binutils-gdb/gdbserver/netbsd-low.cc

1245 lines
31 KiB
C++
Raw Permalink Normal View History

/* Copyright (C) 2020-2024 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 "target.h"
#include "netbsd-low.h"
#include "nat/netbsd-nat.h"
#include <sys/param.h>
#include <sys/types.h>
#include <sys/ptrace.h>
#include <sys/sysctl.h>
#include <limits.h>
#include <unistd.h>
#include <signal.h>
#include <elf.h>
#include <type_traits>
#include "gdbsupport/eintr.h"
#include "gdbsupport/gdb_wait.h"
#include "gdbsupport/filestuff.h"
#include "gdbsupport/common-inferior.h"
#include "nat/fork-inferior.h"
#include "hostio.h"
int using_threads = 1;
/* Callback used by fork_inferior to start tracing the inferior. */
static void
netbsd_ptrace_fun ()
{
/* Switch child to its own process group so that signals won't
directly affect GDBserver. */
if (setpgid (0, 0) < 0)
trace_start_error_with_name (("setpgid"));
if (ptrace (PT_TRACE_ME, 0, nullptr, 0) < 0)
trace_start_error_with_name (("ptrace"));
/* If GDBserver is connected to gdb via stdio, redirect the inferior's
stdout to stderr so that inferior i/o doesn't corrupt the connection.
Also, redirect stdin to /dev/null. */
if (remote_connection_is_stdio ())
{
if (close (0) < 0)
trace_start_error_with_name (("close"));
if (open ("/dev/null", O_RDONLY) < 0)
trace_start_error_with_name (("open"));
if (dup2 (2, 1) < 0)
trace_start_error_with_name (("dup2"));
if (write (2, "stdin/stdout redirected\n",
sizeof ("stdin/stdout redirected\n") - 1) < 0)
{
/* Errors ignored. */
}
}
}
/* Implement the create_inferior method of the target_ops vector. */
int
netbsd_process_target::create_inferior (const char *program,
const std::vector<char *> &program_args)
{
std::string str_program_args = construct_inferior_arguments (program_args);
pid_t pid = fork_inferior (program, str_program_args.c_str (),
get_environ ()->envp (), netbsd_ptrace_fun,
nullptr, nullptr, nullptr, nullptr);
add_process (pid, 0);
post_fork_inferior (pid, program);
return pid;
}
/* Implement the post_create_inferior target_ops method. */
void
netbsd_process_target::post_create_inferior ()
{
pid_t pid = current_process ()->pid;
netbsd_nat::enable_proc_events (pid);
low_arch_setup ();
}
/* Implement the attach target_ops method. */
int
netbsd_process_target::attach (unsigned long pid)
{
/* Unimplemented. */
return -1;
}
/* Returns true if GDB is interested in any child syscalls. */
static bool
gdb_catching_syscalls_p (pid_t pid)
{
struct process_info *proc = find_process_pid (pid);
return !proc->syscalls_to_catch.empty ();
}
/* Implement the resume target_ops method. */
void
netbsd_process_target::resume (struct thread_resume *resume_info, size_t n)
{
ptid_t resume_ptid = resume_info[0].thread;
const int signal = resume_info[0].sig;
const bool step = resume_info[0].kind == resume_step;
if (resume_ptid == minus_one_ptid)
resume_ptid = ptid_of (current_thread);
const pid_t pid = resume_ptid.pid ();
const lwpid_t lwp = resume_ptid.lwp ();
regcache_invalidate_pid (pid);
auto fn
= [&] (ptid_t ptid)
{
if (step)
{
if (ptid.lwp () == lwp || n != 1)
{
if (ptrace (PT_SETSTEP, pid, NULL, ptid.lwp ()) == -1)
perror_with_name (("ptrace"));
if (ptrace (PT_RESUME, pid, NULL, ptid.lwp ()) == -1)
perror_with_name (("ptrace"));
}
else
{
if (ptrace (PT_CLEARSTEP, pid, NULL, ptid.lwp ()) == -1)
perror_with_name (("ptrace"));
if (ptrace (PT_SUSPEND, pid, NULL, ptid.lwp ()) == -1)
perror_with_name (("ptrace"));
}
}
else
{
if (ptrace (PT_CLEARSTEP, pid, NULL, ptid.lwp ()) == -1)
perror_with_name (("ptrace"));
if (ptrace (PT_RESUME, pid, NULL, ptid.lwp ()) == -1)
perror_with_name (("ptrace"));
}
};
netbsd_nat::for_each_thread (pid, fn);
int request = gdb_catching_syscalls_p (pid) ? PT_CONTINUE : PT_SYSCALL;
errno = 0;
ptrace (request, pid, (void *)1, signal);
if (errno)
perror_with_name (("ptrace"));
}
/* Returns true if GDB is interested in the reported SYSNO syscall. */
static bool
netbsd_catch_this_syscall (int sysno)
{
struct process_info *proc = current_process ();
if (proc->syscalls_to_catch.empty ())
return false;
if (proc->syscalls_to_catch[0] == ANY_SYSCALL)
return true;
for (int iter : proc->syscalls_to_catch)
if (iter == sysno)
return true;
return false;
}
/* Helper function for child_wait and the derivatives of child_wait.
HOSTSTATUS is the waitstatus from wait() or the equivalent; store our
translation of that in OURSTATUS. */
static void
netbsd_store_waitstatus (struct target_waitstatus *ourstatus, int hoststatus)
{
if (WIFEXITED (hoststatus))
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_exited (WEXITSTATUS (hoststatus));
else if (!WIFSTOPPED (hoststatus))
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_signalled (gdb_signal_from_host (WTERMSIG (hoststatus)));
else
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_stopped (gdb_signal_from_host (WSTOPSIG (hoststatus)));
}
/* Implement a safe wrapper around waitpid(). */
static pid_t
Make target_wait options use enum flags This changes TARGET_WNOHANG to be a member of an enum, rather than a define, and also adds a DEF_ENUM_FLAGS_TYPE for this type. Then, it changes target_wait and the various target wait methods to use this type rather than "int". This didn't catch any bugs, but it seems like a decent cleanup nevertheless. I did not change deprecated_target_wait_hook, since that's only used out-of-tree (by Insight), and there didn't seem to be a need. I can't build some of these targets, so I modified them on a best-effort basis. I don't think this patch should go in before the release branch is made. gdb/ChangeLog 2020-09-18 Tom Tromey <tromey@adacore.com> * windows-nat.c (struct windows_nat_target) <wait>: Update. (windows_nat_target::wait): Update. * target/wait.h (enum target_wait_flag): New. Use DEF_ENUM_FLAGS_TYPE. * target/target.h (target_wait): Change type of options. * target.h (target_options_to_string, default_target_wait): Update. (struct target_ops) <wait>: Change type of options. * target.c (target_wait, default_target_wait, do_option): Change type of "options". (target_options_to_string): Likewise. * target-delegates.c: Rebuild. * target-debug.h (target_debug_print_target_wait_flags): Rename from target_debug_print_options. * sol-thread.c (class sol_thread_target) <wait>: Update. (sol_thread_target::wait): Update. * rs6000-nat.c (class rs6000_nat_target) <wait>: Update. (rs6000_nat_target::wait): Update. * remote.c (class remote_target) <wait, wait_ns, wait_as>: Update. (remote_target::wait_ns, remote_target::wait_as): Change type of "options". (remote_target::wait): Update. * remote-sim.c (struct gdbsim_target) <wait>: Update. (gdbsim_target::wait): Update. * record-full.c (class record_full_base_target) <wait>: Update. (record_full_wait_1): Change type of "options". (record_full_base_target::wait): Update. * record-btrace.c (class record_btrace_target) <wait>: Update. (record_btrace_target::wait): Update. * ravenscar-thread.c (struct ravenscar_thread_target) <wait>: Update. (ravenscar_thread_target::wait): Update. * procfs.c (class procfs_target) <wait>: Update. (procfs_target::wait): Update. * obsd-nat.h (class obsd_nat_target) <wait>: Update. * obsd-nat.c (obsd_nat_target::wait): Update. * nto-procfs.c (struct nto_procfs_target) <wait>: Update. (nto_procfs_target::wait): Update. * nbsd-nat.h (struct nbsd_nat_target) <wait>: Update. * nbsd-nat.c (nbsd_wait): Change type of "options". (nbsd_nat_target::wait): Update. * linux-thread-db.c (class thread_db_target) <wait>: Update. (thread_db_target::wait): Update. * linux-nat.h (class linux_nat_target) <wait>: Update. * linux-nat.c (linux_nat_target::wait): Update. (linux_nat_wait_1): Update. * infrun.c (do_target_wait_1, do_target_wait): Change type of "options". * inf-ptrace.h (struct inf_ptrace_target) <wait>: Update. * inf-ptrace.c (inf_ptrace_target::wait): Update. * go32-nat.c (struct go32_nat_target) <wait>: Update. (go32_nat_target::wait): Update. * gnu-nat.h (struct gnu_nat_target) <wait>: Update. * gnu-nat.c (gnu_nat_target::wait): Update. * fbsd-nat.h (class fbsd_nat_target) <wait>: Update. * fbsd-nat.c (fbsd_nat_target::wait): Update. * darwin-nat.h (class darwin_nat_target) <wait>: Update. * darwin-nat.c (darwin_nat_target::wait): Update. * bsd-uthread.c (struct bsd_uthread_target) <wait>: Update. (bsd_uthread_target::wait): Update. * aix-thread.c (class aix_thread_target) <wait>: Update. (aix_thread_target::wait): Update. gdbserver/ChangeLog 2020-09-18 Tom Tromey <tromey@adacore.com> * netbsd-low.h (class netbsd_process_target) <wait>: Update. * netbsd-low.cc (netbsd_waitpid, netbsd_wait) (netbsd_process_target::wait): Change type of target_options. * win32-low.h (class win32_process_target) <wait>: Update. * win32-low.cc (win32_process_target::wait): Update. * target.h (class process_stratum_target) <wait>: Update. (mywait): Update. * target.cc (mywait, target_wait): Change type of "options". * linux-low.h (class linux_process_target) <wait, wait_1>: Update. * linux-low.cc (linux_process_target::wait) (linux_process_target::wait_1): Update.
2020-09-19 04:20:44 +08:00
netbsd_waitpid (ptid_t ptid, struct target_waitstatus *ourstatus,
target_wait_flags target_options)
{
int status;
Make target_wait options use enum flags This changes TARGET_WNOHANG to be a member of an enum, rather than a define, and also adds a DEF_ENUM_FLAGS_TYPE for this type. Then, it changes target_wait and the various target wait methods to use this type rather than "int". This didn't catch any bugs, but it seems like a decent cleanup nevertheless. I did not change deprecated_target_wait_hook, since that's only used out-of-tree (by Insight), and there didn't seem to be a need. I can't build some of these targets, so I modified them on a best-effort basis. I don't think this patch should go in before the release branch is made. gdb/ChangeLog 2020-09-18 Tom Tromey <tromey@adacore.com> * windows-nat.c (struct windows_nat_target) <wait>: Update. (windows_nat_target::wait): Update. * target/wait.h (enum target_wait_flag): New. Use DEF_ENUM_FLAGS_TYPE. * target/target.h (target_wait): Change type of options. * target.h (target_options_to_string, default_target_wait): Update. (struct target_ops) <wait>: Change type of options. * target.c (target_wait, default_target_wait, do_option): Change type of "options". (target_options_to_string): Likewise. * target-delegates.c: Rebuild. * target-debug.h (target_debug_print_target_wait_flags): Rename from target_debug_print_options. * sol-thread.c (class sol_thread_target) <wait>: Update. (sol_thread_target::wait): Update. * rs6000-nat.c (class rs6000_nat_target) <wait>: Update. (rs6000_nat_target::wait): Update. * remote.c (class remote_target) <wait, wait_ns, wait_as>: Update. (remote_target::wait_ns, remote_target::wait_as): Change type of "options". (remote_target::wait): Update. * remote-sim.c (struct gdbsim_target) <wait>: Update. (gdbsim_target::wait): Update. * record-full.c (class record_full_base_target) <wait>: Update. (record_full_wait_1): Change type of "options". (record_full_base_target::wait): Update. * record-btrace.c (class record_btrace_target) <wait>: Update. (record_btrace_target::wait): Update. * ravenscar-thread.c (struct ravenscar_thread_target) <wait>: Update. (ravenscar_thread_target::wait): Update. * procfs.c (class procfs_target) <wait>: Update. (procfs_target::wait): Update. * obsd-nat.h (class obsd_nat_target) <wait>: Update. * obsd-nat.c (obsd_nat_target::wait): Update. * nto-procfs.c (struct nto_procfs_target) <wait>: Update. (nto_procfs_target::wait): Update. * nbsd-nat.h (struct nbsd_nat_target) <wait>: Update. * nbsd-nat.c (nbsd_wait): Change type of "options". (nbsd_nat_target::wait): Update. * linux-thread-db.c (class thread_db_target) <wait>: Update. (thread_db_target::wait): Update. * linux-nat.h (class linux_nat_target) <wait>: Update. * linux-nat.c (linux_nat_target::wait): Update. (linux_nat_wait_1): Update. * infrun.c (do_target_wait_1, do_target_wait): Change type of "options". * inf-ptrace.h (struct inf_ptrace_target) <wait>: Update. * inf-ptrace.c (inf_ptrace_target::wait): Update. * go32-nat.c (struct go32_nat_target) <wait>: Update. (go32_nat_target::wait): Update. * gnu-nat.h (struct gnu_nat_target) <wait>: Update. * gnu-nat.c (gnu_nat_target::wait): Update. * fbsd-nat.h (class fbsd_nat_target) <wait>: Update. * fbsd-nat.c (fbsd_nat_target::wait): Update. * darwin-nat.h (class darwin_nat_target) <wait>: Update. * darwin-nat.c (darwin_nat_target::wait): Update. * bsd-uthread.c (struct bsd_uthread_target) <wait>: Update. (bsd_uthread_target::wait): Update. * aix-thread.c (class aix_thread_target) <wait>: Update. (aix_thread_target::wait): Update. gdbserver/ChangeLog 2020-09-18 Tom Tromey <tromey@adacore.com> * netbsd-low.h (class netbsd_process_target) <wait>: Update. * netbsd-low.cc (netbsd_waitpid, netbsd_wait) (netbsd_process_target::wait): Change type of target_options. * win32-low.h (class win32_process_target) <wait>: Update. * win32-low.cc (win32_process_target::wait): Update. * target.h (class process_stratum_target) <wait>: Update. (mywait): Update. * target.cc (mywait, target_wait): Change type of "options". * linux-low.h (class linux_process_target) <wait, wait_1>: Update. * linux-low.cc (linux_process_target::wait) (linux_process_target::wait_1): Update.
2020-09-19 04:20:44 +08:00
int options = (target_options & TARGET_WNOHANG) ? WNOHANG : 0;
pid_t pid
= gdb::handle_eintr (-1, ::waitpid, ptid.pid (), &status, options);
if (pid == -1)
perror_with_name (_("Child process unexpectedly missing"));
netbsd_store_waitstatus (ourstatus, status);
return pid;
}
/* Implement the wait target_ops method.
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
netbsd_wait (ptid_t ptid, struct target_waitstatus *ourstatus,
Make target_wait options use enum flags This changes TARGET_WNOHANG to be a member of an enum, rather than a define, and also adds a DEF_ENUM_FLAGS_TYPE for this type. Then, it changes target_wait and the various target wait methods to use this type rather than "int". This didn't catch any bugs, but it seems like a decent cleanup nevertheless. I did not change deprecated_target_wait_hook, since that's only used out-of-tree (by Insight), and there didn't seem to be a need. I can't build some of these targets, so I modified them on a best-effort basis. I don't think this patch should go in before the release branch is made. gdb/ChangeLog 2020-09-18 Tom Tromey <tromey@adacore.com> * windows-nat.c (struct windows_nat_target) <wait>: Update. (windows_nat_target::wait): Update. * target/wait.h (enum target_wait_flag): New. Use DEF_ENUM_FLAGS_TYPE. * target/target.h (target_wait): Change type of options. * target.h (target_options_to_string, default_target_wait): Update. (struct target_ops) <wait>: Change type of options. * target.c (target_wait, default_target_wait, do_option): Change type of "options". (target_options_to_string): Likewise. * target-delegates.c: Rebuild. * target-debug.h (target_debug_print_target_wait_flags): Rename from target_debug_print_options. * sol-thread.c (class sol_thread_target) <wait>: Update. (sol_thread_target::wait): Update. * rs6000-nat.c (class rs6000_nat_target) <wait>: Update. (rs6000_nat_target::wait): Update. * remote.c (class remote_target) <wait, wait_ns, wait_as>: Update. (remote_target::wait_ns, remote_target::wait_as): Change type of "options". (remote_target::wait): Update. * remote-sim.c (struct gdbsim_target) <wait>: Update. (gdbsim_target::wait): Update. * record-full.c (class record_full_base_target) <wait>: Update. (record_full_wait_1): Change type of "options". (record_full_base_target::wait): Update. * record-btrace.c (class record_btrace_target) <wait>: Update. (record_btrace_target::wait): Update. * ravenscar-thread.c (struct ravenscar_thread_target) <wait>: Update. (ravenscar_thread_target::wait): Update. * procfs.c (class procfs_target) <wait>: Update. (procfs_target::wait): Update. * obsd-nat.h (class obsd_nat_target) <wait>: Update. * obsd-nat.c (obsd_nat_target::wait): Update. * nto-procfs.c (struct nto_procfs_target) <wait>: Update. (nto_procfs_target::wait): Update. * nbsd-nat.h (struct nbsd_nat_target) <wait>: Update. * nbsd-nat.c (nbsd_wait): Change type of "options". (nbsd_nat_target::wait): Update. * linux-thread-db.c (class thread_db_target) <wait>: Update. (thread_db_target::wait): Update. * linux-nat.h (class linux_nat_target) <wait>: Update. * linux-nat.c (linux_nat_target::wait): Update. (linux_nat_wait_1): Update. * infrun.c (do_target_wait_1, do_target_wait): Change type of "options". * inf-ptrace.h (struct inf_ptrace_target) <wait>: Update. * inf-ptrace.c (inf_ptrace_target::wait): Update. * go32-nat.c (struct go32_nat_target) <wait>: Update. (go32_nat_target::wait): Update. * gnu-nat.h (struct gnu_nat_target) <wait>: Update. * gnu-nat.c (gnu_nat_target::wait): Update. * fbsd-nat.h (class fbsd_nat_target) <wait>: Update. * fbsd-nat.c (fbsd_nat_target::wait): Update. * darwin-nat.h (class darwin_nat_target) <wait>: Update. * darwin-nat.c (darwin_nat_target::wait): Update. * bsd-uthread.c (struct bsd_uthread_target) <wait>: Update. (bsd_uthread_target::wait): Update. * aix-thread.c (class aix_thread_target) <wait>: Update. (aix_thread_target::wait): Update. gdbserver/ChangeLog 2020-09-18 Tom Tromey <tromey@adacore.com> * netbsd-low.h (class netbsd_process_target) <wait>: Update. * netbsd-low.cc (netbsd_waitpid, netbsd_wait) (netbsd_process_target::wait): Change type of target_options. * win32-low.h (class win32_process_target) <wait>: Update. * win32-low.cc (win32_process_target::wait): Update. * target.h (class process_stratum_target) <wait>: Update. (mywait): Update. * target.cc (mywait, target_wait): Change type of "options". * linux-low.h (class linux_process_target) <wait, wait_1>: Update. * linux-low.cc (linux_process_target::wait) (linux_process_target::wait_1): Update.
2020-09-19 04:20:44 +08:00
target_wait_flags target_options)
{
pid_t pid = netbsd_waitpid (ptid, ourstatus, target_options);
ptid_t wptid = ptid_t (pid);
if (pid == 0)
{
gdb_assert (target_options & TARGET_WNOHANG);
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_ignore ();
return null_ptid;
}
gdb_assert (pid != -1);
/* If the child stopped, keep investigating its status. */
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
if (ourstatus->kind () != TARGET_WAITKIND_STOPPED)
return wptid;
/* Extract the event and thread that received a signal. */
ptrace_siginfo_t psi;
if (ptrace (PT_GET_SIGINFO, pid, &psi, sizeof (psi)) == -1)
perror_with_name (("ptrace"));
/* Pick child's siginfo_t. */
siginfo_t *si = &psi.psi_siginfo;
lwpid_t lwp = psi.psi_lwpid;
int signo = si->si_signo;
const int code = si->si_code;
/* Construct PTID with a specified thread that received the event.
If a signal was targeted to the whole process, lwp is 0. */
wptid = ptid_t (pid, lwp, 0);
/* Bail out on non-debugger oriented signals. */
if (signo != SIGTRAP)
return wptid;
/* Stop examining non-debugger oriented SIGTRAP codes. */
if (code <= SI_USER || code == SI_NOINFO)
return wptid;
/* Process state for threading events. */
ptrace_state_t pst = {};
if (code == TRAP_LWP)
if (ptrace (PT_GET_PROCESS_STATE, pid, &pst, sizeof (pst)) == -1)
perror_with_name (("ptrace"));
if (code == TRAP_LWP && pst.pe_report_event == PTRACE_LWP_EXIT)
{
/* If GDB attaches to a multi-threaded process, exiting
threads might be skipped during post_attach that
have not yet reported their PTRACE_LWP_EXIT event.
Ignore exited events for an unknown LWP. */
thread_info *thr = find_thread_ptid (wptid);
if (thr == nullptr)
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_spurious ();
else
{
/* NetBSD does not store an LWP exit status. */
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_thread_exited (0);
remove_thread (thr);
}
return wptid;
}
if (find_thread_ptid (ptid_t (pid)))
switch_to_thread (find_thread_ptid (wptid));
if (code == TRAP_LWP && pst.pe_report_event == PTRACE_LWP_CREATE)
{
/* If GDB attaches to a multi-threaded process, newborn
threads might be added by nbsd_add_threads that have
not yet reported their PTRACE_LWP_CREATE event. Ignore
born events for an already-known LWP. */
if (find_thread_ptid (wptid))
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_spurious ();
else
{
add_thread (wptid, NULL);
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_thread_created ();
}
return wptid;
}
if (code == TRAP_EXEC)
{
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_execd
(make_unique_xstrdup (netbsd_nat::pid_to_exec_file (pid)));
return wptid;
}
if (code == TRAP_TRACE)
return wptid;
if (code == TRAP_SCE || code == TRAP_SCX)
{
int sysnum = si->si_sysnum;
if (!netbsd_catch_this_syscall(sysnum))
{
/* If the core isn't interested in this event, ignore it. */
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_spurious ();
return wptid;
}
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
if (code == TRAP_SCE)
ourstatus->set_syscall_entry (sysnum);
else
ourstatus->set_syscall_return (sysnum);
return wptid;
}
if (code == TRAP_BRKPT)
{
#ifdef PTRACE_BREAKPOINT_ADJ
CORE_ADDR pc;
struct reg r;
ptrace (PT_GETREGS, pid, &r, psi.psi_lwpid);
pc = PTRACE_REG_PC (&r);
PTRACE_REG_SET_PC (&r, pc - PTRACE_BREAKPOINT_ADJ);
ptrace (PT_SETREGS, pid, &r, psi.psi_lwpid);
#endif
return wptid;
}
/* Unclassified SIGTRAP event. */
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
ourstatus->set_spurious ();
return wptid;
}
/* Implement the wait target_ops method. */
ptid_t
netbsd_process_target::wait (ptid_t ptid, struct target_waitstatus *ourstatus,
Make target_wait options use enum flags This changes TARGET_WNOHANG to be a member of an enum, rather than a define, and also adds a DEF_ENUM_FLAGS_TYPE for this type. Then, it changes target_wait and the various target wait methods to use this type rather than "int". This didn't catch any bugs, but it seems like a decent cleanup nevertheless. I did not change deprecated_target_wait_hook, since that's only used out-of-tree (by Insight), and there didn't seem to be a need. I can't build some of these targets, so I modified them on a best-effort basis. I don't think this patch should go in before the release branch is made. gdb/ChangeLog 2020-09-18 Tom Tromey <tromey@adacore.com> * windows-nat.c (struct windows_nat_target) <wait>: Update. (windows_nat_target::wait): Update. * target/wait.h (enum target_wait_flag): New. Use DEF_ENUM_FLAGS_TYPE. * target/target.h (target_wait): Change type of options. * target.h (target_options_to_string, default_target_wait): Update. (struct target_ops) <wait>: Change type of options. * target.c (target_wait, default_target_wait, do_option): Change type of "options". (target_options_to_string): Likewise. * target-delegates.c: Rebuild. * target-debug.h (target_debug_print_target_wait_flags): Rename from target_debug_print_options. * sol-thread.c (class sol_thread_target) <wait>: Update. (sol_thread_target::wait): Update. * rs6000-nat.c (class rs6000_nat_target) <wait>: Update. (rs6000_nat_target::wait): Update. * remote.c (class remote_target) <wait, wait_ns, wait_as>: Update. (remote_target::wait_ns, remote_target::wait_as): Change type of "options". (remote_target::wait): Update. * remote-sim.c (struct gdbsim_target) <wait>: Update. (gdbsim_target::wait): Update. * record-full.c (class record_full_base_target) <wait>: Update. (record_full_wait_1): Change type of "options". (record_full_base_target::wait): Update. * record-btrace.c (class record_btrace_target) <wait>: Update. (record_btrace_target::wait): Update. * ravenscar-thread.c (struct ravenscar_thread_target) <wait>: Update. (ravenscar_thread_target::wait): Update. * procfs.c (class procfs_target) <wait>: Update. (procfs_target::wait): Update. * obsd-nat.h (class obsd_nat_target) <wait>: Update. * obsd-nat.c (obsd_nat_target::wait): Update. * nto-procfs.c (struct nto_procfs_target) <wait>: Update. (nto_procfs_target::wait): Update. * nbsd-nat.h (struct nbsd_nat_target) <wait>: Update. * nbsd-nat.c (nbsd_wait): Change type of "options". (nbsd_nat_target::wait): Update. * linux-thread-db.c (class thread_db_target) <wait>: Update. (thread_db_target::wait): Update. * linux-nat.h (class linux_nat_target) <wait>: Update. * linux-nat.c (linux_nat_target::wait): Update. (linux_nat_wait_1): Update. * infrun.c (do_target_wait_1, do_target_wait): Change type of "options". * inf-ptrace.h (struct inf_ptrace_target) <wait>: Update. * inf-ptrace.c (inf_ptrace_target::wait): Update. * go32-nat.c (struct go32_nat_target) <wait>: Update. (go32_nat_target::wait): Update. * gnu-nat.h (struct gnu_nat_target) <wait>: Update. * gnu-nat.c (gnu_nat_target::wait): Update. * fbsd-nat.h (class fbsd_nat_target) <wait>: Update. * fbsd-nat.c (fbsd_nat_target::wait): Update. * darwin-nat.h (class darwin_nat_target) <wait>: Update. * darwin-nat.c (darwin_nat_target::wait): Update. * bsd-uthread.c (struct bsd_uthread_target) <wait>: Update. (bsd_uthread_target::wait): Update. * aix-thread.c (class aix_thread_target) <wait>: Update. (aix_thread_target::wait): Update. gdbserver/ChangeLog 2020-09-18 Tom Tromey <tromey@adacore.com> * netbsd-low.h (class netbsd_process_target) <wait>: Update. * netbsd-low.cc (netbsd_waitpid, netbsd_wait) (netbsd_process_target::wait): Change type of target_options. * win32-low.h (class win32_process_target) <wait>: Update. * win32-low.cc (win32_process_target::wait): Update. * target.h (class process_stratum_target) <wait>: Update. (mywait): Update. * target.cc (mywait, target_wait): Change type of "options". * linux-low.h (class linux_process_target) <wait, wait_1>: Update. * linux-low.cc (linux_process_target::wait) (linux_process_target::wait_1): Update.
2020-09-19 04:20:44 +08:00
target_wait_flags target_options)
{
while (true)
{
ptid_t wptid = netbsd_wait (ptid, ourstatus, target_options);
/* Register thread in the gdbcore if a thread was not reported earlier.
This is required after ::create_inferior, when the gdbcore does not
know about the first internal thread.
This may also happen on attach, when an event is registered on a thread
that was not fully initialized during the attach stage. */
if (wptid.lwp () != 0 && !find_thread_ptid (wptid)
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
&& ourstatus->kind () != TARGET_WAITKIND_THREAD_EXITED)
add_thread (wptid, nullptr);
gdb, gdbserver: make target_waitstatus safe I stumbled on a bug caused by the fact that a code path read target_waitstatus::value::sig (expecting it to contain a gdb_signal value) while target_waitstatus::kind was TARGET_WAITKIND_FORKED. This meant that the active union field was in fact target_waitstatus::value::related_pid, and contained a ptid. The read signal value was therefore garbage, and that caused GDB to crash soon after. Or, since that GDB was built with ubsan, this nice error message: /home/simark/src/binutils-gdb/gdb/linux-nat.c:1271:12: runtime error: load of value 2686365, which is not a valid value for type 'gdb_signal' Despite being a large-ish change, I think it would be nice to make target_waitstatus safe against that kind of bug. As already done elsewhere (e.g. dynamic_prop), validate that the type of value read from the union matches what is supposed to be the active field. - Make the kind and value of target_waitstatus private. - Make the kind initialized to TARGET_WAITKIND_IGNORE on target_waitstatus construction. This is what most users appear to do explicitly. - Add setters, one for each kind. Each setter takes as a parameter the data associated to that kind, if any. This makes it impossible to forget to attach the associated data. - Add getters, one for each associated data type. Each getter validates that the data type fetched by the user matches the wait status kind. - Change "integer" to "exit_status", "related_pid" to "child_ptid", just because that's more precise terminology. - Fix all users. That last point is semi-mechanical. There are a lot of obvious changes, but some less obvious ones. For example, it's not possible to set the kind at some point and the associated data later, as some users did. But in any case, the intent of the code should not change in this patch. This was tested on x86-64 Linux (unix, native-gdbserver and native-extended-gdbserver boards). It was built-tested on x86-64 FreeBSD, NetBSD, MinGW and macOS. The rest of the changes to native files was done as a best effort. If I forgot any place to update in these files, it should be easy to fix (unless the change happens to reveal an actual bug). Change-Id: I0ae967df1ff6e28de78abbe3ac9b4b2ff4ad03b7
2021-10-22 04:12:04 +08:00
switch (ourstatus->kind ())
{
case TARGET_WAITKIND_EXITED:
case TARGET_WAITKIND_STOPPED:
case TARGET_WAITKIND_SIGNALLED:
case TARGET_WAITKIND_FORKED:
case TARGET_WAITKIND_VFORKED:
case TARGET_WAITKIND_EXECD:
case TARGET_WAITKIND_VFORK_DONE:
case TARGET_WAITKIND_SYSCALL_ENTRY:
case TARGET_WAITKIND_SYSCALL_RETURN:
/* Pass the result to the generic code. */
return wptid;
case TARGET_WAITKIND_THREAD_CREATED:
case TARGET_WAITKIND_THREAD_EXITED:
/* The core needlessly stops on these events. */
[[fallthrough]];
case TARGET_WAITKIND_SPURIOUS:
/* Spurious events are unhandled by the gdbserver core. */
if (ptrace (PT_CONTINUE, current_process ()->pid, (void *) 1, 0)
== -1)
perror_with_name (("ptrace"));
break;
default:
error (("Unknown stopped status"));
}
}
}
/* Implement the kill target_ops method. */
int
netbsd_process_target::kill (process_info *process)
{
pid_t pid = process->pid;
if (ptrace (PT_KILL, pid, nullptr, 0) == -1)
return -1;
int status;
if (gdb::handle_eintr (-1, ::waitpid, pid, &status, 0) == -1)
return -1;
mourn (process);
return 0;
}
/* Implement the detach target_ops method. */
int
netbsd_process_target::detach (process_info *process)
{
pid_t pid = process->pid;
ptrace (PT_DETACH, pid, (void *) 1, 0);
mourn (process);
return 0;
}
/* Implement the mourn target_ops method. */
void
netbsd_process_target::mourn (struct process_info *proc)
{
for_each_thread (proc->pid, remove_thread);
remove_process (proc);
}
/* Implement the join target_ops method. */
void
netbsd_process_target::join (int pid)
{
/* The PT_DETACH is sufficient to detach from the process.
So no need to do anything extra. */
}
/* Implement the thread_alive target_ops method. */
bool
netbsd_process_target::thread_alive (ptid_t ptid)
{
return netbsd_nat::thread_alive (ptid);
}
/* Implement the fetch_registers target_ops method. */
void
netbsd_process_target::fetch_registers (struct regcache *regcache, int regno)
{
const netbsd_regset_info *regset = get_regs_info ();
ptid_t inferior_ptid = ptid_of (current_thread);
while (regset->size >= 0)
{
std::vector<char> buf;
buf.resize (regset->size);
int res = ptrace (regset->get_request, inferior_ptid.pid (), buf.data (),
inferior_ptid.lwp ());
if (res == -1)
perror_with_name (("ptrace"));
regset->store_function (regcache, buf.data ());
regset++;
}
}
/* Implement the store_registers target_ops method. */
void
netbsd_process_target::store_registers (struct regcache *regcache, int regno)
{
const netbsd_regset_info *regset = get_regs_info ();
ptid_t inferior_ptid = ptid_of (current_thread);
while (regset->size >= 0)
{
std::vector<char> buf;
buf.resize (regset->size);
int res = ptrace (regset->get_request, inferior_ptid.pid (), buf.data (),
inferior_ptid.lwp ());
if (res == -1)
perror_with_name (("ptrace"));
/* Then overlay our cached registers on that. */
regset->fill_function (regcache, buf.data ());
/* Only now do we write the register set. */
res = ptrace (regset->set_request, inferior_ptid.pid (), buf. data (),
inferior_ptid.lwp ());
if (res == -1)
perror_with_name (("ptrace"));
regset++;
}
}
/* Implement the read_memory target_ops method. */
int
netbsd_process_target::read_memory (CORE_ADDR memaddr, unsigned char *myaddr,
int size)
{
pid_t pid = current_process ()->pid;
return netbsd_nat::read_memory (pid, myaddr, memaddr, size, nullptr);
}
/* Implement the write_memory target_ops method. */
int
netbsd_process_target::write_memory (CORE_ADDR memaddr,
const unsigned char *myaddr, int size)
{
pid_t pid = current_process ()->pid;
return netbsd_nat::write_memory (pid, myaddr, memaddr, size, nullptr);
}
/* Implement the request_interrupt target_ops method. */
void
netbsd_process_target::request_interrupt ()
{
ptid_t inferior_ptid = ptid_of (get_first_thread ());
::kill (inferior_ptid.pid (), SIGINT);
}
/* Read the AUX Vector for the specified PID, wrapping the ptrace(2) call
with the PIOD_READ_AUXV operation and using the PT_IO standard input
and output arguments. */
static size_t
netbsd_read_auxv(pid_t pid, void *offs, void *addr, size_t len)
{
struct ptrace_io_desc pio;
pio.piod_op = PIOD_READ_AUXV;
pio.piod_offs = offs;
pio.piod_addr = addr;
pio.piod_len = len;
if (ptrace (PT_IO, pid, &pio, 0) == -1)
perror_with_name (("ptrace"));
return pio.piod_len;
}
/* Copy LEN bytes from inferior's auxiliary vector starting at OFFSET
to debugger memory starting at MYADDR. */
int
netbsd_process_target::read_auxv (int pid, CORE_ADDR offset,
unsigned char *myaddr, unsigned int len)
{
return netbsd_read_auxv (pid, (void *) (intptr_t) offset, myaddr, len);
}
bool
netbsd_process_target::supports_z_point_type (char z_type)
{
switch (z_type)
{
case Z_PACKET_SW_BP:
return true;
case Z_PACKET_HW_BP:
case Z_PACKET_WRITE_WP:
case Z_PACKET_READ_WP:
case Z_PACKET_ACCESS_WP:
default:
return false; /* Not supported. */
}
}
/* Insert {break/watch}point at address ADDR. SIZE is not used. */
int
netbsd_process_target::insert_point (enum raw_bkpt_type type, CORE_ADDR addr,
int size, struct raw_breakpoint *bp)
{
switch (type)
{
case raw_bkpt_type_sw:
return insert_memory_breakpoint (bp);
case raw_bkpt_type_hw:
case raw_bkpt_type_write_wp:
case raw_bkpt_type_read_wp:
case raw_bkpt_type_access_wp:
default:
return 1; /* Not supported. */
}
}
/* Remove {break/watch}point at address ADDR. SIZE is not used. */
int
netbsd_process_target::remove_point (enum raw_bkpt_type type, CORE_ADDR addr,
int size, struct raw_breakpoint *bp)
{
switch (type)
{
case raw_bkpt_type_sw:
return remove_memory_breakpoint (bp);
case raw_bkpt_type_hw:
case raw_bkpt_type_write_wp:
case raw_bkpt_type_read_wp:
case raw_bkpt_type_access_wp:
default:
return 1; /* Not supported. */
}
}
/* Implement the stopped_by_sw_breakpoint target_ops method. */
bool
netbsd_process_target::stopped_by_sw_breakpoint ()
{
ptrace_siginfo_t psi;
pid_t pid = current_process ()->pid;
if (ptrace (PT_GET_SIGINFO, pid, &psi, sizeof (psi)) == -1)
perror_with_name (("ptrace"));
return psi.psi_siginfo.si_signo == SIGTRAP &&
psi.psi_siginfo.si_code == TRAP_BRKPT;
}
/* Implement the supports_stopped_by_sw_breakpoint target_ops method. */
bool
netbsd_process_target::supports_stopped_by_sw_breakpoint ()
{
return true;
}
/* Implement the supports_qxfer_siginfo target_ops method. */
bool
netbsd_process_target::supports_qxfer_siginfo ()
{
return true;
}
/* Implement the qxfer_siginfo target_ops method. */
int
netbsd_process_target::qxfer_siginfo (const char *annex, unsigned char *readbuf,
unsigned const char *writebuf,
CORE_ADDR offset, int len)
{
if (current_thread == nullptr)
return -1;
pid_t pid = current_process ()->pid;
return netbsd_nat::qxfer_siginfo(pid, annex, readbuf, writebuf, offset, len);
}
/* Implement the supports_non_stop target_ops method. */
bool
netbsd_process_target::supports_non_stop ()
{
return false;
}
/* Implement the supports_multi_process target_ops method. */
bool
netbsd_process_target::supports_multi_process ()
{
return true;
}
/* Check if fork events are supported. */
bool
netbsd_process_target::supports_fork_events ()
{
return false;
}
/* Check if vfork events are supported. */
bool
netbsd_process_target::supports_vfork_events ()
{
return false;
}
/* Check if exec events are supported. */
bool
netbsd_process_target::supports_exec_events ()
{
return true;
}
/* Implement the supports_disable_randomization target_ops method. */
bool
netbsd_process_target::supports_disable_randomization ()
{
return false;
}
/* Extract &phdr and num_phdr in the inferior. Return 0 on success. */
template <typename T>
int get_phdr_phnum_from_proc_auxv (const pid_t pid,
CORE_ADDR *phdr_memaddr, int *num_phdr)
{
typedef typename std::conditional<sizeof(T) == sizeof(int64_t),
Aux64Info, Aux32Info>::type auxv_type;
const size_t auxv_size = sizeof (auxv_type);
const size_t auxv_buf_size = 128 * sizeof (auxv_type);
std::vector<char> auxv_buf;
auxv_buf.resize (auxv_buf_size);
netbsd_read_auxv (pid, nullptr, auxv_buf.data (), auxv_buf_size);
*phdr_memaddr = 0;
*num_phdr = 0;
for (char *buf = auxv_buf.data ();
buf < (auxv_buf.data () + auxv_buf_size);
buf += auxv_size)
{
auxv_type *const aux = (auxv_type *) buf;
switch (aux->a_type)
{
case AT_PHDR:
*phdr_memaddr = aux->a_v;
break;
case AT_PHNUM:
*num_phdr = aux->a_v;
break;
}
if (*phdr_memaddr != 0 && *num_phdr != 0)
break;
}
if (*phdr_memaddr == 0 || *num_phdr == 0)
{
warning ("Unexpected missing AT_PHDR and/or AT_PHNUM: "
"phdr_memaddr = %s, phdr_num = %d",
core_addr_to_string (*phdr_memaddr), *num_phdr);
return 2;
}
return 0;
}
/* Return &_DYNAMIC (via PT_DYNAMIC) in the inferior, or 0 if not present. */
template <typename T>
static CORE_ADDR
get_dynamic (const pid_t pid)
{
typedef typename std::conditional<sizeof(T) == sizeof(int64_t),
Elf64_Phdr, Elf32_Phdr>::type phdr_type;
const int phdr_size = sizeof (phdr_type);
CORE_ADDR phdr_memaddr;
int num_phdr;
if (get_phdr_phnum_from_proc_auxv<T> (pid, &phdr_memaddr, &num_phdr))
return 0;
std::vector<unsigned char> phdr_buf;
phdr_buf.resize (num_phdr * phdr_size);
if (netbsd_nat::read_memory (pid, phdr_buf.data (), phdr_memaddr,
phdr_buf.size (), nullptr))
return 0;
/* Compute relocation: it is expected to be 0 for "regular" executables,
non-zero for PIE ones. */
CORE_ADDR relocation = -1;
for (int i = 0; relocation == -1 && i < num_phdr; i++)
{
phdr_type *const p = (phdr_type *) (phdr_buf.data () + i * phdr_size);
if (p->p_type == PT_PHDR)
relocation = phdr_memaddr - p->p_vaddr;
}
if (relocation == -1)
{
/* PT_PHDR is optional, but necessary for PIE in general. Fortunately
any real world executables, including PIE executables, have always
PT_PHDR present. PT_PHDR is not present in some shared libraries or
in fpc (Free Pascal 2.4) binaries but neither of those have a need for
or present DT_DEBUG anyway (fpc binaries are statically linked).
Therefore if there exists DT_DEBUG there is always also PT_PHDR.
GDB could find RELOCATION also from AT_ENTRY - e_entry. */
return 0;
}
for (int i = 0; i < num_phdr; i++)
{
phdr_type *const p = (phdr_type *) (phdr_buf.data () + i * phdr_size);
if (p->p_type == PT_DYNAMIC)
return p->p_vaddr + relocation;
}
return 0;
}
/* Return &_r_debug in the inferior, or -1 if not present. Return value
can be 0 if the inferior does not yet have the library list initialized.
We look for DT_MIPS_RLD_MAP first. MIPS executables use this instead of
DT_DEBUG, although they sometimes contain an unused DT_DEBUG entry too. */
template <typename T>
static CORE_ADDR
get_r_debug (const pid_t pid)
{
typedef typename std::conditional<sizeof(T) == sizeof(int64_t),
Elf64_Dyn, Elf32_Dyn>::type dyn_type;
const int dyn_size = sizeof (dyn_type);
unsigned char buf[sizeof (dyn_type)]; /* The larger of the two. */
CORE_ADDR map = -1;
CORE_ADDR dynamic_memaddr = get_dynamic<T> (pid);
if (dynamic_memaddr == 0)
return map;
while (netbsd_nat::read_memory (pid, buf, dynamic_memaddr, dyn_size, nullptr)
== 0)
{
dyn_type *const dyn = (dyn_type *) buf;
#if defined DT_MIPS_RLD_MAP
union
{
T map;
unsigned char buf[sizeof (T)];
}
rld_map;
if (dyn->d_tag == DT_MIPS_RLD_MAP)
{
if (netbsd_nat::read_memory (pid, rld_map.buf, dyn->d_un.d_val,
sizeof (rld_map.buf), nullptr) == 0)
return rld_map.map;
else
break;
}
#endif /* DT_MIPS_RLD_MAP */
if (dyn->d_tag == DT_DEBUG && map == -1)
map = dyn->d_un.d_val;
if (dyn->d_tag == DT_NULL)
break;
dynamic_memaddr += dyn_size;
}
return map;
}
/* Read one pointer from MEMADDR in the inferior. */
static int
read_one_ptr (const pid_t pid, CORE_ADDR memaddr, CORE_ADDR *ptr, int ptr_size)
{
/* Go through a union so this works on either big or little endian
hosts, when the inferior's pointer size is smaller than the size
of CORE_ADDR. It is assumed the inferior's endianness is the
same of the superior's. */
union
{
CORE_ADDR core_addr;
unsigned int ui;
unsigned char uc;
} addr;
int ret = netbsd_nat::read_memory (pid, &addr.uc, memaddr, ptr_size, nullptr);
if (ret == 0)
{
if (ptr_size == sizeof (CORE_ADDR))
*ptr = addr.core_addr;
else if (ptr_size == sizeof (unsigned int))
*ptr = addr.ui;
else
gdb_assert_not_reached ("unhandled pointer size");
}
return ret;
}
/* Construct qXfer:libraries-svr4:read reply. */
template <typename T>
int
netbsd_qxfer_libraries_svr4 (const pid_t pid, const char *annex,
unsigned char *readbuf,
unsigned const char *writebuf,
CORE_ADDR offset, int len)
{
struct link_map_offsets
{
/* Offset and size of r_debug.r_version. */
int r_version_offset;
/* Offset and size of r_debug.r_map. */
int r_map_offset;
/* Offset to l_addr field in struct link_map. */
int l_addr_offset;
/* Offset to l_name field in struct link_map. */
int l_name_offset;
/* Offset to l_ld field in struct link_map. */
int l_ld_offset;
/* Offset to l_next field in struct link_map. */
int l_next_offset;
/* Offset to l_prev field in struct link_map. */
int l_prev_offset;
};
static const struct link_map_offsets lmo_32bit_offsets =
{
0, /* r_version offset. */
4, /* r_debug.r_map offset. */
0, /* l_addr offset in link_map. */
4, /* l_name offset in link_map. */
8, /* l_ld offset in link_map. */
12, /* l_next offset in link_map. */
16 /* l_prev offset in link_map. */
};
static const struct link_map_offsets lmo_64bit_offsets =
{
0, /* r_version offset. */
8, /* r_debug.r_map offset. */
0, /* l_addr offset in link_map. */
8, /* l_name offset in link_map. */
16, /* l_ld offset in link_map. */
24, /* l_next offset in link_map. */
32 /* l_prev offset in link_map. */
};
CORE_ADDR lm_addr = 0, lm_prev = 0;
CORE_ADDR l_name, l_addr, l_ld, l_next, l_prev;
int header_done = 0;
const struct link_map_offsets *lmo
= ((sizeof (T) == sizeof (int64_t))
? &lmo_64bit_offsets : &lmo_32bit_offsets);
int ptr_size = sizeof (T);
while (annex[0] != '\0')
{
const char *sep = strchr (annex, '=');
if (sep == nullptr)
break;
int name_len = sep - annex;
CORE_ADDR *addrp;
if (name_len == 5 && startswith (annex, "start"))
addrp = &lm_addr;
else if (name_len == 4 && startswith (annex, "prev"))
addrp = &lm_prev;
else
{
annex = strchr (sep, ';');
if (annex == nullptr)
break;
annex++;
continue;
}
annex = decode_address_to_semicolon (addrp, sep + 1);
}
if (lm_addr == 0)
{
CORE_ADDR r_debug = get_r_debug<T> (pid);
/* We failed to find DT_DEBUG. Such situation will not change
for this inferior - do not retry it. Report it to GDB as
E01, see for the reasons at the GDB solib-svr4.c side. */
if (r_debug == (CORE_ADDR) -1)
return -1;
if (r_debug != 0)
{
CORE_ADDR map_offset = r_debug + lmo->r_map_offset;
if (read_one_ptr (pid, map_offset, &lm_addr, ptr_size) != 0)
warning ("unable to read r_map from %s",
core_addr_to_string (map_offset));
}
}
std::string document = "<library-list-svr4 version=\"1.0\"";
while (lm_addr
&& read_one_ptr (pid, lm_addr + lmo->l_name_offset,
&l_name, ptr_size) == 0
&& read_one_ptr (pid, lm_addr + lmo->l_addr_offset,
&l_addr, ptr_size) == 0
&& read_one_ptr (pid, lm_addr + lmo->l_ld_offset,
&l_ld, ptr_size) == 0
&& read_one_ptr (pid, lm_addr + lmo->l_prev_offset,
&l_prev, ptr_size) == 0
&& read_one_ptr (pid, lm_addr + lmo->l_next_offset,
&l_next, ptr_size) == 0)
{
if (lm_prev != l_prev)
{
warning ("Corrupted shared library list: 0x%lx != 0x%lx",
(long) lm_prev, (long) l_prev);
break;
}
/* Ignore the first entry even if it has valid name as the first entry
corresponds to the main executable. The first entry should not be
skipped if the dynamic loader was loaded late by a static executable
(see solib-svr4.c parameter ignore_first). But in such case the main
executable does not have PT_DYNAMIC present and this function already
exited above due to failed get_r_debug. */
if (lm_prev == 0)
string_appendf (document, " main-lm=\"0x%lx\"",
(unsigned long) lm_addr);
else
{
unsigned char libname[PATH_MAX];
/* Not checking for error because reading may stop before
we've got PATH_MAX worth of characters. */
libname[0] = '\0';
netbsd_nat::read_memory (pid, libname, l_name, sizeof (libname) - 1,
nullptr);
libname[sizeof (libname) - 1] = '\0';
if (libname[0] != '\0')
{
if (!header_done)
{
/* Terminate `<library-list-svr4'. */
document += '>';
header_done = 1;
}
string_appendf (document, "<library name=\"");
xml_escape_text_append (document, (char *) libname);
string_appendf (document, "\" lm=\"0x%lx\" "
"l_addr=\"0x%lx\" l_ld=\"0x%lx\"/>",
(unsigned long) lm_addr, (unsigned long) l_addr,
(unsigned long) l_ld);
}
}
lm_prev = lm_addr;
lm_addr = l_next;
}
if (!header_done)
{
/* Empty list; terminate `<library-list-svr4'. */
document += "/>";
}
else
document += "</library-list-svr4>";
int document_len = document.length ();
if (offset < document_len)
document_len -= offset;
else
document_len = 0;
if (len > document_len)
len = document_len;
memcpy (readbuf, document.data () + offset, len);
return len;
}
/* Return true if FILE is a 64-bit ELF file,
false if the file is not a 64-bit ELF file,
and error if the file is not accessible or doesn't exist. */
static bool
elf_64_file_p (const char *file)
{
int fd = gdb::handle_eintr (-1, ::open, file, O_RDONLY);
if (fd < 0)
perror_with_name (("open"));
Elf64_Ehdr header;
ssize_t ret = gdb::handle_eintr (-1, ::read, fd, &header, sizeof (header));
if (ret == -1)
perror_with_name (("read"));
gdb::handle_eintr (-1, ::close, fd);
if (ret != sizeof (header))
error ("Cannot read ELF file header: %s", file);
if (header.e_ident[EI_MAG0] != ELFMAG0
|| header.e_ident[EI_MAG1] != ELFMAG1
|| header.e_ident[EI_MAG2] != ELFMAG2
|| header.e_ident[EI_MAG3] != ELFMAG3)
error ("Unrecognized ELF file header: %s", file);
return header.e_ident[EI_CLASS] == ELFCLASS64;
}
/* Construct qXfer:libraries-svr4:read reply. */
int
netbsd_process_target::qxfer_libraries_svr4 (const char *annex,
unsigned char *readbuf,
unsigned const char *writebuf,
CORE_ADDR offset, int len)
{
if (writebuf != nullptr)
return -2;
if (readbuf == nullptr)
return -1;
struct process_info *proc = current_process ();
pid_t pid = proc->pid;
bool is_elf64 = elf_64_file_p (netbsd_nat::pid_to_exec_file (pid));
if (is_elf64)
return netbsd_qxfer_libraries_svr4<int64_t> (pid, annex, readbuf,
writebuf, offset, len);
else
return netbsd_qxfer_libraries_svr4<int32_t> (pid, annex, readbuf,
writebuf, offset, len);
}
/* Implement the supports_qxfer_libraries_svr4 target_ops method. */
bool
netbsd_process_target::supports_qxfer_libraries_svr4 ()
{
return true;
}
/* Return the name of a file that can be opened to get the symbols for
the child process identified by PID. */
const char *
netbsd_process_target::pid_to_exec_file (pid_t pid)
{
return netbsd_nat::pid_to_exec_file (pid);
}
/* Implementation of the target_ops method "supports_pid_to_exec_file". */
bool
netbsd_process_target::supports_pid_to_exec_file ()
{
return true;
}
/* Implementation of the target_ops method "supports_hardware_single_step". */
bool
netbsd_process_target::supports_hardware_single_step ()
{
return true;
}
/* Implementation of the target_ops method "sw_breakpoint_from_kind". */
const gdb_byte *
netbsd_process_target::sw_breakpoint_from_kind (int kind, int *size)
{
static gdb_byte brkpt[PTRACE_BREAKPOINT_SIZE] = {*PTRACE_BREAKPOINT};
*size = PTRACE_BREAKPOINT_SIZE;
return brkpt;
}
/* Implement the thread_name target_ops method. */
const char *
netbsd_process_target::thread_name (ptid_t ptid)
{
return netbsd_nat::thread_name (ptid);
}
/* Implement the supports_catch_syscall target_ops method. */
bool
netbsd_process_target::supports_catch_syscall ()
{
return true;
}
/* Implement the supports_read_auxv target_ops method. */
bool
netbsd_process_target::supports_read_auxv ()
{
return true;
}
void
initialize_low ()
{
set_target_ops (the_netbsd_target);
}