
bfd_set_error (bfd_error_wrong_format);
return NULL;
}
- /* maybe you should alloc space for the whole core chunk over here!! FIXMEmgo */
- tmpptr = (char*)bfd_zalloc (abfd, sizeof (Rs6kCorData));
+ /* Issue warning if the core file was truncated during writing. */
+ if (c_flag & CORE_TRUNC)
+ _bfd_error_handler (_("%pB: warning core file truncated"), abfd);
+
+ /* Allocate core file header. */
+#ifndef BFD64
+ size = CORE_NEW (core) ? sizeof (core.new_dump) : sizeof (core.old);
+#else
+ size = sizeof (core.new_dump);
+#endif
+ tmpptr = (char *) bfd_zalloc (abfd, (bfd_size_type) size);
if (!tmpptr)
+ return NULL;
+
+ /* Copy core file header. */
+ memcpy (tmpptr, &core, size);
+ set_tdata (abfd, tmpptr);
+
+ /* Set architecture. */
+ if (CORE_NEW (core))
{
- bfd_set_error (bfd_error_no_memory);
- return NULL;
+ enum bfd_architecture arch;
+ unsigned long mach;
+
+ switch (CNEW_IMPL (core.new_dump))
+ {
+ case POWER_RS1:
+ case POWER_RSC:
+ case POWER_RS2:
+ arch = bfd_arch_rs6000;
+ mach = bfd_mach_rs6k;
+ break;
+ default:
+ arch = bfd_arch_powerpc;
+ mach = bfd_mach_ppc;
+ break;
+ }
+ bfd_default_set_arch_mach (abfd, arch, mach);
}
-
- set_tdata (abfd, tmpptr);
- /* Copy core file header. */
- core_hdr (abfd) = coredata;
+ /* .stack section. */
+ if (!make_bfd_asection (abfd, ".stack",
+ SEC_ALLOC | SEC_LOAD | SEC_HAS_CONTENTS,
+ c_size, c_stackend - c_size, c_stack))
+ goto fail;
- /* .stack section. */
- if ((core_stacksec (abfd) = (asection*) bfd_zalloc (abfd, sizeof (asection)))
- == NULL) {
- bfd_set_error (bfd_error_no_memory);
- /* bfd_release (abfd, ???? ) */
- return NULL;
- }
- core_stacksec (abfd)->name = ".stack";
- core_stacksec (abfd)->flags = SEC_ALLOC + SEC_LOAD + SEC_HAS_CONTENTS;
- core_stacksec (abfd)->_raw_size = coredata.c_size;
- core_stacksec (abfd)->vma = STACK_END_ADDR - coredata.c_size;
- core_stacksec (abfd)->filepos = (int)coredata.c_stack; /*???? */
-
- /* .reg section for GPRs and special registers. */
- if ((core_regsec (abfd) = (asection*) bfd_zalloc (abfd, sizeof (asection)))
- == NULL) {
- bfd_set_error (bfd_error_no_memory);
- /* bfd_release (abfd, ???? ) */
- return NULL;
- }
- core_regsec (abfd)->name = ".reg";
- core_regsec (abfd)->flags = SEC_HAS_CONTENTS;
- core_regsec (abfd)->_raw_size = (32 + NUM_OF_SPEC_REGS) * 4;
- core_regsec (abfd)->vma = 0; /* not used?? */
- core_regsec (abfd)->filepos =
- (char*)&coredata.SAVE_FIELD - (char*)&coredata;
-
- /* .reg2 section for FPRs (floating point registers). */
- if ((core_reg2sec (abfd) = (asection*) bfd_zalloc (abfd, sizeof (asection)))
- == NULL) {
- bfd_set_error (bfd_error_no_memory);
- /* bfd_release (abfd, ???? ) */
- return NULL;
- }
- core_reg2sec (abfd)->name = ".reg2";
- core_reg2sec (abfd)->flags = SEC_HAS_CONTENTS;
- core_reg2sec (abfd)->_raw_size = 8 * 32; /* 32 FPRs. */
- core_reg2sec (abfd)->vma = 0; /* not used?? */
- core_reg2sec (abfd)->filepos =
- (char*)&coredata.SAVE_FIELD.fpr[0] - (char*)&coredata;
-
- if ((core_ldinfosec (abfd) = (asection*) bfd_zalloc (abfd, sizeof (asection)))
- == NULL) {
- bfd_set_error (bfd_error_no_memory);
- /* bfd_release (abfd, ???? ) */
- return NULL;
- }
- core_ldinfosec (abfd)->name = ".ldinfo";
- core_ldinfosec (abfd)->flags = SEC_HAS_CONTENTS;
- /* To actually find out how long this section is in this particular
+ /* .reg section for all registers. */
+ if (!make_bfd_asection (abfd, ".reg",
+ SEC_HAS_CONTENTS,
+ c_regsize, (bfd_vma) 0, c_regoff))
+ goto fail;
+
+ /* .ldinfo section.
+ To actually find out how long this section is in this particular
core dump would require going down the whole list of struct ld_info's.
See if we can just fake it. */
- core_ldinfosec (abfd)->_raw_size = 0x7fffffff;
- /* Not relevant for ldinfo section. */
- core_ldinfosec (abfd)->vma = 0;
- core_ldinfosec (abfd)->filepos = (file_ptr) coredata.c_tab;
-
- /* set up section chain here. */
- abfd->section_count = 4;
- abfd->sections = core_stacksec (abfd);
- core_stacksec (abfd)->next = core_regsec(abfd);
- core_regsec (abfd)->next = core_reg2sec (abfd);
- core_reg2sec (abfd)->next = core_ldinfosec (abfd);
- core_ldinfosec (abfd)->next = NULL;
-
- if (coredata.c_flag & FULL_CORE)
+ if (!make_bfd_asection (abfd, ".ldinfo",
+ SEC_HAS_CONTENTS,
+ c_lsize, (bfd_vma) 0, c_loader))
+ goto fail;
+
+#ifndef CORE_VERSION_1
+ /* .data section if present.
+ AIX 3 dumps the complete data section and sets FULL_CORE if the
+ ulimit is large enough, otherwise the data section is omitted.
+ AIX 4 sets FULL_CORE even if the core file is truncated, we have
+ to examine core.c_datasize below to find out the actual size of
+ the .data section. */
+ if (c_flag & FULL_CORE)
{
- asection *sec = (asection *) bfd_zalloc (abfd, sizeof (asection));
- if (sec == NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return NULL;
- }
- sec->name = ".data";
- sec->flags = SEC_ALLOC | SEC_LOAD | SEC_HAS_CONTENTS;
- sec->_raw_size = coredata.CORE_DATA_SIZE_FIELD;
- sec->vma = CDATA_ADDR (coredata.CORE_DATA_SIZE_FIELD);
- sec->filepos = (int)coredata.c_stack + coredata.c_size;
-
- sec->next = abfd->sections;
- abfd->sections = sec;
- ++abfd->section_count;
+ if (!make_bfd_asection (abfd, ".data",
+ SEC_ALLOC | SEC_LOAD | SEC_HAS_CONTENTS,
+ (bfd_size_type) core.old.c_u.u_dsize,
+ (bfd_vma)
+ CDATA_ADDR (core.old.c_u.u_dsize),
+ c_stack + c_size))
+ goto fail;
}
+#endif
- return abfd->xvec; /* this is garbage for now. */
-}
+#ifdef CORE_VERSION_1
+ /* AIX 4 adds data sections from loaded objects to the core file,
+ which can be found by examining ldinfo, and anonymously mmapped
+ regions. */
+ {
+ LdInfo ldinfo;
+ bfd_size_type ldi_datasize;
+ file_ptr ldi_core;
+ uint ldi_next;
+ bfd_vma ldi_dataorg;
+ bfd_vma core_dataorg;
+
+ /* Fields from new and old core structures. */
+ bfd_size_type c_datasize, c_vmregions;
+ file_ptr c_data, c_vmm;
+
+ if (CORE_NEW (core))
+ {
+ c_datasize = CNEW_DATASIZE (core.new_dump);
+ c_data = (file_ptr) core.new_dump.c_data;
+ c_vmregions = core.new_dump.c_vmregions;
+ c_vmm = (file_ptr) core.new_dump.c_vmm;
+ }
+#ifndef BFD64
+ else
+ {
+ c_datasize = core.old.c_datasize;
+ c_data = (file_ptr) (ptr_to_uint) core.old.c_data;
+ c_vmregions = core.old.c_vmregions;
+ c_vmm = (file_ptr) (ptr_to_uint) core.old.c_vmm;
+ }
+#endif
+ /* .data section from executable. */
+ if (c_datasize)
+ {
+ /* If Large Memory Model is used, then the .data segment should start from
+ BDATAORG which has been defined in the system header files. */
+
+ if (c_flag & CORE_BIGDATA)
+ core_dataorg = BDATAORG;
+ else
+ core_dataorg = CDATA_ADDR (c_datasize);
+
+ if (!make_bfd_asection (abfd, ".data",
+ SEC_ALLOC | SEC_LOAD | SEC_HAS_CONTENTS,
+ c_datasize,
+ (bfd_vma) core_dataorg,
+ c_data))
+ goto fail;
+ }
+
+ /* .data sections from loaded objects. */
+ if (proc64)
+ size = (unsigned long) ((LdInfo *) 0)->l64.ldinfo_filename;
+ else
+ size = (unsigned long) ((LdInfo *) 0)->l32.ldinfo_filename;
+
+ while (1)
+ {
+ if (bfd_seek (abfd, c_loader, SEEK_SET) != 0)
+ goto fail;
+ if (bfd_bread (&ldinfo, size, abfd) != size)
+ goto fail;
+
+ if (proc64)
+ {
+ ldi_core = ldinfo.l64.ldinfo_core;
+ ldi_datasize = ldinfo.l64.ldinfo_datasize;
+ ldi_dataorg = (bfd_vma) ldinfo.l64.ldinfo_dataorg;
+ ldi_next = ldinfo.l64.ldinfo_next;
+ }
+ else
+ {
+ ldi_core = ldinfo.l32.ldinfo_core;
+ ldi_datasize = ldinfo.l32.ldinfo_datasize;
+ ldi_dataorg = (bfd_vma) (ptr_to_uint) ldinfo.l32.ldinfo_dataorg;
+ ldi_next = ldinfo.l32.ldinfo_next;
+ }
+
+ if (ldi_core)
+ if (!make_bfd_asection (abfd, ".data",
+ SEC_ALLOC | SEC_LOAD | SEC_HAS_CONTENTS,
+ ldi_datasize, ldi_dataorg, ldi_core))
+ goto fail;
+
+ if (ldi_next == 0)
+ break;
+ c_loader += ldi_next;
+ }
+
+ /* .vmdata sections from anonymously mmapped regions. */
+ if (c_vmregions)
+ {
+ bfd_size_type i;
+
+ if (bfd_seek (abfd, c_vmm, SEEK_SET) != 0)
+ goto fail;
+
+ for (i = 0; i < c_vmregions; i++)
+ {
+ VmInfo vminfo;
+ bfd_size_type vminfo_size;
+ file_ptr vminfo_offset;
+ bfd_vma vminfo_addr;
+
+#ifndef BFD64
+ size = CORE_NEW (core) ? sizeof (vminfo.new_dump) : sizeof (vminfo.old);
+#else
+ size = sizeof (vminfo.new_dump);
+#endif
+ if (bfd_bread (&vminfo, size, abfd) != size)
+ goto fail;
+
+ if (CORE_NEW (core))
+ {
+ vminfo_addr = (bfd_vma) vminfo.new_dump.vminfo_addr;
+ vminfo_size = vminfo.new_dump.vminfo_size;
+ vminfo_offset = vminfo.new_dump.vminfo_offset;
+ }
+#ifndef BFD64
+ else
+ {
+ vminfo_addr = (bfd_vma) (ptr_to_uint) vminfo.old.vminfo_addr;
+ vminfo_size = vminfo.old.vminfo_size;
+ vminfo_offset = vminfo.old.vminfo_offset;
+ }
+#endif
+
+ if (vminfo_offset)
+ if (!make_bfd_asection (abfd, ".vmdata",
+ SEC_ALLOC | SEC_LOAD | SEC_HAS_CONTENTS,
+ vminfo_size, vminfo_addr,
+ vminfo_offset))
+ goto fail;
+ }
+ }
+ }
+#endif
+
+ return _bfd_no_cleanup;
+
+ fail:
+ bfd_release (abfd, abfd->tdata.any);
+ abfd->tdata.any = NULL;
+ bfd_section_list_clear (abfd);
+ return NULL;
+}
+/* Return `TRUE' if given core is from the given executable. */
-/* return `true' if given core is from the given executable.. */
-boolean
-rs6000coff_core_file_matches_executable_p (core_bfd, exec_bfd)
- bfd *core_bfd;
- bfd *exec_bfd;
+bfd_boolean
+rs6000coff_core_file_matches_executable_p (bfd *core_bfd, bfd *exec_bfd)
{
- FILE *fd;
- struct core_dump coredata;
- struct ld_info ldinfo;
- char pathname [1024];
+ CoreHdr core;
+ bfd_size_type size;
+ char *path, *s;
+ size_t alloc;
const char *str1, *str2;
+ bfd_boolean ret;
+ file_ptr c_loader;
+
+ if (!read_hdr (core_bfd, &core))
+ return FALSE;
+
+ if (CORE_NEW (core))
+ c_loader = CNEW_LOADER (core.new_dump);
+#ifndef BFD64
+ else
+ c_loader = (file_ptr) (ptr_to_uint) COLD_LOADER (core.old);
+#endif
+
+ if (CORE_NEW (core) && CNEW_PROC64 (core.new_dump))
+ size = (int) ((LdInfo *) 0)->l64.ldinfo_filename;
+ else
+ size = (int) ((LdInfo *) 0)->l32.ldinfo_filename;
+
+ if (bfd_seek (core_bfd, c_loader + size, SEEK_SET) != 0)
+ return FALSE;
- /* Use bfd_xxx routines, rather than O/S primitives, do error checking!!
- FIXMEmgo */
- /* Actually should be able to use bfd_get_section_contents now that
- we have a .ldinfo section. */
- fd = fopen (core_bfd->filename, FOPEN_RB);
-
- fread (&coredata, sizeof (struct core_dump), 1, fd);
- fseek (fd, (long)coredata.c_tab, 0);
- fread (&ldinfo, (char*)&ldinfo.ldinfo_filename[0] - (char*)&ldinfo.ldinfo_next,
- 1, fd);
- fscanf (fd, "%s", pathname);
-
- str1 = strrchr (pathname, '/');
+ alloc = 100;
+ path = bfd_malloc ((bfd_size_type) alloc);
+ if (path == NULL)
+ return FALSE;
+ s = path;
+
+ while (1)
+ {
+ if (bfd_bread (s, (bfd_size_type) 1, core_bfd) != 1)
+ {
+ free (path);
+ return FALSE;
+ }
+ if (*s == '\0')
+ break;
+ ++s;
+ if (s == path + alloc)
+ {
+ char *n;
+
+ alloc *= 2;
+ n = bfd_realloc (path, (bfd_size_type) alloc);
+ if (n == NULL)
+ {
+ free (path);
+ return FALSE;
+ }
+ s = n + (path - s);
+ path = n;
+ }
+ }
+
+ str1 = strrchr (path, '/');
str2 = strrchr (exec_bfd->filename, '/');
/* step over character '/' */
- str1 = str1 ? str1+1 : &pathname[0];
- str2 = str2 ? str2+1 : exec_bfd->filename;
+ str1 = str1 != NULL ? str1 + 1 : path;
+ str2 = str2 != NULL ? str2 + 1 : exec_bfd->filename;
+
+ if (strcmp (str1, str2) == 0)
+ ret = TRUE;
+ else
+ ret = FALSE;
+
+ free (path);
- fclose (fd);
- return strcmp (str1, str2) == 0;
+ return ret;
}
char *
-rs6000coff_core_file_failing_command (abfd)
- bfd *abfd;
+rs6000coff_core_file_failing_command (bfd *abfd)
{
- char *com = core_hdr (abfd).CORE_COMM_FIELD;
+ CoreHdr *core = core_hdr (abfd);
+#ifndef BFD64
+ char *com = CORE_NEW (*core) ?
+ CNEW_COMM (core->new_dump) : COLD_COMM (core->old);
+#else
+ char *com = CNEW_COMM (core->new_dump);
+#endif
+
if (*com)
return com;
else
}
int
-rs6000coff_core_file_failing_signal (abfd)
- bfd *abfd;
-{
- return core_hdr (abfd).c_signo;
-}
-
-
-boolean
-rs6000coff_get_section_contents (abfd, section, location, offset, count)
- bfd *abfd;
- sec_ptr section;
- PTR location;
- file_ptr offset;
- int count;
+rs6000coff_core_file_failing_signal (bfd *abfd)
{
- if (count == 0)
- return true;
-
- /* Reading a core file's sections will be slightly different. For the
- rest of them we can use bfd_generic_get_section_contents () I suppose. */
- /* Make sure this routine works for any bfd and any section. FIXMEmgo. */
-
- if (abfd->format == bfd_core && strcmp (section->name, ".reg") == 0) {
-
- struct mstsave mstatus;
- int regoffset = (char*)&mstatus.gpr[0] - (char*)&mstatus;
-
- /* Assert that the only way this code will be executed is reading the
- whole section. */
- if (offset || count != (sizeof(mstatus.gpr) + (4 * NUM_OF_SPEC_REGS)))
- fprintf (stderr, "ERROR! in rs6000coff_get_section_contents()\n");
-
- /* for `.reg' section, `filepos' is a pointer to the `mstsave' structure
- in the core file. */
-
- /* read GPR's into the location. */
- if ( bfd_seek(abfd, section->filepos + regoffset, SEEK_SET) == -1
- || bfd_read(location, sizeof (mstatus.gpr), 1, abfd) != sizeof (mstatus.gpr))
- return (false); /* on error */
-
- /* increment location to the beginning of special registers in the section,
- reset register offset value to the beginning of first special register
- in mstsave structure, and read special registers. */
-
- location = (PTR) ((char*)location + sizeof (mstatus.gpr));
- regoffset = (char*)&mstatus.iar - (char*)&mstatus;
-
- if ( bfd_seek(abfd, section->filepos + regoffset, SEEK_SET) == -1
- || bfd_read(location, 4 * NUM_OF_SPEC_REGS, 1, abfd) !=
- 4 * NUM_OF_SPEC_REGS)
- return (false); /* on error */
-
- /* increment location address, and read the special registers.. */
- /* FIXMEmgo */
- return (true);
- }
-
- /* else, use default bfd section content transfer. */
- else
- return _bfd_generic_get_section_contents
- (abfd, section, location, offset, count);
+ CoreHdr *core = core_hdr (abfd);
+#ifndef BFD64
+ return CORE_NEW (*core) ? core->new_dump.c_signo : core->old.c_signo;
+#else
+ return core->new_dump.c_signo;
+#endif
}
#endif /* AIX_CORE */
projects / deliverable / binutils-gdb.git / blobdiff