Separate the name of the igen program from the options used to run it.
This allows us to avoid duplicating ../igen/igen in Makefiles and reuse
the existing setting in the common Makefile. This also allows us to
easily harmonize the use of EXEEXT between igen/local.mk and the common
makefiles when cross-compiling for e.g. Windows.
Building GDB with current git (future 13) Clang runs into these two
issues:
#1:
src/gdb/symtab.h:1139:3: error: definition of implicit copy assignment operator for 'symbol' is deprecated because it has a user-declared copy constructor [-Werror,-Wdeprecated-copy]
symbol (const symbol &) = default;
^
#2:
src/gdb/dwarf2/read.c:834:23: error: definition of implicit copy constructor for 'partial_die_info' is deprecated because it has a user-declared copy assignment operator [-Werror,-Wdeprecated-copy]
partial_die_info& operator=(const partial_die_info& rhs) = delete;
^
Fix them by adding the explicit defaulted versions of copy ctor and
copy-assign op appropriately.
gdb/ChangeLog:
yyyy-mm-dd Pedro Alves <pedro@palves.net>
* dwarf2/read.c (struct partial_die_info): Add defaulted copy
ctor.
* symtab.h (struct symbol): Add defaulted copy assignment
operator.
Compiling GDB with current git Clang (future 13) runs into this:
src/gdb/completer.c:287:18: error: variable 'found_quote' set but not used [-Werror,-Wunused-but-set-variable]
int scan, end, found_quote, delimiter, pass_next, isbrk;
^
gdb_rl_find_completion_word came to life as a modified (stripped down)
version of readline's internal _rl_find_completion_word function.
When I added it, I don't remember whether I realized that
'found_quote' wasn't really necessary. Maybe I kept it thinking of
keeping the source code in sync with readline? I don't recall
anymore. Since the function is already stripped down compared to the
original, stripping it down some more doesn't hurt.
So fix the issue by removing the unnecessary code.
gdb/ChangeLog:
yyyy-mm-dd Pedro Alves <pedro@palves.net>
* completer.c (RL_QF_SINGLE_QUOTE, RL_QF_DOUBLE_QUOTE)
(RL_QF_BACKSLASH, RL_QF_OTHER_QUOTE): Delete.
(gdb_rl_find_completion_word): Remove write-only 'found_quote'
local.
Since GDB is written in C++ now, we don't need struct/union typedefs
any more. Remove them from nat/amd64-linux-siginfo.c.
gdb/ChangeLog:
yyyy-mm-dd Pedro Alves <pedro@palves.net>
* nat/amd64-linux-siginfo.c (union nat_sigval): Rename to ...
(nat_sigval_t): ... this and remove typedef of same name.
(struct nat_siginfo): Rename to ...
(nat_siginfo_t): ... this and remove typedef of same name.
(struct compat_sigval): Rename to ...
(compat_sigval_t): ... this and remove typedef of same name.
(struct compat_siginfo): Rename to ...
(compat_siginfo_t): ... this and remove typedef of same name.
(struct compat_x32_siginfo): Rename to ...
(compat_x32_siginfo_t): ... this and remove typedef of same name.
(amd64_linux_siginfo_fixup_common): Adjust.
Compiling GDB with current git Clang (future 13) fails with (among
other problems), this issue:
$ make nat/amd64-linux-siginfo.o
CXX nat/amd64-linux-siginfo.o
src/gdb/nat/amd64-linux-siginfo.c:590:35: warning: passing 4-byte aligned argument to 8-byte aligned parameter 1 of 'compat_x32_siginfo_from_siginfo' may result in an unaligned pointer access [-Walign-mismatch]
compat_x32_siginfo_from_siginfo ((struct compat_x32_siginfo *) inf,
^
1 warning generated.
The problem is that:
- The flagged code is casting to "struct compat_x32_siginfo" pointer
directly instead of to a pointer to the compat_x32_siginfo_t
typedef. The called function is declared with a
compat_x32_siginfo_t typedef pointer parameter.
- Only the typedef has the __aligned__ attribute.
Fix this by moving the attribute to the struct, so both struct and
typedef have the same alignment.
The next patch removes the typedefs.
gdb/ChangeLog:
yyyy-mm-dd Pedro Alves <pedro@palves.net>
* nat/amd64-linux-siginfo.c (compat_x32_siginfo_t): Move
__attribute__ __aligned__ from the typedef to the struct.
While doing some changes, I managed to break this test in a way that
running to main didn't work. However, it didn't produce any FAIL. I
noticed because I diff'ed the results and saw some PASSes unexpectedly
disappear, but that's a bit fragile. Add a fail in case this test fails
to run to main. Ideally, I think that runto_main should by default
produce a FAIL when it fails (the opposite of the existing logic), but
that's a project of its own, so just fix this test for the moment.
gdb/testsuite/ChangeLog:
* gdb.base/continue-all-already-running.exp: Call fail if can't
run to main.
Change-Id: I84b816a126c92ac579ed5ebbe39b017bd5cb7096
It was spotted that if type_align returned 0 then it was possible to
trigger a divide by zero exception within GDB. It turns out this will
only happen in an edge case where GDB is unable to figure out the
alignment of a field within a structure.
The attached test generates some non-standard, probably broken, DWARF,
that triggers this condition, and then fixes this issue by throwing an
exception when this case occurs.
gdb/ChangeLog:
PR gdb/27847
* amd64-tdep.c (amd64_has_unaligned_fields): Move call to
type_align, and spot case where the alignment is unknown.
gdb/testsuite/ChangeLog:
PR gdb/27847
* gdb.dwarf2/dw2-weird-type-len.c: New file.
* gdb.dwarf2/dw2-weird-type-len.exp: New file.
Use proc_with_prefix for test_command_line_attach_run, as we do for
other procs.
gdb/testsuite/ChangeLog:
* gdb.base/attach.exp (test_command_line_attach_run): Use
proc_with_prefix.
Change-Id: I47b61bb91b6ac0570ad7556a824f874ea50c7cda
The past commit d1e93af64a6b ("gdb: set current thread in
sparc_{fetch,collect}_inferior_registers (PR gdb/27147)") changed
sparc_fetch_inferior_registers and sparc_store_inferior_registers to
look up the thread corresponding to the regcache's ptid and make it the
current thread. The reason being that down the call chain, some
functions (like sparc_supply_rwindow) can do some memory reads or write,
through target_read_memory/target_write_memory, and those rely on the
current global context.
There is one small problem with this approach: when debugging a
multi-threaded program, the regcache for a new thread is created just
before the corresponding thread_info is created. In fact, the regcache
is created somewhere during the call to thread_from_lwp, which is
responsible for creating the thread_info:
#8 0x0000010000ab9968 in internal_error (file=0x10000bfca20 "/home/simark/src/binutils-gdb/gdb/thread.c", line=1346, fmt=0x10000bfc918 "%s: Assertion `%s' failed.") at /home/simark/src/binutils-gdb/gdbsupport/errors.cc:55
#9 0x0000010000827f3c in switch_to_thread (thr=0x0) at /home/simark/src/binutils-gdb/gdb/thread.c:1346
#10 0x0000010000753444 in sparc_fetch_inferior_registers (proc_target=0x10000fa8cb0 <the_sparc64_linux_nat_target>, regcache=0x10000ff03c0, regnum=-1) at /home/simark/src/binutils-gdb/gdb/sparc-nat.c:175
#11 0x000001000075b908 in sparc64_linux_nat_target::fetch_registers (this=0x10000fa8cb0 <the_sparc64_linux_nat_target>, regcache=0x10000ff03c0, regnum=-1) at /home/simark/src/binutils-gdb/gdb/sparc64-linux-nat.c:38
#12 0x00000100007fe6f4 in target_ops::fetch_registers (this=0x10000f7feb0 <the_thread_db_target>, arg0=0x10000ff03c0, arg1=-1) at /home/simark/src/binutils-gdb/gdb/target-delegates.c:496
#13 0x00000100008162a0 in target_fetch_registers (regcache=0x10000ff03c0, regno=-1) at /home/simark/src/binutils-gdb/gdb/target.c:3287
#14 0x000001000060a4bc in ps_lgetregs (ph=0x10001264368, lwpid=458727, gregset=0x7feff97d388) at /home/simark/src/binutils-gdb/gdb/proc-service.c:158
#15 0xffff800103e32420 in __td_ta_lookup_th_unique (ta_arg=0x100012d7080, lwpid=<optimized out>, th=0x7feff97d7c8) at td_ta_map_lwp2thr.c:119
#16 0xffff800103e32604 in td_ta_map_lwp2thr (ta_arg=0x100012d7080, lwpid=<optimized out>, th=0x7feff97d7c8) at td_ta_map_lwp2thr.c:207
#17 0x000001000051fee8 in thread_from_lwp (stopped=0x100011a3650, ptid=...) at /home/simark/src/binutils-gdb/gdb/linux-thread-db.c:415
#18 0x0000010000520150 in thread_db_notice_clone (parent=..., child=...) at /home/simark/src/binutils-gdb/gdb/linux-thread-db.c:446
#19 0x00000100005068a8 in linux_handle_extended_wait (lp=0x10001230700, status=4479) at /home/simark/src/binutils-gdb/gdb/linux-nat.c:1978
#20 0x000001000050a278 in linux_nat_filter_event (lwpid=458724, status=198015) at /home/simark/src/binutils-gdb/gdb/linux-nat.c:2913
#21 0x000001000050b818 in linux_nat_wait_1 (ptid=..., ourstatus=0x7feff97e8d0, target_options=...) at /home/simark/src/binutils-gdb/gdb/linux-nat.c:3194
#22 0x000001000050ca4c in linux_nat_target::wait (this=0x10000fa8cb0 <the_sparc64_linux_nat_target>, ptid=..., ourstatus=0x7feff97e8d0, target_options=...) at /home/simark/src/binutils-gdb/gdb/linux-nat.c:3432
#23 0x00000100005237ec in thread_db_target::wait (this=0x10000f7feb0 <the_thread_db_target>, ptid=..., ourstatus=0x7feff97e8d0, options=...) at /home/simark/src/binutils-gdb/gdb/linux-thread-db.c:1379
#24 0x00000100007fa668 in target_wait (ptid=..., status=0x7feff97e8d0, options=...) at /home/simark/src/binutils-gdb/gdb/target.c:2000
#25 0x00000100004adb0c in do_target_wait_1 (inf=0x10001173170, ptid=..., status=0x7feff97e8d0, options=...) at /home/simark/src/binutils-gdb/gdb/infrun.c:3464
#26 0x00000100004add48 in operator() (__closure=0x7feff97e658, inf=0x10001173170) at /home/simark/src/binutils-gdb/gdb/infrun.c:3527
#27 0x00000100004ae15c in do_target_wait (wait_ptid=..., ecs=0x7feff97e8a8, options=...) at /home/simark/src/binutils-gdb/gdb/infrun.c:3540
#28 0x00000100004af254 in fetch_inferior_event () at /home/simark/src/binutils-gdb/gdb/infrun.c:3880
#29 0x0000010000486ef8 in inferior_event_handler (event_type=INF_REG_EVENT) at /home/simark/src/binutils-gdb/gdb/inf-loop.c:42
The problem is that while sparc_fetch_inferior_registers runs and is
asked to read the registers of a given ptid, there isn't a thread_info
with that ptid yet. So, find_thread_ptid returns nullptr, and
switch_to_thread gives an internal error.
Fix this by only setting inferior_ptid, instead of the whole global
context, as switch_to_thread does. This is sufficient for
target_read_memory / target_write_memory to work down the line.
Ideally, it would be nice to be able to pass the ptid down the whole
call chain and to target_read_memory / target_write_memory, so that this
setting of inferior_ptid would not be necessary. But this is not going
to happen soon.
This fixes running a multi-threaded program, which would hit the
internal error show in the call stack above.
gdb/ChangeLog:
PR gdb/27899
* sparc-nat.c (sparc_fetch_inferior_registers): Set
inferior_ptid instead of using switch_to_thread.
(sparc_store_inferior_registers): Likewise.
Change-Id: I0b6ddb3af9b11f67b10ee46a734fb82ecc6462d5
With a testsuite setup modified to make expect wait a little bit longer for
gdb output (see PR27957), I reliably run into:
...
27 return SYSCALL_CANCEL (nanosleep, requested_time, remaining);^M
(gdb) ^M
Thread 2 "run-attach-whil" stopped.^M
0x00007f13c85a74c0 in __GI___nanosleep () at nanosleep.c:27^M
27 return SYSCALL_CANCEL (nanosleep, requested_time, remaining);^M
FAIL: gdb.base/run-attach-while-running.exp: threaded=1: \
run-or-attach=attach: non-stop=on: test: attach to process (timeout)
...
The problem is that we're trying to match the gdb_prompt using gdb_test which
uses '$gdb_prompt $'. The terminating '$' prevents the match.
Fix this by rewriting the gdb_test into a gdb_test_multiple and dropping the
'$'.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2021-06-07 Tom de Vries <tdevries@suse.de>
PR testsuite/27908
* gdb.base/run-attach-while-running.exp (test): Don't match prompt
after attach using '$gdb_prompt $'.
The regexp matching state machine in gdb.base/info-types.exp.tcl handles
"File .*:" lines individually. Now that we do line-by-line matching, that's
no longer necessary.
Simplify accordingly.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2021-06-07 Tom de Vries <tdevries@suse.de>
* gdb.base/info-types.exp.tcl: Remove "File .*" matching from regexp
matching state machine.
With a testsuite setup modified to make expect wait a little bit longer for
gdb output (see PR27957), I run into:
...
PASS: gdb.base/new-ui-pending-input.exp: spawn gdb
print 1^M
print 2^M
print 3^M
(gdb) $1 = 1^M
(gdb) $2 = 2^M
(gdb) $3 = 3^M
(gdb) PASS: gdb.base/new-ui-pending-input.exp: initial prompt on extra console
^M
Breakpoint 1, main () at new-ui-pending-input.c:25^M
25 return 0; /* set breakpoint here */^M
FAIL: gdb.base/new-ui-pending-input.exp: print 1 on extra console (timeout)
...
Usually, I get a pass instead. In the passing case, the "initial prompt"
PASS is after the first prompt:
...
PASS: gdb.base/new-ui-pending-input.exp: spawn gdb
print 1^M
print 2^M
print 3^M
(gdb) PASS: gdb.base/new-ui-pending-input.exp: initial prompt on extra console
...
while in the failing case, that PASS is after the fourth prompt.
The regexp doesn't match the first prompt because it terminates with a '$':
...
-re "$gdb_prompt $" {
...
Fix this by removing the terminating '$' and prefixing the regex with
something unique to the first prompt: "print 3".
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2021-06-07 Tom de Vries <tdevries@suse.de>
* gdb.base/new-ui-pending-input.exp
(test_command_line_new_ui_pending_input): Fix regexp for "initial
prompt on extra console".
Recently, I've rewritten gdb.base/info-types.exp.tcl to do processing
line-by-line (commit 2129a942553).
When building gdb with -O0, this test passes for me, but with -O2 I sometimes
run into:
...
FAIL: gdb.base/info-types-c.exp: info types (state == 1)
...
The output the failing gdb_test_multiple is trying to parse is:
...
(gdb) info types^M
All defined types:^M
^M
File src/gdb/testsuite/gdb.base/info-types.c:^M
52: typedef enum {...} anon_enum_t;^M
...
(gdb)
...
and the FAIL with state == 1 means that:
- the regexp for "All defined types:" did trigger, and that
- the regexp for "File .*info-types.c" didn't trigger.
This is due to the fact that the empty line inbetween is supposed to be
matched by the "random line" regexp "^\r\n(\[^\r\n\]*)(?=\r\n)", which doesn't
happen because instead the earlier regexp consuming the gdb prompt matches.
Fix this by moving the "random line" regexp up. [ Note that that regexp does
not consume the gdb prompt, because it requires (but doesn't consume) a
terminating "\r\n".
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2021-06-07 Tom de Vries <tdevries@suse.de>
* gdb.base/info-types.exp.tcl (run_test): Move "random line" regexp
up.
When AVX512 support was added, symbol quotation was not paid attention
to. Just like the (base,index,scale) specifier gets parsed from the end
of the expression, the {...} also wants parsing from the end; in no case
is the first { found a guarantee of a masking or broadcasting specifier.
When d02603dc201f ("Allow symbol and label names to be enclosed in
double quotes") added the check for a double quote to the loop body
there, it didn't go quite far enough: Parentheses inside quotes
shouldn't be counted, and character restrictions also shouldn't apply
inside quoted regions.
In i386_att_operand(), which needs adjustment to remain in sync, besides
respecting double quotes now, also change the logic such that we don't
count parentheses anymore: Finding any opening or closing parenthesis or
any double quote means we're done, because the subsequent parsing code
wouldn't accept (extra) instances of these anyway.
Note that in parse_operands() this mimics get_symbol_name()'s
questionable behavior of treating \ specially only when ahead of ". (The
behavior is suspicious because the meaning of \\ then is ambiguous. It
is in particular impossible to have a (quoted) symbol name end in a
single \.) I would have used get_symbol_name() here, if that didn't
require fiddling with input_line_pointer.
So far only - was permitted, but +, !, and ~ ought to be treated the
same.
Rather than adding them to digit_chars[], which was at least odd to have
held - so far, drop this array and its wrapper macro for being used just
once.
While adjusting this logic, also include [ in the characters which may
start a displacement expression - gas generally treats [] as equivalent
to ().
When d02603dc201f ("Allow symbol and label names to be enclosed in
double quotes") added the check for a leading double quote to
i386_att_operand(), it missed a second similar check after having found
a segment override. To avoid the two checks going out of sync again,
introduce an inline helper.
This needs to happen before checking of what may legitimately start a
memory operand (like is done when there's no segment override). Plus a
second '*' shouldn't be permitted when one was already found before the
segment override.
While subsequent processing in AT&T mode relies on this simplistic early
checking, Intel mode hasn't been for quite a long time (or perhaps never
really did).
Neither masking nor broadcast are possible here, and RC/SAE get dealt
with elsewhere.
This also fixes gas crashes (i.e. "Fatal error: unable to continue with
assembly"), since the return path being removed failed to restore
input_line_pointer from save_input_line_pointer.
Non-64-bit code should get handled the same with or without BFD64. This
wasn't the case though in a number of situations (and quite likely there
are more that I haven't spotted yet).
It's not very nice to tie the check in md_apply_fix() to object_64bit,
but afaict at that time we have no record anymore of the mode an insn
was assembled in (it might also have been data). This doesn't look to be
the first inconsistency of this kind, though. In x86_cons() it's even
less clear what the right approach would be: flag_code shouldn't matter
for data emission, but instead we'd need to know from which mode(s) the
data actually gets accessed. On this basis, signed_cons() also gets
adjusted.
Older gcc reports:
.../bfd/dwarf2.c: In function 'read_ranges':
.../bfd/dwarf2.c:3107: error: comparison between signed and unsigned
.../bfd/dwarf2.c: In function 'read_rnglists':
.../bfd/dwarf2.c:3189: error: comparison between signed and unsigned
Similarly for binutils/dwarf.c. Arrange for the left sides of the > to
also be unsigned quantities.
I get some random timeouts in this test due to big debug info taking a
lot of time to read through gdbserver. When host and target are on the
same machine, clear the sysroot parameter so that GDB reads the files
from the local file system, as we already do in many tests.
I agree with what Pedro says here:
https://sourceware.org/pipermail/gdb-patches/2019-March/156568.html
that if this is bad for us, it's also bad for users, so we should be
fixing the slowness instead. But so far nobody seems to be working on
it, and the testsuite timeouts are getting in the way, so I think this
"set sysroot" is a net positive for now.
Without this patch, the test takes over 2 minutes to run (most of it
"downloading" libc debug info), with it it takes 10 seconds.
gdb/testsuite/ChangeLog:
* gdb.server/stop-reply-no-thread-multi.exp: Clear sysroot when
host and target are local.
Change-Id: Ieb6304f0e56b4575af450913de4210c667c6bf7b
Due to the SIGPIPE the gdb process is killed here, which is
not helpful.
2021-06-05 Bernd Edlinger <bernd.edlinger@hotmail.de>
* compile/compile.c (scoped_ignore_sigpipe): New helper class.
(compile_to_object): Ignore SIGPIPE before calling the plugin.
Oops, I botched the last patch to fix data-directory/Makefile
rebuilding, by copying the config.status line and forgetting to update
the directory name. This one fixes the problem for real.
gdb/ChangeLog
2021-06-05 Tom Tromey <tom@tromey.com>
* data-directory/Makefile.in (Makefile): Use correct directory
name.
My recent changes to the gdb build removed a special case for
data-directory/Makefile, but neglected to update the rule in that
Makefile. This patch fixes the error by rewriting this rule.
gdb/ChangeLog
2021-06-05 Tom Tromey <tom@tromey.com>
* data-directory/Makefile.in (Makefile): Rewrite.
Shahab Vahedi pointed out that the patch to remove
gdb/testsuite/configure regressed the site.exp creation a bit -- it
left an unresolved configure substitution. Andrew Burgess pointed out
that the patch removed the call to ACX_NONCANONICAL_TARGET, which
caused this problem.
This patch adds ACX_NONCANONICAL_TARGET to gdb's configure, and fixes
the bug.
gdb/ChangeLog
2021-06-05 Tom Tromey <tromey@adacore.com>
* configure: Rebuild.
* configure.ac: Add ACX_NONCANONICAL_TARGET.
This provides a space to generate things that we only need to build
once per-arch. Some day that will be all of common/, but for now,
we move the version.c management in.
Implement ARC target support for passing options to the disassembler
through the command interface. e.g.:
gdb> set disassembler-options cpu=hs38_linux ...
gdb/ChangeLog:
* NEWS: Document 'set disassembler-options' support for the ARC
target.
* arc-tdep.c (arc_gdbarch_init): Set
'gdbarch_valid_disassembler_options'.
gdb/doc/ChangeLog:
* gdb.texinfo (Source and Machine Code): Document 'set
disassembler-options' support for the ARC target.
gdb/testsuite/ChangeLog:
* gdb.arch/arc-disassembler-options.exp: New test.
* gdb.arch/arc-disassembler-options.s: New test source.
With -fgnat-encodings=minimal, an internal version (these patches will
be upstreamed in the near future) of the Ada compiler can emit DWARF
for an array where the bound comes from a variable, like:
<1><12a7>: Abbrev Number: 7 (DW_TAG_array_type)
<12a8> DW_AT_name : (indirect string, offset: 0x1ae9): pck__my_array
[...]
<2><12b4>: Abbrev Number: 8 (DW_TAG_subrange_type)
<12b5> DW_AT_type : <0x1294>
<12b9> DW_AT_upper_bound : <0x1277>
With the upper bound DIE being:
<1><1277>: Abbrev Number: 2 (DW_TAG_variable)
<1278> DW_AT_name : (indirect string, offset: 0x1a4d): pck__my_length___U
<127c> DW_AT_type : <0x128f>
<1280> DW_AT_external : 1
<1280> DW_AT_artificial : 1
<1280> DW_AT_declaration : 1
Note that the variable is just a declaration -- in this situation, the
variable comes from another compilation unit, and must be found when
trying to compute the array bound.
This patch adds a new PROP_VARIABLE_NAME kind, to enable this search.
This same scenario can occur with DW_OP_GNU_variable_value, so this
patch adds support for that as well.
gdb/ChangeLog
2021-06-04 Tom Tromey <tromey@adacore.com>
* dwarf2/read.h (dwarf2_fetch_die_type_sect_off): Add 'var_name'
parameter.
* dwarf2/loc.c (dwarf2_evaluate_property) <case
PROP_VARIABLE_NAME>: New case.
(compute_var_value): New function.
(sect_variable_value): Use compute_var_value.
* dwarf2/read.c (attr_to_dynamic_prop): Handle DW_TAG_variable.
(var_decl_name): New function.
(dwarf2_fetch_die_type_sect_off): Add 'var_name' parameter.
* gdbtypes.h (enum dynamic_prop_kind) <PROP_VARIABLE_NAME>: New
constant.
(union dynamic_prop_data) <variable_name>: New member.
(struct dynamic_prop) <variable_name, set_variable_name>: New
methods.
gdb/testsuite/ChangeLog
2021-06-04 Tom Tromey <tromey@adacore.com>
* gdb.ada/array_of_symbolic_length.exp: New file.
* gdb.ada/array_of_symbolic_length/foo.adb: New file.
* gdb.ada/array_of_symbolic_length/gl.adb: New file.
* gdb.ada/array_of_symbolic_length/gl.ads: New file.
* gdb.ada/array_of_symbolic_length/pck.adb: New file.
* gdb.ada/array_of_symbolic_length/pck.ads: New file.
I needed more debug output from:
remote_target::select_thread_for_ambiguous_stop_reply
I thought this would be useful for others too.
gdb/ChangeLog:
* remote.c (remote_target)
<select_thread_for_ambiguous_stop_reply>: Add additional debug
output.
If the TUI window object implements the click method, it is called for each
mouse click event in this window.
gdb/ChangeLog:
2021-06-04 Hannes Domani <ssbssa@yahoo.de>
* python/py-tui.c (class tui_py_window): Add click function.
(tui_py_window::click): Likewise.
gdb/doc/ChangeLog:
2021-06-04 Hannes Domani <ssbssa@yahoo.de>
* python.texi (TUI Windows In Python): Document Window.click.
Implements an overridable tui_win_info::click method whose arguments
are the mouse coordinates inside the specific window, and the mouse
button clicked.
And if the curses implementation supports 5 buttons, the 4th and 5th
buttons are used for scrolling.
gdb/ChangeLog:
2021-06-04 Hannes Domani <ssbssa@yahoo.de>
* ser-mingw.c (console_select_thread): Handle MOUSE_EVENT.
* tui/tui-data.h (struct tui_win_info): Add click function.
* tui/tui-io.c (tui_prep_terminal): Enable mouse events.
(tui_deprep_terminal): Disable mouse events.
(tui_dispatch_ctrl_char): Handle KEY_MOUSE.
* tui/tui.c (tui_disable): Disable mouse events.
Evaluating expressions from within an inferior exit event handler can
cause a crash:
echo "int main() { return 0; }" > repro.c
gcc -g repro.c -o repro
./gdb -q --ex "set language c++" --ex "python gdb.events.exited.connect(lambda _: gdb.execute('set \$_a=0'))" --ex "run" repro
Reading symbols from repro...
Starting program: /home/mhov/repos/binutils-gdb-master/install-bad/bin/repro
[Inferior 1 (process 1974779) exited normally]
../../gdb/thread.c:72: internal-error: thread_info* inferior_thread(): Assertion `current_thread_ != nullptr' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n) [answered Y; input not from terminal]
This is a bug, please report it. For instructions, see:
<https://www.gnu.org/software/gdb/bugs/>.
Backtrace
0 in internal_error of ../../gdbsupport/errors.cc:51
1 in inferior_thread of ../../gdb/thread.c:72
2 in expression::evaluate of ../../gdb/eval.c:98
3 in evaluate_expression of ../../gdb/eval.c:115
4 in set_command of ../../gdb/printcmd.c:1502
5 in do_const_cfunc of ../../gdb/cli/cli-decode.c:101
6 in cmd_func of ../../gdb/cli/cli-decode.c:2181
7 in execute_command of ../../gdb/top.c:670
...
22 in python_inferior_exit of ../../gdb/python/py-inferior.c:182
In `expression::evaluate (...)' there is a call to `inferior_thread
()' that is guarded by `target_has_execution ()':
struct value *
expression::evaluate (struct type *expect_type, enum noside noside)
{
gdb::optional<enable_thread_stack_temporaries> stack_temporaries;
if (target_has_execution ()
&& language_defn->la_language == language_cplus
&& !thread_stack_temporaries_enabled_p (inferior_thread ()))
stack_temporaries.emplace (inferior_thread ());
The `target_has_execution ()' guard maps onto `inf->pid' and the
`inferior_thread ()' call assumes that `current_thread_' is set to
something meaningful:
struct thread_info*
inferior_thread (void)
{
gdb_assert (current_thread_ != nullptr);
return current_thread_;
}
In other words, it is assumed that if `inf->pid' is set then
`current_thread_' must also be set. This does not hold at the point
where inferior exit observers are notified:
- `generic_mourn_inferior (...)'
- `switch_to_no_thread ()'
- `current_thread_ = nullptr;'
- `exit_inferior (...)'
- `gdb::observers::inferior_exit.notify (...)'
- `inf->pid = 0'
The inferior exit notification means that a Python handler can get a
chance to run while `current_thread' has been cleared and the
`inf->pid' has not been cleared. Since the Python handler can call any
GDB command with `gdb.execute(...)' (in my case `gdb.execute("set
$_a=0")' we can end up evaluating expressions and asserting in
`evaluate_subexp (...)'.
This patch adds a test in `evaluate_subexp (...)' to check the global
`inferior_ptid' which is reset at the same time as `current_thread_'.
Checking `inferior_ptid' at the same time as `target_has_execution ()'
seems to be a common pattern:
$ git grep -n -e inferior_ptid --and -e target_has_execution
gdb/breakpoint.c:2998: && (inferior_ptid == null_ptid || !target_has_execution ()))
gdb/breakpoint.c:3054: && (inferior_ptid == null_ptid || !target_has_execution ()))
gdb/breakpoint.c:4587: if (inferior_ptid == null_ptid || !target_has_execution ())
gdb/infcmd.c:360: if (inferior_ptid != null_ptid && target_has_execution ())
gdb/infcmd.c:2380: /* FIXME: This should not really be inferior_ptid (or target_has_execution).
gdb/infrun.c:3438: if (!target_has_execution () || inferior_ptid == null_ptid)
gdb/remote.c:11961: if (!target_has_execution () || inferior_ptid == null_ptid)
gdb/solib.c:725: if (target_has_execution () && inferior_ptid != null_ptid)
The testsuite has been run on 5.4.0-59-generic x86_64 GNU/Linux:
- Ubuntu 20.04.1 LTS
- gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0
- DejaGnu version 1.6.2
- Expect version 5.45.4
- Tcl version 8.6
- Native configuration: x86_64-pc-linux-gnu
- Target: unix
Results show a few XFAIL in
gdb.threads/attach-many-short-lived-threads.exp. The existing
py-events.exp tests are skipped for native-gdbserver and fail for
native-extended-gdbserver, but the new tests pass with
native-extended-gdbserver when run without the existing tests.
gdb/ChangeLog:
2021-06-03 Magne Hov <mhov@undo.io>
PR python/27841
* eval.c (expression::evaluate): Check inferior_ptid.
gdb/testsuite/ChangeLog:
2021-06-03 Magne Hov <mhov@undo.io>
PR python/27841
* gdb.python/py-events.exp: Extend inferior exit tests.
* gdb.python/py-events.py: Print inferior exit PID.
gdb/ChangeLog:
yyyy-mm-dd Pedro Alves <pedro@palves.net>
* MAINTAINERS (The Official FSF-appointed GDB Maintainers): Remove
affiliation.
(Global Maintainers): Update my address.
(Write After Approval): Remove stale entry.
Change-Id: I3266fedeebfa6800faa2217baf6c032408e84902
Provide a description for si_code values as a sigcode-meaning field.
For signals raised by a system call, provide the pid and user ID of
the sending process. For signals raised by a POSIX timer exparation,
provide the id of the timer. For signals raised by a POSIX message
queue, provide the id of the message queue. For SIGCHLD provide the
pid and user ID of the child process along with the exit status or
relevant signal number.
Sample output for SIGUSR1 raised by kill():
before:
Program received signal SIGUSR1, User defined signal 1.
kill () at kill.S:4
4 RSYSCALL(kill)
after:
Program received signal SIGUSR1, User defined signal 1.
Sent by kill() from pid 30529 and user 1001.
kill () at kill.S:4
4 RSYSCALL(kill)
SIGCHLD for exited process:
before:
Program received signal SIGCHLD, Child status changed.
after:
Program received signal SIGCHLD, Child status changed.
Child has exited: pid 31929, uid 1001, exit status 0.
SIGALRM raised by a POSIX timer (timer_create):
before:
Program received signal SIGALRM, Alarm clock.
after:
Program received signal SIGALRM, Alarm clock.
Timer expired: timerid 3.
gdb/ChangeLog:
* fbsd-tdep.c (FBSD_SI_USER, FBSD_SI_QUEUE, FBSD_SI_TIMER)
(FBSD_SI_ASYNCIO, FBSD_SI_MESGQ, FBSD_SI_KERNEL, FBSD_SI_LWP)
(FBSD_ILL_ILLOPC, FBSD_ILL_ILLOPN, FBSD_ILL_ILLADR)
(FBSD_ILL_ILLTRP, FBSD_ILL_PRVOPC, FBSD_ILL_PRVREG)
(FBSD_ILL_COPROC, FBSD_ILL_BADSTK, FBSD_BUS_ADRALN)
(FBSD_BUS_ADRERR, FBSD_BUS_OBJERR, FBSD_BUS_OOMERR)
(FBSD_SEGV_MAPERR, FBSD_SEGV_ACCERR, FBSD_SEGV_PKUERR)
(FBSD_FPE_INTOVF, FBSD_FPE_INTDIV, FBSD_FPE_FLTDIV)
(FBSD_FPE_FLTOVF, FBSD_FPE_FLTUND, FBSD_FPE_FLTRES)
(FBSD_FPE_FLTINV, FBSD_FPE_FLTSUB, FBSD_TRAP_BRKPT)
(FBSD_TRAP_TRACE, FBSD_TRAP_DTRACE, FBSD_TRAP_CAP)
(FBSD_CLD_EXITED, FBSD_CLD_KILLED, FBSD_CLD_DUMPED)
(FBSD_CLD_TRAPPED, FBSD_CLD_STOPPED, FBSD_CLD_CONTINUED)
(FBSD_POLL_IN, FBSD_POLL_OUT, FBSD_POLL_MSG, FBSD_POLL_ERR)
(FBSD_POLL_PRI, FBSD_POLL_HUP, fbsd_signal_cause)
(fbsd_report_signal_info): New.
(fbsd_init_abi): Use fbsd_report_signal_info as gdbarch
report_signal_info method.
