According to the DWARF3 standard, a function always has a name attribute
(Section 3.3 - Subroutine and Entry Point Entries). The only exception
is when a DW_AT_abstract_origin attribute is provided, in which case
the name may be inherited from the referenced DIE.
The problem occured because our compiler generated a subprogram DIE
for a nested function where the name attribute was missing (and no
abstract-origin either). Our code in add_partial_symbol is not
prepared to deal with the situation, and happily just tries to compute
the length of the (NULL) function name.
This normally cannot happen, because there is already a guard in
scan_partial_symbols, where we (silently!) ignore anonymous dies,
including anonymous subprograms. Unfortunately, there is a flaw that
affects Ada and other languages that allow nested subprograms. For
nested subprograms, we do not go through scan_partial_symbols and
thus we are missing the name check.
This patch adds the name check in the nested subprogram case. It also
adds a complaint which is emitted during the psymtab->symtab conversion
phase.
gdb/ChangeLog:
* dwarf2read.c (add_partial_subprogram): Make sure the subprogram
DIE has a name before creating the associated partial symbol.
(read_func_scope): Emit a complaint if the subprogram does not
have a name or when we can't extract the subprogram PC bounds.
gdb/testsuite/ChangeLog:
* gdb.dwarf2/dw2-anonymous-func.S: New file.
* gdb.dwarf2/dw2-anonymous-func.exp: New testcase.
Tested on x86_64-linux, no regression. Note that the testcase also
verifies that the psymtab->symtab conversion does not crash (this is
the purpose of the "list file1.txt:1" test.
* infcmd.c (until_command): Use ERROR_NO_INFERIOR. Ensure there's
a valid selected thread, and that it is not running.
(advance_command): Ditto.
(finish_command): Ditto.
gdb/testsuite/
* gdb.base/default.exp: Adjust.
2010-03-18 H.J. Lu <hongjiu.lu@intel.com>
PR binutils/11396
* libcoff-in.h (pe_tdata): Add dont_strip_reloc.
* libcoff.h: Regenerated.
* peXXigen.c (_bfd_XXi_only_swap_filehdr_out): Clear F_RELFLG
if dont_strip_reloc is set.
(_bfd_XX_bfd_copy_private_bfd_data_common): Set
dont_strip_reloc on output if there is no .reloc and
IMAGE_FILE_RELOCS_STRIPPED isn't set in input.
* ax-gdb.h (struct axs_value): New field optimized_out.
(gen_trace_for_var): Add gdbarch argument.
* ax-gdb.c (gen_trace_static_fields): New function.
(gen_traced_pop): Call it, add gdbarch argument.
(gen_trace_for_expr): Update call to it.
(gen_trace_for_var): Ditto, and report optimized-out variables.
(gen_struct_ref_recursive): Check for optimized-out value.
(gen_struct_elt_for_reference): Ditto.
(gen_static_field): Pass gdbarch instead of expression, assume
optimization if field not found.
(gen_var_ref): Set the optimized_out flag.
(gen_expr): Error on optimized-out variable.
* tracepoint.c (collect_symbol): Handle struct-valued vars as
expressions, skip optimized-out variables with computed locations.
* dwarf2loc.c (dwarf2_tracepoint_var_ref): Flag instead of
erroring out if location expression missing.
(loclist_tracepoint_var_ref): Don't error out here.
The common code sets up an autoconf option --enable-sim-trace which adds
-DTRACE= to CPPFLAGS. This causes warnings in the building of some files
that declare a local TRACE() helper macro. So punt it from hw-ports.c
(since it isn't actually used) and convert hw-properties.c to HW_TRACE().
This is to prevent an internal error during an Ada task switch. A task
switch is simply a thread switch under the hood. What we do is collect
the info from the Ada Task Control Block, deduce the associated thread
ptid, and then switch to that thread. If the thread ptid computation
routine has not been implemented for the target, of if there is a bug,
then we end up computing a bogus ptid which GDB does not know about,
which eventually leads to an assertion failure:
(gdb) task 1
[New Thread 5715]
/[...]/gdb/thread.c:595: internal-error: is_thread_state:
Assertion `tp' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n)
When this happens, it's just nicer for the user to print an error
message, and cancel the task switch. After this patch is applied,
this is what we get:
(gdb) task 1
[New Thread 10250]
Unable to compute thread ID for task 1.
Cannot switch to this task.
gdb/ChangeLog:
* ada-tasks.c (task_command_1): Check that the task ptid is valid
before doing the associated thread switch.
* stringpool.cc (Stringpool_template::Stringpool_template): Initialize
offset_.
(Stringpool_template::new_key_offset): New method.
(Stringpool_template::add_string): Assign offsets when adding new
strings.
(Stringpool_template::set_string_offsets): Do not set string offsets
when not optimizing.
* stringpool.h (Chunked_vector::Chunked_vector): Initialize data
member size_.
(Chunked_vector::clear): Clear size_.
(Chunked_vector::reserve): Call reserve method of all Element_vectors.
(Chunked_vector::size): Return size_.
(Chunked_vector::push_back): Use size_ to find insert position.
(Chunked_vector::size_): New data member.
(Stringpool_template::set_no_zero_null): Assert string set is empty.
(Stringpool_template::new_key_offset): New method declaration.
(Stringpool_template::offset_): New data member.
* ax-gdb.c: Include cp-support.h.
(find_field): Remove.
(gen_primitive_field): New function.
(gen_struct_ref_recursive): New function.
(gen_struct_ref): Rewrite to call gen_struct_ref_recursive instead
of find_field.
(gen_static_field): New function.
(gen_struct_elt_for_reference): New.
(gen_namespace_elt): New.
(gen_maybe_namespace_elt): New.
(gen_aggregate_elt_ref): New.
(gen_expr): Add OP_SCOPE, display opcode name in error message.
