Use the inferior parameter now available in jit_inferior_created_hook.
It is passed down to jit_inferior_init, which uses it as much as
possible instead of the current inferior or current program space.
gdb/ChangeLog:
* jit.c (jit_reader_load_command): Pass current inferior.
(jit_inferior_init): Change parameter type to inferior, use it.
(jit_inferior_created): Remove.
(jit_inferior_created_hook): Pass inferior parameter down.
(_initialize_jit): Use jit_inferior_created_hook instead of
jit_inferior_created.
* jit.h (jit_inferior_created_hook): Add inferior parameter.
* infrun.c (follow_exec): Pass inferior to
jit_inferior_created_hook.
Change-Id: If3a2114a933370dd313d5abd623136d273cdb8fa
Debugging with "maintenance set target-async off" on Linux has been
broken since 5b6d1e4fa4 ("Multi-target support").
The issue is easy to reproduce:
$ ./gdb -q --data-directory=data-directory -nx ./test
Reading symbols from ./test...
(gdb) maintenance set target-async off
(gdb) start
Temporary breakpoint 1 at 0x1151: file test.c, line 5.
Starting program: /home/simark/build/binutils-gdb/gdb/test
... and it hangs there.
The difference between pre-5b6d1e4fa4f and 5b6d1e4fa4 is that
fetch_inferior_event now calls target_wait with TARGET_WNOHANG for
non-async-capable targets, whereas it didn't before.
For non-async-capable targets, this is how it's expected to work when
resuming execution:
1. we call resume
2. the infrun async handler is marked in prepare_to_wait, to immediately
wake up the event loop when we get back to it
3. fetch_inferior_event calls the target's wait method without
TARGET_WNOHANG, effectively blocking until the target has something
to report
However, since we call the target's wait method with TARGET_WNOHANG,
this happens:
1. we call resume
2. the infrun async handler is marked in prepare_to_wait, to immediately
wake up the event loop when we get back to it
3. fetch_inferior_event calls the target's wait method with
TARGET_WNOHANG, the target has nothing to report yet
4. we go back to blocking on the event loop
5. SIGCHLD finally arrives, but the event loop is not woken up, because
we are not in async mode. Normally, we should have been stuck in
waitpid the SIGCHLD would have unblocked us.
We end up in this situation because these two necessary conditions are
met:
1. GDB uses the TARGET_WNOHANG option with a target that can't do async.
I don't think this makes sense. I mean, it's technically possible,
the doc for TARGET_WNOHANG is:
/* Return immediately if there's no event already queued. If this
options is not requested, target_wait blocks waiting for an
event. */
TARGET_WNOHANG = 1,
... which isn't in itself necessarily incompatible with synchronous
targets. It could be possible for a target to support non-blocking
polls, while not having a way to asynchronously wake up the event
loop, which is also necessary to support async. But as of today,
we don't expect GDB and sync targets to work this way.
2. The linux-nat target, even in the mode where it emulates a
synchronous target (with "maintenance set target-async off") respects
TARGET_WNOHANG. Other non-async targets, such as windows_nat_target,
simply don't check / support TARGET_WNOHANG, so their wait method is
always blocking.
Fix the first issue by avoiding using TARGET_WNOHANG on non-async
targets, in do_target_wait_1. Add an assert in target_wait to verify it
doesn't happen.
The new test gdb.base/maint-target-async-off.exp is a simple test that
just tries running to main and then to the end of the program, with
"maintenance set target-async off".
gdb/ChangeLog:
PR gdb/26642
* infrun.c (do_target_wait_1): Clear TARGET_WNOHANG if the
target can't do async.
* target.c (target_wait): Assert that we don't pass
TARGET_WNOHANG to a target that can't async.
gdb/testsuite/ChangeLog:
PR gdb/26642
* gdb.base/maint-target-async-off.c: New test.
* gdb.base/maint-target-async-off.exp: New test.
Change-Id: I69ad3a14598863d21338a8c4e78700a58ce7ad86
The following patch needs to output debug prints from gdbsupport code.
Move debug_prefixed_vprintf so that it is possible to use it from
gdbsupport.
gdb/ChangeLog:
* debug.c (debug_prefixed_vprintf): Move to gdbsupport.
* debug.h: Remove.
* infrun.c: Include gdbsupport/common-debug.h.
* linux-nat.c: Likewise.
gdbsupport/ChangeLog:
* common-debug.cc (debug_prefixed_vprintf): Move here.
* common-debug.h (debug_prefixed_vprintf): Move here.
Change-Id: I5170065fc10a7a49c0f1bba67c691decb2cf3bcb
Assign names to async event/signal handlers. They will be used in debug
messages when file handlers are invoked.
Unlike in the previous patch, the names are not copied in the structure,
since we don't need to (all names are string literals for the moment).
gdb/ChangeLog:
* async-event.h (create_async_signal_handler): Add name
parameter.
(create_async_event_handler): Likewise.
* async-event.c (struct async_signal_handler) <name>: New field.
(struct async_event_handler) <name>: New field.
(create_async_signal_handler): Assign name.
(create_async_event_handler): Assign name.
* event-top.c (async_init_signals): Pass name when creating
handler.
* infrun.c (_initialize_infrun): Likewise.
* record-btrace.c (record_btrace_push_target): Likewise.
* record-full.c (record_full_open): Likewise.
* remote-notif.c (remote_notif_state_allocate): Likewise.
* remote.c (remote_target::open_1): Likewise.
* tui/tui-win.c (tui_initialize_win): Likewise.
Change-Id: Icd9d9f775542ae5fc2cd148c12f481e7885936d5
I noticed that non of the listeners of the inferior_created observable
used either of the arguments. Remove them. This in turn allows
removing the target parameter of post_create_inferior.
Tested only by rebuilding.
gdb/ChangeLog:
* observable.h <inferior_created>: Remove parameters. Update all
listeners.
* inferior.h (post_create_inferior): Remove target parameter.
Update all callers.
Change-Id: I8944cefdc4447ed5347dc927b75abf1e7a0e27e6
This changes the object-like macro target_have_steppable_watchpoint
into an inline function.
gdb/ChangeLog
2020-09-28 Tom Tromey <tom@tromey.com>
* infrun.c (displaced_step_fixup, thread_still_needs_step_over)
(handle_signal_stop): Update.
* procfs.c (procfs_target::insert_watchpoint): Update.
* target.h (target_have_steppable_watchpoint): Now a function.
This changes the object-like macro target_can_lock_scheduler into an
inline function.
gdb/ChangeLog
2020-09-28 Tom Tromey <tom@tromey.com>
* infrun.c (set_schedlock_func): Update.
* target.h (target_can_lock_scheduler): Now a function.
This changes target_can_execute_reverse from an object-like macro to
an inline function.
gdb/ChangeLog
2020-09-28 Tom Tromey <tom@tromey.com>
* mi/mi-main.c (exec_reverse_continue)
(mi_cmd_list_target_features): Update.
* infrun.c (set_exec_direction_func): Update.
* target.c (default_execution_direction): Update.
* reverse.c (exec_reverse_once): Update.
* target.h (target_can_execute_reverse): Now a function.
PR gdb/26598 notes that, before commit bcfe6157ca ("Use the linkage
name if it exists"), the "skip" command would match the demangled name
of a symbol, but now only matches the linkage name.
This patch fixes this regression. I looked at all calls to
function_name_is_marked_for_skip, and only one used the linkage name.
2020-09-16 Tom Tromey <tromey@adacore.com>
PR gdb/26598:
* infrun.c (fill_in_stop_func): Use find_pc_partial_function_sym.
gdb/testsuite/ChangeLog
2020-09-16 Tom Tromey <tromey@adacore.com>
PR gdb/26598:
* gdb.base/skipcxx.exp: New file.
* gdb.base/skipcxx.cc: New file.
I noticed this while testing the GDB in the context of the upcoming
GDB 10 release branching, because part of the process involves setting
development to False, which in turn changes the default for including
unittest to false as well. As a result, without this patch, we get
compilation errors in infrun.c such as:
infrun.c:9219:5: error: `scoped_mock_context' was not declared in this scope
This patch fixes it by bracketing the unitttest in namespace selftest
with an #if GDB_SELF_TEST.
gdb/ChangeLog:
* infrun.c (namespace selftests): Only define #if GDB_SELF_TEST.
Tested on x86_64-linux, with and without self-tests.
Use the available `switch_to_target_no_thread` function to switch the
target. This is a refactoring.
gdb/ChangeLog:
2020-09-07 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* infrun.c (fetch_inferior_event): Use
`switch_to_target_no_thread` to switch the target.
To help ensure that all debug statements have the same format, introduce
the debug_prefixed_vprintf helper. Implement linux_nat_debug_printf_1
and infrun_debug_printf_1 with it.
I would eventually like to style the module and function name with some
color, to help them stick out, but I don't really know how to do that
yet, it can always be done later.
gdb/ChangeLog:
* debug.h: New file.
* debug.c (debug_prefixed_vprintf): New function.
* infrun.c (infrun_debug_printf_1): Use debug_prefixed_vprintf.
* linux-nat.c (linux_nat_debug_printf_1): Likewise.
Change-Id: Iccc290a2dc6b5fffcbe1c2866ed8d804ad380764
Introduce this macro to print debug statements in the infrun.c file,
same idea as what was done in 9327494e0e ("gdb: add
linux_nat_debug_printf macro").
Although in this case, there are places outside infrun.c that print
debug statements if debug_infrun is set. So the macro has to be
declared in the header file, so that it can be used in these other
files.
Note one special case. In stop_all_threads, I've used an explicit
if (debug_infrun)
infrun_debug_printf_1 ("stop_all_threads", "done");
for the message in the SCOPE_EXIT. Otherwise, the message appears like
this:
[infrun] operator(): done
Until we find a better solution for extracting a meaningful function
name for lambda functions, I think it's fine to handle these special
cases manually, they are quite rare.
Some tests need to be updated, because they rely on some infrun debug
statements.
gdb/ChangeLog:
* infrun.h (infrun_debug_printf_1): New function declaration.
(infrun_debug_printf): New macro.
* infrun.c (infrun_debug_printf_1): Use infrun_debug_printf
throughout.
(infrun_debug_printf): New function.
* breakpoint.c (should_be_inserted): Use infrun_debug_printf.
(handle_jit_event): Likewise.
gdb/testsuite/ChangeLog:
* gdb.base/gdb-sigterm.exp (do_test): Update expected regexp.
* gdb.threads/signal-while-stepping-over-bp-other-thread.exp:
Likewise.
* gdb.threads/stepi-random-signal.exp: Likewise.
Change-Id: I66433c8a9caa64c8525ab57c593022b9d1956d5c
I noticed what I think is a potential bug. I did not observe it nor was
I able to reproduce it using actual debugging. It's quite unlikely,
because it involves multi-target and ptid clashes. I added selftests
that demonstrate it though.
The thread_ptid_changed observer says that thread with OLD_PTID now has
NEW_PTID. Now, if for some reason we happen to have two targets
defining a thread with OLD_PTID, the observers don't know which thread
this is about.
regcache::regcache_thread_ptid_changed changes all regcaches with
OLD_PTID. If there is a regcache for a thread with ptid OLD_PTID, but
that belongs to a different target, this regcache will be erroneously
changed.
Similarly, infrun_thread_ptid_changed updates inferior_ptid if
inferior_ptid matches OLD_PTID. But if inferior_ptid currently refers
not to the thread is being changed, but to a thread with the same ptid
belonging to a different target, then inferior_ptid will erroneously be
changed.
This patch adds a `process_stratum_target *` parameter to the
`thread_ptid_changed` observable and makes the two observers use it.
Tests for both are added, which would fail if the corresponding fix
wasn't done.
gdb/ChangeLog:
* observable.h (thread_ptid_changed): Add parameter
`process_stratum_target *`.
* infrun.c (infrun_thread_ptid_changed): Add parameter
`process_stratum_target *` and use it.
(selftests): New namespace.
(infrun_thread_ptid_changed): New function.
(_initialize_infrun): Register selftest.
* regcache.c (regcache_thread_ptid_changed): Add parameter
`process_stratum_target *` and use it.
(regcache_thread_ptid_changed): New function.
(_initialize_regcache): Register selftest.
* thread.c (thread_change_ptid): Pass target to
thread_ptid_changed observable.
Change-Id: I0599e61224b6d154a7b55088a894cb88298c3c71
This is a more general version of the existing handle_segmentation_fault
hook that is able to report information for an arbitrary signal, not
just SIGSEGV.
gdb/ChangeLog:
* gdbarch.c: Regenerate.
* gdbarch.h: Regenerate.
* gdbarch.sh (report_signal_info): New method.
* infrun.c (print_signal_received_reason): Invoke gdbarch
report_signal_info hook if present.
When interrupting a program in non-stop, the program gets interrupted
correctly, but GDB busy loops (the event loop is always woken up).
Here is how to reproduce it:
1. Start GDB: ./gdb -nx --data-directory=data-directory -ex "set non-stop 1" --args /bin/sleep 60
2. Run the program with "run"
3. Interrupt with ^C.
4. Look into htop, see GDB taking 100% CPU
Debugging `handle_file_event`, we see that the event source that wakes
up the event loop is the linux-nat one:
(top-gdb) p file_ptr.proc
$5 = (handler_func *) 0xb9cccd <handle_target_event(int, gdb_client_data)>
^^^^^^^^^^^^^^^^^^^
|
\-- the linux-nat callback
Debugging fetch_inferior_event and do_target_wait, we see that we
don't actually call `wait` on the linux-nat target, because
inferior_matches returns false:
auto inferior_matches = [&wait_ptid] (inferior *inf)
{
return (inf->process_target () != NULL
&& (threads_are_executing (inf->process_target ())
|| threads_are_resumed_pending_p (inf))
&& ptid_t (inf->pid).matches (wait_ptid));
};
because `threads_are_executing` is false.
What happens is:
1. User types ctrl-c, that writes in the linux-nat pipe, waking up
the event source.
2. linux-nat's wait gets called, the SIGINT event is returned, but
before returning, it marks the pipe again, in order for wait to
get called again:
/* If we requested any event, and something came out, assume there
may be more. If we requested a specific lwp or process, also
assume there may be more. */
if (target_is_async_p ()
&& ((ourstatus->kind != TARGET_WAITKIND_IGNORE
&& ourstatus->kind != TARGET_WAITKIND_NO_RESUMED)
|| ptid != minus_one_ptid))
async_file_mark ();
3. The SIGINT event is handled, the program is stopped, the stop
notification is printed.
4. The event loop is woken up again because of the `async_file_mark`
of step 2.
5. Because `inferior_matches` returns false, we never call
linux-nat's wait, so the pipe stays readable.
6. Goto 4.
Pedro says:
This commit fixes it by letting do_target_wait call target_wait even
if threads_are_executing is false. This will normally result in the
target returning TARGET_WAITKIND_NO_RESUMED, and _not_ marking its
event source again. This results in infrun only calling into the
target only once (i.e., breaking the busy loop).
Note that the busy loop bug didn't trigger in all-stop mode because
all-stop handles this by unregistering the target from the event loop
as soon as it was all stopped -- see
inf-loop.c:inferior_event_handler's INF_EXEC_COMPLETE handling. If we
remove that non-stop check from inferior_event_handler, and replace
the target_has_execution check for threads_are_executing instead, it
also fixes the issue for non-stop. I considered that as the final
solution, but decided that the solution proposed here instead is just
simpler and more future-proof design. With the
TARGET_WAITKIND_NO_RESUMED handling fixes done in the previous
patches, I think it should be possible to always keep the target
registered in the event loop, meaning we could eliminate the
target_async(0) call from inferior_event_handler as well as most of
the target_async(1) calls in the target backends. That would allow in
the future e.g., the remote target reporting asynchronous
notifications even if all threads are stopped. I haven't attempted
that, though.
gdb/ChangeLog:
yyyy-mm-dd Simon Marchi <simon.marchi@polymtl.ca>
Pedro Alves <pedro@palves.net>
PR gdb/26199
* infrun.c (threads_are_resumed_pending_p): Delete.
(do_target_wait): Remove threads_are_executing and
threads_are_resumed_pending_p checks from the inferior_matches
lambda. Update comments.
Let's consider the same use case as in the previous commit:
Say you have two inferiors 1 and 2, each connected to a different
target, A and B.
Now say you set inferior 2 running, with "continue &".
Now you select a thread of inferior 1, say thread 1.2, and continue in
the foreground. All other threads of inferior 1 are left stopped.
Thread 1.2 exits, and thus target A has no other resumed thread, so it
reports TARGET_WAITKIND_NO_RESUMED.
At this point, because the threads of inferior 2 are still executing
the TARGET_WAITKIND_NO_RESUMED event is ignored.
Now, the user types Ctrl-C. Because GDB had previously put inferior 1
in the foreground, the kernel sends the SIGINT to that inferior.
However, no thread in that inferior is executing right now, so ptrace
never intercepts the SIGINT -- it is never dequeued by any thread.
The result is that GDB's CLI is stuck. There's no way to get back the
prompt (unless inferior 2 happens to report some event).
The fix in this commit is to make handle_no_resumed give the terminal
to some other inferior that still has threads executing so that a
subsequent Ctrl-C reaches that target first (and then GDB intercepts
the SIGINT). This is a bit hacky, but seems like the best we can do
with the current design.
I think that putting all native inferiors in their own session would
help fixing this in a clean way, since with that a Ctrl-C on GDB's
terminal will _always_ reach GDB first, and then GDB can decide how to
pause the inferior. But that's a much larger change.
The testcase added by the following patch needs this fix.
gdb/ChangeLog:
PR gdb/26199
* infrun.c (handle_no_resumed): Transfer terminal to inferior with
executing threads.
handle_no_resumed is currently not considering multiple targets.
Say you have two inferiors 1 and 2, each connected to a different
target, A and B.
Now say you set inferior 2 running, with "continue &".
Now you select a thread of inferior 1, say thread 1.2, and continue in
the foreground. All other threads of inferior 1 are left stopped.
Thread 1.2 exits, and thus target A has no other resumed thread, so it
reports TARGET_WAITKIND_NO_RESUMED.
At this point, if both inferiors were running in the same target,
handle_no_resumed would realize that threads of inferior 2 are still
executing, so the TARGET_WAITKIND_NO_RESUMED event should be ignored.
But because handle_no_resumed only walks the threads of the current
target, it misses noticing that threads of inferior 2 are still
executing. The fix is just to walk over all threads of all targets.
A testcase covering the use case above will be added in a following
patch. It can't be added yet because it depends on yet another fix to
handle_no_resumed not included here.
gdb/ChangeLog:
PR gdb/26199
* infrun.c (handle_no_resumed): Handle multiple targets.
If we hit the synchronous execution command case described by
handle_no_resumed, and handle_no_resumed determines that the event
should be ignored, because it found a thread that is executing, we end
up in prepare_to_wait.
There, if the current target is not registered in the event loop right
now, we call mark_infrun_async_event_handler. With that event handler
marked, the event loop calls again into fetch_inferior_event, which
calls target_wait, which returns TARGET_WAITKIND_NO_RESUMED, and we
end up in handle_no_resumed, again ignoring the event and marking
infrun_async_event_handler. The result is that GDB is now always
keeping the CPU 100% busy in this loop, even though it continues to be
able to react to input and to real target events, because we still go
through the event-loop.
The problem is that marking of the infrun_async_event_handler in
prepare_to_wait. That is there to handle targets that don't support
asynchronous execution. So the correct predicate is whether async
execution is supported, not whether the target is async right now.
gdb/ChangeLog:
PR gdb/26199
* infrun.c (prepare_to_wait): Check target_can_async_p instead of
target_is_async_p.
I noticed that fetch_inferior_event receives the client_data parameter
from its caller, inferior_event_handler, but doesn't actually need it.
This patch removes it. In turn, inferior_event_handler doesn't use its
parameter, so remove it too.
The `data` argument used when registering
remote_async_inferior_event_handler is changed to NULL, to avoid
confusion. It could make people think that the value passed is used
somewhere, when in fact it's not.
gdb/ChangeLog:
* inf-loop.c (inferior_event_handler): Remove client_data param.
* inf-loop.h (inferior_event_handler): Likewise.
* infcmd.c (step_1): Adjust.
* infrun.c (proceed): Adjust.
(fetch_inferior_event): Remove client_data param.
(infrun_async_inferior_event_handler): Adjust.
* infrun.h (fetch_inferior_event): Remove `void *` param.
* linux-nat.c (handle_target_event): Adjust.
* record-btrace.c (record_btrace_handle_async_inferior_event):
Adjust.
* record-full.c (record_full_async_inferior_event_handler):
Adjust.
* remote.c (remote_async_inferior_event_handler): Adjust.
Change-Id: I3c2aa1eb0ea3e0985df096660d2dcd794674f2ea
gdb/ChangeLog:
2020-06-18 Pedro Alves <palves@redhat.com>
* infrun.c (generic_mourn_inferior): Use switch_to_thread instead
of writing to inferior_ptid.
(scoped_restore_exited_inferior): Delete.
(handle_vfork_child_exec_or_exit): Simplify using
scoped_restore_current_pspace_and_thread. Use switch_to_thread
instead of writing to inferior_ptid.
(THREAD_STOPPED_BY): Delete.
(thread_stopped_by_watchpoint, thread_stopped_by_sw_breakpoint)
(thread_stopped_by_hw_breakpoint): Delete.
(save_waitstatus): Use
scoped_restore_current_thread+switch_to_thread, and call
target_stopped_by_watchpoint instead of
thread_stopped_by_watchpoint, target_stopped_by_sw_breakpoint
instead of thread_stopped_by_sw_breakpoint, and
target_stopped_by_hw_breakpoint instead of
thread_stopped_by_hw_breakpoint.
(handle_inferior_event)
<TARGET_WAITKIND_EXITED/TARGET_WAITKIND_SIGNALLED>: Don't write to
inferior_ptid directly, nor
set_current_inferior/set_current_program_space. Use
switch_to_thread / switch_to_inferior_no_thread instead.
Continuing my goal of removing the "ALL_*" iterator macros, this
removes ALL_UIS, replacing it with an iterator adaptor.
gdb/ChangeLog
2020-05-16 Tom Tromey <tom@tromey.com>
* top.c (quit_force): Update.
* infrun.c (handle_no_resumed): Update.
* top.h (all_uis): New function.
(ALL_UIS): Remove.
[Simon: I send this patch on behalf of Laurent Morichetti, I added the
commit message and performance measurement stuff.
Also, this patch is better viewed with "git show -w".]
stop_all_threads, in infrun.c, is used to stop all running threads on
targets that are always non-stop. It's used, for example, when the
program hits a breakpoint while GDB is set to "non-stop off". It sends
a stop request for each running thread, then collects one wait event for
each.
Since new threads can spawn while we are stopping the threads, it's
written in a way where it makes multiple such "send stop requests to
running threads & collect wait events" passes. The function completes
when it has made two passes where it hasn't seen any running threads.
With the way it's written right now is, it iterates on the thread list,
sending a stop request for each running thread. It then waits for a
single event, after which it iterates through the thread list again. It
sends stop requests for any running threads that's been created since
the last iteration. It then consumes another single wait event.
This makes it so we iterate on O(n^2) threads in total, where n is the
number of threads. This patch changes the function to reduce it to
O(n). This starts to have an impact when dealing with multiple
thousands of threads (see numbers below). At each pass, we know the
number of outstanding stop requests we have sent, for which we need to
collect a stop event. We can therefore loop to collect this many stop
events before proceeding to the next pass and iterate on the thread list
again.
To check the performance improvements with this patch, I made an
x86/Linux program with a large number of idle threads (varying from 1000
to 10000). The program's main thread hits a breakpoint once all these
threads have started, which causes stop_all_threads to be called to stop
all these threads. I measured (by patching stop_all_threads):
- the execution time of stop_all_threads
- the total number of threads we iterate on during the complete
execution of the function (the total number of times we execute the
"for (thread_info *t : all_non_exited_threads ())" loop)
These are the execution times, in milliseconds:
# threads before after
1000 226 106
2000 997 919
3000 3461 2323
4000 4330 3570
5000 8642 6600
6000 9918 8039
7000 12662 10930
8000 16652 11222
9000 21561 15875
10000 26613 20019
Note that I very unscientifically executed each case only once.
These are the number of loop executions:
# threads before after
1000 1003002 3003
2000 4006002 6003
3000 9009002 9003
4000 16012002 12003
5000 25015002 15003
6000 36018002 18003
7000 49021002 21003
8000 64024002 24003
9000 81027002 27003
10000 100030002 30003
This last table shows pretty well the O(n^2) vs O(n) behaviors.
Reg-tested on x86 GNU/Linux (Ubuntu 16.04).
gdb/ChangeLog:
YYYY-MM-DD Laurent Morichetti <Laurent.Morichetti@amd.com>
YYYY-MM-DD Simon Marchi <simon.marchi@efficios.com>
* infrun.c (stop_all_threads): Collect multiple wait events at
each pass.
In stop_all_threads, GDB sends signals to other threads in an attempt
to stop them. While in a typical scenario the expected wait status is
TARGET_WAITKIND_STOPPED, it is possible that the thread GDB attempted
to stop has already terminated. If so, a waitstatus other than
TARGET_WAITKIND_STOPPED would be received. Handle this case
appropriately.
If a wait status that denotes thread termination is ignored, GDB goes
into an infinite loop in stop_all_threads.
E.g.:
$ gdb ./a.out
(gdb) start
...
(gdb) add-inferior -exec ./a.out
...
(gdb) inferior 2
...
(gdb) start
...
(gdb) set schedule-multiple on
(gdb) set debug infrun 2
(gdb) continue
Continuing.
infrun: clear_proceed_status_thread (process 10449)
infrun: clear_proceed_status_thread (process 10453)
infrun: proceed (addr=0xffffffffffffffff, signal=GDB_SIGNAL_DEFAULT)
infrun: proceed: resuming process 10449
infrun: resume (step=0, signal=GDB_SIGNAL_0), trap_expected=0, current thread [process 10449] at 0x55555555514e
infrun: infrun_async(1)
infrun: prepare_to_wait
infrun: proceed: resuming process 10453
infrun: resume (step=0, signal=GDB_SIGNAL_0), trap_expected=0, current thread [process 10453] at 0x55555555514e
infrun: prepare_to_wait
infrun: Found 2 inferiors, starting at #0
infrun: target_wait (-1.0.0, status) =
infrun: 10449.10449.0 [process 10449],
infrun: status->kind = exited, status = 0
infrun: handle_inferior_event status->kind = exited, status = 0
[Inferior 1 (process 10449) exited normally]
infrun: stop_waiting
infrun: stop_all_threads
infrun: stop_all_threads, pass=0, iterations=0
infrun: process 10453 executing, need stop
infrun: target_wait (-1.0.0, status) =
infrun: 10453.10453.0 [process 10453],
infrun: status->kind = exited, status = 0
infrun: stop_all_threads status->kind = exited, status = 0 process 10453
infrun: process 10453 executing, already stopping
infrun: target_wait (-1.0.0, status) =
infrun: -1.0.0 [process -1],
infrun: status->kind = no-resumed
infrun: infrun_async(0)
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
infrun: stop_all_threads status->kind = no-resumed process -1
infrun: process 10453 executing, already stopping
...
And this polling goes on forever. This patch prevents the infinite
looping behavior. For the same scenario above, we obtain the
following behavior:
...
(gdb) continue
Continuing.
infrun: clear_proceed_status_thread (process 31229)
infrun: clear_proceed_status_thread (process 31233)
infrun: proceed (addr=0xffffffffffffffff, signal=GDB_SIGNAL_DEFAULT)
infrun: proceed: resuming process 31229
infrun: resume (step=0, signal=GDB_SIGNAL_0), trap_expected=0, current thread [process 31229] at 0x55555555514e
infrun: infrun_async(1)
infrun: prepare_to_wait
infrun: proceed: resuming process 31233
infrun: resume (step=0, signal=GDB_SIGNAL_0), trap_expected=0, current thread [process 31233] at 0x55555555514e
infrun: prepare_to_wait
infrun: Found 2 inferiors, starting at #0
infrun: target_wait (-1.0.0, status) =
infrun: 31229.31229.0 [process 31229],
infrun: status->kind = exited, status = 0
infrun: handle_inferior_event status->kind = exited, status = 0
[Inferior 1 (process 31229) exited normally]
infrun: stop_waiting
infrun: stop_all_threads
infrun: stop_all_threads, pass=0, iterations=0
infrun: process 31233 executing, need stop
infrun: target_wait (-1.0.0, status) =
infrun: 31233.31233.0 [process 31233],
infrun: status->kind = exited, status = 0
infrun: stop_all_threads status->kind = exited, status = 0 process 31233
infrun: saving status status->kind = exited, status = 0 for 31233.31233.0
infrun: process 31233 not executing
infrun: stop_all_threads, pass=1, iterations=1
infrun: process 31233 not executing
infrun: stop_all_threads done
(gdb)
The exit event from Inferior 1 is received and shown to the user.
The exit event from Inferior 2 is not displayed, but kept pending.
(gdb) info inferiors
Num Description Connection Executable
* 1 <null> a.out
2 process 31233 1 (native) a.out
(gdb) inferior 2
[Switching to inferior 2 [process 31233] (a.out)]
[Switching to thread 2.1 (process 31233)]
Couldn't get registers: No such process.
(gdb) continue
Continuing.
infrun: clear_proceed_status_thread (process 31233)
infrun: clear_proceed_status_thread: thread process 31233 has pending wait status status->kind = exited, status = 0 (currently_stepping=0).
infrun: proceed (addr=0xffffffffffffffff, signal=GDB_SIGNAL_DEFAULT)
infrun: proceed: resuming process 31233
infrun: resume: thread process 31233 has pending wait status status->kind = exited, status = 0 (currently_stepping=0).
infrun: prepare_to_wait
infrun: Using pending wait status status->kind = exited, status = 0 for process 31233.
infrun: target_wait (-1.0.0, status) =
infrun: 31233.31233.0 [process 31233],
infrun: status->kind = exited, status = 0
infrun: handle_inferior_event status->kind = exited, status = 0
[Inferior 2 (process 31233) exited normally]
infrun: stop_waiting
(gdb) info inferiors
Num Description Connection Executable
1 <null> a.out
* 2 <null> a.out
(gdb)
When a process exits and we leave the process exit event pending, we
need to make sure that at least one thread is left listed in the
inferior's thread list. This is necessary in order to make sure we
have a thread that we can later resume, so the process exit event can
be collected/reported.
When native debugging, the GNU/Linux back end already makes sure that
the last LWP isn't deleted.
When remote debugging against GNU/Linux GDBserver, the GNU/Linux
GDBserver backend also makes sure that the last thread isn't deleted
until the process exit event is reported to GDBserver core.
However, between the backend reporting the process exit event to
GDBserver core, and GDB consuming the event, GDB may update the thread
list and find no thread left in the process. The process exit event
will be pending somewhere in GDBserver's stop reply queue, or
gdb/remote.c's queue, or whathever other event queue inbetween
GDBserver and infrun.c's handle_inferior_event.
This patch tweaks remote.c's target_update_thread_list implementation
to avoid deleting the last thread of an inferior.
In the past, this case of inferior-with-no-threads led to a special
case at the bottom of handle_no_resumed, where it reads:
/* Note however that we may find no resumed thread because the whole
process exited meanwhile (thus updating the thread list results
in an empty thread list). In this case we know we'll be getting
a process exit event shortly. */
for (inferior *inf : all_non_exited_inferiors (ecs->target))
In current master, that code path is still reachable with the
gdb.threads/continue-pending-after-query.exp testcase, when tested
against GDBserver, with "maint set target-non-stop" forced "on".
With this patch, the scenario that loop was concerned about is still
properly handled, because the loop above it finds the process's last
thread with "executing" set to true, and thus the handle_no_resumed
function still returns true.
Since GNU/Linux native and remote are the only targets that support
non-stop mode, and with this patch, we always make sure the inferior
has at least one thread, this patch also removes that "inferior with
no threads" special case handling from handle_no_resumed.
Since remote.c now has a special case where we treat a thread that has
already exited as if it was still alive, we might need to tweak
remote.c's target_thread_alive implementation to return true for that
thread without querying the remote side (which would say "no, not
alive"). After inspecting all the target_thread_alive calls in the
codebase, it seems that only the one from prune_threads could result
in that thread being accidentally deleted. There's only one call to
prune_threads in GDB's common code, so this patch handles this by
replacing the prune_threads call with a delete_exited_threads call.
This seems like an improvement anyway, because we'll still be doing
what the comment suggests we want to do, and, we avoid remote protocol
traffic.
Regression-tested on X86_64 Linux.
gdb/ChangeLog:
2020-05-14 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
Tom de Vries <tdevries@suse.de>
Pedro Alves <palves@redhat.com>
PR threads/25478
* infrun.c (stop_all_threads): Do NOT ignore
TARGET_WAITKIND_NO_RESUMED, TARGET_WAITKIND_THREAD_EXITED,
TARGET_WAITKIND_EXITED, TARGET_WAITKIND_SIGNALLED wait statuses
received.
(handle_no_resumed): Remove code handling a live inferior with no
threads.
* remote.c (has_single_non_exited_thread): New.
(remote_target::update_thread_list): Do not delete a thread if is
the last thread of the process.
* thread.c (thread_select): Call delete_exited_threads instead of
prune_threads.
gdb/testsuite/ChangeLog:
2020-05-14 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
Pedro Alves <palves@redhat.com>
* gdb.multi/multi-exit.c: New file.
* gdb.multi/multi-exit.exp: New file.
* gdb.multi/multi-kill.c: New file.
* gdb.multi/multi-kill.exp: New file.
In stop_all_threads, the thread events of the current top target are
enabled at the beginning of the function and then disabled at the end
(at scope exit time). Because there may be multiple targets whose
thread lists will be updated and whose threads are stopped,
enable/disable thread events for all targets.
This update caused a change in the annotations. In particular, a
"frames-invalid" annotation is printed one more time due to switching
the current inferior. Hence, gdb.base/annota1.exp and
gdb.cp/annota2.exp tests are also updated.
Regression-tested on X86_64 Linux using the default board file and the
native-extended-gdbserver board file.
gdb/ChangeLog:
2020-05-14 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* infrun.c (stop_all_threads): Enable/disable thread events of all
targets. Move a debug message denoting the end of the function
into the SCOPED_EXIT block.
gdb/testsuite/ChangeLog:
2020-05-14 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* gdb.base/annota1.exp: Update the expected output.
* gdb.cp/annota2.exp: Ditto.
This is a refactoring. The extracted function is placed deliberately
before 'stop_all_threads' because the function will be re-used there
in a subsequent patch for handling an exit status kind received from
a thread that GDB attempted to stop.
gdb/ChangeLog:
2020-05-14 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* infrun.c (handle_inferior_event): Extract out a piece of code
into...
(mark_non_executing_threads): ...this new function.
Change-Id: I2b088f4a724f4260cb37068264964525cf62a118
In infrun.c's resume_1 function, move the definition of the local
variable PC down to its first use. This is useful if the thread we want
to resume is already gone with a pending exit event, because we avoid
the error we would see otherwise when trying to read the PC.
gdb/ChangeLog:
2020-05-14 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* infrun.c (resume_1): Move a 'regcache_read_pc' call down to first
use.
It possible that a thread whose PC we attempt to read is already dead.
In this case, 'regcache_read_pc' errors out. This impacts the
"proceed" execution flow, where GDB quits early before having a chance
to check if there exists a pending event. To remedy, keep going with
a 0 value for the PC if 'regcache_read_pc' fails. Because the value
of PC before resuming a thread is mostly used for storing and checking
the next time the thread stops, this tolerance is expected to be
harmless for a dead thread/process.
gdb/ChangeLog:
2020-05-14 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* regcache.c (regcache_read_pc_protected): New function
implementation that returns 0 if the PC cannot read via
'regcache_read_pc'.
* infrun.c (proceed): Call 'regcache_read_pc_protected'
instead of 'regcache_read_pc'.
(keep_going_pass_signal): Ditto.
gdbsupport/ChangeLog:
2020-05-14 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* common-regcache.h (regcache_read_pc_protected): New function
declaration.
When running test-case gdb.threads/fork-child-threads.exp with gcc-8 instead
of gcc-7, we have:
...
(gdb) next^M
[Attaching after Thread 0x7ffff7fae740 (LWP 27574) fork to child process \
27578]^M
[New inferior 2 (process 27578)]^M
[Detaching after fork from parent process 27574]^M
[Inferior 1 (process 27574) detached]^M
[Thread debugging using libthread_db enabled]^M
Using host libthread_db library "/lib64/libthread_db.so.1".^M
[Switching to Thread 0x7ffff7fae740 (LWP 27578)]^M
-main () at src/gdb/testsuite/gdb.threads/fork-child-threads.c:41^M
+main () at src/gdb/testsuite/gdb.threads/fork-child-threads.c:34^M
-41 i = pthread_create (&thread, NULL, start, NULL);^M
+34 switch (fork ())^M
-(gdb) PASS: gdb.threads/fork-child-threads.exp: next over fork
+(gdb) FAIL: gdb.threads/fork-child-threads.exp: next over fork
...
This is due to the fact that gcc-8 generates more precise line info, making
the instruction after the call to fork a "recommended breakpoint location".
However, it is a bug because next is supposed to move to the next source
line.
The problem is that in process_event_stop_test we hit this code:
...
if ((ecs->event_thread->suspend.stop_pc == stop_pc_sal.pc)
&& (ecs->event_thread->current_line != stop_pc_sal.line
|| ecs->event_thread->current_symtab != stop_pc_sal.symtab))
{
if (stop_pc_sal.is_stmt)
{
/* We are at the start of a different line. So stop. Note that
we don't stop if we step into the middle of a different line.
That is said to make things like for (;;) statements work
better. */
if (debug_infrun)
fprintf_unfiltered (gdb_stdlog,
"infrun: stepped to a different line\n");
end_stepping_range (ecs);
return;
}
...
because current_line and current_symtab have initial values:
...
(gdb) p ecs->event_thread->current_line
$8 = 0
(gdb) p ecs->event_thread->current_symtab
$9 = (symtab *) 0x0
...
Fix this in follow_fork by copying current_line and current_symtab from
parent thread to child thread.
Tested on x86_64-linux, with gcc 7.5.0 and gcc 10.0.1.
gdb/ChangeLog:
2020-05-08 Tom de Vries <tdevries@suse.de>
* infrun.c (follow_fork): Copy current_line and current_symtab to
child thread.
In infrun.c's 'displaced_step_fixup', as part of the 'finish_step_over'
flow, switch to the eventing thread *before* calling
'displaced_step_restore', because down in the flow ptid-dependent
memory accesses are used via current_inferior() and current_top_target().
Without this patch, the problem is exposed with the scenario below:
$ gdb -q
(gdb) maint set target-non-stop on
(gdb) file a.out
Reading symbols from a.out...
(gdb) set remote exec-file a.out
(gdb) target extended-remote | gdbserver --once --multi -
...
(gdb) add-inferior
[New inferior 2]
Added inferior 2 on connection 1 (extended-remote ...)
(gdb) inferior 2
[Switching to inferior 2 [<null>] (<noexec>)]
(gdb) file a.out
Reading symbols from a.out...
(gdb) set remote exec-file a.out
(gdb) run
...
Cannot access memory at address 0x555555555042
(gdb)
The problem is, down inside 'displaced_step_restore', GDB wants to
access the memory for inferior 2 because of an internal breakpoint.
However, the current inferior and inferior_ptid are out of sync.
While inferior_ptid correctly points to the process of inf 2 that was
just started, current_inferior points to inf 1. Then, the attempt to
access the memory fails, because target_has_execution results in false
since inf 1 was not started. I was not able to simplify the failing
scenario, but it shows the problem.
After this patch, we get
... same steps above...
(gdb) run
...
[Inferior 2 (process 28652) exited normally]
(gdb)
Regression-tested on X86_64 Linux with `make check`s default board file
and also `--target_board=native-extended-gdbserver`. In fact, the bug
fixed by this patch was exposed when using the native-extended-gdbserver
board file.
gdb/ChangeLog:
2020-04-21 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* infrun.c (displaced_step_fixup): Switch to the event_thread
before calling displaced_step_restore, not after.
gdb/testsuite/ChangeLog:
2020-04-21 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* gdb.multi/run-only-second-inf.c: New file.
* gdb.multi/run-only-second-inf.exp: New file.
This patch splits out some gdb-specific code from event-loop, into new
files async-event.[ch]. Strictly speaking this code could perhaps be
put into gdbsupport/, but because gdbserver does not currently use it,
it seemed better, for size reasons, to split it out.
gdb/ChangeLog
2020-04-13 Tom Tromey <tom@tromey.com>
* tui/tui-win.c: Include async-event.h.
* remote.c: Include async-event.h.
* remote-notif.c: Include async-event.h.
* record-full.c: Include async-event.h.
* record-btrace.c: Include async-event.h.
* infrun.c: Include async-event.h.
* event-top.c: Include async-event.h.
* event-loop.h: Move some declarations to async-event.h.
* event-loop.c: Don't include ser-event.h or top.h. Move some
code to async-event.c.
* async-event.h: New file.
* async-event.c: New file.
* Makefile.in (COMMON_SFILES): Add async-event.c.
(HFILES_NO_SRCDIR): Add async-event.h.
This moves gdb_select.h to gdbsupport/, so it can be used by other
code there.
gdb/ChangeLog
2020-04-13 Tom Tromey <tom@tromey.com>
* gdb_select.h: Move to ../gdbsupport/.
* event-loop.c: Update include path.
* top.c: Update include path.
* ser-base.c: Update include path.
* ui-file.c: Update include path.
* ser-tcp.c: Update include path.
* guile/scm-ports.c: Update include path.
* posix-hdep.c: Update include path.
* ser-unix.c: Update include path.
* gdb_usleep.c: Update include path.
* mingw-hdep.c: Update include path.
* inflow.c: Update include path.
* infrun.c: Update include path.
* event-top.c: Update include path.
gdbsupport/ChangeLog
2020-04-13 Tom Tromey <tom@tromey.com>
* gdb_select.h: Move from ../gdb/.
Stop all threads not only if the current target is non-stop, but also
if there exists a non-stop target.
The multi-target patch (5b6d1e4fa4 "Multi-target support") made the
following change to gdb/inf-child.c:
void
inf_child_target::maybe_unpush_target ()
{
- if (!inf_child_explicitly_opened && !have_inferiors ())
+ if (!inf_child_explicitly_opened)
unpush_target (this);
}
If we are in all-stop mode with multiple inferiors, and an exit event
is received from an inferior, target_mourn_inferior() gets to this
point and without the have_inferiors() check, the target is unpushed.
This leads to having exec_ops as the top target.
Here is a test scenario. Two executables, ./a.out returns
immediately; ./sleepy just sleeps.
$ gdb ./sleepy
(gdb) start
...
(gdb) add-inferior -exec ./a.out
...
(gdb) inferior 2
[Switching to inferior 2..
(gdb) start
...
(gdb) set schedule-multiple on
(gdb) set debug infrun 1
(gdb) continue
At this point, the exit event is received from ./a.out. Normally,
this would lead to stop_all_threads() to also stop ./sleepy, but this
doesn't happen, because target_is_non_stop_p() returns false. And it
returns false because the top target is no longer the process target;
it is the exec_ops.
This patch modifies 'stop_waiting' to call 'stop_all_threads' if there
exists a non-stop target, not just when the current top target is
non-stop.
Tested on X86_64 Linux.
gdb/ChangeLog:
2020-04-01 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* infrun.c (stop_all_threads): Update assertion, plus when
stopping threads, take into account that we might be trying
to stop an all-stop target.
(stop_waiting): Call 'stop_all_threads' if there exists a
non-stop target.
gdb/testsuite/ChangeLog:
2020-04-01 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* gdb.multi/stop-all-on-exit.c: New test.
* gdb.multi/stop-all-on-exit.exp: New file.
This commit brings support for the DWARF line table is_stmt field to
GDB. The is_stmt field is used by the compiler when a single source
line is split into multiple assembler instructions, especially if the
assembler instructions are interleaved with instruction from other
source lines.
The compiler will set the is_stmt flag false from some instructions
from the source lines, these instructions are not a good place to
insert a breakpoint in order to stop at the source line.
Instructions which are marked with the is_stmt flag true are a good
place to insert a breakpoint for that source line.
Currently GDB ignores all instructions for which is_stmt is false.
This is fine in a lot of cases, however, there are some cases where
this means the debug experience is not as good as it could be.
Consider stopping at a random instruction, currently this instruction
will be attributed to the last line table entry before this point for
which is_stmt was true - as these are the only line table entries that
GDB tracks. This can easily be incorrect in code with even a low
level of optimisation.
With is_stmt tracking in place, when stopping at a random instruction
we now attribute the instruction back to the real source line, even
when is_stmt is false for that instruction in the line table.
When inserting breakpoints we still select line table entries for
which is_stmt is true, so the breakpoint placing behaviour should not
change.
When stepping though code (at the line level, not the instruction
level) we will still stop at instruction where is_stmt is true, I
think this is more likely to be the desired behaviour.
Instruction stepping is, of course, unchanged, stepping one
instruction at a time, but we should now report more accurate line
table information with each instruction step.
The original motivation for this work was a patch posted by Bernd
here:
https://sourceware.org/ml/gdb-patches/2019-11/msg00792.html
As part of that thread it was suggested that many issues would be
resolved if GDB supported line table views, this isn't something I've
attempted in this patch, though reading the spec, it seems like this
would be a useful feature to support in GDB in the future. The spec
is here:
http://dwarfstd.org/ShowIssue.php?issue=170427.1
And Bernd gives a brief description of the benefits here:
https://sourceware.org/ml/gdb-patches/2020-01/msg00147.html
With that all said, I think that there is benefit to having proper
is_stmt support regardless of whether we have views support, so I
think we should consider getting this in first, and then building view
support on top of this.
The gdb.cp/step-and-next-inline.exp test is based off a test proposed
by Bernd Edlinger in this message:
https://sourceware.org/ml/gdb-patches/2019-12/msg00842.html
gdb/ChangeLog:
* buildsym-legacy.c (record_line): Pass extra parameter to
record_line.
* buildsym.c (buildsym_compunit::record_line): Take an extra
parameter, reduce duplication in the line table, and record the
is_stmt flag in the line table.
* buildsym.h (buildsym_compunit::record_line): Add extra
parameter.
* disasm.c (do_mixed_source_and_assembly_deprecated): Ignore
non-statement lines.
* dwarf2/read.c (dwarf_record_line_1): Add extra parameter, pass
this to the symtab builder.
(dwarf_finish_line): Pass extra parameter to dwarf_record_line_1.
(lnp_state_machine::record_line): Pass a suitable is_stmt flag
through to dwarf_record_line_1.
* infrun.c (process_event_stop_test): When stepping, don't stop at
a non-statement instruction, and only refresh the step info when
we land in the middle of a line's range. Also add an extra
comment.
* jit.c (jit_symtab_line_mapping_add_impl): Initialise is_stmt
field.
* record-btrace.c (btrace_find_line_range): Only record lines
marked as is-statement.
* stack.c (frame_show_address): Show the frame address if we are
in a non-statement sal.
* symmisc.c (dump_symtab_1): Print the is_stmt flag.
(maintenance_print_one_line_table): Print a header for the is_stmt
column, and include is_stmt information in the output.
* symtab.c (find_pc_sect_line): Find lines marked as statements in
preference to non-statements.
(find_pcs_for_symtab_line): Prefer is-statement entries.
(find_line_common): Likewise.
* symtab.h (struct linetable_entry): Add is_stmt field.
(struct symtab_and_line): Likewise.
* xcoffread.c (arrange_linetable): Initialise is_stmt field when
arranging the line table.
gdb/testsuite/ChangeLog:
* gdb.cp/step-and-next-inline.cc: New file.
* gdb.cp/step-and-next-inline.exp: New file.
* gdb.cp/step-and-next-inline.h: New file.
* gdb.dwarf2/dw2-is-stmt.c: New file.
* gdb.dwarf2/dw2-is-stmt.exp: New file.
* gdb.dwarf2/dw2-is-stmt-2.c: New file.
* gdb.dwarf2/dw2-is-stmt-2.exp: New file.
* gdb.dwarf2/dw2-ranges-base.exp: Update line table pattern.
[ Migrating this from Gerrit: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/321 ]
I noticed that some functions in infcmd and infrun call each other and
all call inferior_thread, while they could just get the thread_info
pointer from their caller. That means less calls to inferior_thread, so
less reliance on global state, since inferior_thread reads
inferior_ptid.
The paths I am unsure about are:
- fetch_inferior_event calls...
- step_command_fsm::should_stop calls...
- prepare_one_step
and
- process_event_stop_test calls...
- set_step_info
Before this patch, prepare_one_step gets the thread pointer using
inferior_thread. After this patch, it gets it from the
execution_control_state structure in fetch_inferior_event. Are we sure
that the thread from the execution_control_state structure is the same
as the one inferior_thread would return? This code path is used when a
thread completes a step, but the user had specified a step count (e.g.
"step 5") so we decide to do one more step. It would be strange (and
even a bug I suppose) if the thread in the ecs structure in
fetch_inferior_event was not the same thread that is prepared to stepped
by prepare_one_step. So I believe passing the ecs thread is fine.
The same logic applies to process_event_stop_test calling
set_step_info.
gdb/ChangeLog:
* infrun.h: Forward-declare thread_info.
(set_step_info): Add thread_info parameter, add doc.
* infrun.c (set_step_info): Add thread_info parameter, move doc
to header.
* infrun.c (process_event_stop_test): Pass thread to
set_step_info call.
* infcmd.c (set_step_frame): Add thread_info pointer, pass it to
set_step_info.
(prepare_one_step): Add thread_info parameter, pass it to
set_step_frame and prepare_one_step (recursive) call.
(step_1): Pass thread to prepare_one_step call.
(step_command_fsm::should_stop): Pass thread to
prepare_one_step.
(until_next_fsm): Pass thread to set_step_frame call.
(finish_command): Pass thread to set_step_info call.
This function returns the result of a quite big condition. I think it
would be more readeable if it was broken up in smaller pieces and
commented. This is what this patch does.
I also introduced gdbarch_supports_displaced_stepping, since it shows
the intent better than checking for gdbarch_displaced_step_copy_insn_p.
I also used that new function in displaced_step_prepare_throw.
I also updated the comment on top of can_use_displaced_stepping, which
seemed a bit outdated with respect to non-stop. The comment likely
dates from before it was possible to have targets that always operate
non-stop under the hood, even when the user-visible mode is all-stop.
No functional changes intended.
gdb/ChangeLog:
* infrun.c (gdbarch_supports_displaced_stepping): New.
(use_displaced_stepping): Break up conditions in smaller pieces.
Use gdbarch_supports_displaced_stepping.
(displaced_step_prepare_throw): Use
gdbarch_supports_displaced_stepping.
With this commit:
commit 5b6d1e4fa4
Date: Fri Jan 10 20:06:08 2020 +0000
Multi-target support
There was a regression in GDB's support for older aspects of the
remote protocol. Specifically, when a target sends the 'S' stop reply
packet (which doesn't include a thread-id) then GDB has to figure out
which thread actually stopped.
Before the above commit GDB figured this out by using inferior_ptid in
process_stop_reply, which contained the ptid of the current
process/thread. This would be fine for single threaded
targets (which is the only place using an S packet makes sense), but
in the general case, relying on inferior_ptid for processing a stop is
wrong - there's no reason to believe that what was GDB's current
thread will be the same thread that just stopped in the target.
With the above commit the inferior_ptid now has the value null_ptid
inside process_stop_reply, this can be seen in do_target_wait, where
we call switch_to_inferior_no_thread before calling do_target_wait_1.
The problem this causes can be seen in the new test that runs
gdbserver using the flag --disable-packet=T, and causes GDB to throw
this assertion:
inferior.c:279: internal-error: inferior* find_inferior_pid(process_stratum_target*, int): Assertion `pid != 0' failed.
A similar problem was fixed in this commit:
commit 3cada74087
Date: Thu Jan 11 00:23:04 2018 +0000
Fix backwards compatibility with old GDBservers (PR remote/22597)
However, this commit deals with the case where the T packet doesn't
include a thread-id, not the S packet case. This commit solves the
problem providing a thread-id at the GDB side if the remote target
doesn't provide one. The thread-id provided comes from
remote_state::general_thread, however, though this does work, I don't
think it is the ideal solution.
The remote_state tracks two threads, the continue_thread and the
general_thread, these are updated when GDB asks the remote target to
switch threads. The general_thread is set before performing things
like register or memory accesses, and the continue_thread is set
before things like continue or step commands. Further, the
general_thread is updated after a target stops to reference the thread
that stopped.
The first thing to note from the above description is that we have a
cycle of dependency, when a T packet arrives without a thread-id we
fill in the thread-id from the general_thread data. The thread-id
from the stop event is then used to set the general_thread. This in
itself feels a little weird.
The second question is why use the general_thread at all? You'd think
given how they are originally set that the continue thread would be a
better choice. The problem with this is that the continue_thread, if
the user just does "continue", will be set to the minus_one_ptid, in
the remote protocol this means all threads. When the stop arrives
with no thread-id and we use continue_thread we end up with a very
similar assertion to before because we now end up trying to lookup a
thread using the minus_one_ptid. By contrast, once GDB has connected
to a remote target the general_thread will be set to a valid
thread-id, after which, if the target is single threaded, and stop
events arrive without a thread-id, everything works fine.
There is one slight weirdness with the above behaviour though. When
GDB first connects to the remote target inferior_ptid is null_ptid,
however, upon connecting we query the remote for its threads. As the
thread information arrives GDB adds the threads to its internal
database, and this process involves setting inferior_ptid to the id of
each new thread in turn. Once we know about all the threads we wait
for a stop event from the remote target to indicate that GDB is now in
control of the target.
The problem is that after adding the new threads we don't reset
inferior_ptid, and the code path we use to wait for a stop event from
the target also doesn't reset inferior_ptid, so it turns out that
during the initial connection inferior_ptid is not null_ptid. This is
lucky, because during the initial connection the general_thread
variable _is_ set to null_ptid.
So, during the initial connection, if the first stop event is missing
a thread-id then we "provide" a thead-id from general_thread. This
turns out to be null_ptid meaning no thread-id is known, and then
during process_stop_reply we fill in the missing thread-id using
inferior_ptid.
This was all discussed on the mailing list here:
https://sourceware.org/ml/gdb-patches/2020-02/msg01011.html
My proposal for a fix then is:
1. Move the call to switch_to_inferior_no_thread into
do_target_wait_1, this means that in all cases where we are waiting
for an inferior the inferior_ptid will be set to null_ptid. This is
good as no wait code should rely on inferior_ptid.
2. Remove the use of general_thread from the 'T' packet processing.
The general_thread read here was only ever correct by chance, and we
shouldn't be using it this way.
3. Remove use of inferior_ptid from process_stop_event as this is
wrong, and will always be null_ptid now anyway.
4. When a stop_event has null_ptid due to a lack of thread-id (either
from a T packet or an S packet) then pick the first non exited thread
in the target and use that. This will be fine for single threaded
targets. A multi-thread or multi-inferior aware remote target
should be using T packets with a thread-id, so we give a warning if
the target is multi-threaded, and we are still missing a thread-id.
5. Extend the existing test that covered the T packet with missing
thread-id to also cover the S packet.
gdb/ChangeLog:
* remote.c (remote_target::remote_parse_stop_reply): Don't use the
general_thread if the stop reply is missing a thread-id.
(remote_target::process_stop_reply): Use the first non-exited
thread if the target didn't pass a thread-id.
* infrun.c (do_target_wait): Move call to
switch_to_inferior_no_thread to ....
(do_target_wait_1): ... here.
gdb/testsuite/ChangeLog:
* gdb.server/stop-reply-no-thread.exp: Add test where T packet is
disabled.
This callback dynamically allocates a specialized displaced_step_closure, and
gives the ownership of the object to its caller. So I think it would make
sense for the callback to return an std::unique_ptr, this is what this patch
implements.
gdb/ChangeLog:
* gdbarch.sh (displaced_step_copy_insn): Change return type to an
std::unique_ptr.
* gdbarch.c: Re-generate.
* gdbarch.h: Re-generate.
* infrun.c (displaced_step_prepare_throw): Adjust to std::unique_ptr
change.
* aarch64-tdep.c (aarch64_displaced_step_copy_insn): Change return
type to std::unique_ptr.
* aarch64-tdep.h (aarch64_displaced_step_copy_insn): Likewise.
* amd64-tdep.c (amd64_displaced_step_copy_insn): Likewise.
* amd64-tdep.h (amd64_displaced_step_copy_insn): Likewise.
* arm-linux-tdep.c (arm_linux_displaced_step_copy_insn): Likewise.
* i386-linux-tdep.c (i386_linux_displaced_step_copy_insn): Likewise.
* i386-tdep.c (i386_displaced_step_copy_insn): Likewise.
* i386-tdep.h (i386_displaced_step_copy_insn): Likewise.
* rs6000-tdep.c (ppc_displaced_step_copy_insn): Likewise.
* s390-tdep.c (s390_displaced_step_copy_insn): Likewise.
displaced_step_inferior_state::reset and displaced_step_clear appear to
have the same goal, but they don't do the same thing.
displaced_step_inferior_state::reset clears more things than
displaced_step_clear, but it misses free'ing the closure, which
displaced_step_clear does.
This patch replaces displaced_step_clear's implementation with just a call to
displaced_step_inferior_state::reset. It then changes
displaced_step_inferior_state::step_closure to be a unique_ptr, to indicate the
fact that displaced_step_inferior_state owns the closure (and so that it is
automatically freed when the field is reset).
The test gdb.base/step-over-syscall.exp caught a problem when doing this, which
I consider to be a latent bug which my cleanup exposes. In
handle_inferior_event, in the TARGET_WAITKIND_FORKED case, if we displaced-step
over a fork syscall, we make sure to restore the memory that we used as a
displaced-stepping buffer in the child. We do so using the
displaced_step_inferior_state of the parent. However, we do it after calling
displaced_step_fixup for the parent, which clears the information in the
parent's displaced_step_inferior_state. It worked fine before, because
displaced_step_clear didn't completely clear the displaced_step_inferior_state
structure, so the required information (in this case the gdbarch) was
still available after clearing.
I fixed it by making GDB restore the child's memory before calling the
displaced_step_fixup on the parent. This way, the data in the
displaced_step_inferior_state structure is still valid when we use it for the
child. This is the error you would get in
gdb.base/step-over-syscall.exp without this fix:
/home/smarchi/src/binutils-gdb/gdb/gdbarch.c:3911: internal-error: ULONGEST gdbarch_max_insn_length(gdbarch*): Assertion `gdbarch != NULL' failed.
gdb/ChangeLog:
* infrun.c (get_displaced_step_closure_by_addr): Adjust to
std::unique_ptr.
(displaced_step_clear): Rename to...
(displaced_step_reset): ... this. Just call displaced->reset ().
(displaced_step_clear_cleanup): Rename to...
(displaced_step_reset_cleanup): ... this.
(displaced_step_prepare_throw): Adjust to std::unique_ptr.
(displaced_step_fixup): Likewise.
(resume_1): Likewise.
(handle_inferior_event): Restore child's memory before calling
displaced_step_fixup on the parent.
* infrun.h (displaced_step_inferior_state) <reset>: Adjust
to std::unique_ptr.
<step_closure>: Change type to std::unique_ptr.
