mirror of
https://sourceware.org/git/binutils-gdb.git
synced 2024-12-21 04:42:53 +08:00
cea6e4f13a
gdb/ChangeLog: * fbsd-nat.c: Include "gdb_wait.h" instead of <sys/wait.h>.
514 lines
14 KiB
C
514 lines
14 KiB
C
/* Native-dependent code for FreeBSD.
|
|
|
|
Copyright (C) 2002-2015 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 <http://www.gnu.org/licenses/>. */
|
|
|
|
#include "defs.h"
|
|
#include "gdbcore.h"
|
|
#include "inferior.h"
|
|
#include "regcache.h"
|
|
#include "regset.h"
|
|
#include "gdbthread.h"
|
|
#include "gdb_wait.h"
|
|
#include <sys/types.h>
|
|
#include <sys/procfs.h>
|
|
#include <sys/ptrace.h>
|
|
#include <sys/sysctl.h>
|
|
#ifdef HAVE_KINFO_GETVMMAP
|
|
#include <sys/user.h>
|
|
#include <libutil.h>
|
|
#endif
|
|
|
|
#include "elf-bfd.h"
|
|
#include "fbsd-nat.h"
|
|
|
|
/* Return the name of a file that can be opened to get the symbols for
|
|
the child process identified by PID. */
|
|
|
|
static char *
|
|
fbsd_pid_to_exec_file (struct target_ops *self, int pid)
|
|
{
|
|
ssize_t len = PATH_MAX;
|
|
static char buf[PATH_MAX];
|
|
char name[PATH_MAX];
|
|
|
|
#ifdef KERN_PROC_PATHNAME
|
|
int mib[4];
|
|
|
|
mib[0] = CTL_KERN;
|
|
mib[1] = KERN_PROC;
|
|
mib[2] = KERN_PROC_PATHNAME;
|
|
mib[3] = pid;
|
|
if (sysctl (mib, 4, buf, &len, NULL, 0) == 0)
|
|
return buf;
|
|
#endif
|
|
|
|
xsnprintf (name, PATH_MAX, "/proc/%d/exe", pid);
|
|
len = readlink (name, buf, PATH_MAX - 1);
|
|
if (len != -1)
|
|
{
|
|
buf[len] = '\0';
|
|
return buf;
|
|
}
|
|
|
|
return NULL;
|
|
}
|
|
|
|
#ifdef HAVE_KINFO_GETVMMAP
|
|
/* Iterate over all the memory regions in the current inferior,
|
|
calling FUNC for each memory region. OBFD is passed as the last
|
|
argument to FUNC. */
|
|
|
|
static int
|
|
fbsd_find_memory_regions (struct target_ops *self,
|
|
find_memory_region_ftype func, void *obfd)
|
|
{
|
|
pid_t pid = ptid_get_pid (inferior_ptid);
|
|
struct kinfo_vmentry *vmentl, *kve;
|
|
uint64_t size;
|
|
struct cleanup *cleanup;
|
|
int i, nitems;
|
|
|
|
vmentl = kinfo_getvmmap (pid, &nitems);
|
|
if (vmentl == NULL)
|
|
perror_with_name (_("Couldn't fetch VM map entries."));
|
|
cleanup = make_cleanup (free, vmentl);
|
|
|
|
for (i = 0; i < nitems; i++)
|
|
{
|
|
kve = &vmentl[i];
|
|
|
|
/* Skip unreadable segments and those where MAP_NOCORE has been set. */
|
|
if (!(kve->kve_protection & KVME_PROT_READ)
|
|
|| kve->kve_flags & KVME_FLAG_NOCOREDUMP)
|
|
continue;
|
|
|
|
/* Skip segments with an invalid type. */
|
|
if (kve->kve_type != KVME_TYPE_DEFAULT
|
|
&& kve->kve_type != KVME_TYPE_VNODE
|
|
&& kve->kve_type != KVME_TYPE_SWAP
|
|
&& kve->kve_type != KVME_TYPE_PHYS)
|
|
continue;
|
|
|
|
size = kve->kve_end - kve->kve_start;
|
|
if (info_verbose)
|
|
{
|
|
fprintf_filtered (gdb_stdout,
|
|
"Save segment, %ld bytes at %s (%c%c%c)\n",
|
|
(long) size,
|
|
paddress (target_gdbarch (), kve->kve_start),
|
|
kve->kve_protection & KVME_PROT_READ ? 'r' : '-',
|
|
kve->kve_protection & KVME_PROT_WRITE ? 'w' : '-',
|
|
kve->kve_protection & KVME_PROT_EXEC ? 'x' : '-');
|
|
}
|
|
|
|
/* Invoke the callback function to create the corefile segment.
|
|
Pass MODIFIED as true, we do not know the real modification state. */
|
|
func (kve->kve_start, size, kve->kve_protection & KVME_PROT_READ,
|
|
kve->kve_protection & KVME_PROT_WRITE,
|
|
kve->kve_protection & KVME_PROT_EXEC, 1, obfd);
|
|
}
|
|
do_cleanups (cleanup);
|
|
return 0;
|
|
}
|
|
#else
|
|
static int
|
|
fbsd_read_mapping (FILE *mapfile, unsigned long *start, unsigned long *end,
|
|
char *protection)
|
|
{
|
|
/* FreeBSD 5.1-RELEASE uses a 256-byte buffer. */
|
|
char buf[256];
|
|
int resident, privateresident;
|
|
unsigned long obj;
|
|
int ret = EOF;
|
|
|
|
/* As of FreeBSD 5.0-RELEASE, the layout is described in
|
|
/usr/src/sys/fs/procfs/procfs_map.c. Somewhere in 5.1-CURRENT a
|
|
new column was added to the procfs map. Therefore we can't use
|
|
fscanf since we need to support older releases too. */
|
|
if (fgets (buf, sizeof buf, mapfile) != NULL)
|
|
ret = sscanf (buf, "%lx %lx %d %d %lx %s", start, end,
|
|
&resident, &privateresident, &obj, protection);
|
|
|
|
return (ret != 0 && ret != EOF);
|
|
}
|
|
|
|
/* Iterate over all the memory regions in the current inferior,
|
|
calling FUNC for each memory region. OBFD is passed as the last
|
|
argument to FUNC. */
|
|
|
|
static int
|
|
fbsd_find_memory_regions (struct target_ops *self,
|
|
find_memory_region_ftype func, void *obfd)
|
|
{
|
|
pid_t pid = ptid_get_pid (inferior_ptid);
|
|
char *mapfilename;
|
|
FILE *mapfile;
|
|
unsigned long start, end, size;
|
|
char protection[4];
|
|
int read, write, exec;
|
|
struct cleanup *cleanup;
|
|
|
|
mapfilename = xstrprintf ("/proc/%ld/map", (long) pid);
|
|
cleanup = make_cleanup (xfree, mapfilename);
|
|
mapfile = fopen (mapfilename, "r");
|
|
if (mapfile == NULL)
|
|
error (_("Couldn't open %s."), mapfilename);
|
|
make_cleanup_fclose (mapfile);
|
|
|
|
if (info_verbose)
|
|
fprintf_filtered (gdb_stdout,
|
|
"Reading memory regions from %s\n", mapfilename);
|
|
|
|
/* Now iterate until end-of-file. */
|
|
while (fbsd_read_mapping (mapfile, &start, &end, &protection[0]))
|
|
{
|
|
size = end - start;
|
|
|
|
read = (strchr (protection, 'r') != 0);
|
|
write = (strchr (protection, 'w') != 0);
|
|
exec = (strchr (protection, 'x') != 0);
|
|
|
|
if (info_verbose)
|
|
{
|
|
fprintf_filtered (gdb_stdout,
|
|
"Save segment, %ld bytes at %s (%c%c%c)\n",
|
|
size, paddress (target_gdbarch (), start),
|
|
read ? 'r' : '-',
|
|
write ? 'w' : '-',
|
|
exec ? 'x' : '-');
|
|
}
|
|
|
|
/* Invoke the callback function to create the corefile segment.
|
|
Pass MODIFIED as true, we do not know the real modification state. */
|
|
func (start, size, read, write, exec, 1, obfd);
|
|
}
|
|
|
|
do_cleanups (cleanup);
|
|
return 0;
|
|
}
|
|
#endif
|
|
|
|
#ifdef PT_LWPINFO
|
|
static ptid_t (*super_wait) (struct target_ops *,
|
|
ptid_t,
|
|
struct target_waitstatus *,
|
|
int);
|
|
|
|
#ifdef TDP_RFPPWAIT
|
|
/*
|
|
To catch fork events, PT_FOLLOW_FORK is set on every traced process
|
|
to enable stops on returns from fork or vfork. Note that both the
|
|
parent and child will always stop, even if system call stops are not
|
|
enabled.
|
|
|
|
After a fork, both the child and parent process will stop and report
|
|
an event. However, there is no guarantee of order. If the parent
|
|
reports its stop first, then fbsd_wait explicitly waits for the new
|
|
child before returning. If the child reports its stop first, then
|
|
the event is saved on a list and ignored until the parent's stop is
|
|
reported. fbsd_wait could have been changed to fetch the parent PID
|
|
of the new child and used that to wait for the parent explicitly.
|
|
However, if two threads in the parent fork at the same time, then
|
|
the wait on the parent might return the "wrong" fork event.
|
|
|
|
The initial version of PT_FOLLOW_FORK did not set PL_FLAG_CHILD for
|
|
the new child process. This flag could be inferred by treating any
|
|
events for an unknown pid as a new child.
|
|
|
|
In addition, the initial version of PT_FOLLOW_FORK did not report a
|
|
stop event for the parent process of a vfork until after the child
|
|
process executed a new program or exited. The kernel was changed to
|
|
defer the wait for exit or exec of the child until after posting the
|
|
stop event shortly after the change to introduce PL_FLAG_CHILD.
|
|
This could be worked around by reporting a vfork event when the
|
|
child event posted and ignoring the subsequent event from the
|
|
parent.
|
|
|
|
This implementation requires both of these fixes for simplicity's
|
|
sake. FreeBSD versions newer than 9.1 contain both fixes.
|
|
*/
|
|
|
|
struct fbsd_fork_child_info
|
|
{
|
|
struct fbsd_fork_child_info *next;
|
|
pid_t child; /* Pid of new child. */
|
|
};
|
|
|
|
static struct fbsd_fork_child_info *fbsd_pending_children;
|
|
|
|
/* Record a new child process event that is reported before the
|
|
corresponding fork event in the parent. */
|
|
|
|
static void
|
|
fbsd_remember_child (pid_t pid)
|
|
{
|
|
struct fbsd_fork_child_info *info;
|
|
|
|
info = xcalloc (1, sizeof *info);
|
|
|
|
info->child = pid;
|
|
info->next = fbsd_pending_children;
|
|
fbsd_pending_children = info;
|
|
}
|
|
|
|
/* Check for a previously-recorded new child process event for PID.
|
|
If one is found, remove it from the list. */
|
|
|
|
static int
|
|
fbsd_is_child_pending (pid_t pid)
|
|
{
|
|
struct fbsd_fork_child_info *info, *prev;
|
|
|
|
prev = NULL;
|
|
for (info = fbsd_pending_children; info; prev = info, info = info->next)
|
|
{
|
|
if (info->child == pid)
|
|
{
|
|
if (prev == NULL)
|
|
fbsd_pending_children = info->next;
|
|
else
|
|
prev->next = info->next;
|
|
xfree (info);
|
|
return 1;
|
|
}
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
/* Fetch the external variant of the kernel's internal process
|
|
structure for the process PID into KP. */
|
|
|
|
static void
|
|
fbsd_fetch_kinfo_proc (pid_t pid, struct kinfo_proc *kp)
|
|
{
|
|
size_t len;
|
|
int mib[4];
|
|
|
|
len = sizeof *kp;
|
|
mib[0] = CTL_KERN;
|
|
mib[1] = KERN_PROC;
|
|
mib[2] = KERN_PROC_PID;
|
|
mib[3] = pid;
|
|
if (sysctl (mib, 4, kp, &len, NULL, 0) == -1)
|
|
perror_with_name (("sysctl"));
|
|
}
|
|
#endif
|
|
|
|
/* Wait for the child specified by PTID to do something. Return the
|
|
process ID of the child, or MINUS_ONE_PTID in case of error; store
|
|
the status in *OURSTATUS. */
|
|
|
|
static ptid_t
|
|
fbsd_wait (struct target_ops *ops,
|
|
ptid_t ptid, struct target_waitstatus *ourstatus,
|
|
int target_options)
|
|
{
|
|
ptid_t wptid;
|
|
|
|
while (1)
|
|
{
|
|
wptid = super_wait (ops, ptid, ourstatus, target_options);
|
|
if (ourstatus->kind == TARGET_WAITKIND_STOPPED)
|
|
{
|
|
struct ptrace_lwpinfo pl;
|
|
pid_t pid;
|
|
int status;
|
|
|
|
pid = ptid_get_pid (wptid);
|
|
if (ptrace (PT_LWPINFO, pid, (caddr_t)&pl, sizeof pl) == -1)
|
|
perror_with_name (("ptrace"));
|
|
|
|
#ifdef TDP_RFPPWAIT
|
|
if (pl.pl_flags & PL_FLAG_FORKED)
|
|
{
|
|
struct kinfo_proc kp;
|
|
pid_t child;
|
|
|
|
child = pl.pl_child_pid;
|
|
ourstatus->kind = TARGET_WAITKIND_FORKED;
|
|
ourstatus->value.related_pid = pid_to_ptid (child);
|
|
|
|
/* Make sure the other end of the fork is stopped too. */
|
|
if (!fbsd_is_child_pending (child))
|
|
{
|
|
pid = waitpid (child, &status, 0);
|
|
if (pid == -1)
|
|
perror_with_name (("waitpid"));
|
|
|
|
gdb_assert (pid == child);
|
|
|
|
if (ptrace (PT_LWPINFO, child, (caddr_t)&pl, sizeof pl) == -1)
|
|
perror_with_name (("ptrace"));
|
|
|
|
gdb_assert (pl.pl_flags & PL_FLAG_CHILD);
|
|
}
|
|
|
|
/* For vfork, the child process will have the P_PPWAIT
|
|
flag set. */
|
|
fbsd_fetch_kinfo_proc (child, &kp);
|
|
if (kp.ki_flag & P_PPWAIT)
|
|
ourstatus->kind = TARGET_WAITKIND_VFORKED;
|
|
|
|
return wptid;
|
|
}
|
|
|
|
if (pl.pl_flags & PL_FLAG_CHILD)
|
|
{
|
|
/* Remember that this child forked, but do not report it
|
|
until the parent reports its corresponding fork
|
|
event. */
|
|
fbsd_remember_child (ptid_get_pid (wptid));
|
|
continue;
|
|
}
|
|
#endif
|
|
|
|
#ifdef PL_FLAG_EXEC
|
|
if (pl.pl_flags & PL_FLAG_EXEC)
|
|
{
|
|
ourstatus->kind = TARGET_WAITKIND_EXECD;
|
|
ourstatus->value.execd_pathname
|
|
= xstrdup (fbsd_pid_to_exec_file (NULL, pid));
|
|
return wptid;
|
|
}
|
|
#endif
|
|
}
|
|
return wptid;
|
|
}
|
|
}
|
|
|
|
#ifdef TDP_RFPPWAIT
|
|
/* Target hook for follow_fork. On entry and at return inferior_ptid is
|
|
the ptid of the followed inferior. */
|
|
|
|
static int
|
|
fbsd_follow_fork (struct target_ops *ops, int follow_child,
|
|
int detach_fork)
|
|
{
|
|
if (!follow_child)
|
|
{
|
|
struct thread_info *tp = inferior_thread ();
|
|
pid_t child_pid = ptid_get_pid (tp->pending_follow.value.related_pid);
|
|
|
|
/* Breakpoints have already been detached from the child by
|
|
infrun.c. */
|
|
|
|
if (ptrace (PT_DETACH, child_pid, (PTRACE_TYPE_ARG3)1, 0) == -1)
|
|
perror_with_name (("ptrace"));
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
static int
|
|
fbsd_insert_fork_catchpoint (struct target_ops *self, int pid)
|
|
{
|
|
return 0;
|
|
}
|
|
|
|
static int
|
|
fbsd_remove_fork_catchpoint (struct target_ops *self, int pid)
|
|
{
|
|
return 0;
|
|
}
|
|
|
|
static int
|
|
fbsd_insert_vfork_catchpoint (struct target_ops *self, int pid)
|
|
{
|
|
return 0;
|
|
}
|
|
|
|
static int
|
|
fbsd_remove_vfork_catchpoint (struct target_ops *self, int pid)
|
|
{
|
|
return 0;
|
|
}
|
|
|
|
/* Enable fork tracing for a specific process.
|
|
|
|
To catch fork events, PT_FOLLOW_FORK is set on every traced process
|
|
to enable stops on returns from fork or vfork. Note that both the
|
|
parent and child will always stop, even if system call stops are
|
|
not enabled. */
|
|
|
|
static void
|
|
fbsd_enable_follow_fork (pid_t pid)
|
|
{
|
|
if (ptrace (PT_FOLLOW_FORK, pid, (PTRACE_TYPE_ARG3)0, 1) == -1)
|
|
perror_with_name (("ptrace"));
|
|
}
|
|
|
|
/* Implement the "to_post_startup_inferior" target_ops method. */
|
|
|
|
static void
|
|
fbsd_post_startup_inferior (struct target_ops *self, ptid_t pid)
|
|
{
|
|
fbsd_enable_follow_fork (ptid_get_pid (pid));
|
|
}
|
|
|
|
/* Implement the "to_post_attach" target_ops method. */
|
|
|
|
static void
|
|
fbsd_post_attach (struct target_ops *self, int pid)
|
|
{
|
|
fbsd_enable_follow_fork (pid);
|
|
}
|
|
#endif
|
|
|
|
#ifdef PL_FLAG_EXEC
|
|
/* If the FreeBSD kernel supports PL_FLAG_EXEC, then traced processes
|
|
will always stop after exec. */
|
|
|
|
static int
|
|
fbsd_insert_exec_catchpoint (struct target_ops *self, int pid)
|
|
{
|
|
return 0;
|
|
}
|
|
|
|
static int
|
|
fbsd_remove_exec_catchpoint (struct target_ops *self, int pid)
|
|
{
|
|
return 0;
|
|
}
|
|
#endif
|
|
#endif
|
|
|
|
void
|
|
fbsd_nat_add_target (struct target_ops *t)
|
|
{
|
|
t->to_pid_to_exec_file = fbsd_pid_to_exec_file;
|
|
t->to_find_memory_regions = fbsd_find_memory_regions;
|
|
#ifdef PT_LWPINFO
|
|
super_wait = t->to_wait;
|
|
t->to_wait = fbsd_wait;
|
|
#ifdef TDP_RFPPWAIT
|
|
t->to_follow_fork = fbsd_follow_fork;
|
|
t->to_insert_fork_catchpoint = fbsd_insert_fork_catchpoint;
|
|
t->to_remove_fork_catchpoint = fbsd_remove_fork_catchpoint;
|
|
t->to_insert_vfork_catchpoint = fbsd_insert_vfork_catchpoint;
|
|
t->to_remove_vfork_catchpoint = fbsd_remove_vfork_catchpoint;
|
|
t->to_post_startup_inferior = fbsd_post_startup_inferior;
|
|
t->to_post_attach = fbsd_post_attach;
|
|
#endif
|
|
#ifdef PL_FLAG_EXEC
|
|
t->to_insert_exec_catchpoint = fbsd_insert_exec_catchpoint;
|
|
t->to_remove_exec_catchpoint = fbsd_remove_exec_catchpoint;
|
|
#endif
|
|
#endif
|
|
add_target (t);
|
|
}
|