Alan Modra [Sun, 12 Jan 2020 23:40:41 +0000 (10:10 +1030)]
ubsan: alpha-vma: timeout
* vms-alpha.c (_bfd_vms_slurp_egsd): Ensure minimum size even
for "ignored" records.
Alan Modra [Sun, 12 Jan 2020 21:42:18 +0000 (08:12 +1030)]
Memory leaks and ineffective bounds checking in wasm_scan
It's always a bad idea to perform arithmetic on an unknown value read
from an object file before comparing against bounds. Code like the
following attempting to bounds check "len", a 64-bit value, isn't
effective because the pointer arithmetic ignores the high 32 bits when
compiled for a 32-bit host.
READ_LEB128 (len, p, end);
if (p + len < p || p + len > end)
goto error_return;
Instead, perform any arithmetic on known values where we don't need to
worry about overflows:
READ_LEB128 (len, p, end);
if (len > (size_t) (end - p))
goto error_return;
I'll note that this check does do things the right way:
READ_LEB128 (symcount, p, end);
/* Sanity check: each symbol has at least two bytes. */
if (symcount > payload_size / 2)
return FALSE;
"symcount * 2 > payload_size" would be wrong since the multiply could
overflow.
* wasm-module.c (wasm_scan_name_function_section): Formatting.
Delete asect name check. Move asect NULL check to wasm_object_p.
Correct bounds check of sizes against end. Replace uses of
bfd_zalloc with bfd_alloc, zeroing only necessary bytes. Use
just one bfd_release.
(wasm_scan): Don't use malloc/strdup for section names,
bfd_alloc instead. Simplify code prefixing section name.
Formatting. Don't attempt to free memory here..
(wasm_object_p): ..do so here. Formatting.
Alan Modra [Sun, 12 Jan 2020 09:46:22 +0000 (20:16 +1030)]
tic4x: sign extension using shifts
Don't do that. Especially don't use shift counts that assume the type
being shifted is 32 bits when the type is long/unsigned long. Also
reverts part of a change I made on 2019-12-11 to tic4x_print_register
that on closer inspection turns out to be unnecessary.
include/
* opcode/tic4x.h (EXTR): Delete.
(EXTRU, EXTRS, INSERTU, INSERTS): Rewrite without zero/sign
extension using shifts. Do trim INSERTU value to specified bitfield.
opcodes/
* tic4x-dis.c (tic4x_print_register): Remove dead code.
gas/
* config/tc-tic4x.c (tic4x_operands_match): Correct tic3x trap
insertion.
Alan Modra [Sat, 11 Jan 2020 02:02:11 +0000 (12:32 +1030)]
ubsan: fr30: left shift of negative value
cpu/
* fr30.cpu (f-disp9, f-disp10, f-s10, f-rel9, f-rel12): Don't
left shift signed values.
opcodes/
* fr30-ibld.c: Regenerate.
Alan Modra [Sat, 11 Jan 2020 01:53:47 +0000 (12:23 +1030)]
ubsan: xgate: left shift of negative value
* xgate-dis.c (print_insn): Don't left shift signed value.
(ripBits): Formatting, use 1u.
GDB Administrator [Mon, 13 Jan 2020 00:00:20 +0000 (00:00 +0000)]
Automatic date update in version.in
Simon Marchi [Sun, 12 Jan 2020 20:11:07 +0000 (15:11 -0500)]
gdb: include gdb_wait.h in gdb_wait.c
When building for mingw with -Wmissing-declarations, I get:
CXX gdbsupport/gdb_wait.o
/binutils-gdb/gdb/gdbsupport/gdb_wait.c:52:1: error: no previous declaration for 'int windows_status_to_termsig(long unsigned int)' [-Wer
ror=missing-declarations]
52 | windows_status_to_termsig (unsigned long status)
| ^~~~~~~~~~~~~~~~~~~~~~~~~
Make gdb_wait.c include gdb_wait.h to fix it.
gdb/ChangeLog:
* gdbsupport/gdb_wait.c: Include gdb_wait.h.
Simon Marchi [Sun, 12 Jan 2020 17:07:30 +0000 (12:07 -0500)]
gdbserver: include linux-arm-tdesc.h in linux-arm-tdesc.c
When building with -Wmissing-declarations, I get:
CXX linux-arm-tdesc.o
/binutils-gdb/gdb/gdbserver/linux-arm-tdesc.c:29:1: error: no previous declaration for 'const target_desc* arm_linux_read_description(arm_fp_type)' [-Werror=missing-declarations]
29 | arm_linux_read_description (arm_fp_type fp_type)
| ^~~~~~~~~~~~~~~~~~~~~~~~~~
/binutils-gdb/gdb/gdbserver/linux-arm-tdesc.c:49:1: error: no previous declaration for 'arm_fp_type arm_linux_get_tdesc_fp_type(const target_desc*)' [-Werror=missing-declarations]
49 | arm_linux_get_tdesc_fp_type (const target_desc *tdesc)
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~
Include linux-arm-tdesc.h in linux-arm-tdesc.c to fix it. And because
linux-arm-tdesc.h uses the arm_fp_type, it should include arch/arm.h as
well.
gdb/gdbserver/ChangeLog:
* linux-arm-tdesc.c: Include linux-arm-tdesc.h.
* linux-arm-tdesc.h: Include arch/arm.h.
Simon Marchi [Sun, 12 Jan 2020 16:06:23 +0000 (11:06 -0500)]
gdbserver: make aarch64_write_goto_address static
This function is only used in this file, so make it static. It fixes
this error, when building with -Wmissing-declarations:
CXX linux-aarch64-low.o
/home/simark/src/binutils-gdb/gdb/gdbserver/linux-aarch64-low.c:2642:1: error: no previous declaration for 'void aarch64_write_goto_address(CORE_ADDR, CORE_ADDR, int)' [-Werror=missing-declarations]
aarch64_write_goto_address (CORE_ADDR from, CORE_ADDR to, int size)
^~~~~~~~~~~~~~~~~~~~~~~~~~
gdb/gdbserver/ChangeLog:
* linux-aarch64-low.c (aarch64_write_goto_address): Make static.
Simon Marchi [Sun, 12 Jan 2020 16:06:21 +0000 (11:06 -0500)]
gdbserver: include aarch32/aarch64 header file in corresponding source file
When building gdbserver for an aarch64 host with -Wmissing-declarations,
I see:
CXX linux-aarch32-tdesc.o
/home/simark/src/binutils-gdb/gdb/gdbserver/linux-aarch32-tdesc.c:28:1: error: no previous declaration for 'const target_desc* aarch32_linux_read_description()' [-Werror=missing-declarations]
aarch32_linux_read_description ()
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/simark/src/binutils-gdb/gdb/gdbserver/linux-aarch32-tdesc.c:43:1: error: no previous declaration for 'bool is_aarch32_linux_description(const target_desc*)' [-Werror=missing-declarations]
is_aarch32_linux_description (const target_desc *tdesc)
^~~~~~~~~~~~~~~~~~~~~~~~~~~~
CXX linux-aarch64-tdesc.o
/home/simark/src/binutils-gdb/gdb/gdbserver/linux-aarch64-tdesc.c:32:1: error: no previous declaration for 'const target_desc* aarch64_linux_read_description(uint64_t, bool)' [-Werror=missing-declarations]
aarch64_linux_read_description (uint64_t vq, bool pauth_p)
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Fix it by including linux-aarch32-tdesc.h in linux-aarch32-tdesc.c and
linux-aarch64-tdesc.h in linux-aarch64-tdesc.c.
gdb/gdbserver/ChangeLog:
* linux-aarch32-tdesc.c: Include linux-aarch32-tdesc.h.
* linux-aarch64-tdesc.c: Include linux-aarch64-tdesc.h.
Pedro Alves [Sun, 12 Jan 2020 00:40:02 +0000 (00:40 +0000)]
Remove last traces of discard_all_inferiors
The multi-target patch should have removed all traces of
discard_all_inferiors, but somehow one use stayed behind along with
the definition of the function.
discard_all_inferiors is bad now because it blindly exits inferiors of
all target connections. It's best to remove it.
gdb/ChangeLog:
2020-01-12 Pedro Alves <palves@redhat.com>
* bsd-kvm.c (bsd_kvm_target::close): Call exit_inferior_silent
directly for the current inferior instead of
discard_all_inferiors.
(discard_all_inferiors): Delete.
GDB Administrator [Sun, 12 Jan 2020 00:01:17 +0000 (00:01 +0000)]
Automatic date update in version.in
Tom Tromey [Sat, 4 Jan 2020 21:35:02 +0000 (14:35 -0700)]
Make TUI borders respect "set style enabled"
When adding support for styling the TUI borders, I neglected to have
this code check cli_styling. As a result, "set style enabled off"
does not affect the borders.
This patch fixes this oversight. While doing this, I found that
running gdb without an executable, enabling the TUI, and then trying
"set style enabled off" would fail with the mysterious "No registers".
The fix for this is to use deprecated_safe_get_selected_frame in
tui_source_window_base::refill.
gdb/ChangeLog
2020-01-11 Tom Tromey <tom@tromey.com>
* tui/tui-wingeneral.c (box_win): Check cli_styling.
* tui/tui-winsource.c (tui_source_window_base::refill): Use
deprecated_safe_get_selected_frame.
Change-Id: I36acda25dd9014d994d366b4a0e8faee9d95d0f8
GDB Administrator [Sat, 11 Jan 2020 00:00:11 +0000 (00:00 +0000)]
Automatic date update in version.in
Tankut Baris Aktemur [Fri, 10 Jan 2020 20:06:17 +0000 (20:06 +0000)]
Switch the inferior before outputting its id in "info inferiors"
GDB uses the 'current_top_target' when displaying the description of
an inferior. This leads to same target being used for each inferior
and, in turn, yields incorrect output when the inferior has a target
that is supposed to give a specialized output. For instance, the
remote target outputs "Remote target" instead of "process XYZ" as the
description if the multi-process feature is not supported or turned
off.
E.g.: Suppose we have a native and a remote target, and the native is
the current inferior. The remote target does not support multi-process.
For "info inferiors", we would expect to see:
~~~
(gdb) i inferiors
Num Description Connection Executable
* 1 process 29060 1 (native) /a/path
2 Remote target 2 (remote ...)
~~~
but instead we get
~~~
(gdb) i inferiors
Num Description Connection Executable
* 1 process 29060 1 (native) /a/path
2 process 42000 2 (remote ...)
~~~
Similarly, if the current inferior is the remote one, we would expect
to see
~~~
(gdb) i inferiors
Num Description Connection Executable
1 process 29060 1 (native) /a/path
* 2 Remote target 2 (remote ...)
~~~
but we get
~~~
(gdb) i inferiors
Num Description Connection Executable
* 1 Remote target 1 (native) /a/path
2 Remote target 2 (remote ...)
~~~
With this patch, we switch to the inferior when outputting its
description, so that the current_top_target will be aligned to the
inferior we are displaying.
For testing, this patch expands the "info inferiors" test for the
multi-target feature. The test was checking for the output of the
info commands after setup, only when the current inferior is the last
added inferior.
This patch does the following to the testcase:
1. The "info inferiors" and "info connections" test is extracted out
from the "setup" procedure to a separate procedure.
2. The test is enriched to check the output after switching to each
inferior, not just the last one.
3. The test is performed twice; one for when the multi-process feature
is turned on, one for off.
gdb/ChangeLog:
2020-01-10 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* inferior.c (print_inferior): Switch inferior before printing it.
gdb/testsuite/ChangeLog:
2020-01-10 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* gdb.multi/multi-target.exp (setup): Factor out "info
connections" and "info inferiors" tests to ...
(test_info_inferiors): ... this new procedure.
(top level): Run new "info-inferiors" tests.
Pedro Alves [Fri, 10 Jan 2020 20:06:16 +0000 (20:06 +0000)]
Switch the inferior too in switch_to_program_space_and_thread
With multi-target, each inferior now has its own target connection.
The problem in switch_to_program_space_and_thread is that in the
current state GDB switches to "no thread" and also sets the program
space but because the inferior is not switched, potentially an
incorrect target remains selected.
Here is a sample scenario that exploits this flow:
On terminal 1, start a gdbserver on a program named foo:
$ gdbserver :1234 ./foo
On terminal 2, start gdb on a program named bar. Suppose foo and bar
are compiled from foo.c and bar.c. They are completely separate. So,
bar.c:2 has no meaning for foo.
$ gdb -q ./bar
Reading symbols from ./bar...
(gdb) add-inferior
[New inferior 2]
Added inferior 2
(gdb) inferior 2
[Switching to inferior 2 [<null>] (<noexec>)]
(gdb) target remote :1234
...
(gdb) set debug remote 2
(gdb) break bar.c:2
Sending packet: $Hgp0.0#ad...Packet received: OK
Sending packet: $m5fa,12#f8...Packet received: E01
Sending packet: $m5fa,1#c6...Packet received: E01
Sending packet: $m5fb,3#c9...Packet received: E01
Sending packet: $m5fe,1#ca...Packet received: E01
Breakpoint 1 at 0x5fe: file bar.c, line 2.
(gdb)
Here we have an unnecessary sending of the packets to the gdbserver.
With this fix in progspace-and-thread.c, we'll get this:
(gdb) break bar.c:2
Breakpoint 1 at 0x5fe: file bar.c, line 2.
(gdb)
Now there is no sending of the packets to gdbserver.
The changes around clear_symtab_users calls are necessary because
otherwise we regress gdb.base/step-over-exit.exp, hitting the new
assertion in switch_to_program_space_and_thread. The problem is, a
forked child terminates, and when GDB decides to auto-purge that
inferior, GDB tries to switch to the pspace of that no-longer-existing
inferior.
The root of the problem is within the program_space destructor:
program_space::~program_space ()
{
...
set_current_program_space (this); # (1)
...
breakpoint_program_space_exit (this); # (2)
...
free_all_objfiles (); # (3)
...
}
We get here from delete_inferior -> delete_program_space.
So we're deleting an inferior, and the inferior to be
deleted is no longer in the inferior list.
At (2), we've deleted all the breakpoints and locations for the
program space being deleted.
The crash happens while doing a breakpoint re-set, called by
clear_symtab_users at the tail end of (3). That is, while recreating
breakpoints for the current program space, which is the program space
we're tearing down. During breakpoint re-set, we try to switch to the
new location's pspace (the current pspace set in (1), so the pspace
we're tearing down) with switch_to_program_space_and_thread, and that
hits the failed assertion. It's the fact that we recreate breakpoints
in the program_space destructor that is the latent bug here. Just
don't do that, and we don't end up in the crash situation.
My first approach to fix this added a symfile_add_flags parameter to
program_space::free_all_objfiles, and then passed that down to
clear_symtab_users. The program_space dtor would then pass down
SYMFILE_DEFER_BP_RESET to free_all_objfiles. I couldn't help feeling
that adding that parameter to free_all_objfiles looked a little
awkward, so I settled on something a little different -- hoist the
clear_symtab_users call to the callers. There are only two callers.
I felt that that didn't look as odd, particularly since
remove_symbol_file_command also does:
objf->unlink ();
clear_symtab_users (0);
I.e., objfile deletion is already separate from calling
clear_symtab_users in some places.
gdb/ChangeLog:
2020-01-10 Aleksandar Paunovic <aleksandar.paunovic@intel.com>
Pedro Alves <palves@redhat.com>
* progspace-and-thread.c (switch_to_program_space_and_thread):
Assert there's an inferior for PSPACE. Use
switch_to_inferior_no_thread to switch the inferior too.
* progspace.c (program_space::~program_space): Call
clear_symtab_users here, with SYMFILE_DEFER_BP_RESET.
(program_space::free_all_objfiles): Don't call clear_symtab_users
here.
* symfile.c (symbol_file_clear): Call clear_symtab_users here.
gdb/testsuite/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdb.server/bkpt-other-inferior.exp: New file.
Pedro Alves [Fri, 10 Jan 2020 20:06:15 +0000 (20:06 +0000)]
Multi-target: NEWS and user manual
This commit documents the new multi-target features in both NEWS and
user manual.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* NEWS: Mention multi-target debugging, "info connections", and
"add-inferior -no-connection".
gdb/doc/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdb.texinfo (Starting): Say "current inferior not connected"
instead of "GDB not connected".
(Inferiors and Programs): Rename node to ...
(Inferiors Connections and Programs): ... this. Update all
references. Talk about multiple target connections. Update "info
inferiors" info to mention the connections column. Describe "info
connections". Document "add-inferior -no-connection".
* guile.texi, python.texi: Update cross references.
Pedro Alves [Fri, 10 Jan 2020 20:06:14 +0000 (20:06 +0000)]
Require always-non-stop for multi-target resumptions
Currently, we can only support resuming multiple targets at the same
time if all targets are in non-stop mode (or user-visible all-stop
mode with target backend in non-stop mode).
This patch makes GDB error out if the user tries to resume more than
one target at the same time and one of the resumed targets isn't in
non-stop mode:
(gdb) info inferiors
Num Description Connection Executable
1 process 15303 1 (native) a.out
* 2 process 15286 2 (extended-remote :9999) a.out
(gdb) set schedule-multiple on
(gdb) c
Continuing.
Connection 2 (extended-remote :9999) does not support multi-target resumption.
This is here later in the series instead of in the main multi-target
patch because it depends the previous patch, which added
process_stratum_target::connection_string().
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* infrun.c: Include "target-connection.h".
(check_multi_target_resumption): New.
(proceed): Call it.
* target-connection.c (make_target_connection_string): Make
extern.
* target-connection.h (make_target_connection_string): Declare.
Pedro Alves [Fri, 10 Jan 2020 20:06:14 +0000 (20:06 +0000)]
Add "info connections" command, "info inferiors" connection number/string
This commit extends the CLI a bit for multi-target, in three ways.
#1 - New "info connections" command.
This is a new command that lists the open connections (process_stratum
targets). For example, if you're debugging two remote connections, a
couple local/native processes, and a core dump, all at the same time,
you might see something like this:
(gdb) info connections
Num What Description
1 remote 192.168.0.1:9999 Remote serial target in gdb-specific protocol
2 remote 192.168.0.2:9998 Remote serial target in gdb-specific protocol
* 3 native Native process
4 core Local core dump file
#2 - New "info inferiors" "Connection" column
You'll also see a new matching "Connection" column in "info
inferiors", showing you which connection an inferior is bound to:
(gdb) info inferiors
Num Description Connection Executable
1 process 18526 1 (remote 192.168.0.1:9999) target:/tmp/a.out
2 process 18531 2 (remote 192.168.0.2:9998) target:/tmp/a.out
3 process 19115 3 (native) /tmp/prog1
4 process 6286 4 (core) myprogram
* 5 process 19122 3 (native) /bin/hello
#3 - Makes "add-inferior" show the inferior's target connection
"add-inferior" now shows you the connection you've just bound the
inferior to, which is the current process_stratum target:
(gdb) add-inferior
[New inferior 2]
Added inferior 2 on connection 1 (extended-remote localhost:2346)
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* Makefile.in (COMMON_SFILES): Add target-connection.c.
* inferior.c (uiout_field_connection): New function.
(print_inferior): Add new "connection-id" column.
(add_inferior_command): Show connection number/string of added
inferior.
* process-stratum-target.h
(process_stratum_target::connection_string): New virtual method.
(process_stratum_target::connection_number): New field.
* remote.c (remote_target::connection_string): New override.
* target-connection.c: New file.
* target-connection.h: New file.
* target.c (decref_target): Remove process_stratum targets from
the connection list.
(target_stack::push): Add process_stratum targets to the
connection list.
gdb/testsuite/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdb.base/kill-detach-inferiors-cmd.exp: Adjust expected output
of "add-inferior".
* gdb.base/quit-live.exp: Likewise.
* gdb.base/remote-exec-file.exp: Likewise.
* gdb.guile/scm-progspace.exp: Likewise.
* gdb.linespec/linespec.exp: Likewise.
* gdb.mi/new-ui-mi-sync.exp: Likewise.
* gdb.mi/user-selected-context-sync.exp: Likewise.
* gdb.multi/multi-target.exp (setup): Add "info connection" and
"info inferiors" tests.
* gdb.multi/remove-inferiors.exp: Adjust expected output of
"add-inferior".
* gdb.multi/watchpoint-multi.exp: Likewise.
* gdb.python/py-inferior.exp: Likewise.
* gdb.server/extended-remote-restart.exp: Likewise.
* gdb.threads/fork-plus-threads.exp: Adjust expected output of
"info inferiors".
* gdb.threads/forking-threads-plus-breakpoint.exp: Likewise.
* gdb.trace/report.exp: Likewise.
Pedro Alves [Fri, 10 Jan 2020 20:06:11 +0000 (20:06 +0000)]
Revert 'Remove unused struct serial::name field'
This commit reverts:
commit
5f5219fc34f7557296272230123a3837960a6f09
Author: Pedro Alves <palves@redhat.com>
AuthorDate: Tue Apr 12 16:49:30 2016 +0100
Remove unused struct serial::name field
The following patches will add uses for the field.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
Revert:
2016-04-12 Pedro Alves <palves@redhat.com>
* serial.c (serial_open, serial_fdopen_ops, do_serial_close):
Remove references to name.
* serial.h (struct serial) <name>: Delete.
Pedro Alves [Fri, 10 Jan 2020 20:06:10 +0000 (20:06 +0000)]
gdbarch-selftests.c: No longer error out if debugging something
Since each inferior has its own target stack, the stratum condition
for the "error out if debugging something" check is always false.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdbarch-selftests.c (register_to_value_test): Remove "target
already pushed" check.
Pedro Alves [Fri, 10 Jan 2020 20:06:09 +0000 (20:06 +0000)]
Add multi-target tests
This adds a testcase exercising multi-target features. It spawns 6
inferiors, like this:
inferior 1 -> native
inferior 2 -> extended-remote 1
inferior 3 -> core
inferior 4 -> native
inferior 5 -> extended-remote 2
inferior 6 -> core
and then tests various details, including:
- running to breakpoints
- interrupting with Ctrl-C and "interrupt -a"
- "next" bouncing between two breakpoints in two threads running in
different targets.
- since we have cores and live inferiors mixed in the same session,
this makes sure that gdb doesn't try to remove a core dump's
threads.
- all-stop and non-stop modes.
This testcase caught a _lot_ of bugs in development.
gdb/testsuite/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdb.multi/multi-target.c: New file.
* gdb.multi/multi-target.exp: New file.
* lib/gdbserver-support.exp (gdb_target_cmd): Handle "Non-stop
mode requested, but remote does not support non-stop".
Pedro Alves [Fri, 10 Jan 2020 20:06:08 +0000 (20:06 +0000)]
Multi-target support
This commit adds multi-target support to GDB. What this means is that
with this commit, GDB can now be connected to different targets at the
same time. E.g., you can debug a live native process and a core dump
at the same time, connect to multiple gdbservers, etc.
Actually, the word "target" is overloaded in gdb. We already have a
target stack, with pushes several target_ops instances on top of one
another. We also have "info target" already, which means something
completely different to what this patch does.
So from here on, I'll be using the "target connections" term, to mean
an open process_stratum target, pushed on a target stack. This patch
makes gdb have multiple target stacks, and multiple process_stratum
targets open simultaneously. The user-visible changes / commands will
also use this terminology, but of course it's all open to debate.
User-interface-wise, not that much changes. The main difference is
that each inferior may have its own target connection.
A target connection (e.g., a target extended-remote connection) may
support debugging multiple processes, just as before.
Say you're debugging against gdbserver in extended-remote mode, and
you do "add-inferior" to prepare to spawn a new process, like:
(gdb) target extended-remote :9999
...
(gdb) start
...
(gdb) add-inferior
Added inferior 2
(gdb) inferior 2
[Switching to inferior 2 [<null>] (<noexec>)]
(gdb) file a.out
...
(gdb) start
...
At this point, you have two inferiors connected to the same gdbserver.
With this commit, GDB will maintain a target stack per inferior,
instead of a global target stack.
To preserve the behavior above, by default, "add-inferior" makes the
new inferior inherit a copy of the target stack of the current
inferior. Same across a fork - the child inherits a copy of the
target stack of the parent. While the target stacks are copied, the
targets themselves are not. Instead, target_ops is made a
refcounted_object, which means that target_ops instances are
refcounted, which each inferior counting for a reference.
What if you want to create an inferior and connect it to some _other_
target? For that, this commit introduces a new "add-inferior
-no-connection" option that makes the new inferior not share the
current inferior's target. So you could do:
(gdb) target extended-remote :9999
Remote debugging using :9999
...
(gdb) add-inferior -no-connection
[New inferior 2]
Added inferior 2
(gdb) inferior 2
[Switching to inferior 2 [<null>] (<noexec>)]
(gdb) info inferiors
Num Description Executable
1 process 18401 target:/home/pedro/tmp/main
* 2 <null>
(gdb) tar extended-remote :10000
Remote debugging using :10000
...
(gdb) info inferiors
Num Description Executable
1 process 18401 target:/home/pedro/tmp/main
* 2 process 18450 target:/home/pedro/tmp/main
(gdb)
A following patch will extended "info inferiors" to include a column
indicating which connection an inferior is bound to, along with a
couple other UI tweaks.
Other than that, debugging is the same as before. Users interact with
inferiors and threads as before. The only difference is that
inferiors may be bound to processes running in different machines.
That's pretty much all there is to it in terms of noticeable UI
changes.
On to implementation.
Since we can be connected to different systems at the same time, a
ptid_t is no longer a unique identifier. Instead a thread can be
identified by a pair of ptid_t and 'process_stratum_target *', the
later being the instance of the process_stratum target that owns the
process/thread. Note that process_stratum_target inherits from
target_ops, and all process_stratum targets inherit from
process_stratum_target. In earlier patches, many places in gdb were
converted to refer to threads by thread_info pointer instead of
ptid_t, but there are still places in gdb where we start with a
pid/tid and need to find the corresponding inferior or thread_info
objects. So you'll see in the patch many places adding a
process_stratum_target parameter to functions that used to take only a
ptid_t.
Since each inferior has its own target stack now, we can always find
the process_stratum target for an inferior. That is done via a
inf->process_target() convenience method.
Since each inferior has its own target stack, we need to handle the
"beneath" calls when servicing target calls. The solution I settled
with is just to make sure to switch the current inferior to the
inferior you want before making a target call. Not relying on global
context is just not feasible in current GDB. Fortunately, there
aren't that many places that need to do that, because generally most
code that calls target methods already has the current context
pointing to the right inferior/thread. Note, to emphasize -- there's
no method to "switch to this target stack". Instead, you switch the
current inferior, and that implicitly switches the target stack.
In some spots, we need to iterate over all inferiors so that we reach
all target stacks.
Native targets are still singletons. There's always only a single
instance of such targets.
Remote targets however, we'll have one instance per remote connection.
The exec target is still a singleton. There's only one instance. I
did not see the point of instanciating more than one exec_target
object.
After vfork, we need to make sure to push the exec target on the new
inferior. See exec_on_vfork.
For type safety, functions that need a {target, ptid} pair to identify
a thread, take a process_stratum_target pointer for target parameter
instead of target_ops *. Some shared code in gdb/nat/ also need to
gain a target pointer parameter. This poses an issue, since gdbserver
doesn't have process_stratum_target, only target_ops. To fix this,
this commit renames gdbserver's target_ops to process_stratum_target.
I think this makes sense. There's no concept of target stack in
gdbserver, and gdbserver's target_ops really implements a
process_stratum-like target.
The thread and inferior iterator functions also gain
process_stratum_target parameters. These are used to be able to
iterate over threads and inferiors of a given target. Following usual
conventions, if the target pointer is null, then we iterate over
threads and inferiors of all targets.
I tried converting "add-inferior" to the gdb::option framework, as a
preparatory patch, but that stumbled on the fact that gdb::option does
not support file options yet, for "add-inferior -exec". I have a WIP
patchset that adds that, but it's not a trivial patch, mainly due to
need to integrate readline's filename completion, so I deferred that
to some other time.
In infrun.c/infcmd.c, the main change is that we need to poll events
out of all targets. See do_target_wait. Right after collecting an
event, we switch the current inferior to an inferior bound to the
target that reported the event, so that target methods can be used
while handling the event. This makes most of the code transparent to
multi-targets. See fetch_inferior_event.
infrun.c:stop_all_threads is interesting -- in this function we need
to stop all threads of all targets. What the function does is send an
asynchronous stop request to all threads, and then synchronously waits
for events, with target_wait, rinse repeat, until all it finds are
stopped threads. Now that we have multiple targets, it's not
efficient to synchronously block in target_wait waiting for events out
of one target. Instead, we implement a mini event loop, with
interruptible_select, select'ing on one file descriptor per target.
For this to work, we need to be able to ask the target for a waitable
file descriptor. Such file descriptors already exist, they are the
descriptors registered in the main event loop with add_file_handler,
inside the target_async implementations. This commit adds a new
target_async_wait_fd target method that just returns the file
descriptor in question. See wait_one / stop_all_threads in infrun.c.
The 'threads_executing' global is made a per-target variable. Since
it is only relevant to process_stratum_target targets, this is where
it is put, instead of in target_ops.
You'll notice that remote.c includes some FIXME notes. These refer to
the fact that the global arrays that hold data for the remote packets
supported are still globals. For example, if we connect to two
different servers/stubs, then each might support different remote
protocol features. They might even be different architectures, like
e.g., one ARM baremetal stub, and a x86 gdbserver, to debug a
host/controller scenario as a single program. That isn't going to
work correctly today, because of said globals. I'm leaving fixing
that for another pass, since it does not appear to be trivial, and I'd
rather land the base work first. It's already useful to be able to
debug multiple instances of the same server (e.g., a distributed
cluster, where you have full control over the servers installed), so I
think as is it's already reasonable incremental progress.
Current limitations:
- You can only resume more that one target at the same time if all
targets support asynchronous debugging, and support non-stop mode.
It should be possible to support mixed all-stop + non-stop
backends, but that is left for another time. This means that
currently in order to do multi-target with gdbserver you need to
issue "maint set target-non-stop on". I would like to make that
mode be the default, but we're not there yet. Note that I'm
talking about how the target backend works, only. User-visible
all-stop mode works just fine.
- As explained above, connecting to different remote servers at the
same time is likely to produce bad results if they don't support the
exact set of RSP features.
FreeBSD updates courtesy of John Baldwin.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
John Baldwin <jhb@FreeBSD.org>
* aarch64-linux-nat.c
(aarch64_linux_nat_target::thread_architecture): Adjust.
* ada-tasks.c (print_ada_task_info): Adjust find_thread_ptid call.
(task_command_1): Likewise.
* aix-thread.c (sync_threadlists, aix_thread_target::resume)
(aix_thread_target::wait, aix_thread_target::fetch_registers)
(aix_thread_target::store_registers)
(aix_thread_target::thread_alive): Adjust.
* amd64-fbsd-tdep.c: Include "inferior.h".
(amd64fbsd_get_thread_local_address): Pass down target.
* amd64-linux-nat.c (ps_get_thread_area): Use ps_prochandle
thread's gdbarch instead of target_gdbarch.
* break-catch-sig.c (signal_catchpoint_print_it): Adjust call to
get_last_target_status.
* break-catch-syscall.c (print_it_catch_syscall): Likewise.
* breakpoint.c (breakpoints_should_be_inserted_now): Consider all
inferiors.
(update_inserted_breakpoint_locations): Skip if inferiors with no
execution.
(update_global_location_list): When handling moribund locations,
find representative inferior for location's pspace, and use thread
count of its process_stratum target.
* bsd-kvm.c (bsd_kvm_target_open): Pass target down.
* bsd-uthread.c (bsd_uthread_target::wait): Use
as_process_stratum_target and adjust thread_change_ptid and
add_thread calls.
(bsd_uthread_target::update_thread_list): Use
as_process_stratum_target and adjust find_thread_ptid,
thread_change_ptid and add_thread calls.
* btrace.c (maint_btrace_packet_history_cmd): Adjust
find_thread_ptid call.
* corelow.c (add_to_thread_list): Adjust add_thread call.
(core_target_open): Adjust add_thread_silent and thread_count
calls.
(core_target::pid_to_str): Adjust find_inferior_ptid call.
* ctf.c (ctf_target_open): Adjust add_thread_silent call.
* event-top.c (async_disconnect): Pop targets from all inferiors.
* exec.c (add_target_sections): Push exec target on all inferiors
sharing the program space.
(remove_target_sections): Remove the exec target from all
inferiors sharing the program space.
(exec_on_vfork): New.
* exec.h (exec_on_vfork): Declare.
* fbsd-nat.c (fbsd_add_threads): Add fbsd_nat_target parameter.
Pass it down.
(fbsd_nat_target::update_thread_list): Adjust.
(fbsd_nat_target::resume): Adjust.
(fbsd_handle_debug_trap): Add fbsd_nat_target parameter. Pass it
down.
(fbsd_nat_target::wait, fbsd_nat_target::post_attach): Adjust.
* fbsd-tdep.c (fbsd_corefile_thread): Adjust
get_thread_arch_regcache call.
* fork-child.c (gdb_startup_inferior): Pass target down to
startup_inferior and set_executing.
* gdbthread.h (struct process_stratum_target): Forward declare.
(add_thread, add_thread_silent, add_thread_with_info)
(in_thread_list): Add process_stratum_target parameter.
(find_thread_ptid(inferior*, ptid_t)): New overload.
(find_thread_ptid, thread_change_ptid): Add process_stratum_target
parameter.
(all_threads()): Delete overload.
(all_threads, all_non_exited_threads): Add process_stratum_target
parameter.
(all_threads_safe): Use brace initialization.
(thread_count): Add process_stratum_target parameter.
(set_resumed, set_running, set_stop_requested, set_executing)
(threads_are_executing, finish_thread_state): Add
process_stratum_target parameter.
(switch_to_thread): Use is_current_thread.
* i386-fbsd-tdep.c: Include "inferior.h".
(i386fbsd_get_thread_local_address): Pass down target.
* i386-linux-nat.c (i386_linux_nat_target::low_resume): Adjust.
* inf-child.c (inf_child_target::maybe_unpush_target): Remove
have_inferiors check.
* inf-ptrace.c (inf_ptrace_target::create_inferior)
(inf_ptrace_target::attach): Adjust.
* infcall.c (run_inferior_call): Adjust.
* infcmd.c (run_command_1): Pass target to
scoped_finish_thread_state.
(proceed_thread_callback): Skip inferiors with no execution.
(continue_command): Rename 'all_threads' local to avoid hiding
'all_threads' function. Adjust get_last_target_status call.
(prepare_one_step): Adjust set_running call.
(signal_command): Use user_visible_resume_target. Compare thread
pointers instead of inferior_ptid.
(info_program_command): Adjust to pass down target.
(attach_command): Mark target's 'thread_executing' flag.
(stop_current_target_threads_ns): New, factored out from ...
(interrupt_target_1): ... this. Switch inferior before making
target calls.
* inferior-iter.h
(struct all_inferiors_iterator, struct all_inferiors_range)
(struct all_inferiors_safe_range)
(struct all_non_exited_inferiors_range): Filter on
process_stratum_target too. Remove explicit.
* inferior.c (inferior::inferior): Push dummy target on target
stack.
(find_inferior_pid, find_inferior_ptid, number_of_live_inferiors):
Add process_stratum_target parameter, and pass it down.
(have_live_inferiors): Adjust.
(switch_to_inferior_and_push_target): New.
(add_inferior_command, clone_inferior_command): Handle
"-no-connection" parameter. Use
switch_to_inferior_and_push_target.
(_initialize_inferior): Mention "-no-connection" option in
the help of "add-inferior" and "clone-inferior" commands.
* inferior.h: Include "process-stratum-target.h".
(interrupt_target_1): Use bool.
(struct inferior) <push_target, unpush_target, target_is_pushed,
find_target_beneath, top_target, process_target, target_at,
m_stack>: New.
(discard_all_inferiors): Delete.
(find_inferior_pid, find_inferior_ptid, number_of_live_inferiors)
(all_inferiors, all_non_exited_inferiors): Add
process_stratum_target parameter.
* infrun.c: Include "gdb_select.h" and <unordered_map>.
(target_last_proc_target): New global.
(follow_fork_inferior): Push target on new inferior. Pass target
to add_thread_silent. Call exec_on_vfork. Handle target's
reference count.
(follow_fork): Adjust get_last_target_status call. Also consider
target.
(follow_exec): Push target on new inferior.
(struct execution_control_state) <target>: New field.
(user_visible_resume_target): New.
(do_target_resume): Call target_async.
(resume_1): Set target's threads_executing flag. Consider resume
target.
(commit_resume_all_targets): New.
(proceed): Also consider resume target. Skip threads of inferiors
with no execution. Commit resumtion in all targets.
(start_remote): Pass current inferior to wait_for_inferior.
(infrun_thread_stop_requested): Consider target as well. Pass
thread_info pointer to clear_inline_frame_state instead of ptid.
(infrun_thread_thread_exit): Consider target as well.
(random_pending_event_thread): New inferior parameter. Use it.
(do_target_wait): Rename to ...
(do_target_wait_1): ... this. Add inferior parameter, and pass it
down.
(threads_are_resumed_pending_p, do_target_wait): New.
(prepare_for_detach): Adjust calls.
(wait_for_inferior): New inferior parameter. Handle it. Use
do_target_wait_1 instead of do_target_wait.
(fetch_inferior_event): Adjust. Switch to representative
inferior. Pass target down.
(set_last_target_status): Add process_stratum_target parameter.
Save target in global.
(get_last_target_status): Add process_stratum_target parameter and
handle it.
(nullify_last_target_wait_ptid): Clear 'target_last_proc_target'.
(context_switch): Check inferior_ptid == null_ptid before calling
inferior_thread().
(get_inferior_stop_soon): Pass down target.
(wait_one): Rename to ...
(poll_one_curr_target): ... this.
(struct wait_one_event): New.
(wait_one): New.
(stop_all_threads): Adjust.
(handle_no_resumed, handle_inferior_event): Adjust to consider the
event's target.
(switch_back_to_stepped_thread): Also consider target.
(print_stop_event): Update.
(normal_stop): Update. Also consider the resume target.
* infrun.h (wait_for_inferior): Remove declaration.
(user_visible_resume_target): New declaration.
(get_last_target_status, set_last_target_status): New
process_stratum_target parameter.
* inline-frame.c (clear_inline_frame_state(ptid_t)): Add
process_stratum_target parameter, and use it.
(clear_inline_frame_state (thread_info*)): New.
* inline-frame.c (clear_inline_frame_state(ptid_t)): Add
process_stratum_target parameter.
(clear_inline_frame_state (thread_info*)): Declare.
* linux-fork.c (delete_checkpoint_command): Pass target down to
find_thread_ptid.
(checkpoint_command): Adjust.
* linux-nat.c (linux_nat_target::follow_fork): Switch to thread
instead of just tweaking inferior_ptid.
(linux_nat_switch_fork): Pass target down to thread_change_ptid.
(exit_lwp): Pass target down to find_thread_ptid.
(attach_proc_task_lwp_callback): Pass target down to
add_thread/set_running/set_executing.
(linux_nat_target::attach): Pass target down to
thread_change_ptid.
(get_detach_signal): Pass target down to find_thread_ptid.
Consider last target status's target.
(linux_resume_one_lwp_throw, resume_lwp)
(linux_handle_syscall_trap, linux_handle_extended_wait, wait_lwp)
(stop_wait_callback, save_stop_reason, linux_nat_filter_event)
(linux_nat_wait_1, resume_stopped_resumed_lwps): Pass target down.
(linux_nat_target::async_wait_fd): New.
(linux_nat_stop_lwp, linux_nat_target::thread_address_space): Pass
target down.
* linux-nat.h (linux_nat_target::async_wait_fd): Declare.
* linux-tdep.c (get_thread_arch_regcache): Pass target down.
* linux-thread-db.c (struct thread_db_info::process_target): New
field.
(add_thread_db_info): Save target.
(get_thread_db_info): New process_stratum_target parameter. Also
match target.
(delete_thread_db_info): New process_stratum_target parameter.
Also match target.
(thread_from_lwp): Adjust to pass down target.
(thread_db_notice_clone): Pass down target.
(check_thread_db_callback): Pass down target.
(try_thread_db_load_1): Always push the thread_db target.
(try_thread_db_load, record_thread): Pass target down.
(thread_db_target::detach): Pass target down. Always unpush the
thread_db target.
(thread_db_target::wait, thread_db_target::mourn_inferior): Pass
target down. Always unpush the thread_db target.
(find_new_threads_callback, thread_db_find_new_threads_2)
(thread_db_target::update_thread_list): Pass target down.
(thread_db_target::pid_to_str): Pass current inferior down.
(thread_db_target::get_thread_local_address): Pass target down.
(thread_db_target::resume, maintenance_check_libthread_db): Pass
target down.
* nto-procfs.c (nto_procfs_target::update_thread_list): Adjust.
* procfs.c (procfs_target::procfs_init_inferior): Declare.
(proc_set_current_signal, do_attach, procfs_target::wait): Adjust.
(procfs_init_inferior): Rename to ...
(procfs_target::procfs_init_inferior): ... this and adjust.
(procfs_target::create_inferior, procfs_notice_thread)
(procfs_do_thread_registers): Adjust.
* ppc-fbsd-tdep.c: Include "inferior.h".
(ppcfbsd_get_thread_local_address): Pass down target.
* proc-service.c (ps_xfer_memory): Switch current inferior and
program space as well.
(get_ps_regcache): Pass target down.
* process-stratum-target.c
(process_stratum_target::thread_address_space)
(process_stratum_target::thread_architecture): Pass target down.
* process-stratum-target.h
(process_stratum_target::threads_executing): New field.
(as_process_stratum_target): New.
* ravenscar-thread.c
(ravenscar_thread_target::update_inferior_ptid): Pass target down.
(ravenscar_thread_target::wait, ravenscar_add_thread): Pass target
down.
* record-btrace.c (record_btrace_target::info_record): Adjust.
(record_btrace_target::record_method)
(record_btrace_target::record_is_replaying)
(record_btrace_target::fetch_registers)
(get_thread_current_frame_id, record_btrace_target::resume)
(record_btrace_target::wait, record_btrace_target::stop): Pass
target down.
* record-full.c (record_full_wait_1): Switch to event thread.
Pass target down.
* regcache.c (regcache::regcache)
(get_thread_arch_aspace_regcache, get_thread_arch_regcache): Add
process_stratum_target parameter and handle it.
(current_thread_target): New global.
(get_thread_regcache): Add process_stratum_target parameter and
handle it. Switch inferior before calling target method.
(get_thread_regcache): Pass target down.
(get_thread_regcache_for_ptid): Pass target down.
(registers_changed_ptid): Add process_stratum_target parameter and
handle it.
(registers_changed_thread, registers_changed): Pass target down.
(test_get_thread_arch_aspace_regcache): New.
(current_regcache_test): Define a couple local test_target_ops
instances and use them for testing.
(readwrite_regcache): Pass process_stratum_target parameter.
(cooked_read_test, cooked_write_test): Pass mock_target down.
* regcache.h (get_thread_regcache, get_thread_arch_regcache)
(get_thread_arch_aspace_regcache): Add process_stratum_target
parameter.
(regcache::target): New method.
(regcache::regcache, regcache::get_thread_arch_aspace_regcache)
(regcache::registers_changed_ptid): Add process_stratum_target
parameter.
(regcache::m_target): New field.
(registers_changed_ptid): Add process_stratum_target parameter.
* remote.c (remote_state::supports_vCont_probed): New field.
(remote_target::async_wait_fd): New method.
(remote_unpush_and_throw): Add remote_target parameter.
(get_current_remote_target): Adjust.
(remote_target::remote_add_inferior): Push target.
(remote_target::remote_add_thread)
(remote_target::remote_notice_new_inferior)
(get_remote_thread_info): Pass target down.
(remote_target::update_thread_list): Skip threads of inferiors
bound to other targets. (remote_target::close): Don't discard
inferiors. (remote_target::add_current_inferior_and_thread)
(remote_target::process_initial_stop_replies)
(remote_target::start_remote)
(remote_target::remote_serial_quit_handler): Pass down target.
(remote_target::remote_unpush_target): New remote_target
parameter. Unpush the target from all inferiors.
(remote_target::remote_unpush_and_throw): New remote_target
parameter. Pass it down.
(remote_target::open_1): Check whether the current inferior has
execution instead of checking whether any inferior is live. Pass
target down.
(remote_target::remote_detach_1): Pass down target. Use
remote_unpush_target.
(extended_remote_target::attach): Pass down target.
(remote_target::remote_vcont_probe): Set supports_vCont_probed.
(remote_target::append_resumption): Pass down target.
(remote_target::append_pending_thread_resumptions)
(remote_target::remote_resume_with_hc, remote_target::resume)
(remote_target::commit_resume): Pass down target.
(remote_target::remote_stop_ns): Check supports_vCont_probed.
(remote_target::interrupt_query)
(remote_target::remove_new_fork_children)
(remote_target::check_pending_events_prevent_wildcard_vcont)
(remote_target::remote_parse_stop_reply)
(remote_target::process_stop_reply): Pass down target.
(first_remote_resumed_thread): New remote_target parameter. Pass
it down.
(remote_target::wait_as): Pass down target.
(unpush_and_perror): New remote_target parameter. Pass it down.
(remote_target::readchar, remote_target::remote_serial_write)
(remote_target::getpkt_or_notif_sane_1)
(remote_target::kill_new_fork_children, remote_target::kill): Pass
down target.
(remote_target::mourn_inferior): Pass down target. Use
remote_unpush_target.
(remote_target::core_of_thread)
(remote_target::remote_btrace_maybe_reopen): Pass down target.
(remote_target::pid_to_exec_file)
(remote_target::thread_handle_to_thread_info): Pass down target.
(remote_target::async_wait_fd): New.
* riscv-fbsd-tdep.c: Include "inferior.h".
(riscv_fbsd_get_thread_local_address): Pass down target.
* sol2-tdep.c (sol2_core_pid_to_str): Pass down target.
* sol-thread.c (sol_thread_target::wait, ps_lgetregs, ps_lsetregs)
(ps_lgetfpregs, ps_lsetfpregs, sol_update_thread_list_callback):
Adjust.
* solib-spu.c (spu_skip_standalone_loader): Pass down target.
* solib-svr4.c (enable_break): Pass down target.
* spu-multiarch.c (parse_spufs_run): Pass down target.
* spu-tdep.c (spu2ppu_sniffer): Pass down target.
* target-delegates.c: Regenerate.
* target.c (g_target_stack): Delete.
(current_top_target): Return the current inferior's top target.
(target_has_execution_1): Refer to the passed-in inferior's top
target.
(target_supports_terminal_ours): Check whether the initial
inferior was already created.
(decref_target): New.
(target_stack::push): Incref/decref the target.
(push_target, push_target, unpush_target): Adjust.
(target_stack::unpush): Defref target.
(target_is_pushed): Return bool. Adjust to refer to the current
inferior's target stack.
(dispose_inferior): Delete, and inline parts ...
(target_preopen): ... here. Only dispose of the current inferior.
(target_detach): Hold strong target reference while detaching.
Pass target down.
(target_thread_name): Add assertion.
(target_resume): Pass down target.
(target_ops::beneath, find_target_at): Adjust to refer to the
current inferior's target stack.
(get_dummy_target): New.
(target_pass_ctrlc): Pass the Ctrl-C to the first inferior that
has a thread running.
(initialize_targets): Rename to ...
(_initialize_target): ... this.
* target.h: Include "gdbsupport/refcounted-object.h".
(struct target_ops): Inherit refcounted_object.
(target_ops::shortname, target_ops::longname): Make const.
(target_ops::async_wait_fd): New method.
(decref_target): Declare.
(struct target_ops_ref_policy): New.
(target_ops_ref): New typedef.
(get_dummy_target): Declare function.
(target_is_pushed): Return bool.
* thread-iter.c (all_matching_threads_iterator::m_inf_matches)
(all_matching_threads_iterator::all_matching_threads_iterator):
Handle filter target.
* thread-iter.h (struct all_matching_threads_iterator, struct
all_matching_threads_range, class all_non_exited_threads_range):
Filter by target too. Remove explicit.
* thread.c (threads_executing): Delete.
(inferior_thread): Pass down current inferior.
(clear_thread_inferior_resources): Pass down thread pointer
instead of ptid_t.
(add_thread_silent, add_thread_with_info, add_thread): Add
process_stratum_target parameter. Use it for thread and inferior
searches.
(is_current_thread): New.
(thread_info::deletable): Use it.
(find_thread_ptid, thread_count, in_thread_list)
(thread_change_ptid, set_resumed, set_running): New
process_stratum_target parameter. Pass it down.
(set_executing): New process_stratum_target parameter. Pass it
down. Adjust reference to 'threads_executing'.
(threads_are_executing): New process_stratum_target parameter.
Adjust reference to 'threads_executing'.
(set_stop_requested, finish_thread_state): New
process_stratum_target parameter. Pass it down.
(switch_to_thread): Also match inferior.
(switch_to_thread): New process_stratum_target parameter. Pass it
down.
(update_threads_executing): Reimplement.
* top.c (quit_force): Pop targets from all inferior.
(gdb_init): Don't call initialize_targets.
* windows-nat.c (windows_nat_target) <get_windows_debug_event>:
Declare.
(windows_add_thread, windows_delete_thread): Adjust.
(get_windows_debug_event): Rename to ...
(windows_nat_target::get_windows_debug_event): ... this. Adjust.
* tracefile-tfile.c (tfile_target_open): Pass down target.
* gdbsupport/common-gdbthread.h (struct process_stratum_target):
Forward declare.
(switch_to_thread): Add process_stratum_target parameter.
* mi/mi-interp.c (mi_on_resume_1): Add process_stratum_target
parameter. Use it.
(mi_on_resume): Pass target down.
* nat/fork-inferior.c (startup_inferior): Add
process_stratum_target parameter. Pass it down.
* nat/fork-inferior.h (startup_inferior): Add
process_stratum_target parameter.
* python/py-threadevent.c (py_get_event_thread): Pass target down.
gdb/gdbserver/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* fork-child.c (post_fork_inferior): Pass target down to
startup_inferior.
* inferiors.c (switch_to_thread): Add process_stratum_target
parameter.
* lynx-low.c (lynx_target_ops): Now a process_stratum_target.
* nto-low.c (nto_target_ops): Now a process_stratum_target.
* linux-low.c (linux_target_ops): Now a process_stratum_target.
* remote-utils.c (prepare_resume_reply): Pass the target to
switch_to_thread.
* target.c (the_target): Now a process_stratum_target.
(done_accessing_memory): Pass the target to switch_to_thread.
(set_target_ops): Ajust to use process_stratum_target.
* target.h (struct target_ops): Rename to ...
(struct process_stratum_target): ... this.
(the_target, set_target_ops): Adjust.
(prepare_to_access_memory): Adjust comment.
* win32-low.c (child_xfer_memory): Adjust to use
process_stratum_target.
(win32_target_ops): Now a process_stratum_target.
Pedro Alves [Fri, 10 Jan 2020 20:05:54 +0000 (20:05 +0000)]
Fix reconnecting to a gdbserver already debugging multiple processes, II
Another bug exposed by gdb.server/extended-remote-restart.exp in the
multi-target work is that remote_target::start_remote can leave
inferior_ptid and current_inferior() out of sync:
(top-gdb) p current_inferior_->pid
$1 = 29541
(top-gdb) p inferior_ptid
$2 = {m_pid = 29540, m_lwp = 29540, m_tid = 0}
This is caused by writing to inferior_ptid directly instead of using
switch_to_thread. Also, "inferior_list->thread_list->ptid" assumes
that we want the first thread of the first inferior, but that inferior
may not have threads, or with multi-target, that target may be
connected to some other target.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* remote.c (remote_target::start_remote): Don't set inferior_ptid
directly. Instead find the first thread in the thread list and
use switch_to_thread.
Pedro Alves [Fri, 10 Jan 2020 20:05:53 +0000 (20:05 +0000)]
Fix reconnecting to a gdbserver already debugging multiple processes, I
The multi-target patch will change the remote target's behavior when:
- the current inferior is connected to an extended-remote target.
- the current inferior is attached to any process.
- some other inferior than than the current one is live.
In current master, we get:
(gdb) tar extended-remote :9999
A program is being debugged already. Kill it? (y or n)
While after multi-target, since each inferior may have its own target
connection, we'll get:
(gdb) tar extended-remote :9999
Already connected to a remote target. Disconnect? (y or n)
That change made gdb.server/extended-remote-restart.exp expose a gdb
bug, because it made "target remote", via gdb_reconnect, just
disconnect from the previous connection, while in current master that
command would kill the inferior before disconnecting. In turn, that
would make a multi-target gdb find processes already running under
control of gdbserver as soon as it reconnects, while in current master
there is never any process around when gdb reconnects, since they'd
all been killed prior to disconnection.
The bug this exposed is that remote_target::remote_add_inferior was
always reusing current_inferior() for the new process, even if the
current inferior was already bound to a process. In the testcase's
case, when we reconnect, the remote is debugging two processes. So
we'd bind the first remote process to the empty current inferior the
first time, and then bind the second remote process to the same
inferior again, essencially losing track of the first process. That
resulted in failed assertions when we look up the inferior for the
first process by PID. The fix is to still prefer binding to the
current inferior (so that plain "target remote" keeps doing what you'd
expect), but not reuse the current inferior if it is already bound to
a process.
This patch tweaks the test to explicitly disconnect before
reconnecting, to avoid GDB killing processes, thus making current GDB
behave the same as it will behave when the multi-target work lands.
That change alone without the GDB fix exposes the bug like so:
(gdb) PASS: gdb.server/extended-remote-restart.exp: kill: 0, follow-child 0: disconnect
target extended-remote localhost:2350
Remote debugging using localhost:2350
src/gdb/thread.c:93: internal-error: thread_info* inferior_thread(): Assertion `tp' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n)
The original bug that the testcase was written for was related to
killing, (git
9d4a934ce604 ("gdb: Fix assert for extended-remote
target (PR gdb/18050)")), but since the testcase tries reconnecting
with both explicitly killing and not explicitly killing, I think we're
covering the original bug with this testcase change.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* remote.c (remote_target::remote_add_inferior): Don't bind a
process to the current inferior if the current inferior is already
bound to a process.
gdb/testsuite/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdb.server/extended-remote-restart.exp (test_reload): Explicitly
disconnect before reconnecting.
Tankut Baris Aktemur [Fri, 10 Jan 2020 20:05:52 +0000 (20:05 +0000)]
Avoid another inferior_ptid reference in gdb/remote.c
The multi-target patch makes inferior_ptid point to null_ptid before
calling into target_wait, which catches bad uses of inferior_ptid,
since the current selected thread in gdb shouldn't have much relation
to the thread that reports an event.
One such bad use is found in remote_target::remote_parse_stop_reply,
where we handle the 'W' or 'X' packets (process exit), and the remote
target does not support the multi-process extensions, i.e., it does
not report the PID of the process that exited.
With the multi-target patch, that would result in a failed assertion,
trying to find the inferior for process pid 0.
gdb/ChangeLog:
2020-01-10 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
Pedro Alves <palves@redhat.com>
* remote.c (remote_target::remote_parse_stop_reply) <W/X packets>:
If no process is specified, return null_ptid instead of
inferior_ptid.
(remote_target::wait_as): Handle TARGET_WAITKIND_EXITED /
TARGET_WAITKIND_SIGNALLED with no pid.
gdb/testsuite/ChangeLog:
2020-01-10 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
Pedro Alves <palves@redhat.com>
* gdb.server/connect-without-multi-process.exp: Also test
continuing to end.
Pedro Alves [Fri, 10 Jan 2020 20:05:52 +0000 (20:05 +0000)]
Tweak handling of remote errors in response to resumption packet
With current master, on a Fedora 27 machine with a kernel with buggy
watchpoint support, I see:
(gdb) PASS: gdb.threads/watchpoint-fork.exp: parent: singlethreaded: hardware breakpoints work
continue
Continuing.
warning: Remote failure reply: E01
Remote communication error. Target disconnected.: Connection reset by peer.
(gdb) FAIL: gdb.threads/watchpoint-fork.exp: parent: singlethreaded: watchpoints work
continue
The program is not being run.
(gdb) FAIL: gdb.threads/watchpoint-fork.exp: parent: singlethreaded: breakpoint after the first fork (the program is no longer running)
The FAILs themselves aren't what's interesting here. What is
interesting is that with the main multi-target patch applied, I was getting this:
(gdb) PASS: gdb.threads/watchpoint-fork.exp: parent: singlethreaded: hardware breakpoints work
continue
Continuing.
warning: Remote failure reply: E01
/home/pedro/brno/pedro/gdb/binutils-gdb-2/build/../src/gdb/inferior.c:285: internal-error: inferior* find_inferior_pid(process_stratum_target*, int): Assertion `pid != 0' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n) FAIL: gdb.threads/watchpoint-fork.exp: parent: singlethreaded: watchpoints work (GDB internal error)
The problem is that in remote_target::wait_as, we're hitting this:
switch (buf[0])
{
case 'E': /* Error of some sort. */
/* We're out of sync with the target now. Did it continue or
not? Not is more likely, so report a stop. */
rs->waiting_for_stop_reply = 0;
warning (_("Remote failure reply: %s"), buf);
status->kind = TARGET_WAITKIND_STOPPED;
status->value.sig = GDB_SIGNAL_0;
break;
which leaves event_ptid as null_ptid. At the end of the function, we then reach:
else if (status->kind != TARGET_WAITKIND_EXITED
&& status->kind != TARGET_WAITKIND_SIGNALLED)
{
if (event_ptid != null_ptid)
record_currthread (rs, event_ptid);
else
event_ptid = inferior_ptid; <<<<< here
}
and the trouble is that with the multi-target patch, we'll get here
with inferior_ptid as null_ptid too. That is done exactly to find
these implicit assumptions that inferior_ptid is a good choice for
default thread, which isn't generaly true.
I first thought of fixing this in the "case 'E'" path, but, given that
this "event_ptid = inferior_ptid" path is also taken when the remote
target does not support threads at all, no thread-related packets or
extensions, it's better to fix it in latter path, to handle all
scenarios that miss reporting a thread.
That's what this patch does.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* remote.c (first_remote_resumed_thread): New.
(remote_target::wait_as): Use it as default event_ptid instead of
inferior_ptid.
Pedro Alves [Fri, 10 Jan 2020 20:05:51 +0000 (20:05 +0000)]
Use all_non_exited_inferiors in infrun.c
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* infrun.c (handle_no_resumed): Use all_non_exited_inferiors.
Pedro Alves [Fri, 10 Jan 2020 20:05:50 +0000 (20:05 +0000)]
tfile_target::close: trace_fd can't be -1
It's not possible to open a tfile target with an invalid trace_fd, and
it's not possible to close a closed target, so this early return is dead.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* tracefile-tfile.c (tfile_target::close): Assert that trace_fd is
not -1.
Pedro Alves [Fri, 10 Jan 2020 20:05:49 +0000 (20:05 +0000)]
Some get_last_target_status tweaks
- Make get_last_target_status arguments optional. A following patch
will add another argument to get_last_target_status (the event's
target), and passing nullptr when we don't care for some piece of
info is handier than creating dummy local variables.
- Declare nullify_last_target_wait_ptid in a header, and remove the
local extern declaration from linux-fork.c.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* break-catch-sig.c (signal_catchpoint_print_it): Don't pass a
ptid to get_last_target_status.
* break-catch-syscall.c (print_it_catch_syscall): Don't pass a
ptid to get_last_target_status.
* infcmd.c (continue_command): Don't pass a target_waitstatus to
get_last_target_status.
(info_program_command): Don't pass a target_waitstatus to
get_last_target_status.
* infrun.c (init_wait_for_inferior): Use
nullify_last_target_wait_ptid.
(get_last_target_status): Handle nullptr arguments.
(nullify_last_target_wait_ptid): Clear target_last_waitstatus.
(print_stop_event): Don't pass a ptid to get_last_target_status.
(normal_stop): Don't pass a ptid to get_last_target_status.
* infrun.h (get_last_target_status, set_last_target_status): Move
comments here and update.
(nullify_last_target_wait_ptid): Declare.
* linux-fork.c (fork_load_infrun_state): Remove local extern
declaration of nullify_last_target_wait_ptid.
* linux-nat.c (get_detach_signal): Don't pass a target_waitstatus
to get_last_target_status.
Pedro Alves [Fri, 10 Jan 2020 20:05:48 +0000 (20:05 +0000)]
switch inferior/thread before calling target methods
Once each inferior has its own target stack, we'll need to make sure
that the right inferior is selected before we call into target
methods.
It kind of sounds worse than it is in practice. Not that many places
need to be concerned.
In thread.c, we add a new switch_to_thread_if_alive function that
centralizes the switching before calls to target_thread_alive. Other
cases are handled with explicit switching.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdbthread.h (scoped_restore_current_thread)
<dont_restore, restore, m_dont_restore>: Declare.
* thread.c (thread_alive): Add assertion. Return bool.
(switch_to_thread_if_alive): New.
(prune_threads): Switch inferior/thread.
(print_thread_info_1): Switch thread before calling target methods.
(scoped_restore_current_thread::restore): New, factored out from
...
(scoped_restore_current_thread::~scoped_restore_current_thread):
... this.
(scoped_restore_current_thread::scoped_restore_current_thread):
Add assertion.
(thread_apply_all_command, thread_select): Use
switch_to_thread_if_alive.
* infrun.c (proceed, restart_threads, handle_signal_stop)
(switch_back_to_stepped_thread): Switch current thread before
calling target methods.
Pedro Alves [Fri, 10 Jan 2020 20:05:47 +0000 (20:05 +0000)]
Introduce switch_to_inferior_no_thread
Several places want to switch context to an inferior and its pspace,
while at the same time switch to "no thread selected". This commit
adds a function that does that, and uses it in a few places.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* inferior.c (switch_to_inferior_no_thread): New function,
factored out from ...
(inferior_command): ... here.
* inferior.h (switch_to_inferior_no_thread): Declare.
* mi/mi-main.c (run_one_inferior): Use
switch_to_inferior_no_thread.
Pedro Alves [Fri, 10 Jan 2020 20:05:46 +0000 (20:05 +0000)]
Delete unnecessary code from kill_command
I believe this comment:
/* Killing off the inferior can leave us with a core file. If
so, print the state we are left in. */
Referred to the fact that a decade ago, by design, GDB would let you
type "run" when debugging a core dump, keeping the core open. That
"run" would push a process_stratum target on the target stack for the
live process, and, the core would remain open -- we used to have a
core_stratum. When the live process was killed/detached or exited,
GDB would go back to debugging the core, since the core_stratum target
was now at the top of the stack. That design had a number of
problems, see here for example:
https://sourceware.org/ml/gdb-patches/2008-08/msg00290.html
In 2010, core_stratum was finaly eliminated and cores now have
process_stratum too, with commit
c0edd9edadfe ("Make core files the
process_stratum."). Pushing a live process on the stack while you're
debugging a core discards the core completely.
I also thought that this might be in use with checkpoints, but it does
not -- "kill" when you have multiple checkpoints kills all the
checkpoints.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* infcmd.c (kill_command): Remove dead code.
Pedro Alves [Fri, 10 Jan 2020 20:05:45 +0000 (20:05 +0000)]
Don't check target is running in remote_target::mourn_inferior
I believe the tail end of remote_target::mourn_inferior is broken, and
it's been broken for too long to even bother trying to fix. Most
probably nobody needs it. If the code is reached and we find the
target is running, we'd need to resync the thread list, at least,
since generic_mourn_inferior got rid of all the threads in the
inferior, otherwise, we'd hit an assertion on the next call to
inferior_thread(), for example. A "correct" fix would probably
involve restarting the whole remote_target::start_remote requence,
exactly as if we had completely disconnected and reconnected from
scratch.
Note that regular stub debugging usually uses plain target remote, but
this code is only reachable in target extended-mode:
- The !remote_multi_process_p check means that it's only reacheable if
the stub does not support multi-process. I.e., there can only ever
be one live process.
- remote_target::mourn_inferior has this at the top:
/* In 'target remote' mode with one inferior, we close the connection. */
if (!rs->extended && number_of_live_inferiors () <= 1)
{
unpush_target (this);
/* remote_close takes care of doing most of the clean up. */
generic_mourn_inferior ();
return;
}
Which means that if we only had one live inferior (which for our
case, must be true), we'll have closed the connection already,
unless we're in extended-remote mode.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* remote.c (remote_target::mourn_inferior): No longer check
whether the target is running.
Pedro Alves [Fri, 10 Jan 2020 20:05:44 +0000 (20:05 +0000)]
Make target_ops::has_execution take an 'inferior *' instead of a ptid_t
With the multi-target work, each inferior will have its own target
stack, so to call a target method, we'll need to make sure that the
inferior in question is the current one, otherwise target->beneath()
calls will find the target beneath in the wrong inferior.
In some places, it's much more convenient to be able to check whether
an inferior has execution without having to switch to it in order to
call target_has_execution on the right inferior/target stack, to avoid
side effects with switching inferior/thread/program space.
The current target_ops::has_execution method takes a ptid_t as
parameter, which, in a multi-target world, isn't sufficient to
identify the target. This patch prepares to address that, by changing
the parameter to an inferior pointer instead. From the inferior,
we'll be able to query its target stack to tell which target is
beneath.
Also adds a new inferior::has_execution() method to make callers a bit
more natural to read.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* corelow.c (core_target::has_execution): Change parameter type to
inferior pointer.
* inferior.c (number_of_live_inferiors): Use
inferior::has_execution instead of target_has_execution_1.
* inferior.h (inferior::has_execution): New.
* linux-thread-db.c (thread_db_target::update_thread_list): Use
inferior::has_execution instead of target_has_execution_1.
* process-stratum-target.c
(process_stratum_target::has_execution): Change parameter type to
inferior pointer. Check the inferior's PID instead of
inferior_ptid.
* process-stratum-target.h
(process_stratum_target::has_execution): Change parameter type to
inferior pointer.
* record-full.c (record_full_core_target::has_execution): Change
parameter type to inferior pointer.
* target.c (target_has_execution_1): Change parameter type to
inferior pointer.
(target_has_execution_current): Adjust.
* target.h (target_ops::has_execution): Change parameter type to
inferior pointer.
(target_has_execution_1): Change parameter type to inferior
pointer. Change return type to bool.
* tracefile.h (tracefile_target::has_execution): Change parameter
type to inferior pointer.
Pedro Alves [Fri, 10 Jan 2020 20:05:43 +0000 (20:05 +0000)]
exceptions.c:print_flush: Remove obsolete check
Commit
20f0d60db4fb ("Avoid crash when calling warning too early"),
added a "current_top_target () != NULL" check to
target_supports_terminal_ours, so this check in exceptions.c is now
obsolete.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* exceptions.c (print_flush): Remove current_top_target() check.
Pedro Alves [Fri, 10 Jan 2020 20:05:42 +0000 (20:05 +0000)]
Make "show remote exec-file" inferior-aware
The "set remote exec-file" setting is per-inferior, but the "show
remote exec-file" command always shows the last set exec-file,
irrespective of the current inferior. E.g.:
# Set inferior 1's exec-file:
(gdb) set remote exec-file prog1
# Add inferior 2, switch to it, and set its exec-file:
(gdb) add-inferior
Added inferior 2
(gdb) inferior 2
(gdb) set remote exec-file prog2
# Switch back to inferior 1, and show its exec-file:
(gdb) inferior 1
(gdb) show remote exec-file
prog2
^^^^^ should show "prog1" instead here.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* remote.c (show_remote_exec_file): Show the current inferior's
exec-file instead of the command variable's value.
gdb/testsuite/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdb.base/remote-exec-file.exp: New file.
Pedro Alves [Fri, 10 Jan 2020 20:05:41 +0000 (20:05 +0000)]
Don't rely on inferior_ptid in record_full_wait
The multi-target patch sets inferior_ptid to null_ptid before handling
a target event, and thus before calling target_wait, in order to catch
places in target_ops::wait implementations that are incorrectly
relying on inferior_ptid (which could otherwise be a ptid of a
different target, for example). That caught this instance in
record-full.c.
Fix it by saving the last resumed ptid, and then using it in
record_full_wait_1, just like how the last "step" argument passed to
record_full_target::resume is handled too.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* record-full.c (record_full_resume_ptid): New global.
(record_full_target::resume): Set it.
(record_full_wait_1): Use record_full_resume_ptid instead of
inferior_ptid.
Pedro Alves [Fri, 10 Jan 2020 20:05:41 +0000 (20:05 +0000)]
Preserve selected thread in all-stop w/ background execution
In non-stop mode, if you resume the program in the background (with
"continue&", for example), then gdb makes sure to not switch the
current thread behind your back. That means that you can be sure that
the commands you type apply to the thread you selected, even if some
other thread that was running in the background hits some event just
while you're typing.
In all-stop mode, however, if you resume the program in the
background, gdb let's the current thread switch behind your back.
This is bogus, of course. All-stop and non-stop background
resumptions should behave the same.
This patch fixes that, and adds a testcase that exposes the bad
behavior in current master.
The fork-running-state.exp changes are necessary because that
preexisting testcase was expecting the old behavior:
Before:
continue &
Continuing.
(gdb)
[Attaching after process 8199 fork to child process 8203]
[New inferior 2 (process 8203)]
info threads
Id Target Id Frame
1.1 process 8199 "fork-running-st" (running)
* 2.1 process 8203 "fork-running-st" (running)
(gdb)
After:
continue &
Continuing.
(gdb)
[Attaching after process 24660 fork to child process 24664]
[New inferior 2 (process 24664)]
info threads
Id Target Id Frame
* 1.1 process 24660 "fork-running-st" (running)
2.1 process 24664 "fork-running-st" (running)
(gdb)
Here we see that before this patch GDB switches current inferior to
the new inferior behind the user's back, as a side effect of handling
the fork.
The delete_exited_threads call in inferior_appeared is there to fix an
issue that Baris found in a previous version of this patch. The
fetch_inferior_event change increases the refcount of the current
thread, and in case the fetched inferior event denotes a thread exit,
the thread will not be deleted right away. A non-deleted but exited
thread stays in the inferior's thread list. This, in turn, causes the
"init_thread_list" call in inferior.c to be skipped. A consequence is
that the global thread ID counter is not restarted if the current
thread exits, and then the inferior is restarted:
(gdb) start
Temporary breakpoint 1 at 0x4004d6: file main.c, line 21.
Starting program: /tmp/main
Temporary breakpoint 1, main () at main.c:21
21 foo ();
(gdb) info threads -gid
Id GId Target Id Frame
* 1 1 process 16106 "main" main () at main.c:21
(gdb) c
Continuing.
[Inferior 1 (process 16106) exited normally]
(gdb) start
Temporary breakpoint 2 at 0x4004d6: file main.c, line 21.
Starting program: /tmp/main
Temporary breakpoint 2, main () at main.c:21
21 foo ();
(gdb) info threads -gid
Id GId Target Id Frame
* 1 2 process 16138 "main" main () at main.c:21
^^^
Notice that GId == 2 above. It should have been "1" instead.
The new tids-git-reset.exp testcase exercises the problem above.
gdb/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdbthread.h (scoped_restore_current_thread)
<dont_restore, restore, m_dont_restore>: Declare.
* thread.c (thread_alive): Add assertion. Return bool.
(switch_to_thread_if_alive): New.
(prune_threads): Switch inferior/thread.
(print_thread_info_1): Switch thread before calling target methods.
(scoped_restore_current_thread::restore): New, factored out from
...
(scoped_restore_current_thread::~scoped_restore_current_thread):
... this.
(scoped_restore_current_thread::scoped_restore_current_thread):
Add assertion.
(thread_apply_all_command, thread_select): Use
switch_to_thread_if_alive.
gdb/testsuite/ChangeLog:
2020-01-10 Pedro Alves <palves@redhat.com>
* gdb.base/fork-running-state.exp (do_test): Adjust expected
output.
* gdb.threads/async.c: New.
* gdb.threads/async.exp: New.
* gdb.multi/tids-gid-reset.c: New.
* gdb.multi/tids-gid-reset.exp: New.
George Barrett [Fri, 10 Jan 2020 19:30:28 +0000 (06:30 +1100)]
Fix handling of null stap semaphores
According to the SystemTap documentation on user-space probes[0], stap
probe points without semaphores are denoted by setting the semaphore
address in the probe's note to zero. At present the code does do a
comparison of the semaphore address against zero, but only after it's
been relocated; as such it will (almost?) always fail, commonly
resulting in GDB trying to overwrite the ELF magic located at the
image's base address.
This commit tests the address as specified in the SDT note rather than
the relocated value in order to correctly detect absent probe
semaphores.
[0]: https://sourceware.org/systemtap/wiki/UserSpaceProbeImplementation
gdb/Changelog:
2020-01-11 George Barrett <bob@bob131.so>
* stap-probe.c (stap_modify_semaphore): Don't check for null
semaphores.
(stap_probe::set_semaphore, stap_probe::clear_semaphore): Check
for null semaphores.
gdb/testsuite/ChangeLog:
2020-01-11 George Barrett <bob@bob131.so>
* gdb.base/stap-probe.c (relocation_marker): Add dummy variable
to help in finding the image relocation offset.
* gdb.base/stap-probe.exp (stap_test): Accept arbitrary compile
options in arguments.
(stap_test_no_debuginfo): Likewise.
(stap-probe-nosem-noopt-pie, stap-probe-nosem-noopt-nopie): Add
test variants.
(stap_test): Add null semaphore relocation test.
George Barrett [Fri, 10 Jan 2020 19:30:01 +0000 (06:30 +1100)]
gdb/testsuite/gdb.base/stap-probe: Minor clean-up
This patch resolves a couple of issues with the test case for SystemTap
user-space probe points:
1. The preprocessor macro guarding the semaphore variables in the C
file is (rather confusingly) named USE_PROBES. This has been
renamed to USE_SEMAPHORES, to better reflect its function.
2. The test procedures in the expect file improperly pass the flag
defining USE_PROBES to prepare_for_testing; as such, the test
binary that's supposed to have probes with semaphores is the same
as the one without. This has also been fixed.
3. No test is performed to check that `info probes' returns
information about probe semaphores. Such a test is included in this
patch.
gdb/testsuite/ChangeLog
2020-01-10 George Barrett <bob@bob131.so>
* gdb.base/stap-probe.c: Rename USE_PROBES to USE_SEMAPHORES.
* gdb.base/stap-probe.exp: Likewise.
(stap_test): Pass argument as an additional flag.
(stap_test_no_debuginfo): Likewise.
(stap_test): Check `info probes stap' output for semaphore
addresses if the test binary is supposed to have them.
Szabolcs Nagy [Thu, 9 Jan 2020 17:20:56 +0000 (17:20 +0000)]
[PR ld/22269] arm: Avoid dynamic relocs for undefweak symbols in static PIE
With static PIE linking undefined weak symbols are resolved to 0, so no
dynamic relocation is needed for them. The UNDEFWEAK_NO_DYNAMIC_RELOC
macro was introduced so this case can be handled easily, but it was not
applied consistently in the first attempt to fix ld/22269 for arm:
commit
95b03e4ad68e7a90f5096b47df595636344b783a
arm: Check UNDEFWEAK_NO_DYNAMIC_RELOC
This patch fixes spurious relative relocs in static PIE binaries against
GOT entries created for undefined weak symbols on arm*-*, this fixes
FAIL: pr22269-1 (static pie undefined weak)
bfd/ChangeLog:
PR ld/22269
* elf32-arm.c (elf32_arm_final_link_relocate): Use
UNDEFWEAK_NO_DYNAMIC_RELOC.
(allocate_dynrelocs_for_symbol): Likewise.
Tamar Christina [Fri, 10 Jan 2020 13:48:57 +0000 (13:48 +0000)]
AArch64: Revert setting of elf class in linker stub.
This changes the fix to PR 25210 by removing the ELF class change.
As it turns out the correct change was only the change in compress.c.
Everything else is unneeded and setting the elf class is making the linker
behave very oddly under LTO. The first stub is correctly written out but for
the rest the suddenly don't have a pointer to the stub section anymore.
This caused SPEC to fail as the program would branch to the stub and it wouldn't
be filled in.
Committed to master under the trivial rule as this is partially reverting a previous commit.
bfd/ChangeLog:
PR 25210
* elfnn-aarch64.c (_bfd_aarch64_create_stub_section): Remove elfclass.
Alan Modra [Fri, 10 Jan 2020 06:59:59 +0000 (17:29 +1030)]
HPUX gas testsuite fixes
* testsuite/gas/elf/pr14891.s: Don't start directives in first column.
* testsuite/gas/elf/pr21661.d: Don't run on hpux.
Alan Modra [Thu, 9 Jan 2020 21:57:33 +0000 (08:27 +1030)]
ubsan: tilepro: signed integer overflow
* tilepro-opc.c (parse_insn_tilepro): Make opval unsigned.
* tilegx-opc.c (parse_insn_tilegx): Likewise. Delete raw_opval.
Alan Modra [Wed, 8 Jan 2020 20:29:42 +0000 (06:59 +1030)]
ubsan: m10300: shift exponent -4
* m10300-dis.c (disassemble): Move extraction of DREG, AREG, RREG,
and XRREG value earlier to avoid a shift with negative exponent.
* m10200-dis.c (disassemble): Similarly.
Alan Modra [Wed, 8 Jan 2020 20:14:16 +0000 (06:44 +1030)]
ubsan: spu: left shift of negative value
Also fixes a real bug. The DECODE_INSN_I9a and DECODE_INSN_I9b both
use UNSIGNED_EXTRACT for 7 low bits of the result, but this was an
unsigned value due to "insn" being unsigned. DECODE_INSN_I9* is
therefore unsigned too, leading to a zero extension in an expression
using a bfd_vma if bfd_vma is 64 bits.
* opcode/spu.h: Formatting.
(UNSIGNED_EXTRACT): Use 1u.
(SIGNED_EXTRACT): Don't sign extend with shifts.
(DECODE_INSN_I9a, DECODE_INSN_I9b): Avoid left shift of signed value.
Keep result signed.
(DECODE_INSN_U9a, DECODE_INSN_U9b): Delete.
Alan Modra [Wed, 8 Jan 2020 20:11:25 +0000 (06:41 +1030)]
ubsan: alpha-coff: signed integer overflow
* coff-alpha.c (alpha_ecoff_object_p): Calculate size in bfd_size_type.
GDB Administrator [Fri, 10 Jan 2020 00:00:24 +0000 (00:00 +0000)]
Automatic date update in version.in
Andrew Burgess [Tue, 7 Jan 2020 11:39:17 +0000 (11:39 +0000)]
gdb/tui: Link source and assembler scrolling .... again
Until recently when the source window was scrolled the assembler
window would scroll in sync - keeping the disassembly for the current
line in view.
This was broken in commit:
commit
b4b49dcbff6b437fa8b4e2fc0c3f27b457f11310
Date: Wed Nov 13 16:47:58 2019 -0700
Don't call tui_show_source from tui_ui_out
This commit restores the synchronised scrolling and also maintains the
horizontal scroll within the source view when it is vertically
scrolled, something that was broken before.
This commit does not mean that scrolling the assembler view scrolls
the source view. The connection this way never existed, though maybe
it should, but I'll leave adding this feature for a separate commit.
gdb/ChangeLog:
* tui/tui-source.c (tui_source_window::do_scroll_vertical): Update
all source windows, and maintain horizontal scroll status while
doing so.
gdb/testsuite/ChangeLog:
* gdb.tui/basic.exp: Add more scrolling tests.
Change-Id: I250114a3bc670040a6a759d41905776771b2f818
Tom Tromey [Sun, 22 Dec 2019 23:52:56 +0000 (16:52 -0700)]
gdb: Fix scrolling in TUI
Hannes Domani pointed out that my previous patch to fix the "list"
command in the TUI instead broke vertical scrolling. While looking at
this, I found that do_scroll_vertical calls print_source_lines, which
seems like a very roundabout way to change the source window. This
patch removes this oddity and fixes the bug at the same time.
I've added a new test case. This is somewhat tricky, because the
obvious approach of sending a dummy command after the scroll did not
work -- due to how the TUI works, sennding a command causes the scroll
to take effect.
gdb/ChangeLog
2019-12-22 Tom Tromey <tom@tromey.com>
PR tui/18932:
* tui/tui-source.c (tui_source_window::do_scroll_vertical): Call
update_source_window, not print_source_lines.
gdb/testsuite/ChangeLog
2019-12-22 Tom Tromey <tom@tromey.com>
PR tui/18932:
* lib/tuiterm.exp (Term::wait_for): Rename from _accept. Return a
meangingful value.
(Term::command, Term::resize): Update.
* gdb.tui/basic.exp: Add scrolling test.
Change-Id: I9636a7c8a8cade37431c6165ee996a9d556ef1c8
Andrew Burgess [Tue, 7 Jan 2020 00:41:08 +0000 (00:41 +0000)]
gdb/tui: Fix 'layout asm' before the inferior has started
Currently if a user starts the tui with 'layout asm' then they will be
presented with the 'src' layout.
What happens is:
1. Layout command enables TUI, selecting the SRC layout by default.
2. As part of tui_enable we call tui_display_main, which calls
tui_get_begin_asm_address, which calls
set_default_source_symtab_and_line. This changes core GDBs
current symtab and line, which triggers a call to the symtab
changed hook tui_symtab_changed, which sets the flag
from_source_symtab.
3. Back in the layout command, the layout is changed from SRC to
ASM. After this the layout command completes and we return to
core GDB which prints the prompt, however...
4. The before prompt hook is called which sees the
from_source_symtab flag is set and forces the SRC window to be
displayed. This switches us back to SRC view.
The solution I propose here is to delay installing the hooks into core
GDB until after we have finished setting up the tui and selecting the
default frame to view. In this way we effectively ignore the first
symtab changed event triggered when making main the default symtab.
gdb/ChangeLog:
* tui/tui.c (tui_enable): Register tui hooks after calling
tui_display_main.
gdb/testsuite/ChangeLog:
* gdb.tui/tui-layout-asm.exp: New file.
Change-Id: I858ab81a17ffb4aa72deb3f36c3755228a9c9d9a
Andrew Burgess [Tue, 7 Jan 2020 00:35:02 +0000 (00:35 +0000)]
gdb/testsuite/tui: Introduce check_box_contents
A new test procedure for matching the contents of one screen box
against a regexp. This can be used to match the contents of one TUI
window against a regexp without any of the borders, or other windows
being included in the matched output (as is currently the case with
check_contents).
This will be used in a later commit.
gdb/testsuite/ChangeLog:
* lib/tuiterm.exp (Term::check_box_contents): New proc.
Change-Id: Icf795bf38dd9295e282a34eecc318a9cdbc73926
Andrew Burgess [Tue, 7 Jan 2020 00:30:35 +0000 (00:30 +0000)]
gdb/testsuite/tui: Split enter_tui into two procs
Split Term::enter_tui into two procedures, a core which does the
setup, but doesn't actually enable tui mode, and the old enter_tui
that calls the new core, and then enables tui mode.
This is going to be useful in a later commit.
gdb/testsuite/ChangeLog:
* lib/tuiterm.exp (Term::prepare_for_tui): New proc.
(Term::enter_tui): Use Term::prepare_for_tui.
Change-Id: I501dfb2ddaa4a4e7246a5ad319ab428e4f42b3af
Andrew Burgess [Tue, 7 Jan 2020 00:26:22 +0000 (00:26 +0000)]
gdb/testsuite/tui: Always dump_screen when asked
The Term::dump_screen routine currently dumps the screen using calls
to 'verbose', this means it will only dump the screen when the
testsuite is running in verbose mode.
However, the Term::dump_screen is most often called when a test fails,
in this case I think it is useful to have the screen dumped even when
we're not in verbose mode.
This commit changes the calls to 'verbose' to be 'verbose -log' so we
always get the screen dump.
gdb/testsuite/ChangeLog:
* lib/tuiterm.exp (Term::dump_screen): Always dump the screen when
called.
Change-Id: I5f0a7f5ac2ece04d6fe6e9c5a28ea2a0dda38955
Andrew Burgess [Thu, 9 Jan 2020 22:38:22 +0000 (22:38 +0000)]
gdb/testsuite: Fix race condition in gdb.base/skip.exp
In this commit:
commit
5024637fac653914d471808288dc3221bc7ec089
Date: Sun Dec 15 11:05:47 2019 +0100
Fix skip.exp test failure observed with gcc-9.2.0
A race condition was introduced into the gdb.base/skip.exp test when
this line:
gdb_test "step" "foo \\(\\) at.*" "step 3"
Was changed to this:
gdb_test "step" "foo \\(\\) at.*" "step 3" "main \\(\\) at .*" "step"
Before the above change we expected GDB to behave like this:
(gdb) step
foo () at /path/to/gdb/testsuite/gdb.base/skip.c:42
42 return 0;
(gdb)
However, when the test is compiled with GCC 9.2.0 we get a different
behaviour, and so we need a second 'step', like this:
(gdb) step
main () at /path/to/gdb.base/skip.c:32
32 x = baz ((bar (), foo ()));
(gdb) step
foo () at /path/to/gdb/testsuite/gdb.base/skip.c:42
42 return 0;
(gdb)
Now the change to the test matches against 'main () at .*', however if
GDB or expect is being slow then we might only get to see output like
this:
(gdb) step
main () at /path/to/g
This will happily match the question pattern, so we send 'step' to GDB
again. Now GDB continues to produce output which expect accepts, we
now see this:
b.base/skip.c:32
32 x = baz ((bar (), foo ()));
(gdb)
This has carried on from where the previous block of output left off.
This doesn't match the final pattern 'foo \\(\\) at.*', but it does
match the prompt pattern that gdb_test_multiple adds, and so we report
the test as failing.
The solution is to simply ensure that the question consumes everything
up to, and including the prompt. This ensures that the prompt can't
then match the failure case. The new test line becomes:
gdb_test "step" "foo \\(\\) at.*" "step 3" \
"main \\(\\) at .*\r\n$gdb_prompt " "step"
gdb/testsuite/ChangeLog:
* gdb.base/skip.exp: Fix race condition in test.
Change-Id: I9f0b0b52ef1b4f980bfaa8fe405ff06d520f3482
Christian Biesinger [Wed, 18 Dec 2019 18:06:43 +0000 (12:06 -0600)]
Don't define _FORTIFY_SOURCE on MinGW
Recent MinGW versions require -lssp when using _FORTIFY_SOURCE, which
gdb does (in common-defs.h)
https://github.com/msys2/MINGW-packages/issues/5868#issuecomment-
544107564
To avoid all the complications with checking for -lssp and making sure it's
linked statically, just don't define it.
gdb/ChangeLog:
2020-01-09 Christian Biesinger <cbiesinger@google.com>
* gdbsupport/common-defs.h: Don't define _FORTIFY_SOURCE on MinGW.
Change-Id: Ide6870ab57198219a2ef78bc675768a789ca2b1d
Simon Marchi [Wed, 8 Jan 2020 21:55:03 +0000 (16:55 -0500)]
Fix indentation in print_thread_info_1
The body of this this big "for" loop is missing an indentation level,
this patch fixes that.
gdb/ChangeLog:
* thread.c (print_thread_info_1): Fix indentation.
Christian Biesinger [Wed, 8 Jan 2020 01:10:40 +0000 (19:10 -0600)]
Fix memory leak of the demangled symbol name
compute_and_set_names would only free the name if we did not find the name
in the hashtable, but it needs to always free it. Solve this by moving the
smart pointer outside the if.
Thanks to PhilippeW for finding this.
gdb/ChangeLog:
2020-01-09 Christian Biesinger <cbiesinger@google.com>
* symtab.c (general_symbol_info::compute_and_set_names): Move the
unique_xmalloc_ptr outside the if to always free the demangled name.
Change-Id: Id7c6b8408432183700ccb5ff634818d6c5a3ac95
Nick Clifton [Thu, 9 Jan 2020 16:51:04 +0000 (16:51 +0000)]
Fix an attempt to free a static pointer when using objcopy's symbol addition feature.
PR 25220
* objcopy.c (empty_name): New variable.
(need_sym_before): Prevent an attempt to free a static variable.
(filter_symbols): Avoid strcmp test by checking for pointer
equality.
Nick Clifton [Thu, 9 Jan 2020 15:49:08 +0000 (15:49 +0000)]
Fix an illegal memory access triggered when trying to examine an input file containing corrupt compressed sections.
PR 25221
* bfd.c (bfd_convert_section_contents): Check for a compress
header size that is larger than the actual section size.
Nick Clifton [Thu, 9 Jan 2020 14:32:49 +0000 (14:32 +0000)]
Fix the cast used to prevent compile time warning about an always false test.
PR 25224
* z80-dis.c (ld_ii_ii): Use correct cast.
Aaron Merey [Thu, 9 Jan 2020 13:37:26 +0000 (13:37 +0000)]
oops - toplevel changelog entry for previous delta.
* config/debuginfod.m4: New file. Add macro AC_DEBUGINFOD. Adds
new configure option --with-debuginfod.
* configure: Regenerate.
* configure.ac: Call AC_DEBUGINFOD.
Nick Clifton [Thu, 9 Jan 2020 13:36:22 +0000 (13:36 +0000)]
Update Traditional Chinese translation for the binutils sub-directory.
* po/zh_TW.po: Updated Traditional Chinese translation.
Aaron Merey [Thu, 9 Jan 2020 13:19:20 +0000 (13:19 +0000)]
Add support for debuginfod to the binutils (disable by default, enabled via a configure time option).
debuginfod is a lightweight web service that indexes ELF/DWARF
debugging resources by build-id and serves them over HTTP. This patch
enables objdump and readelf to query debuginfod servers when they are
otherwise not able to find separate debug files. Binutils can be built
with debuginfod using the --with-debuginfod configure option. This
requires that libdebuginfod be installed and found at configure time.
debuginfod is packaged with elfutils, starting with version 0.178. For
more information see https://sourceware.org/elfutils/.
toplevel* config/debuginfod.m4: New file. Add macro AC_DEBUGINFOD. Adds
new configure option --with-debuginfod.
* configure: Regenerate.
* configure.ac: Call AC_DEBUGINFOD.
binutils* Makefile.am (readelf_LDADD, objdump_LDADD): Add libdebuginfod.
* Makefile.in: Regenerate.
* NEWS: Update.
* config.in: Regenerate.
* configure: Regenerate.
* configure.ac: Call AC_DEBUGINFOD.
* doc/Makefile.in: Regenerate.
* doc/binutils.texi: Add section on using binutils
with debuginfod.
* dwarf.c (debuginfod_fetch_separate_debug_info): New function.
Query debuginfod servers for the target debug file.
(load_separate_debug_info): Call
debuginfod_fetch_separate_debug_info if configured with
debuginfod.
(load_separate_debug_files): Add file argument to
load_separate_debug_info calls.
* dwarf.h (get_build_id): Add declaration.
* objdump.c (get_build_id): New function. Get build-id of file.
* readelf.c (get_build_id): Likewise.
* testsuite/binutils-all/debuginfod.exp: New tests.
* testsuite/binutils-all/linkdebug.s: Add .note.gnu.build-id
section.
Sergey Belyashov [Thu, 9 Jan 2020 11:47:44 +0000 (11:47 +0000)]
Fix compile time warnings about comparisons always being false.
PR 25224
gas * config/tc-z80.c (emit_ld_m_rr): Use integer types when checking
opcode byte values.
(emit_ld_r_r): Likewise.
(emit_ld_rr_m): Likewise.
(emit_ld_rr_nn): Likewise.
opcodes * z80-dis.c (ld_ii_ii): Use character constant when checking
opcode byte value.
Jan Beulich [Thu, 9 Jan 2020 10:40:04 +0000 (11:40 +0100)]
x86: refine when to trigger optimizations
Checking just the base opcode without also checking this isn't a VEX
encoding, and without there being other insn properties avoiding a match
once respective VEX/XOP/EXEX-encoded insns would appear, is at least
dangerous. Add respective checks. At the same time there's no real need
to check the extension opcode to be None for the 0xA8 form - there's
nothing it can be confused with, and non-VEX-and-alike forms also can't
appear.
Jan Beulich [Thu, 9 Jan 2020 10:39:33 +0000 (11:39 +0100)]
x86-64: assert sane internal state for REX conversions
For the comments about "hi" registers to be really applicable, RegRex
may not be set on the respective registers. Assert this is the case.
Jan Beulich [Thu, 9 Jan 2020 10:38:59 +0000 (11:38 +0100)]
x86: consistently convert to byte registers for TEST w/ imm optimization
Commit
ac0ab1842d ("i386: Also check R12-R15 registers when optimizing
testq to testb") didn't go quite far enough: In order to avoid confusing
other code registers would better be converted to byte ones uniformly.
Jan Beulich [Thu, 9 Jan 2020 10:38:01 +0000 (11:38 +0100)]
x86: SYSENTER/SYSEXIT are unavailable in 64-bit mode on AMD
The disassembler change is such that in default mode we'd disassemble
the insns (for there not ebing any conflicts), but when AMD64 mode was
explicitly requested, we'd show them as "(bad)".
GDB Administrator [Thu, 9 Jan 2020 00:00:29 +0000 (00:00 +0000)]
Automatic date update in version.in
Tom Tromey [Mon, 6 Jan 2020 21:34:52 +0000 (14:34 -0700)]
Change section_offsets to a std::vector
This changes section_offsets to be specialization of a std::vector and
updates all the users. It also removes the ANOFFSET and
SIZEOF_N_SECTION_OFFSETS macros.
Most of this is just a generic sort of cleanup, that reduces the
number of lines of code. However, a couple spots were doing weird
things.
objfile_relocate did:
- std::vector<struct section_offsets>
- new_debug_offsets (SIZEOF_N_SECTION_OFFSETS (debug_objfile->num_sections));
... which seems to greatly over-estimate the number of elements
needed.
This appeared in set_objfile_default_section_offset:
- std::vector<struct section_offsets> offsets (objf->num_sections,
- { { offset } });
... which makes sense due to type safety, but is also actively
confusing given that section_offsets was previously also a kind of
vector type.
Tested on x86-64 Fedora 30.
gdb/ChangeLog
2020-01-08 Tom Tromey <tromey@adacore.com>
* xcoffread.c (enter_line_range, read_xcoff_symtab)
(process_xcoff_symbol, xcoff_symfile_offsets): Update.
* symtab.h (MSYMBOL_VALUE_ADDRESS): Update.
(struct section_offsets, ANOFFSET, SIZEOF_N_SECTION_OFFSETS):
Remove.
(section_offsets): New typedef.
* symtab.c (fixup_section, get_msymbol_address): Update.
* symmisc.c (dump_msymbols): Update.
* symfile.h (relative_addr_info_to_section_offsets)
(symfile_map_offsets_to_segments): Update.
* symfile.c (build_section_addr_info_from_objfile)
(init_objfile_sect_indices): Update.
(struct place_section_arg): Change type of "offsets".
(place_section): Update.
(relative_addr_info_to_section_offsets): Change type of
"section_offsets". Remove "num_sections" parameter.
(default_symfile_offsets, syms_from_objfile_1)
(set_objfile_default_section_offset): Update.
(reread_symbols): No need to preserve section offsets by hand.
(symfile_map_offsets_to_segments): Change type of "offsets".
* stap-probe.c (relocate_address): Update.
* stabsread.h (process_one_symbol): Update.
* solib-target.c (struct lm_info_target) <offsets>: Change type.
(solib_target_relocate_section_addresses): Update.
* solib-svr4.c (enable_break, svr4_relocate_main_executable):
Update.
* solib-frv.c (frv_relocate_main_executable): Update.
* solib-dsbt.c (dsbt_relocate_main_executable): Update.
* solib-aix.c (solib_aix_get_section_offsets): Change return
type.
(solib_aix_solib_create_inferior_hook): Update.
* remote.c (remote_target::get_offsets): Update.
* psymtab.c (find_pc_sect_psymtab): Update.
* psympriv.h (struct partial_symbol) <address, text_low,
text_high>: Update.
* objfiles.h (obj_section_offset): Update.
(struct objfile) <section_offsets>: Change type.
<num_sections>: Remove.
(objfile_relocate): Update.
* objfiles.c (entry_point_address_query): Update
(relocate_one_symbol): Change type of "section_offsets".
(objfile_relocate1, objfile_relocate1): Change type of
"new_offsets".
(objfile_rebase1): Update.
* mipsread.c (mipscoff_symfile_read): Update.
(read_alphacoff_dynamic_symtab): Remove "section_offsets"
parameter.
* mdebugread.c (parse_symbol): Change type of "section_offsets".
(parse_external, psymtab_to_symtab_1): Update.
* machoread.c (macho_symfile_offsets): Update.
* ia64-tdep.c (ia64_find_unwind_table): Update.
* hppa-tdep.c (read_unwind_info): Update.
* hppa-bsd-tdep.c (hppabsd_find_global_pointer): Update.
* dwarf2read.c (create_addrmap_from_index)
(create_addrmap_from_aranges, dw2_find_pc_sect_compunit_symtab)
(process_psymtab_comp_unit_reader, add_partial_symbol)
(add_partial_subprogram, process_full_comp_unit)
(read_file_scope, read_func_scope, read_lexical_block_scope)
(read_call_site_scope, dwarf2_rnglists_process)
(dwarf2_ranges_process, dwarf2_ranges_read)
(dwarf_decode_lines_1, var_decode_location, new_symbol)
(dwarf2_fetch_die_loc_sect_off, dwarf2_per_cu_text_offset):
Update.
* dwarf2-frame.c (execute_cfa_program, dwarf2_frame_find_fde):
Update.
* dtrace-probe.c (dtrace_probe::get_relocated_address): Update.
* dbxread.c (read_dbx_symtab, read_ofile_symtab): Update.
(process_one_symbol): Change type of "section_offsets".
* ctfread.c (get_objfile_text_range): Update.
* coffread.c (coff_symtab_read, enter_linenos)
(process_coff_symbol): Update.
* coff-pe-read.c (add_pe_forwarded_sym): Update.
* amd64-windows-tdep.c (amd64_windows_find_unwind_info): Update.
Change-Id: I147eb967e9b44d82f4048039de7bb44b80cd72fb
Tom Tromey [Thu, 31 Oct 2019 16:46:18 +0000 (10:46 -0600)]
Use std::string in dwarf2read.c
This replaces two instances of manual string management in
dwarf2read.c with std::string.
gdb/ChangeLog
2020-01-08 Tom Tromey <tromey@adacore.com>
* dwarf2read.c (parse_macro_definition): Use std::string.
(parse_macro_definition): Likewise.
Change-Id: Iec437100105484aa4a116fb5d651d7ed52ee9d81
Tom Tromey [Thu, 31 Oct 2019 16:37:35 +0000 (10:37 -0600)]
Use std::vector in abbrev_table_read_table
This removes some manual memory management from
abbrev_table_read_table, replacing it with a std::vector.
gdb/ChangeLog
2020-01-08 Tom Tromey <tromey@adacore.com>
* dwarf2read.c (abbrev_table_read_table): Use std::vector.
(ATTR_ALLOC_CHUNK): Remove.
Change-Id: I0b0e70ac2281d89a78f4d6a642700c9f0506871d
Tom Tromey [Thu, 31 Oct 2019 16:34:40 +0000 (10:34 -0600)]
Use unique_xmalloc_ptr in fixup_go_packaging
This changes fixup_go_packaging to use unique_xmalloc_ptr. I kept
this patch separate as it is slightly more complicated than the
previous changes.
gdb/ChangeLog
2020-01-08 Tom Tromey <tromey@adacore.com>
* dwarf2read.c (fixup_go_packaging): Use unique_xmalloc_ptr.
Change-Id: I0c553d0c6579db478c27bc40fc21133a61e1a4d9
Tom Tromey [Thu, 31 Oct 2019 16:31:28 +0000 (10:31 -0600)]
Remove some explicit memory management from dwarf2read.c
I noticed a few spots in dwarf2read.c that could be improved by moving
to unique_xmalloc_ptr or, in one case, std::vector.
gdb/ChangeLog
2020-01-08 Tom Tromey <tromey@adacore.com>
* dwarf2read.c (add_partial_symbol): Use unique_xmalloc_ptr.
(dwarf2_compute_name, open_dwo_file): Likewise.
(process_enumeration_scope): Use std::vector.
(guess_partial_die_structure_name): Use unique_xmalloc_ptr.
(partial_die_info::fixup, dwarf2_start_subfile)
(guess_full_die_structure_name, dwarf2_name): Likewise.
(determine_prefix): Update.
(guess_full_die_structure_name): Make return type const.
(partial_die_full_name): Return unique_xmalloc_ptr.
(DW_FIELD_ALLOC_CHUNK): Remove.
Change-Id: I1cb278c608041ef36ef1f77c7e7565c921038d08
Nick Clifton [Wed, 8 Jan 2020 17:00:54 +0000 (17:00 +0000)]
Document the fact that the assembler's alignment pseudo-ops can be issued without any argumemtns.
PR 25284
* doc/as.texi (Align): Document the fact that all arguments can be
omitted.
(Balign): Likewise.
(P2align): Likewise.
Nick Clifton [Wed, 8 Jan 2020 16:30:20 +0000 (16:30 +0000)]
Make the assembler generate an error if there is an attempt to define a section with the same name as an already defined symbol.
PR 14891
* config/obj-elf.c (obj_elf_section): Fail if the section name is
already defined as a different symbol type.
* testsuite/gas/elf/pr14891.s: New test source file.
* testsuite/gas/elf/pr14891.d: New test driver.
* testsuite/gas/elf/pr14891.s: New test expected error output.
* testsuite/gas/elf/elf.exp: Run the new test.
Alan Modra [Wed, 8 Jan 2020 01:12:36 +0000 (11:42 +1030)]
ubsan: z8k: index 10 out of bounds for type 'unsigned int const[10]'
The fix is the additional ARRAY_SIZE test, the rest just tidies
variable types rather than adding a cast to avoid warnings.
opcodes/
* z8k-dis.c: Include libiberty.h
(instr_data_s): Make max_fetched unsigned.
(z8k_lookup_instr): Make nibl_index and tabl_index unsigned.
Don't exceed byte_info bounds.
(output_instr): Make num_bytes unsigned.
(unpack_instr): Likewise for nibl_count and loop.
* z8kgen.c (gas <opcode_entry_type>): Make noperands, length and
idx unsigned.
* z8k-opc.h: Regenerate.
gas/
* config/tc-z8k.c (md_begin): Make idx unsigned.
(get_specific): Likewise for this_index.
Alan Modra [Wed, 8 Jan 2020 01:06:01 +0000 (11:36 +1030)]
PR25351 .ARM.attributes not found for symbol
PR 25351
* elflink.c (bfd_elf_final_link): Call _bfd_fix_excluded_sec_syms
after removing sections.
GDB Administrator [Wed, 8 Jan 2020 00:00:28 +0000 (00:00 +0000)]
Automatic date update in version.in
Tom Tromey [Thu, 19 Dec 2019 18:35:22 +0000 (11:35 -0700)]
Add constructor to stap_static_probe_ops
PR build/24937 concerns an error given by the clang provided by a
particular version of macOS. In particular, it reports
error: default initialization of an object of const type 'const
stap_static_probe_ops' without a user-provided default constructor
Although (at least according to sources I found online) this was
resolved as a bug in the standard, it seemed simple enough to work
around this.
Given that this is a trivial build fix, I think it should go on the
gdb 9 branch as well.
gdb/ChangeLog
2020-01-07 Tom Tromey <tromey@adacore.com>
PR build/24937:
* stap-probe.c (class stap_static_probe_ops): Add constructor.
Change-Id: I18f180c17850f420e9b66afc67f9cb3d8dceb0b3
Jon Turney [Sun, 29 Dec 2019 22:26:41 +0000 (22:26 +0000)]
Enable styling by default on Cygwin
Cygwin meets the expectations of gdb for styling (if TERM is set and not
'DUMB', the terminal supports 'ANSI' (ECMA-48) escape sequences.
gdb/ChangeLog:
2020-01-02 Jon Turney <jon.turney@dronecode.org.uk>
* cli/cli-style.c: Set cli_styling to 'true' in the Cygwin build.
Claudiu Zissulescu [Tue, 7 Jan 2020 13:29:16 +0000 (15:29 +0200)]
[ARC] Improve parsing instruction operands.
We use gas' expression function to parse the operands of an
instruction in a generic way. There are situations when we have labels
and registers having the same name as well as the substraction sign
doesn't always stands for the arithmetical operation but for the
register range (e.g. enter instruction). This patch improves parsing
symbols found in a instruction operand, cleans up code and avoids
using default or undefined variables.
gas/
xxxx-xx-xx Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_reloc_symbol): New function.
(tokenize_arguments): Clean up, use parse_reloc_symbol function.
(md_operand): Set X_md to absent.
(arc_parse_name): Check for X_md.
Shahab Vahedi [Tue, 7 Jan 2020 13:25:15 +0000 (15:25 +0200)]
[ARC] Add finer details for LLOCK and SCOND
This patch changes the "class" of LLOCK/SCOND from "MEMORY" to
"LLOCK/SCOND" respectively. Moreover, it corrects the "data_size_mode".
These changes are necessary for GDB's atmoic sequence handler.
Signed-off-by: Shahab Vahedi <shahab@synopsys.com>
GDB Administrator [Tue, 7 Jan 2020 00:00:19 +0000 (00:00 +0000)]
Automatic date update in version.in
Jim Wilson [Mon, 6 Jan 2020 23:34:50 +0000 (15:34 -0800)]
RISC-V: Fix weak function call reloc overflow on llvm build.
bfd/
PR 25205
* elfnn-riscv.c (riscv_elf_relocate_section) <R_RISCV_CALL>: Add
check for !bfd_link_pic (info).
<R_RISCV_CALL_PLT>: Move next to R_RISCV_CALL.
<R_RISCV_JAL>: Add comment.
(_bfd_riscv_relax_section): For plt.offset check, add check for
bfd_link_pic (info). Add comment.
Change-Id: Ie769bc3d5adf096a51df5cc12efe3d50e80acb8f
Alan Modra [Mon, 6 Jan 2020 22:59:54 +0000 (09:29 +1030)]
Re: Prefer object over notype symbols when disassembling
Reverts unnecessary mips testsuite changes made by commit
660df28acf.
* testsuite/ld-mips-elf/eh-frame5.s,
* testsuite/ld-mips-elf/ehdr_start-new.s,
* testsuite/ld-mips-elf/ehdr_start-o32.s,
* testsuite/ld-mips-elf/mips16-call-global-1.s,
* testsuite/ld-mips-elf/mips16-intermix-1.s,
* testsuite/ld-mips-elf/mips16-pic-1b.s,
* testsuite/ld-mips-elf/mips16-pic-4c.s,
* testsuite/ld-mips-elf/no-shared-1-n64.s,
* testsuite/ld-mips-elf/no-shared-1-o32.s,
* testsuite/ld-mips-elf/pic-and-nonpic-1b-micromips.s,
* testsuite/ld-mips-elf/pic-and-nonpic-1b.s,
* testsuite/ld-mips-elf/pic-and-nonpic-2a.s,
* testsuite/ld-mips-elf/pic-and-nonpic-3b.s,
* testsuite/ld-mips-elf/pic-and-nonpic-4b.s,
* testsuite/ld-mips-elf/pic-and-nonpic-5a.s,
* testsuite/ld-mips-elf/pic-and-nonpic-6-n32c.s,
* testsuite/ld-mips-elf/pic-and-nonpic-6-n64c.s,
* testsuite/ld-mips-elf/pic-and-nonpic-6-o32c.s,
* testsuite/ld-mips-elf/pie.s,
* testsuite/ld-mips-elf/relax-jalr.s: Revert 2019-12-17 change.
Andrew Burgess [Thu, 26 Dec 2019 20:56:01 +0000 (20:56 +0000)]
gdb: Fix backtrace with disassemble-next-line on
In this commit:
commit
ec8e2b6d3051f0b4b2a8eee9917898e95046c62f
Date: Fri Jun 14 23:43:00 2019 +0100
gdb: Don't allow annotations to influence what else GDB prints
A change was accidentally made that moved a call to do_gdb_disassembly
out of an if block guarded by 'if (source_print && sal.symtab)'. The
result was that if a user has 'set disassemble-next-line on' then the
backtrace would now include some disassembly of a few instructions in
each frame.
This change was not intentional, but was not spotted by any tests.
This commit restores the old behaviour and adds a test to ensure this
doesn't break again in the future.
gdb/ChangeLog:
* stack.c (print_frame_info): Move disassemble_next_line code
inside source_print block.
gdb/testsuite/ChangeLog:
* gdb.base/backtrace.c: New file.
* gdb.base/backtrace.exp: New file.
Change-Id: I47c52a202fa74be138382646b695827940178689
Eli Zaretskii [Mon, 6 Jan 2020 19:54:21 +0000 (21:54 +0200)]
Fix MinGW native compilation of gdb/gdbsupport/gdb_wait.c
gdb/ChangeLog
2020-01-06 Eli Zaretskii <eliz@gnu.org>
* gdbsupport/gdb_wait.c: Include <signal.h> instead of
gdb/signals.h, as we are now using native signal symbols.
Shahab Vahedi [Mon, 6 Jan 2020 14:27:32 +0000 (15:27 +0100)]
GDB: Fix the overflow in addr/line_is_displayed()
In tui_disasm_window::addr_is_displayed(), there can be situations
where "content" is empty. For instance, it can happen when the
"content" was not filled in tui_disasm_window::set_contents(),
because tui_disassemble() threw an exception. Usually this exception
is the result of fetching invalid PC addresses like the ones beyond
the end of the program.
Having "content.size ()" zero leads to an overflow in this condition
check inside tui_disasm_window::addr_is_displayed():
int i = 0;
while (i < content.size () - threshold ...) {
... content[i] ...
}
"threshold" is 2 and there are times that "content.size ()" is 0.
This results into an overflow and the loop is entered whereas it
should have been skipped. Finally, "content[i]" access leads to
a segmentation fault.
Same problem applies to tui_source_window::line_is_displayed().
The issue has been discussed at length in bug 25345:
https://sourceware.org/bugzilla/show_bug.cgi?id=25345
This commit avoids the segmentation faults with an early check:
if (content.size () < SCROLL_THRESHOLD)
return false;
Moreover, those functions have been overhauled to a leaner code.
gdb/ChangeLog:
2020-01-06 Shahab Vahedi <shahab@synopsys.com>
* tui/tui-disasm.c (tui_disasm_window::addr_is_displayed): Avoid
overflow by an early check of content vs threshold.
* tui/tui-source.c (tui_source_window::line_is_displayed):
Likewise.
Eli Zaretskii [Mon, 6 Jan 2020 19:25:23 +0000 (21:25 +0200)]
Mention the recent fix of $_exitsignal on MS-Windows in gdb/NEWS
gdb/ChangeLog
2020-01-06 Eli Zaretskii <eliz@gnu.org>
* NEWS: Mention the recent fix of $_exitsignal on MS-Windows.
Jon Turney [Thu, 2 Jan 2020 00:25:56 +0000 (00:25 +0000)]
Fix a crash with a malformed PE header
Don't try to read the PE export table when no section contains the RVA
for it.
(I have a PE executable [1] packed with UPX, where the export table data
directory entry contains a RVA which doesn't correspond to any section.
Mistakenly trying to debug this with gdb makes it crash.)
[1] https://cygwin.com/setup/setup-2.898.x86_64.exe
gdb/ChangeLog:
2020-01-02 Jon Turney <jon.turney@dronecode.org.uk>
* coff-pe-read.c (read_pe_exported_syms): Don't try to read the
export table if no section contains it's RVA.
Eli Zaretskii [Mon, 6 Jan 2020 18:22:15 +0000 (20:22 +0200)]
Fix a typo in gdb/windows-tdep.c
gdb/ChangeLog
2020-01-06 Eli Zaretskii <eliz@gnu.org>
* windows-tdep.c: Fix a typo in WINDOWS_SIGABRT.
Nick Clifton [Mon, 6 Jan 2020 16:24:23 +0000 (16:24 +0000)]
Add -fcommon compiler command line option to linker tests that need common symbols.
PR 25327
* testsuite/ld-elf/shared.exp: Add -fcommon option to compiler
command line when building libcomm1.o and pr13250 tests.
* testsuite/ld-plugin/lto.exp: Likewise for pr20267 tests.
Hannes Domani [Sun, 22 Dec 2019 13:58:22 +0000 (14:58 +0100)]
Fix search in TUI
The variable last_line_listed is never set when print_source_lines_base is
called in TUI mode, so the search always started from the last line printed
outside of TUI mode.
gdb/ChangeLog:
2020-01-06 Hannes Domani <ssbssa@yahoo.de>
* source.c (print_source_lines_base): Set last_line_listed.
Nick Clifton [Mon, 6 Jan 2020 14:00:34 +0000 (14:00 +0000)]
Fix test of the fix for PR19579 when compiling with gcc-10.
PR 25326
* testsuite/ld-elf/shared.exp: Add -fcommon option to compiler
command line when building pr19579 binaries.
Shahab Vahedi [Mon, 6 Jan 2020 12:54:18 +0000 (13:54 +0100)]
GDB: Remove trailing spaces in tui-disasm.c
A few trailing spaces are removed.
gdb/ChangeLog:
2020-01-06 Shahab Vahedi <shahab@synopsys.com>
* tui/tui-disasm.c: Remove trailing spaces.
Eli Zaretskii [Mon, 6 Jan 2020 11:51:54 +0000 (11:51 +0000)]
Improve process exit status macros on MinGW
When a Windows program is terminated by a fatal exception, its exit
code is the value of that exception, as defined by the various
EXCEPTION_* symbols in the Windows API headers. This commit emulates
WTERMSIG etc. by translating the fatal exception codes to more-or-less
equivalent Posix signals.
gdb/ChangeLog:
2020-01-06 Eli Zaretskii <eliz@gnu.org>
Pedro Alves <palves@redhat.com>
* Makefile.in (COMMON_SFILES): Add gdbsupport/gdb_wait.c.
* windows-tdep.c: New enumeration of WINDOWS_SIG* signals.
(windows_gdb_signal_to_target): New function, uses the above
enumeration to convert GDB internal signal codes to equivalent
Windows codes.
(windows_init_abi): Call set_gdbarch_gdb_signal_to_target.
* windows-nat.c: Include "gdb_wait.h".
(get_windows_debug_event): Extract the fatal exception from the
exit status and convert to the equivalent Posix signal number.
* cli/cli-cmds.c (exit_status_set_internal_vars): Account for the
possibility that WTERMSIG returns GDB_SIGNAL_UNKNOWN.
* gdbsupport/gdb_wait.c: New file, implements
windows_status_to_termsig.
* gdbsupport/gdb_wait.h (WIFEXITED, WIFSIGNALED, WEXITSTATUS)
(WTERMSIG) [__MINGW32__]: Separate definitions for MinGW.
gdb/gdbserver/ChangeLog:
2020-01-06 Eli Zaretskii <eliz@gnu.org>
Pedro Alves <palves@redhat.com>
* win32-low.c (get_child_debug_event): Extract the fatal exception
from the exit status and convert to the equivalent Posix signal
number.
(win32_wait): Allow TARGET_WAITKIND_SIGNALLED status as well.
* Makefile.in (OBS, SFILES): Add gdb_wait.[co].
Alan Modra [Mon, 6 Jan 2020 09:21:22 +0000 (19:51 +1030)]
bfd_check_format: ignore errors from coff_real_object_p
Since 1993-11-05 git commit
c188b0bec3b, bfd_check_format has failed
if any of the target object_p functions returns false with any error
but bfd_error_wrong_format. That's just weird. There is really no
reason why coff_real_object_p should be fixed to only return that
error instead of numerous other possible errors. Even an out of
memory condition for one target doesn't necessarily mean other targets
can't match, assuming the failing target nicely returns all memory it
might have used.
* format.c (bfd_check_format_matches): Ignore bfd_error on target
match failures. Don't init to bfd_error_wrong_format before
calling _bfd_check_format.