/* Definitions for BFD wrappers used by GDB.
Copyright (C) 2011-2020 Free Software Foundation, Inc.
This file is part of GDB.
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.
You should have received a copy of the GNU General Public License
along with this program. If not, see . */
#include "defs.h"
#include "gdb_bfd.h"
#include "ui-out.h"
#include "gdbcmd.h"
#include "hashtab.h"
#include "gdbsupport/filestuff.h"
#ifdef HAVE_MMAP
#include
#ifndef MAP_FAILED
#define MAP_FAILED ((void *) -1)
#endif
#endif
#include "target.h"
#include "gdb/fileio.h"
#include "inferior.h"
/* An object of this type is stored in the section's user data when
mapping a section. */
struct gdb_bfd_section_data
{
/* Size of the data. */
bfd_size_type size;
/* If the data was mmapped, this is the length of the map. */
bfd_size_type map_len;
/* The data. If NULL, the section data has not been read. */
void *data;
/* If the data was mmapped, this is the map address. */
void *map_addr;
};
/* A hash table holding every BFD that gdb knows about. This is not
to be confused with 'gdb_bfd_cache', which is used for sharing
BFDs; in contrast, this hash is used just to implement
"maint info bfd". */
static htab_t all_bfds;
/* An object of this type is stored in each BFD's user data. */
struct gdb_bfd_data
{
gdb_bfd_data (bfd *abfd)
: mtime (bfd_get_mtime (abfd)),
size (bfd_get_size (abfd)),
relocation_computed (0),
needs_relocations (0),
crc_computed (0)
{
struct stat buf;
if (bfd_stat (abfd, &buf) == 0)
{
inode = buf.st_ino;
device_id = buf.st_dev;
}
else
{
/* The stat failed. */
inode = 0;
device_id = 0;
}
}
~gdb_bfd_data ()
{
}
/* The reference count. */
int refc = 1;
/* The mtime of the BFD at the point the cache entry was made. */
time_t mtime;
/* The file size (in bytes) at the point the cache entry was made. */
off_t size;
/* The inode of the file at the point the cache entry was made. */
ino_t inode;
/* The device id of the file at the point the cache entry was made. */
dev_t device_id;
/* This is true if we have determined whether this BFD has any
sections requiring relocation. */
unsigned int relocation_computed : 1;
/* This is true if any section needs relocation. */
unsigned int needs_relocations : 1;
/* This is true if we have successfully computed the file's CRC. */
unsigned int crc_computed : 1;
/* The file's CRC. */
unsigned long crc = 0;
/* If the BFD comes from an archive, this points to the archive's
BFD. Otherwise, this is NULL. */
bfd *archive_bfd = nullptr;
/* Table of all the bfds this bfd has included. */
std::vector included_bfds;
/* The registry. */
REGISTRY_FIELDS = {};
};
#define GDB_BFD_DATA_ACCESSOR(ABFD) \
((struct gdb_bfd_data *) bfd_usrdata (ABFD))
DEFINE_REGISTRY (bfd, GDB_BFD_DATA_ACCESSOR)
/* A hash table storing all the BFDs maintained in the cache. */
static htab_t gdb_bfd_cache;
/* When true gdb will reuse an existing bfd object if the filename,
modification time, and file size all match. */
static bool bfd_sharing = true;
static void
show_bfd_sharing (struct ui_file *file, int from_tty,
struct cmd_list_element *c, const char *value)
{
fprintf_filtered (file, _("BFD sharing is %s.\n"), value);
}
/* When non-zero debugging of the bfd caches is enabled. */
static unsigned int debug_bfd_cache;
static void
show_bfd_cache_debug (struct ui_file *file, int from_tty,
struct cmd_list_element *c, const char *value)
{
fprintf_filtered (file, _("BFD cache debugging is %s.\n"), value);
}
/* The type of an object being looked up in gdb_bfd_cache. We use
htab's capability of storing one kind of object (BFD in this case)
and using a different sort of object for searching. */
struct gdb_bfd_cache_search
{
/* The filename. */
const char *filename;
/* The mtime. */
time_t mtime;
/* The file size (in bytes). */
off_t size;
/* The inode of the file. */
ino_t inode;
/* The device id of the file. */
dev_t device_id;
};
/* A hash function for BFDs. */
static hashval_t
hash_bfd (const void *b)
{
const bfd *abfd = (const struct bfd *) b;
/* It is simplest to just hash the filename. */
return htab_hash_string (bfd_get_filename (abfd));
}
/* An equality function for BFDs. Note that this expects the caller
to search using struct gdb_bfd_cache_search only, not BFDs. */
static int
eq_bfd (const void *a, const void *b)
{
const bfd *abfd = (const struct bfd *) a;
const struct gdb_bfd_cache_search *s
= (const struct gdb_bfd_cache_search *) b;
struct gdb_bfd_data *gdata = (struct gdb_bfd_data *) bfd_usrdata (abfd);
return (gdata->mtime == s->mtime
&& gdata->size == s->size
&& gdata->inode == s->inode
&& gdata->device_id == s->device_id
&& strcmp (bfd_get_filename (abfd), s->filename) == 0);
}
/* See gdb_bfd.h. */
int
is_target_filename (const char *name)
{
return startswith (name, TARGET_SYSROOT_PREFIX);
}
/* See gdb_bfd.h. */
int
gdb_bfd_has_target_filename (struct bfd *abfd)
{
return is_target_filename (bfd_get_filename (abfd));
}
/* Return the system error number corresponding to ERRNUM. */
static int
fileio_errno_to_host (int errnum)
{
switch (errnum)
{
case FILEIO_EPERM:
return EPERM;
case FILEIO_ENOENT:
return ENOENT;
case FILEIO_EINTR:
return EINTR;
case FILEIO_EIO:
return EIO;
case FILEIO_EBADF:
return EBADF;
case FILEIO_EACCES:
return EACCES;
case FILEIO_EFAULT:
return EFAULT;
case FILEIO_EBUSY:
return EBUSY;
case FILEIO_EEXIST:
return EEXIST;
case FILEIO_ENODEV:
return ENODEV;
case FILEIO_ENOTDIR:
return ENOTDIR;
case FILEIO_EISDIR:
return EISDIR;
case FILEIO_EINVAL:
return EINVAL;
case FILEIO_ENFILE:
return ENFILE;
case FILEIO_EMFILE:
return EMFILE;
case FILEIO_EFBIG:
return EFBIG;
case FILEIO_ENOSPC:
return ENOSPC;
case FILEIO_ESPIPE:
return ESPIPE;
case FILEIO_EROFS:
return EROFS;
case FILEIO_ENOSYS:
return ENOSYS;
case FILEIO_ENAMETOOLONG:
return ENAMETOOLONG;
}
return -1;
}
/* Wrapper for target_fileio_open suitable for passing as the
OPEN_FUNC argument to gdb_bfd_openr_iovec. The supplied
OPEN_CLOSURE is unused. */
static void *
gdb_bfd_iovec_fileio_open (struct bfd *abfd, void *inferior)
{
const char *filename = bfd_get_filename (abfd);
int fd, target_errno;
int *stream;
gdb_assert (is_target_filename (filename));
fd = target_fileio_open_warn_if_slow ((struct inferior *) inferior,
filename
+ strlen (TARGET_SYSROOT_PREFIX),
FILEIO_O_RDONLY, 0,
&target_errno);
if (fd == -1)
{
errno = fileio_errno_to_host (target_errno);
bfd_set_error (bfd_error_system_call);
return NULL;
}
stream = XCNEW (int);
*stream = fd;
return stream;
}
/* Wrapper for target_fileio_pread suitable for passing as the
PREAD_FUNC argument to gdb_bfd_openr_iovec. */
static file_ptr
gdb_bfd_iovec_fileio_pread (struct bfd *abfd, void *stream, void *buf,
file_ptr nbytes, file_ptr offset)
{
int fd = *(int *) stream;
int target_errno;
file_ptr pos, bytes;
pos = 0;
while (nbytes > pos)
{
QUIT;
bytes = target_fileio_pread (fd, (gdb_byte *) buf + pos,
nbytes - pos, offset + pos,
&target_errno);
if (bytes == 0)
/* Success, but no bytes, means end-of-file. */
break;
if (bytes == -1)
{
errno = fileio_errno_to_host (target_errno);
bfd_set_error (bfd_error_system_call);
return -1;
}
pos += bytes;
}
return pos;
}
/* Wrapper for target_fileio_close suitable for passing as the
CLOSE_FUNC argument to gdb_bfd_openr_iovec. */
static int
gdb_bfd_iovec_fileio_close (struct bfd *abfd, void *stream)
{
int fd = *(int *) stream;
int target_errno;
xfree (stream);
/* Ignore errors on close. These may happen with remote
targets if the connection has already been torn down. */
target_fileio_close (fd, &target_errno);
/* Zero means success. */
return 0;
}
/* Wrapper for target_fileio_fstat suitable for passing as the
STAT_FUNC argument to gdb_bfd_openr_iovec. */
static int
gdb_bfd_iovec_fileio_fstat (struct bfd *abfd, void *stream,
struct stat *sb)
{
int fd = *(int *) stream;
int target_errno;
int result;
result = target_fileio_fstat (fd, sb, &target_errno);
if (result == -1)
{
errno = fileio_errno_to_host (target_errno);
bfd_set_error (bfd_error_system_call);
}
return result;
}
/* See gdb_bfd.h. */
gdb_bfd_ref_ptr
gdb_bfd_open (const char *name, const char *target, int fd)
{
hashval_t hash;
void **slot;
bfd *abfd;
struct gdb_bfd_cache_search search;
struct stat st;
if (is_target_filename (name))
{
if (!target_filesystem_is_local ())
{
gdb_assert (fd == -1);
return gdb_bfd_openr_iovec (name, target,
gdb_bfd_iovec_fileio_open,
current_inferior (),
gdb_bfd_iovec_fileio_pread,
gdb_bfd_iovec_fileio_close,
gdb_bfd_iovec_fileio_fstat);
}
name += strlen (TARGET_SYSROOT_PREFIX);
}
if (gdb_bfd_cache == NULL)
gdb_bfd_cache = htab_create_alloc (1, hash_bfd, eq_bfd, NULL,
xcalloc, xfree);
if (fd == -1)
{
fd = gdb_open_cloexec (name, O_RDONLY | O_BINARY, 0);
if (fd == -1)
{
bfd_set_error (bfd_error_system_call);
return NULL;
}
}
search.filename = name;
if (fstat (fd, &st) < 0)
{
/* Weird situation here. */
search.mtime = 0;
search.size = 0;
search.inode = 0;
search.device_id = 0;
}
else
{
search.mtime = st.st_mtime;
search.size = st.st_size;
search.inode = st.st_ino;
search.device_id = st.st_dev;
}
/* Note that this must compute the same result as hash_bfd. */
hash = htab_hash_string (name);
/* Note that we cannot use htab_find_slot_with_hash here, because
opening the BFD may fail; and this would violate hashtab
invariants. */
abfd = (struct bfd *) htab_find_with_hash (gdb_bfd_cache, &search, hash);
if (bfd_sharing && abfd != NULL)
{
if (debug_bfd_cache)
fprintf_unfiltered (gdb_stdlog,
"Reusing cached bfd %s for %s\n",
host_address_to_string (abfd),
bfd_get_filename (abfd));
close (fd);
return gdb_bfd_ref_ptr::new_reference (abfd);
}
abfd = bfd_fopen (name, target, FOPEN_RB, fd);
if (abfd == NULL)
return NULL;
if (debug_bfd_cache)
fprintf_unfiltered (gdb_stdlog,
"Creating new bfd %s for %s\n",
host_address_to_string (abfd),
bfd_get_filename (abfd));
if (bfd_sharing)
{
slot = htab_find_slot_with_hash (gdb_bfd_cache, &search, hash, INSERT);
gdb_assert (!*slot);
*slot = abfd;
}
return gdb_bfd_ref_ptr::new_reference (abfd);
}
/* A helper function that releases any section data attached to the
BFD. */
static void
free_one_bfd_section (bfd *abfd, asection *sectp, void *ignore)
{
struct gdb_bfd_section_data *sect
= (struct gdb_bfd_section_data *) bfd_section_userdata (sectp);
if (sect != NULL && sect->data != NULL)
{
#ifdef HAVE_MMAP
if (sect->map_addr != NULL)
{
int res;
res = munmap (sect->map_addr, sect->map_len);
gdb_assert (res == 0);
}
else
#endif
xfree (sect->data);
}
}
/* Close ABFD, and warn if that fails. */
static int
gdb_bfd_close_or_warn (struct bfd *abfd)
{
int ret;
const char *name = bfd_get_filename (abfd);
bfd_map_over_sections (abfd, free_one_bfd_section, NULL);
ret = bfd_close (abfd);
if (!ret)
warning (_("cannot close \"%s\": %s"),
name, bfd_errmsg (bfd_get_error ()));
return ret;
}
/* See gdb_bfd.h. */
void
gdb_bfd_ref (struct bfd *abfd)
{
struct gdb_bfd_data *gdata;
void **slot;
if (abfd == NULL)
return;
gdata = (struct gdb_bfd_data *) bfd_usrdata (abfd);
if (debug_bfd_cache)
fprintf_unfiltered (gdb_stdlog,
"Increase reference count on bfd %s (%s)\n",
host_address_to_string (abfd),
bfd_get_filename (abfd));
if (gdata != NULL)
{
gdata->refc += 1;
return;
}
/* Ask BFD to decompress sections in bfd_get_full_section_contents. */
abfd->flags |= BFD_DECOMPRESS;
gdata = new gdb_bfd_data (abfd);
bfd_set_usrdata (abfd, gdata);
bfd_alloc_data (abfd);
/* This is the first we've seen it, so add it to the hash table. */
slot = htab_find_slot (all_bfds, abfd, INSERT);
gdb_assert (slot && !*slot);
*slot = abfd;
}
/* See gdb_bfd.h. */
void
gdb_bfd_unref (struct bfd *abfd)
{
struct gdb_bfd_data *gdata;
struct gdb_bfd_cache_search search;
bfd *archive_bfd;
if (abfd == NULL)
return;
gdata = (struct gdb_bfd_data *) bfd_usrdata (abfd);
gdb_assert (gdata->refc >= 1);
gdata->refc -= 1;
if (gdata->refc > 0)
{
if (debug_bfd_cache)
fprintf_unfiltered (gdb_stdlog,
"Decrease reference count on bfd %s (%s)\n",
host_address_to_string (abfd),
bfd_get_filename (abfd));
return;
}
if (debug_bfd_cache)
fprintf_unfiltered (gdb_stdlog,
"Delete final reference count on bfd %s (%s)\n",
host_address_to_string (abfd),
bfd_get_filename (abfd));
archive_bfd = gdata->archive_bfd;
search.filename = bfd_get_filename (abfd);
if (gdb_bfd_cache && search.filename)
{
hashval_t hash = htab_hash_string (search.filename);
void **slot;
search.mtime = gdata->mtime;
search.size = gdata->size;
search.inode = gdata->inode;
search.device_id = gdata->device_id;
slot = htab_find_slot_with_hash (gdb_bfd_cache, &search, hash,
NO_INSERT);
if (slot && *slot)
htab_clear_slot (gdb_bfd_cache, slot);
}
bfd_free_data (abfd);
delete gdata;
bfd_set_usrdata (abfd, NULL); /* Paranoia. */
htab_remove_elt (all_bfds, abfd);
gdb_bfd_close_or_warn (abfd);
gdb_bfd_unref (archive_bfd);
}
/* A helper function that returns the section data descriptor
associated with SECTION. If no such descriptor exists, a new one
is allocated and cleared. */
static struct gdb_bfd_section_data *
get_section_descriptor (asection *section)
{
struct gdb_bfd_section_data *result;
result = (struct gdb_bfd_section_data *) bfd_section_userdata (section);
if (result == NULL)
{
result = ((struct gdb_bfd_section_data *)
bfd_zalloc (section->owner, sizeof (*result)));
bfd_set_section_userdata (section, result);
}
return result;
}
/* See gdb_bfd.h. */
const gdb_byte *
gdb_bfd_map_section (asection *sectp, bfd_size_type *size)
{
bfd *abfd;
struct gdb_bfd_section_data *descriptor;
bfd_byte *data;
gdb_assert ((sectp->flags & SEC_RELOC) == 0);
gdb_assert (size != NULL);
abfd = sectp->owner;
descriptor = get_section_descriptor (sectp);
/* If the data was already read for this BFD, just reuse it. */
if (descriptor->data != NULL)
goto done;
#ifdef HAVE_MMAP
if (!bfd_is_section_compressed (abfd, sectp))
{
/* The page size, used when mmapping. */
static int pagesize;
if (pagesize == 0)
pagesize = getpagesize ();
/* Only try to mmap sections which are large enough: we don't want
to waste space due to fragmentation. */
if (bfd_section_size (sectp) > 4 * pagesize)
{
descriptor->size = bfd_section_size (sectp);
descriptor->data = bfd_mmap (abfd, 0, descriptor->size, PROT_READ,
MAP_PRIVATE, sectp->filepos,
&descriptor->map_addr,
&descriptor->map_len);
if ((caddr_t)descriptor->data != MAP_FAILED)
{
#if HAVE_POSIX_MADVISE
posix_madvise (descriptor->map_addr, descriptor->map_len,
POSIX_MADV_WILLNEED);
#endif
goto done;
}
/* On failure, clear out the section data and try again. */
memset (descriptor, 0, sizeof (*descriptor));
}
}
#endif /* HAVE_MMAP */
/* Handle compressed sections, or ordinary uncompressed sections in
the no-mmap case. */
descriptor->size = bfd_section_size (sectp);
descriptor->data = NULL;
data = NULL;
if (!bfd_get_full_section_contents (abfd, sectp, &data))
{
warning (_("Can't read data for section '%s' in file '%s'"),
bfd_section_name (sectp),
bfd_get_filename (abfd));
/* Set size to 0 to prevent further attempts to read the invalid
section. */
*size = 0;
return NULL;
}
descriptor->data = data;
done:
gdb_assert (descriptor->data != NULL);
*size = descriptor->size;
return (const gdb_byte *) descriptor->data;
}
/* Return 32-bit CRC for ABFD. If successful store it to *FILE_CRC_RETURN and
return 1. Otherwise print a warning and return 0. ABFD seek position is
not preserved. */
static int
get_file_crc (bfd *abfd, unsigned long *file_crc_return)
{
unsigned long file_crc = 0;
if (bfd_seek (abfd, 0, SEEK_SET) != 0)
{
warning (_("Problem reading \"%s\" for CRC: %s"),
bfd_get_filename (abfd), bfd_errmsg (bfd_get_error ()));
return 0;
}
for (;;)
{
gdb_byte buffer[8 * 1024];
bfd_size_type count;
count = bfd_bread (buffer, sizeof (buffer), abfd);
if (count == (bfd_size_type) -1)
{
warning (_("Problem reading \"%s\" for CRC: %s"),
bfd_get_filename (abfd), bfd_errmsg (bfd_get_error ()));
return 0;
}
if (count == 0)
break;
file_crc = bfd_calc_gnu_debuglink_crc32 (file_crc, buffer, count);
}
*file_crc_return = file_crc;
return 1;
}
/* See gdb_bfd.h. */
int
gdb_bfd_crc (struct bfd *abfd, unsigned long *crc_out)
{
struct gdb_bfd_data *gdata = (struct gdb_bfd_data *) bfd_usrdata (abfd);
if (!gdata->crc_computed)
gdata->crc_computed = get_file_crc (abfd, &gdata->crc);
if (gdata->crc_computed)
*crc_out = gdata->crc;
return gdata->crc_computed;
}
/* See gdb_bfd.h. */
gdb_bfd_ref_ptr
gdb_bfd_fopen (const char *filename, const char *target, const char *mode,
int fd)
{
bfd *result = bfd_fopen (filename, target, mode, fd);
return gdb_bfd_ref_ptr::new_reference (result);
}
/* See gdb_bfd.h. */
gdb_bfd_ref_ptr
gdb_bfd_openr (const char *filename, const char *target)
{
bfd *result = bfd_openr (filename, target);
return gdb_bfd_ref_ptr::new_reference (result);
}
/* See gdb_bfd.h. */
gdb_bfd_ref_ptr
gdb_bfd_openw (const char *filename, const char *target)
{
bfd *result = bfd_openw (filename, target);
return gdb_bfd_ref_ptr::new_reference (result);
}
/* See gdb_bfd.h. */
gdb_bfd_ref_ptr
gdb_bfd_openr_iovec (const char *filename, const char *target,
void *(*open_func) (struct bfd *nbfd,
void *open_closure),
void *open_closure,
file_ptr (*pread_func) (struct bfd *nbfd,
void *stream,
void *buf,
file_ptr nbytes,
file_ptr offset),
int (*close_func) (struct bfd *nbfd,
void *stream),
int (*stat_func) (struct bfd *abfd,
void *stream,
struct stat *sb))
{
bfd *result = bfd_openr_iovec (filename, target,
open_func, open_closure,
pread_func, close_func, stat_func);
return gdb_bfd_ref_ptr::new_reference (result);
}
/* See gdb_bfd.h. */
void
gdb_bfd_mark_parent (bfd *child, bfd *parent)
{
struct gdb_bfd_data *gdata;
gdb_bfd_ref (child);
/* No need to stash the filename here, because we also keep a
reference on the parent archive. */
gdata = (struct gdb_bfd_data *) bfd_usrdata (child);
if (gdata->archive_bfd == NULL)
{
gdata->archive_bfd = parent;
gdb_bfd_ref (parent);
}
else
gdb_assert (gdata->archive_bfd == parent);
}
/* See gdb_bfd.h. */
gdb_bfd_ref_ptr
gdb_bfd_openr_next_archived_file (bfd *archive, bfd *previous)
{
bfd *result = bfd_openr_next_archived_file (archive, previous);
if (result)
gdb_bfd_mark_parent (result, archive);
return gdb_bfd_ref_ptr (result);
}
/* See gdb_bfd.h. */
void
gdb_bfd_record_inclusion (bfd *includer, bfd *includee)
{
struct gdb_bfd_data *gdata;
gdata = (struct gdb_bfd_data *) bfd_usrdata (includer);
gdata->included_bfds.push_back (gdb_bfd_ref_ptr::new_reference (includee));
}
gdb_static_assert (ARRAY_SIZE (_bfd_std_section) == 4);
/* See gdb_bfd.h. */
int
gdb_bfd_section_index (bfd *abfd, asection *section)
{
if (section == NULL)
return -1;
else if (section == bfd_com_section_ptr)
return bfd_count_sections (abfd);
else if (section == bfd_und_section_ptr)
return bfd_count_sections (abfd) + 1;
else if (section == bfd_abs_section_ptr)
return bfd_count_sections (abfd) + 2;
else if (section == bfd_ind_section_ptr)
return bfd_count_sections (abfd) + 3;
return section->index;
}
/* See gdb_bfd.h. */
int
gdb_bfd_count_sections (bfd *abfd)
{
return bfd_count_sections (abfd) + 4;
}
/* See gdb_bfd.h. */
int
gdb_bfd_requires_relocations (bfd *abfd)
{
struct gdb_bfd_data *gdata = (struct gdb_bfd_data *) bfd_usrdata (abfd);
if (gdata->relocation_computed == 0)
{
asection *sect;
for (sect = abfd->sections; sect != NULL; sect = sect->next)
if ((sect->flags & SEC_RELOC) != 0)
{
gdata->needs_relocations = 1;
break;
}
gdata->relocation_computed = 1;
}
return gdata->needs_relocations;
}
/* See gdb_bfd.h. */
bool
gdb_bfd_get_full_section_contents (bfd *abfd, asection *section,
gdb::byte_vector *contents)
{
bfd_size_type section_size = bfd_section_size (section);
contents->resize (section_size);
return bfd_get_section_contents (abfd, section, contents->data (), 0,
section_size);
}
/* A callback for htab_traverse that prints a single BFD. */
static int
print_one_bfd (void **slot, void *data)
{
bfd *abfd = (struct bfd *) *slot;
struct gdb_bfd_data *gdata = (struct gdb_bfd_data *) bfd_usrdata (abfd);
struct ui_out *uiout = (struct ui_out *) data;
ui_out_emit_tuple tuple_emitter (uiout, NULL);
uiout->field_signed ("refcount", gdata->refc);
uiout->field_string ("addr", host_address_to_string (abfd));
uiout->field_string ("filename", bfd_get_filename (abfd));
uiout->text ("\n");
return 1;
}
/* Implement the 'maint info bfd' command. */
static void
maintenance_info_bfds (const char *arg, int from_tty)
{
struct ui_out *uiout = current_uiout;
ui_out_emit_table table_emitter (uiout, 3, -1, "bfds");
uiout->table_header (10, ui_left, "refcount", "Refcount");
uiout->table_header (18, ui_left, "addr", "Address");
uiout->table_header (40, ui_left, "filename", "Filename");
uiout->table_body ();
htab_traverse (all_bfds, print_one_bfd, uiout);
}
void _initialize_gdb_bfd ();
void
_initialize_gdb_bfd ()
{
all_bfds = htab_create_alloc (10, htab_hash_pointer, htab_eq_pointer,
NULL, xcalloc, xfree);
add_cmd ("bfds", class_maintenance, maintenance_info_bfds, _("\
List the BFDs that are currently open."),
&maintenanceinfolist);
add_setshow_boolean_cmd ("bfd-sharing", no_class,
&bfd_sharing, _("\
Set whether gdb will share bfds that appear to be the same file."), _("\
Show whether gdb will share bfds that appear to be the same file."), _("\
When enabled gdb will reuse existing bfds rather than reopening the\n\
same file. To decide if two files are the same then gdb compares the\n\
filename, file size, file modification time, and file inode."),
NULL,
&show_bfd_sharing,
&maintenance_set_cmdlist,
&maintenance_show_cmdlist);
add_setshow_zuinteger_cmd ("bfd-cache", class_maintenance,
&debug_bfd_cache, _("\
Set bfd cache debugging."), _("\
Show bfd cache debugging."), _("\
When non-zero, bfd cache specific debugging is enabled."),
NULL,
&show_bfd_cache_debug,
&setdebuglist, &showdebuglist);
}